1 ############################################################
3 # perltidy - a perl script indenter and formatter
5 # Copyright (c) 2000-2007 by Steve Hancock
6 # Distributed under the GPL license agreement; see file COPYING
8 # This program is free software; you can redistribute it and/or modify
9 # it under the terms of the GNU General Public License as published by
10 # the Free Software Foundation; either version 2 of the License, or
11 # (at your option) any later version.
13 # This program is distributed in the hope that it will be useful,
14 # but WITHOUT ANY WARRANTY; without even the implied warranty of
15 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 # GNU General Public License for more details.
18 # You should have received a copy of the GNU General Public License
19 # along with this program; if not, write to the Free Software
20 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
22 # For brief instructions instructions, try 'perltidy -h'.
23 # For more complete documentation, try 'man perltidy'
24 # or visit http://perltidy.sourceforge.net
26 # This script is an example of the default style. It was formatted with:
31 # Michael Cartmell supplied code for adaptation to VMS and helped with
33 # Hugh S. Myers supplied sub streamhandle and the supporting code to
34 # create a Perl::Tidy module which can operate on strings, arrays, etc.
35 # Yves Orton supplied coding to help detect Windows versions.
36 # Axel Rose supplied a patch for MacPerl.
37 # Sebastien Aperghis-Tramoni supplied a patch for the defined or operator.
38 # Dan Tyrell contributed a patch for binary I/O.
39 # Ueli Hugenschmidt contributed a patch for -fpsc
40 # Many others have supplied key ideas, suggestions, and bug reports;
41 # see the CHANGES file.
43 ############################################################
46 use 5.004; # need IO::File from 5.004 or later
47 BEGIN { $^W = 1; } # turn on warnings
61 @ISA = qw( Exporter );
62 @EXPORT = qw( &perltidy );
68 ( $VERSION = q($Id: Tidy.pm,v 1.73 2007/12/05 17:51:17 perltidy Exp $) ) =~ s/^.*\s+(\d+)\/(\d+)\/(\d+).*$/$1$2$3/; # all one line for MakeMaker
73 # given filename and mode (r or w), create an object which:
74 # has a 'getline' method if mode='r', and
75 # has a 'print' method if mode='w'.
76 # The objects also need a 'close' method.
78 # How the object is made:
80 # if $filename is: Make object using:
81 # ---------------- -----------------
82 # '-' (STDIN if mode = 'r', STDOUT if mode='w')
84 # ARRAY ref Perl::Tidy::IOScalarArray (formerly IO::ScalarArray)
85 # STRING ref Perl::Tidy::IOScalar (formerly IO::Scalar)
87 # (check for 'print' method for 'w' mode)
88 # (check for 'getline' method for 'r' mode)
89 my $ref = ref( my $filename = shift );
96 if ( $ref eq 'ARRAY' ) {
97 $New = sub { Perl::Tidy::IOScalarArray->new(@_) };
99 elsif ( $ref eq 'SCALAR' ) {
100 $New = sub { Perl::Tidy::IOScalar->new(@_) };
104 # Accept an object with a getline method for reading. Note:
105 # IO::File is built-in and does not respond to the defined
106 # operator. If this causes trouble, the check can be
107 # skipped and we can just let it crash if there is no
109 if ( $mode =~ /[rR]/ ) {
110 if ( $ref eq 'IO::File' || defined &{ $ref . "::getline" } ) {
111 $New = sub { $filename };
114 $New = sub { undef };
116 ------------------------------------------------------------------------
117 No 'getline' method is defined for object of class $ref
118 Please check your call to Perl::Tidy::perltidy. Trace follows.
119 ------------------------------------------------------------------------
124 # Accept an object with a print method for writing.
125 # See note above about IO::File
126 if ( $mode =~ /[wW]/ ) {
127 if ( $ref eq 'IO::File' || defined &{ $ref . "::print" } ) {
128 $New = sub { $filename };
131 $New = sub { undef };
133 ------------------------------------------------------------------------
134 No 'print' method is defined for object of class $ref
135 Please check your call to Perl::Tidy::perltidy. Trace follows.
136 ------------------------------------------------------------------------
145 if ( $filename eq '-' ) {
146 $New = sub { $mode eq 'w' ? *STDOUT : *STDIN }
149 $New = sub { IO::File->new(@_) };
152 $fh = $New->( $filename, $mode )
153 or warn "Couldn't open file:$filename in mode:$mode : $!\n";
154 return $fh, ( $ref or $filename );
157 sub find_input_line_ending {
159 # Peek at a file and return first line ending character.
160 # Quietly return undef in case of any trouble.
161 my ($input_file) = @_;
164 # silently ignore input from object or stdin
165 if ( ref($input_file) || $input_file eq '-' ) {
168 open( INFILE, $input_file ) || return $ending;
172 read( INFILE, $buf, 1024 );
174 if ( $buf && $buf =~ /([\012\015]+)/ ) {
178 if ( $test =~ /^(\015\012)+$/ ) { $ending = "\015\012" }
181 elsif ( $test =~ /^\015+$/ ) { $ending = "\015" }
184 elsif ( $test =~ /^\012+$/ ) { $ending = "\012" }
198 # concatenate a path and file basename
199 # returns undef in case of error
201 BEGIN { eval "require File::Spec"; $missing_file_spec = $@; }
203 # use File::Spec if we can
204 unless ($missing_file_spec) {
205 return File::Spec->catfile(@_);
208 # Perl 5.004 systems may not have File::Spec so we'll make
209 # a simple try. We assume File::Basename is available.
210 # return undef if not successful.
212 my $path = join '/', @_;
213 my $test_file = $path . $name;
214 my ( $test_name, $test_path ) = fileparse($test_file);
215 return $test_file if ( $test_name eq $name );
216 return undef if ( $^O eq 'VMS' );
218 # this should work at least for Windows and Unix:
219 $test_file = $path . '/' . $name;
220 ( $test_name, $test_path ) = fileparse($test_file);
221 return $test_file if ( $test_name eq $name );
225 sub make_temporary_filename {
227 # Make a temporary filename.
229 # The POSIX tmpnam() function tends to be unreliable for non-unix
230 # systems (at least for the win32 systems that I've tested), so use
231 # a pre-defined name. A slight disadvantage of this is that two
232 # perltidy runs in the same working directory may conflict.
233 # However, the chance of that is small and managable by the user.
234 # An alternative would be to check for the file's existance and use,
235 # say .TMP0, .TMP1, etc, but that scheme has its own problems. So,
237 my $name = "perltidy.TMP";
238 if ( $^O =~ /win32|dos/i || $^O eq 'VMS' || $^O eq 'MacOs' ) {
241 eval "use POSIX qw(tmpnam)";
242 if ($@) { return $name }
245 # just make a couple of tries before giving up and using the default
247 my $tmpname = tmpnam();
248 my $fh = IO::File->new( $tmpname, O_RDWR | O_CREAT | O_EXCL );
258 # Here is a map of the flow of data from the input source to the output
261 # LineSource-->Tokenizer-->Formatter-->VerticalAligner-->FileWriter-->
262 # input groups output
263 # lines tokens lines of lines lines
266 # The names correspond to the package names responsible for the unit processes.
268 # The overall process is controlled by the "main" package.
270 # LineSource is the stream of input lines
272 # Tokenizer analyzes a line and breaks it into tokens, peeking ahead
273 # if necessary. A token is any section of the input line which should be
274 # manipulated as a single entity during formatting. For example, a single
275 # ',' character is a token, and so is an entire side comment. It handles
276 # the complexities of Perl syntax, such as distinguishing between '<<' as
277 # a shift operator and as a here-document, or distinguishing between '/'
278 # as a divide symbol and as a pattern delimiter.
280 # Formatter inserts and deletes whitespace between tokens, and breaks
281 # sequences of tokens at appropriate points as output lines. It bases its
282 # decisions on the default rules as modified by any command-line options.
284 # VerticalAligner collects groups of lines together and tries to line up
285 # certain tokens, such as '=>', '#', and '=' by adding whitespace.
287 # FileWriter simply writes lines to the output stream.
289 # The Logger package, not shown, records significant events and warning
290 # messages. It writes a .LOG file, which may be saved with a
291 # '-log' or a '-g' flag.
295 # variables needed by interrupt handler:
299 # this routine may be called to give a status report if interrupted. If a
300 # parameter is given, it will call exit with that parameter. This is no
301 # longer used because it works under Unix but not under Windows.
302 sub interrupt_handler {
304 my $exit_flag = shift;
305 print STDERR "perltidy interrupted";
307 my $input_line_number =
308 Perl::Tidy::Tokenizer::get_input_line_number();
309 print STDERR " at line $input_line_number";
313 if ( ref $input_file ) { print STDERR " of reference to:" }
314 else { print STDERR " of file:" }
315 print STDERR " $input_file";
318 exit $exit_flag if defined($exit_flag);
325 destination => undef,
332 dump_options => undef,
333 dump_options_type => undef,
334 dump_getopt_flags => undef,
335 dump_options_category => undef,
336 dump_options_range => undef,
337 dump_abbreviations => undef,
340 # don't overwrite callers ARGV
345 if ( my @bad_keys = grep { !exists $defaults{$_} } keys %input_hash ) {
347 my @good_keys = sort keys %defaults;
348 @bad_keys = sort @bad_keys;
350 ------------------------------------------------------------------------
351 Unknown perltidy parameter : (@bad_keys)
352 perltidy only understands : (@good_keys)
353 ------------------------------------------------------------------------
358 my $get_hash_ref = sub {
360 my $hash_ref = $input_hash{$key};
361 if ( defined($hash_ref) ) {
362 unless ( ref($hash_ref) eq 'HASH' ) {
363 my $what = ref($hash_ref);
365 $what ? "but is ref to $what" : "but is not a reference";
367 ------------------------------------------------------------------------
368 error in call to perltidy:
369 -$key must be reference to HASH $but_is
370 ------------------------------------------------------------------------
377 %input_hash = ( %defaults, %input_hash );
378 my $argv = $input_hash{'argv'};
379 my $destination_stream = $input_hash{'destination'};
380 my $errorfile_stream = $input_hash{'errorfile'};
381 my $logfile_stream = $input_hash{'logfile'};
382 my $perltidyrc_stream = $input_hash{'perltidyrc'};
383 my $source_stream = $input_hash{'source'};
384 my $stderr_stream = $input_hash{'stderr'};
385 my $user_formatter = $input_hash{'formatter'};
387 # various dump parameters
388 my $dump_options_type = $input_hash{'dump_options_type'};
389 my $dump_options = $get_hash_ref->('dump_options');
390 my $dump_getopt_flags = $get_hash_ref->('dump_getopt_flags');
391 my $dump_options_category = $get_hash_ref->('dump_options_category');
392 my $dump_abbreviations = $get_hash_ref->('dump_abbreviations');
393 my $dump_options_range = $get_hash_ref->('dump_options_range');
395 # validate dump_options_type
396 if ( defined($dump_options) ) {
397 unless ( defined($dump_options_type) ) {
398 $dump_options_type = 'perltidyrc';
400 unless ( $dump_options_type =~ /^(perltidyrc|full)$/ ) {
402 ------------------------------------------------------------------------
403 Please check value of -dump_options_type in call to perltidy;
404 saw: '$dump_options_type'
405 expecting: 'perltidyrc' or 'full'
406 ------------------------------------------------------------------------
412 $dump_options_type = "";
415 if ($user_formatter) {
417 # if the user defines a formatter, there is no output stream,
418 # but we need a null stream to keep coding simple
419 $destination_stream = Perl::Tidy::DevNull->new();
422 # see if ARGV is overridden
423 if ( defined($argv) ) {
425 my $rargv = ref $argv;
426 if ( $rargv eq 'SCALAR' ) { $argv = $$argv; $rargv = undef }
430 if ( $rargv eq 'ARRAY' ) {
435 ------------------------------------------------------------------------
436 Please check value of -argv in call to perltidy;
437 it must be a string or ref to ARRAY but is: $rargv
438 ------------------------------------------------------------------------
445 my ( $rargv, $msg ) = parse_args($argv);
448 Error parsing this string passed to to perltidy with 'argv':
456 # redirect STDERR if requested
457 if ($stderr_stream) {
458 my ( $fh_stderr, $stderr_file ) =
459 Perl::Tidy::streamhandle( $stderr_stream, 'w' );
460 if ($fh_stderr) { *STDERR = $fh_stderr }
463 ------------------------------------------------------------------------
464 Unable to redirect STDERR to $stderr_stream
465 Please check value of -stderr in call to perltidy
466 ------------------------------------------------------------------------
471 my $rpending_complaint;
472 $$rpending_complaint = "";
473 my $rpending_logfile_message;
474 $$rpending_logfile_message = "";
476 my ( $is_Windows, $Windows_type ) =
477 look_for_Windows($rpending_complaint);
479 # VMS file names are restricted to a 40.40 format, so we append _tdy
480 # instead of .tdy, etc. (but see also sub check_vms_filename)
483 if ( $^O eq 'VMS' ) {
489 $dot_pattern = '\.'; # must escape for use in regex
492 # handle command line options
493 my ( $rOpts, $config_file, $rraw_options, $saw_extrude, $roption_string,
494 $rexpansion, $roption_category, $roption_range )
495 = process_command_line(
496 $perltidyrc_stream, $is_Windows, $Windows_type,
497 $rpending_complaint, $dump_options_type,
500 # return or exit immediately after all dumps
503 # Getopt parameters and their flags
504 if ( defined($dump_getopt_flags) ) {
506 foreach my $op ( @{$roption_string} ) {
515 if ( $opt =~ /(.*)(!|=.*|:.*)$/ ) {
519 $dump_getopt_flags->{$opt} = $flag;
523 if ( defined($dump_options_category) ) {
525 %{$dump_options_category} = %{$roption_category};
528 if ( defined($dump_options_range) ) {
530 %{$dump_options_range} = %{$roption_range};
533 if ( defined($dump_abbreviations) ) {
535 %{$dump_abbreviations} = %{$rexpansion};
538 if ( defined($dump_options) ) {
540 %{$dump_options} = %{$rOpts};
543 return if ($quit_now);
545 # make printable string of options for this run as possible diagnostic
546 my $readable_options = readable_options( $rOpts, $roption_string );
548 # dump from command line
549 if ( $rOpts->{'dump-options'} ) {
550 print STDOUT $readable_options;
554 check_options( $rOpts, $is_Windows, $Windows_type,
555 $rpending_complaint );
557 if ($user_formatter) {
558 $rOpts->{'format'} = 'user';
561 # there must be one entry here for every possible format
562 my %default_file_extension = (
568 # be sure we have a valid output format
569 unless ( exists $default_file_extension{ $rOpts->{'format'} } ) {
570 my $formats = join ' ',
571 sort map { "'" . $_ . "'" } keys %default_file_extension;
572 my $fmt = $rOpts->{'format'};
573 die "-format='$fmt' but must be one of: $formats\n";
576 my $output_extension =
577 make_extension( $rOpts->{'output-file-extension'},
578 $default_file_extension{ $rOpts->{'format'} }, $dot );
580 my $backup_extension =
581 make_extension( $rOpts->{'backup-file-extension'}, 'bak', $dot );
583 my $html_toc_extension =
584 make_extension( $rOpts->{'html-toc-extension'}, 'toc', $dot );
586 my $html_src_extension =
587 make_extension( $rOpts->{'html-src-extension'}, 'src', $dot );
589 # check for -b option;
590 my $in_place_modify = $rOpts->{'backup-and-modify-in-place'}
591 && $rOpts->{'format'} eq 'tidy' # silently ignore unless beautify mode
592 && @ARGV > 0; # silently ignore if standard input;
593 # this allows -b to be in a .perltidyrc file
594 # without error messages when running from an editor
596 # turn off -b with warnings in case of conflicts with other options
597 if ($in_place_modify) {
598 if ( $rOpts->{'standard-output'} ) {
599 warn "Ignoring -b; you may not use -b and -st together\n";
600 $in_place_modify = 0;
602 if ($destination_stream) {
604 "Ignoring -b; you may not specify a destination array and -b together\n";
605 $in_place_modify = 0;
607 if ($source_stream) {
609 "Ignoring -b; you may not specify a source array and -b together\n";
610 $in_place_modify = 0;
612 if ( $rOpts->{'outfile'} ) {
613 warn "Ignoring -b; you may not use -b and -o together\n";
614 $in_place_modify = 0;
616 if ( defined( $rOpts->{'output-path'} ) ) {
617 warn "Ignoring -b; you may not use -b and -opath together\n";
618 $in_place_modify = 0;
622 Perl::Tidy::Formatter::check_options($rOpts);
623 if ( $rOpts->{'format'} eq 'html' ) {
624 Perl::Tidy::HtmlWriter->check_options($rOpts);
627 # make the pattern of file extensions that we shouldn't touch
628 my $forbidden_file_extensions = "(($dot_pattern)(LOG|DEBUG|ERR|TEE)";
629 if ($output_extension) {
630 my $ext = quotemeta($output_extension);
631 $forbidden_file_extensions .= "|$ext";
633 if ( $in_place_modify && $backup_extension ) {
634 my $ext = quotemeta($backup_extension);
635 $forbidden_file_extensions .= "|$ext";
637 $forbidden_file_extensions .= ')$';
639 # Create a diagnostics object if requested;
640 # This is only useful for code development
641 my $diagnostics_object = undef;
642 if ( $rOpts->{'DIAGNOSTICS'} ) {
643 $diagnostics_object = Perl::Tidy::Diagnostics->new();
646 # no filenames should be given if input is from an array
647 if ($source_stream) {
650 "You may not specify any filenames when a source array is given\n";
653 # we'll stuff the source array into ARGV
654 unshift( @ARGV, $source_stream );
656 # No special treatment for source stream which is a filename.
657 # This will enable checks for binary files and other bad stuff.
658 $source_stream = undef unless ref($source_stream);
661 # use stdin by default if no source array and no args
663 unshift( @ARGV, '-' ) unless @ARGV;
666 # loop to process all files in argument list
667 my $number_of_files = @ARGV;
668 my $formatter = undef;
670 while ( $input_file = shift @ARGV ) {
672 my $input_file_permissions;
674 #---------------------------------------------------------------
675 # determine the input file name
676 #---------------------------------------------------------------
677 if ($source_stream) {
678 $fileroot = "perltidy";
680 elsif ( $input_file eq '-' ) { # '-' indicates input from STDIN
681 $fileroot = "perltidy"; # root name to use for .ERR, .LOG, etc
682 $in_place_modify = 0;
685 $fileroot = $input_file;
686 unless ( -e $input_file ) {
688 # file doesn't exist - check for a file glob
689 if ( $input_file =~ /([\?\*\[\{])/ ) {
691 # Windows shell may not remove quotes, so do it
692 my $input_file = $input_file;
693 if ( $input_file =~ /^\'(.+)\'$/ ) { $input_file = $1 }
694 if ( $input_file =~ /^\"(.+)\"$/ ) { $input_file = $1 }
695 my $pattern = fileglob_to_re($input_file);
697 if ( !$@ && opendir( DIR, './' ) ) {
699 grep { /$pattern/ && !-d $_ } readdir(DIR);
702 unshift @ARGV, @files;
707 print "skipping file: '$input_file': no matches found\n";
711 unless ( -f $input_file ) {
712 print "skipping file: $input_file: not a regular file\n";
716 unless ( ( -T $input_file ) || $rOpts->{'force-read-binary'} ) {
718 "skipping file: $input_file: Non-text (override with -f)\n";
722 # we should have a valid filename now
723 $fileroot = $input_file;
724 $input_file_permissions = ( stat $input_file )[2] & 07777;
726 if ( $^O eq 'VMS' ) {
727 ( $fileroot, $dot ) = check_vms_filename($fileroot);
730 # add option to change path here
731 if ( defined( $rOpts->{'output-path'} ) ) {
733 my ( $base, $old_path ) = fileparse($fileroot);
734 my $new_path = $rOpts->{'output-path'};
735 unless ( -d $new_path ) {
736 unless ( mkdir $new_path, 0777 ) {
737 die "unable to create directory $new_path: $!\n";
740 my $path = $new_path;
741 $fileroot = catfile( $path, $base );
744 ------------------------------------------------------------------------
745 Problem combining $new_path and $base to make a filename; check -opath
746 ------------------------------------------------------------------------
752 # Skip files with same extension as the output files because
753 # this can lead to a messy situation with files like
754 # script.tdy.tdy.tdy ... or worse problems ... when you
755 # rerun perltidy over and over with wildcard input.
758 && ( $input_file =~ /$forbidden_file_extensions/o
759 || $input_file eq 'DIAGNOSTICS' )
762 print "skipping file: $input_file: wrong extension\n";
766 # the 'source_object' supplies a method to read the input file
768 Perl::Tidy::LineSource->new( $input_file, $rOpts,
769 $rpending_logfile_message );
770 next unless ($source_object);
772 # register this file name with the Diagnostics package
773 $diagnostics_object->set_input_file($input_file)
774 if $diagnostics_object;
776 #---------------------------------------------------------------
777 # determine the output file name
778 #---------------------------------------------------------------
779 my $output_file = undef;
780 my $actual_output_extension;
782 if ( $rOpts->{'outfile'} ) {
784 if ( $number_of_files <= 1 ) {
786 if ( $rOpts->{'standard-output'} ) {
787 die "You may not use -o and -st together\n";
789 elsif ($destination_stream) {
791 "You may not specify a destination array and -o together\n";
793 elsif ( defined( $rOpts->{'output-path'} ) ) {
794 die "You may not specify -o and -opath together\n";
796 elsif ( defined( $rOpts->{'output-file-extension'} ) ) {
797 die "You may not specify -o and -oext together\n";
799 $output_file = $rOpts->{outfile};
801 # make sure user gives a file name after -o
802 if ( $output_file =~ /^-/ ) {
803 die "You must specify a valid filename after -o\n";
806 # do not overwrite input file with -o
807 if ( defined($input_file_permissions)
808 && ( $output_file eq $input_file ) )
811 "Use 'perltidy -b $input_file' to modify in-place\n";
815 die "You may not use -o with more than one input file\n";
818 elsif ( $rOpts->{'standard-output'} ) {
819 if ($destination_stream) {
821 "You may not specify a destination array and -st together\n";
825 if ( $number_of_files <= 1 ) {
828 die "You may not use -st with more than one input file\n";
831 elsif ($destination_stream) {
832 $output_file = $destination_stream;
834 elsif ($source_stream) { # source but no destination goes to stdout
837 elsif ( $input_file eq '-' ) {
841 if ($in_place_modify) {
842 $output_file = IO::File->new_tmpfile()
843 or die "cannot open temp file for -b option: $!\n";
846 $actual_output_extension = $output_extension;
847 $output_file = $fileroot . $output_extension;
851 # the 'sink_object' knows how to write the output file
852 my $tee_file = $fileroot . $dot . "TEE";
854 my $line_separator = $rOpts->{'output-line-ending'};
855 if ( $rOpts->{'preserve-line-endings'} ) {
856 $line_separator = find_input_line_ending($input_file);
859 # Eventually all I/O may be done with binmode, but for now it is
860 # only done when a user requests a particular line separator
861 # through the -ple or -ole flags
863 if ( defined($line_separator) ) { $binmode = 1 }
864 else { $line_separator = "\n" }
867 Perl::Tidy::LineSink->new( $output_file, $tee_file,
868 $line_separator, $rOpts, $rpending_logfile_message, $binmode );
870 #---------------------------------------------------------------
871 # initialize the error logger
872 #---------------------------------------------------------------
873 my $warning_file = $fileroot . $dot . "ERR";
874 if ($errorfile_stream) { $warning_file = $errorfile_stream }
875 my $log_file = $fileroot . $dot . "LOG";
876 if ($logfile_stream) { $log_file = $logfile_stream }
879 Perl::Tidy::Logger->new( $rOpts, $log_file, $warning_file,
881 write_logfile_header(
882 $rOpts, $logger_object, $config_file,
883 $rraw_options, $Windows_type, $readable_options,
885 if ($$rpending_logfile_message) {
886 $logger_object->write_logfile_entry($$rpending_logfile_message);
888 if ($$rpending_complaint) {
889 $logger_object->complain($$rpending_complaint);
892 #---------------------------------------------------------------
893 # initialize the debug object, if any
894 #---------------------------------------------------------------
895 my $debugger_object = undef;
896 if ( $rOpts->{DEBUG} ) {
898 Perl::Tidy::Debugger->new( $fileroot . $dot . "DEBUG" );
901 #---------------------------------------------------------------
902 # create a formatter for this file : html writer or pretty printer
903 #---------------------------------------------------------------
905 # we have to delete any old formatter because, for safety,
906 # the formatter will check to see that there is only one.
909 if ($user_formatter) {
910 $formatter = $user_formatter;
912 elsif ( $rOpts->{'format'} eq 'html' ) {
914 Perl::Tidy::HtmlWriter->new( $fileroot, $output_file,
915 $actual_output_extension, $html_toc_extension,
916 $html_src_extension );
918 elsif ( $rOpts->{'format'} eq 'tidy' ) {
919 $formatter = Perl::Tidy::Formatter->new(
920 logger_object => $logger_object,
921 diagnostics_object => $diagnostics_object,
922 sink_object => $sink_object,
926 die "I don't know how to do -format=$rOpts->{'format'}\n";
929 unless ($formatter) {
930 die "Unable to continue with $rOpts->{'format'} formatting\n";
933 #---------------------------------------------------------------
934 # create the tokenizer for this file
935 #---------------------------------------------------------------
936 $tokenizer = undef; # must destroy old tokenizer
937 $tokenizer = Perl::Tidy::Tokenizer->new(
938 source_object => $source_object,
939 logger_object => $logger_object,
940 debugger_object => $debugger_object,
941 diagnostics_object => $diagnostics_object,
942 starting_level => $rOpts->{'starting-indentation-level'},
943 tabs => $rOpts->{'tabs'},
944 indent_columns => $rOpts->{'indent-columns'},
945 look_for_hash_bang => $rOpts->{'look-for-hash-bang'},
946 look_for_autoloader => $rOpts->{'look-for-autoloader'},
947 look_for_selfloader => $rOpts->{'look-for-selfloader'},
948 trim_qw => $rOpts->{'trim-qw'},
951 #---------------------------------------------------------------
953 #---------------------------------------------------------------
954 process_this_file( $tokenizer, $formatter );
956 #---------------------------------------------------------------
957 # close the input source and report errors
958 #---------------------------------------------------------------
959 $source_object->close_input_file();
961 # get file names to use for syntax check
962 my $ifname = $source_object->get_input_file_copy_name();
963 my $ofname = $sink_object->get_output_file_copy();
965 #---------------------------------------------------------------
966 # handle the -b option (backup and modify in-place)
967 #---------------------------------------------------------------
968 if ($in_place_modify) {
969 unless ( -f $input_file ) {
971 # oh, oh, no real file to backup ..
972 # shouldn't happen because of numerous preliminary checks
974 "problem with -b backing up input file '$input_file': not a file\n";
976 my $backup_name = $input_file . $backup_extension;
977 if ( -f $backup_name ) {
980 "unable to remove previous '$backup_name' for -b option; check permissions: $!\n";
982 rename( $input_file, $backup_name )
984 "problem renaming $input_file to $backup_name for -b option: $!\n";
985 $ifname = $backup_name;
987 seek( $output_file, 0, 0 )
988 or die "unable to rewind tmp file for -b option: $!\n";
990 my $fout = IO::File->new("> $input_file")
992 "problem opening $input_file for write for -b option; check directory permissions: $!\n";
995 while ( $line = $output_file->getline() ) {
999 $output_file = $input_file;
1000 $ofname = $input_file;
1003 #---------------------------------------------------------------
1004 # clean up and report errors
1005 #---------------------------------------------------------------
1006 $sink_object->close_output_file() if $sink_object;
1007 $debugger_object->close_debug_file() if $debugger_object;
1009 my $infile_syntax_ok = 0; # -1 no 0=don't know 1 yes
1012 if ($input_file_permissions) {
1014 # give output script same permissions as input script, but
1015 # make it user-writable or else we can't run perltidy again.
1016 # Thus we retain whatever executable flags were set.
1017 if ( $rOpts->{'format'} eq 'tidy' ) {
1018 chmod( $input_file_permissions | 0600, $output_file );
1021 # else use default permissions for html and any other format
1024 if ( $logger_object && $rOpts->{'check-syntax'} ) {
1026 check_syntax( $ifname, $ofname, $logger_object, $rOpts );
1030 $logger_object->finish( $infile_syntax_ok, $formatter )
1032 } # end of loop to process all files
1033 } # end of main program
1036 sub fileglob_to_re {
1038 # modified (corrected) from version in find2perl
1040 $x =~ s#([./^\$()])#\\$1#g; # escape special characters
1041 $x =~ s#\*#.*#g; # '*' -> '.*'
1042 $x =~ s#\?#.#g; # '?' -> '.'
1043 "^$x\\z"; # match whole word
1046 sub make_extension {
1048 # Make a file extension, including any leading '.' if necessary
1049 # The '.' may actually be an '_' under VMS
1050 my ( $extension, $default, $dot ) = @_;
1052 # Use the default if none specified
1053 $extension = $default unless ($extension);
1055 # Only extensions with these leading characters get a '.'
1056 # This rule gives the user some freedom
1057 if ( $extension =~ /^[a-zA-Z0-9]/ ) {
1058 $extension = $dot . $extension;
1063 sub write_logfile_header {
1065 $rOpts, $logger_object, $config_file,
1066 $rraw_options, $Windows_type, $readable_options
1068 $logger_object->write_logfile_entry(
1069 "perltidy version $VERSION log file on a $^O system, OLD_PERL_VERSION=$]\n"
1071 if ($Windows_type) {
1072 $logger_object->write_logfile_entry("Windows type is $Windows_type\n");
1074 my $options_string = join( ' ', @$rraw_options );
1077 $logger_object->write_logfile_entry(
1078 "Found Configuration File >>> $config_file \n");
1080 $logger_object->write_logfile_entry(
1081 "Configuration and command line parameters for this run:\n");
1082 $logger_object->write_logfile_entry("$options_string\n");
1084 if ( $rOpts->{'DEBUG'} || $rOpts->{'show-options'} ) {
1085 $rOpts->{'logfile'} = 1; # force logfile to be saved
1086 $logger_object->write_logfile_entry(
1087 "Final parameter set for this run\n");
1088 $logger_object->write_logfile_entry(
1089 "------------------------------------\n");
1091 $logger_object->write_logfile_entry($readable_options);
1093 $logger_object->write_logfile_entry(
1094 "------------------------------------\n");
1096 $logger_object->write_logfile_entry(
1097 "To find error messages search for 'WARNING' with your editor\n");
1100 sub generate_options {
1102 ######################################################################
1103 # Generate and return references to:
1104 # @option_string - the list of options to be passed to Getopt::Long
1105 # @defaults - the list of default options
1106 # %expansion - a hash showing how all abbreviations are expanded
1107 # %category - a hash giving the general category of each option
1108 # %option_range - a hash giving the valid ranges of certain options
1110 # Note: a few options are not documented in the man page and usage
1111 # message. This is because these are experimental or debug options and
1112 # may or may not be retained in future versions.
1114 # Here are the undocumented flags as far as I know. Any of them
1115 # may disappear at any time. They are mainly for fine-tuning
1118 # fll --> fuzzy-line-length # a trivial parameter which gets
1119 # turned off for the extrude option
1120 # which is mainly for debugging
1121 # chk --> check-multiline-quotes # check for old bug; to be deleted
1122 # scl --> short-concatenation-item-length # helps break at '.'
1123 # recombine # for debugging line breaks
1124 # valign # for debugging vertical alignment
1125 # I --> DIAGNOSTICS # for debugging
1126 ######################################################################
1128 # here is a summary of the Getopt codes:
1129 # <none> does not take an argument
1130 # =s takes a mandatory string
1131 # :s takes an optional string (DO NOT USE - filenames will get eaten up)
1132 # =i takes a mandatory integer
1133 # :i takes an optional integer (NOT RECOMMENDED - can cause trouble)
1134 # ! does not take an argument and may be negated
1135 # i.e., -foo and -nofoo are allowed
1136 # a double dash signals the end of the options list
1138 #---------------------------------------------------------------
1139 # Define the option string passed to GetOptions.
1140 #---------------------------------------------------------------
1142 my @option_string = ();
1144 my %option_category = ();
1145 my %option_range = ();
1146 my $rexpansion = \%expansion;
1148 # names of categories in manual
1149 # leading integers will allow sorting
1150 my @category_name = (
1152 '1. Basic formatting options',
1153 '2. Code indentation control',
1154 '3. Whitespace control',
1155 '4. Comment controls',
1156 '5. Linebreak controls',
1157 '6. Controlling list formatting',
1158 '7. Retaining or ignoring existing line breaks',
1159 '8. Blank line control',
1160 '9. Other controls',
1162 '11. pod2html options',
1163 '12. Controlling HTML properties',
1167 # These options are parsed directly by perltidy:
1170 # However, they are included in the option set so that they will
1171 # be seen in the options dump.
1173 # These long option names have no abbreviations or are treated specially
1174 @option_string = qw(
1183 my $category = 13; # Debugging
1184 foreach (@option_string) {
1185 my $opt = $_; # must avoid changing the actual flag
1187 $option_category{$opt} = $category_name[$category];
1190 $category = 11; # HTML
1191 $option_category{html} = $category_name[$category];
1193 # routine to install and check options
1194 my $add_option = sub {
1195 my ( $long_name, $short_name, $flag ) = @_;
1196 push @option_string, $long_name . $flag;
1197 $option_category{$long_name} = $category_name[$category];
1199 if ( $expansion{$short_name} ) {
1200 my $existing_name = $expansion{$short_name}[0];
1202 "redefining abbreviation $short_name for $long_name; already used for $existing_name\n";
1204 $expansion{$short_name} = [$long_name];
1205 if ( $flag eq '!' ) {
1206 my $nshort_name = 'n' . $short_name;
1207 my $nolong_name = 'no' . $long_name;
1208 if ( $expansion{$nshort_name} ) {
1209 my $existing_name = $expansion{$nshort_name}[0];
1211 "attempting to redefine abbreviation $nshort_name for $nolong_name; already used for $existing_name\n";
1213 $expansion{$nshort_name} = [$nolong_name];
1218 # Install long option names which have a simple abbreviation.
1219 # Options with code '!' get standard negation ('no' for long names,
1220 # 'n' for abbreviations). Categories follow the manual.
1222 ###########################
1223 $category = 0; # I/O_Control
1224 ###########################
1225 $add_option->( 'backup-and-modify-in-place', 'b', '!' );
1226 $add_option->( 'backup-file-extension', 'bext', '=s' );
1227 $add_option->( 'force-read-binary', 'f', '!' );
1228 $add_option->( 'format', 'fmt', '=s' );
1229 $add_option->( 'logfile', 'log', '!' );
1230 $add_option->( 'logfile-gap', 'g', ':i' );
1231 $add_option->( 'outfile', 'o', '=s' );
1232 $add_option->( 'output-file-extension', 'oext', '=s' );
1233 $add_option->( 'output-path', 'opath', '=s' );
1234 $add_option->( 'profile', 'pro', '=s' );
1235 $add_option->( 'quiet', 'q', '!' );
1236 $add_option->( 'standard-error-output', 'se', '!' );
1237 $add_option->( 'standard-output', 'st', '!' );
1238 $add_option->( 'warning-output', 'w', '!' );
1240 # options which are both toggle switches and values moved here
1241 # to hide from tidyview (which does not show category 0 flags):
1242 # -ole moved here from category 1
1243 # -sil moved here from category 2
1244 $add_option->( 'output-line-ending', 'ole', '=s' );
1245 $add_option->( 'starting-indentation-level', 'sil', '=i' );
1247 ########################################
1248 $category = 1; # Basic formatting options
1249 ########################################
1250 $add_option->( 'check-syntax', 'syn', '!' );
1251 $add_option->( 'entab-leading-whitespace', 'et', '=i' );
1252 $add_option->( 'indent-columns', 'i', '=i' );
1253 $add_option->( 'maximum-line-length', 'l', '=i' );
1254 $add_option->( 'perl-syntax-check-flags', 'pscf', '=s' );
1255 $add_option->( 'preserve-line-endings', 'ple', '!' );
1256 $add_option->( 'tabs', 't', '!' );
1258 ########################################
1259 $category = 2; # Code indentation control
1260 ########################################
1261 $add_option->( 'continuation-indentation', 'ci', '=i' );
1262 $add_option->( 'line-up-parentheses', 'lp', '!' );
1263 $add_option->( 'outdent-keyword-list', 'okwl', '=s' );
1264 $add_option->( 'outdent-keywords', 'okw', '!' );
1265 $add_option->( 'outdent-labels', 'ola', '!' );
1266 $add_option->( 'outdent-long-quotes', 'olq', '!' );
1267 $add_option->( 'indent-closing-brace', 'icb', '!' );
1268 $add_option->( 'closing-token-indentation', 'cti', '=i' );
1269 $add_option->( 'closing-paren-indentation', 'cpi', '=i' );
1270 $add_option->( 'closing-brace-indentation', 'cbi', '=i' );
1271 $add_option->( 'closing-square-bracket-indentation', 'csbi', '=i' );
1272 $add_option->( 'brace-left-and-indent', 'bli', '!' );
1273 $add_option->( 'brace-left-and-indent-list', 'blil', '=s' );
1275 ########################################
1276 $category = 3; # Whitespace control
1277 ########################################
1278 $add_option->( 'add-semicolons', 'asc', '!' );
1279 $add_option->( 'add-whitespace', 'aws', '!' );
1280 $add_option->( 'block-brace-tightness', 'bbt', '=i' );
1281 $add_option->( 'brace-tightness', 'bt', '=i' );
1282 $add_option->( 'delete-old-whitespace', 'dws', '!' );
1283 $add_option->( 'delete-semicolons', 'dsm', '!' );
1284 $add_option->( 'nospace-after-keyword', 'nsak', '=s' );
1285 $add_option->( 'nowant-left-space', 'nwls', '=s' );
1286 $add_option->( 'nowant-right-space', 'nwrs', '=s' );
1287 $add_option->( 'paren-tightness', 'pt', '=i' );
1288 $add_option->( 'space-after-keyword', 'sak', '=s' );
1289 $add_option->( 'space-for-semicolon', 'sfs', '!' );
1290 $add_option->( 'space-function-paren', 'sfp', '!' );
1291 $add_option->( 'space-keyword-paren', 'skp', '!' );
1292 $add_option->( 'space-terminal-semicolon', 'sts', '!' );
1293 $add_option->( 'square-bracket-tightness', 'sbt', '=i' );
1294 $add_option->( 'square-bracket-vertical-tightness', 'sbvt', '=i' );
1295 $add_option->( 'square-bracket-vertical-tightness-closing', 'sbvtc', '=i' );
1296 $add_option->( 'trim-qw', 'tqw', '!' );
1297 $add_option->( 'want-left-space', 'wls', '=s' );
1298 $add_option->( 'want-right-space', 'wrs', '=s' );
1300 ########################################
1301 $category = 4; # Comment controls
1302 ########################################
1303 $add_option->( 'closing-side-comment-else-flag', 'csce', '=i' );
1304 $add_option->( 'closing-side-comment-interval', 'csci', '=i' );
1305 $add_option->( 'closing-side-comment-list', 'cscl', '=s' );
1306 $add_option->( 'closing-side-comment-maximum-text', 'csct', '=i' );
1307 $add_option->( 'closing-side-comment-prefix', 'cscp', '=s' );
1308 $add_option->( 'closing-side-comment-warnings', 'cscw', '!' );
1309 $add_option->( 'closing-side-comments', 'csc', '!' );
1310 $add_option->( 'format-skipping', 'fs', '!' );
1311 $add_option->( 'format-skipping-begin', 'fsb', '=s' );
1312 $add_option->( 'format-skipping-end', 'fse', '=s' );
1313 $add_option->( 'hanging-side-comments', 'hsc', '!' );
1314 $add_option->( 'indent-block-comments', 'ibc', '!' );
1315 $add_option->( 'indent-spaced-block-comments', 'isbc', '!' );
1316 $add_option->( 'fixed-position-side-comment', 'fpsc', '=i' );
1317 $add_option->( 'minimum-space-to-comment', 'msc', '=i' );
1318 $add_option->( 'outdent-long-comments', 'olc', '!' );
1319 $add_option->( 'outdent-static-block-comments', 'osbc', '!' );
1320 $add_option->( 'static-block-comment-prefix', 'sbcp', '=s' );
1321 $add_option->( 'static-block-comments', 'sbc', '!' );
1322 $add_option->( 'static-side-comment-prefix', 'sscp', '=s' );
1323 $add_option->( 'static-side-comments', 'ssc', '!' );
1325 ########################################
1326 $category = 5; # Linebreak controls
1327 ########################################
1328 $add_option->( 'add-newlines', 'anl', '!' );
1329 $add_option->( 'block-brace-vertical-tightness', 'bbvt', '=i' );
1330 $add_option->( 'block-brace-vertical-tightness-list', 'bbvtl', '=s' );
1331 $add_option->( 'brace-vertical-tightness', 'bvt', '=i' );
1332 $add_option->( 'brace-vertical-tightness-closing', 'bvtc', '=i' );
1333 $add_option->( 'cuddled-else', 'ce', '!' );
1334 $add_option->( 'delete-old-newlines', 'dnl', '!' );
1335 $add_option->( 'opening-brace-always-on-right', 'bar', '!' );
1336 $add_option->( 'opening-brace-on-new-line', 'bl', '!' );
1337 $add_option->( 'opening-hash-brace-right', 'ohbr', '!' );
1338 $add_option->( 'opening-paren-right', 'opr', '!' );
1339 $add_option->( 'opening-square-bracket-right', 'osbr', '!' );
1340 $add_option->( 'opening-sub-brace-on-new-line', 'sbl', '!' );
1341 $add_option->( 'paren-vertical-tightness', 'pvt', '=i' );
1342 $add_option->( 'paren-vertical-tightness-closing', 'pvtc', '=i' );
1343 $add_option->( 'stack-closing-hash-brace', 'schb', '!' );
1344 $add_option->( 'stack-closing-paren', 'scp', '!' );
1345 $add_option->( 'stack-closing-square-bracket', 'scsb', '!' );
1346 $add_option->( 'stack-opening-hash-brace', 'sohb', '!' );
1347 $add_option->( 'stack-opening-paren', 'sop', '!' );
1348 $add_option->( 'stack-opening-square-bracket', 'sosb', '!' );
1349 $add_option->( 'vertical-tightness', 'vt', '=i' );
1350 $add_option->( 'vertical-tightness-closing', 'vtc', '=i' );
1351 $add_option->( 'want-break-after', 'wba', '=s' );
1352 $add_option->( 'want-break-before', 'wbb', '=s' );
1353 $add_option->( 'break-after-all-operators', 'baao', '!' );
1354 $add_option->( 'break-before-all-operators', 'bbao', '!' );
1355 $add_option->( 'keep-interior-semicolons', 'kis', '!' );
1357 ########################################
1358 $category = 6; # Controlling list formatting
1359 ########################################
1360 $add_option->( 'break-at-old-comma-breakpoints', 'boc', '!' );
1361 $add_option->( 'comma-arrow-breakpoints', 'cab', '=i' );
1362 $add_option->( 'maximum-fields-per-table', 'mft', '=i' );
1364 ########################################
1365 $category = 7; # Retaining or ignoring existing line breaks
1366 ########################################
1367 $add_option->( 'break-at-old-keyword-breakpoints', 'bok', '!' );
1368 $add_option->( 'break-at-old-logical-breakpoints', 'bol', '!' );
1369 $add_option->( 'break-at-old-ternary-breakpoints', 'bot', '!' );
1370 $add_option->( 'ignore-old-breakpoints', 'iob', '!' );
1372 ########################################
1373 $category = 8; # Blank line control
1374 ########################################
1375 $add_option->( 'blanks-before-blocks', 'bbb', '!' );
1376 $add_option->( 'blanks-before-comments', 'bbc', '!' );
1377 $add_option->( 'blanks-before-subs', 'bbs', '!' );
1378 $add_option->( 'long-block-line-count', 'lbl', '=i' );
1379 $add_option->( 'maximum-consecutive-blank-lines', 'mbl', '=i' );
1380 $add_option->( 'swallow-optional-blank-lines', 'sob', '!' );
1382 ########################################
1383 $category = 9; # Other controls
1384 ########################################
1385 $add_option->( 'delete-block-comments', 'dbc', '!' );
1386 $add_option->( 'delete-closing-side-comments', 'dcsc', '!' );
1387 $add_option->( 'delete-pod', 'dp', '!' );
1388 $add_option->( 'delete-side-comments', 'dsc', '!' );
1389 $add_option->( 'tee-block-comments', 'tbc', '!' );
1390 $add_option->( 'tee-pod', 'tp', '!' );
1391 $add_option->( 'tee-side-comments', 'tsc', '!' );
1392 $add_option->( 'look-for-autoloader', 'lal', '!' );
1393 $add_option->( 'look-for-hash-bang', 'x', '!' );
1394 $add_option->( 'look-for-selfloader', 'lsl', '!' );
1395 $add_option->( 'pass-version-line', 'pvl', '!' );
1397 ########################################
1398 $category = 13; # Debugging
1399 ########################################
1400 $add_option->( 'DEBUG', 'D', '!' );
1401 $add_option->( 'DIAGNOSTICS', 'I', '!' );
1402 $add_option->( 'check-multiline-quotes', 'chk', '!' );
1403 $add_option->( 'dump-defaults', 'ddf', '!' );
1404 $add_option->( 'dump-long-names', 'dln', '!' );
1405 $add_option->( 'dump-options', 'dop', '!' );
1406 $add_option->( 'dump-profile', 'dpro', '!' );
1407 $add_option->( 'dump-short-names', 'dsn', '!' );
1408 $add_option->( 'dump-token-types', 'dtt', '!' );
1409 $add_option->( 'dump-want-left-space', 'dwls', '!' );
1410 $add_option->( 'dump-want-right-space', 'dwrs', '!' );
1411 $add_option->( 'fuzzy-line-length', 'fll', '!' );
1412 $add_option->( 'help', 'h', '' );
1413 $add_option->( 'short-concatenation-item-length', 'scl', '=i' );
1414 $add_option->( 'show-options', 'opt', '!' );
1415 $add_option->( 'version', 'v', '' );
1417 #---------------------------------------------------------------------
1419 # The Perl::Tidy::HtmlWriter will add its own options to the string
1420 Perl::Tidy::HtmlWriter->make_getopt_long_names( \@option_string );
1422 ########################################
1423 # Set categories 10, 11, 12
1424 ########################################
1425 # Based on their known order
1426 $category = 12; # HTML properties
1427 foreach my $opt (@option_string) {
1428 my $long_name = $opt;
1429 $long_name =~ s/(!|=.*|:.*)$//;
1430 unless ( defined( $option_category{$long_name} ) ) {
1431 if ( $long_name =~ /^html-linked/ ) {
1432 $category = 10; # HTML options
1434 elsif ( $long_name =~ /^pod2html/ ) {
1435 $category = 11; # Pod2html
1437 $option_category{$long_name} = $category_name[$category];
1441 #---------------------------------------------------------------
1442 # Assign valid ranges to certain options
1443 #---------------------------------------------------------------
1444 # In the future, these may be used to make preliminary checks
1445 # hash keys are long names
1446 # If key or value is undefined:
1447 # strings may have any value
1448 # integer ranges are >=0
1449 # If value is defined:
1450 # value is [qw(any valid words)] for strings
1451 # value is [min, max] for integers
1452 # if min is undefined, there is no lower limit
1453 # if max is undefined, there is no upper limit
1454 # Parameters not listed here have defaults
1456 'format' => [ 'tidy', 'html', 'user' ],
1457 'output-line-ending' => [ 'dos', 'win', 'mac', 'unix' ],
1459 'block-brace-tightness' => [ 0, 2 ],
1460 'brace-tightness' => [ 0, 2 ],
1461 'paren-tightness' => [ 0, 2 ],
1462 'square-bracket-tightness' => [ 0, 2 ],
1464 'block-brace-vertical-tightness' => [ 0, 2 ],
1465 'brace-vertical-tightness' => [ 0, 2 ],
1466 'brace-vertical-tightness-closing' => [ 0, 2 ],
1467 'paren-vertical-tightness' => [ 0, 2 ],
1468 'paren-vertical-tightness-closing' => [ 0, 2 ],
1469 'square-bracket-vertical-tightness' => [ 0, 2 ],
1470 'square-bracket-vertical-tightness-closing' => [ 0, 2 ],
1471 'vertical-tightness' => [ 0, 2 ],
1472 'vertical-tightness-closing' => [ 0, 2 ],
1474 'closing-brace-indentation' => [ 0, 3 ],
1475 'closing-paren-indentation' => [ 0, 3 ],
1476 'closing-square-bracket-indentation' => [ 0, 3 ],
1477 'closing-token-indentation' => [ 0, 3 ],
1479 'closing-side-comment-else-flag' => [ 0, 2 ],
1480 'comma-arrow-breakpoints' => [ 0, 3 ],
1483 # Note: we could actually allow negative ci if someone really wants it:
1484 # $option_range{'continuation-indentation'} = [ undef, undef ];
1486 #---------------------------------------------------------------
1487 # Assign default values to the above options here, except
1488 # for 'outfile' and 'help'.
1489 # These settings should approximate the perlstyle(1) suggestions.
1490 #---------------------------------------------------------------
1495 blanks-before-blocks
1496 blanks-before-comments
1498 block-brace-tightness=0
1499 block-brace-vertical-tightness=0
1501 brace-vertical-tightness-closing=0
1502 brace-vertical-tightness=0
1503 break-at-old-logical-breakpoints
1504 break-at-old-ternary-breakpoints
1505 break-at-old-keyword-breakpoints
1506 comma-arrow-breakpoints=1
1508 closing-side-comment-interval=6
1509 closing-side-comment-maximum-text=20
1510 closing-side-comment-else-flag=0
1511 closing-paren-indentation=0
1512 closing-brace-indentation=0
1513 closing-square-bracket-indentation=0
1514 continuation-indentation=2
1518 hanging-side-comments
1519 indent-block-comments
1521 long-block-line-count=8
1524 maximum-consecutive-blank-lines=1
1525 maximum-fields-per-table=0
1526 maximum-line-length=80
1527 minimum-space-to-comment=4
1528 nobrace-left-and-indent
1530 nodelete-old-whitespace
1535 nostatic-side-comments
1536 noswallow-optional-blank-lines
1541 outdent-long-comments
1543 paren-vertical-tightness-closing=0
1544 paren-vertical-tightness=0
1548 short-concatenation-item-length=8
1550 square-bracket-tightness=1
1551 square-bracket-vertical-tightness-closing=0
1552 square-bracket-vertical-tightness=0
1553 static-block-comments
1556 backup-file-extension=bak
1560 html-table-of-contents
1564 push @defaults, "perl-syntax-check-flags=-c -T";
1566 #---------------------------------------------------------------
1567 # Define abbreviations which will be expanded into the above primitives.
1568 # These may be defined recursively.
1569 #---------------------------------------------------------------
1572 'freeze-newlines' => [qw(noadd-newlines nodelete-old-newlines)],
1573 'fnl' => [qw(freeze-newlines)],
1574 'freeze-whitespace' => [qw(noadd-whitespace nodelete-old-whitespace)],
1575 'fws' => [qw(freeze-whitespace)],
1576 'indent-only' => [qw(freeze-newlines freeze-whitespace)],
1577 'outdent-long-lines' => [qw(outdent-long-quotes outdent-long-comments)],
1578 'nooutdent-long-lines' =>
1579 [qw(nooutdent-long-quotes nooutdent-long-comments)],
1580 'noll' => [qw(nooutdent-long-lines)],
1581 'io' => [qw(indent-only)],
1582 'delete-all-comments' =>
1583 [qw(delete-block-comments delete-side-comments delete-pod)],
1584 'nodelete-all-comments' =>
1585 [qw(nodelete-block-comments nodelete-side-comments nodelete-pod)],
1586 'dac' => [qw(delete-all-comments)],
1587 'ndac' => [qw(nodelete-all-comments)],
1588 'gnu' => [qw(gnu-style)],
1589 'pbp' => [qw(perl-best-practices)],
1590 'tee-all-comments' =>
1591 [qw(tee-block-comments tee-side-comments tee-pod)],
1592 'notee-all-comments' =>
1593 [qw(notee-block-comments notee-side-comments notee-pod)],
1594 'tac' => [qw(tee-all-comments)],
1595 'ntac' => [qw(notee-all-comments)],
1596 'html' => [qw(format=html)],
1597 'nhtml' => [qw(format=tidy)],
1598 'tidy' => [qw(format=tidy)],
1600 'break-after-comma-arrows' => [qw(cab=0)],
1601 'nobreak-after-comma-arrows' => [qw(cab=1)],
1602 'baa' => [qw(cab=0)],
1603 'nbaa' => [qw(cab=1)],
1605 'break-at-old-trinary-breakpoints' => [qw(bot)],
1607 'cti=0' => [qw(cpi=0 cbi=0 csbi=0)],
1608 'cti=1' => [qw(cpi=1 cbi=1 csbi=1)],
1609 'cti=2' => [qw(cpi=2 cbi=2 csbi=2)],
1610 'icp' => [qw(cpi=2 cbi=2 csbi=2)],
1611 'nicp' => [qw(cpi=0 cbi=0 csbi=0)],
1613 'closing-token-indentation=0' => [qw(cpi=0 cbi=0 csbi=0)],
1614 'closing-token-indentation=1' => [qw(cpi=1 cbi=1 csbi=1)],
1615 'closing-token-indentation=2' => [qw(cpi=2 cbi=2 csbi=2)],
1616 'indent-closing-paren' => [qw(cpi=2 cbi=2 csbi=2)],
1617 'noindent-closing-paren' => [qw(cpi=0 cbi=0 csbi=0)],
1619 'vt=0' => [qw(pvt=0 bvt=0 sbvt=0)],
1620 'vt=1' => [qw(pvt=1 bvt=1 sbvt=1)],
1621 'vt=2' => [qw(pvt=2 bvt=2 sbvt=2)],
1623 'vertical-tightness=0' => [qw(pvt=0 bvt=0 sbvt=0)],
1624 'vertical-tightness=1' => [qw(pvt=1 bvt=1 sbvt=1)],
1625 'vertical-tightness=2' => [qw(pvt=2 bvt=2 sbvt=2)],
1627 'vtc=0' => [qw(pvtc=0 bvtc=0 sbvtc=0)],
1628 'vtc=1' => [qw(pvtc=1 bvtc=1 sbvtc=1)],
1629 'vtc=2' => [qw(pvtc=2 bvtc=2 sbvtc=2)],
1631 'vertical-tightness-closing=0' => [qw(pvtc=0 bvtc=0 sbvtc=0)],
1632 'vertical-tightness-closing=1' => [qw(pvtc=1 bvtc=1 sbvtc=1)],
1633 'vertical-tightness-closing=2' => [qw(pvtc=2 bvtc=2 sbvtc=2)],
1635 'otr' => [qw(opr ohbr osbr)],
1636 'opening-token-right' => [qw(opr ohbr osbr)],
1637 'notr' => [qw(nopr nohbr nosbr)],
1638 'noopening-token-right' => [qw(nopr nohbr nosbr)],
1640 'sot' => [qw(sop sohb sosb)],
1641 'nsot' => [qw(nsop nsohb nsosb)],
1642 'stack-opening-tokens' => [qw(sop sohb sosb)],
1643 'nostack-opening-tokens' => [qw(nsop nsohb nsosb)],
1645 'sct' => [qw(scp schb scsb)],
1646 'stack-closing-tokens' => => [qw(scp schb scsb)],
1647 'nsct' => [qw(nscp nschb nscsb)],
1648 'nostack-opening-tokens' => [qw(nscp nschb nscsb)],
1650 # 'mangle' originally deleted pod and comments, but to keep it
1651 # reversible, it no longer does. But if you really want to
1652 # delete them, just use:
1655 # An interesting use for 'mangle' is to do this:
1656 # perltidy -mangle myfile.pl -st | perltidy -o myfile.pl.new
1657 # which will form as many one-line blocks as possible
1663 delete-old-whitespace
1666 maximum-consecutive-blank-lines=0
1667 maximum-line-length=100000
1671 noblanks-before-blocks
1672 noblanks-before-subs
1677 # 'extrude' originally deleted pod and comments, but to keep it
1678 # reversible, it no longer does. But if you really want to
1679 # delete them, just use
1682 # An interesting use for 'extrude' is to do this:
1683 # perltidy -extrude myfile.pl -st | perltidy -o myfile.pl.new
1684 # which will break up all one-line blocks.
1691 delete-old-whitespace
1694 maximum-consecutive-blank-lines=0
1695 maximum-line-length=1
1698 noblanks-before-blocks
1699 noblanks-before-subs
1706 # this style tries to follow the GNU Coding Standards (which do
1707 # not really apply to perl but which are followed by some perl
1711 lp bl noll pt=2 bt=2 sbt=2 cpi=1 csbi=1 cbi=1
1715 # Style suggested in Damian Conway's Perl Best Practices
1716 'perl-best-practices' => [
1717 qw(l=78 i=4 ci=4 st se vt=2 cti=0 pt=1 bt=1 sbt=1 bbt=1 nsfs nolq),
1718 q(wbb=% + - * / x != == >= <= =~ !~ < > | & = **= += *= &= <<= &&= -= /= |= >>= ||= //= .= %= ^= x=)
1721 # Additional styles can be added here
1724 Perl::Tidy::HtmlWriter->make_abbreviated_names( \%expansion );
1726 # Uncomment next line to dump all expansions for debugging:
1727 # dump_short_names(\%expansion);
1729 \@option_string, \@defaults, \%expansion,
1730 \%option_category, \%option_range
1733 } # end of generate_options
1735 sub process_command_line {
1738 $perltidyrc_stream, $is_Windows, $Windows_type,
1739 $rpending_complaint, $dump_options_type
1745 $roption_string, $rdefaults, $rexpansion,
1746 $roption_category, $roption_range
1747 ) = generate_options();
1749 #---------------------------------------------------------------
1750 # set the defaults by passing the above list through GetOptions
1751 #---------------------------------------------------------------
1757 # do not load the defaults if we are just dumping perltidyrc
1758 unless ( $dump_options_type eq 'perltidyrc' ) {
1759 for $i (@$rdefaults) { push @ARGV, "--" . $i }
1762 # Patch to save users Getopt::Long configuration
1763 # and set to Getopt::Long defaults. Use eval to avoid
1764 # breaking old versions of Perl without these routines.
1766 eval { $glc = Getopt::Long::Configure() };
1768 eval { Getopt::Long::ConfigDefaults() };
1770 else { $glc = undef }
1772 if ( !GetOptions( \%Opts, @$roption_string ) ) {
1773 die "Programming Bug: error in setting default options";
1776 # Patch to put the previous Getopt::Long configuration back
1777 eval { Getopt::Long::Configure($glc) } if defined $glc;
1781 my @raw_options = ();
1782 my $config_file = "";
1783 my $saw_ignore_profile = 0;
1784 my $saw_extrude = 0;
1785 my $saw_dump_profile = 0;
1788 #---------------------------------------------------------------
1789 # Take a first look at the command-line parameters. Do as many
1790 # immediate dumps as possible, which can avoid confusion if the
1791 # perltidyrc file has an error.
1792 #---------------------------------------------------------------
1793 foreach $i (@ARGV) {
1796 if ( $i =~ /^-(npro|noprofile|no-profile)$/ ) {
1797 $saw_ignore_profile = 1;
1800 # note: this must come before -pro and -profile, below:
1801 elsif ( $i =~ /^-(dump-profile|dpro)$/ ) {
1802 $saw_dump_profile = 1;
1804 elsif ( $i =~ /^-(pro|profile)=(.+)/ ) {
1807 "Only one -pro=filename allowed, using '$2' instead of '$config_file'\n";
1810 unless ( -e $config_file ) {
1811 warn "cannot find file given with -pro=$config_file: $!\n";
1815 elsif ( $i =~ /^-(pro|profile)=?$/ ) {
1816 die "usage: -pro=filename or --profile=filename, no spaces\n";
1818 elsif ( $i =~ /^-extrude$/ ) {
1821 elsif ( $i =~ /^-(help|h|HELP|H)$/ ) {
1825 elsif ( $i =~ /^-(version|v)$/ ) {
1829 elsif ( $i =~ /^-(dump-defaults|ddf)$/ ) {
1830 dump_defaults(@$rdefaults);
1833 elsif ( $i =~ /^-(dump-long-names|dln)$/ ) {
1834 dump_long_names(@$roption_string);
1837 elsif ( $i =~ /^-(dump-short-names|dsn)$/ ) {
1838 dump_short_names($rexpansion);
1841 elsif ( $i =~ /^-(dump-token-types|dtt)$/ ) {
1842 Perl::Tidy::Tokenizer->dump_token_types(*STDOUT);
1847 if ( $saw_dump_profile && $saw_ignore_profile ) {
1848 warn "No profile to dump because of -npro\n";
1852 #---------------------------------------------------------------
1853 # read any .perltidyrc configuration file
1854 #---------------------------------------------------------------
1855 unless ($saw_ignore_profile) {
1857 # resolve possible conflict between $perltidyrc_stream passed
1858 # as call parameter to perltidy and -pro=filename on command
1860 if ($perltidyrc_stream) {
1863 Conflict: a perltidyrc configuration file was specified both as this
1864 perltidy call parameter: $perltidyrc_stream
1865 and with this -profile=$config_file.
1866 Using -profile=$config_file.
1870 $config_file = $perltidyrc_stream;
1874 # look for a config file if we don't have one yet
1875 my $rconfig_file_chatter;
1876 $$rconfig_file_chatter = "";
1878 find_config_file( $is_Windows, $Windows_type, $rconfig_file_chatter,
1879 $rpending_complaint )
1880 unless $config_file;
1882 # open any config file
1885 ( $fh_config, $config_file ) =
1886 Perl::Tidy::streamhandle( $config_file, 'r' );
1887 unless ($fh_config) {
1888 $$rconfig_file_chatter .=
1889 "# $config_file exists but cannot be opened\n";
1893 if ($saw_dump_profile) {
1894 if ($saw_dump_profile) {
1895 dump_config_file( $fh_config, $config_file,
1896 $rconfig_file_chatter );
1903 my ( $rconfig_list, $death_message ) =
1904 read_config_file( $fh_config, $config_file, $rexpansion );
1905 die $death_message if ($death_message);
1907 # process any .perltidyrc parameters right now so we can
1909 if (@$rconfig_list) {
1910 local @ARGV = @$rconfig_list;
1912 expand_command_abbreviations( $rexpansion, \@raw_options,
1915 if ( !GetOptions( \%Opts, @$roption_string ) ) {
1917 "Error in this config file: $config_file \nUse -npro to ignore this file, -h for help'\n";
1920 # Anything left in this local @ARGV is an error and must be
1921 # invalid bare words from the configuration file. We cannot
1922 # check this earlier because bare words may have been valid
1923 # values for parameters. We had to wait for GetOptions to have
1927 my $str = "\'" . pop(@ARGV) . "\'";
1928 while ( my $param = pop(@ARGV) ) {
1929 if ( length($str) < 70 ) {
1930 $str .= ", '$param'";
1938 There are $count unrecognized values in the configuration file '$config_file':
1940 Use leading dashes for parameters. Use -npro to ignore this file.
1944 # Undo any options which cause premature exit. They are not
1945 # appropriate for a config file, and it could be hard to
1946 # diagnose the cause of the premature exit.
1955 dump-want-left-space
1956 dump-want-right-space
1964 if ( defined( $Opts{$_} ) ) {
1966 warn "ignoring --$_ in config file: $config_file\n";
1973 #---------------------------------------------------------------
1974 # now process the command line parameters
1975 #---------------------------------------------------------------
1976 expand_command_abbreviations( $rexpansion, \@raw_options, $config_file );
1978 if ( !GetOptions( \%Opts, @$roption_string ) ) {
1979 die "Error on command line; for help try 'perltidy -h'\n";
1982 return ( \%Opts, $config_file, \@raw_options, $saw_extrude, $roption_string,
1983 $rexpansion, $roption_category, $roption_range );
1984 } # end of process_command_line
1988 my ( $rOpts, $is_Windows, $Windows_type, $rpending_complaint ) = @_;
1990 #---------------------------------------------------------------
1991 # check and handle any interactions among the basic options..
1992 #---------------------------------------------------------------
1994 # Since -vt, -vtc, and -cti are abbreviations, but under
1995 # msdos, an unquoted input parameter like vtc=1 will be
1996 # seen as 2 parameters, vtc and 1, so the abbreviations
1997 # won't be seen. Therefore, we will catch them here if
2000 if ( defined $rOpts->{'vertical-tightness'} ) {
2001 my $vt = $rOpts->{'vertical-tightness'};
2002 $rOpts->{'paren-vertical-tightness'} = $vt;
2003 $rOpts->{'square-bracket-vertical-tightness'} = $vt;
2004 $rOpts->{'brace-vertical-tightness'} = $vt;
2007 if ( defined $rOpts->{'vertical-tightness-closing'} ) {
2008 my $vtc = $rOpts->{'vertical-tightness-closing'};
2009 $rOpts->{'paren-vertical-tightness-closing'} = $vtc;
2010 $rOpts->{'square-bracket-vertical-tightness-closing'} = $vtc;
2011 $rOpts->{'brace-vertical-tightness-closing'} = $vtc;
2014 if ( defined $rOpts->{'closing-token-indentation'} ) {
2015 my $cti = $rOpts->{'closing-token-indentation'};
2016 $rOpts->{'closing-square-bracket-indentation'} = $cti;
2017 $rOpts->{'closing-brace-indentation'} = $cti;
2018 $rOpts->{'closing-paren-indentation'} = $cti;
2021 # In quiet mode, there is no log file and hence no way to report
2022 # results of syntax check, so don't do it.
2023 if ( $rOpts->{'quiet'} ) {
2024 $rOpts->{'check-syntax'} = 0;
2027 # can't check syntax if no output
2028 if ( $rOpts->{'format'} ne 'tidy' ) {
2029 $rOpts->{'check-syntax'} = 0;
2032 # Never let Windows 9x/Me systems run syntax check -- this will prevent a
2033 # wide variety of nasty problems on these systems, because they cannot
2034 # reliably run backticks. Don't even think about changing this!
2035 if ( $rOpts->{'check-syntax'}
2037 && ( !$Windows_type || $Windows_type =~ /^(9|Me)/ ) )
2039 $rOpts->{'check-syntax'} = 0;
2042 # It's really a bad idea to check syntax as root unless you wrote
2043 # the script yourself. FIXME: not sure if this works with VMS
2044 unless ($is_Windows) {
2046 if ( $< == 0 && $rOpts->{'check-syntax'} ) {
2047 $rOpts->{'check-syntax'} = 0;
2048 $$rpending_complaint .=
2049 "Syntax check deactivated for safety; you shouldn't run this as root\n";
2053 # see if user set a non-negative logfile-gap
2054 if ( defined( $rOpts->{'logfile-gap'} ) && $rOpts->{'logfile-gap'} >= 0 ) {
2056 # a zero gap will be taken as a 1
2057 if ( $rOpts->{'logfile-gap'} == 0 ) {
2058 $rOpts->{'logfile-gap'} = 1;
2061 # setting a non-negative logfile gap causes logfile to be saved
2062 $rOpts->{'logfile'} = 1;
2065 # not setting logfile gap, or setting it negative, causes default of 50
2067 $rOpts->{'logfile-gap'} = 50;
2070 # set short-cut flag when only indentation is to be done.
2071 # Note that the user may or may not have already set the
2073 if ( !$rOpts->{'add-whitespace'}
2074 && !$rOpts->{'delete-old-whitespace'}
2075 && !$rOpts->{'add-newlines'}
2076 && !$rOpts->{'delete-old-newlines'} )
2078 $rOpts->{'indent-only'} = 1;
2081 # -isbc implies -ibc
2082 if ( $rOpts->{'indent-spaced-block-comments'} ) {
2083 $rOpts->{'indent-block-comments'} = 1;
2086 # -bli flag implies -bl
2087 if ( $rOpts->{'brace-left-and-indent'} ) {
2088 $rOpts->{'opening-brace-on-new-line'} = 1;
2091 if ( $rOpts->{'opening-brace-always-on-right'}
2092 && $rOpts->{'opening-brace-on-new-line'} )
2095 Conflict: you specified both 'opening-brace-always-on-right' (-bar) and
2096 'opening-brace-on-new-line' (-bl). Ignoring -bl.
2098 $rOpts->{'opening-brace-on-new-line'} = 0;
2101 # it simplifies things if -bl is 0 rather than undefined
2102 if ( !defined( $rOpts->{'opening-brace-on-new-line'} ) ) {
2103 $rOpts->{'opening-brace-on-new-line'} = 0;
2106 # -sbl defaults to -bl if not defined
2107 if ( !defined( $rOpts->{'opening-sub-brace-on-new-line'} ) ) {
2108 $rOpts->{'opening-sub-brace-on-new-line'} =
2109 $rOpts->{'opening-brace-on-new-line'};
2112 # set shortcut flag if no blanks to be written
2113 unless ( $rOpts->{'maximum-consecutive-blank-lines'} ) {
2114 $rOpts->{'swallow-optional-blank-lines'} = 1;
2117 if ( $rOpts->{'entab-leading-whitespace'} ) {
2118 if ( $rOpts->{'entab-leading-whitespace'} < 0 ) {
2119 warn "-et=n must use a positive integer; ignoring -et\n";
2120 $rOpts->{'entab-leading-whitespace'} = undef;
2123 # entab leading whitespace has priority over the older 'tabs' option
2124 if ( $rOpts->{'tabs'} ) { $rOpts->{'tabs'} = 0; }
2128 sub expand_command_abbreviations {
2130 # go through @ARGV and expand any abbreviations
2132 my ( $rexpansion, $rraw_options, $config_file ) = @_;
2135 # set a pass limit to prevent an infinite loop;
2136 # 10 should be plenty, but it may be increased to allow deeply
2137 # nested expansions.
2138 my $max_passes = 10;
2141 # keep looping until all expansions have been converted into actual
2143 for ( my $pass_count = 0 ; $pass_count <= $max_passes ; $pass_count++ ) {
2145 my $abbrev_count = 0;
2147 # loop over each item in @ARGV..
2148 foreach $word (@ARGV) {
2150 # convert any leading 'no-' to just 'no'
2151 if ( $word =~ /^(-[-]?no)-(.*)/ ) { $word = $1 . $2 }
2153 # if it is a dash flag (instead of a file name)..
2154 if ( $word =~ /^-[-]?([\w\-]+)(.*)/ ) {
2159 # save the raw input for debug output in case of circular refs
2160 if ( $pass_count == 0 ) {
2161 push( @$rraw_options, $word );
2164 # recombine abbreviation and flag, if necessary,
2165 # to allow abbreviations with arguments such as '-vt=1'
2166 if ( $rexpansion->{ $abr . $flags } ) {
2167 $abr = $abr . $flags;
2171 # if we see this dash item in the expansion hash..
2172 if ( $rexpansion->{$abr} ) {
2175 # stuff all of the words that it expands to into the
2176 # new arg list for the next pass
2177 foreach my $abbrev ( @{ $rexpansion->{$abr} } ) {
2178 next unless $abbrev; # for safety; shouldn't happen
2179 push( @new_argv, '--' . $abbrev . $flags );
2183 # not in expansion hash, must be actual long name
2185 push( @new_argv, $word );
2189 # not a dash item, so just save it for the next pass
2191 push( @new_argv, $word );
2193 } # end of this pass
2195 # update parameter list @ARGV to the new one
2197 last unless ( $abbrev_count > 0 );
2199 # make sure we are not in an infinite loop
2200 if ( $pass_count == $max_passes ) {
2202 "I'm tired. We seem to be in an infinite loop trying to expand aliases.\n";
2203 print STDERR "Here are the raw options\n";
2205 print STDERR "(@$rraw_options)\n";
2206 my $num = @new_argv;
2209 print STDERR "After $max_passes passes here is ARGV\n";
2210 print STDERR "(@new_argv)\n";
2213 print STDERR "After $max_passes passes ARGV has $num entries\n";
2218 Please check your configuration file $config_file for circular-references.
2219 To deactivate it, use -npro.
2224 Program bug - circular-references in the %expansion hash, probably due to
2225 a recent program change.
2228 } # end of check for circular references
2229 } # end of loop over all passes
2232 # Debug routine -- this will dump the expansion hash
2233 sub dump_short_names {
2234 my $rexpansion = shift;
2236 List of short names. This list shows how all abbreviations are
2237 translated into other abbreviations and, eventually, into long names.
2238 New abbreviations may be defined in a .perltidyrc file.
2239 For a list of all long names, use perltidy --dump-long-names (-dln).
2240 --------------------------------------------------------------------------
2242 foreach my $abbrev ( sort keys %$rexpansion ) {
2243 my @list = @{ $$rexpansion{$abbrev} };
2244 print STDOUT "$abbrev --> @list\n";
2248 sub check_vms_filename {
2250 # given a valid filename (the perltidy input file)
2251 # create a modified filename and separator character
2254 # Contributed by Michael Cartmell
2256 my ( $base, $path ) = fileparse( $_[0] );
2258 # remove explicit ; version
2259 $base =~ s/;-?\d*$//
2261 # remove explicit . version ie two dots in filename NB ^ escapes a dot
2262 or $base =~ s/( # begin capture $1
2263 (?:^|[^^])\. # match a dot not preceded by a caret
2264 (?: # followed by nothing
2266 .*[^^] # anything ending in a non caret
2269 \.-?\d*$ # match . version number
2272 # normalise filename, if there are no unescaped dots then append one
2273 $base .= '.' unless $base =~ /(?:^|[^^])\./;
2275 # if we don't already have an extension then we just append the extention
2276 my $separator = ( $base =~ /\.$/ ) ? "" : "_";
2277 return ( $path . $base, $separator );
2282 # TODO: are these more standard names?
2283 # Win32s Win95 Win98 WinMe WinNT3.51 WinNT4 Win2000 WinXP/.Net Win2003
2285 # Returns a string that determines what MS OS we are on.
2286 # Returns win32s,95,98,Me,NT3.51,NT4,2000,XP/.Net,Win2003
2287 # Returns blank string if not an MS system.
2288 # Original code contributed by: Yves Orton
2289 # We need to know this to decide where to look for config files
2291 my $rpending_complaint = shift;
2293 return $os unless $^O =~ /win32|dos/i; # is it a MS box?
2295 # Systems built from Perl source may not have Win32.pm
2296 # But probably have Win32::GetOSVersion() anyway so the
2297 # following line is not 'required':
2298 # return $os unless eval('require Win32');
2300 # Use the standard API call to determine the version
2301 my ( $undef, $major, $minor, $build, $id );
2302 eval { ( $undef, $major, $minor, $build, $id ) = Win32::GetOSVersion() };
2305 # NAME ID MAJOR MINOR
2306 # Windows NT 4 2 4 0
2307 # Windows 2000 2 5 0
2309 # Windows Server 2003 2 5 2
2311 return "win32s" unless $id; # If id==0 then its a win32s box.
2312 $os = { # Magic numbers from MSDN
2313 # documentation of GetOSVersion
2320 0 => "2000", # or NT 4, see below
2327 # If $os is undefined, the above code is out of date. Suggested updates
2329 unless ( defined $os ) {
2331 $$rpending_complaint .= <<EOS;
2332 Error trying to discover Win_OS_Type: $id:$major:$minor Has no name of record!
2333 We won't be able to look for a system-wide config file.
2337 # Unfortunately the logic used for the various versions isnt so clever..
2338 # so we have to handle an outside case.
2339 return ( $os eq "2000" && $major != 5 ) ? "NT4" : $os;
2344 ( $^O !~ /win32|dos/i )
2347 && ( $^O ne 'MacOS' );
2350 sub look_for_Windows {
2352 # determine Windows sub-type and location of
2353 # system-wide configuration files
2354 my $rpending_complaint = shift;
2355 my $is_Windows = ( $^O =~ /win32|dos/i );
2356 my $Windows_type = Win_OS_Type($rpending_complaint) if $is_Windows;
2357 return ( $is_Windows, $Windows_type );
2360 sub find_config_file {
2362 # look for a .perltidyrc configuration file
2363 my ( $is_Windows, $Windows_type, $rconfig_file_chatter,
2364 $rpending_complaint ) = @_;
2366 $$rconfig_file_chatter .= "# Config file search...system reported as:";
2368 $$rconfig_file_chatter .= "Windows $Windows_type\n";
2371 $$rconfig_file_chatter .= " $^O\n";
2374 # sub to check file existance and record all tests
2375 my $exists_config_file = sub {
2376 my $config_file = shift;
2377 return 0 unless $config_file;
2378 $$rconfig_file_chatter .= "# Testing: $config_file\n";
2379 return -f $config_file;
2384 # look in current directory first
2385 $config_file = ".perltidyrc";
2386 return $config_file if $exists_config_file->($config_file);
2388 # Default environment vars.
2389 my @envs = qw(PERLTIDY HOME);
2391 # Check the NT/2k/XP locations, first a local machine def, then a
2393 push @envs, qw(USERPROFILE HOMESHARE) if $^O =~ /win32/i;
2395 # Now go through the enviornment ...
2396 foreach my $var (@envs) {
2397 $$rconfig_file_chatter .= "# Examining: \$ENV{$var}";
2398 if ( defined( $ENV{$var} ) ) {
2399 $$rconfig_file_chatter .= " = $ENV{$var}\n";
2401 # test ENV{ PERLTIDY } as file:
2402 if ( $var eq 'PERLTIDY' ) {
2403 $config_file = "$ENV{$var}";
2404 return $config_file if $exists_config_file->($config_file);
2407 # test ENV as directory:
2408 $config_file = catfile( $ENV{$var}, ".perltidyrc" );
2409 return $config_file if $exists_config_file->($config_file);
2412 $$rconfig_file_chatter .= "\n";
2416 # then look for a system-wide definition
2417 # where to look varies with OS
2420 if ($Windows_type) {
2421 my ( $os, $system, $allusers ) =
2422 Win_Config_Locs( $rpending_complaint, $Windows_type );
2424 # Check All Users directory, if there is one.
2426 $config_file = catfile( $allusers, ".perltidyrc" );
2427 return $config_file if $exists_config_file->($config_file);
2430 # Check system directory.
2431 $config_file = catfile( $system, ".perltidyrc" );
2432 return $config_file if $exists_config_file->($config_file);
2436 # Place to add customization code for other systems
2437 elsif ( $^O eq 'OS2' ) {
2439 elsif ( $^O eq 'MacOS' ) {
2441 elsif ( $^O eq 'VMS' ) {
2444 # Assume some kind of Unix
2447 $config_file = "/usr/local/etc/perltidyrc";
2448 return $config_file if $exists_config_file->($config_file);
2450 $config_file = "/etc/perltidyrc";
2451 return $config_file if $exists_config_file->($config_file);
2454 # Couldn't find a config file
2458 sub Win_Config_Locs {
2460 # In scalar context returns the OS name (95 98 ME NT3.51 NT4 2000 XP),
2461 # or undef if its not a win32 OS. In list context returns OS, System
2462 # Directory, and All Users Directory. All Users will be empty on a
2463 # 9x/Me box. Contributed by: Yves Orton.
2465 my $rpending_complaint = shift;
2466 my $os = (@_) ? shift : Win_OS_Type();
2472 if ( $os =~ /9[58]|Me/ ) {
2473 $system = "C:/Windows";
2475 elsif ( $os =~ /NT|XP|200?/ ) {
2476 $system = ( $os =~ /XP/ ) ? "C:/Windows/" : "C:/WinNT/";
2479 ? "C:/WinNT/profiles/All Users/"
2480 : "C:/Documents and Settings/All Users/";
2484 # This currently would only happen on a win32s computer. I dont have
2485 # one to test, so I am unsure how to proceed. Suggestions welcome!
2486 $$rpending_complaint .=
2487 "I dont know a sensible place to look for config files on an $os system.\n";
2490 return wantarray ? ( $os, $system, $allusers ) : $os;
2493 sub dump_config_file {
2495 my $config_file = shift;
2496 my $rconfig_file_chatter = shift;
2497 print STDOUT "$$rconfig_file_chatter";
2499 print STDOUT "# Dump of file: '$config_file'\n";
2500 while ( my $line = $fh->getline() ) { print STDOUT $line }
2501 eval { $fh->close() };
2504 print STDOUT "# ...no config file found\n";
2508 sub read_config_file {
2510 my ( $fh, $config_file, $rexpansion ) = @_;
2511 my @config_list = ();
2513 # file is bad if non-empty $death_message is returned
2514 my $death_message = "";
2518 while ( my $line = $fh->getline() ) {
2521 next if $line =~ /^\s*#/; # skip full-line comment
2522 ( $line, $death_message ) =
2523 strip_comment( $line, $config_file, $line_no );
2524 last if ($death_message);
2525 $line =~ s/^\s*(.*?)\s*$/$1/; # trim both ends
2528 # look for something of the general form
2533 if ( $line =~ /^((\w+)\s*\{)?([^}]*)(\})?$/ ) {
2534 my ( $newname, $body, $curly ) = ( $2, $3, $4 );
2536 # handle a new alias definition
2540 "No '}' seen after $name and before $newname in config file $config_file line $.\n";
2545 if ( ${$rexpansion}{$name} ) {
2547 my @names = sort keys %$rexpansion;
2549 "Here is a list of all installed aliases\n(@names)\n"
2550 . "Attempting to redefine alias ($name) in config file $config_file line $.\n";
2553 ${$rexpansion}{$name} = [];
2559 my ( $rbody_parts, $msg ) = parse_args($body);
2561 $death_message = <<EOM;
2562 Error reading file '$config_file' at line number $line_no.
2564 Please fix this line or use -npro to avoid reading this file
2571 # remove leading dashes if this is an alias
2572 foreach (@$rbody_parts) { s/^\-+//; }
2573 push @{ ${$rexpansion}{$name} }, @$rbody_parts;
2576 push( @config_list, @$rbody_parts );
2583 "Unexpected '}' seen in config file $config_file line $.\n";
2590 eval { $fh->close() };
2591 return ( \@config_list, $death_message );
2596 my ( $instr, $config_file, $line_no ) = @_;
2599 # nothing to do if no comments
2600 if ( $instr !~ /#/ ) {
2601 return ( $instr, $msg );
2604 # use simple method of no quotes
2605 elsif ( $instr !~ /['"]/ ) {
2606 $instr =~ s/\s*\#.*$//; # simple trim
2607 return ( $instr, $msg );
2610 # handle comments and quotes
2612 my $quote_char = "";
2615 # looking for ending quote character
2617 if ( $instr =~ /\G($quote_char)/gc ) {
2621 elsif ( $instr =~ /\G(.)/gc ) {
2625 # error..we reached the end without seeing the ending quote char
2628 Error reading file $config_file at line number $line_no.
2629 Did not see ending quote character <$quote_char> in this text:
2631 Please fix this line or use -npro to avoid reading this file
2637 # accumulating characters and looking for start of a quoted string
2639 if ( $instr =~ /\G([\"\'])/gc ) {
2643 elsif ( $instr =~ /\G#/gc ) {
2646 elsif ( $instr =~ /\G(.)/gc ) {
2654 return ( $outstr, $msg );
2659 # Parse a command string containing multiple string with possible
2660 # quotes, into individual commands. It might look like this, for example:
2662 # -wba=" + - " -some-thing -wbb='. && ||'
2664 # There is no need, at present, to handle escaped quote characters.
2665 # (They are not perltidy tokens, so needn't be in strings).
2668 my @body_parts = ();
2669 my $quote_char = "";
2674 # looking for ending quote character
2676 if ( $body =~ /\G($quote_char)/gc ) {
2679 elsif ( $body =~ /\G(.)/gc ) {
2683 # error..we reached the end without seeing the ending quote char
2685 if ( length($part) ) { push @body_parts, $part; }
2687 Did not see ending quote character <$quote_char> in this text:
2694 # accumulating characters and looking for start of a quoted string
2696 if ( $body =~ /\G([\"\'])/gc ) {
2699 elsif ( $body =~ /\G(\s+)/gc ) {
2700 if ( length($part) ) { push @body_parts, $part; }
2703 elsif ( $body =~ /\G(.)/gc ) {
2707 if ( length($part) ) { push @body_parts, $part; }
2712 return ( \@body_parts, $msg );
2715 sub dump_long_names {
2717 my @names = sort @_;
2719 # Command line long names (passed to GetOptions)
2720 #---------------------------------------------------------------
2721 # here is a summary of the Getopt codes:
2722 # <none> does not take an argument
2723 # =s takes a mandatory string
2724 # :s takes an optional string
2725 # =i takes a mandatory integer
2726 # :i takes an optional integer
2727 # ! does not take an argument and may be negated
2728 # i.e., -foo and -nofoo are allowed
2729 # a double dash signals the end of the options list
2731 #---------------------------------------------------------------
2734 foreach (@names) { print STDOUT "$_\n" }
2738 my @defaults = sort @_;
2739 print STDOUT "Default command line options:\n";
2740 foreach (@_) { print STDOUT "$_\n" }
2743 sub readable_options {
2745 # return options for this run as a string which could be
2746 # put in a perltidyrc file
2747 my ( $rOpts, $roption_string ) = @_;
2749 my $rGetopt_flags = \%Getopt_flags;
2750 my $readable_options = "# Final parameter set for this run.\n";
2751 $readable_options .=
2752 "# See utility 'perltidyrc_dump.pl' for nicer formatting.\n";
2753 foreach my $opt ( @{$roption_string} ) {
2755 if ( $opt =~ /(.*)(!|=.*)$/ ) {
2759 if ( defined( $rOpts->{$opt} ) ) {
2760 $rGetopt_flags->{$opt} = $flag;
2763 foreach my $key ( sort keys %{$rOpts} ) {
2764 my $flag = $rGetopt_flags->{$key};
2765 my $value = $rOpts->{$key};
2769 if ( $flag =~ /^=/ ) {
2770 if ( $value !~ /^\d+$/ ) { $value = '"' . $value . '"' }
2771 $suffix = "=" . $value;
2773 elsif ( $flag =~ /^!/ ) {
2774 $prefix .= "no" unless ($value);
2779 $readable_options .=
2780 "# ERROR in dump_options: unrecognized flag $flag for $key\n";
2783 $readable_options .= $prefix . $key . $suffix . "\n";
2785 return $readable_options;
2790 This is perltidy, v$VERSION
2792 Copyright 2000-2007, Steve Hancock
2794 Perltidy is free software and may be copied under the terms of the GNU
2795 General Public License, which is included in the distribution files.
2797 Complete documentation for perltidy can be found using 'man perltidy'
2798 or on the internet at http://perltidy.sourceforge.net.
2805 This is perltidy version $VERSION, a perl script indenter. Usage:
2807 perltidy [ options ] file1 file2 file3 ...
2808 (output goes to file1.tdy, file2.tdy, file3.tdy, ...)
2809 perltidy [ options ] file1 -o outfile
2810 perltidy [ options ] file1 -st >outfile
2811 perltidy [ options ] <infile >outfile
2813 Options have short and long forms. Short forms are shown; see
2814 man pages for long forms. Note: '=s' indicates a required string,
2815 and '=n' indicates a required integer.
2819 -o=file name of the output file (only if single input file)
2820 -oext=s change output extension from 'tdy' to s
2821 -opath=path change path to be 'path' for output files
2822 -b backup original to .bak and modify file in-place
2823 -bext=s change default backup extension from 'bak' to s
2824 -q deactivate error messages (for running under editor)
2825 -w include non-critical warning messages in the .ERR error output
2826 -syn run perl -c to check syntax (default under unix systems)
2827 -log save .LOG file, which has useful diagnostics
2828 -f force perltidy to read a binary file
2829 -g like -log but writes more detailed .LOG file, for debugging scripts
2830 -opt write the set of options actually used to a .LOG file
2831 -npro ignore .perltidyrc configuration command file
2832 -pro=file read configuration commands from file instead of .perltidyrc
2833 -st send output to standard output, STDOUT
2834 -se send error output to standard error output, STDERR
2835 -v display version number to standard output and quit
2838 -i=n use n columns per indentation level (default n=4)
2839 -t tabs: use one tab character per indentation level, not recommeded
2840 -nt no tabs: use n spaces per indentation level (default)
2841 -et=n entab leading whitespace n spaces per tab; not recommended
2842 -io "indent only": just do indentation, no other formatting.
2843 -sil=n set starting indentation level to n; use if auto detection fails
2844 -ole=s specify output line ending (s=dos or win, mac, unix)
2845 -ple keep output line endings same as input (input must be filename)
2848 -fws freeze whitespace; this disables all whitespace changes
2849 and disables the following switches:
2850 -bt=n sets brace tightness, n= (0 = loose, 1=default, 2 = tight)
2851 -bbt same as -bt but for code block braces; same as -bt if not given
2852 -bbvt block braces vertically tight; use with -bl or -bli
2853 -bbvtl=s make -bbvt to apply to selected list of block types
2854 -pt=n paren tightness (n=0, 1 or 2)
2855 -sbt=n square bracket tightness (n=0, 1, or 2)
2856 -bvt=n brace vertical tightness,
2857 n=(0=open, 1=close unless multiple steps on a line, 2=always close)
2858 -pvt=n paren vertical tightness (see -bvt for n)
2859 -sbvt=n square bracket vertical tightness (see -bvt for n)
2860 -bvtc=n closing brace vertical tightness:
2861 n=(0=open, 1=sometimes close, 2=always close)
2862 -pvtc=n closing paren vertical tightness, see -bvtc for n.
2863 -sbvtc=n closing square bracket vertical tightness, see -bvtc for n.
2864 -ci=n sets continuation indentation=n, default is n=2 spaces
2865 -lp line up parentheses, brackets, and non-BLOCK braces
2866 -sfs add space before semicolon in for( ; ; )
2867 -aws allow perltidy to add whitespace (default)
2868 -dws delete all old non-essential whitespace
2869 -icb indent closing brace of a code block
2870 -cti=n closing indentation of paren, square bracket, or non-block brace:
2871 n=0 none, =1 align with opening, =2 one full indentation level
2872 -icp equivalent to -cti=2
2873 -wls=s want space left of tokens in string; i.e. -nwls='+ - * /'
2874 -wrs=s want space right of tokens in string;
2875 -sts put space before terminal semicolon of a statement
2876 -sak=s put space between keywords given in s and '(';
2877 -nsak=s no space between keywords in s and '('; i.e. -nsak='my our local'
2880 -fnl freeze newlines; this disables all line break changes
2881 and disables the following switches:
2882 -anl add newlines; ok to introduce new line breaks
2883 -bbs add blank line before subs and packages
2884 -bbc add blank line before block comments
2885 -bbb add blank line between major blocks
2886 -sob swallow optional blank lines
2887 -ce cuddled else; use this style: '} else {'
2888 -dnl delete old newlines (default)
2889 -mbl=n maximum consecutive blank lines (default=1)
2890 -l=n maximum line length; default n=80
2891 -bl opening brace on new line
2892 -sbl opening sub brace on new line. value of -bl is used if not given.
2893 -bli opening brace on new line and indented
2894 -bar opening brace always on right, even for long clauses
2895 -vt=n vertical tightness (requires -lp); n controls break after opening
2896 token: 0=never 1=no break if next line balanced 2=no break
2897 -vtc=n vertical tightness of closing container; n controls if closing
2898 token starts new line: 0=always 1=not unless list 1=never
2899 -wba=s want break after tokens in string; i.e. wba=': .'
2900 -wbb=s want break before tokens in string
2902 Following Old Breakpoints
2903 -kis keep interior semicolons. Allows multiple statements per line.
2904 -boc break at old comma breaks: turns off all automatic list formatting
2905 -bol break at old logical breakpoints: or, and, ||, && (default)
2906 -bok break at old list keyword breakpoints such as map, sort (default)
2907 -bot break at old conditional (ternary ?:) operator breakpoints (default)
2908 -cab=n break at commas after a comma-arrow (=>):
2909 n=0 break at all commas after =>
2910 n=1 stable: break unless this breaks an existing one-line container
2911 n=2 break only if a one-line container cannot be formed
2912 n=3 do not treat commas after => specially at all
2915 -ibc indent block comments (default)
2916 -isbc indent spaced block comments; may indent unless no leading space
2917 -msc=n minimum desired spaces to side comment, default 4
2918 -fpsc=n fix position for side comments; default 0;
2919 -csc add or update closing side comments after closing BLOCK brace
2920 -dcsc delete closing side comments created by a -csc command
2921 -cscp=s change closing side comment prefix to be other than '## end'
2922 -cscl=s change closing side comment to apply to selected list of blocks
2923 -csci=n minimum number of lines needed to apply a -csc tag, default n=6
2924 -csct=n maximum number of columns of appended text, default n=20
2925 -cscw causes warning if old side comment is overwritten with -csc
2927 -sbc use 'static block comments' identified by leading '##' (default)
2928 -sbcp=s change static block comment identifier to be other than '##'
2929 -osbc outdent static block comments
2931 -ssc use 'static side comments' identified by leading '##' (default)
2932 -sscp=s change static side comment identifier to be other than '##'
2934 Delete selected text
2935 -dac delete all comments AND pod
2936 -dbc delete block comments
2937 -dsc delete side comments
2940 Send selected text to a '.TEE' file
2941 -tac tee all comments AND pod
2942 -tbc tee block comments
2943 -tsc tee side comments
2947 -olq outdent long quoted strings (default)
2948 -olc outdent a long block comment line
2949 -ola outdent statement labels
2950 -okw outdent control keywords (redo, next, last, goto, return)
2951 -okwl=s specify alternative keywords for -okw command
2954 -mft=n maximum fields per table; default n=40
2955 -x do not format lines before hash-bang line (i.e., for VMS)
2956 -asc allows perltidy to add a ';' when missing (default)
2957 -dsm allows perltidy to delete an unnecessary ';' (default)
2959 Combinations of other parameters
2960 -gnu attempt to follow GNU Coding Standards as applied to perl
2961 -mangle remove as many newlines as possible (but keep comments and pods)
2962 -extrude insert as many newlines as possible
2964 Dump and die, debugging
2965 -dop dump options used in this run to standard output and quit
2966 -ddf dump default options to standard output and quit
2967 -dsn dump all option short names to standard output and quit
2968 -dln dump option long names to standard output and quit
2969 -dpro dump whatever configuration file is in effect to standard output
2970 -dtt dump all token types to standard output and quit
2973 -html write an html file (see 'man perl2web' for many options)
2974 Note: when -html is used, no indentation or formatting are done.
2975 Hint: try perltidy -html -css=mystyle.css filename.pl
2976 and edit mystyle.css to change the appearance of filename.html.
2977 -nnn gives line numbers
2978 -pre only writes out <pre>..</pre> code section
2979 -toc places a table of contents to subs at the top (default)
2980 -pod passes pod text through pod2html (default)
2981 -frm write html as a frame (3 files)
2982 -text=s extra extension for table of contents if -frm, default='toc'
2983 -sext=s extra extension for file content if -frm, default='src'
2985 A prefix of "n" negates short form toggle switches, and a prefix of "no"
2986 negates the long forms. For example, -nasc means don't add missing
2989 If you are unable to see this entire text, try "perltidy -h | more"
2990 For more detailed information, and additional options, try "man perltidy",
2991 or go to the perltidy home page at http://perltidy.sourceforge.net
2996 sub process_this_file {
2998 my ( $truth, $beauty ) = @_;
3000 # loop to process each line of this file
3001 while ( my $line_of_tokens = $truth->get_line() ) {
3002 $beauty->write_line($line_of_tokens);
3006 eval { $beauty->finish_formatting() };
3007 $truth->report_tokenization_errors();
3012 # Use 'perl -c' to make sure that we did not create bad syntax
3013 # This is a very good independent check for programming errors
3015 # Given names of the input and output files, ($ifname, $ofname),
3016 # we do the following:
3017 # - check syntax of the input file
3018 # - if bad, all done (could be an incomplete code snippet)
3019 # - if infile syntax ok, then check syntax of the output file;
3020 # - if outfile syntax bad, issue warning; this implies a code bug!
3021 # - set and return flag "infile_syntax_ok" : =-1 bad 0 unknown 1 good
3023 my ( $ifname, $ofname, $logger_object, $rOpts ) = @_;
3024 my $infile_syntax_ok = 0;
3025 my $line_of_dashes = '-' x 42 . "\n";
3027 my $flags = $rOpts->{'perl-syntax-check-flags'};
3029 # be sure we invoke perl with -c
3030 # note: perl will accept repeated flags like '-c -c'. It is safest
3031 # to append another -c than try to find an interior bundled c, as
3032 # in -Tc, because such a 'c' might be in a quoted string, for example.
3033 if ( $flags !~ /(^-c|\s+-c)/ ) { $flags .= " -c" }
3035 # be sure we invoke perl with -x if requested
3036 # same comments about repeated parameters applies
3037 if ( $rOpts->{'look-for-hash-bang'} ) {
3038 if ( $flags !~ /(^-x|\s+-x)/ ) { $flags .= " -x" }
3041 # this shouldn't happen unless a termporary file couldn't be made
3042 if ( $ifname eq '-' ) {
3043 $logger_object->write_logfile_entry(
3044 "Cannot run perl -c on STDIN and STDOUT\n");
3045 return $infile_syntax_ok;
3048 $logger_object->write_logfile_entry(
3049 "checking input file syntax with perl $flags\n");
3050 $logger_object->write_logfile_entry($line_of_dashes);
3052 # Not all operating systems/shells support redirection of the standard
3054 my $error_redirection = ( $^O eq 'VMS' ) ? "" : '2>&1';
3056 my $perl_output = do_syntax_check( $ifname, $flags, $error_redirection );
3057 $logger_object->write_logfile_entry("$perl_output\n");
3059 if ( $perl_output =~ /syntax\s*OK/ ) {
3060 $infile_syntax_ok = 1;
3061 $logger_object->write_logfile_entry($line_of_dashes);
3062 $logger_object->write_logfile_entry(
3063 "checking output file syntax with perl $flags ...\n");
3064 $logger_object->write_logfile_entry($line_of_dashes);
3067 do_syntax_check( $ofname, $flags, $error_redirection );
3068 $logger_object->write_logfile_entry("$perl_output\n");
3070 unless ( $perl_output =~ /syntax\s*OK/ ) {
3071 $logger_object->write_logfile_entry($line_of_dashes);
3072 $logger_object->warning(
3073 "The output file has a syntax error when tested with perl $flags $ofname !\n"
3075 $logger_object->warning(
3076 "This implies an error in perltidy; the file $ofname is bad\n");
3077 $logger_object->report_definite_bug();
3079 # the perl version number will be helpful for diagnosing the problem
3080 $logger_object->write_logfile_entry(
3081 qx/perl -v $error_redirection/ . "\n" );
3086 # Only warn of perl -c syntax errors. Other messages,
3087 # such as missing modules, are too common. They can be
3088 # seen by running with perltidy -w
3089 $logger_object->complain("A syntax check using perl $flags gives: \n");
3090 $logger_object->complain($line_of_dashes);
3091 $logger_object->complain("$perl_output\n");
3092 $logger_object->complain($line_of_dashes);
3093 $infile_syntax_ok = -1;
3094 $logger_object->write_logfile_entry($line_of_dashes);
3095 $logger_object->write_logfile_entry(
3096 "The output file will not be checked because of input file problems\n"
3099 return $infile_syntax_ok;
3102 sub do_syntax_check {
3103 my ( $fname, $flags, $error_redirection ) = @_;
3105 # We have to quote the filename in case it has unusual characters
3106 # or spaces. Example: this filename #CM11.pm# gives trouble.
3107 $fname = '"' . $fname . '"';
3109 # Under VMS something like -T will become -t (and an error) so we
3110 # will put quotes around the flags. Double quotes seem to work on
3111 # Unix/Windows/VMS, but this may not work on all systems. (Single
3112 # quotes do not work under Windows). It could become necessary to
3113 # put double quotes around each flag, such as: -"c" -"T"
3114 # We may eventually need some system-dependent coding here.
3115 $flags = '"' . $flags . '"';
3117 # now wish for luck...
3118 return qx/perl $flags $fname $error_redirection/;
3121 #####################################################################
3123 # This is a stripped down version of IO::Scalar
3124 # Given a reference to a scalar, it supplies either:
3125 # a getline method which reads lines (mode='r'), or
3126 # a print method which reads lines (mode='w')
3128 #####################################################################
3129 package Perl::Tidy::IOScalar;
3133 my ( $package, $rscalar, $mode ) = @_;
3134 my $ref = ref $rscalar;
3135 if ( $ref ne 'SCALAR' ) {
3137 ------------------------------------------------------------------------
3138 expecting ref to SCALAR but got ref to ($ref); trace follows:
3139 ------------------------------------------------------------------------
3143 if ( $mode eq 'w' ) {
3145 return bless [ $rscalar, $mode ], $package;
3147 elsif ( $mode eq 'r' ) {
3149 # Convert a scalar to an array.
3150 # This avoids looking for "\n" on each call to getline
3151 my @array = map { $_ .= "\n" } split /\n/, ${$rscalar};
3153 return bless [ \@array, $mode, $i_next ], $package;
3157 ------------------------------------------------------------------------
3158 expecting mode = 'r' or 'w' but got mode ($mode); trace follows:
3159 ------------------------------------------------------------------------
3166 my $mode = $self->[1];
3167 if ( $mode ne 'r' ) {
3169 ------------------------------------------------------------------------
3170 getline call requires mode = 'r' but mode = ($mode); trace follows:
3171 ------------------------------------------------------------------------
3174 my $i = $self->[2]++;
3175 ##my $line = $self->[0]->[$i];
3176 return $self->[0]->[$i];
3181 my $mode = $self->[1];
3182 if ( $mode ne 'w' ) {
3184 ------------------------------------------------------------------------
3185 print call requires mode = 'w' but mode = ($mode); trace follows:
3186 ------------------------------------------------------------------------
3189 ${ $self->[0] } .= $_[0];
3191 sub close { return }
3193 #####################################################################
3195 # This is a stripped down version of IO::ScalarArray
3196 # Given a reference to an array, it supplies either:
3197 # a getline method which reads lines (mode='r'), or
3198 # a print method which reads lines (mode='w')
3200 # NOTE: this routine assumes that that there aren't any embedded
3201 # newlines within any of the array elements. There are no checks
3204 #####################################################################
3205 package Perl::Tidy::IOScalarArray;
3209 my ( $package, $rarray, $mode ) = @_;
3210 my $ref = ref $rarray;
3211 if ( $ref ne 'ARRAY' ) {
3213 ------------------------------------------------------------------------
3214 expecting ref to ARRAY but got ref to ($ref); trace follows:
3215 ------------------------------------------------------------------------
3219 if ( $mode eq 'w' ) {
3221 return bless [ $rarray, $mode ], $package;
3223 elsif ( $mode eq 'r' ) {
3225 return bless [ $rarray, $mode, $i_next ], $package;
3229 ------------------------------------------------------------------------
3230 expecting mode = 'r' or 'w' but got mode ($mode); trace follows:
3231 ------------------------------------------------------------------------
3238 my $mode = $self->[1];
3239 if ( $mode ne 'r' ) {
3241 ------------------------------------------------------------------------
3242 getline requires mode = 'r' but mode = ($mode); trace follows:
3243 ------------------------------------------------------------------------
3246 my $i = $self->[2]++;
3247 return $self->[0]->[$i];
3252 my $mode = $self->[1];
3253 if ( $mode ne 'w' ) {
3255 ------------------------------------------------------------------------
3256 print requires mode = 'w' but mode = ($mode); trace follows:
3257 ------------------------------------------------------------------------
3260 push @{ $self->[0] }, $_[0];
3262 sub close { return }
3264 #####################################################################
3266 # the Perl::Tidy::LineSource class supplies an object with a 'get_line()' method
3267 # which returns the next line to be parsed
3269 #####################################################################
3271 package Perl::Tidy::LineSource;
3275 my ( $class, $input_file, $rOpts, $rpending_logfile_message ) = @_;
3276 my $input_file_copy = undef;
3279 my $input_line_ending;
3280 if ( $rOpts->{'preserve-line-endings'} ) {
3281 $input_line_ending = Perl::Tidy::find_input_line_ending($input_file);
3284 ( my $fh, $input_file ) = Perl::Tidy::streamhandle( $input_file, 'r' );
3285 return undef unless $fh;
3287 # in order to check output syntax when standard output is used,
3288 # or when it is an object, we have to make a copy of the file
3289 if ( ( $input_file eq '-' || ref $input_file ) && $rOpts->{'check-syntax'} )
3292 # Turning off syntax check when input output is used.
3293 # The reason is that temporary files cause problems on
3295 $rOpts->{'check-syntax'} = 0;
3296 $input_file_copy = '-';
3298 $$rpending_logfile_message .= <<EOM;
3299 Note: --syntax check will be skipped because standard input is used
3306 _fh_copy => $fh_copy,
3307 _filename => $input_file,
3308 _input_file_copy => $input_file_copy,
3309 _input_line_ending => $input_line_ending,
3310 _rinput_buffer => [],
3315 sub get_input_file_copy_name {
3317 my $ifname = $self->{_input_file_copy};
3319 $ifname = $self->{_filename};
3324 sub close_input_file {
3326 eval { $self->{_fh}->close() };
3327 eval { $self->{_fh_copy}->close() } if $self->{_fh_copy};
3333 my $fh = $self->{_fh};
3334 my $fh_copy = $self->{_fh_copy};
3335 my $rinput_buffer = $self->{_rinput_buffer};
3337 if ( scalar(@$rinput_buffer) ) {
3338 $line = shift @$rinput_buffer;
3341 $line = $fh->getline();
3343 # patch to read raw mac files under unix, dos
3344 # see if the first line has embedded \r's
3345 if ( $line && !$self->{_started} ) {
3346 if ( $line =~ /[\015][^\015\012]/ ) {
3348 # found one -- break the line up and store in a buffer
3349 @$rinput_buffer = map { $_ . "\n" } split /\015/, $line;
3350 my $count = @$rinput_buffer;
3351 $line = shift @$rinput_buffer;
3353 $self->{_started}++;
3356 if ( $line && $fh_copy ) { $fh_copy->print($line); }
3360 #####################################################################
3362 # the Perl::Tidy::LineSink class supplies a write_line method for
3363 # actual file writing
3365 #####################################################################
3367 package Perl::Tidy::LineSink;
3371 my ( $class, $output_file, $tee_file, $line_separator, $rOpts,
3372 $rpending_logfile_message, $binmode )
3375 my $fh_copy = undef;
3377 my $output_file_copy = "";
3378 my $output_file_open = 0;
3380 if ( $rOpts->{'format'} eq 'tidy' ) {
3381 ( $fh, $output_file ) = Perl::Tidy::streamhandle( $output_file, 'w' );
3382 unless ($fh) { die "Cannot write to output stream\n"; }
3383 $output_file_open = 1;
3385 if ( ref($fh) eq 'IO::File' ) {
3388 if ( $output_file eq '-' ) { binmode STDOUT }
3392 # in order to check output syntax when standard output is used,
3393 # or when it is an object, we have to make a copy of the file
3394 if ( $output_file eq '-' || ref $output_file ) {
3395 if ( $rOpts->{'check-syntax'} ) {
3397 # Turning off syntax check when standard output is used.
3398 # The reason is that temporary files cause problems on
3400 $rOpts->{'check-syntax'} = 0;
3401 $output_file_copy = '-';
3402 $$rpending_logfile_message .= <<EOM;
3403 Note: --syntax check will be skipped because standard output is used
3411 _fh_copy => $fh_copy,
3413 _output_file => $output_file,
3414 _output_file_open => $output_file_open,
3415 _output_file_copy => $output_file_copy,
3417 _tee_file => $tee_file,
3418 _tee_file_opened => 0,
3419 _line_separator => $line_separator,
3420 _binmode => $binmode,
3427 my $fh = $self->{_fh};
3428 my $fh_copy = $self->{_fh_copy};
3430 my $output_file_open = $self->{_output_file_open};
3432 $_[0] .= $self->{_line_separator};
3434 $fh->print( $_[0] ) if ( $self->{_output_file_open} );
3435 print $fh_copy $_[0] if ( $fh_copy && $self->{_output_file_copy} );
3437 if ( $self->{_tee_flag} ) {
3438 unless ( $self->{_tee_file_opened} ) { $self->really_open_tee_file() }
3439 my $fh_tee = $self->{_fh_tee};
3440 print $fh_tee $_[0];
3444 sub get_output_file_copy {
3446 my $ofname = $self->{_output_file_copy};
3448 $ofname = $self->{_output_file};
3455 $self->{_tee_flag} = 1;
3460 $self->{_tee_flag} = 0;
3463 sub really_open_tee_file {
3465 my $tee_file = $self->{_tee_file};
3467 $fh_tee = IO::File->new(">$tee_file")
3468 or die("couldn't open TEE file $tee_file: $!\n");
3469 binmode $fh_tee if $self->{_binmode};
3470 $self->{_tee_file_opened} = 1;
3471 $self->{_fh_tee} = $fh_tee;
3474 sub close_output_file {
3476 eval { $self->{_fh}->close() } if $self->{_output_file_open};
3477 eval { $self->{_fh_copy}->close() } if ( $self->{_output_file_copy} );
3478 $self->close_tee_file();
3481 sub close_tee_file {
3484 if ( $self->{_tee_file_opened} ) {
3485 eval { $self->{_fh_tee}->close() };
3486 $self->{_tee_file_opened} = 0;
3490 #####################################################################
3492 # The Perl::Tidy::Diagnostics class writes the DIAGNOSTICS file, which is
3493 # useful for program development.
3495 # Only one such file is created regardless of the number of input
3496 # files processed. This allows the results of processing many files
3497 # to be summarized in a single file.
3499 #####################################################################
3501 package Perl::Tidy::Diagnostics;
3507 _write_diagnostics_count => 0,
3508 _last_diagnostic_file => "",
3514 sub set_input_file {
3516 $self->{_input_file} = $_[0];
3519 # This is a diagnostic routine which is useful for program development.
3520 # Output from debug messages go to a file named DIAGNOSTICS, where
3521 # they are labeled by file and line. This allows many files to be
3522 # scanned at once for some particular condition of interest.
3523 sub write_diagnostics {
3526 unless ( $self->{_write_diagnostics_count} ) {
3527 open DIAGNOSTICS, ">DIAGNOSTICS"
3528 or death("couldn't open DIAGNOSTICS: $!\n");
3531 my $last_diagnostic_file = $self->{_last_diagnostic_file};
3532 my $input_file = $self->{_input_file};
3533 if ( $last_diagnostic_file ne $input_file ) {
3534 print DIAGNOSTICS "\nFILE:$input_file\n";
3536 $self->{_last_diagnostic_file} = $input_file;
3537 my $input_line_number = Perl::Tidy::Tokenizer::get_input_line_number();
3538 print DIAGNOSTICS "$input_line_number:\t@_";
3539 $self->{_write_diagnostics_count}++;
3542 #####################################################################
3544 # The Perl::Tidy::Logger class writes the .LOG and .ERR files
3546 #####################################################################
3548 package Perl::Tidy::Logger;
3553 my ( $rOpts, $log_file, $warning_file, $saw_extrude ) = @_;
3555 # remove any old error output file
3556 unless ( ref($warning_file) ) {
3557 if ( -e $warning_file ) { unlink($warning_file) }
3561 _log_file => $log_file,
3562 _fh_warnings => undef,
3564 _fh_warnings => undef,
3565 _last_input_line_written => 0,
3566 _at_end_of_file => 0,
3568 _block_log_output => 0,
3569 _line_of_tokens => undef,
3570 _output_line_number => undef,
3571 _wrote_line_information_string => 0,
3572 _wrote_column_headings => 0,
3573 _warning_file => $warning_file,
3574 _warning_count => 0,
3575 _complaint_count => 0,
3576 _saw_code_bug => -1, # -1=no 0=maybe 1=for sure
3577 _saw_brace_error => 0,
3578 _saw_extrude => $saw_extrude,
3579 _output_array => [],
3583 sub close_log_file {
3586 if ( $self->{_fh_warnings} ) {
3587 eval { $self->{_fh_warnings}->close() };
3588 $self->{_fh_warnings} = undef;
3592 sub get_warning_count {
3594 return $self->{_warning_count};
3597 sub get_use_prefix {
3599 return $self->{_use_prefix};
3602 sub block_log_output {
3604 $self->{_block_log_output} = 1;
3607 sub unblock_log_output {
3609 $self->{_block_log_output} = 0;
3612 sub interrupt_logfile {
3614 $self->{_use_prefix} = 0;
3615 $self->warning("\n");
3616 $self->write_logfile_entry( '#' x 24 . " WARNING " . '#' x 25 . "\n" );
3619 sub resume_logfile {
3621 $self->write_logfile_entry( '#' x 60 . "\n" );
3622 $self->{_use_prefix} = 1;
3625 sub we_are_at_the_last_line {
3627 unless ( $self->{_wrote_line_information_string} ) {
3628 $self->write_logfile_entry("Last line\n\n");
3630 $self->{_at_end_of_file} = 1;
3633 # record some stuff in case we go down in flames
3636 my ( $line_of_tokens, $output_line_number ) = @_;
3637 my $input_line = $line_of_tokens->{_line_text};
3638 my $input_line_number = $line_of_tokens->{_line_number};
3640 # save line information in case we have to write a logfile message
3641 $self->{_line_of_tokens} = $line_of_tokens;
3642 $self->{_output_line_number} = $output_line_number;
3643 $self->{_wrote_line_information_string} = 0;
3645 my $last_input_line_written = $self->{_last_input_line_written};
3646 my $rOpts = $self->{_rOpts};
3649 ( $input_line_number - $last_input_line_written ) >=
3650 $rOpts->{'logfile-gap'}
3652 || ( $input_line =~ /^\s*(sub|package)\s+(\w+)/ )
3655 my $rlevels = $line_of_tokens->{_rlevels};
3656 my $structural_indentation_level = $$rlevels[0];
3657 $self->{_last_input_line_written} = $input_line_number;
3658 ( my $out_str = $input_line ) =~ s/^\s*//;
3661 $out_str = ( '.' x $structural_indentation_level ) . $out_str;
3663 if ( length($out_str) > 35 ) {
3664 $out_str = substr( $out_str, 0, 35 ) . " ....";
3666 $self->logfile_output( "", "$out_str\n" );
3670 sub write_logfile_entry {
3673 # add leading >>> to avoid confusing error mesages and code
3674 $self->logfile_output( ">>>", "@_" );
3677 sub write_column_headings {
3680 $self->{_wrote_column_headings} = 1;
3681 my $routput_array = $self->{_output_array};
3682 push @{$routput_array}, <<EOM;
3683 The nesting depths in the table below are at the start of the lines.
3684 The indicated output line numbers are not always exact.
3685 ci = levels of continuation indentation; bk = 1 if in BLOCK, 0 if not.
3687 in:out indent c b nesting code + messages; (messages begin with >>>)
3688 lines levels i k (code begins with one '.' per indent level)
3689 ------ ----- - - -------- -------------------------------------------
3693 sub make_line_information_string {
3695 # make columns of information when a logfile message needs to go out
3697 my $line_of_tokens = $self->{_line_of_tokens};
3698 my $input_line_number = $line_of_tokens->{_line_number};
3699 my $line_information_string = "";
3700 if ($input_line_number) {
3702 my $output_line_number = $self->{_output_line_number};
3703 my $brace_depth = $line_of_tokens->{_curly_brace_depth};
3704 my $paren_depth = $line_of_tokens->{_paren_depth};
3705 my $square_bracket_depth = $line_of_tokens->{_square_bracket_depth};
3706 my $python_indentation_level =
3707 $line_of_tokens->{_python_indentation_level};
3708 my $rlevels = $line_of_tokens->{_rlevels};
3709 my $rnesting_tokens = $line_of_tokens->{_rnesting_tokens};
3710 my $rci_levels = $line_of_tokens->{_rci_levels};
3711 my $rnesting_blocks = $line_of_tokens->{_rnesting_blocks};
3713 my $structural_indentation_level = $$rlevels[0];
3715 $self->write_column_headings() unless $self->{_wrote_column_headings};
3717 # keep logfile columns aligned for scripts up to 999 lines;
3718 # for longer scripts it doesn't really matter
3719 my $extra_space = "";
3721 ( $input_line_number < 10 ) ? " "
3722 : ( $input_line_number < 100 ) ? " "
3725 ( $output_line_number < 10 ) ? " "
3726 : ( $output_line_number < 100 ) ? " "
3729 # there are 2 possible nesting strings:
3730 # the original which looks like this: (0 [1 {2
3731 # the new one, which looks like this: {{[
3732 # the new one is easier to read, and shows the order, but
3733 # could be arbitrarily long, so we use it unless it is too long
3734 my $nesting_string =
3735 "($paren_depth [$square_bracket_depth {$brace_depth";
3736 my $nesting_string_new = $$rnesting_tokens[0];
3738 my $ci_level = $$rci_levels[0];
3739 if ( $ci_level > 9 ) { $ci_level = '*' }
3740 my $bk = ( $$rnesting_blocks[0] =~ /1$/ ) ? '1' : '0';
3742 if ( length($nesting_string_new) <= 8 ) {
3744 $nesting_string_new . " " x ( 8 - length($nesting_string_new) );
3746 if ( $python_indentation_level < 0 ) { $python_indentation_level = 0 }
3747 $line_information_string =
3748 "L$input_line_number:$output_line_number$extra_space i$python_indentation_level:$structural_indentation_level $ci_level $bk $nesting_string";
3750 return $line_information_string;
3753 sub logfile_output {
3755 my ( $prompt, $msg ) = @_;
3756 return if ( $self->{_block_log_output} );
3758 my $routput_array = $self->{_output_array};
3759 if ( $self->{_at_end_of_file} || !$self->{_use_prefix} ) {
3760 push @{$routput_array}, "$msg";
3763 my $line_information_string = $self->make_line_information_string();
3764 $self->{_wrote_line_information_string} = 1;
3766 if ($line_information_string) {
3767 push @{$routput_array}, "$line_information_string $prompt$msg";
3770 push @{$routput_array}, "$msg";
3775 sub get_saw_brace_error {
3777 return $self->{_saw_brace_error};
3780 sub increment_brace_error {
3782 $self->{_saw_brace_error}++;
3787 use constant BRACE_WARNING_LIMIT => 10;
3788 my $saw_brace_error = $self->{_saw_brace_error};
3790 if ( $saw_brace_error < BRACE_WARNING_LIMIT ) {
3794 $self->{_saw_brace_error} = $saw_brace_error;
3796 if ( $saw_brace_error == BRACE_WARNING_LIMIT ) {
3797 $self->warning("No further warnings of this type will be given\n");
3803 # handle non-critical warning messages based on input flag
3805 my $rOpts = $self->{_rOpts};
3807 # these appear in .ERR output only if -w flag is used
3808 if ( $rOpts->{'warning-output'} ) {
3812 # otherwise, they go to the .LOG file
3814 $self->{_complaint_count}++;
3815 $self->write_logfile_entry(@_);
3821 # report errors to .ERR file (or stdout)
3823 use constant WARNING_LIMIT => 50;
3825 my $rOpts = $self->{_rOpts};
3826 unless ( $rOpts->{'quiet'} ) {
3828 my $warning_count = $self->{_warning_count};
3829 unless ($warning_count) {
3830 my $warning_file = $self->{_warning_file};
3832 if ( $rOpts->{'standard-error-output'} ) {
3833 $fh_warnings = *STDERR;
3836 ( $fh_warnings, my $filename ) =
3837 Perl::Tidy::streamhandle( $warning_file, 'w' );
3838 $fh_warnings or die("couldn't open $filename $!\n");
3839 warn "## Please see file $filename\n";
3841 $self->{_fh_warnings} = $fh_warnings;
3844 my $fh_warnings = $self->{_fh_warnings};
3845 if ( $warning_count < WARNING_LIMIT ) {
3846 if ( $self->get_use_prefix() > 0 ) {
3847 my $input_line_number =
3848 Perl::Tidy::Tokenizer::get_input_line_number();
3849 $fh_warnings->print("$input_line_number:\t@_");
3850 $self->write_logfile_entry("WARNING: @_");
3853 $fh_warnings->print(@_);
3854 $self->write_logfile_entry(@_);
3858 $self->{_warning_count} = $warning_count;
3860 if ( $warning_count == WARNING_LIMIT ) {
3861 $fh_warnings->print("No further warnings will be given\n");
3866 # programming bug codes:
3868 # 0 = maybe, not sure.
3870 sub report_possible_bug {
3872 my $saw_code_bug = $self->{_saw_code_bug};
3873 $self->{_saw_code_bug} = ( $saw_code_bug < 0 ) ? 0 : $saw_code_bug;
3876 sub report_definite_bug {
3878 $self->{_saw_code_bug} = 1;
3881 sub ask_user_for_bug_report {
3884 my ( $infile_syntax_ok, $formatter ) = @_;
3885 my $saw_code_bug = $self->{_saw_code_bug};
3886 if ( ( $saw_code_bug == 0 ) && ( $infile_syntax_ok == 1 ) ) {
3887 $self->warning(<<EOM);
3889 You may have encountered a code bug in perltidy. If you think so, and
3890 the problem is not listed in the BUGS file at
3891 http://perltidy.sourceforge.net, please report it so that it can be
3892 corrected. Include the smallest possible script which has the problem,
3893 along with the .LOG file. See the manual pages for contact information.
3898 elsif ( $saw_code_bug == 1 ) {
3899 if ( $self->{_saw_extrude} ) {
3900 $self->warning(<<EOM);
3902 You may have encountered a bug in perltidy. However, since you are using the
3903 -extrude option, the problem may be with perl or one of its modules, which have
3904 occasional problems with this type of file. If you believe that the
3905 problem is with perltidy, and the problem is not listed in the BUGS file at
3906 http://perltidy.sourceforge.net, please report it so that it can be corrected.
3907 Include the smallest possible script which has the problem, along with the .LOG
3908 file. See the manual pages for contact information.
3913 $self->warning(<<EOM);
3915 Oops, you seem to have encountered a bug in perltidy. Please check the
3916 BUGS file at http://perltidy.sourceforge.net. If the problem is not
3917 listed there, please report it so that it can be corrected. Include the
3918 smallest possible script which produces this message, along with the
3919 .LOG file if appropriate. See the manual pages for contact information.
3920 Your efforts are appreciated.
3923 my $added_semicolon_count = 0;
3925 $added_semicolon_count =
3926 $formatter->get_added_semicolon_count();
3928 if ( $added_semicolon_count > 0 ) {
3929 $self->warning(<<EOM);
3931 The log file shows that perltidy added $added_semicolon_count semicolons.
3932 Please rerun with -nasc to see if that is the cause of the syntax error. Even
3933 if that is the problem, please report it so that it can be fixed.
3943 # called after all formatting to summarize errors
3945 my ( $infile_syntax_ok, $formatter ) = @_;
3947 my $rOpts = $self->{_rOpts};
3948 my $warning_count = $self->{_warning_count};
3949 my $saw_code_bug = $self->{_saw_code_bug};
3952 ( $saw_code_bug == 0 && $infile_syntax_ok == 1 )
3953 || $saw_code_bug == 1
3954 || $rOpts->{'logfile'};
3955 my $log_file = $self->{_log_file};
3956 if ($warning_count) {
3957 if ($save_logfile) {
3958 $self->block_log_output(); # avoid echoing this to the logfile
3960 "The logfile $log_file may contain useful information\n");
3961 $self->unblock_log_output();
3964 if ( $self->{_complaint_count} > 0 ) {
3966 "To see $self->{_complaint_count} non-critical warnings rerun with -w\n"
3970 if ( $self->{_saw_brace_error}
3971 && ( $rOpts->{'logfile-gap'} > 1 || !$save_logfile ) )
3973 $self->warning("To save a full .LOG file rerun with -g\n");
3976 $self->ask_user_for_bug_report( $infile_syntax_ok, $formatter );
3978 if ($save_logfile) {
3979 my $log_file = $self->{_log_file};
3980 my ( $fh, $filename ) = Perl::Tidy::streamhandle( $log_file, 'w' );
3982 my $routput_array = $self->{_output_array};
3983 foreach ( @{$routput_array} ) { $fh->print($_) }
3984 eval { $fh->close() };
3989 #####################################################################
3991 # The Perl::Tidy::DevNull class supplies a dummy print method
3993 #####################################################################
3995 package Perl::Tidy::DevNull;
3996 sub new { return bless {}, $_[0] }
3997 sub print { return }
3998 sub close { return }
4000 #####################################################################
4002 # The Perl::Tidy::HtmlWriter class writes a copy of the input stream in html
4004 #####################################################################
4006 package Perl::Tidy::HtmlWriter;
4016 %short_to_long_names
4020 $missing_html_entities
4023 # replace unsafe characters with HTML entity representation if HTML::Entities
4025 { eval "use HTML::Entities"; $missing_html_entities = $@; }
4029 my ( $class, $input_file, $html_file, $extension, $html_toc_extension,
4030 $html_src_extension )
4033 my $html_file_opened = 0;
4035 ( $html_fh, my $html_filename ) =
4036 Perl::Tidy::streamhandle( $html_file, 'w' );
4038 warn("can't open $html_file: $!\n");
4041 $html_file_opened = 1;
4043 if ( !$input_file || $input_file eq '-' || ref($input_file) ) {
4044 $input_file = "NONAME";
4047 # write the table of contents to a string
4049 my $html_toc_fh = Perl::Tidy::IOScalar->new( \$toc_string, 'w' );
4052 my @pre_string_stack;
4053 if ( $rOpts->{'html-pre-only'} ) {
4055 # pre section goes directly to the output stream
4056 $html_pre_fh = $html_fh;
4057 $html_pre_fh->print( <<"PRE_END");
4063 # pre section go out to a temporary string
4065 $html_pre_fh = Perl::Tidy::IOScalar->new( \$pre_string, 'w' );
4066 push @pre_string_stack, \$pre_string;
4069 # pod text gets diverted if the 'pod2html' is used
4072 if ( $rOpts->{'pod2html'} ) {
4073 if ( $rOpts->{'html-pre-only'} ) {
4074 undef $rOpts->{'pod2html'};
4077 eval "use Pod::Html";
4080 "unable to find Pod::Html; cannot use pod2html\n-npod disables this message\n";
4081 undef $rOpts->{'pod2html'};
4084 $html_pod_fh = Perl::Tidy::IOScalar->new( \$pod_string, 'w' );
4091 if ( $rOpts->{'frames'} ) {
4092 unless ($extension) {
4094 "cannot use frames without a specified output extension; ignoring -frm\n";
4095 undef $rOpts->{'frames'};
4098 $toc_filename = $input_file . $html_toc_extension . $extension;
4099 $src_filename = $input_file . $html_src_extension . $extension;
4103 # ----------------------------------------------------------
4104 # Output is now directed as follows:
4105 # html_toc_fh <-- table of contents items
4106 # html_pre_fh <-- the <pre> section of formatted code, except:
4107 # html_pod_fh <-- pod goes here with the pod2html option
4108 # ----------------------------------------------------------
4110 my $title = $rOpts->{'title'};
4112 ( $title, my $path ) = fileparse($input_file);
4114 my $toc_item_count = 0;
4115 my $in_toc_package = "";
4118 _input_file => $input_file, # name of input file
4119 _title => $title, # title, unescaped
4120 _html_file => $html_file, # name of .html output file
4121 _toc_filename => $toc_filename, # for frames option
4122 _src_filename => $src_filename, # for frames option
4123 _html_file_opened => $html_file_opened, # a flag
4124 _html_fh => $html_fh, # the output stream
4125 _html_pre_fh => $html_pre_fh, # pre section goes here
4126 _rpre_string_stack => \@pre_string_stack, # stack of pre sections
4127 _html_pod_fh => $html_pod_fh, # pod goes here if pod2html
4128 _rpod_string => \$pod_string, # string holding pod
4129 _pod_cut_count => 0, # how many =cut's?
4130 _html_toc_fh => $html_toc_fh, # fh for table of contents
4131 _rtoc_string => \$toc_string, # string holding toc
4132 _rtoc_item_count => \$toc_item_count, # how many toc items
4133 _rin_toc_package => \$in_toc_package, # package name
4134 _rtoc_name_count => {}, # hash to track unique names
4135 _rpackage_stack => [], # stack to check for package
4137 _rlast_level => \$last_level, # brace indentation level
4143 # Add an item to the html table of contents.
4144 # This is called even if no table of contents is written,
4145 # because we still want to put the anchors in the <pre> text.
4146 # We are given an anchor name and its type; types are:
4147 # 'package', 'sub', '__END__', '__DATA__', 'EOF'
4148 # There must be an 'EOF' call at the end to wrap things up.
4150 my ( $name, $type ) = @_;
4151 my $html_toc_fh = $self->{_html_toc_fh};
4152 my $html_pre_fh = $self->{_html_pre_fh};
4153 my $rtoc_name_count = $self->{_rtoc_name_count};
4154 my $rtoc_item_count = $self->{_rtoc_item_count};
4155 my $rlast_level = $self->{_rlast_level};
4156 my $rin_toc_package = $self->{_rin_toc_package};
4157 my $rpackage_stack = $self->{_rpackage_stack};
4159 # packages contain sublists of subs, so to avoid errors all package
4160 # items are written and finished with the following routines
4161 my $end_package_list = sub {
4162 if ($$rin_toc_package) {
4163 $html_toc_fh->print("</ul>\n</li>\n");
4164 $$rin_toc_package = "";
4168 my $start_package_list = sub {
4169 my ( $unique_name, $package ) = @_;
4170 if ($$rin_toc_package) { $end_package_list->() }
4171 $html_toc_fh->print(<<EOM);
4172 <li><a href=\"#$unique_name\">package $package</a>
4175 $$rin_toc_package = $package;
4178 # start the table of contents on the first item
4179 unless ($$rtoc_item_count) {
4181 # but just quit if we hit EOF without any other entries
4182 # in this case, there will be no toc
4183 return if ( $type eq 'EOF' );
4184 $html_toc_fh->print( <<"TOC_END");
4185 <!-- BEGIN CODE INDEX --><a name="code-index"></a>
4189 $$rtoc_item_count++;
4191 # make a unique anchor name for this location:
4192 # - packages get a 'package-' prefix
4193 # - subs use their names
4194 my $unique_name = $name;
4195 if ( $type eq 'package' ) { $unique_name = "package-$name" }
4197 # append '-1', '-2', etc if necessary to make unique; this will
4198 # be unique because subs and packages cannot have a '-'
4199 if ( my $count = $rtoc_name_count->{ lc $unique_name }++ ) {
4200 $unique_name .= "-$count";
4203 # - all names get terminal '-' if pod2html is used, to avoid
4204 # conflicts with anchor names created by pod2html
4205 if ( $rOpts->{'pod2html'} ) { $unique_name .= '-' }
4207 # start/stop lists of subs
4208 if ( $type eq 'sub' ) {
4209 my $package = $rpackage_stack->[$$rlast_level];
4210 unless ($package) { $package = 'main' }
4212 # if we're already in a package/sub list, be sure its the right
4213 # package or else close it
4214 if ( $$rin_toc_package && $$rin_toc_package ne $package ) {
4215 $end_package_list->();
4218 # start a package/sub list if necessary
4219 unless ($$rin_toc_package) {
4220 $start_package_list->( $unique_name, $package );
4224 # now write an entry in the toc for this item
4225 if ( $type eq 'package' ) {
4226 $start_package_list->( $unique_name, $name );
4228 elsif ( $type eq 'sub' ) {
4229 $html_toc_fh->print("<li><a href=\"#$unique_name\">$name</a></li>\n");
4232 $end_package_list->();
4233 $html_toc_fh->print("<li><a href=\"#$unique_name\">$name</a></li>\n");
4236 # write the anchor in the <pre> section
4237 $html_pre_fh->print("<a name=\"$unique_name\"></a>");
4239 # end the table of contents, if any, on the end of file
4240 if ( $type eq 'EOF' ) {
4241 $html_toc_fh->print( <<"TOC_END");
4243 <!-- END CODE INDEX -->
4250 # This is the official list of tokens which may be identified by the
4251 # user. Long names are used as getopt keys. Short names are
4252 # convenient short abbreviations for specifying input. Short names
4253 # somewhat resemble token type characters, but are often different
4254 # because they may only be alphanumeric, to allow command line
4255 # input. Also, note that because of case insensitivity of html,
4256 # this table must be in a single case only (I've chosen to use all
4258 # When adding NEW_TOKENS: update this hash table
4259 # short names => long names
4260 %short_to_long_names = (
4270 'pu' => 'punctuation',
4271 'i' => 'identifier',
4273 'h' => 'here-doc-target',
4274 'hh' => 'here-doc-text',
4276 'sc' => 'semicolon',
4277 'm' => 'subroutine',
4281 # Now we have to map actual token types into one of the above short
4282 # names; any token types not mapped will get 'punctuation'
4285 # The values of this hash table correspond to the keys of the
4286 # previous hash table.
4287 # The keys of this hash table are token types and can be seen
4288 # by running with --dump-token-types (-dtt).
4290 # When adding NEW_TOKENS: update this hash table
4291 # $type => $short_name
4292 %token_short_names = (
4317 # These token types will all be called identifiers for now
4318 # FIXME: need to separate user defined modules as separate type
4319 my @identifier = qw" i t U C Y Z G :: ";
4320 @token_short_names{@identifier} = ('i') x scalar(@identifier);
4322 # These token types will be called 'structure'
4323 my @structure = qw" { } ";
4324 @token_short_names{@structure} = ('s') x scalar(@structure);
4326 # OLD NOTES: save for reference
4327 # Any of these could be added later if it would be useful.
4328 # For now, they will by default become punctuation
4329 # my @list = qw" L R [ ] ";
4330 # @token_long_names{@list} = ('non-structure') x scalar(@list);
4333 # / /= * *= ** **= + += - -= % %= = ++ -- << <<= >> >>= pp p m mm
4335 # @token_long_names{@list} = ('math') x scalar(@list);
4337 # my @list = qw" & &= ~ ~= ^ ^= | |= ";
4338 # @token_long_names{@list} = ('bit') x scalar(@list);
4340 # my @list = qw" == != < > <= <=> ";
4341 # @token_long_names{@list} = ('numerical-comparison') x scalar(@list);
4343 # my @list = qw" && || ! &&= ||= //= ";
4344 # @token_long_names{@list} = ('logical') x scalar(@list);
4346 # my @list = qw" . .= =~ !~ x x= ";
4347 # @token_long_names{@list} = ('string-operators') x scalar(@list);
4350 # my @list = qw" .. -> <> ... \ ? ";
4351 # @token_long_names{@list} = ('misc-operators') x scalar(@list);
4355 sub make_getopt_long_names {
4357 my ($rgetopt_names) = @_;
4358 while ( my ( $short_name, $name ) = each %short_to_long_names ) {
4359 push @$rgetopt_names, "html-color-$name=s";
4360 push @$rgetopt_names, "html-italic-$name!";
4361 push @$rgetopt_names, "html-bold-$name!";
4363 push @$rgetopt_names, "html-color-background=s";
4364 push @$rgetopt_names, "html-linked-style-sheet=s";
4365 push @$rgetopt_names, "nohtml-style-sheets";
4366 push @$rgetopt_names, "html-pre-only";
4367 push @$rgetopt_names, "html-line-numbers";
4368 push @$rgetopt_names, "html-entities!";
4369 push @$rgetopt_names, "stylesheet";
4370 push @$rgetopt_names, "html-table-of-contents!";
4371 push @$rgetopt_names, "pod2html!";
4372 push @$rgetopt_names, "frames!";
4373 push @$rgetopt_names, "html-toc-extension=s";
4374 push @$rgetopt_names, "html-src-extension=s";
4376 # Pod::Html parameters:
4377 push @$rgetopt_names, "backlink=s";
4378 push @$rgetopt_names, "cachedir=s";
4379 push @$rgetopt_names, "htmlroot=s";
4380 push @$rgetopt_names, "libpods=s";
4381 push @$rgetopt_names, "podpath=s";
4382 push @$rgetopt_names, "podroot=s";
4383 push @$rgetopt_names, "title=s";
4385 # Pod::Html parameters with leading 'pod' which will be removed
4386 # before the call to Pod::Html
4387 push @$rgetopt_names, "podquiet!";
4388 push @$rgetopt_names, "podverbose!";
4389 push @$rgetopt_names, "podrecurse!";
4390 push @$rgetopt_names, "podflush";
4391 push @$rgetopt_names, "podheader!";
4392 push @$rgetopt_names, "podindex!";
4395 sub make_abbreviated_names {
4397 # We're appending things like this to the expansion list:
4398 # 'hcc' => [qw(html-color-comment)],
4399 # 'hck' => [qw(html-color-keyword)],
4402 my ($rexpansion) = @_;
4404 # abbreviations for color/bold/italic properties
4405 while ( my ( $short_name, $long_name ) = each %short_to_long_names ) {
4406 ${$rexpansion}{"hc$short_name"} = ["html-color-$long_name"];
4407 ${$rexpansion}{"hb$short_name"} = ["html-bold-$long_name"];
4408 ${$rexpansion}{"hi$short_name"} = ["html-italic-$long_name"];
4409 ${$rexpansion}{"nhb$short_name"} = ["nohtml-bold-$long_name"];
4410 ${$rexpansion}{"nhi$short_name"} = ["nohtml-italic-$long_name"];
4413 # abbreviations for all other html options
4414 ${$rexpansion}{"hcbg"} = ["html-color-background"];
4415 ${$rexpansion}{"pre"} = ["html-pre-only"];
4416 ${$rexpansion}{"toc"} = ["html-table-of-contents"];
4417 ${$rexpansion}{"ntoc"} = ["nohtml-table-of-contents"];
4418 ${$rexpansion}{"nnn"} = ["html-line-numbers"];
4419 ${$rexpansion}{"hent"} = ["html-entities"];
4420 ${$rexpansion}{"nhent"} = ["nohtml-entities"];
4421 ${$rexpansion}{"css"} = ["html-linked-style-sheet"];
4422 ${$rexpansion}{"nss"} = ["nohtml-style-sheets"];
4423 ${$rexpansion}{"ss"} = ["stylesheet"];
4424 ${$rexpansion}{"pod"} = ["pod2html"];
4425 ${$rexpansion}{"npod"} = ["nopod2html"];
4426 ${$rexpansion}{"frm"} = ["frames"];
4427 ${$rexpansion}{"nfrm"} = ["noframes"];
4428 ${$rexpansion}{"text"} = ["html-toc-extension"];
4429 ${$rexpansion}{"sext"} = ["html-src-extension"];
4434 # This will be called once after options have been parsed
4438 # X11 color names for default settings that seemed to look ok
4439 # (these color names are only used for programming clarity; the hex
4440 # numbers are actually written)
4441 use constant ForestGreen => "#228B22";
4442 use constant SaddleBrown => "#8B4513";
4443 use constant magenta4 => "#8B008B";
4444 use constant IndianRed3 => "#CD5555";
4445 use constant DeepSkyBlue4 => "#00688B";
4446 use constant MediumOrchid3 => "#B452CD";
4447 use constant black => "#000000";
4448 use constant white => "#FFFFFF";
4449 use constant red => "#FF0000";
4451 # set default color, bold, italic properties
4452 # anything not listed here will be given the default (punctuation) color --
4453 # these types currently not listed and get default: ws pu s sc cm co p
4454 # When adding NEW_TOKENS: add an entry here if you don't want defaults
4456 # set_default_properties( $short_name, default_color, bold?, italic? );
4457 set_default_properties( 'c', ForestGreen, 0, 0 );
4458 set_default_properties( 'pd', ForestGreen, 0, 1 );
4459 set_default_properties( 'k', magenta4, 1, 0 ); # was SaddleBrown
4460 set_default_properties( 'q', IndianRed3, 0, 0 );
4461 set_default_properties( 'hh', IndianRed3, 0, 1 );
4462 set_default_properties( 'h', IndianRed3, 1, 0 );
4463 set_default_properties( 'i', DeepSkyBlue4, 0, 0 );
4464 set_default_properties( 'w', black, 0, 0 );
4465 set_default_properties( 'n', MediumOrchid3, 0, 0 );
4466 set_default_properties( 'v', MediumOrchid3, 0, 0 );
4467 set_default_properties( 'j', IndianRed3, 1, 0 );
4468 set_default_properties( 'm', red, 1, 0 );
4470 set_default_color( 'html-color-background', white );
4471 set_default_color( 'html-color-punctuation', black );
4473 # setup property lookup tables for tokens based on their short names
4474 # every token type has a short name, and will use these tables
4475 # to do the html markup
4476 while ( my ( $short_name, $long_name ) = each %short_to_long_names ) {
4477 $html_color{$short_name} = $rOpts->{"html-color-$long_name"};
4478 $html_bold{$short_name} = $rOpts->{"html-bold-$long_name"};
4479 $html_italic{$short_name} = $rOpts->{"html-italic-$long_name"};
4482 # write style sheet to STDOUT and die if requested
4483 if ( defined( $rOpts->{'stylesheet'} ) ) {
4484 write_style_sheet_file('-');
4488 # make sure user gives a file name after -css
4489 if ( defined( $rOpts->{'html-linked-style-sheet'} ) ) {
4490 $css_linkname = $rOpts->{'html-linked-style-sheet'};
4491 if ( $css_linkname =~ /^-/ ) {
4492 die "You must specify a valid filename after -css\n";
4496 # check for conflict
4497 if ( $css_linkname && $rOpts->{'nohtml-style-sheets'} ) {
4498 $rOpts->{'nohtml-style-sheets'} = 0;
4499 warning("You can't specify both -css and -nss; -nss ignored\n");
4502 # write a style sheet file if necessary
4503 if ($css_linkname) {
4505 # if the selected filename exists, don't write, because user may
4506 # have done some work by hand to create it; use backup name instead
4507 # Also, this will avoid a potential disaster in which the user
4508 # forgets to specify the style sheet, like this:
4509 # perltidy -html -css myfile1.pl myfile2.pl
4510 # This would cause myfile1.pl to parsed as the style sheet by GetOpts
4511 my $css_filename = $css_linkname;
4512 unless ( -e $css_filename ) {
4513 write_style_sheet_file($css_filename);
4516 $missing_html_entities = 1 unless $rOpts->{'html-entities'};
4519 sub write_style_sheet_file {
4521 my $css_filename = shift;
4523 unless ( $fh = IO::File->new("> $css_filename") ) {
4524 die "can't open $css_filename: $!\n";
4526 write_style_sheet_data($fh);
4527 eval { $fh->close };
4530 sub write_style_sheet_data {
4532 # write the style sheet data to an open file handle
4535 my $bg_color = $rOpts->{'html-color-background'};
4536 my $text_color = $rOpts->{'html-color-punctuation'};
4538 # pre-bgcolor is new, and may not be defined
4539 my $pre_bg_color = $rOpts->{'html-pre-color-background'};
4540 $pre_bg_color = $bg_color unless $pre_bg_color;
4542 $fh->print(<<"EOM");
4543 /* default style sheet generated by perltidy */
4544 body {background: $bg_color; color: $text_color}
4545 pre { color: $text_color;
4546 background: $pre_bg_color;
4547 font-family: courier;
4552 foreach my $short_name ( sort keys %short_to_long_names ) {
4553 my $long_name = $short_to_long_names{$short_name};
4555 my $abbrev = '.' . $short_name;
4556 if ( length($short_name) == 1 ) { $abbrev .= ' ' } # for alignment
4557 my $color = $html_color{$short_name};
4558 if ( !defined($color) ) { $color = $text_color }
4559 $fh->print("$abbrev \{ color: $color;");
4561 if ( $html_bold{$short_name} ) {
4562 $fh->print(" font-weight:bold;");
4565 if ( $html_italic{$short_name} ) {
4566 $fh->print(" font-style:italic;");
4568 $fh->print("} /* $long_name */\n");
4572 sub set_default_color {
4574 # make sure that options hash $rOpts->{$key} contains a valid color
4575 my ( $key, $color ) = @_;
4576 if ( $rOpts->{$key} ) { $color = $rOpts->{$key} }
4577 $rOpts->{$key} = check_RGB($color);
4582 # if color is a 6 digit hex RGB value, prepend a #, otherwise
4583 # assume that it is a valid ascii color name
4585 if ( $color =~ /^[0-9a-fA-F]{6,6}$/ ) { $color = "#$color" }
4589 sub set_default_properties {
4590 my ( $short_name, $color, $bold, $italic ) = @_;
4592 set_default_color( "html-color-$short_to_long_names{$short_name}", $color );
4594 $key = "html-bold-$short_to_long_names{$short_name}";
4595 $rOpts->{$key} = ( defined $rOpts->{$key} ) ? $rOpts->{$key} : $bold;
4596 $key = "html-italic-$short_to_long_names{$short_name}";
4597 $rOpts->{$key} = ( defined $rOpts->{$key} ) ? $rOpts->{$key} : $italic;
4602 # Use Pod::Html to process the pod and make the page
4603 # then merge the perltidy code sections into it.
4604 # return 1 if success, 0 otherwise
4606 my ( $pod_string, $css_string, $toc_string, $rpre_string_stack ) = @_;
4607 my $input_file = $self->{_input_file};
4608 my $title = $self->{_title};
4609 my $success_flag = 0;
4611 # don't try to use pod2html if no pod
4612 unless ($pod_string) {
4613 return $success_flag;
4616 # Pod::Html requires a real temporary filename
4617 # If we are making a frame, we have a name available
4618 # Otherwise, we have to fine one
4620 if ( $rOpts->{'frames'} ) {
4621 $tmpfile = $self->{_toc_filename};
4624 $tmpfile = Perl::Tidy::make_temporary_filename();
4626 my $fh_tmp = IO::File->new( $tmpfile, 'w' );
4628 warn "unable to open temporary file $tmpfile; cannot use pod2html\n";
4629 return $success_flag;
4632 #------------------------------------------------------------------
4633 # Warning: a temporary file is open; we have to clean up if
4634 # things go bad. From here on all returns should be by going to
4635 # RETURN so that the temporary file gets unlinked.
4636 #------------------------------------------------------------------
4638 # write the pod text to the temporary file
4639 $fh_tmp->print($pod_string);
4642 # Hand off the pod to pod2html.
4643 # Note that we can use the same temporary filename for input and output
4644 # because of the way pod2html works.
4648 push @args, "--infile=$tmpfile", "--outfile=$tmpfile", "--title=$title";
4651 # Flags with string args:
4652 # "backlink=s", "cachedir=s", "htmlroot=s", "libpods=s",
4653 # "podpath=s", "podroot=s"
4654 # Note: -css=s is handled by perltidy itself
4655 foreach $kw (qw(backlink cachedir htmlroot libpods podpath podroot)) {
4656 if ( $rOpts->{$kw} ) { push @args, "--$kw=$rOpts->{$kw}" }
4659 # Toggle switches; these have extra leading 'pod'
4660 # "header!", "index!", "recurse!", "quiet!", "verbose!"
4661 foreach $kw (qw(podheader podindex podrecurse podquiet podverbose)) {
4662 my $kwd = $kw; # allows us to strip 'pod'
4663 if ( $rOpts->{$kw} ) { $kwd =~ s/^pod//; push @args, "--$kwd" }
4664 elsif ( defined( $rOpts->{$kw} ) ) {
4666 push @args, "--no$kwd";
4672 if ( $rOpts->{$kw} ) { $kw =~ s/^pod//; push @args, "--$kw" }
4674 # Must clean up if pod2html dies (it can);
4675 # Be careful not to overwrite callers __DIE__ routine
4676 local $SIG{__DIE__} = sub {
4678 unlink $tmpfile if -e $tmpfile;
4684 $fh_tmp = IO::File->new( $tmpfile, 'r' );
4687 # this error shouldn't happen ... we just used this filename
4688 warn "unable to open temporary file $tmpfile; cannot use pod2html\n";
4692 my $html_fh = $self->{_html_fh};
4697 # This routine will write the html selectively and store the toc
4698 my $html_print = sub {
4700 $html_fh->print($_) unless ($no_print);
4701 if ($in_toc) { push @toc, $_ }
4705 # loop over lines of html output from pod2html and merge in
4706 # the necessary perltidy html sections
4707 my ( $saw_body, $saw_index, $saw_body_end );
4708 while ( my $line = $fh_tmp->getline() ) {
4710 if ( $line =~ /^\s*<html>\s*$/i ) {
4711 my $date = localtime;
4712 $html_print->("<!-- Generated by perltidy on $date -->\n");
4713 $html_print->($line);
4716 # Copy the perltidy css, if any, after <body> tag
4717 elsif ( $line =~ /^\s*<body.*>\s*$/i ) {
4719 $html_print->($css_string) if $css_string;
4720 $html_print->($line);
4722 # add a top anchor and heading
4723 $html_print->("<a name=\"-top-\"></a>\n");
4724 $title = escape_html($title);
4725 $html_print->("<h1>$title</h1>\n");
4727 elsif ( $line =~ /^\s*<!-- INDEX BEGIN -->\s*$/i ) {
4730 # when frames are used, an extra table of contents in the
4731 # contents panel is confusing, so don't print it
4732 $no_print = $rOpts->{'frames'}
4733 || !$rOpts->{'html-table-of-contents'};
4734 $html_print->("<h2>Doc Index:</h2>\n") if $rOpts->{'frames'};
4735 $html_print->($line);
4738 # Copy the perltidy toc, if any, after the Pod::Html toc
4739 elsif ( $line =~ /^\s*<!-- INDEX END -->\s*$/i ) {
4741 $html_print->($line);
4743 $html_print->("<hr />\n") if $rOpts->{'frames'};
4744 $html_print->("<h2>Code Index:</h2>\n");
4745 my @toc = map { $_ .= "\n" } split /\n/, $toc_string;
4746 $html_print->(@toc);
4752 # Copy one perltidy section after each marker
4753 elsif ( $line =~ /^(.*)<!-- pERLTIDY sECTION -->(.*)$/ ) {
4755 $html_print->($1) if $1;
4757 # Intermingle code and pod sections if we saw multiple =cut's.
4758 if ( $self->{_pod_cut_count} > 1 ) {
4759 my $rpre_string = shift(@$rpre_string_stack);
4760 if ($$rpre_string) {
4761 $html_print->('<pre>');
4762 $html_print->($$rpre_string);
4763 $html_print->('</pre>');
4767 # shouldn't happen: we stored a string before writing
4770 "Problem merging html stream with pod2html; order may be wrong\n";
4772 $html_print->($line);
4775 # If didn't see multiple =cut lines, we'll put the pod out first
4776 # and then the code, because it's less confusing.
4779 # since we are not intermixing code and pod, we don't need
4780 # or want any <hr> lines which separated pod and code
4781 $html_print->($line) unless ( $line =~ /^\s*<hr>\s*$/i );
4785 # Copy any remaining code section before the </body> tag
4786 elsif ( $line =~ /^\s*<\/body>\s*$/i ) {
4788 if (@$rpre_string_stack) {
4789 unless ( $self->{_pod_cut_count} > 1 ) {
4790 $html_print->('<hr />');
4792 while ( my $rpre_string = shift(@$rpre_string_stack) ) {
4793 $html_print->('<pre>');
4794 $html_print->($$rpre_string);
4795 $html_print->('</pre>');
4798 $html_print->($line);
4801 $html_print->($line);
4806 unless ($saw_body) {
4807 warn "Did not see <body> in pod2html output\n";
4810 unless ($saw_body_end) {
4811 warn "Did not see </body> in pod2html output\n";
4814 unless ($saw_index) {
4815 warn "Did not find INDEX END in pod2html output\n";
4820 eval { $html_fh->close() };
4822 # note that we have to unlink tmpfile before making frames
4823 # because the tmpfile may be one of the names used for frames
4824 unlink $tmpfile if -e $tmpfile;
4825 if ( $success_flag && $rOpts->{'frames'} ) {
4826 $self->make_frame( \@toc );
4828 return $success_flag;
4833 # Make a frame with table of contents in the left panel
4834 # and the text in the right panel.
4836 # $html_filename contains the no-frames html output
4837 # $rtoc is a reference to an array with the table of contents
4840 my $input_file = $self->{_input_file};
4841 my $html_filename = $self->{_html_file};
4842 my $toc_filename = $self->{_toc_filename};
4843 my $src_filename = $self->{_src_filename};
4844 my $title = $self->{_title};
4845 $title = escape_html($title);
4847 # FUTURE input parameter:
4848 my $top_basename = "";
4850 # We need to produce 3 html files:
4851 # 1. - the table of contents
4852 # 2. - the contents (source code) itself
4853 # 3. - the frame which contains them
4855 # get basenames for relative links
4856 my ( $toc_basename, $toc_path ) = fileparse($toc_filename);
4857 my ( $src_basename, $src_path ) = fileparse($src_filename);
4859 # 1. Make the table of contents panel, with appropriate changes
4860 # to the anchor names
4861 my $src_frame_name = 'SRC';
4863 write_toc_html( $title, $toc_filename, $src_basename, $rtoc,
4866 # 2. The current .html filename is renamed to be the contents panel
4867 rename( $html_filename, $src_filename )
4868 or die "Cannot rename $html_filename to $src_filename:$!\n";
4870 # 3. Then use the original html filename for the frame
4872 $title, $html_filename, $top_basename,
4873 $toc_basename, $src_basename, $src_frame_name
4877 sub write_toc_html {
4879 # write a separate html table of contents file for frames
4880 my ( $title, $toc_filename, $src_basename, $rtoc, $src_frame_name ) = @_;
4881 my $fh = IO::File->new( $toc_filename, 'w' )
4882 or die "Cannot open $toc_filename:$!\n";
4886 <title>$title</title>
4889 <h1><a href=\"$src_basename#-top-" target="$src_frame_name">$title</a></h1>
4893 change_anchor_names( $rtoc, $src_basename, "$src_frame_name" );
4894 $fh->print( join "", @$rtoc );
4903 sub write_frame_html {
4905 # write an html file to be the table of contents frame
4907 $title, $frame_filename, $top_basename,
4908 $toc_basename, $src_basename, $src_frame_name
4911 my $fh = IO::File->new( $frame_filename, 'w' )
4912 or die "Cannot open $toc_basename:$!\n";
4915 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Frameset//EN"
4916 "http://www.w3.org/TR/xhtml1/DTD/xhtml1-frameset.dtd">
4917 <?xml version="1.0" encoding="iso-8859-1" ?>
4918 <html xmlns="http://www.w3.org/1999/xhtml">
4920 <title>$title</title>
4924 # two left panels, one right, if master index file
4925 if ($top_basename) {
4927 <frameset cols="20%,80%">
4928 <frameset rows="30%,70%">
4929 <frame src = "$top_basename" />
4930 <frame src = "$toc_basename" />
4935 # one left panels, one right, if no master index file
4938 <frameset cols="20%,*">
4939 <frame src = "$toc_basename" />
4943 <frame src = "$src_basename" name = "$src_frame_name" />
4946 <p>If you see this message, you are using a non-frame-capable web client.</p>
4947 <p>This document contains:</p>
4949 <li><a href="$toc_basename">A table of contents</a></li>
4950 <li><a href="$src_basename">The source code</a></li>
4959 sub change_anchor_names {
4961 # add a filename and target to anchors
4962 # also return the first anchor
4963 my ( $rlines, $filename, $target ) = @_;
4965 foreach my $line (@$rlines) {
4967 # We're looking for lines like this:
4968 # <LI><A HREF="#synopsis">SYNOPSIS</A></LI>
4969 # ---- - -------- -----------------
4971 if ( $line =~ /^(.*)<a(.*)href\s*=\s*"([^#]*)#([^"]+)"[^>]*>(.*)$/i ) {
4975 my $href = "$filename#$name";
4976 $line = "$pre<a href=\"$href\" target=\"$target\">$post\n";
4977 unless ($first_anchor) { $first_anchor = $href }
4980 return $first_anchor;
4983 sub close_html_file {
4985 return unless $self->{_html_file_opened};
4987 my $html_fh = $self->{_html_fh};
4988 my $rtoc_string = $self->{_rtoc_string};
4990 # There are 3 basic paths to html output...
4992 # ---------------------------------
4993 # Path 1: finish up if in -pre mode
4994 # ---------------------------------
4995 if ( $rOpts->{'html-pre-only'} ) {
4996 $html_fh->print( <<"PRE_END");
4999 eval { $html_fh->close() };
5004 $self->add_toc_item( 'EOF', 'EOF' );
5006 my $rpre_string_stack = $self->{_rpre_string_stack};
5008 # Patch to darken the <pre> background color in case of pod2html and
5009 # interleaved code/documentation. Otherwise, the distinction
5010 # between code and documentation is blurred.
5011 if ( $rOpts->{pod2html}
5012 && $self->{_pod_cut_count} >= 1
5013 && $rOpts->{'html-color-background'} eq '#FFFFFF' )
5015 $rOpts->{'html-pre-color-background'} = '#F0F0F0';
5018 # put the css or its link into a string, if used
5020 my $fh_css = Perl::Tidy::IOScalar->new( \$css_string, 'w' );
5022 # use css linked to another file
5023 if ( $rOpts->{'html-linked-style-sheet'} ) {
5025 qq(<link rel="stylesheet" href="$css_linkname" type="text/css" />)
5029 # use css embedded in this file
5030 elsif ( !$rOpts->{'nohtml-style-sheets'} ) {
5031 $fh_css->print( <<'ENDCSS');
5032 <style type="text/css">
5035 write_style_sheet_data($fh_css);
5036 $fh_css->print( <<"ENDCSS");
5042 # -----------------------------------------------------------
5043 # path 2: use pod2html if requested
5044 # If we fail for some reason, continue on to path 3
5045 # -----------------------------------------------------------
5046 if ( $rOpts->{'pod2html'} ) {
5047 my $rpod_string = $self->{_rpod_string};
5048 $self->pod_to_html( $$rpod_string, $css_string, $$rtoc_string,
5049 $rpre_string_stack )
5053 # --------------------------------------------------
5054 # path 3: write code in html, with pod only in italics
5055 # --------------------------------------------------
5056 my $input_file = $self->{_input_file};
5057 my $title = escape_html($input_file);
5058 my $date = localtime;
5059 $html_fh->print( <<"HTML_START");
5060 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
5061 "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
5062 <!-- Generated by perltidy on $date -->
5063 <html xmlns="http://www.w3.org/1999/xhtml">
5065 <title>$title</title>
5068 # output the css, if used
5070 $html_fh->print($css_string);
5071 $html_fh->print( <<"ENDCSS");
5078 $html_fh->print( <<"HTML_START");
5080 <body bgcolor=\"$rOpts->{'html-color-background'}\" text=\"$rOpts->{'html-color-punctuation'}\">
5084 $html_fh->print("<a name=\"-top-\"></a>\n");
5085 $html_fh->print( <<"EOM");
5089 # copy the table of contents
5091 && !$rOpts->{'frames'}
5092 && $rOpts->{'html-table-of-contents'} )
5094 $html_fh->print($$rtoc_string);
5097 # copy the pre section(s)
5098 my $fname_comment = $input_file;
5099 $fname_comment =~ s/--+/-/g; # protect HTML comment tags
5100 $html_fh->print( <<"END_PRE");
5102 <!-- contents of filename: $fname_comment -->
5106 foreach my $rpre_string (@$rpre_string_stack) {
5107 $html_fh->print($$rpre_string);
5110 # and finish the html page
5111 $html_fh->print( <<"HTML_END");
5116 eval { $html_fh->close() }; # could be object without close method
5118 if ( $rOpts->{'frames'} ) {
5119 my @toc = map { $_ .= "\n" } split /\n/, $$rtoc_string;
5120 $self->make_frame( \@toc );
5126 my ( $rtokens, $rtoken_type, $rlevels ) = @_;
5127 my ( @colored_tokens, $j, $string, $type, $token, $level );
5128 my $rlast_level = $self->{_rlast_level};
5129 my $rpackage_stack = $self->{_rpackage_stack};
5131 for ( $j = 0 ; $j < @$rtoken_type ; $j++ ) {
5132 $type = $$rtoken_type[$j];
5133 $token = $$rtokens[$j];
5134 $level = $$rlevels[$j];
5135 $level = 0 if ( $level < 0 );
5137 #-------------------------------------------------------
5138 # Update the package stack. The package stack is needed to keep
5139 # the toc correct because some packages may be declared within
5140 # blocks and go out of scope when we leave the block.
5141 #-------------------------------------------------------
5142 if ( $level > $$rlast_level ) {
5143 unless ( $rpackage_stack->[ $level - 1 ] ) {
5144 $rpackage_stack->[ $level - 1 ] = 'main';
5146 $rpackage_stack->[$level] = $rpackage_stack->[ $level - 1 ];
5148 elsif ( $level < $$rlast_level ) {
5149 my $package = $rpackage_stack->[$level];
5150 unless ($package) { $package = 'main' }
5152 # if we change packages due to a nesting change, we
5153 # have to make an entry in the toc
5154 if ( $package ne $rpackage_stack->[ $level + 1 ] ) {
5155 $self->add_toc_item( $package, 'package' );
5158 $$rlast_level = $level;
5160 #-------------------------------------------------------
5161 # Intercept a sub name here; split it
5162 # into keyword 'sub' and sub name; and add an
5164 #-------------------------------------------------------
5165 if ( $type eq 'i' && $token =~ /^(sub\s+)(\w.*)$/ ) {
5166 $token = $self->markup_html_element( $1, 'k' );
5167 push @colored_tokens, $token;
5171 # but don't include sub declarations in the toc;
5172 # these wlll have leading token types 'i;'
5173 my $signature = join "", @$rtoken_type;
5174 unless ( $signature =~ /^i;/ ) {
5175 my $subname = $token;
5176 $subname =~ s/[\s\(].*$//; # remove any attributes and prototype
5177 $self->add_toc_item( $subname, 'sub' );
5181 #-------------------------------------------------------
5182 # Intercept a package name here; split it
5183 # into keyword 'package' and name; add to the toc,
5184 # and update the package stack
5185 #-------------------------------------------------------
5186 if ( $type eq 'i' && $token =~ /^(package\s+)(\w.*)$/ ) {
5187 $token = $self->markup_html_element( $1, 'k' );
5188 push @colored_tokens, $token;
5191 $self->add_toc_item( "$token", 'package' );
5192 $rpackage_stack->[$level] = $token;
5195 $token = $self->markup_html_element( $token, $type );
5196 push @colored_tokens, $token;
5198 return ( \@colored_tokens );
5201 sub markup_html_element {
5203 my ( $token, $type ) = @_;
5205 return $token if ( $type eq 'b' ); # skip a blank token
5206 return $token if ( $token =~ /^\s*$/ ); # skip a blank line
5207 $token = escape_html($token);
5209 # get the short abbreviation for this token type
5210 my $short_name = $token_short_names{$type};
5211 if ( !defined($short_name) ) {
5212 $short_name = "pu"; # punctuation is default
5215 # handle style sheets..
5216 if ( !$rOpts->{'nohtml-style-sheets'} ) {
5217 if ( $short_name ne 'pu' ) {
5218 $token = qq(<span class="$short_name">) . $token . "</span>";
5222 # handle no style sheets..
5224 my $color = $html_color{$short_name};
5226 if ( $color && ( $color ne $rOpts->{'html-color-punctuation'} ) ) {
5227 $token = qq(<font color="$color">) . $token . "</font>";
5229 if ( $html_italic{$short_name} ) { $token = "<i>$token</i>" }
5230 if ( $html_bold{$short_name} ) { $token = "<b>$token</b>" }
5238 if ($missing_html_entities) {
5239 $token =~ s/\&/&/g;
5240 $token =~ s/\</</g;
5241 $token =~ s/\>/>/g;
5242 $token =~ s/\"/"/g;
5245 HTML::Entities::encode_entities($token);
5250 sub finish_formatting {
5252 # called after last line
5254 $self->close_html_file();
5261 return unless $self->{_html_file_opened};
5262 my $html_pre_fh = $self->{_html_pre_fh};
5263 my ($line_of_tokens) = @_;
5264 my $line_type = $line_of_tokens->{_line_type};
5265 my $input_line = $line_of_tokens->{_line_text};
5266 my $line_number = $line_of_tokens->{_line_number};
5269 # markup line of code..
5271 if ( $line_type eq 'CODE' ) {
5272 my $rtoken_type = $line_of_tokens->{_rtoken_type};
5273 my $rtokens = $line_of_tokens->{_rtokens};
5274 my $rlevels = $line_of_tokens->{_rlevels};
5276 if ( $input_line =~ /(^\s*)/ ) {
5282 my ($rcolored_tokens) =
5283 $self->markup_tokens( $rtokens, $rtoken_type, $rlevels );
5284 $html_line .= join '', @$rcolored_tokens;
5287 # markup line of non-code..
5290 if ( $line_type eq 'HERE' ) { $line_character = 'H' }
5291 elsif ( $line_type eq 'HERE_END' ) { $line_character = 'h' }
5292 elsif ( $line_type eq 'FORMAT' ) { $line_character = 'H' }
5293 elsif ( $line_type eq 'FORMAT_END' ) { $line_character = 'h' }
5294 elsif ( $line_type eq 'SYSTEM' ) { $line_character = 'c' }
5295 elsif ( $line_type eq 'END_START' ) {
5296 $line_character = 'k';
5297 $self->add_toc_item( '__END__', '__END__' );
5299 elsif ( $line_type eq 'DATA_START' ) {
5300 $line_character = 'k';
5301 $self->add_toc_item( '__DATA__', '__DATA__' );
5303 elsif ( $line_type =~ /^POD/ ) {
5304 $line_character = 'P';
5305 if ( $rOpts->{'pod2html'} ) {
5306 my $html_pod_fh = $self->{_html_pod_fh};
5307 if ( $line_type eq 'POD_START' ) {
5309 my $rpre_string_stack = $self->{_rpre_string_stack};
5310 my $rpre_string = $rpre_string_stack->[-1];
5312 # if we have written any non-blank lines to the
5313 # current pre section, start writing to a new output
5315 if ( $$rpre_string =~ /\S/ ) {
5318 Perl::Tidy::IOScalar->new( \$pre_string, 'w' );
5319 $self->{_html_pre_fh} = $html_pre_fh;
5320 push @$rpre_string_stack, \$pre_string;
5322 # leave a marker in the pod stream so we know
5323 # where to put the pre section we just
5325 my $for_html = '=for html'; # don't confuse pod utils
5326 $html_pod_fh->print(<<EOM);
5329 <!-- pERLTIDY sECTION -->
5334 # otherwise, just clear the current string and start
5338 $html_pod_fh->print("\n");
5341 $html_pod_fh->print( $input_line . "\n" );
5342 if ( $line_type eq 'POD_END' ) {
5343 $self->{_pod_cut_count}++;
5344 $html_pod_fh->print("\n");
5349 else { $line_character = 'Q' }
5350 $html_line = $self->markup_html_element( $input_line, $line_character );
5353 # add the line number if requested
5354 if ( $rOpts->{'html-line-numbers'} ) {
5356 ( $line_number < 10 ) ? " "
5357 : ( $line_number < 100 ) ? " "
5358 : ( $line_number < 1000 ) ? " "
5360 $html_line = $extra_space . $line_number . " " . $html_line;
5364 $html_pre_fh->print("$html_line\n");
5367 #####################################################################
5369 # The Perl::Tidy::Formatter package adds indentation, whitespace, and
5370 # line breaks to the token stream
5372 # WARNING: This is not a real class for speed reasons. Only one
5373 # Formatter may be used.
5375 #####################################################################
5377 package Perl::Tidy::Formatter;
5381 # Caution: these debug flags produce a lot of output
5382 # They should all be 0 except when debugging small scripts
5383 use constant FORMATTER_DEBUG_FLAG_BOND => 0;
5384 use constant FORMATTER_DEBUG_FLAG_BREAK => 0;
5385 use constant FORMATTER_DEBUG_FLAG_CI => 0;
5386 use constant FORMATTER_DEBUG_FLAG_FLUSH => 0;
5387 use constant FORMATTER_DEBUG_FLAG_FORCE => 0;
5388 use constant FORMATTER_DEBUG_FLAG_LIST => 0;
5389 use constant FORMATTER_DEBUG_FLAG_NOBREAK => 0;
5390 use constant FORMATTER_DEBUG_FLAG_OUTPUT => 0;
5391 use constant FORMATTER_DEBUG_FLAG_SPARSE => 0;
5392 use constant FORMATTER_DEBUG_FLAG_STORE => 0;
5393 use constant FORMATTER_DEBUG_FLAG_UNDOBP => 0;
5394 use constant FORMATTER_DEBUG_FLAG_WHITE => 0;
5396 my $debug_warning = sub {
5397 print "FORMATTER_DEBUGGING with key $_[0]\n";
5400 FORMATTER_DEBUG_FLAG_BOND && $debug_warning->('BOND');
5401 FORMATTER_DEBUG_FLAG_BREAK && $debug_warning->('BREAK');
5402 FORMATTER_DEBUG_FLAG_CI && $debug_warning->('CI');
5403 FORMATTER_DEBUG_FLAG_FLUSH && $debug_warning->('FLUSH');
5404 FORMATTER_DEBUG_FLAG_FORCE && $debug_warning->('FORCE');
5405 FORMATTER_DEBUG_FLAG_LIST && $debug_warning->('LIST');
5406 FORMATTER_DEBUG_FLAG_NOBREAK && $debug_warning->('NOBREAK');
5407 FORMATTER_DEBUG_FLAG_OUTPUT && $debug_warning->('OUTPUT');
5408 FORMATTER_DEBUG_FLAG_SPARSE && $debug_warning->('SPARSE');
5409 FORMATTER_DEBUG_FLAG_STORE && $debug_warning->('STORE');
5410 FORMATTER_DEBUG_FLAG_UNDOBP && $debug_warning->('UNDOBP');
5411 FORMATTER_DEBUG_FLAG_WHITE && $debug_warning->('WHITE');
5418 $max_gnu_stack_index
5419 $gnu_position_predictor
5420 $line_start_index_to_go
5421 $last_indentation_written
5422 $last_unadjusted_indentation
5425 $saw_VERSION_in_this_file
5430 $gnu_sequence_number
5431 $last_output_indentation
5437 @type_sequence_to_go
5438 @container_environment_to_go
5439 @bond_strength_to_go
5440 @forced_breakpoint_to_go
5443 @leading_spaces_to_go
5444 @reduced_spaces_to_go
5445 @matching_token_to_go
5447 @nesting_blocks_to_go
5449 @nesting_depth_to_go
5451 @old_breakpoint_to_go
5455 %saved_opening_indentation
5458 $comma_count_in_batch
5459 $old_line_count_in_batch
5460 $last_nonblank_index_to_go
5461 $last_nonblank_type_to_go
5462 $last_nonblank_token_to_go
5463 $last_last_nonblank_index_to_go
5464 $last_last_nonblank_type_to_go
5465 $last_last_nonblank_token_to_go
5466 @nonblank_lines_at_depth
5470 $in_format_skipping_section
5471 $format_skipping_pattern_begin
5472 $format_skipping_pattern_end
5474 $forced_breakpoint_count
5475 $forced_breakpoint_undo_count
5476 @forced_breakpoint_undo_stack
5477 %postponed_breakpoint
5481 $first_embedded_tab_at
5482 $last_embedded_tab_at
5483 $deleted_semicolon_count
5484 $first_deleted_semicolon_at
5485 $last_deleted_semicolon_at
5486 $added_semicolon_count
5487 $first_added_semicolon_at
5488 $last_added_semicolon_at
5489 $first_tabbing_disagreement
5490 $last_tabbing_disagreement
5491 $in_tabbing_disagreement
5492 $tabbing_disagreement_count
5496 $last_line_leading_type
5497 $last_line_leading_level
5498 $last_last_line_leading_level
5501 %block_opening_line_number
5502 $csc_new_statement_ok
5503 $accumulating_text_for_block
5505 $rleading_block_if_elsif_text
5506 $leading_block_text_level
5507 $leading_block_text_length_exceeded
5508 $leading_block_text_line_length
5509 $leading_block_text_line_number
5510 $closing_side_comment_prefix_pattern
5511 $closing_side_comment_list_pattern
5513 $last_nonblank_token
5515 $last_last_nonblank_token
5516 $last_last_nonblank_type
5517 $last_nonblank_block_type
5520 %is_if_brace_follower
5521 %space_after_keyword
5524 %is_last_next_redo_return
5525 %is_other_brace_follower
5526 %is_else_brace_follower
5527 %is_anon_sub_brace_follower
5528 %is_anon_sub_1_brace_follower
5530 %is_sort_map_grep_eval
5531 %is_sort_map_grep_eval_do
5532 %is_block_without_semicolon
5537 %is_if_unless_and_or_last_next_redo_return
5538 %is_until_while_for_if_elsif_else
5544 $is_static_block_comment
5545 $index_start_one_line_block
5546 $semicolons_before_block_self_destruct
5547 $index_max_forced_break
5550 $vertical_aligner_object
5555 $last_line_had_side_comment
5558 $static_block_comment_pattern
5559 $static_side_comment_pattern
5560 %opening_vertical_tightness
5561 %closing_vertical_tightness
5562 %closing_token_indentation
5564 %opening_token_right
5565 %stack_opening_token
5566 %stack_closing_token
5568 $block_brace_vertical_tightness_pattern
5571 $rOpts_add_whitespace
5572 $rOpts_block_brace_tightness
5573 $rOpts_block_brace_vertical_tightness
5574 $rOpts_brace_left_and_indent
5575 $rOpts_comma_arrow_breakpoints
5576 $rOpts_break_at_old_keyword_breakpoints
5577 $rOpts_break_at_old_comma_breakpoints
5578 $rOpts_break_at_old_logical_breakpoints
5579 $rOpts_break_at_old_ternary_breakpoints
5580 $rOpts_closing_side_comment_else_flag
5581 $rOpts_closing_side_comment_maximum_text
5582 $rOpts_continuation_indentation
5584 $rOpts_delete_old_whitespace
5585 $rOpts_fuzzy_line_length
5586 $rOpts_indent_columns
5587 $rOpts_line_up_parentheses
5588 $rOpts_maximum_fields_per_table
5589 $rOpts_maximum_line_length
5590 $rOpts_short_concatenation_item_length
5591 $rOpts_swallow_optional_blank_lines
5592 $rOpts_ignore_old_breakpoints
5593 $rOpts_format_skipping
5594 $rOpts_space_function_paren
5595 $rOpts_space_keyword_paren
5596 $rOpts_keep_interior_semicolons
5598 $half_maximum_line_length
5602 %is_keyword_returning_list
5606 %right_bond_strength
5623 # default list of block types for which -bli would apply
5624 $bli_list_string = 'if else elsif unless while for foreach do : sub';
5627 .. :: << >> ** && .. || // -> => += -= .= %= &= |= ^= *= <>
5628 <= >= == =~ !~ != ++ -- /= x=
5630 @is_digraph{@_} = (1) x scalar(@_);
5632 @_ = qw( ... **= <<= >>= &&= ||= //= <=> );
5633 @is_trigraph{@_} = (1) x scalar(@_);
5636 = **= += *= &= <<= &&=
5637 -= /= |= >>= ||= //=
5641 @is_assignment{@_} = (1) x scalar(@_);
5651 @is_keyword_returning_list{@_} = (1) x scalar(@_);
5653 @_ = qw(is if unless and or err last next redo return);
5654 @is_if_unless_and_or_last_next_redo_return{@_} = (1) x scalar(@_);
5656 # always break after a closing curly of these block types:
5657 @_ = qw(until while for if elsif else);
5658 @is_until_while_for_if_elsif_else{@_} = (1) x scalar(@_);
5660 @_ = qw(last next redo return);
5661 @is_last_next_redo_return{@_} = (1) x scalar(@_);
5663 @_ = qw(sort map grep);
5664 @is_sort_map_grep{@_} = (1) x scalar(@_);
5666 @_ = qw(sort map grep eval);
5667 @is_sort_map_grep_eval{@_} = (1) x scalar(@_);
5669 @_ = qw(sort map grep eval do);
5670 @is_sort_map_grep_eval_do{@_} = (1) x scalar(@_);
5673 @is_if_unless{@_} = (1) x scalar(@_);
5675 @_ = qw(and or err);
5676 @is_and_or{@_} = (1) x scalar(@_);
5678 # Identify certain operators which often occur in chains.
5679 # Note: the minus (-) causes a side effect of padding of the first line in
5680 # something like this (by sub set_logical_padding):
5681 # Checkbutton => 'Transmission checked',
5682 # -variable => \$TRANS
5683 # This usually improves appearance so it seems ok.
5684 @_ = qw(&& || and or : ? . + - * /);
5685 @is_chain_operator{@_} = (1) x scalar(@_);
5687 # We can remove semicolons after blocks preceded by these keywords
5689 qw(BEGIN END CHECK INIT AUTOLOAD DESTROY UNITCHECK continue if elsif else
5690 unless while until for foreach);
5691 @is_block_without_semicolon{@_} = (1) x scalar(@_);
5693 # 'L' is token for opening { at hash key
5695 @is_opening_type{@_} = (1) x scalar(@_);
5697 # 'R' is token for closing } at hash key
5699 @is_closing_type{@_} = (1) x scalar(@_);
5702 @is_opening_token{@_} = (1) x scalar(@_);
5705 @is_closing_token{@_} = (1) x scalar(@_);
5709 use constant WS_YES => 1;
5710 use constant WS_OPTIONAL => 0;
5711 use constant WS_NO => -1;
5713 # Token bond strengths.
5714 use constant NO_BREAK => 10000;
5715 use constant VERY_STRONG => 100;
5716 use constant STRONG => 2.1;
5717 use constant NOMINAL => 1.1;
5718 use constant WEAK => 0.8;
5719 use constant VERY_WEAK => 0.55;
5721 # values for testing indexes in output array
5722 use constant UNDEFINED_INDEX => -1;
5724 # Maximum number of little messages; probably need not be changed.
5725 use constant MAX_NAG_MESSAGES => 6;
5727 # increment between sequence numbers for each type
5728 # For example, ?: pairs might have numbers 7,11,15,...
5729 use constant TYPE_SEQUENCE_INCREMENT => 4;
5733 # methods to count instances
5735 sub get_count { $_count; }
5736 sub _increment_count { ++$_count }
5737 sub _decrement_count { --$_count }
5742 # trim leading and trailing whitespace from a string
5750 # given a string containing words separated by whitespace,
5751 # return the list of words
5756 return split( /\s+/, $str );
5759 # interface to Perl::Tidy::Logger routines
5761 if ($logger_object) {
5762 $logger_object->warning(@_);
5767 if ($logger_object) {
5768 $logger_object->complain(@_);
5772 sub write_logfile_entry {
5773 if ($logger_object) {
5774 $logger_object->write_logfile_entry(@_);
5779 if ($logger_object) {
5780 $logger_object->black_box(@_);
5784 sub report_definite_bug {
5785 if ($logger_object) {
5786 $logger_object->report_definite_bug();
5790 sub get_saw_brace_error {
5791 if ($logger_object) {
5792 $logger_object->get_saw_brace_error();
5796 sub we_are_at_the_last_line {
5797 if ($logger_object) {
5798 $logger_object->we_are_at_the_last_line();
5802 # interface to Perl::Tidy::Diagnostics routine
5803 sub write_diagnostics {
5805 if ($diagnostics_object) {
5806 $diagnostics_object->write_diagnostics(@_);
5810 sub get_added_semicolon_count {
5812 return $added_semicolon_count;
5816 $_[0]->_decrement_count();
5823 # we are given an object with a write_line() method to take lines
5825 sink_object => undef,
5826 diagnostics_object => undef,
5827 logger_object => undef,
5829 my %args = ( %defaults, @_ );
5831 $logger_object = $args{logger_object};
5832 $diagnostics_object = $args{diagnostics_object};
5834 # we create another object with a get_line() and peek_ahead() method
5835 my $sink_object = $args{sink_object};
5836 $file_writer_object =
5837 Perl::Tidy::FileWriter->new( $sink_object, $rOpts, $logger_object );
5839 # initialize the leading whitespace stack to negative levels
5840 # so that we can never run off the end of the stack
5841 $gnu_position_predictor = 0; # where the current token is predicted to be
5842 $max_gnu_stack_index = 0;
5843 $max_gnu_item_index = -1;
5844 $gnu_stack[0] = new_lp_indentation_item( 0, -1, -1, 0, 0 );
5845 @gnu_item_list = ();
5846 $last_output_indentation = 0;
5847 $last_indentation_written = 0;
5848 $last_unadjusted_indentation = 0;
5849 $last_leading_token = "";
5851 $saw_VERSION_in_this_file = !$rOpts->{'pass-version-line'};
5852 $saw_END_or_DATA_ = 0;
5854 @block_type_to_go = ();
5855 @type_sequence_to_go = ();
5856 @container_environment_to_go = ();
5857 @bond_strength_to_go = ();
5858 @forced_breakpoint_to_go = ();
5859 @lengths_to_go = (); # line length to start of ith token
5861 @matching_token_to_go = ();
5862 @mate_index_to_go = ();
5863 @nesting_blocks_to_go = ();
5864 @ci_levels_to_go = ();
5865 @nesting_depth_to_go = (0);
5866 @nobreak_to_go = ();
5867 @old_breakpoint_to_go = ();
5870 @leading_spaces_to_go = ();
5871 @reduced_spaces_to_go = ();
5874 @has_broken_sublist = ();
5875 @want_comma_break = ();
5878 $first_tabbing_disagreement = 0;
5879 $last_tabbing_disagreement = 0;
5880 $tabbing_disagreement_count = 0;
5881 $in_tabbing_disagreement = 0;
5882 $input_line_tabbing = undef;
5884 $last_line_type = "";
5885 $last_last_line_leading_level = 0;
5886 $last_line_leading_level = 0;
5887 $last_line_leading_type = '#';
5889 $last_nonblank_token = ';';
5890 $last_nonblank_type = ';';
5891 $last_last_nonblank_token = ';';
5892 $last_last_nonblank_type = ';';
5893 $last_nonblank_block_type = "";
5894 $last_output_level = 0;
5895 $looking_for_else = 0;
5896 $embedded_tab_count = 0;
5897 $first_embedded_tab_at = 0;
5898 $last_embedded_tab_at = 0;
5899 $deleted_semicolon_count = 0;
5900 $first_deleted_semicolon_at = 0;
5901 $last_deleted_semicolon_at = 0;
5902 $added_semicolon_count = 0;
5903 $first_added_semicolon_at = 0;
5904 $last_added_semicolon_at = 0;
5905 $last_line_had_side_comment = 0;
5906 $is_static_block_comment = 0;
5907 %postponed_breakpoint = ();
5909 # variables for adding side comments
5910 %block_leading_text = ();
5911 %block_opening_line_number = ();
5912 $csc_new_statement_ok = 1;
5914 %saved_opening_indentation = ();
5915 $in_format_skipping_section = 0;
5917 reset_block_text_accumulator();
5919 prepare_for_new_input_lines();
5921 $vertical_aligner_object =
5922 Perl::Tidy::VerticalAligner->initialize( $rOpts, $file_writer_object,
5923 $logger_object, $diagnostics_object );
5925 if ( $rOpts->{'entab-leading-whitespace'} ) {
5926 write_logfile_entry(
5927 "Leading whitespace will be entabbed with $rOpts->{'entab-leading-whitespace'} spaces per tab\n"
5930 elsif ( $rOpts->{'tabs'} ) {
5931 write_logfile_entry("Indentation will be with a tab character\n");
5934 write_logfile_entry(
5935 "Indentation will be with $rOpts->{'indent-columns'} spaces\n");
5938 # This was the start of a formatter referent, but object-oriented
5939 # coding has turned out to be too slow here.
5940 $formatter_self = {};
5942 bless $formatter_self, $class;
5944 # Safety check..this is not a class yet
5945 if ( _increment_count() > 1 ) {
5947 "Attempt to create more than 1 object in $class, which is not a true class yet\n";
5949 return $formatter_self;
5952 sub prepare_for_new_input_lines {
5954 $gnu_sequence_number++; # increment output batch counter
5955 %last_gnu_equals = ();
5956 %gnu_comma_count = ();
5957 %gnu_arrow_count = ();
5958 $line_start_index_to_go = 0;
5959 $max_gnu_item_index = UNDEFINED_INDEX;
5960 $index_max_forced_break = UNDEFINED_INDEX;
5961 $max_index_to_go = UNDEFINED_INDEX;
5962 $last_nonblank_index_to_go = UNDEFINED_INDEX;
5963 $last_nonblank_type_to_go = '';
5964 $last_nonblank_token_to_go = '';
5965 $last_last_nonblank_index_to_go = UNDEFINED_INDEX;
5966 $last_last_nonblank_type_to_go = '';
5967 $last_last_nonblank_token_to_go = '';
5968 $forced_breakpoint_count = 0;
5969 $forced_breakpoint_undo_count = 0;
5970 $rbrace_follower = undef;
5971 $lengths_to_go[0] = 0;
5972 $old_line_count_in_batch = 1;
5973 $comma_count_in_batch = 0;
5974 $starting_in_quote = 0;
5976 destroy_one_line_block();
5982 my ($line_of_tokens) = @_;
5984 my $line_type = $line_of_tokens->{_line_type};
5985 my $input_line = $line_of_tokens->{_line_text};
5987 # _line_type codes are:
5988 # SYSTEM - system-specific code before hash-bang line
5989 # CODE - line of perl code (including comments)
5990 # POD_START - line starting pod, such as '=head'
5991 # POD - pod documentation text
5992 # POD_END - last line of pod section, '=cut'
5993 # HERE - text of here-document
5994 # HERE_END - last line of here-doc (target word)
5995 # FORMAT - format section
5996 # FORMAT_END - last line of format section, '.'
5997 # DATA_START - __DATA__ line
5998 # DATA - unidentified text following __DATA__
5999 # END_START - __END__ line
6000 # END - unidentified text following __END__
6001 # ERROR - we are in big trouble, probably not a perl script
6003 # put a blank line after an =cut which comes before __END__ and __DATA__
6004 # (required by podchecker)
6005 if ( $last_line_type eq 'POD_END' && !$saw_END_or_DATA_ ) {
6006 $file_writer_object->reset_consecutive_blank_lines();
6007 if ( $input_line !~ /^\s*$/ ) { want_blank_line() }
6010 # handle line of code..
6011 if ( $line_type eq 'CODE' ) {
6013 # let logger see all non-blank lines of code
6014 if ( $input_line !~ /^\s*$/ ) {
6015 my $output_line_number =
6016 $vertical_aligner_object->get_output_line_number();
6017 black_box( $line_of_tokens, $output_line_number );
6019 print_line_of_tokens($line_of_tokens);
6022 # handle line of non-code..
6028 if ( $line_type =~ /^POD/ ) {
6030 # Pod docs should have a preceding blank line. But be
6031 # very careful in __END__ and __DATA__ sections, because:
6032 # 1. the user may be using this section for any purpose whatsoever
6033 # 2. the blank counters are not active there
6034 # It should be safe to request a blank line between an
6035 # __END__ or __DATA__ and an immediately following '=head'
6036 # type line, (types END_START and DATA_START), but not for
6037 # any other lines of type END or DATA.
6038 if ( $rOpts->{'delete-pod'} ) { $skip_line = 1; }
6039 if ( $rOpts->{'tee-pod'} ) { $tee_line = 1; }
6041 && $line_type eq 'POD_START'
6042 && $last_line_type !~ /^(END|DATA)$/ )
6048 # leave the blank counters in a predictable state
6049 # after __END__ or __DATA__
6050 elsif ( $line_type =~ /^(END_START|DATA_START)$/ ) {
6051 $file_writer_object->reset_consecutive_blank_lines();
6052 $saw_END_or_DATA_ = 1;
6055 # write unindented non-code line
6056 if ( !$skip_line ) {
6057 if ($tee_line) { $file_writer_object->tee_on() }
6058 write_unindented_line($input_line);
6059 if ($tee_line) { $file_writer_object->tee_off() }
6062 $last_line_type = $line_type;
6065 sub create_one_line_block {
6066 $index_start_one_line_block = $_[0];
6067 $semicolons_before_block_self_destruct = $_[1];
6070 sub destroy_one_line_block {
6071 $index_start_one_line_block = UNDEFINED_INDEX;
6072 $semicolons_before_block_self_destruct = 0;
6075 sub leading_spaces_to_go {
6077 # return the number of indentation spaces for a token in the output stream;
6078 # these were previously stored by 'set_leading_whitespace'.
6080 return get_SPACES( $leading_spaces_to_go[ $_[0] ] );
6086 # return the number of leading spaces associated with an indentation
6087 # variable $indentation is either a constant number of spaces or an object
6088 # with a get_SPACES method.
6089 my $indentation = shift;
6090 return ref($indentation) ? $indentation->get_SPACES() : $indentation;
6093 sub get_RECOVERABLE_SPACES {
6095 # return the number of spaces (+ means shift right, - means shift left)
6096 # that we would like to shift a group of lines with the same indentation
6097 # to get them to line up with their opening parens
6098 my $indentation = shift;
6099 return ref($indentation) ? $indentation->get_RECOVERABLE_SPACES() : 0;
6102 sub get_AVAILABLE_SPACES_to_go {
6104 my $item = $leading_spaces_to_go[ $_[0] ];
6106 # return the number of available leading spaces associated with an
6107 # indentation variable. $indentation is either a constant number of
6108 # spaces or an object with a get_AVAILABLE_SPACES method.
6109 return ref($item) ? $item->get_AVAILABLE_SPACES() : 0;
6112 sub new_lp_indentation_item {
6114 # this is an interface to the IndentationItem class
6115 my ( $spaces, $level, $ci_level, $available_spaces, $align_paren ) = @_;
6117 # A negative level implies not to store the item in the item_list
6119 if ( $level >= 0 ) { $index = ++$max_gnu_item_index; }
6121 my $item = Perl::Tidy::IndentationItem->new(
6123 $ci_level, $available_spaces,
6124 $index, $gnu_sequence_number,
6125 $align_paren, $max_gnu_stack_index,
6126 $line_start_index_to_go,
6129 if ( $level >= 0 ) {
6130 $gnu_item_list[$max_gnu_item_index] = $item;
6136 sub set_leading_whitespace {
6138 # This routine defines leading whitespace
6139 # given: the level and continuation_level of a token,
6140 # define: space count of leading string which would apply if it
6141 # were the first token of a new line.
6143 my ( $level, $ci_level, $in_continued_quote ) = @_;
6145 # modify for -bli, which adds one continuation indentation for
6147 if ( $rOpts_brace_left_and_indent
6148 && $max_index_to_go == 0
6149 && $block_type_to_go[$max_index_to_go] =~ /$bli_pattern/o )
6154 # patch to avoid trouble when input file has negative indentation.
6155 # other logic should catch this error.
6156 if ( $level < 0 ) { $level = 0 }
6158 #-------------------------------------------
6159 # handle the standard indentation scheme
6160 #-------------------------------------------
6161 unless ($rOpts_line_up_parentheses) {
6163 $ci_level * $rOpts_continuation_indentation +
6164 $level * $rOpts_indent_columns;
6166 ( $ci_level == 0 ) ? 0 : $rOpts_continuation_indentation;
6168 if ($in_continued_quote) {
6172 $leading_spaces_to_go[$max_index_to_go] = $space_count;
6173 $reduced_spaces_to_go[$max_index_to_go] = $space_count - $ci_spaces;
6177 #-------------------------------------------------------------
6178 # handle case of -lp indentation..
6179 #-------------------------------------------------------------
6181 # The continued_quote flag means that this is the first token of a
6182 # line, and it is the continuation of some kind of multi-line quote
6183 # or pattern. It requires special treatment because it must have no
6184 # added leading whitespace. So we create a special indentation item
6185 # which is not in the stack.
6186 if ($in_continued_quote) {
6187 my $space_count = 0;
6188 my $available_space = 0;
6189 $level = -1; # flag to prevent storing in item_list
6190 $leading_spaces_to_go[$max_index_to_go] =
6191 $reduced_spaces_to_go[$max_index_to_go] =
6192 new_lp_indentation_item( $space_count, $level, $ci_level,
6193 $available_space, 0 );
6197 # get the top state from the stack
6198 my $space_count = $gnu_stack[$max_gnu_stack_index]->get_SPACES();
6199 my $current_level = $gnu_stack[$max_gnu_stack_index]->get_LEVEL();
6200 my $current_ci_level = $gnu_stack[$max_gnu_stack_index]->get_CI_LEVEL();
6202 my $type = $types_to_go[$max_index_to_go];
6203 my $token = $tokens_to_go[$max_index_to_go];
6204 my $total_depth = $nesting_depth_to_go[$max_index_to_go];
6206 if ( $type eq '{' || $type eq '(' ) {
6208 $gnu_comma_count{ $total_depth + 1 } = 0;
6209 $gnu_arrow_count{ $total_depth + 1 } = 0;
6211 # If we come to an opening token after an '=' token of some type,
6212 # see if it would be helpful to 'break' after the '=' to save space
6213 my $last_equals = $last_gnu_equals{$total_depth};
6214 if ( $last_equals && $last_equals > $line_start_index_to_go ) {
6216 # find the position if we break at the '='
6217 my $i_test = $last_equals;
6218 if ( $types_to_go[ $i_test + 1 ] eq 'b' ) { $i_test++ }
6221 ##my $too_close = ($i_test==$max_index_to_go-1);
6223 my $test_position = total_line_length( $i_test, $max_index_to_go );
6227 # the equals is not just before an open paren (testing)
6230 # if we are beyond the midpoint
6231 $gnu_position_predictor > $half_maximum_line_length
6233 # or we are beyont the 1/4 point and there was an old
6234 # break at the equals
6236 $gnu_position_predictor > $half_maximum_line_length / 2
6238 $old_breakpoint_to_go[$last_equals]
6239 || ( $last_equals > 0
6240 && $old_breakpoint_to_go[ $last_equals - 1 ] )
6241 || ( $last_equals > 1
6242 && $types_to_go[ $last_equals - 1 ] eq 'b'
6243 && $old_breakpoint_to_go[ $last_equals - 2 ] )
6249 # then make the switch -- note that we do not set a real
6250 # breakpoint here because we may not really need one; sub
6251 # scan_list will do that if necessary
6252 $line_start_index_to_go = $i_test + 1;
6253 $gnu_position_predictor = $test_position;
6258 # Check for decreasing depth ..
6259 # Note that one token may have both decreasing and then increasing
6260 # depth. For example, (level, ci) can go from (1,1) to (2,0). So,
6261 # in this example we would first go back to (1,0) then up to (2,0)
6263 if ( $level < $current_level || $ci_level < $current_ci_level ) {
6265 # loop to find the first entry at or completely below this level
6266 my ( $lev, $ci_lev );
6268 if ($max_gnu_stack_index) {
6270 # save index of token which closes this level
6271 $gnu_stack[$max_gnu_stack_index]->set_CLOSED($max_index_to_go);
6273 # Undo any extra indentation if we saw no commas
6274 my $available_spaces =
6275 $gnu_stack[$max_gnu_stack_index]->get_AVAILABLE_SPACES();
6277 my $comma_count = 0;
6278 my $arrow_count = 0;
6279 if ( $type eq '}' || $type eq ')' ) {
6280 $comma_count = $gnu_comma_count{$total_depth};
6281 $arrow_count = $gnu_arrow_count{$total_depth};
6282 $comma_count = 0 unless $comma_count;
6283 $arrow_count = 0 unless $arrow_count;
6285 $gnu_stack[$max_gnu_stack_index]->set_COMMA_COUNT($comma_count);
6286 $gnu_stack[$max_gnu_stack_index]->set_ARROW_COUNT($arrow_count);
6288 if ( $available_spaces > 0 ) {
6290 if ( $comma_count <= 0 || $arrow_count > 0 ) {
6292 my $i = $gnu_stack[$max_gnu_stack_index]->get_INDEX();
6294 $gnu_stack[$max_gnu_stack_index]
6295 ->get_SEQUENCE_NUMBER();
6297 # Be sure this item was created in this batch. This
6298 # should be true because we delete any available
6299 # space from open items at the end of each batch.
6300 if ( $gnu_sequence_number != $seqno
6301 || $i > $max_gnu_item_index )
6304 "Program bug with -lp. seqno=$seqno should be $gnu_sequence_number and i=$i should be less than max=$max_gnu_item_index\n"
6306 report_definite_bug();
6310 if ( $arrow_count == 0 ) {
6312 ->permanently_decrease_AVAILABLE_SPACES(
6317 ->tentatively_decrease_AVAILABLE_SPACES(
6324 $j <= $max_gnu_item_index ;
6329 ->decrease_SPACES($available_spaces);
6336 --$max_gnu_stack_index;
6337 $lev = $gnu_stack[$max_gnu_stack_index]->get_LEVEL();
6338 $ci_lev = $gnu_stack[$max_gnu_stack_index]->get_CI_LEVEL();
6340 # stop when we reach a level at or below the current level
6341 if ( $lev <= $level && $ci_lev <= $ci_level ) {
6343 $gnu_stack[$max_gnu_stack_index]->get_SPACES();
6344 $current_level = $lev;
6345 $current_ci_level = $ci_lev;
6350 # reached bottom of stack .. should never happen because
6351 # only negative levels can get here, and $level was forced
6352 # to be positive above.
6355 "program bug with -lp: stack_error. level=$level; lev=$lev; ci_level=$ci_level; ci_lev=$ci_lev; rerun with -nlp\n"
6357 report_definite_bug();
6363 # handle increasing depth
6364 if ( $level > $current_level || $ci_level > $current_ci_level ) {
6366 # Compute the standard incremental whitespace. This will be
6367 # the minimum incremental whitespace that will be used. This
6368 # choice results in a smooth transition between the gnu-style
6369 # and the standard style.
6370 my $standard_increment =
6371 ( $level - $current_level ) * $rOpts_indent_columns +
6372 ( $ci_level - $current_ci_level ) * $rOpts_continuation_indentation;
6374 # Now we have to define how much extra incremental space
6375 # ("$available_space") we want. This extra space will be
6376 # reduced as necessary when long lines are encountered or when
6377 # it becomes clear that we do not have a good list.
6378 my $available_space = 0;
6379 my $align_paren = 0;
6382 # initialization on empty stack..
6383 if ( $max_gnu_stack_index == 0 ) {
6384 $space_count = $level * $rOpts_indent_columns;
6387 # if this is a BLOCK, add the standard increment
6388 elsif ($last_nonblank_block_type) {
6389 $space_count += $standard_increment;
6392 # if last nonblank token was not structural indentation,
6393 # just use standard increment
6394 elsif ( $last_nonblank_type ne '{' ) {
6395 $space_count += $standard_increment;
6398 # otherwise use the space to the first non-blank level change token
6401 $space_count = $gnu_position_predictor;
6403 my $min_gnu_indentation =
6404 $gnu_stack[$max_gnu_stack_index]->get_SPACES();
6406 $available_space = $space_count - $min_gnu_indentation;
6407 if ( $available_space >= $standard_increment ) {
6408 $min_gnu_indentation += $standard_increment;
6410 elsif ( $available_space > 1 ) {
6411 $min_gnu_indentation += $available_space + 1;
6413 elsif ( $last_nonblank_token =~ /^[\{\[\(]$/ ) {
6414 if ( ( $tightness{$last_nonblank_token} < 2 ) ) {
6415 $min_gnu_indentation += 2;
6418 $min_gnu_indentation += 1;
6422 $min_gnu_indentation += $standard_increment;
6424 $available_space = $space_count - $min_gnu_indentation;
6426 if ( $available_space < 0 ) {
6427 $space_count = $min_gnu_indentation;
6428 $available_space = 0;
6433 # update state, but not on a blank token
6434 if ( $types_to_go[$max_index_to_go] ne 'b' ) {
6436 $gnu_stack[$max_gnu_stack_index]->set_HAVE_CHILD(1);
6438 ++$max_gnu_stack_index;
6439 $gnu_stack[$max_gnu_stack_index] =
6440 new_lp_indentation_item( $space_count, $level, $ci_level,
6441 $available_space, $align_paren );
6443 # If the opening paren is beyond the half-line length, then
6444 # we will use the minimum (standard) indentation. This will
6445 # help avoid problems associated with running out of space
6446 # near the end of a line. As a result, in deeply nested
6447 # lists, there will be some indentations which are limited
6448 # to this minimum standard indentation. But the most deeply
6449 # nested container will still probably be able to shift its
6450 # parameters to the right for proper alignment, so in most
6451 # cases this will not be noticable.
6452 if ( $available_space > 0
6453 && $space_count > $half_maximum_line_length )
6455 $gnu_stack[$max_gnu_stack_index]
6456 ->tentatively_decrease_AVAILABLE_SPACES($available_space);
6461 # Count commas and look for non-list characters. Once we see a
6462 # non-list character, we give up and don't look for any more commas.
6463 if ( $type eq '=>' ) {
6464 $gnu_arrow_count{$total_depth}++;
6466 # tentatively treating '=>' like '=' for estimating breaks
6467 # TODO: this could use some experimentation
6468 $last_gnu_equals{$total_depth} = $max_index_to_go;
6471 elsif ( $type eq ',' ) {
6472 $gnu_comma_count{$total_depth}++;
6475 elsif ( $is_assignment{$type} ) {
6476 $last_gnu_equals{$total_depth} = $max_index_to_go;
6479 # this token might start a new line
6480 # if this is a non-blank..
6481 if ( $type ne 'b' ) {
6486 # this is the first nonblank token of the line
6487 $max_index_to_go == 1 && $types_to_go[0] eq 'b'
6489 # or previous character was one of these:
6490 || $last_nonblank_type_to_go =~ /^([\:\?\,f])$/
6492 # or previous character was opening and this does not close it
6493 || ( $last_nonblank_type_to_go eq '{' && $type ne '}' )
6494 || ( $last_nonblank_type_to_go eq '(' and $type ne ')' )
6496 # or this token is one of these:
6497 || $type =~ /^([\.]|\|\||\&\&)$/
6499 # or this is a closing structure
6500 || ( $last_nonblank_type_to_go eq '}'
6501 && $last_nonblank_token_to_go eq $last_nonblank_type_to_go )
6503 # or previous token was keyword 'return'
6504 || ( $last_nonblank_type_to_go eq 'k'
6505 && ( $last_nonblank_token_to_go eq 'return' && $type ne '{' ) )
6507 # or starting a new line at certain keywords is fine
6509 && $is_if_unless_and_or_last_next_redo_return{$token} )
6511 # or this is after an assignment after a closing structure
6513 $is_assignment{$last_nonblank_type_to_go}
6515 $last_last_nonblank_type_to_go =~ /^[\}\)\]]$/
6517 # and it is significantly to the right
6518 || $gnu_position_predictor > $half_maximum_line_length
6523 check_for_long_gnu_style_lines();
6524 $line_start_index_to_go = $max_index_to_go;
6526 # back up 1 token if we want to break before that type
6527 # otherwise, we may strand tokens like '?' or ':' on a line
6528 if ( $line_start_index_to_go > 0 ) {
6529 if ( $last_nonblank_type_to_go eq 'k' ) {
6531 if ( $want_break_before{$last_nonblank_token_to_go} ) {
6532 $line_start_index_to_go--;
6535 elsif ( $want_break_before{$last_nonblank_type_to_go} ) {
6536 $line_start_index_to_go--;
6542 # remember the predicted position of this token on the output line
6543 if ( $max_index_to_go > $line_start_index_to_go ) {
6544 $gnu_position_predictor =
6545 total_line_length( $line_start_index_to_go, $max_index_to_go );
6548 $gnu_position_predictor = $space_count +
6549 token_sequence_length( $max_index_to_go, $max_index_to_go );
6552 # store the indentation object for this token
6553 # this allows us to manipulate the leading whitespace
6554 # (in case we have to reduce indentation to fit a line) without
6555 # having to change any token values
6556 $leading_spaces_to_go[$max_index_to_go] = $gnu_stack[$max_gnu_stack_index];
6557 $reduced_spaces_to_go[$max_index_to_go] =
6558 ( $max_gnu_stack_index > 0 && $ci_level )
6559 ? $gnu_stack[ $max_gnu_stack_index - 1 ]
6560 : $gnu_stack[$max_gnu_stack_index];
6564 sub check_for_long_gnu_style_lines {
6566 # look at the current estimated maximum line length, and
6567 # remove some whitespace if it exceeds the desired maximum
6569 # this is only for the '-lp' style
6570 return unless ($rOpts_line_up_parentheses);
6572 # nothing can be done if no stack items defined for this line
6573 return if ( $max_gnu_item_index == UNDEFINED_INDEX );
6575 # see if we have exceeded the maximum desired line length
6576 # keep 2 extra free because they are needed in some cases
6577 # (result of trial-and-error testing)
6579 $gnu_position_predictor - $rOpts_maximum_line_length + 2;
6581 return if ( $spaces_needed < 0 );
6583 # We are over the limit, so try to remove a requested number of
6584 # spaces from leading whitespace. We are only allowed to remove
6585 # from whitespace items created on this batch, since others have
6586 # already been used and cannot be undone.
6587 my @candidates = ();
6590 # loop over all whitespace items created for the current batch
6591 for ( $i = 0 ; $i <= $max_gnu_item_index ; $i++ ) {
6592 my $item = $gnu_item_list[$i];
6594 # item must still be open to be a candidate (otherwise it
6595 # cannot influence the current token)
6596 next if ( $item->get_CLOSED() >= 0 );
6598 my $available_spaces = $item->get_AVAILABLE_SPACES();
6600 if ( $available_spaces > 0 ) {
6601 push( @candidates, [ $i, $available_spaces ] );
6605 return unless (@candidates);
6607 # sort by available whitespace so that we can remove whitespace
6608 # from the maximum available first
6609 @candidates = sort { $b->[1] <=> $a->[1] } @candidates;
6611 # keep removing whitespace until we are done or have no more
6613 foreach $candidate (@candidates) {
6614 my ( $i, $available_spaces ) = @{$candidate};
6615 my $deleted_spaces =
6616 ( $available_spaces > $spaces_needed )
6618 : $available_spaces;
6620 # remove the incremental space from this item
6621 $gnu_item_list[$i]->decrease_AVAILABLE_SPACES($deleted_spaces);
6625 # update the leading whitespace of this item and all items
6626 # that came after it
6627 for ( ; $i <= $max_gnu_item_index ; $i++ ) {
6629 my $old_spaces = $gnu_item_list[$i]->get_SPACES();
6630 if ( $old_spaces > $deleted_spaces ) {
6631 $gnu_item_list[$i]->decrease_SPACES($deleted_spaces);
6634 # shouldn't happen except for code bug:
6636 my $level = $gnu_item_list[$i_debug]->get_LEVEL();
6637 my $ci_level = $gnu_item_list[$i_debug]->get_CI_LEVEL();
6638 my $old_level = $gnu_item_list[$i]->get_LEVEL();
6639 my $old_ci_level = $gnu_item_list[$i]->get_CI_LEVEL();
6641 "program bug with -lp: want to delete $deleted_spaces from item $i, but old=$old_spaces deleted: lev=$level ci=$ci_level deleted: level=$old_level ci=$ci_level\n"
6643 report_definite_bug();
6646 $gnu_position_predictor -= $deleted_spaces;
6647 $spaces_needed -= $deleted_spaces;
6648 last unless ( $spaces_needed > 0 );
6652 sub finish_lp_batch {
6654 # This routine is called once after each each output stream batch is
6655 # finished to undo indentation for all incomplete -lp
6656 # indentation levels. It is too risky to leave a level open,
6657 # because then we can't backtrack in case of a long line to follow.
6658 # This means that comments and blank lines will disrupt this
6659 # indentation style. But the vertical aligner may be able to
6660 # get the space back if there are side comments.
6662 # this is only for the 'lp' style
6663 return unless ($rOpts_line_up_parentheses);
6665 # nothing can be done if no stack items defined for this line
6666 return if ( $max_gnu_item_index == UNDEFINED_INDEX );
6668 # loop over all whitespace items created for the current batch
6670 for ( $i = 0 ; $i <= $max_gnu_item_index ; $i++ ) {
6671 my $item = $gnu_item_list[$i];
6673 # only look for open items
6674 next if ( $item->get_CLOSED() >= 0 );
6676 # Tentatively remove all of the available space
6677 # (The vertical aligner will try to get it back later)
6678 my $available_spaces = $item->get_AVAILABLE_SPACES();
6679 if ( $available_spaces > 0 ) {
6681 # delete incremental space for this item
6683 ->tentatively_decrease_AVAILABLE_SPACES($available_spaces);
6685 # Reduce the total indentation space of any nodes that follow
6686 # Note that any such nodes must necessarily be dependents
6688 foreach ( $i + 1 .. $max_gnu_item_index ) {
6689 $gnu_item_list[$_]->decrease_SPACES($available_spaces);
6696 sub reduce_lp_indentation {
6698 # reduce the leading whitespace at token $i if possible by $spaces_needed
6699 # (a large value of $spaces_needed will remove all excess space)
6700 # NOTE: to be called from scan_list only for a sequence of tokens
6701 # contained between opening and closing parens/braces/brackets
6703 my ( $i, $spaces_wanted ) = @_;
6704 my $deleted_spaces = 0;
6706 my $item = $leading_spaces_to_go[$i];
6707 my $available_spaces = $item->get_AVAILABLE_SPACES();
6710 $available_spaces > 0
6711 && ( ( $spaces_wanted <= $available_spaces )
6712 || !$item->get_HAVE_CHILD() )
6716 # we'll remove these spaces, but mark them as recoverable
6718 $item->tentatively_decrease_AVAILABLE_SPACES($spaces_wanted);
6721 return $deleted_spaces;
6724 sub token_sequence_length {
6726 # return length of tokens ($ifirst .. $ilast) including first & last
6727 # returns 0 if $ifirst > $ilast
6730 return 0 if ( $ilast < 0 || $ifirst > $ilast );
6731 return $lengths_to_go[ $ilast + 1 ] if ( $ifirst < 0 );
6732 return $lengths_to_go[ $ilast + 1 ] - $lengths_to_go[$ifirst];
6735 sub total_line_length {
6737 # return length of a line of tokens ($ifirst .. $ilast)
6740 if ( $ifirst < 0 ) { $ifirst = 0 }
6742 return leading_spaces_to_go($ifirst) +
6743 token_sequence_length( $ifirst, $ilast );
6746 sub excess_line_length {
6748 # return number of characters by which a line of tokens ($ifirst..$ilast)
6749 # exceeds the allowable line length.
6752 if ( $ifirst < 0 ) { $ifirst = 0 }
6753 return leading_spaces_to_go($ifirst) +
6754 token_sequence_length( $ifirst, $ilast ) - $rOpts_maximum_line_length;
6757 sub finish_formatting {
6759 # flush buffer and write any informative messages
6763 $file_writer_object->decrement_output_line_number()
6764 ; # fix up line number since it was incremented
6765 we_are_at_the_last_line();
6766 if ( $added_semicolon_count > 0 ) {
6767 my $first = ( $added_semicolon_count > 1 ) ? "First" : "";
6769 ( $added_semicolon_count > 1 ) ? "semicolons were" : "semicolon was";
6770 write_logfile_entry("$added_semicolon_count $what added:\n");
6771 write_logfile_entry(
6772 " $first at input line $first_added_semicolon_at\n");
6774 if ( $added_semicolon_count > 1 ) {
6775 write_logfile_entry(
6776 " Last at input line $last_added_semicolon_at\n");
6778 write_logfile_entry(" (Use -nasc to prevent semicolon addition)\n");
6779 write_logfile_entry("\n");
6782 if ( $deleted_semicolon_count > 0 ) {
6783 my $first = ( $deleted_semicolon_count > 1 ) ? "First" : "";
6785 ( $deleted_semicolon_count > 1 )
6788 write_logfile_entry(
6789 "$deleted_semicolon_count unnecessary $what deleted:\n");
6790 write_logfile_entry(
6791 " $first at input line $first_deleted_semicolon_at\n");
6793 if ( $deleted_semicolon_count > 1 ) {
6794 write_logfile_entry(
6795 " Last at input line $last_deleted_semicolon_at\n");
6797 write_logfile_entry(" (Use -ndsc to prevent semicolon deletion)\n");
6798 write_logfile_entry("\n");
6801 if ( $embedded_tab_count > 0 ) {
6802 my $first = ( $embedded_tab_count > 1 ) ? "First" : "";
6804 ( $embedded_tab_count > 1 )
6805 ? "quotes or patterns"
6806 : "quote or pattern";
6807 write_logfile_entry("$embedded_tab_count $what had embedded tabs:\n");
6808 write_logfile_entry(
6809 "This means the display of this script could vary with device or software\n"
6811 write_logfile_entry(" $first at input line $first_embedded_tab_at\n");
6813 if ( $embedded_tab_count > 1 ) {
6814 write_logfile_entry(
6815 " Last at input line $last_embedded_tab_at\n");
6817 write_logfile_entry("\n");
6820 if ($first_tabbing_disagreement) {
6821 write_logfile_entry(
6822 "First indentation disagreement seen at input line $first_tabbing_disagreement\n"
6826 if ($in_tabbing_disagreement) {
6827 write_logfile_entry(
6828 "Ending with indentation disagreement which started at input line $in_tabbing_disagreement\n"
6833 if ($last_tabbing_disagreement) {
6835 write_logfile_entry(
6836 "Last indentation disagreement seen at input line $last_tabbing_disagreement\n"
6840 write_logfile_entry("No indentation disagreement seen\n");
6843 write_logfile_entry("\n");
6845 $vertical_aligner_object->report_anything_unusual();
6847 $file_writer_object->report_line_length_errors();
6852 # This routine is called to check the Opts hash after it is defined
6855 my ( $tabbing_string, $tab_msg );
6857 make_static_block_comment_pattern();
6858 make_static_side_comment_pattern();
6859 make_closing_side_comment_prefix();
6860 make_closing_side_comment_list_pattern();
6861 $format_skipping_pattern_begin =
6862 make_format_skipping_pattern( 'format-skipping-begin', '#<<<' );
6863 $format_skipping_pattern_end =
6864 make_format_skipping_pattern( 'format-skipping-end', '#>>>' );
6866 # If closing side comments ARE selected, then we can safely
6867 # delete old closing side comments unless closing side comment
6868 # warnings are requested. This is a good idea because it will
6869 # eliminate any old csc's which fall below the line count threshold.
6870 # We cannot do this if warnings are turned on, though, because we
6871 # might delete some text which has been added. So that must
6872 # be handled when comments are created.
6873 if ( $rOpts->{'closing-side-comments'} ) {
6874 if ( !$rOpts->{'closing-side-comment-warnings'} ) {
6875 $rOpts->{'delete-closing-side-comments'} = 1;
6879 # If closing side comments ARE NOT selected, but warnings ARE
6880 # selected and we ARE DELETING csc's, then we will pretend to be
6881 # adding with a huge interval. This will force the comments to be
6882 # generated for comparison with the old comments, but not added.
6883 elsif ( $rOpts->{'closing-side-comment-warnings'} ) {
6884 if ( $rOpts->{'delete-closing-side-comments'} ) {
6885 $rOpts->{'delete-closing-side-comments'} = 0;
6886 $rOpts->{'closing-side-comments'} = 1;
6887 $rOpts->{'closing-side-comment-interval'} = 100000000;
6892 make_block_brace_vertical_tightness_pattern();
6894 if ( $rOpts->{'line-up-parentheses'} ) {
6896 if ( $rOpts->{'indent-only'}
6897 || !$rOpts->{'add-newlines'}
6898 || !$rOpts->{'delete-old-newlines'} )
6901 -----------------------------------------------------------------------
6902 Conflict: -lp conflicts with -io, -fnl, -nanl, or -ndnl; ignoring -lp
6904 The -lp indentation logic requires that perltidy be able to coordinate
6905 arbitrarily large numbers of line breakpoints. This isn't possible
6906 with these flags. Sometimes an acceptable workaround is to use -wocb=3
6907 -----------------------------------------------------------------------
6909 $rOpts->{'line-up-parentheses'} = 0;
6913 # At present, tabs are not compatable with the line-up-parentheses style
6914 # (it would be possible to entab the total leading whitespace
6915 # just prior to writing the line, if desired).
6916 if ( $rOpts->{'line-up-parentheses'} && $rOpts->{'tabs'} ) {
6918 Conflict: -t (tabs) cannot be used with the -lp option; ignoring -t; see -et.
6920 $rOpts->{'tabs'} = 0;
6923 # Likewise, tabs are not compatable with outdenting..
6924 if ( $rOpts->{'outdent-keywords'} && $rOpts->{'tabs'} ) {
6926 Conflict: -t (tabs) cannot be used with the -okw options; ignoring -t; see -et.
6928 $rOpts->{'tabs'} = 0;
6931 if ( $rOpts->{'outdent-labels'} && $rOpts->{'tabs'} ) {
6933 Conflict: -t (tabs) cannot be used with the -ola option; ignoring -t; see -et.
6935 $rOpts->{'tabs'} = 0;
6938 if ( !$rOpts->{'space-for-semicolon'} ) {
6939 $want_left_space{'f'} = -1;
6942 if ( $rOpts->{'space-terminal-semicolon'} ) {
6943 $want_left_space{';'} = 1;
6946 # implement outdenting preferences for keywords
6947 %outdent_keyword = ();
6948 unless ( @_ = split_words( $rOpts->{'outdent-keyword-okl'} ) ) {
6949 @_ = qw(next last redo goto return); # defaults
6952 # FUTURE: if not a keyword, assume that it is an identifier
6954 if ( $Perl::Tidy::Tokenizer::is_keyword{$_} ) {
6955 $outdent_keyword{$_} = 1;
6958 warn "ignoring '$_' in -okwl list; not a perl keyword";
6962 # implement user whitespace preferences
6963 if ( @_ = split_words( $rOpts->{'want-left-space'} ) ) {
6964 @want_left_space{@_} = (1) x scalar(@_);
6967 if ( @_ = split_words( $rOpts->{'want-right-space'} ) ) {
6968 @want_right_space{@_} = (1) x scalar(@_);
6971 if ( @_ = split_words( $rOpts->{'nowant-left-space'} ) ) {
6972 @want_left_space{@_} = (-1) x scalar(@_);
6975 if ( @_ = split_words( $rOpts->{'nowant-right-space'} ) ) {
6976 @want_right_space{@_} = (-1) x scalar(@_);
6978 if ( $rOpts->{'dump-want-left-space'} ) {
6979 dump_want_left_space(*STDOUT);
6983 if ( $rOpts->{'dump-want-right-space'} ) {
6984 dump_want_right_space(*STDOUT);
6988 # default keywords for which space is introduced before an opening paren
6989 # (at present, including them messes up vertical alignment)
6990 @_ = qw(my local our and or err eq ne if else elsif until
6991 unless while for foreach return switch case given when);
6992 @space_after_keyword{@_} = (1) x scalar(@_);
6994 # allow user to modify these defaults
6995 if ( @_ = split_words( $rOpts->{'space-after-keyword'} ) ) {
6996 @space_after_keyword{@_} = (1) x scalar(@_);
6999 if ( @_ = split_words( $rOpts->{'nospace-after-keyword'} ) ) {
7000 @space_after_keyword{@_} = (0) x scalar(@_);
7003 # implement user break preferences
7004 my @all_operators = qw(% + - * / x != == >= <= =~ !~ < > | &
7005 = **= += *= &= <<= &&= -= /= |= >>= ||= //= .= %= ^= x=
7006 . : ? && || and or err xor
7009 my $break_after = sub {
7010 foreach my $tok (@_) {
7011 if ( $tok eq '?' ) { $tok = ':' } # patch to coordinate ?/:
7012 my $lbs = $left_bond_strength{$tok};
7013 my $rbs = $right_bond_strength{$tok};
7014 if ( defined($lbs) && defined($rbs) && $lbs < $rbs ) {
7015 ( $right_bond_strength{$tok}, $left_bond_strength{$tok} ) =
7021 my $break_before = sub {
7022 foreach my $tok (@_) {
7023 my $lbs = $left_bond_strength{$tok};
7024 my $rbs = $right_bond_strength{$tok};
7025 if ( defined($lbs) && defined($rbs) && $rbs < $lbs ) {
7026 ( $right_bond_strength{$tok}, $left_bond_strength{$tok} ) =
7032 $break_after->(@all_operators) if ( $rOpts->{'break-after-all-operators'} );
7033 $break_before->(@all_operators)
7034 if ( $rOpts->{'break-before-all-operators'} );
7036 $break_after->( split_words( $rOpts->{'want-break-after'} ) );
7037 $break_before->( split_words( $rOpts->{'want-break-before'} ) );
7039 # make note if breaks are before certain key types
7040 %want_break_before = ();
7041 foreach my $tok ( @all_operators, ',' ) {
7042 $want_break_before{$tok} =
7043 $left_bond_strength{$tok} < $right_bond_strength{$tok};
7046 # Coordinate ?/: breaks, which must be similar
7047 if ( !$want_break_before{':'} ) {
7048 $want_break_before{'?'} = $want_break_before{':'};
7049 $right_bond_strength{'?'} = $right_bond_strength{':'} + 0.01;
7050 $left_bond_strength{'?'} = NO_BREAK;
7053 # Define here tokens which may follow the closing brace of a do statement
7054 # on the same line, as in:
7055 # } while ( $something);
7056 @_ = qw(until while unless if ; : );
7058 @is_do_follower{@_} = (1) x scalar(@_);
7060 # These tokens may follow the closing brace of an if or elsif block.
7061 # In other words, for cuddled else we want code to look like:
7062 # } elsif ( $something) {
7064 if ( $rOpts->{'cuddled-else'} ) {
7065 @_ = qw(else elsif);
7066 @is_if_brace_follower{@_} = (1) x scalar(@_);
7069 %is_if_brace_follower = ();
7072 # nothing can follow the closing curly of an else { } block:
7073 %is_else_brace_follower = ();
7075 # what can follow a multi-line anonymous sub definition closing curly:
7076 @_ = qw# ; : => or and && || ~~ !~~ ) #;
7078 @is_anon_sub_brace_follower{@_} = (1) x scalar(@_);
7080 # what can follow a one-line anonynomous sub closing curly:
7081 # one-line anonumous subs also have ']' here...
7082 # see tk3.t and PP.pm
7083 @_ = qw# ; : => or and && || ) ] ~~ !~~ #;
7085 @is_anon_sub_1_brace_follower{@_} = (1) x scalar(@_);
7087 # What can follow a closing curly of a block
7088 # which is not an if/elsif/else/do/sort/map/grep/eval/sub
7089 # Testfiles: 'Toolbar.pm', 'Menubar.pm', bless.t, '3rules.pl'
7090 @_ = qw# ; : => or and && || ) #;
7093 # allow cuddled continue if cuddled else is specified
7094 if ( $rOpts->{'cuddled-else'} ) { push @_, 'continue'; }
7096 @is_other_brace_follower{@_} = (1) x scalar(@_);
7098 $right_bond_strength{'{'} = WEAK;
7099 $left_bond_strength{'{'} = VERY_STRONG;
7101 # make -l=0 equal to -l=infinite
7102 if ( !$rOpts->{'maximum-line-length'} ) {
7103 $rOpts->{'maximum-line-length'} = 1000000;
7106 # make -lbl=0 equal to -lbl=infinite
7107 if ( !$rOpts->{'long-block-line-count'} ) {
7108 $rOpts->{'long-block-line-count'} = 1000000;
7111 my $ole = $rOpts->{'output-line-ending'};
7120 unless ( $rOpts->{'output-line-ending'} = $endings{$ole} ) {
7121 my $str = join " ", keys %endings;
7123 Unrecognized line ending '$ole'; expecting one of: $str
7126 if ( $rOpts->{'preserve-line-endings'} ) {
7127 warn "Ignoring -ple; conflicts with -ole\n";
7128 $rOpts->{'preserve-line-endings'} = undef;
7132 # hashes used to simplify setting whitespace
7134 '{' => $rOpts->{'brace-tightness'},
7135 '}' => $rOpts->{'brace-tightness'},
7136 '(' => $rOpts->{'paren-tightness'},
7137 ')' => $rOpts->{'paren-tightness'},
7138 '[' => $rOpts->{'square-bracket-tightness'},
7139 ']' => $rOpts->{'square-bracket-tightness'},
7148 # frequently used parameters
7149 $rOpts_add_newlines = $rOpts->{'add-newlines'};
7150 $rOpts_add_whitespace = $rOpts->{'add-whitespace'};
7151 $rOpts_block_brace_tightness = $rOpts->{'block-brace-tightness'};
7152 $rOpts_block_brace_vertical_tightness =
7153 $rOpts->{'block-brace-vertical-tightness'};
7154 $rOpts_brace_left_and_indent = $rOpts->{'brace-left-and-indent'};
7155 $rOpts_comma_arrow_breakpoints = $rOpts->{'comma-arrow-breakpoints'};
7156 $rOpts_break_at_old_ternary_breakpoints =
7157 $rOpts->{'break-at-old-ternary-breakpoints'};
7158 $rOpts_break_at_old_comma_breakpoints =
7159 $rOpts->{'break-at-old-comma-breakpoints'};
7160 $rOpts_break_at_old_keyword_breakpoints =
7161 $rOpts->{'break-at-old-keyword-breakpoints'};
7162 $rOpts_break_at_old_logical_breakpoints =
7163 $rOpts->{'break-at-old-logical-breakpoints'};
7164 $rOpts_closing_side_comment_else_flag =
7165 $rOpts->{'closing-side-comment-else-flag'};
7166 $rOpts_closing_side_comment_maximum_text =
7167 $rOpts->{'closing-side-comment-maximum-text'};
7168 $rOpts_continuation_indentation = $rOpts->{'continuation-indentation'};
7169 $rOpts_cuddled_else = $rOpts->{'cuddled-else'};
7170 $rOpts_delete_old_whitespace = $rOpts->{'delete-old-whitespace'};
7171 $rOpts_fuzzy_line_length = $rOpts->{'fuzzy-line-length'};
7172 $rOpts_indent_columns = $rOpts->{'indent-columns'};
7173 $rOpts_line_up_parentheses = $rOpts->{'line-up-parentheses'};
7174 $rOpts_maximum_fields_per_table = $rOpts->{'maximum-fields-per-table'};
7175 $rOpts_maximum_line_length = $rOpts->{'maximum-line-length'};
7176 $rOpts_short_concatenation_item_length =
7177 $rOpts->{'short-concatenation-item-length'};
7178 $rOpts_swallow_optional_blank_lines =
7179 $rOpts->{'swallow-optional-blank-lines'};
7180 $rOpts_ignore_old_breakpoints = $rOpts->{'ignore-old-breakpoints'};
7181 $rOpts_format_skipping = $rOpts->{'format-skipping'};
7182 $rOpts_space_function_paren = $rOpts->{'space-function-paren'};
7183 $rOpts_space_keyword_paren = $rOpts->{'space-keyword-paren'};
7184 $rOpts_keep_interior_semicolons = $rOpts->{'keep-interior-semicolons'};
7185 $half_maximum_line_length = $rOpts_maximum_line_length / 2;
7187 # Note that both opening and closing tokens can access the opening
7188 # and closing flags of their container types.
7189 %opening_vertical_tightness = (
7190 '(' => $rOpts->{'paren-vertical-tightness'},
7191 '{' => $rOpts->{'brace-vertical-tightness'},
7192 '[' => $rOpts->{'square-bracket-vertical-tightness'},
7193 ')' => $rOpts->{'paren-vertical-tightness'},
7194 '}' => $rOpts->{'brace-vertical-tightness'},
7195 ']' => $rOpts->{'square-bracket-vertical-tightness'},
7198 %closing_vertical_tightness = (
7199 '(' => $rOpts->{'paren-vertical-tightness-closing'},
7200 '{' => $rOpts->{'brace-vertical-tightness-closing'},
7201 '[' => $rOpts->{'square-bracket-vertical-tightness-closing'},
7202 ')' => $rOpts->{'paren-vertical-tightness-closing'},
7203 '}' => $rOpts->{'brace-vertical-tightness-closing'},
7204 ']' => $rOpts->{'square-bracket-vertical-tightness-closing'},
7207 # assume flag for '>' same as ')' for closing qw quotes
7208 %closing_token_indentation = (
7209 ')' => $rOpts->{'closing-paren-indentation'},
7210 '}' => $rOpts->{'closing-brace-indentation'},
7211 ']' => $rOpts->{'closing-square-bracket-indentation'},
7212 '>' => $rOpts->{'closing-paren-indentation'},
7215 %opening_token_right = (
7216 '(' => $rOpts->{'opening-paren-right'},
7217 '{' => $rOpts->{'opening-hash-brace-right'},
7218 '[' => $rOpts->{'opening-square-bracket-right'},
7221 %stack_opening_token = (
7222 '(' => $rOpts->{'stack-opening-paren'},
7223 '{' => $rOpts->{'stack-opening-hash-brace'},
7224 '[' => $rOpts->{'stack-opening-square-bracket'},
7227 %stack_closing_token = (
7228 ')' => $rOpts->{'stack-closing-paren'},
7229 '}' => $rOpts->{'stack-closing-hash-brace'},
7230 ']' => $rOpts->{'stack-closing-square-bracket'},
7234 sub make_static_block_comment_pattern {
7236 # create the pattern used to identify static block comments
7237 $static_block_comment_pattern = '^\s*##';
7239 # allow the user to change it
7240 if ( $rOpts->{'static-block-comment-prefix'} ) {
7241 my $prefix = $rOpts->{'static-block-comment-prefix'};
7242 $prefix =~ s/^\s*//;
7243 my $pattern = $prefix;
7245 # user may give leading caret to force matching left comments only
7246 if ( $prefix !~ /^\^#/ ) {
7247 if ( $prefix !~ /^#/ ) {
7249 "ERROR: the -sbcp prefix is '$prefix' but must begin with '#' or '^#'\n";
7251 $pattern = '^\s*' . $prefix;
7253 eval "'##'=~/$pattern/";
7256 "ERROR: the -sbc prefix '$prefix' causes the invalid regex '$pattern'\n";
7258 $static_block_comment_pattern = $pattern;
7262 sub make_format_skipping_pattern {
7263 my ( $opt_name, $default ) = @_;
7264 my $param = $rOpts->{$opt_name};
7265 unless ($param) { $param = $default }
7267 if ( $param !~ /^#/ ) {
7268 die "ERROR: the $opt_name parameter '$param' must begin with '#'\n";
7270 my $pattern = '^' . $param . '\s';
7271 eval "'#'=~/$pattern/";
7274 "ERROR: the $opt_name parameter '$param' causes the invalid regex '$pattern'\n";
7279 sub make_closing_side_comment_list_pattern {
7281 # turn any input list into a regex for recognizing selected block types
7282 $closing_side_comment_list_pattern = '^\w+';
7283 if ( defined( $rOpts->{'closing-side-comment-list'} )
7284 && $rOpts->{'closing-side-comment-list'} )
7286 $closing_side_comment_list_pattern =
7287 make_block_pattern( '-cscl', $rOpts->{'closing-side-comment-list'} );
7291 sub make_bli_pattern {
7293 if ( defined( $rOpts->{'brace-left-and-indent-list'} )
7294 && $rOpts->{'brace-left-and-indent-list'} )
7296 $bli_list_string = $rOpts->{'brace-left-and-indent-list'};
7299 $bli_pattern = make_block_pattern( '-blil', $bli_list_string );
7302 sub make_block_brace_vertical_tightness_pattern {
7304 # turn any input list into a regex for recognizing selected block types
7305 $block_brace_vertical_tightness_pattern =
7306 '^((if|else|elsif|unless|while|for|foreach|do|\w+:)$|sub)';
7308 if ( defined( $rOpts->{'block-brace-vertical-tightness-list'} )
7309 && $rOpts->{'block-brace-vertical-tightness-list'} )
7311 $block_brace_vertical_tightness_pattern =
7312 make_block_pattern( '-bbvtl',
7313 $rOpts->{'block-brace-vertical-tightness-list'} );
7317 sub make_block_pattern {
7319 # given a string of block-type keywords, return a regex to match them
7320 # The only tricky part is that labels are indicated with a single ':'
7321 # and the 'sub' token text may have additional text after it (name of
7326 # input string: "if else elsif unless while for foreach do : sub";
7327 # pattern: '^((if|else|elsif|unless|while|for|foreach|do|\w+:)$|sub)';
7329 my ( $abbrev, $string ) = @_;
7330 my @list = split_words($string);
7336 if ( $i eq 'sub' ) {
7338 elsif ( $i eq ':' ) {
7339 push @words, '\w+:';
7341 elsif ( $i =~ /^\w/ ) {
7345 warn "unrecognized block type $i after $abbrev, ignoring\n";
7348 my $pattern = '(' . join( '|', @words ) . ')$';
7349 if ( $seen{'sub'} ) {
7350 $pattern = '(' . $pattern . '|sub)';
7352 $pattern = '^' . $pattern;
7356 sub make_static_side_comment_pattern {
7358 # create the pattern used to identify static side comments
7359 $static_side_comment_pattern = '^##';
7361 # allow the user to change it
7362 if ( $rOpts->{'static-side-comment-prefix'} ) {
7363 my $prefix = $rOpts->{'static-side-comment-prefix'};
7364 $prefix =~ s/^\s*//;
7365 my $pattern = '^' . $prefix;
7366 eval "'##'=~/$pattern/";
7369 "ERROR: the -sscp prefix '$prefix' causes the invalid regex '$pattern'\n";
7371 $static_side_comment_pattern = $pattern;
7375 sub make_closing_side_comment_prefix {
7377 # Be sure we have a valid closing side comment prefix
7378 my $csc_prefix = $rOpts->{'closing-side-comment-prefix'};
7379 my $csc_prefix_pattern;
7380 if ( !defined($csc_prefix) ) {
7381 $csc_prefix = '## end';
7382 $csc_prefix_pattern = '^##\s+end';
7385 my $test_csc_prefix = $csc_prefix;
7386 if ( $test_csc_prefix !~ /^#/ ) {
7387 $test_csc_prefix = '#' . $test_csc_prefix;
7390 # make a regex to recognize the prefix
7391 my $test_csc_prefix_pattern = $test_csc_prefix;
7393 # escape any special characters
7394 $test_csc_prefix_pattern =~ s/([^#\s\w])/\\$1/g;
7396 $test_csc_prefix_pattern = '^' . $test_csc_prefix_pattern;
7398 # allow exact number of intermediate spaces to vary
7399 $test_csc_prefix_pattern =~ s/\s+/\\s\+/g;
7401 # make sure we have a good pattern
7402 # if we fail this we probably have an error in escaping
7404 eval "'##'=~/$test_csc_prefix_pattern/";
7407 # shouldn't happen..must have screwed up escaping, above
7408 report_definite_bug();
7410 "Program Error: the -cscp prefix '$csc_prefix' caused the invalid regex '$csc_prefix_pattern'\n";
7412 # just warn and keep going with defaults
7413 warn "Please consider using a simpler -cscp prefix\n";
7414 warn "Using default -cscp instead; please check output\n";
7417 $csc_prefix = $test_csc_prefix;
7418 $csc_prefix_pattern = $test_csc_prefix_pattern;
7421 $rOpts->{'closing-side-comment-prefix'} = $csc_prefix;
7422 $closing_side_comment_prefix_pattern = $csc_prefix_pattern;
7425 sub dump_want_left_space {
7429 These values are the main control of whitespace to the left of a token type;
7430 They may be altered with the -wls parameter.
7431 For a list of token types, use perltidy --dump-token-types (-dtt)
7432 1 means the token wants a space to its left
7433 -1 means the token does not want a space to its left
7434 ------------------------------------------------------------------------
7436 foreach ( sort keys %want_left_space ) {
7437 print $fh "$_\t$want_left_space{$_}\n";
7441 sub dump_want_right_space {
7445 These values are the main control of whitespace to the right of a token type;
7446 They may be altered with the -wrs parameter.
7447 For a list of token types, use perltidy --dump-token-types (-dtt)
7448 1 means the token wants a space to its right
7449 -1 means the token does not want a space to its right
7450 ------------------------------------------------------------------------
7452 foreach ( sort keys %want_right_space ) {
7453 print $fh "$_\t$want_right_space{$_}\n";
7457 { # begin is_essential_whitespace
7459 my %is_sort_grep_map;
7464 @_ = qw(sort grep map);
7465 @is_sort_grep_map{@_} = (1) x scalar(@_);
7467 @_ = qw(for foreach);
7468 @is_for_foreach{@_} = (1) x scalar(@_);
7472 sub is_essential_whitespace {
7474 # Essential whitespace means whitespace which cannot be safely deleted
7475 # without risking the introduction of a syntax error.
7476 # We are given three tokens and their types:
7477 # ($tokenl, $typel) is the token to the left of the space in question
7478 # ($tokenr, $typer) is the token to the right of the space in question
7479 # ($tokenll, $typell) is previous nonblank token to the left of $tokenl
7481 # This is a slow routine but is not needed too often except when -mangle
7484 # Note: This routine should almost never need to be changed. It is
7485 # for avoiding syntax problems rather than for formatting.
7486 my ( $tokenll, $typell, $tokenl, $typel, $tokenr, $typer ) = @_;
7490 # never combine two bare words or numbers
7491 # examples: and ::ok(1)
7493 # for bla::bla:: abc
7494 # example is "%overload:: and" in files Dumpvalue.pm or colonbug.pl
7495 # $input eq"quit" to make $inputeq"quit"
7496 # my $size=-s::SINK if $file; <==OK but we won't do it
7497 # don't join something like: for bla::bla:: abc
7498 # example is "%overload:: and" in files Dumpvalue.pm or colonbug.pl
7499 ( ( $tokenl =~ /([\'\w]|\:\:)$/ ) && ( $tokenr =~ /^([\'\w]|\:\:)/ ) )
7501 # do not combine a number with a concatination dot
7502 # example: pom.caputo:
7503 # $vt100_compatible ? "\e[0;0H" : ('-' x 78 . "\n");
7504 || ( ( $typel eq 'n' ) && ( $tokenr eq '.' ) )
7505 || ( ( $typer eq 'n' ) && ( $tokenl eq '.' ) )
7507 # do not join a minus with a bare word, because you might form
7508 # a file test operator. Example from Complex.pm:
7509 # if (CORE::abs($z - i) < $eps); "z-i" would be taken as a file test.
7510 || ( ( $tokenl eq '-' ) && ( $tokenr =~ /^[_A-Za-z]$/ ) )
7512 # and something like this could become ambiguous without space
7514 # use constant III=>1;
7518 || ( ( $tokenl eq '-' )
7519 && ( $typer =~ /^[wC]$/ && $tokenr =~ /^[_A-Za-z]/ ) )
7521 # '= -' should not become =- or you will get a warning
7523 # || ($tokenr eq '-')
7525 # keep a space between a quote and a bareword to prevent the
7526 # bareword from becomming a quote modifier.
7527 || ( ( $typel eq 'Q' ) && ( $tokenr =~ /^[a-zA-Z_]/ ) )
7529 # keep a space between a token ending in '$' and any word;
7530 # this caused trouble: "die @$ if $@"
7531 || ( ( $typel eq 'i' && $tokenl =~ /\$$/ )
7532 && ( $tokenr =~ /^[a-zA-Z_]/ ) )
7534 # perl is very fussy about spaces before <<
7535 || ( $tokenr =~ /^\<\</ )
7537 # avoid combining tokens to create new meanings. Example:
7538 # $a+ +$b must not become $a++$b
7539 || ( $is_digraph{ $tokenl . $tokenr } )
7540 || ( $is_trigraph{ $tokenl . $tokenr } )
7542 # another example: do not combine these two &'s:
7543 # allow_options & &OPT_EXECCGI
7544 || ( $is_digraph{ $tokenl . substr( $tokenr, 0, 1 ) } )
7546 # don't combine $$ or $# with any alphanumeric
7547 # (testfile mangle.t with --mangle)
7548 || ( ( $tokenl =~ /^\$[\$\#]$/ ) && ( $tokenr =~ /^\w/ ) )
7550 # retain any space after possible filehandle
7551 # (testfiles prnterr1.t with --extrude and mangle.t with --mangle)
7552 || ( $typel eq 'Z' )
7554 # Perl is sensitive to whitespace after the + here:
7555 # $b = xvals $a + 0.1 * yvals $a;
7556 || ( $typell eq 'Z' && $typel =~ /^[\/\?\+\-\*]$/ )
7558 # keep paren separate in 'use Foo::Bar ()'
7562 && $tokenll eq 'use' )
7564 # keep any space between filehandle and paren:
7565 # file mangle.t with --mangle:
7566 || ( $typel eq 'Y' && $tokenr eq '(' )
7568 # retain any space after here doc operator ( hereerr.t)
7569 || ( $typel eq 'h' )
7571 # be careful with a space around ++ and --, to avoid ambiguity as to
7572 # which token it applies
7573 || ( ( $typer =~ /^(pp|mm)$/ ) && ( $tokenl !~ /^[\;\{\(\[]/ ) )
7574 || ( ( $typel =~ /^(\+\+|\-\-)$/ ) && ( $tokenr !~ /^[\;\}\)\]]/ ) )
7576 # need space after foreach my; for example, this will fail in
7577 # older versions of Perl:
7578 # foreach my$ft(@filetypes)...
7583 && $is_for_foreach{$tokenll}
7587 # must have space between grep and left paren; "grep(" will fail
7588 || ( $tokenr eq '(' && $is_sort_grep_map{$tokenl} )
7590 # don't stick numbers next to left parens, as in:
7591 #use Mail::Internet 1.28 (); (see Entity.pm, Head.pm, Test.pm)
7592 || ( ( $typel eq 'n' ) && ( $tokenr eq '(' ) )
7594 # We must be sure that a space between a ? and a quoted string
7595 # remains if the space before the ? remains. [Loca.pm, lockarea]
7597 # $b=join $comma ? ',' : ':', @_; # ok
7598 # $b=join $comma?',' : ':', @_; # ok!
7599 # $b=join $comma ?',' : ':', @_; # error!
7600 # Not really required:
7601 ## || ( ( $typel eq '?' ) && ( $typer eq 'Q' ) )
7603 # do not remove space between an '&' and a bare word because
7604 # it may turn into a function evaluation, like here
7605 # between '&' and 'O_ACCMODE', producing a syntax error [File.pm]
7606 # $opts{rdonly} = (($opts{mode} & O_ACCMODE) == O_RDONLY);
7607 || ( ( $typel eq '&' ) && ( $tokenr =~ /^[a-zA-Z_]/ ) )
7609 ; # the value of this long logic sequence is the result we want
7614 sub set_white_space_flag {
7616 # This routine examines each pair of nonblank tokens and
7617 # sets values for array @white_space_flag.
7619 # $white_space_flag[$j] is a flag indicating whether a white space
7620 # BEFORE token $j is needed, with the following values:
7622 # -1 do not want a space before token $j
7623 # 0 optional space or $j is a whitespace
7624 # 1 want a space before token $j
7627 # The values for the first token will be defined based
7628 # upon the contents of the "to_go" output array.
7630 # Note: retain debug print statements because they are usually
7631 # required after adding new token types.
7635 # initialize these global hashes, which control the use of
7636 # whitespace around tokens:
7641 # %space_after_keyword
7643 # Many token types are identical to the tokens themselves.
7644 # See the tokenizer for a complete list. Here are some special types:
7646 # f = semicolon in for statement
7649 # Note that :: is excluded since it should be contained in an identifier
7650 # Note that '->' is excluded because it never gets space
7651 # parentheses and brackets are excluded since they are handled specially
7652 # curly braces are included but may be overridden by logic, such as
7655 # NEW_TOKENS: create a whitespace rule here. This can be as
7656 # simple as adding your new letter to @spaces_both_sides, for
7660 @is_opening_type{@_} = (1) x scalar(@_);
7663 @is_closing_type{@_} = (1) x scalar(@_);
7665 my @spaces_both_sides = qw"
7666 + - * / % ? = . : x < > | & ^ .. << >> ** && .. || // => += -=
7667 .= %= x= &= |= ^= *= <> <= >= == =~ !~ /= != ... <<= >>= ~~ !~~
7668 &&= ||= //= <=> A k f w F n C Y U G v
7671 my @spaces_left_side = qw"
7672 t ! ~ m p { \ h pp mm Z j
7674 push( @spaces_left_side, '#' ); # avoids warning message
7676 my @spaces_right_side = qw"
7677 ; } ) ] R J ++ -- **=
7679 push( @spaces_right_side, ',' ); # avoids warning message
7680 @want_left_space{@spaces_both_sides} = (1) x scalar(@spaces_both_sides);
7681 @want_right_space{@spaces_both_sides} =
7682 (1) x scalar(@spaces_both_sides);
7683 @want_left_space{@spaces_left_side} = (1) x scalar(@spaces_left_side);
7684 @want_right_space{@spaces_left_side} = (-1) x scalar(@spaces_left_side);
7685 @want_left_space{@spaces_right_side} =
7686 (-1) x scalar(@spaces_right_side);
7687 @want_right_space{@spaces_right_side} =
7688 (1) x scalar(@spaces_right_side);
7689 $want_left_space{'L'} = WS_NO;
7690 $want_left_space{'->'} = WS_NO;
7691 $want_right_space{'->'} = WS_NO;
7692 $want_left_space{'**'} = WS_NO;
7693 $want_right_space{'**'} = WS_NO;
7695 # hash type information must stay tightly bound
7697 $binary_ws_rules{'i'}{'L'} = WS_NO;
7698 $binary_ws_rules{'i'}{'{'} = WS_YES;
7699 $binary_ws_rules{'k'}{'{'} = WS_YES;
7700 $binary_ws_rules{'U'}{'{'} = WS_YES;
7701 $binary_ws_rules{'i'}{'['} = WS_NO;
7702 $binary_ws_rules{'R'}{'L'} = WS_NO;
7703 $binary_ws_rules{'R'}{'{'} = WS_NO;
7704 $binary_ws_rules{'t'}{'L'} = WS_NO;
7705 $binary_ws_rules{'t'}{'{'} = WS_NO;
7706 $binary_ws_rules{'}'}{'L'} = WS_NO;
7707 $binary_ws_rules{'}'}{'{'} = WS_NO;
7708 $binary_ws_rules{'$'}{'L'} = WS_NO;
7709 $binary_ws_rules{'$'}{'{'} = WS_NO;
7710 $binary_ws_rules{'@'}{'L'} = WS_NO;
7711 $binary_ws_rules{'@'}{'{'} = WS_NO;
7712 $binary_ws_rules{'='}{'L'} = WS_YES;
7714 # the following includes ') {'
7715 # as in : if ( xxx ) { yyy }
7716 $binary_ws_rules{']'}{'L'} = WS_NO;
7717 $binary_ws_rules{']'}{'{'} = WS_NO;
7718 $binary_ws_rules{')'}{'{'} = WS_YES;
7719 $binary_ws_rules{')'}{'['} = WS_NO;
7720 $binary_ws_rules{']'}{'['} = WS_NO;
7721 $binary_ws_rules{']'}{'{'} = WS_NO;
7722 $binary_ws_rules{'}'}{'['} = WS_NO;
7723 $binary_ws_rules{'R'}{'['} = WS_NO;
7725 $binary_ws_rules{']'}{'++'} = WS_NO;
7726 $binary_ws_rules{']'}{'--'} = WS_NO;
7727 $binary_ws_rules{')'}{'++'} = WS_NO;
7728 $binary_ws_rules{')'}{'--'} = WS_NO;
7730 $binary_ws_rules{'R'}{'++'} = WS_NO;
7731 $binary_ws_rules{'R'}{'--'} = WS_NO;
7733 ########################################################
7734 # should no longer be necessary (see niek.pl)
7735 ##$binary_ws_rules{'k'}{':'} = WS_NO; # keep colon with label
7736 ##$binary_ws_rules{'w'}{':'} = WS_NO;
7737 ########################################################
7738 $binary_ws_rules{'i'}{'Q'} = WS_YES;
7739 $binary_ws_rules{'n'}{'('} = WS_YES; # occurs in 'use package n ()'
7741 # FIXME: we need to split 'i' into variables and functions
7742 # and have no space for functions but space for variables. For now,
7743 # I have a special patch in the special rules below
7744 $binary_ws_rules{'i'}{'('} = WS_NO;
7746 $binary_ws_rules{'w'}{'('} = WS_NO;
7747 $binary_ws_rules{'w'}{'{'} = WS_YES;
7749 my ( $jmax, $rtokens, $rtoken_type, $rblock_type ) = @_;
7750 my ( $last_token, $last_type, $last_block_type, $token, $type,
7752 my (@white_space_flag);
7753 my $j_tight_closing_paren = -1;
7755 if ( $max_index_to_go >= 0 ) {
7756 $token = $tokens_to_go[$max_index_to_go];
7757 $type = $types_to_go[$max_index_to_go];
7758 $block_type = $block_type_to_go[$max_index_to_go];
7766 # loop over all tokens
7769 for ( $j = 0 ; $j <= $jmax ; $j++ ) {
7771 if ( $$rtoken_type[$j] eq 'b' ) {
7772 $white_space_flag[$j] = WS_OPTIONAL;
7776 # set a default value, to be changed as needed
7778 $last_token = $token;
7780 $last_block_type = $block_type;
7781 $token = $$rtokens[$j];
7782 $type = $$rtoken_type[$j];
7783 $block_type = $$rblock_type[$j];
7785 #---------------------------------------------------------------
7787 # handle space on the inside of opening braces
7788 #---------------------------------------------------------------
7791 if ( $is_opening_type{$last_type} ) {
7793 $j_tight_closing_paren = -1;
7795 # let's keep empty matched braces together: () {} []
7797 if ( $token eq $matching_token{$last_token} ) {
7807 # we're considering the right of an opening brace
7808 # tightness = 0 means always pad inside with space
7809 # tightness = 1 means pad inside if "complex"
7810 # tightness = 2 means never pad inside with space
7813 if ( $last_type eq '{'
7814 && $last_token eq '{'
7815 && $last_block_type )
7817 $tightness = $rOpts_block_brace_tightness;
7819 else { $tightness = $tightness{$last_token} }
7821 if ( $tightness <= 0 ) {
7824 elsif ( $tightness > 1 ) {
7829 # Patch to count '-foo' as single token so that
7830 # each of $a{-foo} and $a{foo} and $a{'foo'} do
7831 # not get spaces with default formatting.
7835 && $last_token eq '{'
7836 && $$rtoken_type[ $j + 1 ] eq 'w' );
7838 # $j_next is where a closing token should be if
7839 # the container has a single token
7841 ( $$rtoken_type[ $j_here + 1 ] eq 'b' )
7844 my $tok_next = $$rtokens[$j_next];
7845 my $type_next = $$rtoken_type[$j_next];
7847 # for tightness = 1, if there is just one token
7848 # within the matching pair, we will keep it tight
7850 $tok_next eq $matching_token{$last_token}
7852 # but watch out for this: [ [ ] (misc.t)
7853 && $last_token ne $token
7857 # remember where to put the space for the closing paren
7858 $j_tight_closing_paren = $j_next;
7866 } # done with opening braces and brackets
7868 if FORMATTER_DEBUG_FLAG_WHITE;
7870 #---------------------------------------------------------------
7872 # handle space on inside of closing brace pairs
7873 #---------------------------------------------------------------
7876 if ( $is_closing_type{$type} ) {
7878 if ( $j == $j_tight_closing_paren ) {
7880 $j_tight_closing_paren = -1;
7885 if ( !defined($ws) ) {
7888 if ( $type eq '}' && $token eq '}' && $block_type ) {
7889 $tightness = $rOpts_block_brace_tightness;
7891 else { $tightness = $tightness{$token} }
7893 $ws = ( $tightness > 1 ) ? WS_NO : WS_YES;
7899 if FORMATTER_DEBUG_FLAG_WHITE;
7901 #---------------------------------------------------------------
7903 # use the binary table
7904 #---------------------------------------------------------------
7905 if ( !defined($ws) ) {
7906 $ws = $binary_ws_rules{$last_type}{$type};
7909 if FORMATTER_DEBUG_FLAG_WHITE;
7911 #---------------------------------------------------------------
7913 # some special cases
7914 #---------------------------------------------------------------
7915 if ( $token eq '(' ) {
7917 # This will have to be tweaked as tokenization changes.
7918 # We usually want a space at '} (', for example:
7919 # map { 1 * $_; } ( $y, $M, $w, $d, $h, $m, $s );
7922 # &{ $_->[1] }( delete $_[$#_]{ $_->[0] } );
7923 # At present, the above & block is marked as type L/R so this case
7924 # won't go through here.
7925 if ( $last_type eq '}' ) { $ws = WS_YES }
7927 # NOTE: some older versions of Perl had occasional problems if
7928 # spaces are introduced between keywords or functions and opening
7929 # parens. So the default is not to do this except is certain
7930 # cases. The current Perl seems to tolerate spaces.
7932 # Space between keyword and '('
7933 elsif ( $last_type eq 'k' ) {
7935 unless ( $rOpts_space_keyword_paren
7936 || $space_after_keyword{$last_token} );
7939 # Space between function and '('
7940 # -----------------------------------------------------
7941 # 'w' and 'i' checks for something like:
7942 # myfun( &myfun( ->myfun(
7943 # -----------------------------------------------------
7944 elsif (( $last_type =~ /^[wU]$/ )
7945 || ( $last_type =~ /^[wi]$/ && $last_token =~ /^(\&|->)/ ) )
7947 $ws = WS_NO unless ($rOpts_space_function_paren);
7950 # space between something like $i and ( in
7951 # for $i ( 0 .. 20 ) {
7952 # FIXME: eventually, type 'i' needs to be split into multiple
7953 # token types so this can be a hardwired rule.
7954 elsif ( $last_type eq 'i' && $last_token =~ /^[\$\%\@]/ ) {
7958 # allow constant function followed by '()' to retain no space
7959 elsif ( $last_type eq 'C' && $$rtokens[ $j + 1 ] eq ')' ) {
7964 # patch for SWITCH/CASE: make space at ']{' optional
7965 # since the '{' might begin a case or when block
7966 elsif ( ( $token eq '{' && $type ne 'L' ) && $last_token eq ']' ) {
7970 # keep space between 'sub' and '{' for anonymous sub definition
7971 if ( $type eq '{' ) {
7972 if ( $last_token eq 'sub' ) {
7976 # this is needed to avoid no space in '){'
7977 if ( $last_token eq ')' && $token eq '{' ) { $ws = WS_YES }
7979 # avoid any space before the brace or bracket in something like
7980 # @opts{'a','b',...}
7981 if ( $last_type eq 'i' && $last_token =~ /^\@/ ) {
7986 elsif ( $type eq 'i' ) {
7988 # never a space before ->
7989 if ( $token =~ /^\-\>/ ) {
7994 # retain any space between '-' and bare word
7995 elsif ( $type eq 'w' || $type eq 'C' ) {
7996 $ws = WS_OPTIONAL if $last_type eq '-';
7998 # never a space before ->
7999 if ( $token =~ /^\-\>/ ) {
8004 # retain any space between '-' and bare word
8005 # example: avoid space between 'USER' and '-' here:
8006 # $myhash{USER-NAME}='steve';
8007 elsif ( $type eq 'm' || $type eq '-' ) {
8008 $ws = WS_OPTIONAL if ( $last_type eq 'w' );
8011 # always space before side comment
8012 elsif ( $type eq '#' ) { $ws = WS_YES if $j > 0 }
8014 # always preserver whatever space was used after a possible
8015 # filehandle (except _) or here doc operator
8018 && ( ( $last_type eq 'Z' && $last_token ne '_' )
8019 || $last_type eq 'h' )
8026 if FORMATTER_DEBUG_FLAG_WHITE;
8028 #---------------------------------------------------------------
8030 # default rules not covered above
8031 #---------------------------------------------------------------
8032 # if we fall through to here,
8033 # look at the pre-defined hash tables for the two tokens, and
8034 # if (they are equal) use the common value
8035 # if (either is zero or undef) use the other
8036 # if (either is -1) use it
8050 if ( !defined($ws) ) {
8051 my $wl = $want_left_space{$type};
8052 my $wr = $want_right_space{$last_type};
8053 if ( !defined($wl) ) { $wl = 0 }
8054 if ( !defined($wr) ) { $wr = 0 }
8055 $ws = ( ( $wl == $wr ) || ( $wl == -1 ) || !$wr ) ? $wl : $wr;
8058 if ( !defined($ws) ) {
8061 "WS flag is undefined for tokens $last_token $token\n");
8064 # Treat newline as a whitespace. Otherwise, we might combine
8065 # 'Send' and '-recipients' here according to the above rules:
8066 # my $msg = new Fax::Send
8067 # -recipients => $to,
8069 if ( $ws == 0 && $j == 0 ) { $ws = 1 }
8074 && ( $last_type !~ /^[Zh]$/ ) )
8077 # If this happens, we have a non-fatal but undesirable
8078 # hole in the above rules which should be patched.
8080 "WS flag is zero for tokens $last_token $token\n");
8082 $white_space_flag[$j] = $ws;
8084 FORMATTER_DEBUG_FLAG_WHITE && do {
8085 my $str = substr( $last_token, 0, 15 );
8086 $str .= ' ' x ( 16 - length($str) );
8087 if ( !defined($ws_1) ) { $ws_1 = "*" }
8088 if ( !defined($ws_2) ) { $ws_2 = "*" }
8089 if ( !defined($ws_3) ) { $ws_3 = "*" }
8090 if ( !defined($ws_4) ) { $ws_4 = "*" }
8092 "WHITE: i=$j $str $last_type $type $ws_1 : $ws_2 : $ws_3 : $ws_4 : $ws \n";
8095 return \@white_space_flag;
8098 { # begin print_line_of_tokens
8105 my $rcontainer_type;
8106 my $rcontainer_environment;
8109 my $rnesting_tokens;
8111 my $rnesting_blocks;
8114 my $python_indentation_level;
8116 # These local token variables are stored by store_token_to_go:
8119 my $container_environment;
8121 my $in_continued_quote;
8124 my $no_internal_newlines;
8130 # routine to pull the jth token from the line of tokens
8133 $token = $$rtokens[$j];
8134 $type = $$rtoken_type[$j];
8135 $block_type = $$rblock_type[$j];
8136 $container_type = $$rcontainer_type[$j];
8137 $container_environment = $$rcontainer_environment[$j];
8138 $type_sequence = $$rtype_sequence[$j];
8139 $level = $$rlevels[$j];
8140 $slevel = $$rslevels[$j];
8141 $nesting_blocks = $$rnesting_blocks[$j];
8142 $ci_level = $$rci_levels[$j];
8148 sub save_current_token {
8151 $block_type, $ci_level,
8152 $container_environment, $container_type,
8153 $in_continued_quote, $level,
8154 $nesting_blocks, $no_internal_newlines,
8156 $type, $type_sequence,
8160 sub restore_current_token {
8162 $block_type, $ci_level,
8163 $container_environment, $container_type,
8164 $in_continued_quote, $level,
8165 $nesting_blocks, $no_internal_newlines,
8167 $type, $type_sequence,
8172 # Routine to place the current token into the output stream.
8173 # Called once per output token.
8174 sub store_token_to_go {
8176 my $flag = $no_internal_newlines;
8177 if ( $_[0] ) { $flag = 1 }
8179 $tokens_to_go[ ++$max_index_to_go ] = $token;
8180 $types_to_go[$max_index_to_go] = $type;
8181 $nobreak_to_go[$max_index_to_go] = $flag;
8182 $old_breakpoint_to_go[$max_index_to_go] = 0;
8183 $forced_breakpoint_to_go[$max_index_to_go] = 0;
8184 $block_type_to_go[$max_index_to_go] = $block_type;
8185 $type_sequence_to_go[$max_index_to_go] = $type_sequence;
8186 $container_environment_to_go[$max_index_to_go] = $container_environment;
8187 $nesting_blocks_to_go[$max_index_to_go] = $nesting_blocks;
8188 $ci_levels_to_go[$max_index_to_go] = $ci_level;
8189 $mate_index_to_go[$max_index_to_go] = -1;
8190 $matching_token_to_go[$max_index_to_go] = '';
8191 $bond_strength_to_go[$max_index_to_go] = 0;
8193 # Note: negative levels are currently retained as a diagnostic so that
8194 # the 'final indentation level' is correctly reported for bad scripts.
8195 # But this means that every use of $level as an index must be checked.
8196 # If this becomes too much of a problem, we might give up and just clip
8198 ## $levels_to_go[$max_index_to_go] = ( $level > 0 ) ? $level : 0;
8199 $levels_to_go[$max_index_to_go] = $level;
8200 $nesting_depth_to_go[$max_index_to_go] = ( $slevel >= 0 ) ? $slevel : 0;
8201 $lengths_to_go[ $max_index_to_go + 1 ] =
8202 $lengths_to_go[$max_index_to_go] + length($token);
8204 # Define the indentation that this token would have if it started
8205 # a new line. We have to do this now because we need to know this
8206 # when considering one-line blocks.
8207 set_leading_whitespace( $level, $ci_level, $in_continued_quote );
8209 if ( $type ne 'b' ) {
8210 $last_last_nonblank_index_to_go = $last_nonblank_index_to_go;
8211 $last_last_nonblank_type_to_go = $last_nonblank_type_to_go;
8212 $last_last_nonblank_token_to_go = $last_nonblank_token_to_go;
8213 $last_nonblank_index_to_go = $max_index_to_go;
8214 $last_nonblank_type_to_go = $type;
8215 $last_nonblank_token_to_go = $token;
8216 if ( $type eq ',' ) {
8217 $comma_count_in_batch++;
8221 FORMATTER_DEBUG_FLAG_STORE && do {
8222 my ( $a, $b, $c ) = caller();
8224 "STORE: from $a $c: storing token $token type $type lev=$level slev=$slevel at $max_index_to_go\n";
8228 sub insert_new_token_to_go {
8230 # insert a new token into the output stream. use same level as
8231 # previous token; assumes a character at max_index_to_go.
8232 save_current_token();
8233 ( $token, $type, $slevel, $no_internal_newlines ) = @_;
8235 if ( $max_index_to_go == UNDEFINED_INDEX ) {
8236 warning("code bug: bad call to insert_new_token_to_go\n");
8238 $level = $levels_to_go[$max_index_to_go];
8240 # FIXME: it seems to be necessary to use the next, rather than
8241 # previous, value of this variable when creating a new blank (align.t)
8242 #my $slevel = $nesting_depth_to_go[$max_index_to_go];
8243 $nesting_blocks = $nesting_blocks_to_go[$max_index_to_go];
8244 $ci_level = $ci_levels_to_go[$max_index_to_go];
8245 $container_environment = $container_environment_to_go[$max_index_to_go];
8246 $in_continued_quote = 0;
8248 $type_sequence = "";
8249 store_token_to_go();
8250 restore_current_token();
8254 sub print_line_of_tokens {
8256 my $line_of_tokens = shift;
8258 # This routine is called once per input line to process all of
8259 # the tokens on that line. This is the first stage of
8262 # Full-line comments and blank lines may be processed immediately.
8264 # For normal lines of code, the tokens are stored one-by-one,
8265 # via calls to 'sub store_token_to_go', until a known line break
8266 # point is reached. Then, the batch of collected tokens is
8267 # passed along to 'sub output_line_to_go' for further
8268 # processing. This routine decides if there should be
8269 # whitespace between each pair of non-white tokens, so later
8270 # routines only need to decide on any additional line breaks.
8271 # Any whitespace is initally a single space character. Later,
8272 # the vertical aligner may expand that to be multiple space
8273 # characters if necessary for alignment.
8275 # extract input line number for error messages
8276 $input_line_number = $line_of_tokens->{_line_number};
8278 $rtoken_type = $line_of_tokens->{_rtoken_type};
8279 $rtokens = $line_of_tokens->{_rtokens};
8280 $rlevels = $line_of_tokens->{_rlevels};
8281 $rslevels = $line_of_tokens->{_rslevels};
8282 $rblock_type = $line_of_tokens->{_rblock_type};
8283 $rcontainer_type = $line_of_tokens->{_rcontainer_type};
8284 $rcontainer_environment = $line_of_tokens->{_rcontainer_environment};
8285 $rtype_sequence = $line_of_tokens->{_rtype_sequence};
8286 $input_line = $line_of_tokens->{_line_text};
8287 $rnesting_tokens = $line_of_tokens->{_rnesting_tokens};
8288 $rci_levels = $line_of_tokens->{_rci_levels};
8289 $rnesting_blocks = $line_of_tokens->{_rnesting_blocks};
8291 $in_continued_quote = $starting_in_quote =
8292 $line_of_tokens->{_starting_in_quote};
8293 $in_quote = $line_of_tokens->{_ending_in_quote};
8294 $ending_in_quote = $in_quote;
8295 $python_indentation_level =
8296 $line_of_tokens->{_python_indentation_level};
8301 my $next_nonblank_token;
8302 my $next_nonblank_token_type;
8303 my $rwhite_space_flag;
8305 $jmax = @$rtokens - 1;
8307 $container_type = "";
8308 $container_environment = "";
8309 $type_sequence = "";
8310 $no_internal_newlines = 1 - $rOpts_add_newlines;
8311 $is_static_block_comment = 0;
8313 # Handle a continued quote..
8314 if ($in_continued_quote) {
8316 # A line which is entirely a quote or pattern must go out
8317 # verbatim. Note: the \n is contained in $input_line.
8319 if ( ( $input_line =~ "\t" ) ) {
8320 note_embedded_tab();
8322 write_unindented_line("$input_line");
8323 $last_line_had_side_comment = 0;
8327 # prior to version 20010406, perltidy had a bug which placed
8328 # continuation indentation before the last line of some multiline
8329 # quotes and patterns -- exactly the lines passing this way.
8330 # To help find affected lines in scripts run with these
8331 # versions, run with '-chk', and it will warn of any quotes or
8332 # patterns which might have been modified by these early
8334 if ( $rOpts->{'check-multiline-quotes'} && $input_line =~ /^ / ) {
8336 "-chk: please check this line for extra leading whitespace\n"
8341 # Write line verbatim if we are in a formatting skip section
8342 if ($in_format_skipping_section) {
8343 write_unindented_line("$input_line");
8344 $last_line_had_side_comment = 0;
8346 # Note: extra space appended to comment simplifies pattern matching
8348 && $$rtoken_type[0] eq '#'
8349 && ( $$rtokens[0] . " " ) =~ /$format_skipping_pattern_end/o )
8351 $in_format_skipping_section = 0;
8352 write_logfile_entry("Exiting formatting skip section\n");
8357 # See if we are entering a formatting skip section
8358 if ( $rOpts_format_skipping
8360 && $$rtoken_type[0] eq '#'
8361 && ( $$rtokens[0] . " " ) =~ /$format_skipping_pattern_begin/o )
8364 $in_format_skipping_section = 1;
8365 write_logfile_entry("Entering formatting skip section\n");
8366 write_unindented_line("$input_line");
8367 $last_line_had_side_comment = 0;
8371 # delete trailing blank tokens
8372 if ( $jmax > 0 && $$rtoken_type[$jmax] eq 'b' ) { $jmax-- }
8374 # Handle a blank line..
8377 # For the 'swallow-optional-blank-lines' option, we delete all
8378 # old blank lines and let the blank line rules generate any
8380 if ( !$rOpts_swallow_optional_blank_lines ) {
8382 $file_writer_object->write_blank_code_line();
8383 $last_line_leading_type = 'b';
8385 $last_line_had_side_comment = 0;
8389 # see if this is a static block comment (starts with ## by default)
8390 my $is_static_block_comment_without_leading_space = 0;
8392 && $$rtoken_type[0] eq '#'
8393 && $rOpts->{'static-block-comments'}
8394 && $input_line =~ /$static_block_comment_pattern/o )
8396 $is_static_block_comment = 1;
8397 $is_static_block_comment_without_leading_space =
8398 substr( $input_line, 0, 1 ) eq '#';
8401 # Check for comments which are line directives
8402 # Treat exactly as static block comments without leading space
8403 # reference: perlsyn, near end, section Plain Old Comments (Not!)
8404 # example: '# line 42 "new_filename.plx"'
8407 && $$rtoken_type[0] eq '#'
8408 && $input_line =~ /^\# \s*
8410 (?:\s("?)([^"]+)\2)? \s*
8414 $is_static_block_comment = 1;
8415 $is_static_block_comment_without_leading_space = 1;
8418 # create a hanging side comment if appropriate
8421 && $$rtoken_type[0] eq '#' # only token is a comment
8422 && $last_line_had_side_comment # last line had side comment
8423 && $input_line =~ /^\s/ # there is some leading space
8424 && !$is_static_block_comment # do not make static comment hanging
8425 && $rOpts->{'hanging-side-comments'} # user is allowing this
8429 # We will insert an empty qw string at the start of the token list
8430 # to force this comment to be a side comment. The vertical aligner
8431 # should then line it up with the previous side comment.
8432 unshift @$rtoken_type, 'q';
8433 unshift @$rtokens, '';
8434 unshift @$rlevels, $$rlevels[0];
8435 unshift @$rslevels, $$rslevels[0];
8436 unshift @$rblock_type, '';
8437 unshift @$rcontainer_type, '';
8438 unshift @$rcontainer_environment, '';
8439 unshift @$rtype_sequence, '';
8440 unshift @$rnesting_tokens, $$rnesting_tokens[0];
8441 unshift @$rci_levels, $$rci_levels[0];
8442 unshift @$rnesting_blocks, $$rnesting_blocks[0];
8446 # remember if this line has a side comment
8447 $last_line_had_side_comment =
8448 ( $jmax > 0 && $$rtoken_type[$jmax] eq '#' );
8450 # Handle a block (full-line) comment..
8451 if ( ( $jmax == 0 ) && ( $$rtoken_type[0] eq '#' ) ) {
8453 if ( $rOpts->{'delete-block-comments'} ) { return }
8455 if ( $rOpts->{'tee-block-comments'} ) {
8456 $file_writer_object->tee_on();
8459 destroy_one_line_block();
8460 output_line_to_go();
8462 # output a blank line before block comments
8464 $last_line_leading_type !~ /^[#b]$/
8465 && $rOpts->{'blanks-before-comments'} # only if allowed
8467 $is_static_block_comment # never before static block comments
8470 flush(); # switching to new output stream
8471 $file_writer_object->write_blank_code_line();
8472 $last_line_leading_type = 'b';
8475 # TRIM COMMENTS -- This could be turned off as a option
8476 $$rtokens[0] =~ s/\s*$//; # trim right end
8479 $rOpts->{'indent-block-comments'}
8480 && ( !$rOpts->{'indent-spaced-block-comments'}
8481 || $input_line =~ /^\s+/ )
8482 && !$is_static_block_comment_without_leading_space
8486 store_token_to_go();
8487 output_line_to_go();
8490 flush(); # switching to new output stream
8491 $file_writer_object->write_code_line( $$rtokens[0] . "\n" );
8492 $last_line_leading_type = '#';
8494 if ( $rOpts->{'tee-block-comments'} ) {
8495 $file_writer_object->tee_off();
8500 # compare input/output indentation except for continuation lines
8501 # (because they have an unknown amount of initial blank space)
8502 # and lines which are quotes (because they may have been outdented)
8503 # Note: this test is placed here because we know the continuation flag
8504 # at this point, which allows us to avoid non-meaningful checks.
8505 my $structural_indentation_level = $$rlevels[0];
8506 compare_indentation_levels( $python_indentation_level,
8507 $structural_indentation_level )
8508 unless ( $python_indentation_level < 0
8509 || ( $$rci_levels[0] > 0 )
8510 || ( ( $python_indentation_level == 0 ) && $$rtoken_type[0] eq 'Q' )
8513 # Patch needed for MakeMaker. Do not break a statement
8514 # in which $VERSION may be calculated. See MakeMaker.pm;
8515 # this is based on the coding in it.
8516 # The first line of a file that matches this will be eval'd:
8517 # /([\$*])(([\w\:\']*)\bVERSION)\b.*\=/
8519 # *VERSION = \'1.01';
8520 # ( $VERSION ) = '$Revision: 1.73 $ ' =~ /\$Revision:\s+([^\s]+)/;
8521 # We will pass such a line straight through without breaking
8522 # it unless -npvl is used
8524 my $is_VERSION_statement = 0;
8527 !$saw_VERSION_in_this_file
8528 && $input_line =~ /VERSION/ # quick check to reject most lines
8529 && $input_line =~ /([\$*])(([\w\:\']*)\bVERSION)\b.*\=/
8532 $saw_VERSION_in_this_file = 1;
8533 $is_VERSION_statement = 1;
8534 write_logfile_entry("passing VERSION line; -npvl deactivates\n");
8535 $no_internal_newlines = 1;
8538 # take care of indentation-only
8539 # NOTE: In previous versions we sent all qw lines out immediately here.
8540 # No longer doing this: also write a line which is entirely a 'qw' list
8541 # to allow stacking of opening and closing tokens. Note that interior
8542 # qw lines will still go out at the end of this routine.
8543 if ( $rOpts->{'indent-only'} ) {
8548 $token = $input_line;
8551 $container_type = "";
8552 $container_environment = "";
8553 $type_sequence = "";
8554 store_token_to_go();
8555 output_line_to_go();
8559 push( @$rtokens, ' ', ' ' ); # making $j+2 valid simplifies coding
8560 push( @$rtoken_type, 'b', 'b' );
8561 ($rwhite_space_flag) =
8562 set_white_space_flag( $jmax, $rtokens, $rtoken_type, $rblock_type );
8564 # find input tabbing to allow checks for tabbing disagreement
8566 ##$input_line_tabbing = "";
8567 ##if ( $input_line =~ /^(\s*)/ ) { $input_line_tabbing = $1; }
8569 # if the buffer hasn't been flushed, add a leading space if
8570 # necessary to keep essential whitespace. This is really only
8571 # necessary if we are squeezing out all ws.
8572 if ( $max_index_to_go >= 0 ) {
8574 $old_line_count_in_batch++;
8577 is_essential_whitespace(
8578 $last_last_nonblank_token,
8579 $last_last_nonblank_type,
8580 $tokens_to_go[$max_index_to_go],
8581 $types_to_go[$max_index_to_go],
8587 my $slevel = $$rslevels[0];
8588 insert_new_token_to_go( ' ', 'b', $slevel,
8589 $no_internal_newlines );
8593 # If we just saw the end of an elsif block, write nag message
8594 # if we do not see another elseif or an else.
8595 if ($looking_for_else) {
8597 unless ( $$rtokens[0] =~ /^(elsif|else)$/ ) {
8598 write_logfile_entry("(No else block)\n");
8600 $looking_for_else = 0;
8603 # This is a good place to kill incomplete one-line blocks
8604 if ( ( $semicolons_before_block_self_destruct == 0 )
8605 && ( $max_index_to_go >= 0 )
8606 && ( $types_to_go[$max_index_to_go] eq ';' )
8607 && ( $$rtokens[0] ne '}' ) )
8609 destroy_one_line_block();
8610 output_line_to_go();
8613 # loop to process the tokens one-by-one
8617 foreach $j ( 0 .. $jmax ) {
8619 # pull out the local values for this token
8622 if ( $type eq '#' ) {
8624 # trim trailing whitespace
8625 # (there is no option at present to prevent this)
8629 $rOpts->{'delete-side-comments'}
8631 # delete closing side comments if necessary
8632 || ( $rOpts->{'delete-closing-side-comments'}
8633 && $token =~ /$closing_side_comment_prefix_pattern/o
8634 && $last_nonblank_block_type =~
8635 /$closing_side_comment_list_pattern/o )
8638 if ( $types_to_go[$max_index_to_go] eq 'b' ) {
8639 unstore_token_to_go();
8645 # If we are continuing after seeing a right curly brace, flush
8646 # buffer unless we see what we are looking for, as in
8648 if ( $rbrace_follower && $type ne 'b' ) {
8650 unless ( $rbrace_follower->{$token} ) {
8651 output_line_to_go();
8653 $rbrace_follower = undef;
8656 $j_next = ( $$rtoken_type[ $j + 1 ] eq 'b' ) ? $j + 2 : $j + 1;
8657 $next_nonblank_token = $$rtokens[$j_next];
8658 $next_nonblank_token_type = $$rtoken_type[$j_next];
8660 #--------------------------------------------------------
8661 # Start of section to patch token text
8662 #--------------------------------------------------------
8664 # Modify certain tokens here for whitespace
8665 # The following is not yet done, but could be:
8667 if ( $type =~ /^[wit]$/ ) {
8670 # change '$ var' to '$var' etc
8671 # '-> new' to '->new'
8672 if ( $token =~ /^([\$\&\%\*\@]|\-\>)\s/ ) {
8676 if ( $token =~ /^sub/ ) { $token =~ s/\s+/ /g }
8679 # change 'LABEL :' to 'LABEL:'
8680 elsif ( $type eq 'J' ) { $token =~ s/\s+//g }
8682 # patch to add space to something like "x10"
8683 # This avoids having to split this token in the pre-tokenizer
8684 elsif ( $type eq 'n' ) {
8685 if ( $token =~ /^x\d+/ ) { $token =~ s/x/x / }
8688 elsif ( $type eq 'Q' ) {
8689 note_embedded_tab() if ( $token =~ "\t" );
8691 # make note of something like '$var = s/xxx/yyy/;'
8692 # in case it should have been '$var =~ s/xxx/yyy/;'
8694 $token =~ /^(s|tr|y|m|\/)/
8695 && $last_nonblank_token =~ /^(=|==|!=)$/
8697 # precededed by simple scalar
8698 && $last_last_nonblank_type eq 'i'
8699 && $last_last_nonblank_token =~ /^\$/
8701 # followed by some kind of termination
8702 # (but give complaint if we can's see far enough ahead)
8703 && $next_nonblank_token =~ /^[; \)\}]$/
8705 # scalar is not decleared
8707 $types_to_go[0] eq 'k'
8708 && $tokens_to_go[0] =~ /^(my|our|local)$/
8712 my $guess = substr( $last_nonblank_token, 0, 1 ) . '~';
8714 "Note: be sure you want '$last_nonblank_token' instead of '$guess' here\n"
8719 # trim blanks from right of qw quotes
8720 # (To avoid trimming qw quotes use -ntqw; the tokenizer handles this)
8721 elsif ( $type eq 'q' ) {
8723 note_embedded_tab() if ( $token =~ "\t" );
8726 #--------------------------------------------------------
8727 # End of section to patch token text
8728 #--------------------------------------------------------
8730 # insert any needed whitespace
8731 if ( ( $type ne 'b' )
8732 && ( $max_index_to_go >= 0 )
8733 && ( $types_to_go[$max_index_to_go] ne 'b' )
8734 && $rOpts_add_whitespace )
8736 my $ws = $$rwhite_space_flag[$j];
8739 insert_new_token_to_go( ' ', 'b', $slevel,
8740 $no_internal_newlines );
8744 # Do not allow breaks which would promote a side comment to a
8745 # block comment. In order to allow a break before an opening
8746 # or closing BLOCK, followed by a side comment, those sections
8747 # of code will handle this flag separately.
8748 my $side_comment_follows = ( $next_nonblank_token_type eq '#' );
8749 my $is_opening_BLOCK =
8753 && $block_type ne 't' );
8754 my $is_closing_BLOCK =
8758 && $block_type ne 't' );
8760 if ( $side_comment_follows
8761 && !$is_opening_BLOCK
8762 && !$is_closing_BLOCK )
8764 $no_internal_newlines = 1;
8767 # We're only going to handle breaking for code BLOCKS at this
8768 # (top) level. Other indentation breaks will be handled by
8769 # sub scan_list, which is better suited to dealing with them.
8770 if ($is_opening_BLOCK) {
8772 # Tentatively output this token. This is required before
8773 # calling starting_one_line_block. We may have to unstore
8774 # it, though, if we have to break before it.
8775 store_token_to_go($side_comment_follows);
8777 # Look ahead to see if we might form a one-line block
8779 starting_one_line_block( $j, $jmax, $level, $slevel,
8780 $ci_level, $rtokens, $rtoken_type, $rblock_type );
8781 clear_breakpoint_undo_stack();
8783 # to simplify the logic below, set a flag to indicate if
8784 # this opening brace is far from the keyword which introduces it
8785 my $keyword_on_same_line = 1;
8786 if ( ( $max_index_to_go >= 0 )
8787 && ( $last_nonblank_type eq ')' ) )
8789 if ( $block_type =~ /^(if|else|elsif)$/
8790 && ( $tokens_to_go[0] eq '}' )
8791 && $rOpts_cuddled_else )
8793 $keyword_on_same_line = 1;
8795 elsif ( ( $slevel < $nesting_depth_to_go[0] ) || $too_long )
8797 $keyword_on_same_line = 0;
8801 # decide if user requested break before '{'
8804 # use -bl flag if not a sub block of any type
8805 $block_type !~ /^sub/
8806 ? $rOpts->{'opening-brace-on-new-line'}
8808 # use -sbl flag unless this is an anonymous sub block
8809 : $block_type !~ /^sub\W*$/
8810 ? $rOpts->{'opening-sub-brace-on-new-line'}
8812 # do not break for anonymous subs
8815 # Break before an opening '{' ...
8821 # and we were unable to start looking for a block,
8822 && $index_start_one_line_block == UNDEFINED_INDEX
8824 # or if it will not be on same line as its keyword, so that
8825 # it will be outdented (eval.t, overload.t), and the user
8826 # has not insisted on keeping it on the right
8827 || ( !$keyword_on_same_line
8828 && !$rOpts->{'opening-brace-always-on-right'} )
8833 # but only if allowed
8834 unless ($no_internal_newlines) {
8836 # since we already stored this token, we must unstore it
8837 unstore_token_to_go();
8839 # then output the line
8840 output_line_to_go();
8842 # and now store this token at the start of a new line
8843 store_token_to_go($side_comment_follows);
8847 # Now update for side comment
8848 if ($side_comment_follows) { $no_internal_newlines = 1 }
8850 # now output this line
8851 unless ($no_internal_newlines) {
8852 output_line_to_go();
8856 elsif ($is_closing_BLOCK) {
8858 # If there is a pending one-line block ..
8859 if ( $index_start_one_line_block != UNDEFINED_INDEX ) {
8861 # we have to terminate it if..
8864 # it is too long (final length may be different from
8865 # initial estimate). note: must allow 1 space for this token
8866 excess_line_length( $index_start_one_line_block,
8867 $max_index_to_go ) >= 0
8869 # or if it has too many semicolons
8870 || ( $semicolons_before_block_self_destruct == 0
8871 && $last_nonblank_type ne ';' )
8874 destroy_one_line_block();
8878 # put a break before this closing curly brace if appropriate
8879 unless ( $no_internal_newlines
8880 || $index_start_one_line_block != UNDEFINED_INDEX )
8883 # add missing semicolon if ...
8884 # there are some tokens
8886 ( $max_index_to_go > 0 )
8888 # and we don't have one
8889 && ( $last_nonblank_type ne ';' )
8891 # patch until some block type issues are fixed:
8892 # Do not add semi-colon for block types '{',
8893 # '}', and ';' because we cannot be sure yet
8894 # that this is a block and not an anonomyous
8895 # hash (blktype.t, blktype1.t)
8896 && ( $block_type !~ /^[\{\};]$/ )
8898 # it seems best not to add semicolons in these
8899 # special block types: sort|map|grep
8900 && ( !$is_sort_map_grep{$block_type} )
8902 # and we are allowed to do so.
8903 && $rOpts->{'add-semicolons'}
8907 save_current_token();
8910 $level = $levels_to_go[$max_index_to_go];
8911 $slevel = $nesting_depth_to_go[$max_index_to_go];
8913 $nesting_blocks_to_go[$max_index_to_go];
8914 $ci_level = $ci_levels_to_go[$max_index_to_go];
8916 $container_type = "";
8917 $container_environment = "";
8918 $type_sequence = "";
8920 # Note - we remove any blank AFTER extracting its
8921 # parameters such as level, etc, above
8922 if ( $types_to_go[$max_index_to_go] eq 'b' ) {
8923 unstore_token_to_go();
8925 store_token_to_go();
8927 note_added_semicolon();
8928 restore_current_token();
8931 # then write out everything before this closing curly brace
8932 output_line_to_go();
8936 # Now update for side comment
8937 if ($side_comment_follows) { $no_internal_newlines = 1 }
8939 # store the closing curly brace
8940 store_token_to_go();
8942 # ok, we just stored a closing curly brace. Often, but
8943 # not always, we want to end the line immediately.
8944 # So now we have to check for special cases.
8946 # if this '}' successfully ends a one-line block..
8947 my $is_one_line_block = 0;
8949 if ( $index_start_one_line_block != UNDEFINED_INDEX ) {
8951 # Remember the type of token just before the
8952 # opening brace. It would be more general to use
8953 # a stack, but this will work for one-line blocks.
8954 $is_one_line_block =
8955 $types_to_go[$index_start_one_line_block];
8957 # we have to actually make it by removing tentative
8958 # breaks that were set within it
8959 undo_forced_breakpoint_stack(0);
8960 set_nobreaks( $index_start_one_line_block,
8961 $max_index_to_go - 1 );
8963 # then re-initialize for the next one-line block
8964 destroy_one_line_block();
8966 # then decide if we want to break after the '}' ..
8967 # We will keep going to allow certain brace followers as in:
8968 # do { $ifclosed = 1; last } unless $losing;
8970 # But make a line break if the curly ends a
8971 # significant block:
8973 $is_block_without_semicolon{$block_type}
8975 # if needless semicolon follows we handle it later
8976 && $next_nonblank_token ne ';'
8979 output_line_to_go() unless ($no_internal_newlines);
8983 # set string indicating what we need to look for brace follower
8985 if ( $block_type eq 'do' ) {
8986 $rbrace_follower = \%is_do_follower;
8988 elsif ( $block_type =~ /^(if|elsif|unless)$/ ) {
8989 $rbrace_follower = \%is_if_brace_follower;
8991 elsif ( $block_type eq 'else' ) {
8992 $rbrace_follower = \%is_else_brace_follower;
8995 # added eval for borris.t
8996 elsif ($is_sort_map_grep_eval{$block_type}
8997 || $is_one_line_block eq 'G' )
8999 $rbrace_follower = undef;
9004 elsif ( $block_type =~ /^sub\W*$/ ) {
9006 if ($is_one_line_block) {
9007 $rbrace_follower = \%is_anon_sub_1_brace_follower;
9010 $rbrace_follower = \%is_anon_sub_brace_follower;
9014 # None of the above: specify what can follow a closing
9015 # brace of a block which is not an
9016 # if/elsif/else/do/sort/map/grep/eval
9018 # 'Toolbar.pm', 'Menubar.pm', bless.t, '3rules.pl', 'break1.t
9020 $rbrace_follower = \%is_other_brace_follower;
9023 # See if an elsif block is followed by another elsif or else;
9025 if ( $block_type eq 'elsif' ) {
9027 if ( $next_nonblank_token_type eq 'b' ) { # end of line?
9028 $looking_for_else = 1; # ok, check on next line
9032 unless ( $next_nonblank_token =~ /^(elsif|else)$/ ) {
9033 write_logfile_entry("No else block :(\n");
9038 # keep going after certain block types (map,sort,grep,eval)
9039 # added eval for borris.t
9045 # if no more tokens, postpone decision until re-entring
9046 elsif ( ( $next_nonblank_token_type eq 'b' )
9047 && $rOpts_add_newlines )
9049 unless ($rbrace_follower) {
9050 output_line_to_go() unless ($no_internal_newlines);
9054 elsif ($rbrace_follower) {
9056 unless ( $rbrace_follower->{$next_nonblank_token} ) {
9057 output_line_to_go() unless ($no_internal_newlines);
9059 $rbrace_follower = undef;
9063 output_line_to_go() unless ($no_internal_newlines);
9066 } # end treatment of closing block token
9069 elsif ( $type eq ';' ) {
9071 # kill one-line blocks with too many semicolons
9072 $semicolons_before_block_self_destruct--;
9074 ( $semicolons_before_block_self_destruct < 0 )
9075 || ( $semicolons_before_block_self_destruct == 0
9076 && $next_nonblank_token_type !~ /^[b\}]$/ )
9079 destroy_one_line_block();
9082 # Remove unnecessary semicolons, but not after bare
9083 # blocks, where it could be unsafe if the brace is
9087 $last_nonblank_token eq '}'
9089 $is_block_without_semicolon{
9090 $last_nonblank_block_type}
9091 || $last_nonblank_block_type =~ /^sub\s+\w/
9092 || $last_nonblank_block_type =~ /^\w+:$/ )
9094 || $last_nonblank_type eq ';'
9099 $rOpts->{'delete-semicolons'}
9101 # don't delete ; before a # because it would promote it
9102 # to a block comment
9103 && ( $next_nonblank_token_type ne '#' )
9106 note_deleted_semicolon();
9108 unless ( $no_internal_newlines
9109 || $index_start_one_line_block != UNDEFINED_INDEX );
9113 write_logfile_entry("Extra ';'\n");
9116 store_token_to_go();
9119 unless ( $no_internal_newlines
9120 || ( $rOpts_keep_interior_semicolons && $j < $jmax )
9121 || ( $next_nonblank_token eq '}' ) );
9125 # handle here_doc target string
9126 elsif ( $type eq 'h' ) {
9127 $no_internal_newlines =
9128 1; # no newlines after seeing here-target
9129 destroy_one_line_block();
9130 store_token_to_go();
9133 # handle all other token types
9136 # if this is a blank...
9137 if ( $type eq 'b' ) {
9139 # make it just one character
9140 $token = ' ' if $rOpts_add_whitespace;
9142 # delete it if unwanted by whitespace rules
9143 # or we are deleting all whitespace
9144 my $ws = $$rwhite_space_flag[ $j + 1 ];
9145 if ( ( defined($ws) && $ws == -1 )
9146 || $rOpts_delete_old_whitespace )
9149 # unless it might make a syntax error
9151 unless is_essential_whitespace(
9152 $last_last_nonblank_token,
9153 $last_last_nonblank_type,
9154 $tokens_to_go[$max_index_to_go],
9155 $types_to_go[$max_index_to_go],
9156 $$rtokens[ $j + 1 ],
9157 $$rtoken_type[ $j + 1 ]
9161 store_token_to_go();
9164 # remember two previous nonblank OUTPUT tokens
9165 if ( $type ne '#' && $type ne 'b' ) {
9166 $last_last_nonblank_token = $last_nonblank_token;
9167 $last_last_nonblank_type = $last_nonblank_type;
9168 $last_nonblank_token = $token;
9169 $last_nonblank_type = $type;
9170 $last_nonblank_block_type = $block_type;
9173 # unset the continued-quote flag since it only applies to the
9174 # first token, and we want to resume normal formatting if
9175 # there are additional tokens on the line
9176 $in_continued_quote = 0;
9178 } # end of loop over all tokens in this 'line_of_tokens'
9180 # we have to flush ..
9183 # if there is a side comment
9184 ( ( $type eq '#' ) && !$rOpts->{'delete-side-comments'} )
9186 # if this line ends in a quote
9187 # NOTE: This is critically important for insuring that quoted lines
9188 # do not get processed by things like -sot and -sct
9191 # if this is a VERSION statement
9192 || $is_VERSION_statement
9194 # to keep a label on one line if that is how it is now
9195 || ( ( $type eq 'J' ) && ( $max_index_to_go == 0 ) )
9197 # if we are instructed to keep all old line breaks
9198 || !$rOpts->{'delete-old-newlines'}
9201 destroy_one_line_block();
9202 output_line_to_go();
9205 # mark old line breakpoints in current output stream
9206 if ( $max_index_to_go >= 0 && !$rOpts_ignore_old_breakpoints ) {
9207 $old_breakpoint_to_go[$max_index_to_go] = 1;
9209 } # end sub print_line_of_tokens
9210 } # end print_line_of_tokens
9212 # sub output_line_to_go sends one logical line of tokens on down the
9213 # pipeline to the VerticalAligner package, breaking the line into continuation
9214 # lines as necessary. The line of tokens is ready to go in the "to_go"
9216 sub output_line_to_go {
9218 # debug stuff; this routine can be called from many points
9219 FORMATTER_DEBUG_FLAG_OUTPUT && do {
9220 my ( $a, $b, $c ) = caller;
9222 "OUTPUT: output_line_to_go called: $a $c $last_nonblank_type $last_nonblank_token, one_line=$index_start_one_line_block, tokens to write=$max_index_to_go\n"
9224 my $output_str = join "", @tokens_to_go[ 0 .. $max_index_to_go ];
9225 write_diagnostics("$output_str\n");
9228 # just set a tentative breakpoint if we might be in a one-line block
9229 if ( $index_start_one_line_block != UNDEFINED_INDEX ) {
9230 set_forced_breakpoint($max_index_to_go);
9234 my $cscw_block_comment;
9235 $cscw_block_comment = add_closing_side_comment()
9236 if ( $rOpts->{'closing-side-comments'} && $max_index_to_go >= 0 );
9238 match_opening_and_closing_tokens();
9240 # tell the -lp option we are outputting a batch so it can close
9241 # any unfinished items in its stack
9244 # If this line ends in a code block brace, set breaks at any
9245 # previous closing code block braces to breakup a chain of code
9246 # blocks on one line. This is very rare but can happen for
9247 # user-defined subs. For example we might be looking at this:
9248 # BOOL { $server_data{uptime} > 0; } NUM { $server_data{load}; } STR {
9249 my $saw_good_break = 0; # flag to force breaks even if short line
9252 # looking for opening or closing block brace
9253 $block_type_to_go[$max_index_to_go]
9255 # but not one of these which are never duplicated on a line:
9256 # until|while|for|if|elsif|else
9257 && !$is_block_without_semicolon{ $block_type_to_go[$max_index_to_go] }
9260 my $lev = $nesting_depth_to_go[$max_index_to_go];
9262 # Walk backwards from the end and
9263 # set break at any closing block braces at the same level.
9264 # But quit if we are not in a chain of blocks.
9265 for ( my $i = $max_index_to_go - 1 ; $i >= 0 ; $i-- ) {
9266 last if ( $levels_to_go[$i] < $lev ); # stop at a lower level
9267 next if ( $levels_to_go[$i] > $lev ); # skip past higher level
9269 if ( $block_type_to_go[$i] ) {
9270 if ( $tokens_to_go[$i] eq '}' ) {
9271 set_forced_breakpoint($i);
9272 $saw_good_break = 1;
9276 # quit if we see anything besides words, function, blanks
9278 elsif ( $types_to_go[$i] !~ /^[\(\)Gwib]$/ ) { last }
9283 my $imax = $max_index_to_go;
9285 # trim any blank tokens
9286 if ( $max_index_to_go >= 0 ) {
9287 if ( $types_to_go[$imin] eq 'b' ) { $imin++ }
9288 if ( $types_to_go[$imax] eq 'b' ) { $imax-- }
9291 # anything left to write?
9292 if ( $imin <= $imax ) {
9294 # add a blank line before certain key types
9295 if ( $last_line_leading_type !~ /^[#b]/ ) {
9297 my $leading_token = $tokens_to_go[$imin];
9298 my $leading_type = $types_to_go[$imin];
9300 # blank lines before subs except declarations and one-liners
9301 # MCONVERSION LOCATION - for sub tokenization change
9302 if ( $leading_token =~ /^(sub\s)/ && $leading_type eq 'i' ) {
9303 $want_blank = ( $rOpts->{'blanks-before-subs'} )
9305 terminal_type( \@types_to_go, \@block_type_to_go, $imin,
9306 $imax ) !~ /^[\;\}]$/
9310 # break before all package declarations
9311 # MCONVERSION LOCATION - for tokenizaton change
9312 elsif ($leading_token =~ /^(package\s)/
9313 && $leading_type eq 'i' )
9315 $want_blank = ( $rOpts->{'blanks-before-subs'} );
9318 # break before certain key blocks except one-liners
9319 if ( $leading_token =~ /^(BEGIN|END)$/ && $leading_type eq 'k' ) {
9320 $want_blank = ( $rOpts->{'blanks-before-subs'} )
9322 terminal_type( \@types_to_go, \@block_type_to_go, $imin,
9327 # Break before certain block types if we haven't had a
9328 # break at this level for a while. This is the
9329 # difficult decision..
9330 elsif ($leading_token =~ /^(unless|if|while|until|for|foreach)$/
9331 && $leading_type eq 'k' )
9333 my $lc = $nonblank_lines_at_depth[$last_line_leading_level];
9334 if ( !defined($lc) ) { $lc = 0 }
9337 $rOpts->{'blanks-before-blocks'}
9338 && $lc >= $rOpts->{'long-block-line-count'}
9339 && $file_writer_object->get_consecutive_nonblank_lines() >=
9340 $rOpts->{'long-block-line-count'}
9342 terminal_type( \@types_to_go, \@block_type_to_go, $imin,
9349 # future: send blank line down normal path to VerticalAligner
9350 Perl::Tidy::VerticalAligner::flush();
9351 $file_writer_object->write_blank_code_line();
9355 # update blank line variables and count number of consecutive
9356 # non-blank, non-comment lines at this level
9357 $last_last_line_leading_level = $last_line_leading_level;
9358 $last_line_leading_level = $levels_to_go[$imin];
9359 if ( $last_line_leading_level < 0 ) { $last_line_leading_level = 0 }
9360 $last_line_leading_type = $types_to_go[$imin];
9361 if ( $last_line_leading_level == $last_last_line_leading_level
9362 && $last_line_leading_type ne 'b'
9363 && $last_line_leading_type ne '#'
9364 && defined( $nonblank_lines_at_depth[$last_line_leading_level] ) )
9366 $nonblank_lines_at_depth[$last_line_leading_level]++;
9369 $nonblank_lines_at_depth[$last_line_leading_level] = 1;
9372 FORMATTER_DEBUG_FLAG_FLUSH && do {
9373 my ( $package, $file, $line ) = caller;
9375 "FLUSH: flushing from $package $file $line, types= $types_to_go[$imin] to $types_to_go[$imax]\n";
9378 # add a couple of extra terminal blank tokens
9381 # set all forced breakpoints for good list formatting
9382 my $is_long_line = excess_line_length( $imin, $max_index_to_go ) > 0;
9385 $max_index_to_go > 0
9388 || $old_line_count_in_batch > 1
9389 || is_unbalanced_batch()
9391 $comma_count_in_batch
9392 && ( $rOpts_maximum_fields_per_table > 0
9393 || $rOpts_comma_arrow_breakpoints == 0 )
9398 $saw_good_break ||= scan_list();
9401 # let $ri_first and $ri_last be references to lists of
9402 # first and last tokens of line fragments to output..
9403 my ( $ri_first, $ri_last );
9405 # write a single line if..
9408 # we aren't allowed to add any newlines
9409 !$rOpts_add_newlines
9411 # or, we don't already have an interior breakpoint
9412 # and we didn't see a good breakpoint
9414 !$forced_breakpoint_count
9417 # and this line is 'short'
9422 @$ri_first = ($imin);
9423 @$ri_last = ($imax);
9426 # otherwise use multiple lines
9429 ( $ri_first, $ri_last, my $colon_count ) =
9430 set_continuation_breaks($saw_good_break);
9432 break_all_chain_tokens( $ri_first, $ri_last );
9434 break_equals( $ri_first, $ri_last );
9436 # now we do a correction step to clean this up a bit
9437 # (The only time we would not do this is for debugging)
9438 if ( $rOpts->{'recombine'} ) {
9439 ( $ri_first, $ri_last ) =
9440 recombine_breakpoints( $ri_first, $ri_last );
9443 insert_final_breaks( $ri_first, $ri_last ) if $colon_count;
9446 # do corrector step if -lp option is used
9448 if ($rOpts_line_up_parentheses) {
9449 $do_not_pad = correct_lp_indentation( $ri_first, $ri_last );
9451 send_lines_to_vertical_aligner( $ri_first, $ri_last, $do_not_pad );
9453 prepare_for_new_input_lines();
9455 # output any new -cscw block comment
9456 if ($cscw_block_comment) {
9458 $file_writer_object->write_code_line( $cscw_block_comment . "\n" );
9462 sub note_added_semicolon {
9463 $last_added_semicolon_at = $input_line_number;
9464 if ( $added_semicolon_count == 0 ) {
9465 $first_added_semicolon_at = $last_added_semicolon_at;
9467 $added_semicolon_count++;
9468 write_logfile_entry("Added ';' here\n");
9471 sub note_deleted_semicolon {
9472 $last_deleted_semicolon_at = $input_line_number;
9473 if ( $deleted_semicolon_count == 0 ) {
9474 $first_deleted_semicolon_at = $last_deleted_semicolon_at;
9476 $deleted_semicolon_count++;
9477 write_logfile_entry("Deleted unnecessary ';'\n"); # i hope ;)
9480 sub note_embedded_tab {
9481 $embedded_tab_count++;
9482 $last_embedded_tab_at = $input_line_number;
9483 if ( !$first_embedded_tab_at ) {
9484 $first_embedded_tab_at = $last_embedded_tab_at;
9487 if ( $embedded_tab_count <= MAX_NAG_MESSAGES ) {
9488 write_logfile_entry("Embedded tabs in quote or pattern\n");
9492 sub starting_one_line_block {
9494 # after seeing an opening curly brace, look for the closing brace
9495 # and see if the entire block will fit on a line. This routine is
9496 # not always right because it uses the old whitespace, so a check
9497 # is made later (at the closing brace) to make sure we really
9498 # have a one-line block. We have to do this preliminary check,
9499 # though, because otherwise we would always break at a semicolon
9500 # within a one-line block if the block contains multiple statements.
9502 my ( $j, $jmax, $level, $slevel, $ci_level, $rtokens, $rtoken_type,
9506 # kill any current block - we can only go 1 deep
9507 destroy_one_line_block();
9510 # 1=distance from start of block to opening brace exceeds line length
9515 # shouldn't happen: there must have been a prior call to
9516 # store_token_to_go to put the opening brace in the output stream
9517 if ( $max_index_to_go < 0 ) {
9518 warning("program bug: store_token_to_go called incorrectly\n");
9519 report_definite_bug();
9523 # cannot use one-line blocks with cuddled else else/elsif lines
9524 if ( ( $tokens_to_go[0] eq '}' ) && $rOpts_cuddled_else ) {
9529 my $block_type = $$rblock_type[$j];
9531 # find the starting keyword for this block (such as 'if', 'else', ...)
9533 if ( $block_type =~ /^[\{\}\;\:]$/ ) {
9534 $i_start = $max_index_to_go;
9537 elsif ( $last_last_nonblank_token_to_go eq ')' ) {
9539 # For something like "if (xxx) {", the keyword "if" will be
9540 # just after the most recent break. This will be 0 unless
9541 # we have just killed a one-line block and are starting another.
9543 $i_start = $index_max_forced_break + 1;
9544 if ( $types_to_go[$i_start] eq 'b' ) {
9548 unless ( $tokens_to_go[$i_start] eq $block_type ) {
9553 # the previous nonblank token should start these block types
9555 ( $last_last_nonblank_token_to_go eq $block_type )
9556 || ( $block_type =~ /^sub/
9557 && $last_last_nonblank_token_to_go =~ /^sub/ )
9560 $i_start = $last_last_nonblank_index_to_go;
9563 # patch for SWITCH/CASE to retain one-line case/when blocks
9564 elsif ( $block_type eq 'case' || $block_type eq 'when' ) {
9565 $i_start = $index_max_forced_break + 1;
9566 if ( $types_to_go[$i_start] eq 'b' ) {
9569 unless ( $tokens_to_go[$i_start] eq $block_type ) {
9578 my $pos = total_line_length( $i_start, $max_index_to_go ) - 1;
9582 # see if length is too long to even start
9583 if ( $pos > $rOpts_maximum_line_length ) {
9587 for ( $i = $j + 1 ; $i <= $jmax ; $i++ ) {
9589 # old whitespace could be arbitrarily large, so don't use it
9590 if ( $$rtoken_type[$i] eq 'b' ) { $pos += 1 }
9591 else { $pos += length( $$rtokens[$i] ) }
9593 # Return false result if we exceed the maximum line length,
9594 if ( $pos > $rOpts_maximum_line_length ) {
9598 # or encounter another opening brace before finding the closing brace.
9599 elsif ($$rtokens[$i] eq '{'
9600 && $$rtoken_type[$i] eq '{'
9601 && $$rblock_type[$i] )
9606 # if we find our closing brace..
9607 elsif ($$rtokens[$i] eq '}'
9608 && $$rtoken_type[$i] eq '}'
9609 && $$rblock_type[$i] )
9612 # be sure any trailing comment also fits on the line
9614 ( $$rtoken_type[ $i + 1 ] eq 'b' ) ? $i + 2 : $i + 1;
9616 if ( $$rtoken_type[$i_nonblank] eq '#' ) {
9617 $pos += length( $$rtokens[$i_nonblank] );
9619 if ( $i_nonblank > $i + 1 ) {
9620 $pos += length( $$rtokens[ $i + 1 ] );
9623 if ( $pos > $rOpts_maximum_line_length ) {
9628 # ok, it's a one-line block
9629 create_one_line_block( $i_start, 20 );
9633 # just keep going for other characters
9638 # Allow certain types of new one-line blocks to form by joining
9639 # input lines. These can be safely done, but for other block types,
9640 # we keep old one-line blocks but do not form new ones. It is not
9641 # always a good idea to make as many one-line blocks as possible,
9642 # so other types are not done. The user can always use -mangle.
9643 if ( $is_sort_map_grep_eval{$block_type} ) {
9644 create_one_line_block( $i_start, 1 );
9650 sub unstore_token_to_go {
9652 # remove most recent token from output stream
9653 if ( $max_index_to_go > 0 ) {
9657 $max_index_to_go = UNDEFINED_INDEX;
9662 sub want_blank_line {
9664 $file_writer_object->want_blank_line();
9667 sub write_unindented_line {
9669 $file_writer_object->write_line( $_[0] );
9674 # If there is a single, long parameter within parens, like this:
9676 # $self->command( "/msg "
9678 # . " You said $1, but did you know that it's square was "
9679 # . $1 * $1 . " ?" );
9681 # we can remove the continuation indentation of the 2nd and higher lines
9682 # to achieve this effect, which is more pleasing:
9684 # $self->command("/msg "
9686 # . " You said $1, but did you know that it's square was "
9687 # . $1 * $1 . " ?");
9689 my ( $line_open, $i_start, $closing_index, $ri_first, $ri_last ) = @_;
9690 my $max_line = @$ri_first - 1;
9692 # must be multiple lines
9693 return unless $max_line > $line_open;
9695 my $lev_start = $levels_to_go[$i_start];
9696 my $ci_start_plus = 1 + $ci_levels_to_go[$i_start];
9698 # see if all additional lines in this container have continuation
9701 my $line_1 = 1 + $line_open;
9702 for ( $n = $line_1 ; $n <= $max_line ; ++$n ) {
9703 my $ibeg = $$ri_first[$n];
9704 my $iend = $$ri_last[$n];
9705 if ( $ibeg eq $closing_index ) { $n--; last }
9706 return if ( $lev_start != $levels_to_go[$ibeg] );
9707 return if ( $ci_start_plus != $ci_levels_to_go[$ibeg] );
9708 last if ( $closing_index <= $iend );
9711 # we can reduce the indentation of all continuation lines
9712 my $continuation_line_count = $n - $line_open;
9713 @ci_levels_to_go[ @$ri_first[ $line_1 .. $n ] ] =
9714 (0) x ($continuation_line_count);
9715 @leading_spaces_to_go[ @$ri_first[ $line_1 .. $n ] ] =
9716 @reduced_spaces_to_go[ @$ri_first[ $line_1 .. $n ] ];
9719 sub set_logical_padding {
9721 # Look at a batch of lines and see if extra padding can improve the
9722 # alignment when there are certain leading operators. Here is an
9723 # example, in which some extra space is introduced before
9724 # '( $year' to make it line up with the subsequent lines:
9726 # if ( ( $Year < 1601 )
9727 # || ( $Year > 2899 )
9728 # || ( $EndYear < 1601 )
9729 # || ( $EndYear > 2899 ) )
9731 # &Error_OutOfRange;
9734 my ( $ri_first, $ri_last ) = @_;
9735 my $max_line = @$ri_first - 1;
9737 my ( $ibeg, $ibeg_next, $ibegm, $iend, $iendm, $ipad, $line, $pad_spaces,
9738 $tok_next, $type_next, $has_leading_op_next, $has_leading_op );
9740 # looking at each line of this batch..
9741 foreach $line ( 0 .. $max_line - 1 ) {
9743 # see if the next line begins with a logical operator
9744 $ibeg = $$ri_first[$line];
9745 $iend = $$ri_last[$line];
9746 $ibeg_next = $$ri_first[ $line + 1 ];
9747 $tok_next = $tokens_to_go[$ibeg_next];
9748 $type_next = $types_to_go[$ibeg_next];
9750 $has_leading_op_next = ( $tok_next =~ /^\w/ )
9751 ? $is_chain_operator{$tok_next} # + - * / : ? && ||
9752 : $is_chain_operator{$type_next}; # and, or
9754 next unless ($has_leading_op_next);
9756 # next line must not be at lesser depth
9758 if ( $nesting_depth_to_go[$ibeg] > $nesting_depth_to_go[$ibeg_next] );
9760 # identify the token in this line to be padded on the left
9763 # handle lines at same depth...
9764 if ( $nesting_depth_to_go[$ibeg] == $nesting_depth_to_go[$ibeg_next] ) {
9766 # if this is not first line of the batch ...
9769 # and we have leading operator..
9770 next if $has_leading_op;
9772 # Introduce padding if..
9773 # 1. the previous line is at lesser depth, or
9774 # 2. the previous line ends in an assignment
9775 # 3. the previous line ends in a 'return'
9776 # 4. the previous line ends in a comma
9777 # Example 1: previous line at lesser depth
9778 # if ( ( $Year < 1601 ) # <- we are here but
9779 # || ( $Year > 2899 ) # list has not yet
9780 # || ( $EndYear < 1601 ) # collapsed vertically
9781 # || ( $EndYear > 2899 ) )
9784 # Example 2: previous line ending in assignment:
9786 # $year % 4 ? 0 # <- We are here
9791 # Example 3: previous line ending in comma:
9798 # be sure levels agree (do not indent after an indented 'if')
9799 next if ( $levels_to_go[$ibeg] ne $levels_to_go[$ibeg_next] );
9801 # allow padding on first line after a comma but only if:
9802 # (1) this is line 2 and
9803 # (2) there are at more than three lines and
9804 # (3) lines 3 and 4 have the same leading operator
9805 # These rules try to prevent padding within a long
9806 # comma-separated list.
9808 if ( $types_to_go[$iendm] eq ','
9812 my $ibeg_next_next = $$ri_first[ $line + 2 ];
9813 my $tok_next_next = $tokens_to_go[$ibeg_next_next];
9814 $ok_comma = $tok_next_next eq $tok_next;
9819 $is_assignment{ $types_to_go[$iendm] }
9821 || ( $nesting_depth_to_go[$ibegm] <
9822 $nesting_depth_to_go[$ibeg] )
9823 || ( $types_to_go[$iendm] eq 'k'
9824 && $tokens_to_go[$iendm] eq 'return' )
9827 # we will add padding before the first token
9831 # for first line of the batch..
9834 # WARNING: Never indent if first line is starting in a
9835 # continued quote, which would change the quote.
9836 next if $starting_in_quote;
9838 # if this is text after closing '}'
9839 # then look for an interior token to pad
9840 if ( $types_to_go[$ibeg] eq '}' ) {
9844 # otherwise, we might pad if it looks really good
9847 # we might pad token $ibeg, so be sure that it
9848 # is at the same depth as the next line.
9850 if ( $nesting_depth_to_go[$ibeg] !=
9851 $nesting_depth_to_go[$ibeg_next] );
9853 # We can pad on line 1 of a statement if at least 3
9854 # lines will be aligned. Otherwise, it
9855 # can look very confusing.
9857 # We have to be careful not to pad if there are too few
9858 # lines. The current rule is:
9859 # (1) in general we require at least 3 consecutive lines
9860 # with the same leading chain operator token,
9861 # (2) but an exception is that we only require two lines
9862 # with leading colons if there are no more lines. For example,
9863 # the first $i in the following snippet would get padding
9864 # by the second rule:
9866 # $i == 1 ? ( "First", "Color" )
9867 # : $i == 2 ? ( "Then", "Rarity" )
9868 # : ( "Then", "Name" );
9870 if ( $max_line > 1 ) {
9871 my $leading_token = $tokens_to_go[$ibeg_next];
9874 # never indent line 1 of a '.' series because
9875 # previous line is most likely at same level.
9876 # TODO: we should also look at the leasing_spaces
9877 # of the last output line and skip if it is same
9879 next if ( $leading_token eq '.' );
9882 foreach my $l ( 2 .. 3 ) {
9883 last if ( $line + $l > $max_line );
9884 my $ibeg_next_next = $$ri_first[ $line + $l ];
9885 if ( $tokens_to_go[$ibeg_next_next] ne
9893 next if ($tokens_differ);
9894 next if ( $count < 3 && $leading_token ne ':' );
9904 # find interior token to pad if necessary
9905 if ( !defined($ipad) ) {
9907 for ( my $i = $ibeg ; ( $i < $iend ) && !$ipad ; $i++ ) {
9909 # find any unclosed container
9911 unless ( $type_sequence_to_go[$i]
9912 && $mate_index_to_go[$i] > $iend );
9914 # find next nonblank token to pad
9916 if ( $types_to_go[$ipad] eq 'b' ) {
9918 last if ( $ipad > $iend );
9924 # next line must not be at greater depth
9925 my $iend_next = $$ri_last[ $line + 1 ];
9927 if ( $nesting_depth_to_go[ $iend_next + 1 ] >
9928 $nesting_depth_to_go[$ipad] );
9930 # lines must be somewhat similar to be padded..
9931 my $inext_next = $ibeg_next + 1;
9932 if ( $types_to_go[$inext_next] eq 'b' ) {
9935 my $type = $types_to_go[$ipad];
9936 my $type_next = $types_to_go[ $ipad + 1 ];
9938 # see if there are multiple continuation lines
9939 my $logical_continuation_lines = 1;
9940 if ( $line + 2 <= $max_line ) {
9941 my $leading_token = $tokens_to_go[$ibeg_next];
9942 my $ibeg_next_next = $$ri_first[ $line + 2 ];
9943 if ( $tokens_to_go[$ibeg_next_next] eq $leading_token
9944 && $nesting_depth_to_go[$ibeg_next] eq
9945 $nesting_depth_to_go[$ibeg_next_next] )
9947 $logical_continuation_lines++;
9951 # see if leading types match
9952 my $types_match = $types_to_go[$inext_next] eq $type;
9953 my $matches_without_bang;
9955 # if first line has leading ! then compare the following token
9956 if ( !$types_match && $type eq '!' ) {
9957 $types_match = $matches_without_bang =
9958 $types_to_go[$inext_next] eq $types_to_go[ $ipad + 1 ];
9963 # either we have multiple continuation lines to follow
9964 # and we are not padding the first token
9965 ( $logical_continuation_lines > 1 && $ipad > 0 )
9973 # and keywords must match if keyword
9976 && $tokens_to_go[$ipad] ne $tokens_to_go[$inext_next]
9982 #----------------------begin special checks--------------
9985 # A check is needed before we can make the pad.
9986 # If we are in a list with some long items, we want each
9987 # item to stand out. So in the following example, the
9988 # first line begining with '$casefold->' would look good
9989 # padded to align with the next line, but then it
9990 # would be indented more than the last line, so we
9994 # $casefold->{code} eq '0041'
9995 # && $casefold->{status} eq 'C'
9996 # && $casefold->{mapping} eq '0061',
10001 # It would be faster, and almost as good, to use a comma
10002 # count, and not pad if comma_count > 1 and the previous
10003 # line did not end with a comma.
10007 my $ibg = $$ri_first[ $line + 1 ];
10008 my $depth = $nesting_depth_to_go[ $ibg + 1 ];
10010 # just use simplified formula for leading spaces to avoid
10011 # needless sub calls
10012 my $lsp = $levels_to_go[$ibg] + $ci_levels_to_go[$ibg];
10014 # look at each line beyond the next ..
10016 foreach $l ( $line + 2 .. $max_line ) {
10017 my $ibg = $$ri_first[$l];
10019 # quit looking at the end of this container
10021 if ( $nesting_depth_to_go[ $ibg + 1 ] < $depth )
10022 || ( $nesting_depth_to_go[$ibg] < $depth );
10024 # cannot do the pad if a later line would be
10026 if ( $levels_to_go[$ibg] + $ci_levels_to_go[$ibg] < $lsp ) {
10032 # don't pad if we end in a broken list
10033 if ( $l == $max_line ) {
10034 my $i2 = $$ri_last[$l];
10035 if ( $types_to_go[$i2] eq '#' ) {
10036 my $i1 = $$ri_first[$l];
10039 terminal_type( \@types_to_go, \@block_type_to_go, $i1,
10046 # a minus may introduce a quoted variable, and we will
10047 # add the pad only if this line begins with a bare word,
10048 # such as for the word 'Button' here:
10050 # Button => "Print letter \"~$_\"",
10051 # -command => [ sub { print "$_[0]\n" }, $_ ],
10052 # -accelerator => "Meta+$_"
10055 # On the other hand, if 'Button' is quoted, it looks best
10058 # 'Button' => "Print letter \"~$_\"",
10059 # -command => [ sub { print "$_[0]\n" }, $_ ],
10060 # -accelerator => "Meta+$_"
10062 if ( $types_to_go[$ibeg_next] eq 'm' ) {
10063 $ok_to_pad = 0 if $types_to_go[$ibeg] eq 'Q';
10066 next unless $ok_to_pad;
10068 #----------------------end special check---------------
10070 my $length_1 = total_line_length( $ibeg, $ipad - 1 );
10071 my $length_2 = total_line_length( $ibeg_next, $inext_next - 1 );
10072 $pad_spaces = $length_2 - $length_1;
10074 # If the first line has a leading ! and the second does
10075 # not, then remove one space to try to align the next
10076 # leading characters, which are often the same. For example:
10078 # || $ts == $self->Holder
10079 # || $self->Holder->Type eq "Arena" )
10081 # This usually helps readability, but if there are subsequent
10082 # ! operators things will still get messed up. For example:
10084 # if ( !exists $Net::DNS::typesbyname{$qtype}
10085 # && exists $Net::DNS::classesbyname{$qtype}
10086 # && !exists $Net::DNS::classesbyname{$qclass}
10087 # && exists $Net::DNS::typesbyname{$qclass} )
10088 # We can't fix that.
10089 if ($matches_without_bang) { $pad_spaces-- }
10091 # make sure this won't change if -lp is used
10092 my $indentation_1 = $leading_spaces_to_go[$ibeg];
10093 if ( ref($indentation_1) ) {
10094 if ( $indentation_1->get_RECOVERABLE_SPACES() == 0 ) {
10095 my $indentation_2 = $leading_spaces_to_go[$ibeg_next];
10096 unless ( $indentation_2->get_RECOVERABLE_SPACES() == 0 ) {
10102 # we might be able to handle a pad of -1 by removing a blank
10104 if ( $pad_spaces < 0 ) {
10106 if ( $pad_spaces == -1 ) {
10107 if ( $ipad > $ibeg && $types_to_go[ $ipad - 1 ] eq 'b' ) {
10108 $tokens_to_go[ $ipad - 1 ] = '';
10114 # now apply any padding for alignment
10115 if ( $ipad >= 0 && $pad_spaces ) {
10117 my $length_t = total_line_length( $ibeg, $iend );
10118 if ( $pad_spaces + $length_t <= $rOpts_maximum_line_length ) {
10119 $tokens_to_go[$ipad] =
10120 ' ' x $pad_spaces . $tokens_to_go[$ipad];
10128 $has_leading_op = $has_leading_op_next;
10129 } # end of loop over lines
10133 sub correct_lp_indentation {
10135 # When the -lp option is used, we need to make a last pass through
10136 # each line to correct the indentation positions in case they differ
10137 # from the predictions. This is necessary because perltidy uses a
10138 # predictor/corrector method for aligning with opening parens. The
10139 # predictor is usually good, but sometimes stumbles. The corrector
10140 # tries to patch things up once the actual opening paren locations
10142 my ( $ri_first, $ri_last ) = @_;
10143 my $do_not_pad = 0;
10145 # Note on flag '$do_not_pad':
10146 # We want to avoid a situation like this, where the aligner inserts
10147 # whitespace before the '=' to align it with a previous '=', because
10148 # otherwise the parens might become mis-aligned in a situation like
10149 # this, where the '=' has become aligned with the previous line,
10150 # pushing the opening '(' forward beyond where we want it.
10152 # $mkFloor::currentRoom = '';
10153 # $mkFloor::c_entry = $c->Entry(
10155 # -relief => 'sunken',
10159 # We leave it to the aligner to decide how to do this.
10161 # first remove continuation indentation if appropriate
10162 my $max_line = @$ri_first - 1;
10164 # looking at each line of this batch..
10165 my ( $ibeg, $iend );
10167 foreach $line ( 0 .. $max_line ) {
10168 $ibeg = $$ri_first[$line];
10169 $iend = $$ri_last[$line];
10171 # looking at each token in this output line..
10173 foreach $i ( $ibeg .. $iend ) {
10175 # How many space characters to place before this token
10176 # for special alignment. Actual padding is done in the
10179 # looking for next unvisited indentation item
10180 my $indentation = $leading_spaces_to_go[$i];
10181 if ( !$indentation->get_MARKED() ) {
10182 $indentation->set_MARKED(1);
10184 # looking for indentation item for which we are aligning
10185 # with parens, braces, and brackets
10186 next unless ( $indentation->get_ALIGN_PAREN() );
10188 # skip closed container on this line
10189 if ( $i > $ibeg ) {
10191 if ( $types_to_go[$im] eq 'b' && $im > $ibeg ) { $im-- }
10192 if ( $type_sequence_to_go[$im]
10193 && $mate_index_to_go[$im] <= $iend )
10199 if ( $line == 1 && $i == $ibeg ) {
10203 # Ok, let's see what the error is and try to fix it
10205 my $predicted_pos = $indentation->get_SPACES();
10206 if ( $i > $ibeg ) {
10208 # token is mid-line - use length to previous token
10209 $actual_pos = total_line_length( $ibeg, $i - 1 );
10211 # for mid-line token, we must check to see if all
10212 # additional lines have continuation indentation,
10213 # and remove it if so. Otherwise, we do not get
10215 my $closing_index = $indentation->get_CLOSED();
10216 if ( $closing_index > $iend ) {
10217 my $ibeg_next = $$ri_first[ $line + 1 ];
10218 if ( $ci_levels_to_go[$ibeg_next] > 0 ) {
10219 undo_lp_ci( $line, $i, $closing_index, $ri_first,
10224 elsif ( $line > 0 ) {
10226 # handle case where token starts a new line;
10227 # use length of previous line
10228 my $ibegm = $$ri_first[ $line - 1 ];
10229 my $iendm = $$ri_last[ $line - 1 ];
10230 $actual_pos = total_line_length( $ibegm, $iendm );
10234 if ( $types_to_go[ $iendm + 1 ] eq 'b' );
10238 # token is first character of first line of batch
10239 $actual_pos = $predicted_pos;
10242 my $move_right = $actual_pos - $predicted_pos;
10244 # done if no error to correct (gnu2.t)
10245 if ( $move_right == 0 ) {
10246 $indentation->set_RECOVERABLE_SPACES($move_right);
10250 # if we have not seen closure for this indentation in
10251 # this batch, we can only pass on a request to the
10253 my $closing_index = $indentation->get_CLOSED();
10255 if ( $closing_index < 0 ) {
10256 $indentation->set_RECOVERABLE_SPACES($move_right);
10260 # If necessary, look ahead to see if there is really any
10261 # leading whitespace dependent on this whitespace, and
10262 # also find the longest line using this whitespace.
10263 # Since it is always safe to move left if there are no
10264 # dependents, we only need to do this if we may have
10265 # dependent nodes or need to move right.
10267 my $right_margin = 0;
10268 my $have_child = $indentation->get_HAVE_CHILD();
10270 my %saw_indentation;
10271 my $line_count = 1;
10272 $saw_indentation{$indentation} = $indentation;
10274 if ( $have_child || $move_right > 0 ) {
10276 my $max_length = 0;
10277 if ( $i == $ibeg ) {
10278 $max_length = total_line_length( $ibeg, $iend );
10281 # look ahead at the rest of the lines of this batch..
10283 foreach $line_t ( $line + 1 .. $max_line ) {
10284 my $ibeg_t = $$ri_first[$line_t];
10285 my $iend_t = $$ri_last[$line_t];
10286 last if ( $closing_index <= $ibeg_t );
10288 # remember all different indentation objects
10289 my $indentation_t = $leading_spaces_to_go[$ibeg_t];
10290 $saw_indentation{$indentation_t} = $indentation_t;
10293 # remember longest line in the group
10294 my $length_t = total_line_length( $ibeg_t, $iend_t );
10295 if ( $length_t > $max_length ) {
10296 $max_length = $length_t;
10299 $right_margin = $rOpts_maximum_line_length - $max_length;
10300 if ( $right_margin < 0 ) { $right_margin = 0 }
10303 my $first_line_comma_count =
10304 grep { $_ eq ',' } @types_to_go[ $ibeg .. $iend ];
10305 my $comma_count = $indentation->get_COMMA_COUNT();
10306 my $arrow_count = $indentation->get_ARROW_COUNT();
10308 # This is a simple approximate test for vertical alignment:
10309 # if we broke just after an opening paren, brace, bracket,
10310 # and there are 2 or more commas in the first line,
10311 # and there are no '=>'s,
10312 # then we are probably vertically aligned. We could set
10313 # an exact flag in sub scan_list, but this is good
10315 my $indentation_count = keys %saw_indentation;
10316 my $is_vertically_aligned =
10318 && $first_line_comma_count > 1
10319 && $indentation_count == 1
10320 && ( $arrow_count == 0 || $arrow_count == $line_count ) );
10322 # Make the move if possible ..
10325 # we can always move left
10328 # but we should only move right if we are sure it will
10329 # not spoil vertical alignment
10330 || ( $comma_count == 0 )
10331 || ( $comma_count > 0 && !$is_vertically_aligned )
10335 ( $move_right <= $right_margin )
10339 foreach ( keys %saw_indentation ) {
10340 $saw_indentation{$_}
10341 ->permanently_decrease_AVAILABLE_SPACES( -$move );
10345 # Otherwise, record what we want and the vertical aligner
10346 # will try to recover it.
10348 $indentation->set_RECOVERABLE_SPACES($move_right);
10353 return $do_not_pad;
10356 # flush is called to output any tokens in the pipeline, so that
10357 # an alternate source of lines can be written in the correct order
10360 destroy_one_line_block();
10361 output_line_to_go();
10362 Perl::Tidy::VerticalAligner::flush();
10365 sub reset_block_text_accumulator {
10367 # save text after 'if' and 'elsif' to append after 'else'
10368 if ($accumulating_text_for_block) {
10370 if ( $accumulating_text_for_block =~ /^(if|elsif)$/ ) {
10371 push @{$rleading_block_if_elsif_text}, $leading_block_text;
10374 $accumulating_text_for_block = "";
10375 $leading_block_text = "";
10376 $leading_block_text_level = 0;
10377 $leading_block_text_length_exceeded = 0;
10378 $leading_block_text_line_number = 0;
10379 $leading_block_text_line_length = 0;
10382 sub set_block_text_accumulator {
10384 $accumulating_text_for_block = $tokens_to_go[$i];
10385 if ( $accumulating_text_for_block !~ /^els/ ) {
10386 $rleading_block_if_elsif_text = [];
10388 $leading_block_text = "";
10389 $leading_block_text_level = $levels_to_go[$i];
10390 $leading_block_text_line_number =
10391 $vertical_aligner_object->get_output_line_number();
10392 $leading_block_text_length_exceeded = 0;
10394 # this will contain the column number of the last character
10395 # of the closing side comment
10396 $leading_block_text_line_length =
10397 length($accumulating_text_for_block) +
10398 length( $rOpts->{'closing-side-comment-prefix'} ) +
10399 $leading_block_text_level * $rOpts_indent_columns + 3;
10402 sub accumulate_block_text {
10405 # accumulate leading text for -csc, ignoring any side comments
10406 if ( $accumulating_text_for_block
10407 && !$leading_block_text_length_exceeded
10408 && $types_to_go[$i] ne '#' )
10411 my $added_length = length( $tokens_to_go[$i] );
10412 $added_length += 1 if $i == 0;
10413 my $new_line_length = $leading_block_text_line_length + $added_length;
10415 # we can add this text if we don't exceed some limits..
10418 # we must not have already exceeded the text length limit
10419 length($leading_block_text) <
10420 $rOpts_closing_side_comment_maximum_text
10423 # the new total line length must be below the line length limit
10424 # or the new length must be below the text length limit
10425 # (ie, we may allow one token to exceed the text length limit)
10426 && ( $new_line_length < $rOpts_maximum_line_length
10427 || length($leading_block_text) + $added_length <
10428 $rOpts_closing_side_comment_maximum_text )
10430 # UNLESS: we are adding a closing paren before the brace we seek.
10431 # This is an attempt to avoid situations where the ... to be
10432 # added are longer than the omitted right paren, as in:
10434 # foreach my $item (@a_rather_long_variable_name_here) {
10436 # } ## end foreach my $item (@a_rather_long_variable_name_here...
10439 $tokens_to_go[$i] eq ')'
10442 $i + 1 <= $max_index_to_go
10443 && $block_type_to_go[ $i + 1 ] eq
10444 $accumulating_text_for_block
10446 || ( $i + 2 <= $max_index_to_go
10447 && $block_type_to_go[ $i + 2 ] eq
10448 $accumulating_text_for_block )
10454 # add an extra space at each newline
10455 if ( $i == 0 ) { $leading_block_text .= ' ' }
10457 # add the token text
10458 $leading_block_text .= $tokens_to_go[$i];
10459 $leading_block_text_line_length = $new_line_length;
10462 # show that text was truncated if necessary
10463 elsif ( $types_to_go[$i] ne 'b' ) {
10464 $leading_block_text_length_exceeded = 1;
10465 $leading_block_text .= '...';
10471 my %is_if_elsif_else_unless_while_until_for_foreach;
10475 # These block types may have text between the keyword and opening
10476 # curly. Note: 'else' does not, but must be included to allow trailing
10477 # if/elsif text to be appended.
10478 # patch for SWITCH/CASE: added 'case' and 'when'
10479 @_ = qw(if elsif else unless while until for foreach case when);
10480 @is_if_elsif_else_unless_while_until_for_foreach{@_} = (1) x scalar(@_);
10483 sub accumulate_csc_text {
10485 # called once per output buffer when -csc is used. Accumulates
10486 # the text placed after certain closing block braces.
10487 # Defines and returns the following for this buffer:
10489 my $block_leading_text = ""; # the leading text of the last '}'
10490 my $rblock_leading_if_elsif_text;
10491 my $i_block_leading_text =
10492 -1; # index of token owning block_leading_text
10493 my $block_line_count = 100; # how many lines the block spans
10494 my $terminal_type = 'b'; # type of last nonblank token
10495 my $i_terminal = 0; # index of last nonblank token
10496 my $terminal_block_type = "";
10498 for my $i ( 0 .. $max_index_to_go ) {
10499 my $type = $types_to_go[$i];
10500 my $block_type = $block_type_to_go[$i];
10501 my $token = $tokens_to_go[$i];
10503 # remember last nonblank token type
10504 if ( $type ne '#' && $type ne 'b' ) {
10505 $terminal_type = $type;
10506 $terminal_block_type = $block_type;
10510 my $type_sequence = $type_sequence_to_go[$i];
10511 if ( $block_type && $type_sequence ) {
10513 if ( $token eq '}' ) {
10515 # restore any leading text saved when we entered this block
10516 if ( defined( $block_leading_text{$type_sequence} ) ) {
10517 ( $block_leading_text, $rblock_leading_if_elsif_text ) =
10518 @{ $block_leading_text{$type_sequence} };
10519 $i_block_leading_text = $i;
10520 delete $block_leading_text{$type_sequence};
10521 $rleading_block_if_elsif_text =
10522 $rblock_leading_if_elsif_text;
10525 # if we run into a '}' then we probably started accumulating
10526 # at something like a trailing 'if' clause..no harm done.
10527 if ( $accumulating_text_for_block
10528 && $levels_to_go[$i] <= $leading_block_text_level )
10530 my $lev = $levels_to_go[$i];
10531 reset_block_text_accumulator();
10534 if ( defined( $block_opening_line_number{$type_sequence} ) )
10536 my $output_line_number =
10537 $vertical_aligner_object->get_output_line_number();
10538 $block_line_count =
10539 $output_line_number -
10540 $block_opening_line_number{$type_sequence} + 1;
10541 delete $block_opening_line_number{$type_sequence};
10545 # Error: block opening line undefined for this line..
10546 # This shouldn't be possible, but it is not a
10547 # significant problem.
10551 elsif ( $token eq '{' ) {
10554 $vertical_aligner_object->get_output_line_number();
10555 $block_opening_line_number{$type_sequence} = $line_number;
10557 if ( $accumulating_text_for_block
10558 && $levels_to_go[$i] == $leading_block_text_level )
10561 if ( $accumulating_text_for_block eq $block_type ) {
10563 # save any leading text before we enter this block
10564 $block_leading_text{$type_sequence} = [
10565 $leading_block_text,
10566 $rleading_block_if_elsif_text
10568 $block_opening_line_number{$type_sequence} =
10569 $leading_block_text_line_number;
10570 reset_block_text_accumulator();
10574 # shouldn't happen, but not a serious error.
10575 # We were accumulating -csc text for block type
10576 # $accumulating_text_for_block and unexpectedly
10577 # encountered a '{' for block type $block_type.
10584 && $csc_new_statement_ok
10585 && $is_if_elsif_else_unless_while_until_for_foreach{$token}
10586 && $token =~ /$closing_side_comment_list_pattern/o )
10588 set_block_text_accumulator($i);
10592 # note: ignoring type 'q' because of tricks being played
10593 # with 'q' for hanging side comments
10594 if ( $type ne 'b' && $type ne '#' && $type ne 'q' ) {
10595 $csc_new_statement_ok =
10596 ( $block_type || $type eq 'J' || $type eq ';' );
10599 && $accumulating_text_for_block
10600 && $levels_to_go[$i] == $leading_block_text_level )
10602 reset_block_text_accumulator();
10605 accumulate_block_text($i);
10610 # Treat an 'else' block specially by adding preceding 'if' and
10611 # 'elsif' text. Otherwise, the 'end else' is not helpful,
10612 # especially for cuddled-else formatting.
10613 if ( $terminal_block_type =~ /^els/ && $rblock_leading_if_elsif_text ) {
10614 $block_leading_text =
10615 make_else_csc_text( $i_terminal, $terminal_block_type,
10616 $block_leading_text, $rblock_leading_if_elsif_text );
10619 return ( $terminal_type, $i_terminal, $i_block_leading_text,
10620 $block_leading_text, $block_line_count );
10624 sub make_else_csc_text {
10626 # create additional -csc text for an 'else' and optionally 'elsif',
10627 # depending on the value of switch
10628 # $rOpts_closing_side_comment_else_flag:
10630 # = 0 add 'if' text to trailing else
10631 # = 1 same as 0 plus:
10632 # add 'if' to 'elsif's if can fit in line length
10633 # add last 'elsif' to trailing else if can fit in one line
10634 # = 2 same as 1 but do not check if exceed line length
10636 # $rif_elsif_text = a reference to a list of all previous closing
10637 # side comments created for this if block
10639 my ( $i_terminal, $block_type, $block_leading_text, $rif_elsif_text ) = @_;
10640 my $csc_text = $block_leading_text;
10642 if ( $block_type eq 'elsif' && $rOpts_closing_side_comment_else_flag == 0 )
10647 my $count = @{$rif_elsif_text};
10648 return $csc_text unless ($count);
10650 my $if_text = '[ if' . $rif_elsif_text->[0];
10652 # always show the leading 'if' text on 'else'
10653 if ( $block_type eq 'else' ) {
10654 $csc_text .= $if_text;
10657 # see if that's all
10658 if ( $rOpts_closing_side_comment_else_flag == 0 ) {
10662 my $last_elsif_text = "";
10663 if ( $count > 1 ) {
10664 $last_elsif_text = ' [elsif' . $rif_elsif_text->[ $count - 1 ];
10665 if ( $count > 2 ) { $last_elsif_text = ' [...' . $last_elsif_text; }
10668 # tentatively append one more item
10669 my $saved_text = $csc_text;
10670 if ( $block_type eq 'else' ) {
10671 $csc_text .= $last_elsif_text;
10674 $csc_text .= ' ' . $if_text;
10677 # all done if no length checks requested
10678 if ( $rOpts_closing_side_comment_else_flag == 2 ) {
10682 # undo it if line length exceeded
10684 length($csc_text) +
10685 length($block_type) +
10686 length( $rOpts->{'closing-side-comment-prefix'} ) +
10687 $levels_to_go[$i_terminal] * $rOpts_indent_columns + 3;
10688 if ( $length > $rOpts_maximum_line_length ) {
10689 $csc_text = $saved_text;
10694 sub add_closing_side_comment {
10696 # add closing side comments after closing block braces if -csc used
10697 my $cscw_block_comment;
10699 #---------------------------------------------------------------
10700 # Step 1: loop through all tokens of this line to accumulate
10701 # the text needed to create the closing side comments. Also see
10702 # how the line ends.
10703 #---------------------------------------------------------------
10705 my ( $terminal_type, $i_terminal, $i_block_leading_text,
10706 $block_leading_text, $block_line_count )
10707 = accumulate_csc_text();
10709 #---------------------------------------------------------------
10710 # Step 2: make the closing side comment if this ends a block
10711 #---------------------------------------------------------------
10712 my $have_side_comment = $i_terminal != $max_index_to_go;
10714 # if this line might end in a block closure..
10716 $terminal_type eq '}'
10721 # the block is long enough
10722 ( $block_line_count >= $rOpts->{'closing-side-comment-interval'} )
10724 # or there is an existing comment to check
10725 || ( $have_side_comment
10726 && $rOpts->{'closing-side-comment-warnings'} )
10729 # .. and if this is one of the types of interest
10730 && $block_type_to_go[$i_terminal] =~
10731 /$closing_side_comment_list_pattern/o
10733 # .. but not an anonymous sub
10734 # These are not normally of interest, and their closing braces are
10735 # often followed by commas or semicolons anyway. This also avoids
10736 # possible erratic output due to line numbering inconsistencies
10737 # in the cases where their closing braces terminate a line.
10738 && $block_type_to_go[$i_terminal] ne 'sub'
10740 # ..and the corresponding opening brace must is not in this batch
10741 # (because we do not need to tag one-line blocks, although this
10742 # should also be caught with a positive -csci value)
10743 && $mate_index_to_go[$i_terminal] < 0
10748 # this is the last token (line doesnt have a side comment)
10749 !$have_side_comment
10751 # or the old side comment is a closing side comment
10752 || $tokens_to_go[$max_index_to_go] =~
10753 /$closing_side_comment_prefix_pattern/o
10758 # then make the closing side comment text
10760 "$rOpts->{'closing-side-comment-prefix'} $block_type_to_go[$i_terminal]";
10762 # append any extra descriptive text collected above
10763 if ( $i_block_leading_text == $i_terminal ) {
10764 $token .= $block_leading_text;
10766 $token =~ s/\s*$//; # trim any trailing whitespace
10768 # handle case of existing closing side comment
10769 if ($have_side_comment) {
10771 # warn if requested and tokens differ significantly
10772 if ( $rOpts->{'closing-side-comment-warnings'} ) {
10773 my $old_csc = $tokens_to_go[$max_index_to_go];
10774 my $new_csc = $token;
10775 $new_csc =~ s/(\.\.\.)\s*$//; # trim trailing '...'
10776 my $new_trailing_dots = $1;
10777 $old_csc =~ s/\.\.\.\s*$//;
10778 $new_csc =~ s/\s+//g; # trim all whitespace
10779 $old_csc =~ s/\s+//g;
10781 # Patch to handle multiple closing side comments at
10782 # else and elsif's. These have become too complicated
10783 # to check, so if we see an indication of
10784 # '[ if' or '[ # elsif', then assume they were made
10786 if ( $block_type_to_go[$i_terminal] eq 'else' ) {
10787 if ( $old_csc =~ /\[\s*elsif/ ) { $old_csc = $new_csc }
10789 elsif ( $block_type_to_go[$i_terminal] eq 'elsif' ) {
10790 if ( $old_csc =~ /\[\s*if/ ) { $old_csc = $new_csc }
10793 # if old comment is contained in new comment,
10794 # only compare the common part.
10795 if ( length($new_csc) > length($old_csc) ) {
10796 $new_csc = substr( $new_csc, 0, length($old_csc) );
10799 # if the new comment is shorter and has been limited,
10800 # only compare the common part.
10801 if ( length($new_csc) < length($old_csc) && $new_trailing_dots )
10803 $old_csc = substr( $old_csc, 0, length($new_csc) );
10806 # any remaining difference?
10807 if ( $new_csc ne $old_csc ) {
10809 # just leave the old comment if we are below the threshold
10810 # for creating side comments
10811 if ( $block_line_count <
10812 $rOpts->{'closing-side-comment-interval'} )
10817 # otherwise we'll make a note of it
10821 "perltidy -cscw replaced: $tokens_to_go[$max_index_to_go]\n"
10824 # save the old side comment in a new trailing block comment
10825 my ( $day, $month, $year ) = (localtime)[ 3, 4, 5 ];
10828 $cscw_block_comment =
10829 "## perltidy -cscw $year-$month-$day: $tokens_to_go[$max_index_to_go]";
10834 # No differences.. we can safely delete old comment if we
10835 # are below the threshold
10836 if ( $block_line_count <
10837 $rOpts->{'closing-side-comment-interval'} )
10840 unstore_token_to_go()
10841 if ( $types_to_go[$max_index_to_go] eq '#' );
10842 unstore_token_to_go()
10843 if ( $types_to_go[$max_index_to_go] eq 'b' );
10848 # switch to the new csc (unless we deleted it!)
10849 $tokens_to_go[$max_index_to_go] = $token if $token;
10852 # handle case of NO existing closing side comment
10855 # insert the new side comment into the output token stream
10857 my $block_type = '';
10858 my $type_sequence = '';
10859 my $container_environment =
10860 $container_environment_to_go[$max_index_to_go];
10861 my $level = $levels_to_go[$max_index_to_go];
10862 my $slevel = $nesting_depth_to_go[$max_index_to_go];
10863 my $no_internal_newlines = 0;
10865 my $nesting_blocks = $nesting_blocks_to_go[$max_index_to_go];
10866 my $ci_level = $ci_levels_to_go[$max_index_to_go];
10867 my $in_continued_quote = 0;
10869 # first insert a blank token
10870 insert_new_token_to_go( ' ', 'b', $slevel, $no_internal_newlines );
10872 # then the side comment
10873 insert_new_token_to_go( $token, $type, $slevel,
10874 $no_internal_newlines );
10877 return $cscw_block_comment;
10880 sub previous_nonblank_token {
10884 return "" if ( $im < 0 );
10885 if ( $types_to_go[$im] eq 'b' ) { $im--; }
10886 return "" if ( $im < 0 );
10887 $name = $tokens_to_go[$im];
10889 # prepend any sub name to an isolated -> to avoid unwanted alignments
10890 # [test case is test8/penco.pl]
10891 if ( $name eq '->' ) {
10893 if ( $im >= 0 && $types_to_go[$im] ne 'b' ) {
10894 $name = $tokens_to_go[$im] . $name;
10900 sub send_lines_to_vertical_aligner {
10902 my ( $ri_first, $ri_last, $do_not_pad ) = @_;
10904 my $rindentation_list = [0]; # ref to indentations for each line
10906 # define the array @matching_token_to_go for the output tokens
10907 # which will be non-blank for each special token (such as =>)
10908 # for which alignment is required.
10909 set_vertical_alignment_markers( $ri_first, $ri_last );
10911 # flush if necessary to avoid unwanted alignment
10912 my $must_flush = 0;
10913 if ( @$ri_first > 1 ) {
10915 # flush before a long if statement
10916 if ( $types_to_go[0] eq 'k' && $tokens_to_go[0] =~ /^(if|unless)$/ ) {
10921 Perl::Tidy::VerticalAligner::flush();
10924 set_logical_padding( $ri_first, $ri_last );
10926 # loop to prepare each line for shipment
10927 my $n_last_line = @$ri_first - 1;
10929 for my $n ( 0 .. $n_last_line ) {
10930 my $ibeg = $$ri_first[$n];
10931 my $iend = $$ri_last[$n];
10933 my ( $rtokens, $rfields, $rpatterns ) =
10934 make_alignment_patterns( $ibeg, $iend );
10936 my ( $indentation, $lev, $level_end, $terminal_type,
10937 $is_semicolon_terminated, $is_outdented_line )
10938 = set_adjusted_indentation( $ibeg, $iend, $rfields, $rpatterns,
10939 $ri_first, $ri_last, $rindentation_list );
10941 # we will allow outdenting of long lines..
10942 my $outdent_long_lines = (
10944 # which are long quotes, if allowed
10945 ( $types_to_go[$ibeg] eq 'Q' && $rOpts->{'outdent-long-quotes'} )
10947 # which are long block comments, if allowed
10949 $types_to_go[$ibeg] eq '#'
10950 && $rOpts->{'outdent-long-comments'}
10952 # but not if this is a static block comment
10953 && !$is_static_block_comment
10958 $nesting_depth_to_go[ $iend + 1 ] - $nesting_depth_to_go[$ibeg];
10960 my $rvertical_tightness_flags =
10961 set_vertical_tightness_flags( $n, $n_last_line, $ibeg, $iend,
10962 $ri_first, $ri_last );
10964 # flush an outdented line to avoid any unwanted vertical alignment
10965 Perl::Tidy::VerticalAligner::flush() if ($is_outdented_line);
10967 my $is_terminal_ternary = 0;
10968 if ( $tokens_to_go[$ibeg] eq ':'
10969 || $n > 0 && $tokens_to_go[ $$ri_last[ $n - 1 ] ] eq ':' )
10971 if ( ( $terminal_type eq ';' && $level_end <= $lev )
10972 || ( $level_end < $lev ) )
10974 $is_terminal_ternary = 1;
10978 # send this new line down the pipe
10979 my $forced_breakpoint = $forced_breakpoint_to_go[$iend];
10980 Perl::Tidy::VerticalAligner::append_line(
10987 $forced_breakpoint_to_go[$iend] || $in_comma_list,
10988 $outdent_long_lines,
10989 $is_terminal_ternary,
10990 $is_semicolon_terminated,
10992 $rvertical_tightness_flags,
10996 $tokens_to_go[$iend] eq ',' && $forced_breakpoint_to_go[$iend];
10998 # flush an outdented line to avoid any unwanted vertical alignment
10999 Perl::Tidy::VerticalAligner::flush() if ($is_outdented_line);
11003 } # end of loop to output each line
11005 # remember indentation of lines containing opening containers for
11006 # later use by sub set_adjusted_indentation
11007 save_opening_indentation( $ri_first, $ri_last, $rindentation_list );
11010 { # begin make_alignment_patterns
11012 my %block_type_map;
11017 # map related block names into a common name to
11019 %block_type_map = (
11030 # map certain keywords to the same 'if' class to align
11031 # long if/elsif sequences. [elsif.pl]
11037 'default' => 'given',
11038 'case' => 'switch',
11040 # treat an 'undef' similar to numbers and quotes
11045 sub make_alignment_patterns {
11047 # Here we do some important preliminary work for the
11048 # vertical aligner. We create three arrays for one
11049 # output line. These arrays contain strings that can
11050 # be tested by the vertical aligner to see if
11051 # consecutive lines can be aligned vertically.
11053 # The three arrays are indexed on the vertical
11054 # alignment fields and are:
11055 # @tokens - a list of any vertical alignment tokens for this line.
11056 # These are tokens, such as '=' '&&' '#' etc which
11057 # we want to might align vertically. These are
11058 # decorated with various information such as
11059 # nesting depth to prevent unwanted vertical
11060 # alignment matches.
11061 # @fields - the actual text of the line between the vertical alignment
11063 # @patterns - a modified list of token types, one for each alignment
11064 # field. These should normally each match before alignment is
11065 # allowed, even when the alignment tokens match.
11066 my ( $ibeg, $iend ) = @_;
11070 my $i_start = $ibeg;
11074 my @container_name = ("");
11075 my @multiple_comma_arrows = (undef);
11077 my $j = 0; # field index
11080 for $i ( $ibeg .. $iend ) {
11082 # Keep track of containers balanced on this line only.
11083 # These are used below to prevent unwanted cross-line alignments.
11084 # Unbalanced containers already avoid aligning across
11085 # container boundaries.
11086 if ( $tokens_to_go[$i] eq '(' ) {
11088 # if container is balanced on this line...
11089 my $i_mate = $mate_index_to_go[$i];
11090 if ( $i_mate > $i && $i_mate <= $iend ) {
11092 my $seqno = $type_sequence_to_go[$i];
11093 my $count = comma_arrow_count($seqno);
11094 $multiple_comma_arrows[$depth] = $count && $count > 1;
11096 # Append the previous token name to make the container name
11097 # more unique. This name will also be given to any commas
11098 # within this container, and it helps avoid undesirable
11099 # alignments of different types of containers.
11100 my $name = previous_nonblank_token($i);
11102 $container_name[$depth] = "+" . $name;
11104 # Make the container name even more unique if necessary.
11105 # If we are not vertically aligning this opening paren,
11106 # append a character count to avoid bad alignment because
11107 # it usually looks bad to align commas within continers
11108 # for which the opening parens do not align. Here
11109 # is an example very BAD alignment of commas (because
11110 # the atan2 functions are not all aligned):
11112 # $X * $RTYSQP1 * atan2( $X, $RTYSQP1 ) +
11113 # $Y * $RTXSQP1 * atan2( $Y, $RTXSQP1 ) -
11114 # $X * atan2( $X, 1 ) -
11115 # $Y * atan2( $Y, 1 );
11117 # On the other hand, it is usually okay to align commas if
11118 # opening parens align, such as:
11119 # glVertex3d( $cx + $s * $xs, $cy, $z );
11120 # glVertex3d( $cx, $cy + $s * $ys, $z );
11121 # glVertex3d( $cx - $s * $xs, $cy, $z );
11122 # glVertex3d( $cx, $cy - $s * $ys, $z );
11124 # To distinguish between these situations, we will
11125 # append the length of the line from the previous matching
11126 # token, or beginning of line, to the function name. This
11127 # will allow the vertical aligner to reject undesirable
11130 # if we are not aligning on this paren...
11131 if ( $matching_token_to_go[$i] eq '' ) {
11133 # Sum length from previous alignment, or start of line.
11134 # Note that we have to sum token lengths here because
11135 # padding has been done and so array $lengths_to_go
11139 join( '', @tokens_to_go[ $i_start .. $i - 1 ] ) );
11140 $len += leading_spaces_to_go($i_start)
11141 if ( $i_start == $ibeg );
11143 # tack length onto the container name to make unique
11144 $container_name[$depth] .= "-" . $len;
11148 elsif ( $tokens_to_go[$i] eq ')' ) {
11149 $depth-- if $depth > 0;
11152 # if we find a new synchronization token, we are done with
11154 if ( $i > $i_start && $matching_token_to_go[$i] ne '' ) {
11156 my $tok = my $raw_tok = $matching_token_to_go[$i];
11158 # make separators in different nesting depths unique
11159 # by appending the nesting depth digit.
11160 if ( $raw_tok ne '#' ) {
11161 $tok .= "$nesting_depth_to_go[$i]";
11164 # also decorate commas with any container name to avoid
11165 # unwanted cross-line alignments.
11166 if ( $raw_tok eq ',' || $raw_tok eq '=>' ) {
11167 if ( $container_name[$depth] ) {
11168 $tok .= $container_name[$depth];
11172 # Patch to avoid aligning leading and trailing if, unless.
11173 # Mark trailing if, unless statements with container names.
11174 # This makes them different from leading if, unless which
11175 # are not so marked at present. If we ever need to name
11176 # them too, we could use ci to distinguish them.
11177 # Example problem to avoid:
11178 # return ( 2, "DBERROR" )
11179 # if ( $retval == 2 );
11180 # if ( scalar @_ ) {
11181 # my ( $a, $b, $c, $d, $e, $f ) = @_;
11183 if ( $raw_tok eq '(' ) {
11184 my $ci = $ci_levels_to_go[$ibeg];
11185 if ( $container_name[$depth] =~ /^\+(if|unless)/
11188 $tok .= $container_name[$depth];
11192 # Decorate block braces with block types to avoid
11193 # unwanted alignments such as the following:
11194 # foreach ( @{$routput_array} ) { $fh->print($_) }
11195 # eval { $fh->close() };
11196 if ( $raw_tok eq '{' && $block_type_to_go[$i] ) {
11197 my $block_type = $block_type_to_go[$i];
11199 # map certain related block types to allow
11200 # else blocks to align
11201 $block_type = $block_type_map{$block_type}
11202 if ( defined( $block_type_map{$block_type} ) );
11204 # remove sub names to allow one-line sub braces to align
11205 # regardless of name
11206 if ( $block_type =~ /^sub / ) { $block_type = 'sub' }
11208 # allow all control-type blocks to align
11209 if ( $block_type =~ /^[A-Z]+$/ ) { $block_type = 'BEGIN' }
11211 $tok .= $block_type;
11214 # concatenate the text of the consecutive tokens to form
11217 join( '', @tokens_to_go[ $i_start .. $i - 1 ] ) );
11219 # store the alignment token for this field
11220 push( @tokens, $tok );
11222 # get ready for the next batch
11225 $patterns[$j] = "";
11228 # continue accumulating tokens
11229 # handle non-keywords..
11230 if ( $types_to_go[$i] ne 'k' ) {
11231 my $type = $types_to_go[$i];
11233 # Mark most things before arrows as a quote to
11234 # get them to line up. Testfile: mixed.pl.
11235 if ( ( $i < $iend - 1 ) && ( $type =~ /^[wnC]$/ ) ) {
11236 my $next_type = $types_to_go[ $i + 1 ];
11237 my $i_next_nonblank =
11238 ( ( $next_type eq 'b' ) ? $i + 2 : $i + 1 );
11240 if ( $types_to_go[$i_next_nonblank] eq '=>' ) {
11243 # Patch to ignore leading minus before words,
11244 # by changing pattern 'mQ' into just 'Q',
11245 # so that we can align things like this:
11246 # Button => "Print letter \"~$_\"",
11247 # -command => [ sub { print "$_[0]\n" }, $_ ],
11248 if ( $patterns[$j] eq 'm' ) { $patterns[$j] = "" }
11252 # patch to make numbers and quotes align
11253 if ( $type eq 'n' ) { $type = 'Q' }
11255 # patch to ignore any ! in patterns
11256 if ( $type eq '!' ) { $type = '' }
11258 $patterns[$j] .= $type;
11261 # for keywords we have to use the actual text
11264 my $tok = $tokens_to_go[$i];
11266 # but map certain keywords to a common string to allow
11268 $tok = $keyword_map{$tok}
11269 if ( defined( $keyword_map{$tok} ) );
11270 $patterns[$j] .= $tok;
11274 # done with this line .. join text of tokens to make the last field
11275 push( @fields, join( '', @tokens_to_go[ $i_start .. $iend ] ) );
11276 return ( \@tokens, \@fields, \@patterns );
11279 } # end make_alignment_patterns
11281 { # begin unmatched_indexes
11283 # closure to keep track of unbalanced containers.
11284 # arrays shared by the routines in this block:
11285 my @unmatched_opening_indexes_in_this_batch;
11286 my @unmatched_closing_indexes_in_this_batch;
11287 my %comma_arrow_count;
11289 sub is_unbalanced_batch {
11290 @unmatched_opening_indexes_in_this_batch +
11291 @unmatched_closing_indexes_in_this_batch;
11294 sub comma_arrow_count {
11296 return $comma_arrow_count{$seqno};
11299 sub match_opening_and_closing_tokens {
11301 # Match up indexes of opening and closing braces, etc, in this batch.
11302 # This has to be done after all tokens are stored because unstoring
11303 # of tokens would otherwise cause trouble.
11305 @unmatched_opening_indexes_in_this_batch = ();
11306 @unmatched_closing_indexes_in_this_batch = ();
11307 %comma_arrow_count = ();
11309 my ( $i, $i_mate, $token );
11310 foreach $i ( 0 .. $max_index_to_go ) {
11311 if ( $type_sequence_to_go[$i] ) {
11312 $token = $tokens_to_go[$i];
11313 if ( $token =~ /^[\(\[\{\?]$/ ) {
11314 push @unmatched_opening_indexes_in_this_batch, $i;
11316 elsif ( $token =~ /^[\)\]\}\:]$/ ) {
11318 $i_mate = pop @unmatched_opening_indexes_in_this_batch;
11319 if ( defined($i_mate) && $i_mate >= 0 ) {
11320 if ( $type_sequence_to_go[$i_mate] ==
11321 $type_sequence_to_go[$i] )
11323 $mate_index_to_go[$i] = $i_mate;
11324 $mate_index_to_go[$i_mate] = $i;
11327 push @unmatched_opening_indexes_in_this_batch,
11329 push @unmatched_closing_indexes_in_this_batch, $i;
11333 push @unmatched_closing_indexes_in_this_batch, $i;
11337 elsif ( $tokens_to_go[$i] eq '=>' ) {
11338 if (@unmatched_opening_indexes_in_this_batch) {
11339 my $j = $unmatched_opening_indexes_in_this_batch[-1];
11340 my $seqno = $type_sequence_to_go[$j];
11341 $comma_arrow_count{$seqno}++;
11347 sub save_opening_indentation {
11349 # This should be called after each batch of tokens is output. It
11350 # saves indentations of lines of all unmatched opening tokens.
11351 # These will be used by sub get_opening_indentation.
11353 my ( $ri_first, $ri_last, $rindentation_list ) = @_;
11355 # we no longer need indentations of any saved indentations which
11356 # are unmatched closing tokens in this batch, because we will
11357 # never encounter them again. So we can delete them to keep
11358 # the hash size down.
11359 foreach (@unmatched_closing_indexes_in_this_batch) {
11360 my $seqno = $type_sequence_to_go[$_];
11361 delete $saved_opening_indentation{$seqno};
11364 # we need to save indentations of any unmatched opening tokens
11365 # in this batch because we may need them in a subsequent batch.
11366 foreach (@unmatched_opening_indexes_in_this_batch) {
11367 my $seqno = $type_sequence_to_go[$_];
11368 $saved_opening_indentation{$seqno} = [
11369 lookup_opening_indentation(
11370 $_, $ri_first, $ri_last, $rindentation_list
11375 } # end unmatched_indexes
11377 sub get_opening_indentation {
11379 # get the indentation of the line which output the opening token
11380 # corresponding to a given closing token in the current output batch.
11383 # $i_closing - index in this line of a closing token ')' '}' or ']'
11385 # $ri_first - reference to list of the first index $i for each output
11386 # line in this batch
11387 # $ri_last - reference to list of the last index $i for each output line
11389 # $rindentation_list - reference to a list containing the indentation
11390 # used for each line.
11393 # -the indentation of the line which contained the opening token
11394 # which matches the token at index $i_opening
11395 # -and its offset (number of columns) from the start of the line
11397 my ( $i_closing, $ri_first, $ri_last, $rindentation_list ) = @_;
11399 # first, see if the opening token is in the current batch
11400 my $i_opening = $mate_index_to_go[$i_closing];
11401 my ( $indent, $offset, $is_leading, $exists );
11403 if ( $i_opening >= 0 ) {
11405 # it is..look up the indentation
11406 ( $indent, $offset, $is_leading ) =
11407 lookup_opening_indentation( $i_opening, $ri_first, $ri_last,
11408 $rindentation_list );
11411 # if not, it should have been stored in the hash by a previous batch
11413 my $seqno = $type_sequence_to_go[$i_closing];
11415 if ( $saved_opening_indentation{$seqno} ) {
11416 ( $indent, $offset, $is_leading ) =
11417 @{ $saved_opening_indentation{$seqno} };
11420 # some kind of serious error
11421 # (example is badfile.t)
11430 # if no sequence number it must be an unbalanced container
11438 return ( $indent, $offset, $is_leading, $exists );
11441 sub lookup_opening_indentation {
11443 # get the indentation of the line in the current output batch
11444 # which output a selected opening token
11447 # $i_opening - index of an opening token in the current output batch
11448 # whose line indentation we need
11449 # $ri_first - reference to list of the first index $i for each output
11450 # line in this batch
11451 # $ri_last - reference to list of the last index $i for each output line
11453 # $rindentation_list - reference to a list containing the indentation
11454 # used for each line. (NOTE: the first slot in
11455 # this list is the last returned line number, and this is
11456 # followed by the list of indentations).
11459 # -the indentation of the line which contained token $i_opening
11460 # -and its offset (number of columns) from the start of the line
11462 my ( $i_opening, $ri_start, $ri_last, $rindentation_list ) = @_;
11464 my $nline = $rindentation_list->[0]; # line number of previous lookup
11466 # reset line location if necessary
11467 $nline = 0 if ( $i_opening < $ri_start->[$nline] );
11469 # find the correct line
11470 unless ( $i_opening > $ri_last->[-1] ) {
11471 while ( $i_opening > $ri_last->[$nline] ) { $nline++; }
11474 # error - token index is out of bounds - shouldn't happen
11477 "non-fatal program bug in lookup_opening_indentation - index out of range\n"
11479 report_definite_bug();
11480 $nline = $#{$ri_last};
11483 $rindentation_list->[0] =
11484 $nline; # save line number to start looking next call
11485 my $ibeg = $ri_start->[$nline];
11486 my $offset = token_sequence_length( $ibeg, $i_opening ) - 1;
11487 my $is_leading = ( $ibeg == $i_opening );
11488 return ( $rindentation_list->[ $nline + 1 ], $offset, $is_leading );
11492 my %is_if_elsif_else_unless_while_until_for_foreach;
11496 # These block types may have text between the keyword and opening
11497 # curly. Note: 'else' does not, but must be included to allow trailing
11498 # if/elsif text to be appended.
11499 # patch for SWITCH/CASE: added 'case' and 'when'
11500 @_ = qw(if elsif else unless while until for foreach case when);
11501 @is_if_elsif_else_unless_while_until_for_foreach{@_} = (1) x scalar(@_);
11504 sub set_adjusted_indentation {
11506 # This routine has the final say regarding the actual indentation of
11507 # a line. It starts with the basic indentation which has been
11508 # defined for the leading token, and then takes into account any
11509 # options that the user has set regarding special indenting and
11512 my ( $ibeg, $iend, $rfields, $rpatterns, $ri_first, $ri_last,
11513 $rindentation_list )
11516 # we need to know the last token of this line
11517 my ( $terminal_type, $i_terminal ) =
11518 terminal_type( \@types_to_go, \@block_type_to_go, $ibeg, $iend );
11520 my $is_outdented_line = 0;
11522 my $is_semicolon_terminated = $terminal_type eq ';'
11523 && $nesting_depth_to_go[$iend] < $nesting_depth_to_go[$ibeg];
11525 ##########################################################
11526 # Section 1: set a flag and a default indentation
11528 # Most lines are indented according to the initial token.
11529 # But it is common to outdent to the level just after the
11530 # terminal token in certain cases...
11531 # adjust_indentation flag:
11532 # 0 - do not adjust
11534 # 2 - vertically align with opening token
11536 ##########################################################
11537 my $adjust_indentation = 0;
11538 my $default_adjust_indentation = $adjust_indentation;
11541 $opening_indentation, $opening_offset,
11542 $is_leading, $opening_exists
11545 # if we are at a closing token of some type..
11546 if ( $types_to_go[$ibeg] =~ /^[\)\}\]]$/ ) {
11548 # get the indentation of the line containing the corresponding
11551 $opening_indentation, $opening_offset,
11552 $is_leading, $opening_exists
11554 = get_opening_indentation( $ibeg, $ri_first, $ri_last,
11555 $rindentation_list );
11557 # First set the default behavior:
11558 # default behavior is to outdent closing lines
11559 # of the form: "); }; ]; )->xxx;"
11561 $is_semicolon_terminated
11563 # and 'cuddled parens' of the form: ")->pack("
11565 $terminal_type eq '('
11566 && $types_to_go[$ibeg] eq ')'
11567 && ( $nesting_depth_to_go[$iend] + 1 ==
11568 $nesting_depth_to_go[$ibeg] )
11572 $adjust_indentation = 1;
11575 # TESTING: outdent something like '),'
11577 $terminal_type eq ','
11579 # allow just one character before the comma
11580 && $i_terminal == $ibeg + 1
11582 # requre LIST environment; otherwise, we may outdent too much --
11583 # this can happen in calls without parentheses (overload.t);
11584 && $container_environment_to_go[$i_terminal] eq 'LIST'
11587 $adjust_indentation = 1;
11590 # undo continuation indentation of a terminal closing token if
11591 # it is the last token before a level decrease. This will allow
11592 # a closing token to line up with its opening counterpart, and
11593 # avoids a indentation jump larger than 1 level.
11594 if ( $types_to_go[$i_terminal] =~ /^[\}\]\)R]$/
11595 && $i_terminal == $ibeg )
11597 my $ci = $ci_levels_to_go[$ibeg];
11598 my $lev = $levels_to_go[$ibeg];
11599 my $next_type = $types_to_go[ $ibeg + 1 ];
11600 my $i_next_nonblank =
11601 ( ( $next_type eq 'b' ) ? $ibeg + 2 : $ibeg + 1 );
11602 if ( $i_next_nonblank <= $max_index_to_go
11603 && $levels_to_go[$i_next_nonblank] < $lev )
11605 $adjust_indentation = 1;
11609 $default_adjust_indentation = $adjust_indentation;
11611 # Now modify default behavior according to user request:
11612 # handle option to indent non-blocks of the form ); }; ];
11613 # But don't do special indentation to something like ')->pack('
11614 if ( !$block_type_to_go[$ibeg] ) {
11615 my $cti = $closing_token_indentation{ $tokens_to_go[$ibeg] };
11617 if ( $i_terminal <= $ibeg + 1
11618 || $is_semicolon_terminated )
11620 $adjust_indentation = 2;
11623 $adjust_indentation = 0;
11626 elsif ( $cti == 2 ) {
11627 if ($is_semicolon_terminated) {
11628 $adjust_indentation = 3;
11631 $adjust_indentation = 0;
11634 elsif ( $cti == 3 ) {
11635 $adjust_indentation = 3;
11639 # handle option to indent blocks
11642 $rOpts->{'indent-closing-brace'}
11644 $i_terminal == $ibeg # isolated terminal '}'
11645 || $is_semicolon_terminated
11649 $adjust_indentation = 3;
11654 # if at ');', '};', '>;', and '];' of a terminal qw quote
11655 elsif ($$rpatterns[0] =~ /^qb*;$/
11656 && $$rfields[0] =~ /^([\)\}\]\>]);$/ )
11658 if ( $closing_token_indentation{$1} == 0 ) {
11659 $adjust_indentation = 1;
11662 $adjust_indentation = 3;
11666 # if line begins with a ':', align it with any
11667 # previous line leading with corresponding ?
11668 elsif ( $types_to_go[$ibeg] eq ':' ) {
11670 $opening_indentation, $opening_offset,
11671 $is_leading, $opening_exists
11673 = get_opening_indentation( $ibeg, $ri_first, $ri_last,
11674 $rindentation_list );
11675 if ($is_leading) { $adjust_indentation = 2; }
11678 ##########################################################
11679 # Section 2: set indentation according to flag set above
11681 # Select the indentation object to define leading
11682 # whitespace. If we are outdenting something like '} } );'
11683 # then we want to use one level below the last token
11684 # ($i_terminal) in order to get it to fully outdent through
11686 ##########################################################
11689 my $level_end = $levels_to_go[$iend];
11691 if ( $adjust_indentation == 0 ) {
11692 $indentation = $leading_spaces_to_go[$ibeg];
11693 $lev = $levels_to_go[$ibeg];
11695 elsif ( $adjust_indentation == 1 ) {
11696 $indentation = $reduced_spaces_to_go[$i_terminal];
11697 $lev = $levels_to_go[$i_terminal];
11700 # handle indented closing token which aligns with opening token
11701 elsif ( $adjust_indentation == 2 ) {
11703 # handle option to align closing token with opening token
11704 $lev = $levels_to_go[$ibeg];
11706 # calculate spaces needed to align with opening token
11708 get_SPACES($opening_indentation) + $opening_offset;
11710 # Indent less than the previous line.
11712 # Problem: For -lp we don't exactly know what it was if there
11713 # were recoverable spaces sent to the aligner. A good solution
11714 # would be to force a flush of the vertical alignment buffer, so
11715 # that we would know. For now, this rule is used for -lp:
11717 # When the last line did not start with a closing token we will
11718 # be optimistic that the aligner will recover everything wanted.
11720 # This rule will prevent us from breaking a hierarchy of closing
11721 # tokens, and in a worst case will leave a closing paren too far
11722 # indented, but this is better than frequently leaving it not
11724 my $last_spaces = get_SPACES($last_indentation_written);
11725 if ( $last_leading_token !~ /^[\}\]\)]$/ ) {
11727 get_RECOVERABLE_SPACES($last_indentation_written);
11730 # reset the indentation to the new space count if it works
11731 # only options are all or none: nothing in-between looks good
11732 $lev = $levels_to_go[$ibeg];
11733 if ( $space_count < $last_spaces ) {
11734 if ($rOpts_line_up_parentheses) {
11735 my $lev = $levels_to_go[$ibeg];
11737 new_lp_indentation_item( $space_count, $lev, 0, 0, 0 );
11740 $indentation = $space_count;
11744 # revert to default if it doesnt work
11746 $space_count = leading_spaces_to_go($ibeg);
11747 if ( $default_adjust_indentation == 0 ) {
11748 $indentation = $leading_spaces_to_go[$ibeg];
11750 elsif ( $default_adjust_indentation == 1 ) {
11751 $indentation = $reduced_spaces_to_go[$i_terminal];
11752 $lev = $levels_to_go[$i_terminal];
11757 # Full indentaion of closing tokens (-icb and -icp or -cti=2)
11760 # handle -icb (indented closing code block braces)
11761 # Updated method for indented block braces: indent one full level if
11762 # there is no continuation indentation. This will occur for major
11763 # structures such as sub, if, else, but not for things like map
11766 # Note: only code blocks without continuation indentation are
11767 # handled here (if, else, unless, ..). In the following snippet,
11768 # the terminal brace of the sort block will have continuation
11769 # indentation as shown so it will not be handled by the coding
11770 # here. We would have to undo the continuation indentation to do
11771 # this, but it probably looks ok as is. This is a possible future
11772 # update for semicolon terminated lines.
11774 # if ($sortby eq 'date' or $sortby eq 'size') {
11776 # $file_data{$a}{$sortby} <=> $file_data{$b}{$sortby}
11781 if ( $block_type_to_go[$ibeg]
11782 && $ci_levels_to_go[$i_terminal] == 0 )
11784 my $spaces = get_SPACES( $leading_spaces_to_go[$i_terminal] );
11785 $indentation = $spaces + $rOpts_indent_columns;
11787 # NOTE: for -lp we could create a new indentation object, but
11788 # there is probably no need to do it
11791 # handle -icp and any -icb block braces which fall through above
11792 # test such as the 'sort' block mentioned above.
11795 # There are currently two ways to handle -icp...
11796 # One way is to use the indentation of the previous line:
11797 # $indentation = $last_indentation_written;
11799 # The other way is to use the indentation that the previous line
11800 # would have had if it hadn't been adjusted:
11801 $indentation = $last_unadjusted_indentation;
11803 # Current method: use the minimum of the two. This avoids
11804 # inconsistent indentation.
11805 if ( get_SPACES($last_indentation_written) <
11806 get_SPACES($indentation) )
11808 $indentation = $last_indentation_written;
11812 # use previous indentation but use own level
11813 # to cause list to be flushed properly
11814 $lev = $levels_to_go[$ibeg];
11817 # remember indentation except for multi-line quotes, which get
11819 unless ( $ibeg == 0 && $starting_in_quote ) {
11820 $last_indentation_written = $indentation;
11821 $last_unadjusted_indentation = $leading_spaces_to_go[$ibeg];
11822 $last_leading_token = $tokens_to_go[$ibeg];
11825 # be sure lines with leading closing tokens are not outdented more
11826 # than the line which contained the corresponding opening token.
11828 #############################################################
11829 # updated per bug report in alex_bug.pl: we must not
11830 # mess with the indentation of closing logical braces so
11831 # we must treat something like '} else {' as if it were
11832 # an isolated brace my $is_isolated_block_brace = (
11833 # $iend == $ibeg ) && $block_type_to_go[$ibeg];
11834 #############################################################
11835 my $is_isolated_block_brace = $block_type_to_go[$ibeg]
11836 && ( $iend == $ibeg
11837 || $is_if_elsif_else_unless_while_until_for_foreach{
11838 $block_type_to_go[$ibeg] } );
11840 # only do this for a ':; which is aligned with its leading '?'
11841 my $is_unaligned_colon = $types_to_go[$ibeg] eq ':' && !$is_leading;
11842 if ( defined($opening_indentation)
11843 && !$is_isolated_block_brace
11844 && !$is_unaligned_colon )
11846 if ( get_SPACES($opening_indentation) > get_SPACES($indentation) ) {
11847 $indentation = $opening_indentation;
11851 # remember the indentation of each line of this batch
11852 push @{$rindentation_list}, $indentation;
11854 # outdent lines with certain leading tokens...
11857 # must be first word of this batch
11863 # certain leading keywords if requested
11865 $rOpts->{'outdent-keywords'}
11866 && $types_to_go[$ibeg] eq 'k'
11867 && $outdent_keyword{ $tokens_to_go[$ibeg] }
11870 # or labels if requested
11871 || ( $rOpts->{'outdent-labels'} && $types_to_go[$ibeg] eq 'J' )
11873 # or static block comments if requested
11874 || ( $types_to_go[$ibeg] eq '#'
11875 && $rOpts->{'outdent-static-block-comments'}
11876 && $is_static_block_comment )
11881 my $space_count = leading_spaces_to_go($ibeg);
11882 if ( $space_count > 0 ) {
11883 $space_count -= $rOpts_continuation_indentation;
11884 $is_outdented_line = 1;
11885 if ( $space_count < 0 ) { $space_count = 0 }
11887 # do not promote a spaced static block comment to non-spaced;
11888 # this is not normally necessary but could be for some
11889 # unusual user inputs (such as -ci = -i)
11890 if ( $types_to_go[$ibeg] eq '#' && $space_count == 0 ) {
11894 if ($rOpts_line_up_parentheses) {
11896 new_lp_indentation_item( $space_count, $lev, 0, 0, 0 );
11899 $indentation = $space_count;
11904 return ( $indentation, $lev, $level_end, $terminal_type,
11905 $is_semicolon_terminated, $is_outdented_line );
11909 sub set_vertical_tightness_flags {
11911 my ( $n, $n_last_line, $ibeg, $iend, $ri_first, $ri_last ) = @_;
11913 # Define vertical tightness controls for the nth line of a batch.
11914 # We create an array of parameters which tell the vertical aligner
11915 # if we should combine this line with the next line to achieve the
11916 # desired vertical tightness. The array of parameters contains:
11918 # [0] type: 1=is opening tok 2=is closing tok 3=is opening block brace
11919 # [1] flag: if opening: 1=no multiple steps, 2=multiple steps ok
11920 # if closing: spaces of padding to use
11921 # [2] sequence number of container
11922 # [3] valid flag: do not append if this flag is false. Will be
11923 # true if appropriate -vt flag is set. Otherwise, Will be
11924 # made true only for 2 line container in parens with -lp
11926 # These flags are used by sub set_leading_whitespace in
11927 # the vertical aligner
11929 my $rvertical_tightness_flags = [ 0, 0, 0, 0, 0, 0 ];
11931 # For non-BLOCK tokens, we will need to examine the next line
11932 # too, so we won't consider the last line.
11933 if ( $n < $n_last_line ) {
11935 # see if last token is an opening token...not a BLOCK...
11936 my $ibeg_next = $$ri_first[ $n + 1 ];
11937 my $token_end = $tokens_to_go[$iend];
11938 my $iend_next = $$ri_last[ $n + 1 ];
11940 $type_sequence_to_go[$iend]
11941 && !$block_type_to_go[$iend]
11942 && $is_opening_token{$token_end}
11944 $opening_vertical_tightness{$token_end} > 0
11946 # allow 2-line method call to be closed up
11947 || ( $rOpts_line_up_parentheses
11948 && $token_end eq '('
11950 && $types_to_go[ $iend - 1 ] ne 'b' )
11955 # avoid multiple jumps in nesting depth in one line if
11957 my $ovt = $opening_vertical_tightness{$token_end};
11958 my $iend_next = $$ri_last[ $n + 1 ];
11961 && ( $nesting_depth_to_go[ $iend_next + 1 ] !=
11962 $nesting_depth_to_go[$ibeg_next] )
11966 # If -vt flag has not been set, mark this as invalid
11967 # and aligner will validate it if it sees the closing paren
11969 my $valid_flag = $ovt;
11970 @{$rvertical_tightness_flags} =
11971 ( 1, $ovt, $type_sequence_to_go[$iend], $valid_flag );
11975 # see if first token of next line is a closing token...
11976 # ..and be sure this line does not have a side comment
11977 my $token_next = $tokens_to_go[$ibeg_next];
11978 if ( $type_sequence_to_go[$ibeg_next]
11979 && !$block_type_to_go[$ibeg_next]
11980 && $is_closing_token{$token_next}
11981 && $types_to_go[$iend] !~ '#' ) # for safety, shouldn't happen!
11983 my $ovt = $opening_vertical_tightness{$token_next};
11984 my $cvt = $closing_vertical_tightness{$token_next};
11987 # never append a trailing line like )->pack(
11988 # because it will throw off later alignment
11990 $nesting_depth_to_go[$ibeg_next] ==
11991 $nesting_depth_to_go[ $iend_next + 1 ] + 1
11996 $container_environment_to_go[$ibeg_next] ne 'LIST'
12000 # allow closing up 2-line method calls
12001 || ( $rOpts_line_up_parentheses
12002 && $token_next eq ')' )
12009 # decide which trailing closing tokens to append..
12011 if ( $cvt == 2 || $iend_next == $ibeg_next ) { $ok = 1 }
12013 my $str = join( '',
12014 @types_to_go[ $ibeg_next + 1 .. $ibeg_next + 2 ] );
12016 # append closing token if followed by comment or ';'
12017 if ( $str =~ /^b?[#;]/ ) { $ok = 1 }
12021 my $valid_flag = $cvt;
12022 @{$rvertical_tightness_flags} = (
12024 $tightness{$token_next} == 2 ? 0 : 1,
12025 $type_sequence_to_go[$ibeg_next], $valid_flag,
12031 # Opening Token Right
12032 # If requested, move an isolated trailing opening token to the end of
12033 # the previous line which ended in a comma. We could do this
12034 # in sub recombine_breakpoints but that would cause problems
12035 # with -lp formatting. The problem is that indentation will
12036 # quickly move far to the right in nested expressions. By
12037 # doing it after indentation has been set, we avoid changes
12038 # to the indentation. Actual movement of the token takes place
12039 # in sub write_leader_and_string.
12041 $opening_token_right{ $tokens_to_go[$ibeg_next] }
12043 # previous line is not opening
12044 # (use -sot to combine with it)
12045 && !$is_opening_token{$token_end}
12047 # previous line ended in one of these
12048 # (add other cases if necessary; '=>' and '.' are not necessary
12049 ##&& ($is_opening_token{$token_end} || $token_end eq ',')
12050 && !$block_type_to_go[$ibeg_next]
12052 # this is a line with just an opening token
12053 && ( $iend_next == $ibeg_next
12054 || $iend_next == $ibeg_next + 2
12055 && $types_to_go[$iend_next] eq '#' )
12057 # looks bad if we align vertically with the wrong container
12058 && $tokens_to_go[$ibeg] ne $tokens_to_go[$ibeg_next]
12061 my $valid_flag = 1;
12062 my $spaces = ( $types_to_go[ $ibeg_next - 1 ] eq 'b' ) ? 1 : 0;
12063 @{$rvertical_tightness_flags} =
12064 ( 2, $spaces, $type_sequence_to_go[$ibeg_next], $valid_flag, );
12067 # Stacking of opening and closing tokens
12069 my $token_beg_next = $tokens_to_go[$ibeg_next];
12071 # patch to make something like 'qw(' behave like an opening paren
12073 if ( $types_to_go[$ibeg_next] eq 'q' ) {
12074 if ( $token_beg_next =~ /^qw\s*([\[\(\{])$/ ) {
12075 $token_beg_next = $1;
12079 if ( $is_closing_token{$token_end}
12080 && $is_closing_token{$token_beg_next} )
12082 $stackable = $stack_closing_token{$token_beg_next}
12083 unless ( $block_type_to_go[$ibeg_next] )
12084 ; # shouldn't happen; just checking
12086 elsif ($is_opening_token{$token_end}
12087 && $is_opening_token{$token_beg_next} )
12089 $stackable = $stack_opening_token{$token_beg_next}
12090 unless ( $block_type_to_go[$ibeg_next] )
12091 ; # shouldn't happen; just checking
12096 my $is_semicolon_terminated;
12097 if ( $n + 1 == $n_last_line ) {
12098 my ( $terminal_type, $i_terminal ) = terminal_type(
12099 \@types_to_go, \@block_type_to_go,
12100 $ibeg_next, $iend_next
12102 $is_semicolon_terminated = $terminal_type eq ';'
12103 && $nesting_depth_to_go[$iend_next] <
12104 $nesting_depth_to_go[$ibeg_next];
12107 # this must be a line with just an opening token
12108 # or end in a semicolon
12110 $is_semicolon_terminated
12111 || ( $iend_next == $ibeg_next
12112 || $iend_next == $ibeg_next + 2
12113 && $types_to_go[$iend_next] eq '#' )
12116 my $valid_flag = 1;
12117 my $spaces = ( $types_to_go[ $ibeg_next - 1 ] eq 'b' ) ? 1 : 0;
12118 @{$rvertical_tightness_flags} =
12119 ( 2, $spaces, $type_sequence_to_go[$ibeg_next], $valid_flag,
12125 # Check for a last line with isolated opening BLOCK curly
12126 elsif ($rOpts_block_brace_vertical_tightness
12128 && $types_to_go[$iend] eq '{'
12129 && $block_type_to_go[$iend] =~
12130 /$block_brace_vertical_tightness_pattern/o )
12132 @{$rvertical_tightness_flags} =
12133 ( 3, $rOpts_block_brace_vertical_tightness, 0, 1 );
12136 # pack in the sequence numbers of the ends of this line
12137 $rvertical_tightness_flags->[4] = get_seqno($ibeg);
12138 $rvertical_tightness_flags->[5] = get_seqno($iend);
12139 return $rvertical_tightness_flags;
12144 # get opening and closing sequence numbers of a token for the vertical
12145 # aligner. Assign qw quotes a value to allow qw opening and closing tokens
12146 # to be treated somewhat like opening and closing tokens for stacking
12147 # tokens by the vertical aligner.
12149 my $seqno = $type_sequence_to_go[$ii];
12150 if ( $types_to_go[$ii] eq 'q' ) {
12153 $seqno = $SEQ_QW if ( $tokens_to_go[$ii] =~ /^qw\s*[\(\{\[]/ );
12156 if ( !$ending_in_quote ) {
12157 $seqno = $SEQ_QW if ( $tokens_to_go[$ii] =~ /[\)\}\]]$/ );
12165 my %is_vertical_alignment_type;
12166 my %is_vertical_alignment_keyword;
12171 = **= += *= &= <<= &&= -= /= |= >>= ||= //= .= %= ^= x=
12172 { ? : => =~ && || // ~~ !~~
12174 @is_vertical_alignment_type{@_} = (1) x scalar(@_);
12176 @_ = qw(if unless and or err eq ne for foreach while until);
12177 @is_vertical_alignment_keyword{@_} = (1) x scalar(@_);
12180 sub set_vertical_alignment_markers {
12182 # This routine takes the first step toward vertical alignment of the
12183 # lines of output text. It looks for certain tokens which can serve as
12184 # vertical alignment markers (such as an '=').
12186 # Method: We look at each token $i in this output batch and set
12187 # $matching_token_to_go[$i] equal to those tokens at which we would
12188 # accept vertical alignment.
12190 # nothing to do if we aren't allowed to change whitespace
12191 if ( !$rOpts_add_whitespace ) {
12192 for my $i ( 0 .. $max_index_to_go ) {
12193 $matching_token_to_go[$i] = '';
12198 my ( $ri_first, $ri_last ) = @_;
12200 # remember the index of last nonblank token before any sidecomment
12201 my $i_terminal = $max_index_to_go;
12202 if ( $types_to_go[$i_terminal] eq '#' ) {
12203 if ( $i_terminal > 0 && $types_to_go[ --$i_terminal ] eq 'b' ) {
12204 if ( $i_terminal > 0 ) { --$i_terminal }
12208 # look at each line of this batch..
12209 my $last_vertical_alignment_before_index;
12210 my $vert_last_nonblank_type;
12211 my $vert_last_nonblank_token;
12212 my $vert_last_nonblank_block_type;
12213 my $max_line = @$ri_first - 1;
12214 my ( $i, $type, $token, $block_type, $alignment_type );
12215 my ( $ibeg, $iend, $line );
12217 foreach $line ( 0 .. $max_line ) {
12218 $ibeg = $$ri_first[$line];
12219 $iend = $$ri_last[$line];
12220 $last_vertical_alignment_before_index = -1;
12221 $vert_last_nonblank_type = '';
12222 $vert_last_nonblank_token = '';
12223 $vert_last_nonblank_block_type = '';
12225 # look at each token in this output line..
12226 foreach $i ( $ibeg .. $iend ) {
12227 $alignment_type = '';
12228 $type = $types_to_go[$i];
12229 $block_type = $block_type_to_go[$i];
12230 $token = $tokens_to_go[$i];
12232 # check for flag indicating that we should not align
12234 if ( $matching_token_to_go[$i] ) {
12235 $matching_token_to_go[$i] = '';
12239 #--------------------------------------------------------
12240 # First see if we want to align BEFORE this token
12241 #--------------------------------------------------------
12243 # The first possible token that we can align before
12244 # is index 2 because: 1) it doesn't normally make sense to
12245 # align before the first token and 2) the second
12246 # token must be a blank if we are to align before
12248 if ( $i < $ibeg + 2 ) { }
12250 # must follow a blank token
12251 elsif ( $types_to_go[ $i - 1 ] ne 'b' ) { }
12253 # align a side comment --
12254 elsif ( $type eq '#' ) {
12258 # it is a static side comment
12260 $rOpts->{'static-side-comments'}
12261 && $token =~ /$static_side_comment_pattern/o
12264 # or a closing side comment
12265 || ( $vert_last_nonblank_block_type
12267 /$closing_side_comment_prefix_pattern/o )
12270 $alignment_type = $type;
12271 } ## Example of a static side comment
12274 # otherwise, do not align two in a row to create a
12276 elsif ( $last_vertical_alignment_before_index == $i - 2 ) { }
12278 # align before one of these keywords
12279 # (within a line, since $i>1)
12280 elsif ( $type eq 'k' ) {
12282 # /^(if|unless|and|or|eq|ne)$/
12283 if ( $is_vertical_alignment_keyword{$token} ) {
12284 $alignment_type = $token;
12288 # align before one of these types..
12289 # Note: add '.' after new vertical aligner is operational
12290 elsif ( $is_vertical_alignment_type{$type} ) {
12291 $alignment_type = $token;
12293 # Do not align a terminal token. Although it might
12294 # occasionally look ok to do this, it has been found to be
12295 # a good general rule. The main problems are:
12296 # (1) that the terminal token (such as an = or :) might get
12297 # moved far to the right where it is hard to see because
12298 # nothing follows it, and
12299 # (2) doing so may prevent other good alignments.
12300 if ( $i == $iend || $i >= $i_terminal ) {
12301 $alignment_type = "";
12304 # Do not align leading ': (' or '. ('. This would prevent
12305 # alignment in something like the following:
12307 # ( $input_line_number < 10 ) ? " "
12308 # : ( $input_line_number < 100 ) ? " "
12312 # ( $case_matters ? $accessor : " lc($accessor) " )
12313 # . ( $yesno ? " eq " : " ne " )
12314 if ( $i == $ibeg + 2
12315 && $types_to_go[$ibeg] =~ /^[\.\:]$/
12316 && $types_to_go[ $i - 1 ] eq 'b' )
12318 $alignment_type = "";
12321 # For a paren after keyword, only align something like this:
12323 # elsif ( $b ) { &b }
12324 if ( $token eq '(' && $vert_last_nonblank_type eq 'k' ) {
12325 $alignment_type = ""
12326 unless $vert_last_nonblank_token =~
12327 /^(if|unless|elsif)$/;
12330 # be sure the alignment tokens are unique
12331 # This didn't work well: reason not determined
12332 # if ($token ne $type) {$alignment_type .= $type}
12335 # NOTE: This is deactivated because it causes the previous
12336 # if/elsif alignment to fail
12337 #elsif ( $type eq '}' && $token eq '}' && $block_type_to_go[$i])
12338 #{ $alignment_type = $type; }
12340 if ($alignment_type) {
12341 $last_vertical_alignment_before_index = $i;
12344 #--------------------------------------------------------
12345 # Next see if we want to align AFTER the previous nonblank
12346 #--------------------------------------------------------
12348 # We want to line up ',' and interior ';' tokens, with the added
12349 # space AFTER these tokens. (Note: interior ';' is included
12350 # because it may occur in short blocks).
12353 # we haven't already set it
12356 # and its not the first token of the line
12359 # and it follows a blank
12360 && $types_to_go[ $i - 1 ] eq 'b'
12362 # and previous token IS one of these:
12363 && ( $vert_last_nonblank_type =~ /^[\,\;]$/ )
12365 # and it's NOT one of these
12366 && ( $type !~ /^[b\#\)\]\}]$/ )
12368 # then go ahead and align
12372 $alignment_type = $vert_last_nonblank_type;
12375 #--------------------------------------------------------
12376 # then store the value
12377 #--------------------------------------------------------
12378 $matching_token_to_go[$i] = $alignment_type;
12379 if ( $type ne 'b' ) {
12380 $vert_last_nonblank_type = $type;
12381 $vert_last_nonblank_token = $token;
12382 $vert_last_nonblank_block_type = $block_type;
12389 sub terminal_type {
12391 # returns type of last token on this line (terminal token), as follows:
12392 # returns # for a full-line comment
12393 # returns ' ' for a blank line
12394 # otherwise returns final token type
12396 my ( $rtype, $rblock_type, $ibeg, $iend ) = @_;
12398 # check for full-line comment..
12399 if ( $$rtype[$ibeg] eq '#' ) {
12400 return wantarray ? ( $$rtype[$ibeg], $ibeg ) : $$rtype[$ibeg];
12404 # start at end and walk bakwards..
12405 for ( my $i = $iend ; $i >= $ibeg ; $i-- ) {
12407 # skip past any side comment and blanks
12408 next if ( $$rtype[$i] eq 'b' );
12409 next if ( $$rtype[$i] eq '#' );
12411 # found it..make sure it is a BLOCK termination,
12412 # but hide a terminal } after sort/grep/map because it is not
12413 # necessarily the end of the line. (terminal.t)
12414 my $terminal_type = $$rtype[$i];
12416 $terminal_type eq '}'
12417 && ( !$$rblock_type[$i]
12418 || ( $is_sort_map_grep_eval_do{ $$rblock_type[$i] } ) )
12421 $terminal_type = 'b';
12423 return wantarray ? ( $terminal_type, $i ) : $terminal_type;
12427 return wantarray ? ( ' ', $ibeg ) : ' ';
12432 my %is_good_keyword_breakpoint;
12433 my %is_lt_gt_le_ge;
12435 sub set_bond_strengths {
12439 @_ = qw(if unless while until for foreach);
12440 @is_good_keyword_breakpoint{@_} = (1) x scalar(@_);
12442 @_ = qw(lt gt le ge);
12443 @is_lt_gt_le_ge{@_} = (1) x scalar(@_);
12445 ###############################################################
12446 # NOTE: NO_BREAK's set here are HINTS which may not be honored;
12447 # essential NO_BREAKS's must be enforced in section 2, below.
12448 ###############################################################
12450 # adding NEW_TOKENS: add a left and right bond strength by
12451 # mimmicking what is done for an existing token type. You
12452 # can skip this step at first and take the default, then
12453 # tweak later to get desired results.
12455 # The bond strengths should roughly follow precenence order where
12456 # possible. If you make changes, please check the results very
12457 # carefully on a variety of scripts.
12459 # no break around possible filehandle
12460 $left_bond_strength{'Z'} = NO_BREAK;
12461 $right_bond_strength{'Z'} = NO_BREAK;
12463 # never put a bare word on a new line:
12464 # example print (STDERR, "bla"); will fail with break after (
12465 $left_bond_strength{'w'} = NO_BREAK;
12467 # blanks always have infinite strength to force breaks after real tokens
12468 $right_bond_strength{'b'} = NO_BREAK;
12470 # try not to break on exponentation
12471 @_ = qw" ** .. ... <=> ";
12472 @left_bond_strength{@_} = (STRONG) x scalar(@_);
12473 @right_bond_strength{@_} = (STRONG) x scalar(@_);
12475 # The comma-arrow has very low precedence but not a good break point
12476 $left_bond_strength{'=>'} = NO_BREAK;
12477 $right_bond_strength{'=>'} = NOMINAL;
12479 # ok to break after label
12480 $left_bond_strength{'J'} = NO_BREAK;
12481 $right_bond_strength{'J'} = NOMINAL;
12482 $left_bond_strength{'j'} = STRONG;
12483 $right_bond_strength{'j'} = STRONG;
12484 $left_bond_strength{'A'} = STRONG;
12485 $right_bond_strength{'A'} = STRONG;
12487 $left_bond_strength{'->'} = STRONG;
12488 $right_bond_strength{'->'} = VERY_STRONG;
12490 # breaking AFTER modulus operator is ok:
12492 @left_bond_strength{@_} = (STRONG) x scalar(@_);
12493 @right_bond_strength{@_} =
12494 ( 0.1 * NOMINAL + 0.9 * STRONG ) x scalar(@_);
12496 # Break AFTER math operators * and /
12498 @left_bond_strength{@_} = (STRONG) x scalar(@_);
12499 @right_bond_strength{@_} = (NOMINAL) x scalar(@_);
12501 # Break AFTER weakest math operators + and -
12502 # Make them weaker than * but a bit stronger than '.'
12504 @left_bond_strength{@_} = (STRONG) x scalar(@_);
12505 @right_bond_strength{@_} =
12506 ( 0.91 * NOMINAL + 0.09 * WEAK ) x scalar(@_);
12508 # breaking BEFORE these is just ok:
12510 @right_bond_strength{@_} = (STRONG) x scalar(@_);
12511 @left_bond_strength{@_} = (NOMINAL) x scalar(@_);
12513 # breaking before the string concatenation operator seems best
12514 # because it can be hard to see at the end of a line
12515 $right_bond_strength{'.'} = STRONG;
12516 $left_bond_strength{'.'} = 0.9 * NOMINAL + 0.1 * WEAK;
12519 @left_bond_strength{@_} = (STRONG) x scalar(@_);
12520 @right_bond_strength{@_} = (NOMINAL) x scalar(@_);
12522 # make these a little weaker than nominal so that they get
12523 # favored for end-of-line characters
12524 @_ = qw"!= == =~ !~ ~~ !~~";
12525 @left_bond_strength{@_} = (STRONG) x scalar(@_);
12526 @right_bond_strength{@_} =
12527 ( 0.9 * NOMINAL + 0.1 * WEAK ) x scalar(@_);
12529 # break AFTER these
12530 @_ = qw" < > | & >= <=";
12531 @left_bond_strength{@_} = (VERY_STRONG) x scalar(@_);
12532 @right_bond_strength{@_} =
12533 ( 0.8 * NOMINAL + 0.2 * WEAK ) x scalar(@_);
12535 # breaking either before or after a quote is ok
12536 # but bias for breaking before a quote
12537 $left_bond_strength{'Q'} = NOMINAL;
12538 $right_bond_strength{'Q'} = NOMINAL + 0.02;
12539 $left_bond_strength{'q'} = NOMINAL;
12540 $right_bond_strength{'q'} = NOMINAL;
12542 # starting a line with a keyword is usually ok
12543 $left_bond_strength{'k'} = NOMINAL;
12545 # we usually want to bond a keyword strongly to what immediately
12546 # follows, rather than leaving it stranded at the end of a line
12547 $right_bond_strength{'k'} = STRONG;
12549 $left_bond_strength{'G'} = NOMINAL;
12550 $right_bond_strength{'G'} = STRONG;
12552 # it is good to break AFTER various assignment operators
12554 = **= += *= &= <<= &&=
12555 -= /= |= >>= ||= //=
12559 @left_bond_strength{@_} = (STRONG) x scalar(@_);
12560 @right_bond_strength{@_} =
12561 ( 0.4 * WEAK + 0.6 * VERY_WEAK ) x scalar(@_);
12563 # break BEFORE '&&' and '||' and '//'
12564 # set strength of '||' to same as '=' so that chains like
12565 # $a = $b || $c || $d will break before the first '||'
12566 $right_bond_strength{'||'} = NOMINAL;
12567 $left_bond_strength{'||'} = $right_bond_strength{'='};
12569 # same thing for '//'
12570 $right_bond_strength{'//'} = NOMINAL;
12571 $left_bond_strength{'//'} = $right_bond_strength{'='};
12573 # set strength of && a little higher than ||
12574 $right_bond_strength{'&&'} = NOMINAL;
12575 $left_bond_strength{'&&'} = $left_bond_strength{'||'} + 0.1;
12577 $left_bond_strength{';'} = VERY_STRONG;
12578 $right_bond_strength{';'} = VERY_WEAK;
12579 $left_bond_strength{'f'} = VERY_STRONG;
12581 # make right strength of for ';' a little less than '='
12582 # to make for contents break after the ';' to avoid this:
12583 # for ( $j = $number_of_fields - 1 ; $j < $item_count ; $j +=
12584 # $number_of_fields )
12585 # and make it weaker than ',' and 'and' too
12586 $right_bond_strength{'f'} = VERY_WEAK - 0.03;
12588 # The strengths of ?/: should be somewhere between
12589 # an '=' and a quote (NOMINAL),
12590 # make strength of ':' slightly less than '?' to help
12591 # break long chains of ? : after the colons
12592 $left_bond_strength{':'} = 0.4 * WEAK + 0.6 * NOMINAL;
12593 $right_bond_strength{':'} = NO_BREAK;
12594 $left_bond_strength{'?'} = $left_bond_strength{':'} + 0.01;
12595 $right_bond_strength{'?'} = NO_BREAK;
12597 $left_bond_strength{','} = VERY_STRONG;
12598 $right_bond_strength{','} = VERY_WEAK;
12600 # Set bond strengths of certain keywords
12601 # make 'or', 'err', 'and' slightly weaker than a ','
12602 $left_bond_strength{'and'} = VERY_WEAK - 0.01;
12603 $left_bond_strength{'or'} = VERY_WEAK - 0.02;
12604 $left_bond_strength{'err'} = VERY_WEAK - 0.02;
12605 $left_bond_strength{'xor'} = NOMINAL;
12606 $right_bond_strength{'and'} = NOMINAL;
12607 $right_bond_strength{'or'} = NOMINAL;
12608 $right_bond_strength{'err'} = NOMINAL;
12609 $right_bond_strength{'xor'} = STRONG;
12612 # patch-its always ok to break at end of line
12613 $nobreak_to_go[$max_index_to_go] = 0;
12615 # adding a small 'bias' to strengths is a simple way to make a line
12616 # break at the first of a sequence of identical terms. For example,
12617 # to force long string of conditional operators to break with
12618 # each line ending in a ':', we can add a small number to the bond
12619 # strength of each ':'
12620 my $colon_bias = 0;
12627 my $code_bias = -.01;
12631 my $last_nonblank_type = $type;
12632 my $last_nonblank_token = $token;
12633 my $delta_bias = 0.0001;
12634 my $list_str = $left_bond_strength{'?'};
12636 my ( $block_type, $i_next, $i_next_nonblank, $next_nonblank_token,
12637 $next_nonblank_type, $next_token, $next_type, $total_nesting_depth,
12640 # preliminary loop to compute bond strengths
12641 for ( my $i = 0 ; $i <= $max_index_to_go ; $i++ ) {
12642 $last_type = $type;
12643 if ( $type ne 'b' ) {
12644 $last_nonblank_type = $type;
12645 $last_nonblank_token = $token;
12647 $type = $types_to_go[$i];
12649 # strength on both sides of a blank is the same
12650 if ( $type eq 'b' && $last_type ne 'b' ) {
12651 $bond_strength_to_go[$i] = $bond_strength_to_go[ $i - 1 ];
12655 $token = $tokens_to_go[$i];
12656 $block_type = $block_type_to_go[$i];
12658 $next_type = $types_to_go[$i_next];
12659 $next_token = $tokens_to_go[$i_next];
12660 $total_nesting_depth = $nesting_depth_to_go[$i_next];
12661 $i_next_nonblank = ( ( $next_type eq 'b' ) ? $i + 2 : $i + 1 );
12662 $next_nonblank_type = $types_to_go[$i_next_nonblank];
12663 $next_nonblank_token = $tokens_to_go[$i_next_nonblank];
12665 # Some token chemistry... The decision about where to break a
12666 # line depends upon a "bond strength" between tokens. The LOWER
12667 # the bond strength, the MORE likely a break. The strength
12668 # values are based on trial-and-error, and need to be tweaked
12669 # occasionally to get desired results. Things to keep in mind
12671 # 1. relative strengths are important. small differences
12672 # in strengths can make big formatting differences.
12673 # 2. each indentation level adds one unit of bond strength
12674 # 3. a value of NO_BREAK makes an unbreakable bond
12675 # 4. a value of VERY_WEAK is the strength of a ','
12676 # 5. values below NOMINAL are considered ok break points
12677 # 6. values above NOMINAL are considered poor break points
12678 # We are computing the strength of the bond between the current
12679 # token and the NEXT token.
12680 my $bond_str = VERY_STRONG; # a default, high strength
12682 #---------------------------------------------------------------
12684 # use minimum of left and right bond strengths if defined;
12685 # digraphs and trigraphs like to break on their left
12686 #---------------------------------------------------------------
12687 my $bsr = $right_bond_strength{$type};
12689 if ( !defined($bsr) ) {
12691 if ( $is_digraph{$type} || $is_trigraph{$type} ) {
12695 $bsr = VERY_STRONG;
12699 # define right bond strengths of certain keywords
12700 if ( $type eq 'k' && defined( $right_bond_strength{$token} ) ) {
12701 $bsr = $right_bond_strength{$token};
12703 elsif ( $token eq 'ne' or $token eq 'eq' ) {
12706 my $bsl = $left_bond_strength{$next_nonblank_type};
12708 # set terminal bond strength to the nominal value
12709 # this will cause good preceding breaks to be retained
12710 if ( $i_next_nonblank > $max_index_to_go ) {
12714 if ( !defined($bsl) ) {
12716 if ( $is_digraph{$next_nonblank_type}
12717 || $is_trigraph{$next_nonblank_type} )
12722 $bsl = VERY_STRONG;
12726 # define right bond strengths of certain keywords
12727 if ( $next_nonblank_type eq 'k'
12728 && defined( $left_bond_strength{$next_nonblank_token} ) )
12730 $bsl = $left_bond_strength{$next_nonblank_token};
12732 elsif ($next_nonblank_token eq 'ne'
12733 or $next_nonblank_token eq 'eq' )
12737 elsif ( $is_lt_gt_le_ge{$next_nonblank_token} ) {
12738 $bsl = 0.9 * NOMINAL + 0.1 * STRONG;
12741 # Note: it might seem that we would want to keep a NO_BREAK if
12742 # either token has this value. This didn't work, because in an
12743 # arrow list, it prevents the comma from separating from the
12744 # following bare word (which is probably quoted by its arrow).
12745 # So necessary NO_BREAK's have to be handled as special cases
12746 # in the final section.
12747 $bond_str = ( $bsr < $bsl ) ? $bsr : $bsl;
12748 my $bond_str_1 = $bond_str;
12750 #---------------------------------------------------------------
12753 #---------------------------------------------------------------
12755 # allow long lines before final { in an if statement, as in:
12760 # Otherwise, the line before the { tends to be too short.
12761 if ( $type eq ')' ) {
12762 if ( $next_nonblank_type eq '{' ) {
12763 $bond_str = VERY_WEAK + 0.03;
12767 elsif ( $type eq '(' ) {
12768 if ( $next_nonblank_type eq '{' ) {
12769 $bond_str = NOMINAL;
12773 # break on something like '} (', but keep this stronger than a ','
12774 # example is in 'howe.pl'
12775 elsif ( $type eq 'R' or $type eq '}' ) {
12776 if ( $next_nonblank_type eq '(' ) {
12777 $bond_str = 0.8 * VERY_WEAK + 0.2 * WEAK;
12781 #-----------------------------------------------------------------
12782 # adjust bond strength bias
12783 #-----------------------------------------------------------------
12785 # TESTING: add any bias set by sub scan_list at old comma
12787 elsif ( $type eq ',' ) {
12788 $bond_str += $bond_strength_to_go[$i];
12791 elsif ( $type eq 'f' ) {
12792 $bond_str += $f_bias;
12793 $f_bias += $delta_bias;
12796 # in long ?: conditionals, bias toward just one set per line (colon.t)
12797 elsif ( $type eq ':' ) {
12798 if ( !$want_break_before{$type} ) {
12799 $bond_str += $colon_bias;
12800 $colon_bias += $delta_bias;
12804 if ( $next_nonblank_type eq ':'
12805 && $want_break_before{$next_nonblank_type} )
12807 $bond_str += $colon_bias;
12808 $colon_bias += $delta_bias;
12811 # if leading '.' is used, align all but 'short' quotes;
12812 # the idea is to not place something like "\n" on a single line.
12813 elsif ( $next_nonblank_type eq '.' ) {
12814 if ( $want_break_before{'.'} ) {
12816 $last_nonblank_type eq '.'
12819 $rOpts_short_concatenation_item_length )
12820 && ( $token !~ /^[\)\]\}]$/ )
12823 $dot_bias += $delta_bias;
12825 $bond_str += $dot_bias;
12828 elsif ($next_nonblank_type eq '&&'
12829 && $want_break_before{$next_nonblank_type} )
12831 $bond_str += $amp_bias;
12832 $amp_bias += $delta_bias;
12834 elsif ($next_nonblank_type eq '||'
12835 && $want_break_before{$next_nonblank_type} )
12837 $bond_str += $bar_bias;
12838 $bar_bias += $delta_bias;
12840 elsif ( $next_nonblank_type eq 'k' ) {
12842 if ( $next_nonblank_token eq 'and'
12843 && $want_break_before{$next_nonblank_token} )
12845 $bond_str += $and_bias;
12846 $and_bias += $delta_bias;
12848 elsif ($next_nonblank_token =~ /^(or|err)$/
12849 && $want_break_before{$next_nonblank_token} )
12851 $bond_str += $or_bias;
12852 $or_bias += $delta_bias;
12855 # FIXME: needs more testing
12856 elsif ( $is_keyword_returning_list{$next_nonblank_token} ) {
12857 $bond_str = $list_str if ( $bond_str > $list_str );
12859 elsif ( $token eq 'err'
12860 && !$want_break_before{$token} )
12862 $bond_str += $or_bias;
12863 $or_bias += $delta_bias;
12868 && !$want_break_before{$type} )
12870 $bond_str += $colon_bias;
12871 $colon_bias += $delta_bias;
12873 elsif ( $type eq '&&'
12874 && !$want_break_before{$type} )
12876 $bond_str += $amp_bias;
12877 $amp_bias += $delta_bias;
12879 elsif ( $type eq '||'
12880 && !$want_break_before{$type} )
12882 $bond_str += $bar_bias;
12883 $bar_bias += $delta_bias;
12885 elsif ( $type eq 'k' ) {
12887 if ( $token eq 'and'
12888 && !$want_break_before{$token} )
12890 $bond_str += $and_bias;
12891 $and_bias += $delta_bias;
12893 elsif ( $token eq 'or'
12894 && !$want_break_before{$token} )
12896 $bond_str += $or_bias;
12897 $or_bias += $delta_bias;
12901 # keep matrix and hash indices together
12902 # but make them a little below STRONG to allow breaking open
12903 # something like {'some-word'}{'some-very-long-word'} at the }{
12905 if ( ( $type eq ']' or $type eq 'R' )
12906 && ( $next_nonblank_type eq '[' or $next_nonblank_type eq 'L' )
12909 $bond_str = 0.9 * STRONG + 0.1 * NOMINAL;
12912 if ( $next_nonblank_token =~ /^->/ ) {
12914 # increase strength to the point where a break in the following
12915 # will be after the opening paren rather than at the arrow:
12917 if ( $type eq 'i' ) {
12918 $bond_str = 1.45 * STRONG;
12921 elsif ( $type =~ /^[\)\]\}R]$/ ) {
12922 $bond_str = 0.1 * STRONG + 0.9 * NOMINAL;
12925 # otherwise make strength before an '->' a little over a '+'
12927 if ( $bond_str <= NOMINAL ) {
12928 $bond_str = NOMINAL + 0.01;
12933 if ( $token eq ')' && $next_nonblank_token eq '[' ) {
12934 $bond_str = 0.2 * STRONG + 0.8 * NOMINAL;
12937 # map1.t -- correct for a quirk in perl
12939 && $next_nonblank_type eq 'i'
12940 && $last_nonblank_type eq 'k'
12941 && $is_sort_map_grep{$last_nonblank_token} )
12943 # /^(sort|map|grep)$/ )
12945 $bond_str = NO_BREAK;
12948 # extrude.t: do not break before paren at:
12950 if ( $last_nonblank_type eq 'F' && $next_nonblank_token eq '(' ) {
12951 $bond_str = NO_BREAK;
12954 # good to break after end of code blocks
12955 if ( $type eq '}' && $block_type ) {
12957 $bond_str = 0.5 * WEAK + 0.5 * VERY_WEAK + $code_bias;
12958 $code_bias += $delta_bias;
12961 if ( $type eq 'k' ) {
12963 # allow certain control keywords to stand out
12964 if ( $next_nonblank_type eq 'k'
12965 && $is_last_next_redo_return{$token} )
12967 $bond_str = 0.45 * WEAK + 0.55 * VERY_WEAK;
12970 # Don't break after keyword my. This is a quick fix for a
12971 # rare problem with perl. An example is this line from file
12973 # foreach my $question( Debian::DebConf::ConfigDb::gettree( $this->{'question'} ) )
12975 if ( $token eq 'my' ) {
12976 $bond_str = NO_BREAK;
12981 # good to break before 'if', 'unless', etc
12982 if ( $is_if_brace_follower{$next_nonblank_token} ) {
12983 $bond_str = VERY_WEAK;
12986 if ( $next_nonblank_type eq 'k' ) {
12988 # keywords like 'unless', 'if', etc, within statements
12990 if ( $is_good_keyword_breakpoint{$next_nonblank_token} ) {
12991 $bond_str = VERY_WEAK / 1.05;
12995 # try not to break before a comma-arrow
12996 elsif ( $next_nonblank_type eq '=>' ) {
12997 if ( $bond_str < STRONG ) { $bond_str = STRONG }
13000 #----------------------------------------------------------------------
13001 # only set NO_BREAK's from here on
13002 #----------------------------------------------------------------------
13003 if ( $type eq 'C' or $type eq 'U' ) {
13005 # use strict requires that bare word and => not be separated
13006 if ( $next_nonblank_type eq '=>' ) {
13007 $bond_str = NO_BREAK;
13010 # Never break between a bareword and a following paren because
13011 # perl may give an error. For example, if a break is placed
13012 # between 'to_filehandle' and its '(' the following line will
13013 # give a syntax error [Carp.pm]: my( $no) =fileno(
13014 # to_filehandle( $in)) ;
13015 if ( $next_nonblank_token eq '(' ) {
13016 $bond_str = NO_BREAK;
13020 # use strict requires that bare word within braces not start new line
13021 elsif ( $type eq 'L' ) {
13023 if ( $next_nonblank_type eq 'w' ) {
13024 $bond_str = NO_BREAK;
13028 # in older version of perl, use strict can cause problems with
13029 # breaks before bare words following opening parens. For example,
13030 # this will fail under older versions if a break is made between
13033 # open( MAIL, "a long filename or command");
13035 elsif ( $type eq '{' ) {
13037 if ( $token eq '(' && $next_nonblank_type eq 'w' ) {
13039 # but it's fine to break if the word is followed by a '=>'
13040 # or if it is obviously a sub call
13041 my $i_next_next_nonblank = $i_next_nonblank + 1;
13042 my $next_next_type = $types_to_go[$i_next_next_nonblank];
13043 if ( $next_next_type eq 'b'
13044 && $i_next_nonblank < $max_index_to_go )
13046 $i_next_next_nonblank++;
13047 $next_next_type = $types_to_go[$i_next_next_nonblank];
13050 ##if ( $next_next_type ne '=>' ) {
13051 # these are ok: '->xxx', '=>', '('
13053 # We'll check for an old breakpoint and keep a leading
13054 # bareword if it was that way in the input file.
13055 # Presumably it was ok that way. For example, the
13056 # following would remain unchanged:
13059 # January, February, March, April,
13060 # May, June, July, August,
13061 # September, October, November, December,
13064 # This should be sufficient:
13065 if ( !$old_breakpoint_to_go[$i]
13066 && ( $next_next_type eq ',' || $next_next_type eq '}' )
13069 $bond_str = NO_BREAK;
13074 elsif ( $type eq 'w' ) {
13076 if ( $next_nonblank_type eq 'R' ) {
13077 $bond_str = NO_BREAK;
13080 # use strict requires that bare word and => not be separated
13081 if ( $next_nonblank_type eq '=>' ) {
13082 $bond_str = NO_BREAK;
13086 # in fact, use strict hates bare words on any new line. For
13087 # example, a break before the underscore here provokes the
13088 # wrath of use strict:
13089 # if ( -r $fn && ( -s _ || $AllowZeroFilesize)) {
13090 elsif ( $type eq 'F' ) {
13091 $bond_str = NO_BREAK;
13094 # use strict does not allow separating type info from trailing { }
13095 # testfile is readmail.pl
13096 elsif ( $type eq 't' or $type eq 'i' ) {
13098 if ( $next_nonblank_type eq 'L' ) {
13099 $bond_str = NO_BREAK;
13103 # Do not break between a possible filehandle and a ? or / and do
13104 # not introduce a break after it if there is no blank
13106 elsif ( $type eq 'Z' ) {
13111 # if there is no blank and we do not want one. Examples:
13112 # print $x++ # do not break after $x
13113 # print HTML"HELLO" # break ok after HTML
13116 && defined( $want_left_space{$next_type} )
13117 && $want_left_space{$next_type} == WS_NO
13120 # or we might be followed by the start of a quote
13121 || $next_nonblank_type =~ /^[\/\?]$/
13124 $bond_str = NO_BREAK;
13128 # Do not break before a possible file handle
13129 if ( $next_nonblank_type eq 'Z' ) {
13130 $bond_str = NO_BREAK;
13133 # As a defensive measure, do not break between a '(' and a
13134 # filehandle. In some cases, this can cause an error. For
13135 # example, the following program works:
13142 # But this program fails:
13150 # This is normally only a problem with the 'extrude' option
13151 if ( $next_nonblank_type eq 'Y' && $token eq '(' ) {
13152 $bond_str = NO_BREAK;
13155 # Breaking before a ++ can cause perl to guess wrong. For
13156 # example the following line will cause a syntax error
13157 # with -extrude if we break between '$i' and '++' [fixstyle2]
13158 # print( ( $i++ & 1 ) ? $_ : ( $change{$_} || $_ ) );
13159 elsif ( $next_nonblank_type eq '++' ) {
13160 $bond_str = NO_BREAK;
13163 # Breaking before a ? before a quote can cause trouble if
13164 # they are not separated by a blank.
13165 # Example: a syntax error occurs if you break before the ? here
13166 # my$logic=join$all?' && ':' || ',@regexps;
13167 # From: Professional_Perl_Programming_Code/multifind.pl
13168 elsif ( $next_nonblank_type eq '?' ) {
13169 $bond_str = NO_BREAK
13170 if ( $types_to_go[ $i_next_nonblank + 1 ] eq 'Q' );
13173 # Breaking before a . followed by a number
13174 # can cause trouble if there is no intervening space
13175 # Example: a syntax error occurs if you break before the .2 here
13176 # $str .= pack($endian.2, ensurrogate($ord));
13177 # From: perl58/Unicode.pm
13178 elsif ( $next_nonblank_type eq '.' ) {
13179 $bond_str = NO_BREAK
13180 if ( $types_to_go[ $i_next_nonblank + 1 ] eq 'n' );
13183 # patch to put cuddled elses back together when on multiple
13184 # lines, as in: } \n else \n { \n
13185 if ($rOpts_cuddled_else) {
13187 if ( ( $token eq 'else' ) && ( $next_nonblank_type eq '{' )
13188 || ( $type eq '}' ) && ( $next_nonblank_token eq 'else' ) )
13190 $bond_str = NO_BREAK;
13194 # keep '}' together with ';'
13195 if ( ( $token eq '}' ) && ( $next_nonblank_type eq ';' ) ) {
13196 $bond_str = NO_BREAK;
13199 # never break between sub name and opening paren
13200 if ( ( $type eq 'w' ) && ( $next_nonblank_token eq '(' ) ) {
13201 $bond_str = NO_BREAK;
13204 #---------------------------------------------------------------
13206 # now take nesting depth into account
13207 #---------------------------------------------------------------
13208 # final strength incorporates the bond strength and nesting depth
13211 if ( defined($bond_str) && !$nobreak_to_go[$i] ) {
13212 if ( $total_nesting_depth > 0 ) {
13213 $strength = $bond_str + $total_nesting_depth;
13216 $strength = $bond_str;
13220 $strength = NO_BREAK;
13223 # always break after side comment
13224 if ( $type eq '#' ) { $strength = 0 }
13226 $bond_strength_to_go[$i] = $strength;
13228 FORMATTER_DEBUG_FLAG_BOND && do {
13229 my $str = substr( $token, 0, 15 );
13230 $str .= ' ' x ( 16 - length($str) );
13232 "BOND: i=$i $str $type $next_nonblank_type depth=$total_nesting_depth strength=$bond_str_1 -> $bond_str -> $strength \n";
13239 sub pad_array_to_go {
13241 # to simplify coding in scan_list and set_bond_strengths, it helps
13242 # to create some extra blank tokens at the end of the arrays
13243 $tokens_to_go[ $max_index_to_go + 1 ] = '';
13244 $tokens_to_go[ $max_index_to_go + 2 ] = '';
13245 $types_to_go[ $max_index_to_go + 1 ] = 'b';
13246 $types_to_go[ $max_index_to_go + 2 ] = 'b';
13247 $nesting_depth_to_go[ $max_index_to_go + 1 ] =
13248 $nesting_depth_to_go[$max_index_to_go];
13251 if ( $is_closing_type{ $types_to_go[$max_index_to_go] } ) {
13252 if ( $nesting_depth_to_go[$max_index_to_go] <= 0 ) {
13254 # shouldn't happen:
13255 unless ( get_saw_brace_error() ) {
13257 "Program bug in scan_list: hit nesting error which should have been caught\n"
13259 report_definite_bug();
13263 $nesting_depth_to_go[ $max_index_to_go + 1 ] -= 1;
13268 elsif ( $is_opening_type{ $types_to_go[$max_index_to_go] } ) {
13269 $nesting_depth_to_go[ $max_index_to_go + 1 ] += 1;
13273 { # begin scan_list
13276 $block_type, $current_depth,
13278 $i_last_nonblank_token, $last_colon_sequence_number,
13279 $last_nonblank_token, $last_nonblank_type,
13280 $last_old_breakpoint_count, $minimum_depth,
13281 $next_nonblank_block_type, $next_nonblank_token,
13282 $next_nonblank_type, $old_breakpoint_count,
13283 $starting_breakpoint_count, $starting_depth,
13289 @breakpoint_stack, @breakpoint_undo_stack,
13290 @comma_index, @container_type,
13291 @identifier_count_stack, @index_before_arrow,
13292 @interrupted_list, @item_count_stack,
13293 @last_comma_index, @last_dot_index,
13294 @last_nonblank_type, @old_breakpoint_count_stack,
13295 @opening_structure_index_stack, @rfor_semicolon_list,
13296 @has_old_logical_breakpoints, @rand_or_list,
13300 # routine to define essential variables when we go 'up' to
13302 sub check_for_new_minimum_depth {
13304 if ( $depth < $minimum_depth ) {
13306 $minimum_depth = $depth;
13308 # these arrays need not retain values between calls
13309 $breakpoint_stack[$depth] = $starting_breakpoint_count;
13310 $container_type[$depth] = "";
13311 $identifier_count_stack[$depth] = 0;
13312 $index_before_arrow[$depth] = -1;
13313 $interrupted_list[$depth] = 1;
13314 $item_count_stack[$depth] = 0;
13315 $last_nonblank_type[$depth] = "";
13316 $opening_structure_index_stack[$depth] = -1;
13318 $breakpoint_undo_stack[$depth] = undef;
13319 $comma_index[$depth] = undef;
13320 $last_comma_index[$depth] = undef;
13321 $last_dot_index[$depth] = undef;
13322 $old_breakpoint_count_stack[$depth] = undef;
13323 $has_old_logical_breakpoints[$depth] = 0;
13324 $rand_or_list[$depth] = [];
13325 $rfor_semicolon_list[$depth] = [];
13326 $i_equals[$depth] = -1;
13328 # these arrays must retain values between calls
13329 if ( !defined( $has_broken_sublist[$depth] ) ) {
13330 $dont_align[$depth] = 0;
13331 $has_broken_sublist[$depth] = 0;
13332 $want_comma_break[$depth] = 0;
13337 # routine to decide which commas to break at within a container;
13339 # $bp_count = number of comma breakpoints set
13340 # $do_not_break_apart = a flag indicating if container need not
13342 sub set_comma_breakpoints {
13346 my $do_not_break_apart = 0;
13349 if ( $item_count_stack[$dd] ) {
13351 # handle commas not in containers...
13352 if ( $dont_align[$dd] ) {
13353 do_uncontained_comma_breaks($dd);
13356 # handle commas within containers...
13358 my $fbc = $forced_breakpoint_count;
13360 # always open comma lists not preceded by keywords,
13361 # barewords, identifiers (that is, anything that doesn't
13362 # look like a function call)
13363 my $must_break_open = $last_nonblank_type[$dd] !~ /^[kwiU]$/;
13365 set_comma_breakpoints_do(
13367 $opening_structure_index_stack[$dd],
13369 $item_count_stack[$dd],
13370 $identifier_count_stack[$dd],
13372 $next_nonblank_type,
13373 $container_type[$dd],
13374 $interrupted_list[$dd],
13375 \$do_not_break_apart,
13378 $bp_count = $forced_breakpoint_count - $fbc;
13379 $do_not_break_apart = 0 if $must_break_open;
13382 return ( $bp_count, $do_not_break_apart );
13385 sub do_uncontained_comma_breaks {
13387 # Handle commas not in containers...
13388 # This is a catch-all routine for commas that we
13389 # don't know what to do with because the don't fall
13390 # within containers. We will bias the bond strength
13391 # to break at commas which ended lines in the input
13392 # file. This usually works better than just trying
13393 # to put as many items on a line as possible. A
13394 # downside is that if the input file is garbage it
13395 # won't work very well. However, the user can always
13396 # prevent following the old breakpoints with the
13400 foreach my $ii ( @{ $comma_index[$dd] } ) {
13401 if ( $old_breakpoint_to_go[$ii] ) {
13402 $bond_strength_to_go[$ii] = $bias;
13404 # reduce bias magnitude to force breaks in order
13409 # Also put a break before the first comma if
13410 # (1) there was a break there in the input, and
13411 # (2) that was exactly one previous break in the input
13413 # For example, we will follow the user and break after
13414 # 'print' in this snippet:
13416 # "conformability (Not the same dimension)\n",
13417 # "\t", $have, " is ", text_unit($hu), "\n",
13418 # "\t", $want, " is ", text_unit($wu), "\n",
13420 my $i_first_comma = $comma_index[$dd]->[0];
13421 if ( $old_breakpoint_to_go[$i_first_comma] ) {
13422 my $level_comma = $levels_to_go[$i_first_comma];
13425 for ( my $ii = $i_first_comma - 1 ; $ii >= 0 ; $ii -= 1 ) {
13426 if ( $old_breakpoint_to_go[$ii] ) {
13428 last if ( $obp_count > 1 );
13430 if ( $levels_to_go[$ii] == $level_comma );
13433 if ( $ibreak >= 0 && $obp_count == 1 ) {
13434 set_forced_breakpoint($ibreak);
13439 my %is_logical_container;
13442 @_ = qw# if elsif unless while and or err not && | || ? : ! #;
13443 @is_logical_container{@_} = (1) x scalar(@_);
13446 sub set_for_semicolon_breakpoints {
13448 foreach ( @{ $rfor_semicolon_list[$dd] } ) {
13449 set_forced_breakpoint($_);
13453 sub set_logical_breakpoints {
13456 $item_count_stack[$dd] == 0
13457 && $is_logical_container{ $container_type[$dd] }
13460 || $has_old_logical_breakpoints[$dd]
13464 # Look for breaks in this order:
13467 foreach my $i ( 0 .. 3 ) {
13468 if ( $rand_or_list[$dd][$i] ) {
13469 foreach ( @{ $rand_or_list[$dd][$i] } ) {
13470 set_forced_breakpoint($_);
13473 # break at any 'if' and 'unless' too
13474 foreach ( @{ $rand_or_list[$dd][4] } ) {
13475 set_forced_breakpoint($_);
13477 $rand_or_list[$dd] = [];
13484 sub is_unbreakable_container {
13486 # never break a container of one of these types
13487 # because bad things can happen (map1.t)
13489 $is_sort_map_grep{ $container_type[$dd] };
13494 # This routine is responsible for setting line breaks for all lists,
13495 # so that hierarchical structure can be displayed and so that list
13496 # items can be vertically aligned. The output of this routine is
13497 # stored in the array @forced_breakpoint_to_go, which is used to set
13498 # final breakpoints.
13500 $starting_depth = $nesting_depth_to_go[0];
13503 $current_depth = $starting_depth;
13505 $last_colon_sequence_number = -1;
13506 $last_nonblank_token = ';';
13507 $last_nonblank_type = ';';
13508 $last_nonblank_block_type = ' ';
13509 $last_old_breakpoint_count = 0;
13510 $minimum_depth = $current_depth + 1; # forces update in check below
13511 $old_breakpoint_count = 0;
13512 $starting_breakpoint_count = $forced_breakpoint_count;
13515 $type_sequence = '';
13517 check_for_new_minimum_depth($current_depth);
13519 my $is_long_line = excess_line_length( 0, $max_index_to_go ) > 0;
13520 my $want_previous_breakpoint = -1;
13522 my $saw_good_breakpoint;
13523 my $i_line_end = -1;
13524 my $i_line_start = -1;
13526 # loop over all tokens in this batch
13527 while ( ++$i <= $max_index_to_go ) {
13528 if ( $type ne 'b' ) {
13529 $i_last_nonblank_token = $i - 1;
13530 $last_nonblank_type = $type;
13531 $last_nonblank_token = $token;
13532 $last_nonblank_block_type = $block_type;
13534 $type = $types_to_go[$i];
13535 $block_type = $block_type_to_go[$i];
13536 $token = $tokens_to_go[$i];
13537 $type_sequence = $type_sequence_to_go[$i];
13538 my $next_type = $types_to_go[ $i + 1 ];
13539 my $next_token = $tokens_to_go[ $i + 1 ];
13540 my $i_next_nonblank = ( ( $next_type eq 'b' ) ? $i + 2 : $i + 1 );
13541 $next_nonblank_type = $types_to_go[$i_next_nonblank];
13542 $next_nonblank_token = $tokens_to_go[$i_next_nonblank];
13543 $next_nonblank_block_type = $block_type_to_go[$i_next_nonblank];
13545 # set break if flag was set
13546 if ( $want_previous_breakpoint >= 0 ) {
13547 set_forced_breakpoint($want_previous_breakpoint);
13548 $want_previous_breakpoint = -1;
13551 $last_old_breakpoint_count = $old_breakpoint_count;
13552 if ( $old_breakpoint_to_go[$i] ) {
13554 $i_line_start = $i_next_nonblank;
13556 $old_breakpoint_count++;
13558 # Break before certain keywords if user broke there and
13559 # this is a 'safe' break point. The idea is to retain
13560 # any preferred breaks for sequential list operations,
13561 # like a schwartzian transform.
13562 if ($rOpts_break_at_old_keyword_breakpoints) {
13564 $next_nonblank_type eq 'k'
13565 && $is_keyword_returning_list{$next_nonblank_token}
13566 && ( $type =~ /^[=\)\]\}Riw]$/
13568 && $is_keyword_returning_list{$token} )
13572 # we actually have to set this break next time through
13573 # the loop because if we are at a closing token (such
13574 # as '}') which forms a one-line block, this break might
13576 $want_previous_breakpoint = $i;
13580 next if ( $type eq 'b' );
13581 $depth = $nesting_depth_to_go[ $i + 1 ];
13583 # safety check - be sure we always break after a comment
13584 # Shouldn't happen .. an error here probably means that the
13585 # nobreak flag did not get turned off correctly during
13587 if ( $type eq '#' ) {
13588 if ( $i != $max_index_to_go ) {
13590 "Non-fatal program bug: backup logic needed to break after a comment\n"
13592 report_definite_bug();
13593 $nobreak_to_go[$i] = 0;
13594 set_forced_breakpoint($i);
13598 # Force breakpoints at certain tokens in long lines.
13599 # Note that such breakpoints will be undone later if these tokens
13600 # are fully contained within parens on a line.
13603 # break before a keyword within a line
13607 # if one of these keywords:
13608 && $token =~ /^(if|unless|while|until|for)$/
13610 # but do not break at something like '1 while'
13611 && ( $last_nonblank_type ne 'n' || $i > 2 )
13613 # and let keywords follow a closing 'do' brace
13614 && $last_nonblank_block_type ne 'do'
13619 # or container is broken (by side-comment, etc)
13620 || ( $next_nonblank_token eq '('
13621 && $mate_index_to_go[$i_next_nonblank] < $i )
13625 set_forced_breakpoint( $i - 1 );
13628 # remember locations of '||' and '&&' for possible breaks if we
13629 # decide this is a long logical expression.
13630 if ( $type eq '||' ) {
13631 push @{ $rand_or_list[$depth][2] }, $i;
13632 ++$has_old_logical_breakpoints[$depth]
13633 if ( ( $i == $i_line_start || $i == $i_line_end )
13634 && $rOpts_break_at_old_logical_breakpoints );
13636 elsif ( $type eq '&&' ) {
13637 push @{ $rand_or_list[$depth][3] }, $i;
13638 ++$has_old_logical_breakpoints[$depth]
13639 if ( ( $i == $i_line_start || $i == $i_line_end )
13640 && $rOpts_break_at_old_logical_breakpoints );
13642 elsif ( $type eq 'f' ) {
13643 push @{ $rfor_semicolon_list[$depth] }, $i;
13645 elsif ( $type eq 'k' ) {
13646 if ( $token eq 'and' ) {
13647 push @{ $rand_or_list[$depth][1] }, $i;
13648 ++$has_old_logical_breakpoints[$depth]
13649 if ( ( $i == $i_line_start || $i == $i_line_end )
13650 && $rOpts_break_at_old_logical_breakpoints );
13653 # break immediately at 'or's which are probably not in a logical
13654 # block -- but we will break in logical breaks below so that
13655 # they do not add to the forced_breakpoint_count
13656 elsif ( $token eq 'or' ) {
13657 push @{ $rand_or_list[$depth][0] }, $i;
13658 ++$has_old_logical_breakpoints[$depth]
13659 if ( ( $i == $i_line_start || $i == $i_line_end )
13660 && $rOpts_break_at_old_logical_breakpoints );
13661 if ( $is_logical_container{ $container_type[$depth] } ) {
13664 if ($is_long_line) { set_forced_breakpoint($i) }
13665 elsif ( ( $i == $i_line_start || $i == $i_line_end )
13666 && $rOpts_break_at_old_logical_breakpoints )
13668 $saw_good_breakpoint = 1;
13672 elsif ( $token eq 'if' || $token eq 'unless' ) {
13673 push @{ $rand_or_list[$depth][4] }, $i;
13674 if ( ( $i == $i_line_start || $i == $i_line_end )
13675 && $rOpts_break_at_old_logical_breakpoints )
13677 set_forced_breakpoint($i);
13681 elsif ( $is_assignment{$type} ) {
13682 $i_equals[$depth] = $i;
13685 if ($type_sequence) {
13687 # handle any postponed closing breakpoints
13688 if ( $token =~ /^[\)\]\}\:]$/ ) {
13689 if ( $type eq ':' ) {
13690 $last_colon_sequence_number = $type_sequence;
13692 # TESTING: retain break at a ':' line break
13693 if ( ( $i == $i_line_start || $i == $i_line_end )
13694 && $rOpts_break_at_old_ternary_breakpoints )
13698 set_forced_breakpoint($i);
13700 # break at previous '='
13701 if ( $i_equals[$depth] > 0 ) {
13702 set_forced_breakpoint( $i_equals[$depth] );
13703 $i_equals[$depth] = -1;
13707 if ( defined( $postponed_breakpoint{$type_sequence} ) ) {
13708 my $inc = ( $type eq ':' ) ? 0 : 1;
13709 set_forced_breakpoint( $i - $inc );
13710 delete $postponed_breakpoint{$type_sequence};
13714 # set breaks at ?/: if they will get separated (and are
13715 # not a ?/: chain), or if the '?' is at the end of the
13717 elsif ( $token eq '?' ) {
13718 my $i_colon = $mate_index_to_go[$i];
13720 $i_colon <= 0 # the ':' is not in this batch
13721 || $i == 0 # this '?' is the first token of the line
13723 $max_index_to_go # or this '?' is the last token
13727 # don't break at a '?' if preceded by ':' on
13728 # this line of previous ?/: pair on this line.
13729 # This is an attempt to preserve a chain of ?/:
13730 # expressions (elsif2.t). And don't break if
13731 # this has a side comment.
13732 set_forced_breakpoint($i)
13734 $type_sequence == (
13735 $last_colon_sequence_number +
13736 TYPE_SEQUENCE_INCREMENT
13738 || $tokens_to_go[$max_index_to_go] eq '#'
13740 set_closing_breakpoint($i);
13745 #print "LISTX sees: i=$i type=$type tok=$token block=$block_type depth=$depth\n";
13747 #------------------------------------------------------------
13748 # Handle Increasing Depth..
13750 # prepare for a new list when depth increases
13751 # token $i is a '(','{', or '['
13752 #------------------------------------------------------------
13753 if ( $depth > $current_depth ) {
13755 $breakpoint_stack[$depth] = $forced_breakpoint_count;
13756 $breakpoint_undo_stack[$depth] = $forced_breakpoint_undo_count;
13757 $has_broken_sublist[$depth] = 0;
13758 $identifier_count_stack[$depth] = 0;
13759 $index_before_arrow[$depth] = -1;
13760 $interrupted_list[$depth] = 0;
13761 $item_count_stack[$depth] = 0;
13762 $last_comma_index[$depth] = undef;
13763 $last_dot_index[$depth] = undef;
13764 $last_nonblank_type[$depth] = $last_nonblank_type;
13765 $old_breakpoint_count_stack[$depth] = $old_breakpoint_count;
13766 $opening_structure_index_stack[$depth] = $i;
13767 $rand_or_list[$depth] = [];
13768 $rfor_semicolon_list[$depth] = [];
13769 $i_equals[$depth] = -1;
13770 $want_comma_break[$depth] = 0;
13771 $container_type[$depth] =
13772 ( $last_nonblank_type =~ /^(k|=>|&&|\|\||\?|\:|\.)$/ )
13773 ? $last_nonblank_token
13775 $has_old_logical_breakpoints[$depth] = 0;
13777 # if line ends here then signal closing token to break
13778 if ( $next_nonblank_type eq 'b' || $next_nonblank_type eq '#' )
13780 set_closing_breakpoint($i);
13783 # Not all lists of values should be vertically aligned..
13784 $dont_align[$depth] =
13786 # code BLOCKS are handled at a higher level
13787 ( $block_type ne "" )
13789 # certain paren lists
13790 || ( $type eq '(' ) && (
13792 # it does not usually look good to align a list of
13793 # identifiers in a parameter list, as in:
13794 # my($var1, $var2, ...)
13795 # (This test should probably be refined, for now I'm just
13796 # testing for any keyword)
13797 ( $last_nonblank_type eq 'k' )
13799 # a trailing '(' usually indicates a non-list
13800 || ( $next_nonblank_type eq '(' )
13803 # patch to outdent opening brace of long if/for/..
13804 # statements (like this one). See similar coding in
13805 # set_continuation breaks. We have also catch it here for
13806 # short line fragments which otherwise will not go through
13807 # set_continuation_breaks.
13811 # if we have the ')' but not its '(' in this batch..
13812 && ( $last_nonblank_token eq ')' )
13813 && $mate_index_to_go[$i_last_nonblank_token] < 0
13815 # and user wants brace to left
13816 && !$rOpts->{'opening-brace-always-on-right'}
13818 && ( $type eq '{' ) # should be true
13819 && ( $token eq '{' ) # should be true
13822 set_forced_breakpoint( $i - 1 );
13826 #------------------------------------------------------------
13827 # Handle Decreasing Depth..
13829 # finish off any old list when depth decreases
13830 # token $i is a ')','}', or ']'
13831 #------------------------------------------------------------
13832 elsif ( $depth < $current_depth ) {
13834 check_for_new_minimum_depth($depth);
13836 # force all outer logical containers to break after we see on
13838 $has_old_logical_breakpoints[$depth] ||=
13839 $has_old_logical_breakpoints[$current_depth];
13841 # Patch to break between ') {' if the paren list is broken.
13842 # There is similar logic in set_continuation_breaks for
13843 # non-broken lists.
13845 && $next_nonblank_block_type
13846 && $interrupted_list[$current_depth]
13847 && $next_nonblank_type eq '{'
13848 && !$rOpts->{'opening-brace-always-on-right'} )
13850 set_forced_breakpoint($i);
13853 #print "LISTY sees: i=$i type=$type tok=$token block=$block_type depth=$depth next=$next_nonblank_type next_block=$next_nonblank_block_type inter=$interrupted_list[$current_depth]\n";
13855 # set breaks at commas if necessary
13856 my ( $bp_count, $do_not_break_apart ) =
13857 set_comma_breakpoints($current_depth);
13859 my $i_opening = $opening_structure_index_stack[$current_depth];
13860 my $saw_opening_structure = ( $i_opening >= 0 );
13862 # this term is long if we had to break at interior commas..
13863 my $is_long_term = $bp_count > 0;
13865 # ..or if the length between opening and closing parens exceeds
13866 # allowed line length
13867 if ( !$is_long_term && $saw_opening_structure ) {
13868 my $i_opening_minus = find_token_starting_list($i_opening);
13870 # Note: we have to allow for one extra space after a
13871 # closing token so that we do not strand a comma or
13872 # semicolon, hence the '>=' here (oneline.t)
13874 excess_line_length( $i_opening_minus, $i ) >= 0;
13877 # We've set breaks after all comma-arrows. Now we have to
13878 # undo them if this can be a one-line block
13879 # (the only breakpoints set will be due to comma-arrows)
13882 # user doesn't require breaking after all comma-arrows
13883 ( $rOpts_comma_arrow_breakpoints != 0 )
13885 # and if the opening structure is in this batch
13886 && $saw_opening_structure
13888 # and either on the same old line
13890 $old_breakpoint_count_stack[$current_depth] ==
13891 $last_old_breakpoint_count
13893 # or user wants to form long blocks with arrows
13894 || $rOpts_comma_arrow_breakpoints == 2
13897 # and we made some breakpoints between the opening and closing
13898 && ( $breakpoint_undo_stack[$current_depth] <
13899 $forced_breakpoint_undo_count )
13901 # and this block is short enough to fit on one line
13902 # Note: use < because need 1 more space for possible comma
13907 undo_forced_breakpoint_stack(
13908 $breakpoint_undo_stack[$current_depth] );
13911 # now see if we have any comma breakpoints left
13912 my $has_comma_breakpoints =
13913 ( $breakpoint_stack[$current_depth] !=
13914 $forced_breakpoint_count );
13916 # update broken-sublist flag of the outer container
13917 $has_broken_sublist[$depth] =
13918 $has_broken_sublist[$depth]
13919 || $has_broken_sublist[$current_depth]
13921 || $has_comma_breakpoints;
13923 # Having come to the closing ')', '}', or ']', now we have to decide if we
13924 # should 'open up' the structure by placing breaks at the opening and
13925 # closing containers. This is a tricky decision. Here are some of the
13926 # basic considerations:
13928 # -If this is a BLOCK container, then any breakpoints will have already
13929 # been set (and according to user preferences), so we need do nothing here.
13931 # -If we have a comma-separated list for which we can align the list items,
13932 # then we need to do so because otherwise the vertical aligner cannot
13933 # currently do the alignment.
13935 # -If this container does itself contain a container which has been broken
13936 # open, then it should be broken open to properly show the structure.
13938 # -If there is nothing to align, and no other reason to break apart,
13939 # then do not do it.
13941 # We will not break open the parens of a long but 'simple' logical expression.
13944 # This is an example of a simple logical expression and its formatting:
13946 # if ( $bigwasteofspace1 && $bigwasteofspace2
13947 # || $bigwasteofspace3 && $bigwasteofspace4 )
13949 # Most people would prefer this than the 'spacey' version:
13952 # $bigwasteofspace1 && $bigwasteofspace2
13953 # || $bigwasteofspace3 && $bigwasteofspace4
13956 # To illustrate the rules for breaking logical expressions, consider:
13960 # and ( exists $ids_excl_uc{$id_uc}
13961 # or grep $id_uc =~ /$_/, @ids_excl_uc ))
13963 # This is on the verge of being difficult to read. The current default is to
13964 # open it up like this:
13969 # and ( exists $ids_excl_uc{$id_uc}
13970 # or grep $id_uc =~ /$_/, @ids_excl_uc )
13973 # This is a compromise which tries to avoid being too dense and to spacey.
13974 # A more spaced version would be:
13980 # exists $ids_excl_uc{$id_uc}
13981 # or grep $id_uc =~ /$_/, @ids_excl_uc
13985 # Some people might prefer the spacey version -- an option could be added. The
13986 # innermost expression contains a long block '( exists $ids_... ')'.
13988 # Here is how the logic goes: We will force a break at the 'or' that the
13989 # innermost expression contains, but we will not break apart its opening and
13990 # closing containers because (1) it contains no multi-line sub-containers itself,
13991 # and (2) there is no alignment to be gained by breaking it open like this
13994 # exists $ids_excl_uc{$id_uc}
13995 # or grep $id_uc =~ /$_/, @ids_excl_uc
13998 # (although this looks perfectly ok and might be good for long expressions). The
13999 # outer 'if' container, though, contains a broken sub-container, so it will be
14000 # broken open to avoid too much density. Also, since it contains no 'or's, there
14001 # will be a forced break at its 'and'.
14003 # set some flags telling something about this container..
14004 my $is_simple_logical_expression = 0;
14005 if ( $item_count_stack[$current_depth] == 0
14006 && $saw_opening_structure
14007 && $tokens_to_go[$i_opening] eq '('
14008 && $is_logical_container{ $container_type[$current_depth] }
14012 # This seems to be a simple logical expression with
14013 # no existing breakpoints. Set a flag to prevent
14015 if ( !$has_comma_breakpoints ) {
14016 $is_simple_logical_expression = 1;
14019 # This seems to be a simple logical expression with
14020 # breakpoints (broken sublists, for example). Break
14021 # at all 'or's and '||'s.
14023 set_logical_breakpoints($current_depth);
14028 && @{ $rfor_semicolon_list[$current_depth] } )
14030 set_for_semicolon_breakpoints($current_depth);
14032 # open up a long 'for' or 'foreach' container to allow
14033 # leading term alignment unless -lp is used.
14034 $has_comma_breakpoints = 1
14035 unless $rOpts_line_up_parentheses;
14040 # breaks for code BLOCKS are handled at a higher level
14043 # we do not need to break at the top level of an 'if'
14045 && !$is_simple_logical_expression
14047 ## modification to keep ': (' containers vertically tight;
14048 ## but probably better to let user set -vt=1 to avoid
14049 ## inconsistency with other paren types
14050 ## && ($container_type[$current_depth] ne ':')
14052 # otherwise, we require one of these reasons for breaking:
14055 # - this term has forced line breaks
14056 $has_comma_breakpoints
14058 # - the opening container is separated from this batch
14059 # for some reason (comment, blank line, code block)
14060 # - this is a non-paren container spanning multiple lines
14061 || !$saw_opening_structure
14063 # - this is a long block contained in another breakable
14066 && $container_environment_to_go[$i_opening] ne
14072 # For -lp option, we must put a breakpoint before
14073 # the token which has been identified as starting
14074 # this indentation level. This is necessary for
14075 # proper alignment.
14076 if ( $rOpts_line_up_parentheses && $saw_opening_structure )
14078 my $item = $leading_spaces_to_go[ $i_opening + 1 ];
14079 if ( $i_opening + 1 < $max_index_to_go
14080 && $types_to_go[ $i_opening + 1 ] eq 'b' )
14082 $item = $leading_spaces_to_go[ $i_opening + 2 ];
14084 if ( defined($item) ) {
14085 my $i_start_2 = $item->get_STARTING_INDEX();
14087 defined($i_start_2)
14089 # we are breaking after an opening brace, paren,
14090 # so don't break before it too
14091 && $i_start_2 ne $i_opening
14095 # Only break for breakpoints at the same
14096 # indentation level as the opening paren
14097 my $test1 = $nesting_depth_to_go[$i_opening];
14098 my $test2 = $nesting_depth_to_go[$i_start_2];
14099 if ( $test2 == $test1 ) {
14100 set_forced_breakpoint( $i_start_2 - 1 );
14106 # break after opening structure.
14107 # note: break before closing structure will be automatic
14108 if ( $minimum_depth <= $current_depth ) {
14110 set_forced_breakpoint($i_opening)
14111 unless ( $do_not_break_apart
14112 || is_unbreakable_container($current_depth) );
14114 # break at '.' of lower depth level before opening token
14115 if ( $last_dot_index[$depth] ) {
14116 set_forced_breakpoint( $last_dot_index[$depth] );
14119 # break before opening structure if preeced by another
14120 # closing structure and a comma. This is normally
14121 # done by the previous closing brace, but not
14122 # if it was a one-line block.
14123 if ( $i_opening > 2 ) {
14125 ( $types_to_go[ $i_opening - 1 ] eq 'b' )
14129 if ( $types_to_go[$i_prev] eq ','
14130 && $types_to_go[ $i_prev - 1 ] =~ /^[\)\}]$/ )
14132 set_forced_breakpoint($i_prev);
14135 # also break before something like ':(' or '?('
14138 $types_to_go[$i_prev] =~ /^([k\:\?]|&&|\|\|)$/ )
14140 my $token_prev = $tokens_to_go[$i_prev];
14141 if ( $want_break_before{$token_prev} ) {
14142 set_forced_breakpoint($i_prev);
14148 # break after comma following closing structure
14149 if ( $next_type eq ',' ) {
14150 set_forced_breakpoint( $i + 1 );
14153 # break before an '=' following closing structure
14155 $is_assignment{$next_nonblank_type}
14156 && ( $breakpoint_stack[$current_depth] !=
14157 $forced_breakpoint_count )
14160 set_forced_breakpoint($i);
14163 # break at any comma before the opening structure Added
14164 # for -lp, but seems to be good in general. It isn't
14165 # obvious how far back to look; the '5' below seems to
14166 # work well and will catch the comma in something like
14167 # push @list, myfunc( $param, $param, ..
14169 my $icomma = $last_comma_index[$depth];
14170 if ( defined($icomma) && ( $i_opening - $icomma ) < 5 ) {
14171 unless ( $forced_breakpoint_to_go[$icomma] ) {
14172 set_forced_breakpoint($icomma);
14175 } # end logic to open up a container
14177 # Break open a logical container open if it was already open
14178 elsif ($is_simple_logical_expression
14179 && $has_old_logical_breakpoints[$current_depth] )
14181 set_logical_breakpoints($current_depth);
14184 # Handle long container which does not get opened up
14185 elsif ($is_long_term) {
14187 # must set fake breakpoint to alert outer containers that
14189 set_fake_breakpoint();
14193 #------------------------------------------------------------
14194 # Handle this token
14195 #------------------------------------------------------------
14197 $current_depth = $depth;
14199 # handle comma-arrow
14200 if ( $type eq '=>' ) {
14201 next if ( $last_nonblank_type eq '=>' );
14202 next if $rOpts_break_at_old_comma_breakpoints;
14203 next if $rOpts_comma_arrow_breakpoints == 3;
14204 $want_comma_break[$depth] = 1;
14205 $index_before_arrow[$depth] = $i_last_nonblank_token;
14209 elsif ( $type eq '.' ) {
14210 $last_dot_index[$depth] = $i;
14213 # Turn off alignment if we are sure that this is not a list
14214 # environment. To be safe, we will do this if we see certain
14215 # non-list tokens, such as ';', and also the environment is
14216 # not a list. Note that '=' could be in any of the = operators
14217 # (lextest.t). We can't just use the reported environment
14218 # because it can be incorrect in some cases.
14219 elsif ( ( $type =~ /^[\;\<\>\~]$/ || $is_assignment{$type} )
14220 && $container_environment_to_go[$i] ne 'LIST' )
14222 $dont_align[$depth] = 1;
14223 $want_comma_break[$depth] = 0;
14224 $index_before_arrow[$depth] = -1;
14227 # now just handle any commas
14228 next unless ( $type eq ',' );
14230 $last_dot_index[$depth] = undef;
14231 $last_comma_index[$depth] = $i;
14233 # break here if this comma follows a '=>'
14234 # but not if there is a side comment after the comma
14235 if ( $want_comma_break[$depth] ) {
14237 if ( $next_nonblank_type =~ /^[\)\}\]R]$/ ) {
14238 $want_comma_break[$depth] = 0;
14239 $index_before_arrow[$depth] = -1;
14243 set_forced_breakpoint($i) unless ( $next_nonblank_type eq '#' );
14245 # break before the previous token if it looks safe
14246 # Example of something that we will not try to break before:
14247 # DBI::SQL_SMALLINT() => $ado_consts->{adSmallInt},
14248 # Also we don't want to break at a binary operator (like +):
14252 # $y - $R, -fill => 'black',
14254 my $ibreak = $index_before_arrow[$depth] - 1;
14256 && $tokens_to_go[ $ibreak + 1 ] !~ /^[\)\}\]]$/ )
14258 if ( $tokens_to_go[$ibreak] eq '-' ) { $ibreak-- }
14259 if ( $types_to_go[$ibreak] eq 'b' ) { $ibreak-- }
14260 if ( $types_to_go[$ibreak] =~ /^[,wiZCUG\(\{\[]$/ ) {
14262 # don't break pointer calls, such as the following:
14263 # File::Spec->curdir => 1,
14264 # (This is tokenized as adjacent 'w' tokens)
14265 if ( $tokens_to_go[ $ibreak + 1 ] !~ /^->/ ) {
14266 set_forced_breakpoint($ibreak);
14271 $want_comma_break[$depth] = 0;
14272 $index_before_arrow[$depth] = -1;
14274 # handle list which mixes '=>'s and ','s:
14275 # treat any list items so far as an interrupted list
14276 $interrupted_list[$depth] = 1;
14280 # break after all commas above starting depth
14281 if ( $depth < $starting_depth && !$dont_align[$depth] ) {
14282 set_forced_breakpoint($i) unless ( $next_nonblank_type eq '#' );
14286 # add this comma to the list..
14287 my $item_count = $item_count_stack[$depth];
14288 if ( $item_count == 0 ) {
14290 # but do not form a list with no opening structure
14293 # open INFILE_COPY, ">$input_file_copy"
14294 # or die ("very long message");
14296 if ( ( $opening_structure_index_stack[$depth] < 0 )
14297 && $container_environment_to_go[$i] eq 'BLOCK' )
14299 $dont_align[$depth] = 1;
14303 $comma_index[$depth][$item_count] = $i;
14304 ++$item_count_stack[$depth];
14305 if ( $last_nonblank_type =~ /^[iR\]]$/ ) {
14306 $identifier_count_stack[$depth]++;
14310 #-------------------------------------------
14311 # end of loop over all tokens in this batch
14312 #-------------------------------------------
14314 # set breaks for any unfinished lists ..
14315 for ( my $dd = $current_depth ; $dd >= $minimum_depth ; $dd-- ) {
14317 $interrupted_list[$dd] = 1;
14318 $has_broken_sublist[$dd] = 1 if ( $dd < $current_depth );
14319 set_comma_breakpoints($dd);
14320 set_logical_breakpoints($dd)
14321 if ( $has_old_logical_breakpoints[$dd] );
14322 set_for_semicolon_breakpoints($dd);
14324 # break open container...
14325 my $i_opening = $opening_structure_index_stack[$dd];
14326 set_forced_breakpoint($i_opening)
14328 is_unbreakable_container($dd)
14330 # Avoid a break which would place an isolated ' or "
14333 && $i_opening >= $max_index_to_go - 2
14334 && $token =~ /^['"]$/ )
14338 # Return a flag indicating if the input file had some good breakpoints.
14339 # This flag will be used to force a break in a line shorter than the
14340 # allowed line length.
14341 if ( $has_old_logical_breakpoints[$current_depth] ) {
14342 $saw_good_breakpoint = 1;
14344 return $saw_good_breakpoint;
14348 sub find_token_starting_list {
14350 # When testing to see if a block will fit on one line, some
14351 # previous token(s) may also need to be on the line; particularly
14352 # if this is a sub call. So we will look back at least one
14353 # token. NOTE: This isn't perfect, but not critical, because
14354 # if we mis-identify a block, it will be wrapped and therefore
14355 # fixed the next time it is formatted.
14356 my $i_opening_paren = shift;
14357 my $i_opening_minus = $i_opening_paren;
14358 my $im1 = $i_opening_paren - 1;
14359 my $im2 = $i_opening_paren - 2;
14360 my $im3 = $i_opening_paren - 3;
14361 my $typem1 = $types_to_go[$im1];
14362 my $typem2 = $im2 >= 0 ? $types_to_go[$im2] : 'b';
14363 if ( $typem1 eq ',' || ( $typem1 eq 'b' && $typem2 eq ',' ) ) {
14364 $i_opening_minus = $i_opening_paren;
14366 elsif ( $tokens_to_go[$i_opening_paren] eq '(' ) {
14367 $i_opening_minus = $im1 if $im1 >= 0;
14369 # walk back to improve length estimate
14370 for ( my $j = $im1 ; $j >= 0 ; $j-- ) {
14371 last if ( $types_to_go[$j] =~ /^[\(\[\{L\}\]\)Rb,]$/ );
14372 $i_opening_minus = $j;
14374 if ( $types_to_go[$i_opening_minus] eq 'b' ) { $i_opening_minus++ }
14376 elsif ( $typem1 eq 'k' ) { $i_opening_minus = $im1 }
14377 elsif ( $typem1 eq 'b' && $im2 >= 0 && $types_to_go[$im2] eq 'k' ) {
14378 $i_opening_minus = $im2;
14380 return $i_opening_minus;
14383 { # begin set_comma_breakpoints_do
14385 my %is_keyword_with_special_leading_term;
14389 # These keywords have prototypes which allow a special leading item
14390 # followed by a list
14392 qw(formline grep kill map printf sprintf push chmod join pack unshift);
14393 @is_keyword_with_special_leading_term{@_} = (1) x scalar(@_);
14396 sub set_comma_breakpoints_do {
14398 # Given a list with some commas, set breakpoints at some of the
14399 # commas, if necessary, to make it easy to read. This list is
14402 $depth, $i_opening_paren, $i_closing_paren,
14403 $item_count, $identifier_count, $rcomma_index,
14404 $next_nonblank_type, $list_type, $interrupted,
14405 $rdo_not_break_apart, $must_break_open,
14408 # nothing to do if no commas seen
14409 return if ( $item_count < 1 );
14410 my $i_first_comma = $$rcomma_index[0];
14411 my $i_true_last_comma = $$rcomma_index[ $item_count - 1 ];
14412 my $i_last_comma = $i_true_last_comma;
14413 if ( $i_last_comma >= $max_index_to_go ) {
14414 $i_last_comma = $$rcomma_index[ --$item_count - 1 ];
14415 return if ( $item_count < 1 );
14418 #---------------------------------------------------------------
14419 # find lengths of all items in the list to calculate page layout
14420 #---------------------------------------------------------------
14421 my $comma_count = $item_count;
14427 my @max_length = ( 0, 0 );
14428 my $first_term_length;
14429 my $i = $i_opening_paren;
14432 for ( my $j = 0 ; $j < $comma_count ; $j++ ) {
14433 $is_odd = 1 - $is_odd;
14434 $i_prev_plus = $i + 1;
14435 $i = $$rcomma_index[$j];
14438 ( $types_to_go[ $i - 1 ] eq 'b' ) ? $i - 2 : $i - 1;
14440 ( $types_to_go[$i_prev_plus] eq 'b' )
14443 push @i_term_begin, $i_term_begin;
14444 push @i_term_end, $i_term_end;
14445 push @i_term_comma, $i;
14447 # note: currently adding 2 to all lengths (for comma and space)
14449 2 + token_sequence_length( $i_term_begin, $i_term_end );
14450 push @item_lengths, $length;
14453 $first_term_length = $length;
14457 if ( $length > $max_length[$is_odd] ) {
14458 $max_length[$is_odd] = $length;
14463 # now we have to make a distinction between the comma count and item
14464 # count, because the item count will be one greater than the comma
14465 # count if the last item is not terminated with a comma
14467 ( $types_to_go[ $i_last_comma + 1 ] eq 'b' )
14468 ? $i_last_comma + 1
14471 ( $types_to_go[ $i_closing_paren - 1 ] eq 'b' )
14472 ? $i_closing_paren - 2
14473 : $i_closing_paren - 1;
14474 my $i_effective_last_comma = $i_last_comma;
14476 my $last_item_length = token_sequence_length( $i_b + 1, $i_e );
14478 if ( $last_item_length > 0 ) {
14480 # add 2 to length because other lengths include a comma and a blank
14481 $last_item_length += 2;
14482 push @item_lengths, $last_item_length;
14483 push @i_term_begin, $i_b + 1;
14484 push @i_term_end, $i_e;
14485 push @i_term_comma, undef;
14487 my $i_odd = $item_count % 2;
14489 if ( $last_item_length > $max_length[$i_odd] ) {
14490 $max_length[$i_odd] = $last_item_length;
14494 $i_effective_last_comma = $i_e + 1;
14496 if ( $types_to_go[ $i_b + 1 ] =~ /^[iR\]]$/ ) {
14497 $identifier_count++;
14501 #---------------------------------------------------------------
14502 # End of length calculations
14503 #---------------------------------------------------------------
14505 #---------------------------------------------------------------
14506 # Compound List Rule 1:
14507 # Break at (almost) every comma for a list containing a broken
14508 # sublist. This has higher priority than the Interrupted List
14510 #---------------------------------------------------------------
14511 if ( $has_broken_sublist[$depth] ) {
14513 # Break at every comma except for a comma between two
14514 # simple, small terms. This prevents long vertical
14515 # columns of, say, just 0's.
14516 my $small_length = 10; # 2 + actual maximum length wanted
14518 # We'll insert a break in long runs of small terms to
14519 # allow alignment in uniform tables.
14520 my $skipped_count = 0;
14521 my $columns = table_columns_available($i_first_comma);
14522 my $fields = int( $columns / $small_length );
14523 if ( $rOpts_maximum_fields_per_table
14524 && $fields > $rOpts_maximum_fields_per_table )
14526 $fields = $rOpts_maximum_fields_per_table;
14528 my $max_skipped_count = $fields - 1;
14530 my $is_simple_last_term = 0;
14531 my $is_simple_next_term = 0;
14532 foreach my $j ( 0 .. $item_count ) {
14533 $is_simple_last_term = $is_simple_next_term;
14534 $is_simple_next_term = 0;
14535 if ( $j < $item_count
14536 && $i_term_end[$j] == $i_term_begin[$j]
14537 && $item_lengths[$j] <= $small_length )
14539 $is_simple_next_term = 1;
14542 if ( $is_simple_last_term
14543 && $is_simple_next_term
14544 && $skipped_count < $max_skipped_count )
14549 $skipped_count = 0;
14550 my $i = $i_term_comma[ $j - 1 ];
14551 last unless defined $i;
14552 set_forced_breakpoint($i);
14556 # always break at the last comma if this list is
14557 # interrupted; we wouldn't want to leave a terminal '{', for
14559 if ($interrupted) { set_forced_breakpoint($i_true_last_comma) }
14563 #my ( $a, $b, $c ) = caller();
14564 #print "LISTX: in set_list $a $c interupt=$interrupted count=$item_count
14565 #i_first = $i_first_comma i_last=$i_last_comma max=$max_index_to_go\n";
14566 #print "depth=$depth has_broken=$has_broken_sublist[$depth] is_multi=$is_multiline opening_paren=($i_opening_paren) \n";
14568 #---------------------------------------------------------------
14569 # Interrupted List Rule:
14570 # A list is is forced to use old breakpoints if it was interrupted
14571 # by side comments or blank lines, or requested by user.
14572 #---------------------------------------------------------------
14573 if ( $rOpts_break_at_old_comma_breakpoints
14575 || $i_opening_paren < 0 )
14577 copy_old_breakpoints( $i_first_comma, $i_true_last_comma );
14581 #---------------------------------------------------------------
14582 # Looks like a list of items. We have to look at it and size it up.
14583 #---------------------------------------------------------------
14585 my $opening_token = $tokens_to_go[$i_opening_paren];
14586 my $opening_environment =
14587 $container_environment_to_go[$i_opening_paren];
14589 #-------------------------------------------------------------------
14590 # Return if this will fit on one line
14591 #-------------------------------------------------------------------
14593 my $i_opening_minus = find_token_starting_list($i_opening_paren);
14595 unless excess_line_length( $i_opening_minus, $i_closing_paren ) > 0;
14597 #-------------------------------------------------------------------
14598 # Now we know that this block spans multiple lines; we have to set
14599 # at least one breakpoint -- real or fake -- as a signal to break
14600 # open any outer containers.
14601 #-------------------------------------------------------------------
14602 set_fake_breakpoint();
14604 # be sure we do not extend beyond the current list length
14605 if ( $i_effective_last_comma >= $max_index_to_go ) {
14606 $i_effective_last_comma = $max_index_to_go - 1;
14609 # Set a flag indicating if we need to break open to keep -lp
14610 # items aligned. This is necessary if any of the list terms
14611 # exceeds the available space after the '('.
14612 my $need_lp_break_open = $must_break_open;
14613 if ( $rOpts_line_up_parentheses && !$must_break_open ) {
14614 my $columns_if_unbroken = $rOpts_maximum_line_length -
14615 total_line_length( $i_opening_minus, $i_opening_paren );
14616 $need_lp_break_open =
14617 ( $max_length[0] > $columns_if_unbroken )
14618 || ( $max_length[1] > $columns_if_unbroken )
14619 || ( $first_term_length > $columns_if_unbroken );
14622 # Specify if the list must have an even number of fields or not.
14623 # It is generally safest to assume an even number, because the
14624 # list items might be a hash list. But if we can be sure that
14625 # it is not a hash, then we can allow an odd number for more
14627 my $odd_or_even = 2; # 1 = odd field count ok, 2 = want even count
14629 if ( $identifier_count >= $item_count - 1
14630 || $is_assignment{$next_nonblank_type}
14631 || ( $list_type && $list_type ne '=>' && $list_type !~ /^[\:\?]$/ )
14637 # do we have a long first term which should be
14638 # left on a line by itself?
14639 my $use_separate_first_term = (
14640 $odd_or_even == 1 # only if we can use 1 field/line
14641 && $item_count > 3 # need several items
14642 && $first_term_length >
14643 2 * $max_length[0] - 2 # need long first term
14644 && $first_term_length >
14645 2 * $max_length[1] - 2 # need long first term
14648 # or do we know from the type of list that the first term should
14650 if ( !$use_separate_first_term ) {
14651 if ( $is_keyword_with_special_leading_term{$list_type} ) {
14652 $use_separate_first_term = 1;
14654 # should the container be broken open?
14655 if ( $item_count < 3 ) {
14656 if ( $i_first_comma - $i_opening_paren < 4 ) {
14657 $$rdo_not_break_apart = 1;
14660 elsif ($first_term_length < 20
14661 && $i_first_comma - $i_opening_paren < 4 )
14663 my $columns = table_columns_available($i_first_comma);
14664 if ( $first_term_length < $columns ) {
14665 $$rdo_not_break_apart = 1;
14672 if ($use_separate_first_term) {
14674 # ..set a break and update starting values
14675 $use_separate_first_term = 1;
14676 set_forced_breakpoint($i_first_comma);
14677 $i_opening_paren = $i_first_comma;
14678 $i_first_comma = $$rcomma_index[1];
14680 return if $comma_count == 1;
14681 shift @item_lengths;
14682 shift @i_term_begin;
14684 shift @i_term_comma;
14687 # if not, update the metrics to include the first term
14689 if ( $first_term_length > $max_length[0] ) {
14690 $max_length[0] = $first_term_length;
14694 # Field width parameters
14695 my $pair_width = ( $max_length[0] + $max_length[1] );
14697 ( $max_length[0] > $max_length[1] ) ? $max_length[0] : $max_length[1];
14699 # Number of free columns across the page width for laying out tables
14700 my $columns = table_columns_available($i_first_comma);
14702 # Estimated maximum number of fields which fit this space
14703 # This will be our first guess
14704 my $number_of_fields_max =
14705 maximum_number_of_fields( $columns, $odd_or_even, $max_width,
14707 my $number_of_fields = $number_of_fields_max;
14709 # Find the best-looking number of fields
14710 # and make this our second guess if possible
14711 my ( $number_of_fields_best, $ri_ragged_break_list,
14712 $new_identifier_count )
14713 = study_list_complexity( \@i_term_begin, \@i_term_end, \@item_lengths,
14716 if ( $number_of_fields_best != 0
14717 && $number_of_fields_best < $number_of_fields_max )
14719 $number_of_fields = $number_of_fields_best;
14722 # ----------------------------------------------------------------------
14723 # If we are crowded and the -lp option is being used, try to
14724 # undo some indentation
14725 # ----------------------------------------------------------------------
14727 $rOpts_line_up_parentheses
14729 $number_of_fields == 0
14730 || ( $number_of_fields == 1
14731 && $number_of_fields != $number_of_fields_best )
14735 my $available_spaces = get_AVAILABLE_SPACES_to_go($i_first_comma);
14736 if ( $available_spaces > 0 ) {
14738 my $spaces_wanted = $max_width - $columns; # for 1 field
14740 if ( $number_of_fields_best == 0 ) {
14741 $number_of_fields_best =
14742 get_maximum_fields_wanted( \@item_lengths );
14745 if ( $number_of_fields_best != 1 ) {
14746 my $spaces_wanted_2 =
14747 1 + $pair_width - $columns; # for 2 fields
14748 if ( $available_spaces > $spaces_wanted_2 ) {
14749 $spaces_wanted = $spaces_wanted_2;
14753 if ( $spaces_wanted > 0 ) {
14754 my $deleted_spaces =
14755 reduce_lp_indentation( $i_first_comma, $spaces_wanted );
14758 if ( $deleted_spaces > 0 ) {
14759 $columns = table_columns_available($i_first_comma);
14760 $number_of_fields_max =
14761 maximum_number_of_fields( $columns, $odd_or_even,
14762 $max_width, $pair_width );
14763 $number_of_fields = $number_of_fields_max;
14765 if ( $number_of_fields_best == 1
14766 && $number_of_fields >= 1 )
14768 $number_of_fields = $number_of_fields_best;
14775 # try for one column if two won't work
14776 if ( $number_of_fields <= 0 ) {
14777 $number_of_fields = int( $columns / $max_width );
14780 # The user can place an upper bound on the number of fields,
14781 # which can be useful for doing maintenance on tables
14782 if ( $rOpts_maximum_fields_per_table
14783 && $number_of_fields > $rOpts_maximum_fields_per_table )
14785 $number_of_fields = $rOpts_maximum_fields_per_table;
14788 # How many columns (characters) and lines would this container take
14789 # if no additional whitespace were added?
14790 my $packed_columns = token_sequence_length( $i_opening_paren + 1,
14791 $i_effective_last_comma + 1 );
14792 if ( $columns <= 0 ) { $columns = 1 } # avoid divide by zero
14793 my $packed_lines = 1 + int( $packed_columns / $columns );
14795 # are we an item contained in an outer list?
14796 my $in_hierarchical_list = $next_nonblank_type =~ /^[\}\,]$/;
14798 if ( $number_of_fields <= 0 ) {
14800 # #---------------------------------------------------------------
14801 # # We're in trouble. We can't find a single field width that works.
14802 # # There is no simple answer here; we may have a single long list
14804 # #---------------------------------------------------------------
14806 # In many cases, it may be best to not force a break if there is just one
14807 # comma, because the standard continuation break logic will do a better
14810 # In the common case that all but one of the terms can fit
14811 # on a single line, it may look better not to break open the
14812 # containing parens. Consider, for example
14816 # sort { $color_value{$::a} <=> $color_value{$::b}; }
14819 # which will look like this with the container broken:
14823 # sort { $color_value{$::a} <=> $color_value{$::b}; } keys %colors
14826 # Here is an example of this rule for a long last term:
14828 # log_message( 0, 256, 128,
14829 # "Number of routes in adj-RIB-in to be considered: $peercount" );
14831 # And here is an example with a long first term:
14834 # "%2d wallclock secs (%$f usr %$f sys + %$f cusr %$f csys = %$f CPU)",
14835 # $r, $pu, $ps, $cu, $cs, $tt
14837 # if $style eq 'all';
14839 my $i_last_comma = $$rcomma_index[ $comma_count - 1 ];
14840 my $long_last_term = excess_line_length( 0, $i_last_comma ) <= 0;
14841 my $long_first_term =
14842 excess_line_length( $i_first_comma + 1, $max_index_to_go ) <= 0;
14844 # break at every comma ...
14847 # if requested by user or is best looking
14848 $number_of_fields_best == 1
14850 # or if this is a sublist of a larger list
14851 || $in_hierarchical_list
14853 # or if multiple commas and we dont have a long first or last
14855 || ( $comma_count > 1
14856 && !( $long_last_term || $long_first_term ) )
14859 foreach ( 0 .. $comma_count - 1 ) {
14860 set_forced_breakpoint( $$rcomma_index[$_] );
14863 elsif ($long_last_term) {
14865 set_forced_breakpoint($i_last_comma);
14866 $$rdo_not_break_apart = 1 unless $must_break_open;
14868 elsif ($long_first_term) {
14870 set_forced_breakpoint($i_first_comma);
14874 # let breaks be defined by default bond strength logic
14879 # --------------------------------------------------------
14880 # We have a tentative field count that seems to work.
14881 # How many lines will this require?
14882 # --------------------------------------------------------
14883 my $formatted_lines = $item_count / ($number_of_fields);
14884 if ( $formatted_lines != int $formatted_lines ) {
14885 $formatted_lines = 1 + int $formatted_lines;
14888 # So far we've been trying to fill out to the right margin. But
14889 # compact tables are easier to read, so let's see if we can use fewer
14890 # fields without increasing the number of lines.
14891 $number_of_fields =
14892 compactify_table( $item_count, $number_of_fields, $formatted_lines,
14895 # How many spaces across the page will we fill?
14896 my $columns_per_line =
14897 ( int $number_of_fields / 2 ) * $pair_width +
14898 ( $number_of_fields % 2 ) * $max_width;
14900 my $formatted_columns;
14902 if ( $number_of_fields > 1 ) {
14903 $formatted_columns =
14904 ( $pair_width * ( int( $item_count / 2 ) ) +
14905 ( $item_count % 2 ) * $max_width );
14908 $formatted_columns = $max_width * $item_count;
14910 if ( $formatted_columns < $packed_columns ) {
14911 $formatted_columns = $packed_columns;
14914 my $unused_columns = $formatted_columns - $packed_columns;
14916 # set some empirical parameters to help decide if we should try to
14917 # align; high sparsity does not look good, especially with few lines
14918 my $sparsity = ($unused_columns) / ($formatted_columns);
14919 my $max_allowed_sparsity =
14920 ( $item_count < 3 ) ? 0.1
14921 : ( $packed_lines == 1 ) ? 0.15
14922 : ( $packed_lines == 2 ) ? 0.4
14925 # Begin check for shortcut methods, which avoid treating a list
14926 # as a table for relatively small parenthesized lists. These
14927 # are usually easier to read if not formatted as tables.
14929 $packed_lines <= 2 # probably can fit in 2 lines
14930 && $item_count < 9 # doesn't have too many items
14931 && $opening_environment eq 'BLOCK' # not a sub-container
14932 && $opening_token eq '(' # is paren list
14936 # Shortcut method 1: for -lp and just one comma:
14937 # This is a no-brainer, just break at the comma.
14939 $rOpts_line_up_parentheses # -lp
14940 && $item_count == 2 # two items, one comma
14941 && !$must_break_open
14944 my $i_break = $$rcomma_index[0];
14945 set_forced_breakpoint($i_break);
14946 $$rdo_not_break_apart = 1;
14947 set_non_alignment_flags( $comma_count, $rcomma_index );
14952 # method 2 is for most small ragged lists which might look
14953 # best if not displayed as a table.
14955 ( $number_of_fields == 2 && $item_count == 3 )
14957 $new_identifier_count > 0 # isn't all quotes
14958 && $sparsity > 0.15
14959 ) # would be fairly spaced gaps if aligned
14963 my $break_count = set_ragged_breakpoints( \@i_term_comma,
14964 $ri_ragged_break_list );
14965 ++$break_count if ($use_separate_first_term);
14967 # NOTE: we should really use the true break count here,
14968 # which can be greater if there are large terms and
14969 # little space, but usually this will work well enough.
14970 unless ($must_break_open) {
14972 if ( $break_count <= 1 ) {
14973 $$rdo_not_break_apart = 1;
14975 elsif ( $rOpts_line_up_parentheses && !$need_lp_break_open )
14977 $$rdo_not_break_apart = 1;
14980 set_non_alignment_flags( $comma_count, $rcomma_index );
14984 } # end shortcut methods
14988 FORMATTER_DEBUG_FLAG_SPARSE && do {
14990 "SPARSE:cols=$columns commas=$comma_count items:$item_count ids=$identifier_count pairwidth=$pair_width fields=$number_of_fields lines packed: $packed_lines packed_cols=$packed_columns fmtd:$formatted_lines cols /line:$columns_per_line unused:$unused_columns fmtd:$formatted_columns sparsity=$sparsity allow=$max_allowed_sparsity\n";
14994 #---------------------------------------------------------------
14995 # Compound List Rule 2:
14996 # If this list is too long for one line, and it is an item of a
14997 # larger list, then we must format it, regardless of sparsity
14998 # (ian.t). One reason that we have to do this is to trigger
14999 # Compound List Rule 1, above, which causes breaks at all commas of
15000 # all outer lists. In this way, the structure will be properly
15002 #---------------------------------------------------------------
15004 # Decide if this list is too long for one line unless broken
15005 my $total_columns = table_columns_available($i_opening_paren);
15006 my $too_long = $packed_columns > $total_columns;
15008 # For a paren list, include the length of the token just before the
15009 # '(' because this is likely a sub call, and we would have to
15010 # include the sub name on the same line as the list. This is still
15011 # imprecise, but not too bad. (steve.t)
15012 if ( !$too_long && $i_opening_paren > 0 && $opening_token eq '(' ) {
15014 $too_long = excess_line_length( $i_opening_minus,
15015 $i_effective_last_comma + 1 ) > 0;
15018 # FIXME: For an item after a '=>', try to include the length of the
15019 # thing before the '=>'. This is crude and should be improved by
15020 # actually looking back token by token.
15021 if ( !$too_long && $i_opening_paren > 0 && $list_type eq '=>' ) {
15022 my $i_opening_minus = $i_opening_paren - 4;
15023 if ( $i_opening_minus >= 0 ) {
15024 $too_long = excess_line_length( $i_opening_minus,
15025 $i_effective_last_comma + 1 ) > 0;
15029 # Always break lists contained in '[' and '{' if too long for 1 line,
15030 # and always break lists which are too long and part of a more complex
15032 my $must_break_open_container = $must_break_open
15034 && ( $in_hierarchical_list || $opening_token ne '(' ) );
15036 #print "LISTX: next=$next_nonblank_type avail cols=$columns packed=$packed_columns must format = $must_break_open_container too-long=$too_long opening=$opening_token list_type=$list_type formatted_lines=$formatted_lines packed=$packed_lines max_sparsity= $max_allowed_sparsity sparsity=$sparsity \n";
15038 #---------------------------------------------------------------
15039 # The main decision:
15040 # Now decide if we will align the data into aligned columns. Do not
15041 # attempt to align columns if this is a tiny table or it would be
15042 # too spaced. It seems that the more packed lines we have, the
15043 # sparser the list that can be allowed and still look ok.
15044 #---------------------------------------------------------------
15046 if ( ( $formatted_lines < 3 && $packed_lines < $formatted_lines )
15047 || ( $formatted_lines < 2 )
15048 || ( $unused_columns > $max_allowed_sparsity * $formatted_columns )
15052 #---------------------------------------------------------------
15053 # too sparse: would look ugly if aligned in a table;
15054 #---------------------------------------------------------------
15056 # use old breakpoints if this is a 'big' list
15057 # FIXME: goal is to improve set_ragged_breakpoints so that
15058 # this is not necessary.
15059 if ( $packed_lines > 2 && $item_count > 10 ) {
15060 write_logfile_entry("List sparse: using old breakpoints\n");
15061 copy_old_breakpoints( $i_first_comma, $i_last_comma );
15064 # let the continuation logic handle it if 2 lines
15067 my $break_count = set_ragged_breakpoints( \@i_term_comma,
15068 $ri_ragged_break_list );
15069 ++$break_count if ($use_separate_first_term);
15071 unless ($must_break_open_container) {
15072 if ( $break_count <= 1 ) {
15073 $$rdo_not_break_apart = 1;
15075 elsif ( $rOpts_line_up_parentheses && !$need_lp_break_open )
15077 $$rdo_not_break_apart = 1;
15080 set_non_alignment_flags( $comma_count, $rcomma_index );
15085 #---------------------------------------------------------------
15086 # go ahead and format as a table
15087 #---------------------------------------------------------------
15088 write_logfile_entry(
15089 "List: auto formatting with $number_of_fields fields/row\n");
15091 my $j_first_break =
15092 $use_separate_first_term ? $number_of_fields : $number_of_fields - 1;
15095 my $j = $j_first_break ;
15096 $j < $comma_count ;
15097 $j += $number_of_fields
15100 my $i = $$rcomma_index[$j];
15101 set_forced_breakpoint($i);
15107 sub set_non_alignment_flags {
15109 # set flag which indicates that these commas should not be
15111 my ( $comma_count, $rcomma_index ) = @_;
15112 foreach ( 0 .. $comma_count - 1 ) {
15113 $matching_token_to_go[ $$rcomma_index[$_] ] = 1;
15117 sub study_list_complexity {
15119 # Look for complex tables which should be formatted with one term per line.
15120 # Returns the following:
15122 # \@i_ragged_break_list = list of good breakpoints to avoid lines
15123 # which are hard to read
15124 # $number_of_fields_best = suggested number of fields based on
15125 # complexity; = 0 if any number may be used.
15127 my ( $ri_term_begin, $ri_term_end, $ritem_lengths, $max_width ) = @_;
15128 my $item_count = @{$ri_term_begin};
15129 my $complex_item_count = 0;
15130 my $number_of_fields_best = $rOpts_maximum_fields_per_table;
15131 my $i_max = @{$ritem_lengths} - 1;
15132 ##my @item_complexity;
15134 my $i_last_last_break = -3;
15135 my $i_last_break = -2;
15136 my @i_ragged_break_list;
15138 my $definitely_complex = 30;
15139 my $definitely_simple = 12;
15140 my $quote_count = 0;
15142 for my $i ( 0 .. $i_max ) {
15143 my $ib = $ri_term_begin->[$i];
15144 my $ie = $ri_term_end->[$i];
15146 # define complexity: start with the actual term length
15147 my $weighted_length = ( $ritem_lengths->[$i] - 2 );
15149 ##TBD: join types here and check for variations
15150 ##my $str=join "", @tokens_to_go[$ib..$ie];
15153 if ( $types_to_go[$ib] =~ /^[qQ]$/ ) {
15157 elsif ( $types_to_go[$ib] =~ /^[w\-]$/ ) {
15161 if ( $ib eq $ie ) {
15162 if ( $is_quote && $tokens_to_go[$ib] =~ /\s/ ) {
15163 $complex_item_count++;
15164 $weighted_length *= 2;
15170 if ( grep { $_ eq 'b' } @types_to_go[ $ib .. $ie ] ) {
15171 $complex_item_count++;
15172 $weighted_length *= 2;
15174 if ( grep { $_ eq '..' } @types_to_go[ $ib .. $ie ] ) {
15175 $weighted_length += 4;
15179 # add weight for extra tokens.
15180 $weighted_length += 2 * ( $ie - $ib );
15182 ## my $BUB = join '', @tokens_to_go[$ib..$ie];
15183 ## print "# COMPLEXITY:$weighted_length $BUB\n";
15185 ##push @item_complexity, $weighted_length;
15187 # now mark a ragged break after this item it if it is 'long and
15189 if ( $weighted_length >= $definitely_complex ) {
15191 # if we broke after the previous term
15192 # then break before it too
15193 if ( $i_last_break == $i - 1
15195 && $i_last_last_break != $i - 2 )
15198 ## FIXME: don't strand a small term
15199 pop @i_ragged_break_list;
15200 push @i_ragged_break_list, $i - 2;
15201 push @i_ragged_break_list, $i - 1;
15204 push @i_ragged_break_list, $i;
15205 $i_last_last_break = $i_last_break;
15206 $i_last_break = $i;
15209 # don't break before a small last term -- it will
15210 # not look good on a line by itself.
15211 elsif ($i == $i_max
15212 && $i_last_break == $i - 1
15213 && $weighted_length <= $definitely_simple )
15215 pop @i_ragged_break_list;
15219 my $identifier_count = $i_max + 1 - $quote_count;
15221 # Need more tuning here..
15222 if ( $max_width > 12
15223 && $complex_item_count > $item_count / 2
15224 && $number_of_fields_best != 2 )
15226 $number_of_fields_best = 1;
15229 return ( $number_of_fields_best, \@i_ragged_break_list, $identifier_count );
15232 sub get_maximum_fields_wanted {
15234 # Not all tables look good with more than one field of items.
15235 # This routine looks at a table and decides if it should be
15236 # formatted with just one field or not.
15237 # This coding is still under development.
15238 my ($ritem_lengths) = @_;
15240 my $number_of_fields_best = 0;
15242 # For just a few items, we tentatively assume just 1 field.
15243 my $item_count = @{$ritem_lengths};
15244 if ( $item_count <= 5 ) {
15245 $number_of_fields_best = 1;
15248 # For larger tables, look at it both ways and see what looks best
15252 my @max_length = ( 0, 0 );
15253 my @last_length_2 = ( undef, undef );
15254 my @first_length_2 = ( undef, undef );
15255 my $last_length = undef;
15256 my $total_variation_1 = 0;
15257 my $total_variation_2 = 0;
15258 my @total_variation_2 = ( 0, 0 );
15259 for ( my $j = 0 ; $j < $item_count ; $j++ ) {
15261 $is_odd = 1 - $is_odd;
15262 my $length = $ritem_lengths->[$j];
15263 if ( $length > $max_length[$is_odd] ) {
15264 $max_length[$is_odd] = $length;
15267 if ( defined($last_length) ) {
15268 my $dl = abs( $length - $last_length );
15269 $total_variation_1 += $dl;
15271 $last_length = $length;
15273 my $ll = $last_length_2[$is_odd];
15274 if ( defined($ll) ) {
15275 my $dl = abs( $length - $ll );
15276 $total_variation_2[$is_odd] += $dl;
15279 $first_length_2[$is_odd] = $length;
15281 $last_length_2[$is_odd] = $length;
15283 $total_variation_2 = $total_variation_2[0] + $total_variation_2[1];
15285 my $factor = ( $item_count > 10 ) ? 1 : ( $item_count > 5 ) ? 0.75 : 0;
15286 unless ( $total_variation_2 < $factor * $total_variation_1 ) {
15287 $number_of_fields_best = 1;
15290 return ($number_of_fields_best);
15293 sub table_columns_available {
15294 my $i_first_comma = shift;
15296 $rOpts_maximum_line_length - leading_spaces_to_go($i_first_comma);
15298 # Patch: the vertical formatter does not line up lines whose lengths
15299 # exactly equal the available line length because of allowances
15300 # that must be made for side comments. Therefore, the number of
15301 # available columns is reduced by 1 character.
15306 sub maximum_number_of_fields {
15308 # how many fields will fit in the available space?
15309 my ( $columns, $odd_or_even, $max_width, $pair_width ) = @_;
15310 my $max_pairs = int( $columns / $pair_width );
15311 my $number_of_fields = $max_pairs * 2;
15312 if ( $odd_or_even == 1
15313 && $max_pairs * $pair_width + $max_width <= $columns )
15315 $number_of_fields++;
15317 return $number_of_fields;
15320 sub compactify_table {
15322 # given a table with a certain number of fields and a certain number
15323 # of lines, see if reducing the number of fields will make it look
15325 my ( $item_count, $number_of_fields, $formatted_lines, $odd_or_even ) = @_;
15326 if ( $number_of_fields >= $odd_or_even * 2 && $formatted_lines > 0 ) {
15330 $min_fields = $number_of_fields ;
15331 $min_fields >= $odd_or_even
15332 && $min_fields * $formatted_lines >= $item_count ;
15333 $min_fields -= $odd_or_even
15336 $number_of_fields = $min_fields;
15339 return $number_of_fields;
15342 sub set_ragged_breakpoints {
15344 # Set breakpoints in a list that cannot be formatted nicely as a
15346 my ( $ri_term_comma, $ri_ragged_break_list ) = @_;
15348 my $break_count = 0;
15349 foreach (@$ri_ragged_break_list) {
15350 my $j = $ri_term_comma->[$_];
15352 set_forced_breakpoint($j);
15356 return $break_count;
15359 sub copy_old_breakpoints {
15360 my ( $i_first_comma, $i_last_comma ) = @_;
15361 for my $i ( $i_first_comma .. $i_last_comma ) {
15362 if ( $old_breakpoint_to_go[$i] ) {
15363 set_forced_breakpoint($i);
15369 my ( $i, $j ) = @_;
15370 if ( $i >= 0 && $i <= $j && $j <= $max_index_to_go ) {
15372 FORMATTER_DEBUG_FLAG_NOBREAK && do {
15373 my ( $a, $b, $c ) = caller();
15375 "NOBREAK: forced_breakpoint $forced_breakpoint_count from $a $c with i=$i max=$max_index_to_go type=$types_to_go[$i]\n"
15379 @nobreak_to_go[ $i .. $j ] = (1) x ( $j - $i + 1 );
15382 # shouldn't happen; non-critical error
15384 FORMATTER_DEBUG_FLAG_NOBREAK && do {
15385 my ( $a, $b, $c ) = caller();
15387 "NOBREAK ERROR: from $a $c with i=$i j=$j max=$max_index_to_go\n"
15393 sub set_fake_breakpoint {
15395 # Just bump up the breakpoint count as a signal that there are breaks.
15396 # This is useful if we have breaks but may want to postpone deciding where
15398 $forced_breakpoint_count++;
15401 sub set_forced_breakpoint {
15404 return unless defined $i && $i >= 0;
15406 # when called with certain tokens, use bond strengths to decide
15407 # if we break before or after it
15408 my $token = $tokens_to_go[$i];
15410 if ( $token =~ /^([\=\.\,\:\?]|and|or|xor|&&|\|\|)$/ ) {
15411 if ( $want_break_before{$token} && $i >= 0 ) { $i-- }
15414 # breaks are forced before 'if' and 'unless'
15415 elsif ( $is_if_unless{$token} ) { $i-- }
15417 if ( $i >= 0 && $i <= $max_index_to_go ) {
15418 my $i_nonblank = ( $types_to_go[$i] ne 'b' ) ? $i : $i - 1;
15420 FORMATTER_DEBUG_FLAG_FORCE && do {
15421 my ( $a, $b, $c ) = caller();
15423 "FORCE forced_breakpoint $forced_breakpoint_count from $a $c with i=$i_nonblank max=$max_index_to_go tok=$tokens_to_go[$i_nonblank] type=$types_to_go[$i_nonblank] nobr=$nobreak_to_go[$i_nonblank]\n";
15426 if ( $i_nonblank >= 0 && $nobreak_to_go[$i_nonblank] == 0 ) {
15427 $forced_breakpoint_to_go[$i_nonblank] = 1;
15429 if ( $i_nonblank > $index_max_forced_break ) {
15430 $index_max_forced_break = $i_nonblank;
15432 $forced_breakpoint_count++;
15433 $forced_breakpoint_undo_stack[ $forced_breakpoint_undo_count++ ] =
15436 # if we break at an opening container..break at the closing
15437 if ( $tokens_to_go[$i_nonblank] =~ /^[\{\[\(\?]$/ ) {
15438 set_closing_breakpoint($i_nonblank);
15444 sub clear_breakpoint_undo_stack {
15445 $forced_breakpoint_undo_count = 0;
15448 sub undo_forced_breakpoint_stack {
15450 my $i_start = shift;
15451 if ( $i_start < 0 ) {
15453 my ( $a, $b, $c ) = caller();
15455 "Program Bug: undo_forced_breakpoint_stack from $a $c has i=$i_start "
15459 while ( $forced_breakpoint_undo_count > $i_start ) {
15461 $forced_breakpoint_undo_stack[ --$forced_breakpoint_undo_count ];
15462 if ( $i >= 0 && $i <= $max_index_to_go ) {
15463 $forced_breakpoint_to_go[$i] = 0;
15464 $forced_breakpoint_count--;
15466 FORMATTER_DEBUG_FLAG_UNDOBP && do {
15467 my ( $a, $b, $c ) = caller();
15469 "UNDOBP: undo forced_breakpoint i=$i $forced_breakpoint_undo_count from $a $c max=$max_index_to_go\n"
15474 # shouldn't happen, but not a critical error
15476 FORMATTER_DEBUG_FLAG_UNDOBP && do {
15477 my ( $a, $b, $c ) = caller();
15479 "Program Bug: undo_forced_breakpoint from $a $c has i=$i but max=$max_index_to_go"
15486 { # begin recombine_breakpoints
15495 @is_amp_amp{@_} = (1) x scalar(@_);
15498 @is_ternary{@_} = (1) x scalar(@_);
15500 @_ = qw( + - * / );
15501 @is_math_op{@_} = (1) x scalar(@_);
15504 sub recombine_breakpoints {
15506 # sub set_continuation_breaks is very liberal in setting line breaks
15507 # for long lines, always setting breaks at good breakpoints, even
15508 # when that creates small lines. Occasionally small line fragments
15509 # are produced which would look better if they were combined.
15510 # That's the task of this routine, recombine_breakpoints.
15512 # $ri_beg = ref to array of BEGinning indexes of each line
15513 # $ri_end = ref to array of ENDing indexes of each line
15514 my ( $ri_beg, $ri_end ) = @_;
15516 my $more_to_do = 1;
15518 # We keep looping over all of the lines of this batch
15519 # until there are no more possible recombinations
15520 my $nmax_last = @$ri_end;
15521 while ($more_to_do) {
15525 my $nmax = @$ri_end - 1;
15527 # safety check for infinite loop
15528 unless ( $nmax < $nmax_last ) {
15530 # shouldn't happen because splice below decreases nmax on each pass:
15531 # but i get paranoid sometimes
15532 die "Program bug-infinite loop in recombine breakpoints\n";
15534 $nmax_last = $nmax;
15536 my $previous_outdentable_closing_paren;
15537 my $leading_amp_count = 0;
15538 my $this_line_is_semicolon_terminated;
15540 # loop over all remaining lines in this batch
15541 for $n ( 1 .. $nmax ) {
15543 #----------------------------------------------------------
15544 # If we join the current pair of lines,
15545 # line $n-1 will become the left part of the joined line
15546 # line $n will become the right part of the joined line
15548 # Here are Indexes of the endpoint tokens of the two lines:
15550 # -----line $n-1--- | -----line $n-----
15551 # $ibeg_1 $iend_1 | $ibeg_2 $iend_2
15554 # We want to decide if we should remove the line break
15555 # betwen the tokens at $iend_1 and $ibeg_2
15557 # We will apply a number of ad-hoc tests to see if joining
15558 # here will look ok. The code will just issue a 'next'
15559 # command if the join doesn't look good. If we get through
15560 # the gauntlet of tests, the lines will be recombined.
15561 #----------------------------------------------------------
15563 # beginning and ending tokens of the lines we are working on
15564 my $ibeg_1 = $$ri_beg[ $n - 1 ];
15565 my $iend_1 = $$ri_end[ $n - 1 ];
15566 my $iend_2 = $$ri_end[$n];
15567 my $ibeg_2 = $$ri_beg[$n];
15569 my $ibeg_nmax = $$ri_beg[$nmax];
15571 # some beginning indexes of other lines, which may not exist
15572 my $ibeg_0 = $n > 1 ? $$ri_beg[ $n - 2 ] : -1;
15573 my $ibeg_3 = $n < $nmax ? $$ri_beg[ $n + 1 ] : -1;
15574 my $ibeg_4 = $n + 2 <= $nmax ? $$ri_beg[ $n + 2 ] : -1;
15578 #my $depth_increase=( $nesting_depth_to_go[$ibeg_2] -
15579 # $nesting_depth_to_go[$ibeg_1] );
15581 ##print "RECOMBINE: n=$n imid=$iend_1 if=$ibeg_1 type=$types_to_go[$ibeg_1] =$tokens_to_go[$ibeg_1] next_type=$types_to_go[$ibeg_2] next_tok=$tokens_to_go[$ibeg_2]\n";
15583 # If line $n is the last line, we set some flags and
15584 # do any special checks for it
15585 if ( $n == $nmax ) {
15587 # a terminal '{' should stay where it is
15588 next if $types_to_go[$ibeg_2] eq '{';
15590 # set flag if statement $n ends in ';'
15591 $this_line_is_semicolon_terminated =
15592 $types_to_go[$iend_2] eq ';'
15594 # with possible side comment
15595 || ( $types_to_go[$iend_2] eq '#'
15596 && $iend_2 - $ibeg_2 >= 2
15597 && $types_to_go[ $iend_2 - 2 ] eq ';'
15598 && $types_to_go[ $iend_2 - 1 ] eq 'b' );
15601 #----------------------------------------------------------
15602 # Section 1: examine token at $iend_1 (right end of first line
15604 #----------------------------------------------------------
15606 # an isolated '}' may join with a ';' terminated segment
15607 if ( $types_to_go[$iend_1] eq '}' ) {
15609 # Check for cases where combining a semicolon terminated
15610 # statement with a previous isolated closing paren will
15611 # allow the combined line to be outdented. This is
15612 # generally a good move. For example, we can join up
15613 # the last two lines here:
15615 # $dev, $ino, $mode, $nlink, $uid, $gid, $rdev,
15616 # $size, $atime, $mtime, $ctime, $blksize, $blocks
15622 # $dev, $ino, $mode, $nlink, $uid, $gid, $rdev,
15623 # $size, $atime, $mtime, $ctime, $blksize, $blocks
15626 # which makes the parens line up.
15628 # Another example, from Joe Matarazzo, probably looks best
15629 # with the 'or' clause appended to the trailing paren:
15630 # $self->some_method(
15633 # ) or die "Some_method didn't work";
15635 $previous_outdentable_closing_paren =
15636 $this_line_is_semicolon_terminated # ends in ';'
15637 && $ibeg_1 == $iend_1 # only one token on last line
15638 && $tokens_to_go[$iend_1] eq
15639 ')' # must be structural paren
15641 # only &&, ||, and : if no others seen
15642 # (but note: our count made below could be wrong
15643 # due to intervening comments)
15644 && ( $leading_amp_count == 0
15645 || $types_to_go[$ibeg_2] !~ /^(:|\&\&|\|\|)$/ )
15647 # but leading colons probably line up with with a
15648 # previous colon or question (count could be wrong).
15649 && $types_to_go[$ibeg_2] ne ':'
15651 # only one step in depth allowed. this line must not
15652 # begin with a ')' itself.
15653 && ( $nesting_depth_to_go[$iend_1] ==
15654 $nesting_depth_to_go[$iend_2] + 1 );
15658 $previous_outdentable_closing_paren
15660 # handle '.' and '?' specially below
15661 || ( $types_to_go[$ibeg_2] =~ /^[\.\?]$/ )
15665 # do not recombine lines with ending &&, ||,
15666 elsif ( $is_amp_amp{ $types_to_go[$iend_1] } ) {
15667 next unless $want_break_before{ $types_to_go[$iend_1] };
15670 # keep a terminal colon
15671 elsif ( $types_to_go[$iend_1] eq ':' ) {
15672 next unless $want_break_before{ $types_to_go[$iend_1] };
15675 # Identify and recombine a broken ?/: chain
15676 elsif ( $types_to_go[$iend_1] eq '?' ) {
15678 # Do not recombine different levels
15680 if ( $levels_to_go[$ibeg_1] ne $levels_to_go[$ibeg_2] );
15682 # do not recombine unless next line ends in :
15683 next unless $types_to_go[$iend_2] eq ':';
15686 # for lines ending in a comma...
15687 elsif ( $types_to_go[$iend_1] eq ',' ) {
15689 # Do not recombine at comma which is following the
15691 # TODO: might be best to make a special flag
15692 next if ( $old_breakpoint_to_go[$iend_1] );
15694 # an isolated '},' may join with an identifier + ';'
15695 # this is useful for the class of a 'bless' statement (bless.t)
15696 if ( $types_to_go[$ibeg_1] eq '}'
15697 && $types_to_go[$ibeg_2] eq 'i' )
15700 unless ( ( $ibeg_1 == ( $iend_1 - 1 ) )
15701 && ( $iend_2 == ( $ibeg_2 + 1 ) )
15702 && $this_line_is_semicolon_terminated );
15704 # override breakpoint
15705 $forced_breakpoint_to_go[$iend_1] = 0;
15711 # do not recombine after a comma unless this will leave
15713 next unless ( $n + 1 >= $nmax );
15715 # do not recombine if there is a change in indentation depth
15718 $levels_to_go[$iend_1] != $levels_to_go[$iend_2] );
15720 # do not recombine a "complex expression" after a
15721 # comma. "complex" means no parens.
15723 foreach my $ii ( $ibeg_2 .. $iend_2 ) {
15724 if ( $tokens_to_go[$ii] eq '(' ) {
15729 next if $saw_paren;
15734 elsif ( $types_to_go[$iend_1] eq '(' ) {
15736 # No longer doing this
15739 elsif ( $types_to_go[$iend_1] eq ')' ) {
15741 # No longer doing this
15744 # keep a terminal for-semicolon
15745 elsif ( $types_to_go[$iend_1] eq 'f' ) {
15749 # if '=' at end of line ...
15750 elsif ( $is_assignment{ $types_to_go[$iend_1] } ) {
15752 my $is_short_quote =
15753 ( $types_to_go[$ibeg_2] eq 'Q'
15754 && $ibeg_2 == $iend_2
15755 && length( $tokens_to_go[$ibeg_2] ) <
15756 $rOpts_short_concatenation_item_length );
15758 ( $types_to_go[$ibeg_1] eq '?'
15759 && ( $ibeg_3 >= 0 && $types_to_go[$ibeg_3] eq ':' ) );
15761 # always join an isolated '=', a short quote, or if this
15762 # will put ?/: at start of adjacent lines
15763 if ( $ibeg_1 != $iend_1
15764 && !$is_short_quote
15771 # unless we can reduce this to two lines
15774 # or three lines, the last with a leading semicolon
15775 || ( $nmax == $n + 2
15776 && $types_to_go[$ibeg_nmax] eq ';' )
15778 # or the next line ends with a here doc
15779 || $types_to_go[$iend_2] eq 'h'
15781 # or the next line ends in an open paren or brace
15782 # and the break hasn't been forced [dima.t]
15783 || ( !$forced_breakpoint_to_go[$iend_1]
15784 && $types_to_go[$iend_2] eq '{' )
15787 # do not recombine if the two lines might align well
15788 # this is a very approximate test for this
15790 && $types_to_go[$ibeg_2] ne
15791 $types_to_go[$ibeg_3] )
15794 # -lp users often prefer this:
15795 # my $title = function($env, $env, $sysarea,
15796 # "bubba Borrower Entry");
15797 # so we will recombine if -lp is used we have ending
15799 if ( !$rOpts_line_up_parentheses
15800 || $types_to_go[$iend_2] ne ',' )
15803 # otherwise, scan the rhs line up to last token for
15804 # complexity. Note that we are not counting the last
15805 # token in case it is an opening paren.
15807 my $depth = $nesting_depth_to_go[$ibeg_2];
15808 for ( my $i = $ibeg_2 + 1 ; $i < $iend_2 ; $i++ ) {
15809 if ( $nesting_depth_to_go[$i] != $depth ) {
15811 last if ( $tv > 1 );
15813 $depth = $nesting_depth_to_go[$i];
15816 # ok to recombine if no level changes before last token
15819 # otherwise, do not recombine if more than two
15821 next if ( $tv > 1 );
15823 # check total complexity of the two adjacent lines
15824 # that will occur if we do this join
15826 ( $n < $nmax ) ? $$ri_end[ $n + 1 ] : $iend_2;
15827 for ( my $i = $iend_2 ; $i <= $istop ; $i++ ) {
15828 if ( $nesting_depth_to_go[$i] != $depth ) {
15830 last if ( $tv > 2 );
15832 $depth = $nesting_depth_to_go[$i];
15835 # do not recombine if total is more than 2 level changes
15836 next if ( $tv > 2 );
15841 unless ( $tokens_to_go[$ibeg_2] =~ /^[\{\(\[]$/ ) {
15842 $forced_breakpoint_to_go[$iend_1] = 0;
15847 elsif ( $types_to_go[$iend_1] eq 'k' ) {
15849 # make major control keywords stand out
15854 #/^(last|next|redo|return)$/
15855 $is_last_next_redo_return{ $tokens_to_go[$iend_1] }
15857 # but only if followed by multiple lines
15861 if ( $is_and_or{ $tokens_to_go[$iend_1] } ) {
15863 unless $want_break_before{ $tokens_to_go[$iend_1] };
15867 # handle trailing + - * /
15868 elsif ( $is_math_op{ $types_to_go[$iend_1] } ) {
15870 # combine lines if next line has single number
15871 # or a short term followed by same operator
15872 my $i_next_nonblank = $ibeg_2;
15873 my $i_next_next = $i_next_nonblank + 1;
15874 $i_next_next++ if ( $types_to_go[$i_next_next] eq 'b' );
15875 my $number_follows = $types_to_go[$i_next_nonblank] eq 'n'
15877 $i_next_nonblank == $iend_2
15878 || ( $i_next_next == $iend_2
15879 && $is_math_op{ $types_to_go[$i_next_next] } )
15880 || $types_to_go[$i_next_next] eq ';'
15883 # find token before last operator of previous line
15884 my $iend_1_minus = $iend_1;
15886 if ( $iend_1_minus > $ibeg_1 );
15888 if ( $types_to_go[$iend_1_minus] eq 'b'
15889 && $iend_1_minus > $ibeg_1 );
15891 my $short_term_follows =
15892 ( $types_to_go[$iend_2] eq $types_to_go[$iend_1]
15893 && $types_to_go[$iend_1_minus] =~ /^[in]$/
15894 && $iend_2 <= $ibeg_2 + 2
15895 && length( $tokens_to_go[$ibeg_2] ) <
15896 $rOpts_short_concatenation_item_length );
15899 unless ( $number_follows || $short_term_follows );
15902 #----------------------------------------------------------
15903 # Section 2: Now examine token at $ibeg_2 (left end of second
15905 #----------------------------------------------------------
15907 # join lines identified above as capable of
15908 # causing an outdented line with leading closing paren
15909 if ($previous_outdentable_closing_paren) {
15910 $forced_breakpoint_to_go[$iend_1] = 0;
15913 # do not recombine lines with leading :
15914 elsif ( $types_to_go[$ibeg_2] eq ':' ) {
15915 $leading_amp_count++;
15916 next if $want_break_before{ $types_to_go[$ibeg_2] };
15919 # handle lines with leading &&, ||
15920 elsif ( $is_amp_amp{ $types_to_go[$ibeg_2] } ) {
15922 $leading_amp_count++;
15924 # ok to recombine if it follows a ? or :
15925 # and is followed by an open paren..
15927 ( $is_ternary{ $types_to_go[$ibeg_1] }
15928 && $tokens_to_go[$iend_2] eq '(' )
15930 # or is followed by a ? or : at same depth
15932 # We are looking for something like this. We can
15933 # recombine the && line with the line above to make the
15934 # structure more clear:
15936 # exists $G->{Attr}->{V}
15937 # && exists $G->{Attr}->{V}->{$u}
15938 # ? %{ $G->{Attr}->{V}->{$u} }
15941 # We should probably leave something like this alone:
15943 # exists $G->{Attr}->{E}
15944 # && exists $G->{Attr}->{E}->{$u}
15945 # && exists $G->{Attr}->{E}->{$u}->{$v}
15946 # ? %{ $G->{Attr}->{E}->{$u}->{$v} }
15948 # so that we either have all of the &&'s (or ||'s)
15949 # on one line, as in the first example, or break at
15950 # each one as in the second example. However, it
15951 # sometimes makes things worse to check for this because
15952 # it prevents multiple recombinations. So this is not done.
15954 && $is_ternary{ $types_to_go[$ibeg_3] }
15955 && $nesting_depth_to_go[$ibeg_3] ==
15956 $nesting_depth_to_go[$ibeg_2] );
15958 next if !$ok && $want_break_before{ $types_to_go[$ibeg_2] };
15959 $forced_breakpoint_to_go[$iend_1] = 0;
15961 # tweak the bond strength to give this joint priority
15966 # Identify and recombine a broken ?/: chain
15967 elsif ( $types_to_go[$ibeg_2] eq '?' ) {
15969 # Do not recombine different levels
15970 my $lev = $levels_to_go[$ibeg_2];
15971 next if ( $lev ne $levels_to_go[$ibeg_1] );
15973 # Do not recombine a '?' if either next line or
15974 # previous line does not start with a ':'. The reasons
15975 # are that (1) no alignment of the ? will be possible
15976 # and (2) the expression is somewhat complex, so the
15977 # '?' is harder to see in the interior of the line.
15978 my $follows_colon =
15979 $ibeg_1 >= 0 && $types_to_go[$ibeg_1] eq ':';
15980 my $precedes_colon =
15981 $ibeg_3 >= 0 && $types_to_go[$ibeg_3] eq ':';
15982 next unless ( $follows_colon || $precedes_colon );
15984 # we will always combining a ? line following a : line
15985 if ( !$follows_colon ) {
15987 # ...otherwise recombine only if it looks like a chain.
15988 # we will just look at a few nearby lines to see if
15989 # this looks like a chain.
15990 my $local_count = 0;
15991 foreach my $ii ( $ibeg_0, $ibeg_1, $ibeg_3, $ibeg_4 ) {
15994 && $types_to_go[$ii] eq ':'
15995 && $levels_to_go[$ii] == $lev;
15997 next unless ( $local_count > 1 );
15999 $forced_breakpoint_to_go[$iend_1] = 0;
16002 # do not recombine lines with leading '.'
16003 elsif ( $types_to_go[$ibeg_2] =~ /^(\.)$/ ) {
16004 my $i_next_nonblank = $ibeg_2 + 1;
16005 if ( $types_to_go[$i_next_nonblank] eq 'b' ) {
16006 $i_next_nonblank++;
16012 # ... unless there is just one and we can reduce
16013 # this to two lines if we do. For example, this
16017 # '($dummy, $pat) = &get_next_tex_cmd;' . '$args .= $pat;'
16019 # looks better than this:
16020 # $bodyA .= '($dummy, $pat) = &get_next_tex_cmd;'
16021 # . '$args .= $pat;'
16026 && $types_to_go[$ibeg_1] ne $types_to_go[$ibeg_2]
16029 # ... or this would strand a short quote , like this
16030 # . "some long qoute"
16032 || ( $types_to_go[$i_next_nonblank] eq 'Q'
16033 && $i_next_nonblank >= $iend_2 - 1
16034 && length( $tokens_to_go[$i_next_nonblank] ) <
16035 $rOpts_short_concatenation_item_length )
16039 # handle leading keyword..
16040 elsif ( $types_to_go[$ibeg_2] eq 'k' ) {
16042 # handle leading "or"
16043 if ( $tokens_to_go[$ibeg_2] eq 'or' ) {
16046 $this_line_is_semicolon_terminated
16049 # following 'if' or 'unless' or 'or'
16050 $types_to_go[$ibeg_1] eq 'k'
16051 && $is_if_unless{ $tokens_to_go[$ibeg_1] }
16053 # important: only combine a very simple or
16054 # statement because the step below may have
16055 # combined a trailing 'and' with this or,
16056 # and we do not want to then combine
16057 # everything together
16058 && ( $iend_2 - $ibeg_2 <= 7 )
16063 # handle leading 'and'
16064 elsif ( $tokens_to_go[$ibeg_2] eq 'and' ) {
16066 # Decide if we will combine a single terminal 'and'
16067 # after an 'if' or 'unless'.
16069 # This looks best with the 'and' on the same
16070 # line as the 'if':
16073 # if $seconds and $nu < 2;
16075 # But this looks better as shown:
16078 # if !$this->{Parents}{$_}
16079 # or $this->{Parents}{$_} eq $_;
16083 $this_line_is_semicolon_terminated
16086 # following 'if' or 'unless' or 'or'
16087 $types_to_go[$ibeg_1] eq 'k'
16088 && ( $is_if_unless{ $tokens_to_go[$ibeg_1] }
16089 || $tokens_to_go[$ibeg_1] eq 'or' )
16094 # handle leading "if" and "unless"
16095 elsif ( $is_if_unless{ $tokens_to_go[$ibeg_2] } ) {
16097 # FIXME: This is still experimental..may not be too useful
16100 $this_line_is_semicolon_terminated
16102 # previous line begins with 'and' or 'or'
16103 && $types_to_go[$ibeg_1] eq 'k'
16104 && $is_and_or{ $tokens_to_go[$ibeg_1] }
16109 # handle all other leading keywords
16112 # keywords look best at start of lines,
16113 # but combine things like "1 while"
16114 unless ( $is_assignment{ $types_to_go[$iend_1] } ) {
16116 if ( ( $types_to_go[$iend_1] ne 'k' )
16117 && ( $tokens_to_go[$ibeg_2] ne 'while' ) );
16122 # similar treatment of && and || as above for 'and' and 'or':
16123 # NOTE: This block of code is currently bypassed because
16124 # of a previous block but is retained for possible future use.
16125 elsif ( $is_amp_amp{ $types_to_go[$ibeg_2] } ) {
16127 # maybe looking at something like:
16128 # unless $TEXTONLY || $item =~ m%</?(hr>|p>|a|img)%i;
16132 $this_line_is_semicolon_terminated
16134 # previous line begins with an 'if' or 'unless' keyword
16135 && $types_to_go[$ibeg_1] eq 'k'
16136 && $is_if_unless{ $tokens_to_go[$ibeg_1] }
16141 # handle leading + - * /
16142 elsif ( $is_math_op{ $types_to_go[$ibeg_2] } ) {
16143 my $i_next_nonblank = $ibeg_2 + 1;
16144 if ( $types_to_go[$i_next_nonblank] eq 'b' ) {
16145 $i_next_nonblank++;
16148 my $i_next_next = $i_next_nonblank + 1;
16149 $i_next_next++ if ( $types_to_go[$i_next_next] eq 'b' );
16152 $types_to_go[$i_next_nonblank] eq 'n'
16153 && ( $i_next_nonblank >= $iend_2 - 1
16154 || $types_to_go[$i_next_next] eq ';' )
16157 my $iend_1_nonblank =
16158 $types_to_go[$iend_1] eq 'b' ? $iend_1 - 1 : $iend_1;
16159 my $iend_2_nonblank =
16160 $types_to_go[$iend_2] eq 'b' ? $iend_2 - 1 : $iend_2;
16162 my $is_short_term =
16163 ( $types_to_go[$ibeg_2] eq $types_to_go[$ibeg_1]
16164 && $types_to_go[$iend_2_nonblank] =~ /^[in]$/
16165 && $types_to_go[$iend_1_nonblank] =~ /^[in]$/
16166 && $iend_2_nonblank <= $ibeg_2 + 2
16167 && length( $tokens_to_go[$iend_2_nonblank] ) <
16168 $rOpts_short_concatenation_item_length );
16170 # Combine these lines if this line is a single
16171 # number, or if it is a short term with same
16172 # operator as the previous line. For example, in
16173 # the following code we will combine all of the
16174 # short terms $A, $B, $C, $D, $E, $F, together
16175 # instead of leaving them one per line:
16177 # $A * $B * $C * $D * $E * $F *
16178 # ( 2. * $eps * $sigma * $area ) *
16179 # ( 1. / $tcold**3 - 1. / $thot**3 );
16180 # This can be important in math-intensive code.
16186 # or if we can reduce this to two lines if we do.
16189 && $types_to_go[$ibeg_1] ne $types_to_go[$ibeg_2] )
16193 # handle line with leading = or similar
16194 elsif ( $is_assignment{ $types_to_go[$ibeg_2] } ) {
16195 next unless $n == 1;
16199 # unless we can reduce this to two lines
16202 # or three lines, the last with a leading semicolon
16203 || ( $nmax == 3 && $types_to_go[$ibeg_nmax] eq ';' )
16205 # or the next line ends with a here doc
16206 || $types_to_go[$iend_2] eq 'h'
16210 #----------------------------------------------------------
16212 # Combine the lines if we arrive here and it is possible
16213 #----------------------------------------------------------
16215 # honor hard breakpoints
16216 next if ( $forced_breakpoint_to_go[$iend_1] > 0 );
16218 my $bs = $bond_strength_to_go[$iend_1] + $bs_tweak;
16220 # combined line cannot be too long
16222 if excess_line_length( $ibeg_1, $iend_2 ) > 0;
16224 # do not recombine if we would skip in indentation levels
16225 if ( $n < $nmax ) {
16226 my $if_next = $$ri_beg[ $n + 1 ];
16229 $levels_to_go[$ibeg_1] < $levels_to_go[$ibeg_2]
16230 && $levels_to_go[$ibeg_2] < $levels_to_go[$if_next]
16232 # but an isolated 'if (' is undesirable
16235 && $iend_1 - $ibeg_1 <= 2
16236 && $types_to_go[$ibeg_1] eq 'k'
16237 && $tokens_to_go[$ibeg_1] eq 'if'
16238 && $tokens_to_go[$iend_1] ne '('
16244 next if ( $bs == NO_BREAK );
16246 # remember the pair with the greatest bond strength
16253 if ( $bs > $bs_best ) {
16260 # recombine the pair with the greatest bond strength
16262 splice @$ri_beg, $n_best, 1;
16263 splice @$ri_end, $n_best - 1, 1;
16265 # keep going if we are still making progress
16269 return ( $ri_beg, $ri_end );
16271 } # end recombine_breakpoints
16273 sub break_all_chain_tokens {
16275 # scan the current breakpoints looking for breaks at certain "chain
16276 # operators" (. : && || + etc) which often occur repeatedly in a long
16277 # statement. If we see a break at any one, break at all similar tokens
16278 # within the same container.
16280 my ( $ri_left, $ri_right ) = @_;
16282 my %saw_chain_type;
16283 my %left_chain_type;
16284 my %right_chain_type;
16285 my %interior_chain_type;
16286 my $nmax = @$ri_right - 1;
16288 # scan the left and right end tokens of all lines
16290 for my $n ( 0 .. $nmax ) {
16291 my $il = $$ri_left[$n];
16292 my $ir = $$ri_right[$n];
16293 my $typel = $types_to_go[$il];
16294 my $typer = $types_to_go[$ir];
16295 $typel = '+' if ( $typel eq '-' ); # treat + and - the same
16296 $typer = '+' if ( $typer eq '-' );
16297 $typel = '*' if ( $typel eq '/' ); # treat * and / the same
16298 $typer = '*' if ( $typer eq '/' );
16299 my $tokenl = $tokens_to_go[$il];
16300 my $tokenr = $tokens_to_go[$ir];
16302 if ( $is_chain_operator{$tokenl} && $want_break_before{$typel} ) {
16303 next if ( $typel eq '?' );
16304 push @{ $left_chain_type{$typel} }, $il;
16305 $saw_chain_type{$typel} = 1;
16308 if ( $is_chain_operator{$tokenr} && !$want_break_before{$typer} ) {
16309 next if ( $typer eq '?' );
16310 push @{ $right_chain_type{$typer} }, $ir;
16311 $saw_chain_type{$typer} = 1;
16315 return unless $count;
16317 # now look for any interior tokens of the same types
16319 for my $n ( 0 .. $nmax ) {
16320 my $il = $$ri_left[$n];
16321 my $ir = $$ri_right[$n];
16322 for ( my $i = $il + 1 ; $i < $ir ; $i++ ) {
16323 my $type = $types_to_go[$i];
16324 $type = '+' if ( $type eq '-' );
16325 $type = '*' if ( $type eq '/' );
16326 if ( $saw_chain_type{$type} ) {
16327 push @{ $interior_chain_type{$type} }, $i;
16332 return unless $count;
16334 # now make a list of all new break points
16337 # loop over all chain types
16338 foreach my $type ( keys %saw_chain_type ) {
16340 # quit if just ONE continuation line with leading . For example--
16341 # print LATEXFILE '\framebox{\parbox[c][' . $h . '][t]{' . $w . '}{'
16343 last if ( $nmax == 1 && $type =~ /^[\.\+]$/ );
16345 # loop over all interior chain tokens
16346 foreach my $itest ( @{ $interior_chain_type{$type} } ) {
16348 # loop over all left end tokens of same type
16349 if ( $left_chain_type{$type} ) {
16350 next if $nobreak_to_go[ $itest - 1 ];
16351 foreach my $i ( @{ $left_chain_type{$type} } ) {
16352 next unless in_same_container( $i, $itest );
16353 push @insert_list, $itest - 1;
16355 # Break at matching ? if this : is at a different level.
16356 # For example, the ? before $THRf_DEAD in the following
16357 # should get a break if its : gets a break.
16360 # ( $_ & 1 ) ? ( $_ & 4 ) ? $THRf_DEAD : $THRf_ZOMBIE
16361 # : ( $_ & 4 ) ? $THRf_R_DETACHED
16362 # : $THRf_R_JOINABLE;
16364 && $levels_to_go[$i] != $levels_to_go[$itest] )
16366 my $i_question = $mate_index_to_go[$itest];
16367 if ( $i_question > 0 ) {
16368 push @insert_list, $i_question - 1;
16375 # loop over all right end tokens of same type
16376 if ( $right_chain_type{$type} ) {
16377 next if $nobreak_to_go[$itest];
16378 foreach my $i ( @{ $right_chain_type{$type} } ) {
16379 next unless in_same_container( $i, $itest );
16380 push @insert_list, $itest;
16382 # break at matching ? if this : is at a different level
16384 && $levels_to_go[$i] != $levels_to_go[$itest] )
16386 my $i_question = $mate_index_to_go[$itest];
16387 if ( $i_question >= 0 ) {
16388 push @insert_list, $i_question;
16397 # insert any new break points
16398 if (@insert_list) {
16399 insert_additional_breaks( \@insert_list, $ri_left, $ri_right );
16405 # Look for assignment operators that could use a breakpoint.
16406 # For example, in the following snippet
16408 # $HOME = $ENV{HOME}
16411 # || die "no home directory for user $<";
16413 # we could break at the = to get this, which is a little nicer:
16418 # || die "no home directory for user $<";
16420 # The logic here follows the logic in set_logical_padding, which
16421 # will add the padding in the second line to improve alignment.
16423 my ( $ri_left, $ri_right ) = @_;
16424 my $nmax = @$ri_right - 1;
16425 return unless ( $nmax >= 2 );
16427 # scan the left ends of first two lines
16430 for my $n ( 1 .. 2 ) {
16431 my $il = $$ri_left[$n];
16432 my $typel = $types_to_go[$il];
16433 my $tokenl = $tokens_to_go[$il];
16435 my $has_leading_op = ( $tokenl =~ /^\w/ )
16436 ? $is_chain_operator{$tokenl} # + - * / : ? && ||
16437 : $is_chain_operator{$typel}; # and, or
16438 return unless ($has_leading_op);
16441 unless ( $tokenl eq $tokbeg
16442 && $nesting_depth_to_go[$il] eq $depth_beg );
16445 $depth_beg = $nesting_depth_to_go[$il];
16448 # now look for any interior tokens of the same types
16449 my $il = $$ri_left[0];
16450 my $ir = $$ri_right[0];
16452 # now make a list of all new break points
16454 for ( my $i = $ir - 1 ; $i > $il ; $i-- ) {
16455 my $type = $types_to_go[$i];
16456 if ( $is_assignment{$type}
16457 && $nesting_depth_to_go[$i] eq $depth_beg )
16459 if ( $want_break_before{$type} ) {
16460 push @insert_list, $i - 1;
16463 push @insert_list, $i;
16468 # Break after a 'return' followed by a chain of operators
16469 # return ( $^O !~ /win32|dos/i )
16470 # && ( $^O ne 'VMS' )
16471 # && ( $^O ne 'OS2' )
16472 # && ( $^O ne 'MacOS' );
16475 # ( $^O !~ /win32|dos/i )
16476 # && ( $^O ne 'VMS' )
16477 # && ( $^O ne 'OS2' )
16478 # && ( $^O ne 'MacOS' );
16480 if ( $types_to_go[$i] eq 'k'
16481 && $tokens_to_go[$i] eq 'return'
16483 && $nesting_depth_to_go[$i] eq $depth_beg )
16485 push @insert_list, $i;
16488 return unless (@insert_list);
16490 # One final check...
16491 # scan second and thrid lines and be sure there are no assignments
16492 # we want to avoid breaking at an = to make something like this:
16494 # $html_icons{"$type-$state"}
16495 # or $icon = $html_icons{$type}
16496 # or $icon = $html_icons{$state} )
16497 for my $n ( 1 .. 2 ) {
16498 my $il = $$ri_left[$n];
16499 my $ir = $$ri_right[$n];
16500 for ( my $i = $il + 1 ; $i <= $ir ; $i++ ) {
16501 my $type = $types_to_go[$i];
16503 if ( $is_assignment{$type}
16504 && $nesting_depth_to_go[$i] eq $depth_beg );
16508 # ok, insert any new break point
16509 if (@insert_list) {
16510 insert_additional_breaks( \@insert_list, $ri_left, $ri_right );
16514 sub insert_final_breaks {
16516 my ( $ri_left, $ri_right ) = @_;
16518 my $nmax = @$ri_right - 1;
16520 # scan the left and right end tokens of all lines
16522 my $i_first_colon = -1;
16523 for my $n ( 0 .. $nmax ) {
16524 my $il = $$ri_left[$n];
16525 my $ir = $$ri_right[$n];
16526 my $typel = $types_to_go[$il];
16527 my $typer = $types_to_go[$ir];
16528 return if ( $typel eq '?' );
16529 return if ( $typer eq '?' );
16530 if ( $typel eq ':' ) { $i_first_colon = $il; last; }
16531 elsif ( $typer eq ':' ) { $i_first_colon = $ir; last; }
16534 # For long ternary chains,
16535 # if the first : we see has its # ? is in the interior
16536 # of a preceding line, then see if there are any good
16537 # breakpoints before the ?.
16538 if ( $i_first_colon > 0 ) {
16539 my $i_question = $mate_index_to_go[$i_first_colon];
16540 if ( $i_question > 0 ) {
16542 for ( my $ii = $i_question - 1 ; $ii >= 0 ; $ii -= 1 ) {
16543 my $token = $tokens_to_go[$ii];
16544 my $type = $types_to_go[$ii];
16546 # For now, a good break is either a comma or a 'return'.
16547 if ( ( $type eq ',' || $type eq 'k' && $token eq 'return' )
16548 && in_same_container( $ii, $i_question ) )
16550 push @insert_list, $ii;
16555 # insert any new break points
16556 if (@insert_list) {
16557 insert_additional_breaks( \@insert_list, $ri_left, $ri_right );
16563 sub in_same_container {
16565 # check to see if tokens at i1 and i2 are in the
16566 # same container, and not separated by a comma, ? or :
16567 my ( $i1, $i2 ) = @_;
16568 my $type = $types_to_go[$i1];
16569 my $depth = $nesting_depth_to_go[$i1];
16570 return unless ( $nesting_depth_to_go[$i2] == $depth );
16571 if ( $i2 < $i1 ) { ( $i1, $i2 ) = ( $i2, $i1 ) }
16573 ###########################################################
16574 # This is potentially a very slow routine and not critical.
16575 # For safety just give up for large differences.
16576 # See test file 'infinite_loop.txt'
16577 # TODO: replace this loop with a data structure
16578 ###########################################################
16579 return if ( $i2-$i1 > 200 );
16581 for ( my $i = $i1 + 1 ; $i < $i2 ; $i++ ) {
16582 next if ( $nesting_depth_to_go[$i] > $depth );
16583 return if ( $nesting_depth_to_go[$i] < $depth );
16585 my $tok = $tokens_to_go[$i];
16586 $tok = ',' if $tok eq '=>'; # treat => same as ,
16588 # Example: we would not want to break at any of these .'s
16589 # : "<A HREF=\"#item_" . htmlify( 0, $s2 ) . "\">$str</A>"
16590 if ( $type ne ':' ) {
16591 return if ( $tok =~ /^[\,\:\?]$/ ) || $tok eq '||' || $tok eq 'or';
16594 return if ( $tok =~ /^[\,]$/ );
16600 sub set_continuation_breaks {
16602 # Define an array of indexes for inserting newline characters to
16603 # keep the line lengths below the maximum desired length. There is
16604 # an implied break after the last token, so it need not be included.
16607 # This routine is part of series of routines which adjust line
16608 # lengths. It is only called if a statement is longer than the
16609 # maximum line length, or if a preliminary scanning located
16610 # desirable break points. Sub scan_list has already looked at
16611 # these tokens and set breakpoints (in array
16612 # $forced_breakpoint_to_go[$i]) where it wants breaks (for example
16613 # after commas, after opening parens, and before closing parens).
16614 # This routine will honor these breakpoints and also add additional
16615 # breakpoints as necessary to keep the line length below the maximum
16616 # requested. It bases its decision on where the 'bond strength' is
16619 # Output: returns references to the arrays:
16622 # which contain the indexes $i of the first and last tokens on each
16625 # In addition, the array:
16626 # $forced_breakpoint_to_go[$i]
16627 # may be updated to be =1 for any index $i after which there must be
16628 # a break. This signals later routines not to undo the breakpoint.
16630 my $saw_good_break = shift;
16631 my @i_first = (); # the first index to output
16632 my @i_last = (); # the last index to output
16633 my @i_colon_breaks = (); # needed to decide if we have to break at ?'s
16634 if ( $types_to_go[0] eq ':' ) { push @i_colon_breaks, 0 }
16636 set_bond_strengths();
16639 my $imax = $max_index_to_go;
16640 if ( $types_to_go[$imin] eq 'b' ) { $imin++ }
16641 if ( $types_to_go[$imax] eq 'b' ) { $imax-- }
16642 my $i_begin = $imin; # index for starting next iteration
16644 my $leading_spaces = leading_spaces_to_go($imin);
16645 my $line_count = 0;
16646 my $last_break_strength = NO_BREAK;
16647 my $i_last_break = -1;
16648 my $max_bias = 0.001;
16649 my $tiny_bias = 0.0001;
16650 my $leading_alignment_token = "";
16651 my $leading_alignment_type = "";
16653 # see if any ?/:'s are in order
16654 my $colons_in_order = 1;
16656 my @colon_list = grep /^[\?\:]$/, @tokens_to_go[ 0 .. $max_index_to_go ];
16657 my $colon_count = @colon_list;
16658 foreach (@colon_list) {
16659 if ( $_ eq $last_tok ) { $colons_in_order = 0; last }
16663 # This is a sufficient but not necessary condition for colon chain
16664 my $is_colon_chain = ( $colons_in_order && @colon_list > 2 );
16666 #-------------------------------------------------------
16667 # BEGINNING of main loop to set continuation breakpoints
16668 # Keep iterating until we reach the end
16669 #-------------------------------------------------------
16670 while ( $i_begin <= $imax ) {
16671 my $lowest_strength = NO_BREAK;
16672 my $starting_sum = $lengths_to_go[$i_begin];
16675 my $lowest_next_token = '';
16676 my $lowest_next_type = 'b';
16677 my $i_lowest_next_nonblank = -1;
16679 #-------------------------------------------------------
16680 # BEGINNING of inner loop to find the best next breakpoint
16681 #-------------------------------------------------------
16682 for ( $i_test = $i_begin ; $i_test <= $imax ; $i_test++ ) {
16683 my $type = $types_to_go[$i_test];
16684 my $token = $tokens_to_go[$i_test];
16685 my $next_type = $types_to_go[ $i_test + 1 ];
16686 my $next_token = $tokens_to_go[ $i_test + 1 ];
16687 my $i_next_nonblank =
16688 ( ( $next_type eq 'b' ) ? $i_test + 2 : $i_test + 1 );
16689 my $next_nonblank_type = $types_to_go[$i_next_nonblank];
16690 my $next_nonblank_token = $tokens_to_go[$i_next_nonblank];
16691 my $next_nonblank_block_type = $block_type_to_go[$i_next_nonblank];
16692 my $strength = $bond_strength_to_go[$i_test];
16693 my $must_break = 0;
16695 # FIXME: TESTING: Might want to be able to break after these
16696 # force an immediate break at certain operators
16697 # with lower level than the start of the line
16700 $next_nonblank_type =~ /^(\.|\&\&|\|\|)$/
16701 || ( $next_nonblank_type eq 'k'
16702 && $next_nonblank_token =~ /^(and|or)$/ )
16704 && ( $nesting_depth_to_go[$i_begin] >
16705 $nesting_depth_to_go[$i_next_nonblank] )
16708 set_forced_breakpoint($i_next_nonblank);
16713 # Try to put a break where requested by scan_list
16714 $forced_breakpoint_to_go[$i_test]
16716 # break between ) { in a continued line so that the '{' can
16718 # See similar logic in scan_list which catches instances
16719 # where a line is just something like ') {'
16721 && ( $token eq ')' )
16722 && ( $next_nonblank_type eq '{' )
16723 && ($next_nonblank_block_type)
16724 && !$rOpts->{'opening-brace-always-on-right'} )
16726 # There is an implied forced break at a terminal opening brace
16727 || ( ( $type eq '{' ) && ( $i_test == $imax ) )
16731 # Forced breakpoints must sometimes be overridden, for example
16732 # because of a side comment causing a NO_BREAK. It is easier
16733 # to catch this here than when they are set.
16734 if ( $strength < NO_BREAK ) {
16735 $strength = $lowest_strength - $tiny_bias;
16740 # quit if a break here would put a good terminal token on
16741 # the next line and we already have a possible break
16744 && ( $next_nonblank_type =~ /^[\;\,]$/ )
16748 $lengths_to_go[ $i_next_nonblank + 1 ] -
16750 ) > $rOpts_maximum_line_length
16754 last if ( $i_lowest >= 0 );
16757 # Avoid a break which would strand a single punctuation
16758 # token. For example, we do not want to strand a leading
16759 # '.' which is followed by a long quoted string.
16762 && ( $i_test == $i_begin )
16763 && ( $i_test < $imax )
16764 && ( $token eq $type )
16768 $lengths_to_go[ $i_test + 1 ] -
16770 ) <= $rOpts_maximum_line_length
16776 if ( ( $i_test < $imax ) && ( $next_type eq 'b' ) ) {
16782 if ( ( $strength <= $lowest_strength ) && ( $strength < NO_BREAK ) )
16785 # break at previous best break if it would have produced
16786 # a leading alignment of certain common tokens, and it
16787 # is different from the latest candidate break
16789 if ($leading_alignment_type);
16791 # Force at least one breakpoint if old code had good
16792 # break It is only called if a breakpoint is required or
16793 # desired. This will probably need some adjustments
16794 # over time. A goal is to try to be sure that, if a new
16795 # side comment is introduced into formated text, then
16796 # the same breakpoints will occur. scbreak.t
16799 $i_test == $imax # we are at the end
16800 && !$forced_breakpoint_count #
16801 && $saw_good_break # old line had good break
16802 && $type =~ /^[#;\{]$/ # and this line ends in
16803 # ';' or side comment
16804 && $i_last_break < 0 # and we haven't made a break
16805 && $i_lowest > 0 # and we saw a possible break
16806 && $i_lowest < $imax - 1 # (but not just before this ;)
16807 && $strength - $lowest_strength < 0.5 * WEAK # and it's good
16810 $lowest_strength = $strength;
16811 $i_lowest = $i_test;
16812 $lowest_next_token = $next_nonblank_token;
16813 $lowest_next_type = $next_nonblank_type;
16814 $i_lowest_next_nonblank = $i_next_nonblank;
16815 last if $must_break;
16817 # set flags to remember if a break here will produce a
16818 # leading alignment of certain common tokens
16819 if ( $line_count > 0
16821 && ( $lowest_strength - $last_break_strength <= $max_bias )
16824 my $i_last_end = $i_begin - 1;
16825 if ( $types_to_go[$i_last_end] eq 'b' ) { $i_last_end -= 1 }
16826 my $tok_beg = $tokens_to_go[$i_begin];
16827 my $type_beg = $types_to_go[$i_begin];
16830 # check for leading alignment of certain tokens
16832 $tok_beg eq $next_nonblank_token
16833 && $is_chain_operator{$tok_beg}
16834 && ( $type_beg eq 'k'
16835 || $type_beg eq $tok_beg )
16836 && $nesting_depth_to_go[$i_begin] >=
16837 $nesting_depth_to_go[$i_next_nonblank]
16840 || ( $tokens_to_go[$i_last_end] eq $token
16841 && $is_chain_operator{$token}
16842 && ( $type eq 'k' || $type eq $token )
16843 && $nesting_depth_to_go[$i_last_end] >=
16844 $nesting_depth_to_go[$i_test] )
16847 $leading_alignment_token = $next_nonblank_token;
16848 $leading_alignment_type = $next_nonblank_type;
16854 ( $i_test >= $imax )
16859 $lengths_to_go[ $i_test + 2 ] -
16861 ) > $rOpts_maximum_line_length
16864 FORMATTER_DEBUG_FLAG_BREAK
16866 "BREAK: testing i = $i_test imax=$imax $types_to_go[$i_test] $next_nonblank_type leading sp=($leading_spaces) next length = $lengths_to_go[$i_test+2] too_long=$too_long str=$strength\n";
16868 # allow one extra terminal token after exceeding line length
16869 # if it would strand this token.
16870 if ( $rOpts_fuzzy_line_length
16872 && ( $i_lowest == $i_test )
16873 && ( length($token) > 1 )
16874 && ( $next_nonblank_type =~ /^[\;\,]$/ ) )
16881 ( $i_test == $imax ) # we're done if no more tokens,
16883 ( $i_lowest >= 0 ) # or no more space and we have a break
16889 #-------------------------------------------------------
16890 # END of inner loop to find the best next breakpoint
16891 # Now decide exactly where to put the breakpoint
16892 #-------------------------------------------------------
16894 # it's always ok to break at imax if no other break was found
16895 if ( $i_lowest < 0 ) { $i_lowest = $imax }
16897 # semi-final index calculation
16898 my $i_next_nonblank = (
16899 ( $types_to_go[ $i_lowest + 1 ] eq 'b' )
16903 my $next_nonblank_type = $types_to_go[$i_next_nonblank];
16904 my $next_nonblank_token = $tokens_to_go[$i_next_nonblank];
16906 #-------------------------------------------------------
16907 # ?/: rule 1 : if a break here will separate a '?' on this
16908 # line from its closing ':', then break at the '?' instead.
16909 #-------------------------------------------------------
16911 foreach $i ( $i_begin + 1 .. $i_lowest - 1 ) {
16912 next unless ( $tokens_to_go[$i] eq '?' );
16914 # do not break if probable sequence of ?/: statements
16915 next if ($is_colon_chain);
16917 # do not break if statement is broken by side comment
16920 $tokens_to_go[$max_index_to_go] eq '#'
16921 && terminal_type( \@types_to_go, \@block_type_to_go, 0,
16922 $max_index_to_go ) !~ /^[\;\}]$/
16925 # no break needed if matching : is also on the line
16927 if ( $mate_index_to_go[$i] >= 0
16928 && $mate_index_to_go[$i] <= $i_next_nonblank );
16931 if ( $want_break_before{'?'} ) { $i_lowest-- }
16935 #-------------------------------------------------------
16936 # END of inner loop to find the best next breakpoint:
16937 # Break the line after the token with index i=$i_lowest
16938 #-------------------------------------------------------
16940 # final index calculation
16941 $i_next_nonblank = (
16942 ( $types_to_go[ $i_lowest + 1 ] eq 'b' )
16946 $next_nonblank_type = $types_to_go[$i_next_nonblank];
16947 $next_nonblank_token = $tokens_to_go[$i_next_nonblank];
16949 FORMATTER_DEBUG_FLAG_BREAK
16950 && print "BREAK: best is i = $i_lowest strength = $lowest_strength\n";
16952 #-------------------------------------------------------
16953 # ?/: rule 2 : if we break at a '?', then break at its ':'
16955 # Note: this rule is also in sub scan_list to handle a break
16956 # at the start and end of a line (in case breaks are dictated
16957 # by side comments).
16958 #-------------------------------------------------------
16959 if ( $next_nonblank_type eq '?' ) {
16960 set_closing_breakpoint($i_next_nonblank);
16962 elsif ( $types_to_go[$i_lowest] eq '?' ) {
16963 set_closing_breakpoint($i_lowest);
16966 #-------------------------------------------------------
16967 # ?/: rule 3 : if we break at a ':' then we save
16968 # its location for further work below. We may need to go
16969 # back and break at its '?'.
16970 #-------------------------------------------------------
16971 if ( $next_nonblank_type eq ':' ) {
16972 push @i_colon_breaks, $i_next_nonblank;
16974 elsif ( $types_to_go[$i_lowest] eq ':' ) {
16975 push @i_colon_breaks, $i_lowest;
16978 # here we should set breaks for all '?'/':' pairs which are
16979 # separated by this line
16983 # save this line segment, after trimming blanks at the ends
16985 ( $types_to_go[$i_begin] eq 'b' ) ? $i_begin + 1 : $i_begin );
16987 ( $types_to_go[$i_lowest] eq 'b' ) ? $i_lowest - 1 : $i_lowest );
16989 # set a forced breakpoint at a container opening, if necessary, to
16990 # signal a break at a closing container. Excepting '(' for now.
16991 if ( $tokens_to_go[$i_lowest] =~ /^[\{\[]$/
16992 && !$forced_breakpoint_to_go[$i_lowest] )
16994 set_closing_breakpoint($i_lowest);
16997 # get ready to go again
16998 $i_begin = $i_lowest + 1;
16999 $last_break_strength = $lowest_strength;
17000 $i_last_break = $i_lowest;
17001 $leading_alignment_token = "";
17002 $leading_alignment_type = "";
17003 $lowest_next_token = '';
17004 $lowest_next_type = 'b';
17006 if ( ( $i_begin <= $imax ) && ( $types_to_go[$i_begin] eq 'b' ) ) {
17010 # update indentation size
17011 if ( $i_begin <= $imax ) {
17012 $leading_spaces = leading_spaces_to_go($i_begin);
17016 #-------------------------------------------------------
17017 # END of main loop to set continuation breakpoints
17018 # Now go back and make any necessary corrections
17019 #-------------------------------------------------------
17021 #-------------------------------------------------------
17022 # ?/: rule 4 -- if we broke at a ':', then break at
17023 # corresponding '?' unless this is a chain of ?: expressions
17024 #-------------------------------------------------------
17025 if (@i_colon_breaks) {
17027 # using a simple method for deciding if we are in a ?/: chain --
17028 # this is a chain if it has multiple ?/: pairs all in order;
17030 # Note that if line starts in a ':' we count that above as a break
17031 my $is_chain = ( $colons_in_order && @i_colon_breaks > 1 );
17033 unless ($is_chain) {
17034 my @insert_list = ();
17035 foreach (@i_colon_breaks) {
17036 my $i_question = $mate_index_to_go[$_];
17037 if ( $i_question >= 0 ) {
17038 if ( $want_break_before{'?'} ) {
17040 if ( $i_question > 0
17041 && $types_to_go[$i_question] eq 'b' )
17047 if ( $i_question >= 0 ) {
17048 push @insert_list, $i_question;
17051 insert_additional_breaks( \@insert_list, \@i_first, \@i_last );
17055 return ( \@i_first, \@i_last, $colon_count );
17058 sub insert_additional_breaks {
17060 # this routine will add line breaks at requested locations after
17061 # sub set_continuation_breaks has made preliminary breaks.
17063 my ( $ri_break_list, $ri_first, $ri_last ) = @_;
17066 my $line_number = 0;
17068 foreach $i_break_left ( sort { $a <=> $b } @$ri_break_list ) {
17070 $i_f = $$ri_first[$line_number];
17071 $i_l = $$ri_last[$line_number];
17072 while ( $i_break_left >= $i_l ) {
17075 # shouldn't happen unless caller passes bad indexes
17076 if ( $line_number >= @$ri_last ) {
17078 "Non-fatal program bug: couldn't set break at $i_break_left\n"
17080 report_definite_bug();
17083 $i_f = $$ri_first[$line_number];
17084 $i_l = $$ri_last[$line_number];
17087 my $i_break_right = $i_break_left + 1;
17088 if ( $types_to_go[$i_break_right] eq 'b' ) { $i_break_right++ }
17090 if ( $i_break_left >= $i_f
17091 && $i_break_left < $i_l
17092 && $i_break_right > $i_f
17093 && $i_break_right <= $i_l )
17095 splice( @$ri_first, $line_number, 1, ( $i_f, $i_break_right ) );
17096 splice( @$ri_last, $line_number, 1, ( $i_break_left, $i_l ) );
17101 sub set_closing_breakpoint {
17103 # set a breakpoint at a matching closing token
17104 # at present, this is only used to break at a ':' which matches a '?'
17105 my $i_break = shift;
17107 if ( $mate_index_to_go[$i_break] >= 0 ) {
17109 # CAUTION: infinite recursion possible here:
17110 # set_closing_breakpoint calls set_forced_breakpoint, and
17111 # set_forced_breakpoint call set_closing_breakpoint
17112 # ( test files attrib.t, BasicLyx.pm.html).
17113 # Don't reduce the '2' in the statement below
17114 if ( $mate_index_to_go[$i_break] > $i_break + 2 ) {
17116 # break before } ] and ), but sub set_forced_breakpoint will decide
17117 # to break before or after a ? and :
17118 my $inc = ( $tokens_to_go[$i_break] eq '?' ) ? 0 : 1;
17119 set_forced_breakpoint( $mate_index_to_go[$i_break] - $inc );
17123 my $type_sequence = $type_sequence_to_go[$i_break];
17124 if ($type_sequence) {
17125 my $closing_token = $matching_token{ $tokens_to_go[$i_break] };
17126 $postponed_breakpoint{$type_sequence} = 1;
17131 # check to see if output line tabbing agrees with input line
17132 # this can be very useful for debugging a script which has an extra
17134 sub compare_indentation_levels {
17136 my ( $python_indentation_level, $structural_indentation_level ) = @_;
17137 if ( ( $python_indentation_level ne $structural_indentation_level ) ) {
17138 $last_tabbing_disagreement = $input_line_number;
17140 if ($in_tabbing_disagreement) {
17143 $tabbing_disagreement_count++;
17145 if ( $tabbing_disagreement_count <= MAX_NAG_MESSAGES ) {
17146 write_logfile_entry(
17147 "Start indentation disagreement: input=$python_indentation_level; output=$structural_indentation_level\n"
17150 $in_tabbing_disagreement = $input_line_number;
17151 $first_tabbing_disagreement = $in_tabbing_disagreement
17152 unless ($first_tabbing_disagreement);
17157 if ($in_tabbing_disagreement) {
17159 if ( $tabbing_disagreement_count <= MAX_NAG_MESSAGES ) {
17160 write_logfile_entry(
17161 "End indentation disagreement from input line $in_tabbing_disagreement\n"
17164 if ( $tabbing_disagreement_count == MAX_NAG_MESSAGES ) {
17165 write_logfile_entry(
17166 "No further tabbing disagreements will be noted\n");
17169 $in_tabbing_disagreement = 0;
17174 #####################################################################
17176 # the Perl::Tidy::IndentationItem class supplies items which contain
17177 # how much whitespace should be used at the start of a line
17179 #####################################################################
17181 package Perl::Tidy::IndentationItem;
17183 # Indexes for indentation items
17184 use constant SPACES => 0; # total leading white spaces
17185 use constant LEVEL => 1; # the indentation 'level'
17186 use constant CI_LEVEL => 2; # the 'continuation level'
17187 use constant AVAILABLE_SPACES => 3; # how many left spaces available
17189 use constant CLOSED => 4; # index where we saw closing '}'
17190 use constant COMMA_COUNT => 5; # how many commas at this level?
17191 use constant SEQUENCE_NUMBER => 6; # output batch number
17192 use constant INDEX => 7; # index in output batch list
17193 use constant HAVE_CHILD => 8; # any dependents?
17194 use constant RECOVERABLE_SPACES => 9; # how many spaces to the right
17195 # we would like to move to get
17196 # alignment (negative if left)
17197 use constant ALIGN_PAREN => 10; # do we want to try to align
17198 # with an opening structure?
17199 use constant MARKED => 11; # if visited by corrector logic
17200 use constant STACK_DEPTH => 12; # indentation nesting depth
17201 use constant STARTING_INDEX => 13; # first token index of this level
17202 use constant ARROW_COUNT => 14; # how many =>'s
17206 # Create an 'indentation_item' which describes one level of leading
17207 # whitespace when the '-lp' indentation is used. We return
17208 # a reference to an anonymous array of associated variables.
17209 # See above constants for storage scheme.
17211 $class, $spaces, $level,
17212 $ci_level, $available_spaces, $index,
17213 $gnu_sequence_number, $align_paren, $stack_depth,
17217 my $arrow_count = 0;
17218 my $comma_count = 0;
17219 my $have_child = 0;
17220 my $want_right_spaces = 0;
17223 $spaces, $level, $ci_level,
17224 $available_spaces, $closed, $comma_count,
17225 $gnu_sequence_number, $index, $have_child,
17226 $want_right_spaces, $align_paren, $marked,
17227 $stack_depth, $starting_index, $arrow_count,
17231 sub permanently_decrease_AVAILABLE_SPACES {
17233 # make a permanent reduction in the available indentation spaces
17234 # at one indentation item. NOTE: if there are child nodes, their
17235 # total SPACES must be reduced by the caller.
17237 my ( $item, $spaces_needed ) = @_;
17238 my $available_spaces = $item->get_AVAILABLE_SPACES();
17239 my $deleted_spaces =
17240 ( $available_spaces > $spaces_needed )
17242 : $available_spaces;
17243 $item->decrease_AVAILABLE_SPACES($deleted_spaces);
17244 $item->decrease_SPACES($deleted_spaces);
17245 $item->set_RECOVERABLE_SPACES(0);
17247 return $deleted_spaces;
17250 sub tentatively_decrease_AVAILABLE_SPACES {
17252 # We are asked to tentatively delete $spaces_needed of indentation
17253 # for a indentation item. We may want to undo this later. NOTE: if
17254 # there are child nodes, their total SPACES must be reduced by the
17256 my ( $item, $spaces_needed ) = @_;
17257 my $available_spaces = $item->get_AVAILABLE_SPACES();
17258 my $deleted_spaces =
17259 ( $available_spaces > $spaces_needed )
17261 : $available_spaces;
17262 $item->decrease_AVAILABLE_SPACES($deleted_spaces);
17263 $item->decrease_SPACES($deleted_spaces);
17264 $item->increase_RECOVERABLE_SPACES($deleted_spaces);
17265 return $deleted_spaces;
17268 sub get_STACK_DEPTH {
17270 return $self->[STACK_DEPTH];
17275 return $self->[SPACES];
17280 return $self->[MARKED];
17284 my ( $self, $value ) = @_;
17285 if ( defined($value) ) {
17286 $self->[MARKED] = $value;
17288 return $self->[MARKED];
17291 sub get_AVAILABLE_SPACES {
17293 return $self->[AVAILABLE_SPACES];
17296 sub decrease_SPACES {
17297 my ( $self, $value ) = @_;
17298 if ( defined($value) ) {
17299 $self->[SPACES] -= $value;
17301 return $self->[SPACES];
17304 sub decrease_AVAILABLE_SPACES {
17305 my ( $self, $value ) = @_;
17306 if ( defined($value) ) {
17307 $self->[AVAILABLE_SPACES] -= $value;
17309 return $self->[AVAILABLE_SPACES];
17312 sub get_ALIGN_PAREN {
17314 return $self->[ALIGN_PAREN];
17317 sub get_RECOVERABLE_SPACES {
17319 return $self->[RECOVERABLE_SPACES];
17322 sub set_RECOVERABLE_SPACES {
17323 my ( $self, $value ) = @_;
17324 if ( defined($value) ) {
17325 $self->[RECOVERABLE_SPACES] = $value;
17327 return $self->[RECOVERABLE_SPACES];
17330 sub increase_RECOVERABLE_SPACES {
17331 my ( $self, $value ) = @_;
17332 if ( defined($value) ) {
17333 $self->[RECOVERABLE_SPACES] += $value;
17335 return $self->[RECOVERABLE_SPACES];
17340 return $self->[CI_LEVEL];
17345 return $self->[LEVEL];
17348 sub get_SEQUENCE_NUMBER {
17350 return $self->[SEQUENCE_NUMBER];
17355 return $self->[INDEX];
17358 sub get_STARTING_INDEX {
17360 return $self->[STARTING_INDEX];
17363 sub set_HAVE_CHILD {
17364 my ( $self, $value ) = @_;
17365 if ( defined($value) ) {
17366 $self->[HAVE_CHILD] = $value;
17368 return $self->[HAVE_CHILD];
17371 sub get_HAVE_CHILD {
17373 return $self->[HAVE_CHILD];
17376 sub set_ARROW_COUNT {
17377 my ( $self, $value ) = @_;
17378 if ( defined($value) ) {
17379 $self->[ARROW_COUNT] = $value;
17381 return $self->[ARROW_COUNT];
17384 sub get_ARROW_COUNT {
17386 return $self->[ARROW_COUNT];
17389 sub set_COMMA_COUNT {
17390 my ( $self, $value ) = @_;
17391 if ( defined($value) ) {
17392 $self->[COMMA_COUNT] = $value;
17394 return $self->[COMMA_COUNT];
17397 sub get_COMMA_COUNT {
17399 return $self->[COMMA_COUNT];
17403 my ( $self, $value ) = @_;
17404 if ( defined($value) ) {
17405 $self->[CLOSED] = $value;
17407 return $self->[CLOSED];
17412 return $self->[CLOSED];
17415 #####################################################################
17417 # the Perl::Tidy::VerticalAligner::Line class supplies an object to
17418 # contain a single output line
17420 #####################################################################
17422 package Perl::Tidy::VerticalAligner::Line;
17429 use constant JMAX => 0;
17430 use constant JMAX_ORIGINAL_LINE => 1;
17431 use constant RTOKENS => 2;
17432 use constant RFIELDS => 3;
17433 use constant RPATTERNS => 4;
17434 use constant INDENTATION => 5;
17435 use constant LEADING_SPACE_COUNT => 6;
17436 use constant OUTDENT_LONG_LINES => 7;
17437 use constant LIST_TYPE => 8;
17438 use constant IS_HANGING_SIDE_COMMENT => 9;
17439 use constant RALIGNMENTS => 10;
17440 use constant MAXIMUM_LINE_LENGTH => 11;
17441 use constant RVERTICAL_TIGHTNESS_FLAGS => 12;
17444 $_index_map{jmax} = JMAX;
17445 $_index_map{jmax_original_line} = JMAX_ORIGINAL_LINE;
17446 $_index_map{rtokens} = RTOKENS;
17447 $_index_map{rfields} = RFIELDS;
17448 $_index_map{rpatterns} = RPATTERNS;
17449 $_index_map{indentation} = INDENTATION;
17450 $_index_map{leading_space_count} = LEADING_SPACE_COUNT;
17451 $_index_map{outdent_long_lines} = OUTDENT_LONG_LINES;
17452 $_index_map{list_type} = LIST_TYPE;
17453 $_index_map{is_hanging_side_comment} = IS_HANGING_SIDE_COMMENT;
17454 $_index_map{ralignments} = RALIGNMENTS;
17455 $_index_map{maximum_line_length} = MAXIMUM_LINE_LENGTH;
17456 $_index_map{rvertical_tightness_flags} = RVERTICAL_TIGHTNESS_FLAGS;
17458 my @_default_data = ();
17459 $_default_data[JMAX] = undef;
17460 $_default_data[JMAX_ORIGINAL_LINE] = undef;
17461 $_default_data[RTOKENS] = undef;
17462 $_default_data[RFIELDS] = undef;
17463 $_default_data[RPATTERNS] = undef;
17464 $_default_data[INDENTATION] = undef;
17465 $_default_data[LEADING_SPACE_COUNT] = undef;
17466 $_default_data[OUTDENT_LONG_LINES] = undef;
17467 $_default_data[LIST_TYPE] = undef;
17468 $_default_data[IS_HANGING_SIDE_COMMENT] = undef;
17469 $_default_data[RALIGNMENTS] = [];
17470 $_default_data[MAXIMUM_LINE_LENGTH] = undef;
17471 $_default_data[RVERTICAL_TIGHTNESS_FLAGS] = undef;
17475 # methods to count object population
17477 sub get_count { $_count; }
17478 sub _increment_count { ++$_count }
17479 sub _decrement_count { --$_count }
17482 # Constructor may be called as a class method
17484 my ( $caller, %arg ) = @_;
17485 my $caller_is_obj = ref($caller);
17486 my $class = $caller_is_obj || $caller;
17488 my $self = bless [], $class;
17490 $self->[RALIGNMENTS] = [];
17493 foreach ( keys %_index_map ) {
17494 $index = $_index_map{$_};
17495 if ( exists $arg{$_} ) { $self->[$index] = $arg{$_} }
17496 elsif ($caller_is_obj) { $self->[$index] = $caller->[$index] }
17497 else { $self->[$index] = $_default_data[$index] }
17500 $self->_increment_count();
17505 $_[0]->_decrement_count();
17508 sub get_jmax { $_[0]->[JMAX] }
17509 sub get_jmax_original_line { $_[0]->[JMAX_ORIGINAL_LINE] }
17510 sub get_rtokens { $_[0]->[RTOKENS] }
17511 sub get_rfields { $_[0]->[RFIELDS] }
17512 sub get_rpatterns { $_[0]->[RPATTERNS] }
17513 sub get_indentation { $_[0]->[INDENTATION] }
17514 sub get_leading_space_count { $_[0]->[LEADING_SPACE_COUNT] }
17515 sub get_outdent_long_lines { $_[0]->[OUTDENT_LONG_LINES] }
17516 sub get_list_type { $_[0]->[LIST_TYPE] }
17517 sub get_is_hanging_side_comment { $_[0]->[IS_HANGING_SIDE_COMMENT] }
17518 sub get_rvertical_tightness_flags { $_[0]->[RVERTICAL_TIGHTNESS_FLAGS] }
17520 sub set_column { $_[0]->[RALIGNMENTS]->[ $_[1] ]->set_column( $_[2] ) }
17521 sub get_alignment { $_[0]->[RALIGNMENTS]->[ $_[1] ] }
17522 sub get_alignments { @{ $_[0]->[RALIGNMENTS] } }
17523 sub get_column { $_[0]->[RALIGNMENTS]->[ $_[1] ]->get_column() }
17525 sub get_starting_column {
17526 $_[0]->[RALIGNMENTS]->[ $_[1] ]->get_starting_column();
17529 sub increment_column {
17530 $_[0]->[RALIGNMENTS]->[ $_[1] ]->increment_column( $_[2] );
17532 sub set_alignments { my $self = shift; @{ $self->[RALIGNMENTS] } = @_; }
17534 sub current_field_width {
17538 return $self->get_column($j);
17541 return $self->get_column($j) - $self->get_column( $j - 1 );
17545 sub field_width_growth {
17548 return $self->get_column($j) - $self->get_starting_column($j);
17551 sub starting_field_width {
17555 return $self->get_starting_column($j);
17558 return $self->get_starting_column($j) -
17559 $self->get_starting_column( $j - 1 );
17563 sub increase_field_width {
17566 my ( $j, $pad ) = @_;
17567 my $jmax = $self->get_jmax();
17568 for my $k ( $j .. $jmax ) {
17569 $self->increment_column( $k, $pad );
17573 sub get_available_space_on_right {
17575 my $jmax = $self->get_jmax();
17576 return $self->[MAXIMUM_LINE_LENGTH] - $self->get_column($jmax);
17579 sub set_jmax { $_[0]->[JMAX] = $_[1] }
17580 sub set_jmax_original_line { $_[0]->[JMAX_ORIGINAL_LINE] = $_[1] }
17581 sub set_rtokens { $_[0]->[RTOKENS] = $_[1] }
17582 sub set_rfields { $_[0]->[RFIELDS] = $_[1] }
17583 sub set_rpatterns { $_[0]->[RPATTERNS] = $_[1] }
17584 sub set_indentation { $_[0]->[INDENTATION] = $_[1] }
17585 sub set_leading_space_count { $_[0]->[LEADING_SPACE_COUNT] = $_[1] }
17586 sub set_outdent_long_lines { $_[0]->[OUTDENT_LONG_LINES] = $_[1] }
17587 sub set_list_type { $_[0]->[LIST_TYPE] = $_[1] }
17588 sub set_is_hanging_side_comment { $_[0]->[IS_HANGING_SIDE_COMMENT] = $_[1] }
17589 sub set_alignment { $_[0]->[RALIGNMENTS]->[ $_[1] ] = $_[2] }
17593 #####################################################################
17595 # the Perl::Tidy::VerticalAligner::Alignment class holds information
17596 # on a single column being aligned
17598 #####################################################################
17599 package Perl::Tidy::VerticalAligner::Alignment;
17607 # Symbolic array indexes
17608 use constant COLUMN => 0; # the current column number
17609 use constant STARTING_COLUMN => 1; # column number when created
17610 use constant MATCHING_TOKEN => 2; # what token we are matching
17611 use constant STARTING_LINE => 3; # the line index of creation
17612 use constant ENDING_LINE => 4; # the most recent line to use it
17613 use constant SAVED_COLUMN => 5; # the most recent line to use it
17614 use constant SERIAL_NUMBER => 6; # unique number for this alignment
17615 # (just its index in an array)
17617 # Correspondence between variables and array indexes
17619 $_index_map{column} = COLUMN;
17620 $_index_map{starting_column} = STARTING_COLUMN;
17621 $_index_map{matching_token} = MATCHING_TOKEN;
17622 $_index_map{starting_line} = STARTING_LINE;
17623 $_index_map{ending_line} = ENDING_LINE;
17624 $_index_map{saved_column} = SAVED_COLUMN;
17625 $_index_map{serial_number} = SERIAL_NUMBER;
17627 my @_default_data = ();
17628 $_default_data[COLUMN] = undef;
17629 $_default_data[STARTING_COLUMN] = undef;
17630 $_default_data[MATCHING_TOKEN] = undef;
17631 $_default_data[STARTING_LINE] = undef;
17632 $_default_data[ENDING_LINE] = undef;
17633 $_default_data[SAVED_COLUMN] = undef;
17634 $_default_data[SERIAL_NUMBER] = undef;
17636 # class population count
17639 sub get_count { $_count; }
17640 sub _increment_count { ++$_count }
17641 sub _decrement_count { --$_count }
17646 my ( $caller, %arg ) = @_;
17647 my $caller_is_obj = ref($caller);
17648 my $class = $caller_is_obj || $caller;
17650 my $self = bless [], $class;
17652 foreach ( keys %_index_map ) {
17653 my $index = $_index_map{$_};
17654 if ( exists $arg{$_} ) { $self->[$index] = $arg{$_} }
17655 elsif ($caller_is_obj) { $self->[$index] = $caller->[$index] }
17656 else { $self->[$index] = $_default_data[$index] }
17658 $self->_increment_count();
17663 $_[0]->_decrement_count();
17666 sub get_column { return $_[0]->[COLUMN] }
17667 sub get_starting_column { return $_[0]->[STARTING_COLUMN] }
17668 sub get_matching_token { return $_[0]->[MATCHING_TOKEN] }
17669 sub get_starting_line { return $_[0]->[STARTING_LINE] }
17670 sub get_ending_line { return $_[0]->[ENDING_LINE] }
17671 sub get_serial_number { return $_[0]->[SERIAL_NUMBER] }
17673 sub set_column { $_[0]->[COLUMN] = $_[1] }
17674 sub set_starting_column { $_[0]->[STARTING_COLUMN] = $_[1] }
17675 sub set_matching_token { $_[0]->[MATCHING_TOKEN] = $_[1] }
17676 sub set_starting_line { $_[0]->[STARTING_LINE] = $_[1] }
17677 sub set_ending_line { $_[0]->[ENDING_LINE] = $_[1] }
17678 sub increment_column { $_[0]->[COLUMN] += $_[1] }
17680 sub save_column { $_[0]->[SAVED_COLUMN] = $_[0]->[COLUMN] }
17681 sub restore_column { $_[0]->[COLUMN] = $_[0]->[SAVED_COLUMN] }
17685 package Perl::Tidy::VerticalAligner;
17687 # The Perl::Tidy::VerticalAligner package collects output lines and
17688 # attempts to line up certain common tokens, such as => and #, which are
17689 # identified by the calling routine.
17691 # There are two main routines: append_line and flush. Append acts as a
17692 # storage buffer, collecting lines into a group which can be vertically
17693 # aligned. When alignment is no longer possible or desirable, it dumps
17694 # the group to flush.
17696 # append_line -----> flush
17704 # Caution: these debug flags produce a lot of output
17705 # They should all be 0 except when debugging small scripts
17707 use constant VALIGN_DEBUG_FLAG_APPEND => 0;
17708 use constant VALIGN_DEBUG_FLAG_APPEND0 => 0;
17709 use constant VALIGN_DEBUG_FLAG_TERNARY => 0;
17711 my $debug_warning = sub {
17712 print "VALIGN_DEBUGGING with key $_[0]\n";
17715 VALIGN_DEBUG_FLAG_APPEND && $debug_warning->('APPEND');
17716 VALIGN_DEBUG_FLAG_APPEND0 && $debug_warning->('APPEND0');
17721 $vertical_aligner_self
17723 $maximum_alignment_index
17727 $previous_minimum_jmax_seen
17728 $previous_maximum_jmax_seen
17729 $maximum_line_index
17734 $last_group_level_written
17735 $last_leading_space_count
17739 $last_comment_column
17740 $last_side_comment_line_number
17741 $last_side_comment_length
17742 $last_side_comment_level
17743 $outdented_line_count
17744 $first_outdented_line_at
17745 $last_outdented_line_at
17746 $diagnostics_object
17748 $file_writer_object
17749 @side_comment_history
17750 $comment_leading_space_count
17751 $is_matching_terminal_line
17758 $cached_line_leading_space_count
17759 $cached_seqno_string
17762 $last_nonblank_seqno_string
17766 $rOpts_maximum_line_length
17767 $rOpts_continuation_indentation
17768 $rOpts_indent_columns
17770 $rOpts_entab_leading_whitespace
17773 $rOpts_fixed_position_side_comment
17774 $rOpts_minimum_space_to_comment
17782 ( $class, $rOpts, $file_writer_object, $logger_object, $diagnostics_object )
17785 # variables describing the entire space group:
17786 $ralignment_list = [];
17788 $last_group_level_written = -1;
17789 $extra_indent_ok = 0; # can we move all lines to the right?
17790 $last_side_comment_length = 0;
17791 $maximum_jmax_seen = 0;
17792 $minimum_jmax_seen = 0;
17793 $previous_minimum_jmax_seen = 0;
17794 $previous_maximum_jmax_seen = 0;
17796 # variables describing each line of the group
17797 @group_lines = (); # list of all lines in group
17799 $outdented_line_count = 0;
17800 $first_outdented_line_at = 0;
17801 $last_outdented_line_at = 0;
17802 $last_side_comment_line_number = 0;
17803 $last_side_comment_level = -1;
17804 $is_matching_terminal_line = 0;
17806 # most recent 3 side comments; [ line number, column ]
17807 $side_comment_history[0] = [ -300, 0 ];
17808 $side_comment_history[1] = [ -200, 0 ];
17809 $side_comment_history[2] = [ -100, 0 ];
17811 # write_leader_and_string cache:
17812 $cached_line_text = "";
17813 $cached_line_type = 0;
17814 $cached_line_flag = 0;
17816 $cached_line_valid = 0;
17817 $cached_line_leading_space_count = 0;
17818 $cached_seqno_string = "";
17820 # string of sequence numbers joined together
17821 $seqno_string = "";
17822 $last_nonblank_seqno_string = "";
17824 # frequently used parameters
17825 $rOpts_indent_columns = $rOpts->{'indent-columns'};
17826 $rOpts_tabs = $rOpts->{'tabs'};
17827 $rOpts_entab_leading_whitespace = $rOpts->{'entab-leading-whitespace'};
17828 $rOpts_fixed_position_side_comment =
17829 $rOpts->{'fixed-position-side-comment'};
17830 $rOpts_minimum_space_to_comment = $rOpts->{'minimum-space-to-comment'};
17831 $rOpts_maximum_line_length = $rOpts->{'maximum-line-length'};
17832 $rOpts_valign = $rOpts->{'valign'};
17834 forget_side_comment();
17836 initialize_for_new_group();
17838 $vertical_aligner_self = {};
17839 bless $vertical_aligner_self, $class;
17840 return $vertical_aligner_self;
17843 sub initialize_for_new_group {
17844 $maximum_line_index = -1; # lines in the current group
17845 $maximum_alignment_index = -1; # alignments in current group
17846 $zero_count = 0; # count consecutive lines without tokens
17847 $current_line = undef; # line being matched for alignment
17848 $group_maximum_gap = 0; # largest gap introduced
17850 $marginal_match = 0;
17851 $comment_leading_space_count = 0;
17852 $last_leading_space_count = 0;
17855 # interface to Perl::Tidy::Diagnostics routines
17856 sub write_diagnostics {
17857 if ($diagnostics_object) {
17858 $diagnostics_object->write_diagnostics(@_);
17862 # interface to Perl::Tidy::Logger routines
17864 if ($logger_object) {
17865 $logger_object->warning(@_);
17869 sub write_logfile_entry {
17870 if ($logger_object) {
17871 $logger_object->write_logfile_entry(@_);
17875 sub report_definite_bug {
17876 if ($logger_object) {
17877 $logger_object->report_definite_bug();
17883 # return the number of leading spaces associated with an indentation
17884 # variable $indentation is either a constant number of spaces or an
17885 # object with a get_SPACES method.
17886 my $indentation = shift;
17887 return ref($indentation) ? $indentation->get_SPACES() : $indentation;
17890 sub get_RECOVERABLE_SPACES {
17892 # return the number of spaces (+ means shift right, - means shift left)
17893 # that we would like to shift a group of lines with the same indentation
17894 # to get them to line up with their opening parens
17895 my $indentation = shift;
17896 return ref($indentation) ? $indentation->get_RECOVERABLE_SPACES() : 0;
17899 sub get_STACK_DEPTH {
17901 my $indentation = shift;
17902 return ref($indentation) ? $indentation->get_STACK_DEPTH() : 0;
17905 sub make_alignment {
17906 my ( $col, $token ) = @_;
17908 # make one new alignment at column $col which aligns token $token
17909 ++$maximum_alignment_index;
17910 my $alignment = new Perl::Tidy::VerticalAligner::Alignment(
17912 starting_column => $col,
17913 matching_token => $token,
17914 starting_line => $maximum_line_index,
17915 ending_line => $maximum_line_index,
17916 serial_number => $maximum_alignment_index,
17918 $ralignment_list->[$maximum_alignment_index] = $alignment;
17922 sub dump_alignments {
17924 "Current Alignments:\ni\ttoken\tstarting_column\tcolumn\tstarting_line\tending_line\n";
17925 for my $i ( 0 .. $maximum_alignment_index ) {
17926 my $column = $ralignment_list->[$i]->get_column();
17927 my $starting_column = $ralignment_list->[$i]->get_starting_column();
17928 my $matching_token = $ralignment_list->[$i]->get_matching_token();
17929 my $starting_line = $ralignment_list->[$i]->get_starting_line();
17930 my $ending_line = $ralignment_list->[$i]->get_ending_line();
17932 "$i\t$matching_token\t$starting_column\t$column\t$starting_line\t$ending_line\n";
17936 sub save_alignment_columns {
17937 for my $i ( 0 .. $maximum_alignment_index ) {
17938 $ralignment_list->[$i]->save_column();
17942 sub restore_alignment_columns {
17943 for my $i ( 0 .. $maximum_alignment_index ) {
17944 $ralignment_list->[$i]->restore_column();
17948 sub forget_side_comment {
17949 $last_comment_column = 0;
17954 # sub append is called to place one line in the current vertical group.
17956 # The input parameters are:
17957 # $level = indentation level of this line
17958 # $rfields = reference to array of fields
17959 # $rpatterns = reference to array of patterns, one per field
17960 # $rtokens = reference to array of tokens starting fields 1,2,..
17962 # Here is an example of what this package does. In this example,
17963 # we are trying to line up both the '=>' and the '#'.
17965 # '18' => 'grave', # \`
17966 # '19' => 'acute', # `'
17967 # '20' => 'caron', # \v
17968 # <-tabs-><f1-><--field 2 ---><-f3->
17971 # col1 col2 col3 col4
17973 # The calling routine has already broken the entire line into 3 fields as
17974 # indicated. (So the work of identifying promising common tokens has
17975 # already been done).
17977 # In this example, there will be 2 tokens being matched: '=>' and '#'.
17978 # They are the leading parts of fields 2 and 3, but we do need to know
17979 # what they are so that we can dump a group of lines when these tokens
17982 # The fields contain the actual characters of each field. The patterns
17983 # are like the fields, but they contain mainly token types instead
17984 # of tokens, so they have fewer characters. They are used to be
17985 # sure we are matching fields of similar type.
17987 # In this example, there will be 4 column indexes being adjusted. The
17988 # first one is always at zero. The interior columns are at the start of
17989 # the matching tokens, and the last one tracks the maximum line length.
17991 # Basically, each time a new line comes in, it joins the current vertical
17992 # group if possible. Otherwise it causes the current group to be dumped
17993 # and a new group is started.
17995 # For each new group member, the column locations are increased, as
17996 # necessary, to make room for the new fields. When the group is finally
17997 # output, these column numbers are used to compute the amount of spaces of
17998 # padding needed for each field.
18000 # Programming note: the fields are assumed not to have any tab characters.
18001 # Tabs have been previously removed except for tabs in quoted strings and
18002 # side comments. Tabs in these fields can mess up the column counting.
18003 # The log file warns the user if there are any such tabs.
18006 $level, $level_end,
18007 $indentation, $rfields,
18008 $rtokens, $rpatterns,
18009 $is_forced_break, $outdent_long_lines,
18010 $is_terminal_ternary, $is_terminal_statement,
18011 $do_not_pad, $rvertical_tightness_flags,
18015 # number of fields is $jmax
18016 # number of tokens between fields is $jmax-1
18017 my $jmax = $#{$rfields};
18019 my $leading_space_count = get_SPACES($indentation);
18021 # set outdented flag to be sure we either align within statements or
18022 # across statement boundaries, but not both.
18023 my $is_outdented = $last_leading_space_count > $leading_space_count;
18024 $last_leading_space_count = $leading_space_count;
18026 # Patch: undo for hanging side comment
18027 my $is_hanging_side_comment =
18028 ( $jmax == 1 && $rtokens->[0] eq '#' && $rfields->[0] =~ /^\s*$/ );
18029 $is_outdented = 0 if $is_hanging_side_comment;
18031 VALIGN_DEBUG_FLAG_APPEND0 && do {
18033 "APPEND0: entering lines=$maximum_line_index new #fields= $jmax, leading_count=$leading_space_count last_cmt=$last_comment_column force=$is_forced_break\n";
18036 # Validate cached line if necessary: If we can produce a container
18037 # with just 2 lines total by combining an existing cached opening
18038 # token with the closing token to follow, then we will mark both
18039 # cached flags as valid.
18040 if ($rvertical_tightness_flags) {
18041 if ( $maximum_line_index <= 0
18042 && $cached_line_type
18044 && $rvertical_tightness_flags->[2]
18045 && $rvertical_tightness_flags->[2] == $cached_seqno )
18047 $rvertical_tightness_flags->[3] ||= 1;
18048 $cached_line_valid ||= 1;
18052 # do not join an opening block brace with an unbalanced line
18053 # unless requested with a flag value of 2
18054 if ( $cached_line_type == 3
18055 && $maximum_line_index < 0
18056 && $cached_line_flag < 2
18057 && $level_jump != 0 )
18059 $cached_line_valid = 0;
18062 # patch until new aligner is finished
18063 if ($do_not_pad) { my_flush() }
18065 # shouldn't happen:
18066 if ( $level < 0 ) { $level = 0 }
18068 # do not align code across indentation level changes
18069 # or if vertical alignment is turned off for debugging
18070 if ( $level != $group_level || $is_outdented || !$rOpts_valign ) {
18072 # we are allowed to shift a group of lines to the right if its
18073 # level is greater than the previous and next group
18075 ( $level < $group_level && $last_group_level_written < $group_level );
18079 # If we know that this line will get flushed out by itself because
18080 # of level changes, we can leave the extra_indent_ok flag set.
18081 # That way, if we get an external flush call, we will still be
18082 # able to do some -lp alignment if necessary.
18083 $extra_indent_ok = ( $is_terminal_statement && $level > $group_level );
18085 $group_level = $level;
18087 # wait until after the above flush to get the leading space
18088 # count because it may have been changed if the -icp flag is in
18090 $leading_space_count = get_SPACES($indentation);
18094 # --------------------------------------------------------------------
18095 # Patch to collect outdentable block COMMENTS
18096 # --------------------------------------------------------------------
18097 my $is_blank_line = "";
18098 my $is_block_comment = ( $jmax == 0 && $rfields->[0] =~ /^#/ );
18099 if ( $group_type eq 'COMMENT' ) {
18103 && $outdent_long_lines
18104 && $leading_space_count == $comment_leading_space_count
18109 $group_lines[ ++$maximum_line_index ] = $rfields->[0];
18117 # --------------------------------------------------------------------
18118 # add dummy fields for terminal ternary
18119 # --------------------------------------------------------------------
18120 my $j_terminal_match;
18121 if ( $is_terminal_ternary && $current_line ) {
18122 $j_terminal_match =
18123 fix_terminal_ternary( $rfields, $rtokens, $rpatterns );
18124 $jmax = @{$rfields} - 1;
18127 # --------------------------------------------------------------------
18128 # add dummy fields for else statement
18129 # --------------------------------------------------------------------
18130 if ( $rfields->[0] =~ /^else\s*$/
18132 && $level_jump == 0 )
18134 $j_terminal_match = fix_terminal_else( $rfields, $rtokens, $rpatterns );
18135 $jmax = @{$rfields} - 1;
18138 # --------------------------------------------------------------------
18139 # Step 1. Handle simple line of code with no fields to match.
18140 # --------------------------------------------------------------------
18141 if ( $jmax <= 0 ) {
18144 if ( $maximum_line_index >= 0
18145 && !get_RECOVERABLE_SPACES( $group_lines[0]->get_indentation() ) )
18148 # flush the current group if it has some aligned columns..
18149 if ( $group_lines[0]->get_jmax() > 1 ) { my_flush() }
18151 # flush current group if we are just collecting side comments..
18154 # ...and we haven't seen a comment lately
18155 ( $zero_count > 3 )
18157 # ..or if this new line doesn't fit to the left of the comments
18158 || ( ( $leading_space_count + length( $$rfields[0] ) ) >
18159 $group_lines[0]->get_column(0) )
18166 # patch to start new COMMENT group if this comment may be outdented
18167 if ( $is_block_comment
18168 && $outdent_long_lines
18169 && $maximum_line_index < 0 )
18171 $group_type = 'COMMENT';
18172 $comment_leading_space_count = $leading_space_count;
18173 $group_lines[ ++$maximum_line_index ] = $rfields->[0];
18177 # just write this line directly if no current group, no side comment,
18178 # and no space recovery is needed.
18179 if ( $maximum_line_index < 0 && !get_RECOVERABLE_SPACES($indentation) )
18181 write_leader_and_string( $leading_space_count, $$rfields[0], 0,
18182 $outdent_long_lines, $rvertical_tightness_flags );
18190 # programming check: (shouldn't happen)
18191 # an error here implies an incorrect call was made
18192 if ( $jmax > 0 && ( $#{$rtokens} != ( $jmax - 1 ) ) ) {
18194 "Program bug in Perl::Tidy::VerticalAligner - number of tokens = $#{$rtokens} should be one less than number of fields: $#{$rfields})\n"
18196 report_definite_bug();
18199 # --------------------------------------------------------------------
18200 # create an object to hold this line
18201 # --------------------------------------------------------------------
18202 my $new_line = new Perl::Tidy::VerticalAligner::Line(
18204 jmax_original_line => $jmax,
18205 rtokens => $rtokens,
18206 rfields => $rfields,
18207 rpatterns => $rpatterns,
18208 indentation => $indentation,
18209 leading_space_count => $leading_space_count,
18210 outdent_long_lines => $outdent_long_lines,
18212 is_hanging_side_comment => $is_hanging_side_comment,
18213 maximum_line_length => $rOpts->{'maximum-line-length'},
18214 rvertical_tightness_flags => $rvertical_tightness_flags,
18217 # Initialize a global flag saying if the last line of the group should
18218 # match end of group and also terminate the group. There should be no
18219 # returns between here and where the flag is handled at the bottom.
18220 my $col_matching_terminal = 0;
18221 if ( defined($j_terminal_match) ) {
18223 # remember the column of the terminal ? or { to match with
18224 $col_matching_terminal = $current_line->get_column($j_terminal_match);
18226 # set global flag for sub decide_if_aligned
18227 $is_matching_terminal_line = 1;
18230 # --------------------------------------------------------------------
18231 # It simplifies things to create a zero length side comment
18233 # --------------------------------------------------------------------
18234 make_side_comment( $new_line, $level_end );
18236 # --------------------------------------------------------------------
18237 # Decide if this is a simple list of items.
18238 # There are 3 list types: none, comma, comma-arrow.
18239 # We use this below to be less restrictive in deciding what to align.
18240 # --------------------------------------------------------------------
18241 if ($is_forced_break) {
18242 decide_if_list($new_line);
18245 if ($current_line) {
18247 # --------------------------------------------------------------------
18248 # Allow hanging side comment to join current group, if any
18249 # This will help keep side comments aligned, because otherwise we
18250 # will have to start a new group, making alignment less likely.
18251 # --------------------------------------------------------------------
18252 join_hanging_comment( $new_line, $current_line )
18253 if $is_hanging_side_comment;
18255 # --------------------------------------------------------------------
18256 # If there is just one previous line, and it has more fields
18257 # than the new line, try to join fields together to get a match with
18258 # the new line. At the present time, only a single leading '=' is
18259 # allowed to be compressed out. This is useful in rare cases where
18260 # a table is forced to use old breakpoints because of side comments,
18261 # and the table starts out something like this:
18262 # my %MonthChars = ('0', 'Jan', # side comment
18265 # Eliminating the '=' field will allow the remaining fields to line up.
18266 # This situation does not occur if there are no side comments
18267 # because scan_list would put a break after the opening '('.
18268 # --------------------------------------------------------------------
18269 eliminate_old_fields( $new_line, $current_line );
18271 # --------------------------------------------------------------------
18272 # If the new line has more fields than the current group,
18273 # see if we can match the first fields and combine the remaining
18274 # fields of the new line.
18275 # --------------------------------------------------------------------
18276 eliminate_new_fields( $new_line, $current_line );
18278 # --------------------------------------------------------------------
18279 # Flush previous group unless all common tokens and patterns match..
18280 # --------------------------------------------------------------------
18281 check_match( $new_line, $current_line );
18283 # --------------------------------------------------------------------
18284 # See if there is space for this line in the current group (if any)
18285 # --------------------------------------------------------------------
18286 if ($current_line) {
18287 check_fit( $new_line, $current_line );
18291 # --------------------------------------------------------------------
18292 # Append this line to the current group (or start new group)
18293 # --------------------------------------------------------------------
18294 accept_line($new_line);
18296 # Future update to allow this to vary:
18297 $current_line = $new_line if ( $maximum_line_index == 0 );
18299 # output this group if it ends in a terminal else or ternary line
18300 if ( defined($j_terminal_match) ) {
18302 # if there is only one line in the group (maybe due to failure to match
18303 # perfectly with previous lines), then align the ? or { of this
18304 # terminal line with the previous one unless that would make the line
18306 if ( $maximum_line_index == 0 ) {
18307 my $col_now = $current_line->get_column($j_terminal_match);
18308 my $pad = $col_matching_terminal - $col_now;
18309 my $padding_available =
18310 $current_line->get_available_space_on_right();
18311 if ( $pad > 0 && $pad <= $padding_available ) {
18312 $current_line->increase_field_width( $j_terminal_match, $pad );
18316 $is_matching_terminal_line = 0;
18319 # --------------------------------------------------------------------
18320 # Step 8. Some old debugging stuff
18321 # --------------------------------------------------------------------
18322 VALIGN_DEBUG_FLAG_APPEND && do {
18323 print "APPEND fields:";
18324 dump_array(@$rfields);
18325 print "APPEND tokens:";
18326 dump_array(@$rtokens);
18327 print "APPEND patterns:";
18328 dump_array(@$rpatterns);
18335 sub join_hanging_comment {
18338 my $jmax = $line->get_jmax();
18339 return 0 unless $jmax == 1; # must be 2 fields
18340 my $rtokens = $line->get_rtokens();
18341 return 0 unless $$rtokens[0] eq '#'; # the second field is a comment..
18342 my $rfields = $line->get_rfields();
18343 return 0 unless $$rfields[0] =~ /^\s*$/; # the first field is empty...
18344 my $old_line = shift;
18345 my $maximum_field_index = $old_line->get_jmax();
18347 unless $maximum_field_index > $jmax; # the current line has more fields
18348 my $rpatterns = $line->get_rpatterns();
18350 $line->set_is_hanging_side_comment(1);
18351 $jmax = $maximum_field_index;
18352 $line->set_jmax($jmax);
18353 $$rfields[$jmax] = $$rfields[1];
18354 $$rtokens[ $jmax - 1 ] = $$rtokens[0];
18355 $$rpatterns[ $jmax - 1 ] = $$rpatterns[0];
18356 for ( my $j = 1 ; $j < $jmax ; $j++ ) {
18357 $$rfields[$j] = " "; # NOTE: caused glitch unless 1 blank, why?
18358 $$rtokens[ $j - 1 ] = "";
18359 $$rpatterns[ $j - 1 ] = "";
18364 sub eliminate_old_fields {
18366 my $new_line = shift;
18367 my $jmax = $new_line->get_jmax();
18368 if ( $jmax > $maximum_jmax_seen ) { $maximum_jmax_seen = $jmax }
18369 if ( $jmax < $minimum_jmax_seen ) { $minimum_jmax_seen = $jmax }
18371 # there must be one previous line
18372 return unless ( $maximum_line_index == 0 );
18374 my $old_line = shift;
18375 my $maximum_field_index = $old_line->get_jmax();
18377 ###############################################
18378 # this line must have fewer fields
18379 return unless $maximum_field_index > $jmax;
18380 ###############################################
18382 # Identify specific cases where field elimination is allowed:
18383 # case=1: both lines have comma-separated lists, and the first
18384 # line has an equals
18385 # case=2: both lines have leading equals
18387 # case 1 is the default
18390 # See if case 2: both lines have leading '='
18391 # We'll require smiliar leading patterns in this case
18392 my $old_rtokens = $old_line->get_rtokens();
18393 my $rtokens = $new_line->get_rtokens();
18394 my $rpatterns = $new_line->get_rpatterns();
18395 my $old_rpatterns = $old_line->get_rpatterns();
18396 if ( $rtokens->[0] =~ /^=\d*$/
18397 && $old_rtokens->[0] eq $rtokens->[0]
18398 && $old_rpatterns->[0] eq $rpatterns->[0] )
18403 # not too many fewer fields in new line for case 1
18404 return unless ( $case != 1 || $maximum_field_index - 2 <= $jmax );
18406 # case 1 must have side comment
18407 my $old_rfields = $old_line->get_rfields();
18410 && length( $$old_rfields[$maximum_field_index] ) == 0 );
18412 my $rfields = $new_line->get_rfields();
18414 my $hid_equals = 0;
18416 my @new_alignments = ();
18417 my @new_fields = ();
18418 my @new_matching_patterns = ();
18419 my @new_matching_tokens = ();
18423 my $current_field = '';
18424 my $current_pattern = '';
18426 # loop over all old tokens
18428 for ( $k = 0 ; $k < $maximum_field_index ; $k++ ) {
18429 $current_field .= $$old_rfields[$k];
18430 $current_pattern .= $$old_rpatterns[$k];
18431 last if ( $j > $jmax - 1 );
18433 if ( $$old_rtokens[$k] eq $$rtokens[$j] ) {
18435 $new_fields[$j] = $current_field;
18436 $new_matching_patterns[$j] = $current_pattern;
18437 $current_field = '';
18438 $current_pattern = '';
18439 $new_matching_tokens[$j] = $$old_rtokens[$k];
18440 $new_alignments[$j] = $old_line->get_alignment($k);
18445 if ( $$old_rtokens[$k] =~ /^\=\d*$/ ) {
18446 last if ( $case == 2 ); # avoid problems with stuff
18447 # like: $a=$b=$c=$d;
18451 if ( $in_match && $case == 1 )
18452 ; # disallow gaps in matching field types in case 1
18456 # Modify the current state if we are successful.
18457 # We must exactly reach the ends of both lists for success.
18458 if ( ( $j == $jmax )
18459 && ( $current_field eq '' )
18460 && ( $case != 1 || $hid_equals ) )
18462 $k = $maximum_field_index;
18463 $current_field .= $$old_rfields[$k];
18464 $current_pattern .= $$old_rpatterns[$k];
18465 $new_fields[$j] = $current_field;
18466 $new_matching_patterns[$j] = $current_pattern;
18468 $new_alignments[$j] = $old_line->get_alignment($k);
18469 $maximum_field_index = $j;
18471 $old_line->set_alignments(@new_alignments);
18472 $old_line->set_jmax($jmax);
18473 $old_line->set_rtokens( \@new_matching_tokens );
18474 $old_line->set_rfields( \@new_fields );
18475 $old_line->set_rpatterns( \@$rpatterns );
18479 # create an empty side comment if none exists
18480 sub make_side_comment {
18481 my $new_line = shift;
18482 my $level_end = shift;
18483 my $jmax = $new_line->get_jmax();
18484 my $rtokens = $new_line->get_rtokens();
18486 # if line does not have a side comment...
18487 if ( ( $jmax == 0 ) || ( $$rtokens[ $jmax - 1 ] ne '#' ) ) {
18488 my $rfields = $new_line->get_rfields();
18489 my $rpatterns = $new_line->get_rpatterns();
18490 $$rtokens[$jmax] = '#';
18491 $$rfields[ ++$jmax ] = '';
18492 $$rpatterns[$jmax] = '#';
18493 $new_line->set_jmax($jmax);
18494 $new_line->set_jmax_original_line($jmax);
18497 # line has a side comment..
18500 # don't remember old side comment location for very long
18501 my $line_number = $vertical_aligner_self->get_output_line_number();
18502 my $rfields = $new_line->get_rfields();
18504 $line_number - $last_side_comment_line_number > 12
18506 # and don't remember comment location across block level changes
18507 || ( $level_end < $last_side_comment_level && $$rfields[0] =~ /^}/ )
18510 forget_side_comment();
18512 $last_side_comment_line_number = $line_number;
18513 $last_side_comment_level = $level_end;
18517 sub decide_if_list {
18521 # A list will be taken to be a line with a forced break in which all
18522 # of the field separators are commas or comma-arrows (except for the
18525 # List separator tokens are things like ',3' or '=>2',
18526 # where the trailing digit is the nesting depth. Allow braces
18527 # to allow nested list items.
18528 my $rtokens = $line->get_rtokens();
18529 my $test_token = $$rtokens[0];
18530 if ( $test_token =~ /^(\,|=>)/ ) {
18531 my $list_type = $test_token;
18532 my $jmax = $line->get_jmax();
18534 foreach ( 1 .. $jmax - 2 ) {
18535 if ( $$rtokens[$_] !~ /^(\,|=>|\{)/ ) {
18540 $line->set_list_type($list_type);
18544 sub eliminate_new_fields {
18546 return unless ( $maximum_line_index >= 0 );
18547 my ( $new_line, $old_line ) = @_;
18548 my $jmax = $new_line->get_jmax();
18550 my $old_rtokens = $old_line->get_rtokens();
18551 my $rtokens = $new_line->get_rtokens();
18552 my $is_assignment =
18553 ( $rtokens->[0] =~ /^=\d*$/ && ( $old_rtokens->[0] eq $rtokens->[0] ) );
18555 # must be monotonic variation
18556 return unless ( $is_assignment || $previous_maximum_jmax_seen <= $jmax );
18558 # must be more fields in the new line
18559 my $maximum_field_index = $old_line->get_jmax();
18560 return unless ( $maximum_field_index < $jmax );
18562 unless ($is_assignment) {
18564 unless ( $old_line->get_jmax_original_line() == $minimum_jmax_seen )
18565 ; # only if monotonic
18567 # never combine fields of a comma list
18569 unless ( $maximum_field_index > 1 )
18570 && ( $new_line->get_list_type() !~ /^,/ );
18573 my $rfields = $new_line->get_rfields();
18574 my $rpatterns = $new_line->get_rpatterns();
18575 my $old_rpatterns = $old_line->get_rpatterns();
18577 # loop over all OLD tokens except comment and check match
18580 for ( $k = 0 ; $k < $maximum_field_index - 1 ; $k++ ) {
18581 if ( ( $$old_rtokens[$k] ne $$rtokens[$k] )
18582 || ( $$old_rpatterns[$k] ne $$rpatterns[$k] ) )
18589 # first tokens agree, so combine extra new tokens
18591 for $k ( $maximum_field_index .. $jmax - 1 ) {
18593 $$rfields[ $maximum_field_index - 1 ] .= $$rfields[$k];
18594 $$rfields[$k] = "";
18595 $$rpatterns[ $maximum_field_index - 1 ] .= $$rpatterns[$k];
18596 $$rpatterns[$k] = "";
18599 $$rtokens[ $maximum_field_index - 1 ] = '#';
18600 $$rfields[$maximum_field_index] = $$rfields[$jmax];
18601 $$rpatterns[$maximum_field_index] = $$rpatterns[$jmax];
18602 $jmax = $maximum_field_index;
18604 $new_line->set_jmax($jmax);
18607 sub fix_terminal_ternary {
18609 # Add empty fields as necessary to align a ternary term
18614 # : $year % 100 ? 1
18615 # : $year % 400 ? 0
18618 # returns 1 if the terminal item should be indented
18620 my ( $rfields, $rtokens, $rpatterns ) = @_;
18622 my $jmax = @{$rfields} - 1;
18623 my $old_line = $group_lines[$maximum_line_index];
18624 my $rfields_old = $old_line->get_rfields();
18626 my $rpatterns_old = $old_line->get_rpatterns();
18627 my $rtokens_old = $old_line->get_rtokens();
18628 my $maximum_field_index = $old_line->get_jmax();
18630 # look for the question mark after the :
18632 my $depth_question;
18634 for ( my $j = 0 ; $j < $maximum_field_index ; $j++ ) {
18635 my $tok = $rtokens_old->[$j];
18636 if ( $tok =~ /^\?(\d+)$/ ) {
18637 $depth_question = $1;
18639 # depth must be correct
18640 next unless ( $depth_question eq $group_level );
18643 if ( $rfields_old->[ $j + 1 ] =~ /^(\?\s*)/ ) {
18644 $pad = " " x length($1);
18647 return; # shouldn't happen
18652 return unless ( defined($jquestion) ); # shouldn't happen
18654 # Now splice the tokens and patterns of the previous line
18655 # into the else line to insure a match. Add empty fields
18657 my $jadd = $jquestion;
18659 # Work on copies of the actual arrays in case we have
18660 # to return due to an error
18661 my @fields = @{$rfields};
18662 my @patterns = @{$rpatterns};
18663 my @tokens = @{$rtokens};
18665 VALIGN_DEBUG_FLAG_TERNARY && do {
18667 print "CURRENT FIELDS=<@{$rfields_old}>\n";
18668 print "CURRENT TOKENS=<@{$rtokens_old}>\n";
18669 print "CURRENT PATTERNS=<@{$rpatterns_old}>\n";
18670 print "UNMODIFIED FIELDS=<@{$rfields}>\n";
18671 print "UNMODIFIED TOKENS=<@{$rtokens}>\n";
18672 print "UNMODIFIED PATTERNS=<@{$rpatterns}>\n";
18675 # handle cases of leading colon on this line
18676 if ( $fields[0] =~ /^(:\s*)(.*)$/ ) {
18678 my ( $colon, $therest ) = ( $1, $2 );
18680 # Handle sub-case of first field with leading colon plus additional code
18681 # This is the usual situation as at the '1' below:
18683 # : $year % 400 ? 0
18687 # Split the first field after the leading colon and insert padding.
18688 # Note that this padding will remain even if the terminal value goes
18689 # out on a separate line. This does not seem to look to bad, so no
18690 # mechanism has been included to undo it.
18691 my $field1 = shift @fields;
18692 unshift @fields, ( $colon, $pad . $therest );
18694 # change the leading pattern from : to ?
18695 return unless ( $patterns[0] =~ s/^\:/?/ );
18697 # install leading tokens and patterns of existing line
18698 unshift( @tokens, @{$rtokens_old}[ 0 .. $jquestion ] );
18699 unshift( @patterns, @{$rpatterns_old}[ 0 .. $jquestion ] );
18701 # insert appropriate number of empty fields
18702 splice( @fields, 1, 0, ('') x $jadd ) if $jadd;
18705 # handle sub-case of first field just equal to leading colon.
18706 # This can happen for example in the example below where
18707 # the leading '(' would create a new alignment token
18708 # : ( $name =~ /[]}]$/ ) ? ( $mname = $name )
18709 # : ( $mname = $name . '->' );
18712 return unless ( $jmax > 0 && $tokens[0] ne '#' ); # shouldn't happen
18714 # prepend a leading ? onto the second pattern
18715 $patterns[1] = "?b" . $patterns[1];
18717 # pad the second field
18718 $fields[1] = $pad . $fields[1];
18720 # install leading tokens and patterns of existing line, replacing
18721 # leading token and inserting appropriate number of empty fields
18722 splice( @tokens, 0, 1, @{$rtokens_old}[ 0 .. $jquestion ] );
18723 splice( @patterns, 1, 0, @{$rpatterns_old}[ 1 .. $jquestion ] );
18724 splice( @fields, 1, 0, ('') x $jadd ) if $jadd;
18728 # Handle case of no leading colon on this line. This will
18729 # be the case when -wba=':' is used. For example,
18730 # $year % 400 ? 0 :
18734 # install leading tokens and patterns of existing line
18735 $patterns[0] = '?' . 'b' . $patterns[0];
18736 unshift( @tokens, @{$rtokens_old}[ 0 .. $jquestion ] );
18737 unshift( @patterns, @{$rpatterns_old}[ 0 .. $jquestion ] );
18739 # insert appropriate number of empty fields
18740 $jadd = $jquestion + 1;
18741 $fields[0] = $pad . $fields[0];
18742 splice( @fields, 0, 0, ('') x $jadd ) if $jadd;
18745 VALIGN_DEBUG_FLAG_TERNARY && do {
18747 print "MODIFIED TOKENS=<@tokens>\n";
18748 print "MODIFIED PATTERNS=<@patterns>\n";
18749 print "MODIFIED FIELDS=<@fields>\n";
18752 # all ok .. update the arrays
18753 @{$rfields} = @fields;
18754 @{$rtokens} = @tokens;
18755 @{$rpatterns} = @patterns;
18757 # force a flush after this line
18761 sub fix_terminal_else {
18763 # Add empty fields as necessary to align a balanced terminal
18764 # else block to a previous if/elsif/unless block,
18767 # if ( 1 || $x ) { print "ok 13\n"; }
18768 # else { print "not ok 13\n"; }
18770 # returns 1 if the else block should be indented
18772 my ( $rfields, $rtokens, $rpatterns ) = @_;
18773 my $jmax = @{$rfields} - 1;
18774 return unless ( $jmax > 0 );
18776 # check for balanced else block following if/elsif/unless
18777 my $rfields_old = $current_line->get_rfields();
18779 # TBD: add handling for 'case'
18780 return unless ( $rfields_old->[0] =~ /^(if|elsif|unless)\s*$/ );
18782 # look for the opening brace after the else, and extrace the depth
18783 my $tok_brace = $rtokens->[0];
18785 if ( $tok_brace =~ /^\{(\d+)/ ) { $depth_brace = $1; }
18787 # probably: "else # side_comment"
18790 my $rpatterns_old = $current_line->get_rpatterns();
18791 my $rtokens_old = $current_line->get_rtokens();
18792 my $maximum_field_index = $current_line->get_jmax();
18794 # be sure the previous if/elsif is followed by an opening paren
18796 my $tok_paren = '(' . $depth_brace;
18797 my $tok_test = $rtokens_old->[$jparen];
18798 return unless ( $tok_test eq $tok_paren ); # shouldn't happen
18800 # Now find the opening block brace
18802 for ( my $j = 1 ; $j < $maximum_field_index ; $j++ ) {
18803 my $tok = $rtokens_old->[$j];
18804 if ( $tok eq $tok_brace ) {
18809 return unless ( defined($jbrace) ); # shouldn't happen
18811 # Now splice the tokens and patterns of the previous line
18812 # into the else line to insure a match. Add empty fields
18814 my $jadd = $jbrace - $jparen;
18815 splice( @{$rtokens}, 0, 0, @{$rtokens_old}[ $jparen .. $jbrace - 1 ] );
18816 splice( @{$rpatterns}, 1, 0, @{$rpatterns_old}[ $jparen + 1 .. $jbrace ] );
18817 splice( @{$rfields}, 1, 0, ('') x $jadd );
18819 # force a flush after this line if it does not follow a case
18821 unless ( $rfields_old->[0] =~ /^case\s*$/ );
18824 { # sub check_match
18825 my %is_good_alignment;
18829 # Vertically aligning on certain "good" tokens is usually okay
18830 # so we can be less restrictive in marginal cases.
18831 @_ = qw( { ? => = );
18833 @is_good_alignment{@_} = (1) x scalar(@_);
18838 # See if the current line matches the current vertical alignment group.
18839 # If not, flush the current group.
18840 my $new_line = shift;
18841 my $old_line = shift;
18843 # uses global variables:
18844 # $previous_minimum_jmax_seen
18845 # $maximum_jmax_seen
18846 # $maximum_line_index
18848 my $jmax = $new_line->get_jmax();
18849 my $maximum_field_index = $old_line->get_jmax();
18851 # flush if this line has too many fields
18852 if ( $jmax > $maximum_field_index ) { goto NO_MATCH }
18854 # flush if adding this line would make a non-monotonic field count
18856 ( $maximum_field_index > $jmax ) # this has too few fields
18858 ( $previous_minimum_jmax_seen <
18859 $jmax ) # and wouldn't be monotonic
18860 || ( $old_line->get_jmax_original_line() != $maximum_jmax_seen )
18867 # otherwise see if this line matches the current group
18868 my $jmax_original_line = $new_line->get_jmax_original_line();
18869 my $is_hanging_side_comment = $new_line->get_is_hanging_side_comment();
18870 my $rtokens = $new_line->get_rtokens();
18871 my $rfields = $new_line->get_rfields();
18872 my $rpatterns = $new_line->get_rpatterns();
18873 my $list_type = $new_line->get_list_type();
18875 my $group_list_type = $old_line->get_list_type();
18876 my $old_rpatterns = $old_line->get_rpatterns();
18877 my $old_rtokens = $old_line->get_rtokens();
18879 my $jlimit = $jmax - 1;
18880 if ( $maximum_field_index > $jmax ) {
18881 $jlimit = $jmax_original_line;
18882 --$jlimit unless ( length( $new_line->get_rfields()->[$jmax] ) );
18885 # handle comma-separated lists ..
18886 if ( $group_list_type && ( $list_type eq $group_list_type ) ) {
18887 for my $j ( 0 .. $jlimit ) {
18888 my $old_tok = $$old_rtokens[$j];
18889 next unless $old_tok;
18890 my $new_tok = $$rtokens[$j];
18891 next unless $new_tok;
18893 # lists always match ...
18894 # unless they would align any '=>'s with ','s
18896 if ( $old_tok =~ /^=>/ && $new_tok =~ /^,/
18897 || $new_tok =~ /^=>/ && $old_tok =~ /^,/ );
18901 # do detailed check for everything else except hanging side comments
18902 elsif ( !$is_hanging_side_comment ) {
18904 my $leading_space_count = $new_line->get_leading_space_count();
18908 my $saw_good_alignment;
18910 for my $j ( 0 .. $jlimit ) {
18912 my $old_tok = $$old_rtokens[$j];
18913 my $new_tok = $$rtokens[$j];
18915 # Note on encoding used for alignment tokens:
18916 # -------------------------------------------
18917 # Tokens are "decorated" with information which can help
18918 # prevent unwanted alignments. Consider for example the
18919 # following two lines:
18920 # local ( $xn, $xd ) = split( '/', &'rnorm(@_) );
18921 # local ( $i, $f ) = &'bdiv( $xn, $xd );
18922 # There are three alignment tokens in each line, a comma,
18923 # an =, and a comma. In the first line these three tokens
18925 # ,4+local-18 =3 ,4+split-7
18926 # and in the second line they are encoded as
18927 # ,4+local-18 =3 ,4+&'bdiv-8
18928 # Tokens always at least have token name and nesting
18929 # depth. So in this example the ='s are at depth 3 and
18930 # the ,'s are at depth 4. This prevents aligning tokens
18931 # of different depths. Commas contain additional
18932 # information, as follows:
18933 # , {depth} + {container name} - {spaces to opening paren}
18934 # This allows us to reject matching the rightmost commas
18935 # in the above two lines, since they are for different
18936 # function calls. This encoding is done in
18937 # 'sub send_lines_to_vertical_aligner'.
18939 # Pick off actual token.
18940 # Everything up to the first digit is the actual token.
18941 my $alignment_token = $new_tok;
18942 if ( $alignment_token =~ /^([^\d]+)/ ) { $alignment_token = $1 }
18944 # see if the decorated tokens match
18945 my $tokens_match = $new_tok eq $old_tok
18947 # Exception for matching terminal : of ternary statement..
18948 # consider containers prefixed by ? and : a match
18949 || ( $new_tok =~ /^,\d*\+\:/ && $old_tok =~ /^,\d*\+\?/ );
18951 # No match if the alignment tokens differ...
18952 if ( !$tokens_match ) {
18954 # ...Unless this is a side comment
18958 # and there is either at least one alignment token
18959 # or this is a single item following a list. This
18960 # latter rule is required for 'December' to join
18961 # the following list:
18963 # '', 'January', 'February', 'March',
18964 # 'April', 'May', 'June', 'July',
18965 # 'August', 'September', 'October', 'November',
18968 # If it doesn't then the -lp formatting will fail.
18969 && ( $j > 0 || $old_tok =~ /^,/ )
18972 $marginal_match = 1
18973 if ( $marginal_match == 0
18974 && $maximum_line_index == 0 );
18981 # Calculate amount of padding required to fit this in.
18982 # $pad is the number of spaces by which we must increase
18983 # the current field to squeeze in this field.
18985 length( $$rfields[$j] ) - $old_line->current_field_width($j);
18986 if ( $j == 0 ) { $pad += $leading_space_count; }
18988 # remember max pads to limit marginal cases
18989 if ( $alignment_token ne '#' ) {
18990 if ( $pad > $max_pad ) { $max_pad = $pad }
18991 if ( $pad < $min_pad ) { $min_pad = $pad }
18993 if ( $is_good_alignment{$alignment_token} ) {
18994 $saw_good_alignment = 1;
18997 # If patterns don't match, we have to be careful...
18998 if ( $$old_rpatterns[$j] ne $$rpatterns[$j] ) {
19000 # flag this as a marginal match since patterns differ
19001 $marginal_match = 1
19002 if ( $marginal_match == 0 && $maximum_line_index == 0 );
19004 # We have to be very careful about aligning commas
19005 # when the pattern's don't match, because it can be
19006 # worse to create an alignment where none is needed
19007 # than to omit one. Here's an example where the ','s
19008 # are not in named continers. The first line below
19009 # should not match the next two:
19010 # ( $a, $b ) = ( $b, $r );
19011 # ( $x1, $x2 ) = ( $x2 - $q * $x1, $x1 );
19012 # ( $y1, $y2 ) = ( $y2 - $q * $y1, $y1 );
19013 if ( $alignment_token eq ',' ) {
19015 # do not align commas unless they are in named containers
19016 goto NO_MATCH unless ( $new_tok =~ /[A-Za-z]/ );
19019 # do not align parens unless patterns match;
19020 # large ugly spaces can occur in math expressions.
19021 elsif ( $alignment_token eq '(' ) {
19023 # But we can allow a match if the parens don't
19024 # require any padding.
19025 if ( $pad != 0 ) { goto NO_MATCH }
19028 # Handle an '=' alignment with different patterns to
19030 elsif ( $alignment_token eq '=' ) {
19032 # It is best to be a little restrictive when
19033 # aligning '=' tokens. Here is an example of
19034 # two lines that we will not align:
19037 # The problem is that one is a 'my' declaration,
19038 # and the other isn't, so they're not very similar.
19039 # We will filter these out by comparing the first
19040 # letter of the pattern. This is crude, but works
19043 substr( $$old_rpatterns[$j], 0, 1 ) ne
19044 substr( $$rpatterns[$j], 0, 1 ) )
19049 # If we pass that test, we'll call it a marginal match.
19050 # Here is an example of a marginal match:
19052 # $op = compile_bblock($op);
19053 # The left tokens are both identifiers, but
19054 # one accesses a hash and the other doesn't.
19055 # We'll let this be a tentative match and undo
19056 # it later if we don't find more than 2 lines
19058 elsif ( $maximum_line_index == 0 ) {
19060 2; # =2 prevents being undone below
19065 # Don't let line with fewer fields increase column widths
19067 if ( $maximum_field_index > $jmax ) {
19069 # Exception: suspend this rule to allow last lines to join
19070 if ( $pad > 0 ) { goto NO_MATCH; }
19072 } ## end for my $j ( 0 .. $jlimit)
19074 # Turn off the "marginal match" flag in some cases...
19075 # A "marginal match" occurs when the alignment tokens agree
19076 # but there are differences in the other tokens (patterns).
19077 # If we leave the marginal match flag set, then the rule is that we
19078 # will align only if there are more than two lines in the group.
19079 # We will turn of the flag if we almost have a match
19080 # and either we have seen a good alignment token or we
19081 # just need a small pad (2 spaces) to fit. These rules are
19082 # the result of experimentation. Tokens which misaligned by just
19083 # one or two characters are annoying. On the other hand,
19084 # large gaps to less important alignment tokens are also annoying.
19085 if ( $marginal_match == 1
19086 && $jmax == $maximum_field_index
19087 && ( $saw_good_alignment || ( $max_pad < 3 && $min_pad > -3 ) )
19090 $marginal_match = 0;
19092 ##print "marginal=$marginal_match saw=$saw_good_alignment jmax=$jmax max=$maximum_field_index maxpad=$max_pad minpad=$min_pad\n";
19095 # We have a match (even if marginal).
19096 # If the current line has fewer fields than the current group
19097 # but otherwise matches, copy the remaining group fields to
19098 # make it a perfect match.
19099 if ( $maximum_field_index > $jmax ) {
19100 my $comment = $$rfields[$jmax];
19101 for $jmax ( $jlimit .. $maximum_field_index ) {
19102 $$rtokens[$jmax] = $$old_rtokens[$jmax];
19103 $$rfields[ ++$jmax ] = '';
19104 $$rpatterns[$jmax] = $$old_rpatterns[$jmax];
19106 $$rfields[$jmax] = $comment;
19107 $new_line->set_jmax($jmax);
19112 ##print "BUBBA: no match jmax=$jmax max=$maximum_field_index $group_list_type lines=$maximum_line_index token=$$old_rtokens[0]\n";
19120 return unless ( $maximum_line_index >= 0 );
19121 my $new_line = shift;
19122 my $old_line = shift;
19124 my $jmax = $new_line->get_jmax();
19125 my $leading_space_count = $new_line->get_leading_space_count();
19126 my $is_hanging_side_comment = $new_line->get_is_hanging_side_comment();
19127 my $rtokens = $new_line->get_rtokens();
19128 my $rfields = $new_line->get_rfields();
19129 my $rpatterns = $new_line->get_rpatterns();
19131 my $group_list_type = $group_lines[0]->get_list_type();
19133 my $padding_so_far = 0;
19134 my $padding_available = $old_line->get_available_space_on_right();
19136 # save current columns in case this doesn't work
19137 save_alignment_columns();
19139 my ( $j, $pad, $eight );
19140 my $maximum_field_index = $old_line->get_jmax();
19141 for $j ( 0 .. $jmax ) {
19143 $pad = length( $$rfields[$j] ) - $old_line->current_field_width($j);
19146 $pad += $leading_space_count;
19149 # remember largest gap of the group, excluding gap to side comment
19151 && $group_maximum_gap < -$pad
19153 && $j < $jmax - 1 )
19155 $group_maximum_gap = -$pad;
19160 ## This patch helps sometimes, but it doesn't check to see if
19161 ## the line is too long even without the side comment. It needs
19163 ##don't let a long token with no trailing side comment push
19164 ##side comments out, or end a group. (sidecmt1.t)
19165 ##next if ($j==$jmax-1 && length($$rfields[$jmax])==0);
19167 # This line will need space; lets see if we want to accept it..
19170 # not if this won't fit
19171 ( $pad > $padding_available )
19173 # previously, there were upper bounds placed on padding here
19174 # (maximum_whitespace_columns), but they were not really helpful
19179 # revert to starting state then flush; things didn't work out
19180 restore_alignment_columns();
19185 # patch to avoid excessive gaps in previous lines,
19186 # due to a line of fewer fields.
19187 # return join( ".",
19188 # $self->{"dfi"}, $self->{"aa"}, $self->rsvd, $self->{"rd"},
19189 # $self->{"area"}, $self->{"id"}, $self->{"sel"} );
19190 next if ( $jmax < $maximum_field_index && $j == $jmax - 1 );
19192 # looks ok, squeeze this field in
19193 $old_line->increase_field_width( $j, $pad );
19194 $padding_available -= $pad;
19196 # remember largest gap of the group, excluding gap to side comment
19197 if ( $pad > $group_maximum_gap && $j > 0 && $j < $jmax - 1 ) {
19198 $group_maximum_gap = $pad;
19205 # The current line either starts a new alignment group or is
19206 # accepted into the current alignment group.
19207 my $new_line = shift;
19208 $group_lines[ ++$maximum_line_index ] = $new_line;
19210 # initialize field lengths if starting new group
19211 if ( $maximum_line_index == 0 ) {
19213 my $jmax = $new_line->get_jmax();
19214 my $rfields = $new_line->get_rfields();
19215 my $rtokens = $new_line->get_rtokens();
19217 my $col = $new_line->get_leading_space_count();
19219 for $j ( 0 .. $jmax ) {
19220 $col += length( $$rfields[$j] );
19222 # create initial alignments for the new group
19224 if ( $j < $jmax ) { $token = $$rtokens[$j] }
19225 my $alignment = make_alignment( $col, $token );
19226 $new_line->set_alignment( $j, $alignment );
19229 $maximum_jmax_seen = $jmax;
19230 $minimum_jmax_seen = $jmax;
19233 # use previous alignments otherwise
19235 my @new_alignments =
19236 $group_lines[ $maximum_line_index - 1 ]->get_alignments();
19237 $new_line->set_alignments(@new_alignments);
19240 # remember group jmax extremes for next call to append_line
19241 $previous_minimum_jmax_seen = $minimum_jmax_seen;
19242 $previous_maximum_jmax_seen = $maximum_jmax_seen;
19247 # debug routine to dump array contents
19252 # flush() sends the current Perl::Tidy::VerticalAligner group down the
19253 # pipeline to Perl::Tidy::FileWriter.
19255 # This is the external flush, which also empties the cache
19258 if ( $maximum_line_index < 0 ) {
19259 if ($cached_line_type) {
19260 $seqno_string = $cached_seqno_string;
19261 entab_and_output( $cached_line_text,
19262 $cached_line_leading_space_count,
19263 $last_group_level_written );
19264 $cached_line_type = 0;
19265 $cached_line_text = "";
19266 $cached_seqno_string = "";
19274 # This is the internal flush, which leaves the cache intact
19277 return if ( $maximum_line_index < 0 );
19279 # handle a group of comment lines
19280 if ( $group_type eq 'COMMENT' ) {
19282 VALIGN_DEBUG_FLAG_APPEND0 && do {
19283 my ( $a, $b, $c ) = caller();
19285 "APPEND0: Flush called from $a $b $c for COMMENT group: lines=$maximum_line_index \n";
19288 my $leading_space_count = $comment_leading_space_count;
19289 my $leading_string = get_leading_string($leading_space_count);
19291 # zero leading space count if any lines are too long
19292 my $max_excess = 0;
19293 for my $i ( 0 .. $maximum_line_index ) {
19294 my $str = $group_lines[$i];
19296 length($str) + $leading_space_count - $rOpts_maximum_line_length;
19297 if ( $excess > $max_excess ) {
19298 $max_excess = $excess;
19302 if ( $max_excess > 0 ) {
19303 $leading_space_count -= $max_excess;
19304 if ( $leading_space_count < 0 ) { $leading_space_count = 0 }
19305 $last_outdented_line_at =
19306 $file_writer_object->get_output_line_number();
19307 unless ($outdented_line_count) {
19308 $first_outdented_line_at = $last_outdented_line_at;
19310 $outdented_line_count += ( $maximum_line_index + 1 );
19313 # write the group of lines
19314 my $outdent_long_lines = 0;
19315 for my $i ( 0 .. $maximum_line_index ) {
19316 write_leader_and_string( $leading_space_count, $group_lines[$i], 0,
19317 $outdent_long_lines, "" );
19321 # handle a group of code lines
19324 VALIGN_DEBUG_FLAG_APPEND0 && do {
19325 my $group_list_type = $group_lines[0]->get_list_type();
19326 my ( $a, $b, $c ) = caller();
19327 my $maximum_field_index = $group_lines[0]->get_jmax();
19329 "APPEND0: Flush called from $a $b $c fields=$maximum_field_index list=$group_list_type lines=$maximum_line_index extra=$extra_indent_ok\n";
19333 # some small groups are best left unaligned
19334 my $do_not_align = decide_if_aligned();
19336 # optimize side comment location
19337 $do_not_align = adjust_side_comment($do_not_align);
19339 # recover spaces for -lp option if possible
19340 my $extra_leading_spaces = get_extra_leading_spaces();
19342 # all lines of this group have the same basic leading spacing
19343 my $group_leader_length = $group_lines[0]->get_leading_space_count();
19345 # add extra leading spaces if helpful
19346 my $min_ci_gap = improve_continuation_indentation( $do_not_align,
19347 $group_leader_length );
19349 # loop to output all lines
19350 for my $i ( 0 .. $maximum_line_index ) {
19351 my $line = $group_lines[$i];
19352 write_vertically_aligned_line( $line, $min_ci_gap, $do_not_align,
19353 $group_leader_length, $extra_leading_spaces );
19356 initialize_for_new_group();
19359 sub decide_if_aligned {
19361 # Do not try to align two lines which are not really similar
19362 return unless $maximum_line_index == 1;
19363 return if ($is_matching_terminal_line);
19365 my $group_list_type = $group_lines[0]->get_list_type();
19367 my $do_not_align = (
19369 # always align lists
19374 # don't align if it was just a marginal match
19377 # don't align two lines with big gap
19378 || $group_maximum_gap > 12
19380 # or lines with differing number of alignment tokens
19381 # TODO: this could be improved. It occasionally rejects
19383 || $previous_maximum_jmax_seen != $previous_minimum_jmax_seen
19387 # But try to convert them into a simple comment group if the first line
19388 # a has side comment
19389 my $rfields = $group_lines[0]->get_rfields();
19390 my $maximum_field_index = $group_lines[0]->get_jmax();
19392 && ( $maximum_line_index > 0 )
19393 && ( length( $$rfields[$maximum_field_index] ) > 0 ) )
19398 return $do_not_align;
19401 sub adjust_side_comment {
19403 my $do_not_align = shift;
19405 # let's see if we can move the side comment field out a little
19406 # to improve readability (the last field is always a side comment field)
19407 my $have_side_comment = 0;
19408 my $first_side_comment_line = -1;
19409 my $maximum_field_index = $group_lines[0]->get_jmax();
19410 for my $i ( 0 .. $maximum_line_index ) {
19411 my $line = $group_lines[$i];
19413 if ( length( $line->get_rfields()->[$maximum_field_index] ) ) {
19414 $have_side_comment = 1;
19415 $first_side_comment_line = $i;
19420 my $kmax = $maximum_field_index + 1;
19422 if ($have_side_comment) {
19424 my $line = $group_lines[0];
19426 # the maximum space without exceeding the line length:
19427 my $avail = $line->get_available_space_on_right();
19429 # try to use the previous comment column
19430 my $side_comment_column = $line->get_column( $kmax - 2 );
19431 my $move = $last_comment_column - $side_comment_column;
19433 ## my $sc_line0 = $side_comment_history[0]->[0];
19434 ## my $sc_col0 = $side_comment_history[0]->[1];
19435 ## my $sc_line1 = $side_comment_history[1]->[0];
19436 ## my $sc_col1 = $side_comment_history[1]->[1];
19437 ## my $sc_line2 = $side_comment_history[2]->[0];
19438 ## my $sc_col2 = $side_comment_history[2]->[1];
19440 ## # FUTURE UPDATES:
19441 ## # Be sure to ignore 'do not align' and '} # end comments'
19442 ## # Find first $move > 0 and $move <= $avail as follows:
19443 ## # 1. try sc_col1 if sc_col1 == sc_col0 && (line-sc_line0) < 12
19444 ## # 2. try sc_col2 if (line-sc_line2) < 12
19445 ## # 3. try min possible space, plus up to 8,
19446 ## # 4. try min possible space
19448 if ( $kmax > 0 && !$do_not_align ) {
19450 # but if this doesn't work, give up and use the minimum space
19451 if ( $move > $avail ) {
19452 $move = $rOpts_minimum_space_to_comment - 1;
19455 # but we want some minimum space to the comment
19456 my $min_move = $rOpts_minimum_space_to_comment - 1;
19458 && $last_side_comment_length > 0
19459 && ( $first_side_comment_line == 0 )
19460 && $group_level == $last_group_level_written )
19465 if ( $move < $min_move ) {
19469 # prevously, an upper bound was placed on $move here,
19470 # (maximum_space_to_comment), but it was not helpful
19472 # don't exceed the available space
19473 if ( $move > $avail ) { $move = $avail }
19475 # we can only increase space, never decrease
19477 $line->increase_field_width( $maximum_field_index - 1, $move );
19480 # remember this column for the next group
19481 $last_comment_column = $line->get_column( $kmax - 2 );
19485 # try to at least line up the existing side comment location
19486 if ( $kmax > 0 && $move > 0 && $move < $avail ) {
19487 $line->increase_field_width( $maximum_field_index - 1, $move );
19491 # reset side comment column if we can't align
19493 forget_side_comment();
19497 return $do_not_align;
19500 sub improve_continuation_indentation {
19501 my ( $do_not_align, $group_leader_length ) = @_;
19503 # See if we can increase the continuation indentation
19504 # to move all continuation lines closer to the next field
19505 # (unless it is a comment).
19507 # '$min_ci_gap'is the extra indentation that we may need to introduce.
19508 # We will only introduce this to fields which already have some ci.
19509 # Without this variable, we would occasionally get something like this
19512 # use overload '+' => \&plus,
19514 # '*' => \&multiply,
19517 # 'atan2' => \&atan2,
19519 # Whereas with this variable, we can shift variables over to get this:
19521 # use overload '+' => \&plus,
19523 # '*' => \&multiply,
19526 # 'atan2' => \&atan2,
19528 ## BUB: Deactivated####################
19529 # The trouble with this patch is that it may, for example,
19530 # move in some 'or's or ':'s, and leave some out, so that the
19531 # left edge alignment suffers.
19533 ###########################################
19535 my $maximum_field_index = $group_lines[0]->get_jmax();
19537 my $min_ci_gap = $rOpts_maximum_line_length;
19538 if ( $maximum_field_index > 1 && !$do_not_align ) {
19540 for my $i ( 0 .. $maximum_line_index ) {
19541 my $line = $group_lines[$i];
19542 my $leading_space_count = $line->get_leading_space_count();
19543 my $rfields = $line->get_rfields();
19546 $line->get_column(0) -
19547 $leading_space_count -
19548 length( $$rfields[0] );
19550 if ( $leading_space_count > $group_leader_length ) {
19551 if ( $gap < $min_ci_gap ) { $min_ci_gap = $gap }
19555 if ( $min_ci_gap >= $rOpts_maximum_line_length ) {
19562 return $min_ci_gap;
19565 sub write_vertically_aligned_line {
19567 my ( $line, $min_ci_gap, $do_not_align, $group_leader_length,
19568 $extra_leading_spaces )
19570 my $rfields = $line->get_rfields();
19571 my $leading_space_count = $line->get_leading_space_count();
19572 my $outdent_long_lines = $line->get_outdent_long_lines();
19573 my $maximum_field_index = $line->get_jmax();
19574 my $rvertical_tightness_flags = $line->get_rvertical_tightness_flags();
19576 # add any extra spaces
19577 if ( $leading_space_count > $group_leader_length ) {
19578 $leading_space_count += $min_ci_gap;
19581 my $str = $$rfields[0];
19583 # loop to concatenate all fields of this line and needed padding
19584 my $total_pad_count = 0;
19586 for $j ( 1 .. $maximum_field_index ) {
19588 # skip zero-length side comments
19590 if ( ( $j == $maximum_field_index )
19591 && ( !defined( $$rfields[$j] ) || ( length( $$rfields[$j] ) == 0 ) )
19594 # compute spaces of padding before this field
19595 my $col = $line->get_column( $j - 1 );
19596 $pad = $col - ( length($str) + $leading_space_count );
19598 if ($do_not_align) {
19600 ( $j < $maximum_field_index )
19602 : $rOpts_minimum_space_to_comment - 1;
19605 # if the -fpsc flag is set, move the side comment to the selected
19606 # column if and only if it is possible, ignoring constraints on
19607 # line length and minimum space to comment
19608 if ( $rOpts_fixed_position_side_comment && $j == $maximum_field_index )
19610 my $newpad = $pad + $rOpts_fixed_position_side_comment - $col - 1;
19611 if ( $newpad >= 0 ) { $pad = $newpad; }
19614 # accumulate the padding
19615 if ( $pad > 0 ) { $total_pad_count += $pad; }
19618 if ( !defined $$rfields[$j] ) {
19619 write_diagnostics("UNDEFined field at j=$j\n");
19622 # only add padding when we have a finite field;
19623 # this avoids extra terminal spaces if we have empty fields
19624 if ( length( $$rfields[$j] ) > 0 ) {
19625 $str .= ' ' x $total_pad_count;
19626 $total_pad_count = 0;
19627 $str .= $$rfields[$j];
19630 $total_pad_count = 0;
19633 # update side comment history buffer
19634 if ( $j == $maximum_field_index ) {
19635 my $lineno = $file_writer_object->get_output_line_number();
19636 shift @side_comment_history;
19637 push @side_comment_history, [ $lineno, $col ];
19641 my $side_comment_length = ( length( $$rfields[$maximum_field_index] ) );
19643 # ship this line off
19644 write_leader_and_string( $leading_space_count + $extra_leading_spaces,
19645 $str, $side_comment_length, $outdent_long_lines,
19646 $rvertical_tightness_flags );
19649 sub get_extra_leading_spaces {
19651 #----------------------------------------------------------
19652 # Define any extra indentation space (for the -lp option).
19654 # If a list has side comments, sub scan_list must dump the
19655 # list before it sees everything. When this happens, it sets
19656 # the indentation to the standard scheme, but notes how
19657 # many spaces it would have liked to use. We may be able
19658 # to recover that space here in the event that that all of the
19659 # lines of a list are back together again.
19660 #----------------------------------------------------------
19662 my $extra_leading_spaces = 0;
19663 if ($extra_indent_ok) {
19664 my $object = $group_lines[0]->get_indentation();
19665 if ( ref($object) ) {
19666 my $extra_indentation_spaces_wanted =
19667 get_RECOVERABLE_SPACES($object);
19669 # all indentation objects must be the same
19671 for $i ( 1 .. $maximum_line_index ) {
19672 if ( $object != $group_lines[$i]->get_indentation() ) {
19673 $extra_indentation_spaces_wanted = 0;
19678 if ($extra_indentation_spaces_wanted) {
19680 # the maximum space without exceeding the line length:
19681 my $avail = $group_lines[0]->get_available_space_on_right();
19682 $extra_leading_spaces =
19683 ( $avail > $extra_indentation_spaces_wanted )
19684 ? $extra_indentation_spaces_wanted
19687 # update the indentation object because with -icp the terminal
19688 # ');' will use the same adjustment.
19689 $object->permanently_decrease_AVAILABLE_SPACES(
19690 -$extra_leading_spaces );
19694 return $extra_leading_spaces;
19697 sub combine_fields {
19699 # combine all fields except for the comment field ( sidecmt.t )
19700 # Uses global variables:
19702 # $maximum_line_index
19704 my $maximum_field_index = $group_lines[0]->get_jmax();
19705 for ( $j = 0 ; $j <= $maximum_line_index ; $j++ ) {
19706 my $line = $group_lines[$j];
19707 my $rfields = $line->get_rfields();
19708 foreach ( 1 .. $maximum_field_index - 1 ) {
19709 $$rfields[0] .= $$rfields[$_];
19711 $$rfields[1] = $$rfields[$maximum_field_index];
19713 $line->set_jmax(1);
19714 $line->set_column( 0, 0 );
19715 $line->set_column( 1, 0 );
19718 $maximum_field_index = 1;
19720 for $j ( 0 .. $maximum_line_index ) {
19721 my $line = $group_lines[$j];
19722 my $rfields = $line->get_rfields();
19723 for $k ( 0 .. $maximum_field_index ) {
19724 my $pad = length( $$rfields[$k] ) - $line->current_field_width($k);
19726 $pad += $group_lines[$j]->get_leading_space_count();
19729 if ( $pad > 0 ) { $line->increase_field_width( $k, $pad ) }
19735 sub get_output_line_number {
19737 # the output line number reported to a caller is the number of items
19738 # written plus the number of items in the buffer
19740 1 + $maximum_line_index + $file_writer_object->get_output_line_number();
19743 sub write_leader_and_string {
19745 my ( $leading_space_count, $str, $side_comment_length, $outdent_long_lines,
19746 $rvertical_tightness_flags )
19749 # handle outdenting of long lines:
19750 if ($outdent_long_lines) {
19753 $side_comment_length +
19754 $leading_space_count -
19755 $rOpts_maximum_line_length;
19756 if ( $excess > 0 ) {
19757 $leading_space_count = 0;
19758 $last_outdented_line_at =
19759 $file_writer_object->get_output_line_number();
19761 unless ($outdented_line_count) {
19762 $first_outdented_line_at = $last_outdented_line_at;
19764 $outdented_line_count++;
19768 # Make preliminary leading whitespace. It could get changed
19769 # later by entabbing, so we have to keep track of any changes
19770 # to the leading_space_count from here on.
19771 my $leading_string =
19772 $leading_space_count > 0 ? ( ' ' x $leading_space_count ) : "";
19774 # Unpack any recombination data; it was packed by
19775 # sub send_lines_to_vertical_aligner. Contents:
19777 # [0] type: 1=opening 2=closing 3=opening block brace
19778 # [1] flag: if opening: 1=no multiple steps, 2=multiple steps ok
19779 # if closing: spaces of padding to use
19780 # [2] sequence number of container
19781 # [3] valid flag: do not append if this flag is false
19783 my ( $open_or_close, $tightness_flag, $seqno, $valid, $seqno_beg,
19785 if ($rvertical_tightness_flags) {
19787 $open_or_close, $tightness_flag, $seqno, $valid, $seqno_beg,
19789 ) = @{$rvertical_tightness_flags};
19792 $seqno_string = $seqno_end;
19794 # handle any cached line ..
19795 # either append this line to it or write it out
19796 if ( length($cached_line_text) ) {
19798 if ( !$cached_line_valid ) {
19799 entab_and_output( $cached_line_text,
19800 $cached_line_leading_space_count,
19801 $last_group_level_written );
19804 # handle cached line with opening container token
19805 elsif ( $cached_line_type == 1 || $cached_line_type == 3 ) {
19807 my $gap = $leading_space_count - length($cached_line_text);
19809 # handle option of just one tight opening per line:
19810 if ( $cached_line_flag == 1 ) {
19811 if ( defined($open_or_close) && $open_or_close == 1 ) {
19817 $leading_string = $cached_line_text . ' ' x $gap;
19818 $leading_space_count = $cached_line_leading_space_count;
19819 $seqno_string = $cached_seqno_string . ':' . $seqno_beg;
19822 entab_and_output( $cached_line_text,
19823 $cached_line_leading_space_count,
19824 $last_group_level_written );
19828 # handle cached line to place before this closing container token
19830 my $test_line = $cached_line_text . ' ' x $cached_line_flag . $str;
19832 if ( length($test_line) <= $rOpts_maximum_line_length ) {
19834 $seqno_string = $cached_seqno_string . ':' . $seqno_beg;
19836 # Patch to outdent closing tokens ending # in ');'
19837 # If we are joining a line like ');' to a previous stacked
19838 # set of closing tokens, then decide if we may outdent the
19839 # combined stack to the indentation of the ');'. Since we
19840 # should not normally outdent any of the other tokens more than
19841 # the indentation of the lines that contained them, we will
19842 # only do this if all of the corresponding opening
19843 # tokens were on the same line. This can happen with
19844 # -sot and -sct. For example, it is ok here:
19845 # __PACKAGE__->load_components( qw(
19850 # But, for example, we do not outdent in this example because
19851 # that would put the closing sub brace out farther than the
19852 # opening sub brace:
19854 # perltidy -sot -sct
19856 # '<Control-f>' => sub {
19858 # my $e = $c->XEvent;
19859 # itemsUnderArea $c;
19862 if ( $str =~ /^\);/ && $cached_line_text =~ /^[\)\}\]\s]*$/ ) {
19864 # The way to tell this is if the stacked sequence numbers
19865 # of this output line are the reverse of the stacked
19866 # sequence numbers of the previous non-blank line of
19867 # sequence numbers. So we can join if the previous
19868 # nonblank string of tokens is the mirror image. For
19869 # example if stack )}] is 13:8:6 then we are looking for a
19870 # leading stack like [{( which is 6:8:13 We only need to
19871 # check the two ends, because the intermediate tokens must
19872 # fall in order. Note on speed: having to split on colons
19873 # and eliminate multiple colons might appear to be slow,
19874 # but it's not an issue because we almost never come
19875 # through here. In a typical file we don't.
19876 $seqno_string =~ s/^:+//;
19877 $last_nonblank_seqno_string =~ s/^:+//;
19878 $seqno_string =~ s/:+/:/g;
19879 $last_nonblank_seqno_string =~ s/:+/:/g;
19881 # how many spaces can we outdent?
19883 $cached_line_leading_space_count - $leading_space_count;
19885 && length($seqno_string)
19886 && length($last_nonblank_seqno_string) ==
19887 length($seqno_string) )
19890 ( split ':', $last_nonblank_seqno_string );
19891 my @seqno_now = ( split ':', $seqno_string );
19892 if ( $seqno_now[-1] == $seqno_last[0]
19893 && $seqno_now[0] == $seqno_last[-1] )
19897 # for absolute safety, be sure we only remove
19899 my $ws = substr( $test_line, 0, $diff );
19900 if ( ( length($ws) == $diff ) && $ws =~ /^\s+$/ ) {
19902 $test_line = substr( $test_line, $diff );
19903 $cached_line_leading_space_count -= $diff;
19906 # shouldn't happen, but not critical:
19908 ## ERROR transferring indentation here
19915 $leading_string = "";
19916 $leading_space_count = $cached_line_leading_space_count;
19919 entab_and_output( $cached_line_text,
19920 $cached_line_leading_space_count,
19921 $last_group_level_written );
19925 $cached_line_type = 0;
19926 $cached_line_text = "";
19928 # make the line to be written
19929 my $line = $leading_string . $str;
19931 # write or cache this line
19932 if ( !$open_or_close || $side_comment_length > 0 ) {
19933 entab_and_output( $line, $leading_space_count, $group_level );
19936 $cached_line_text = $line;
19937 $cached_line_type = $open_or_close;
19938 $cached_line_flag = $tightness_flag;
19939 $cached_seqno = $seqno;
19940 $cached_line_valid = $valid;
19941 $cached_line_leading_space_count = $leading_space_count;
19942 $cached_seqno_string = $seqno_string;
19945 $last_group_level_written = $group_level;
19946 $last_side_comment_length = $side_comment_length;
19947 $extra_indent_ok = 0;
19950 sub entab_and_output {
19951 my ( $line, $leading_space_count, $level ) = @_;
19953 # The line is currently correct if there is no tabbing (recommended!)
19954 # We may have to lop off some leading spaces and replace with tabs.
19955 if ( $leading_space_count > 0 ) {
19957 # Nothing to do if no tabs
19958 if ( !( $rOpts_tabs || $rOpts_entab_leading_whitespace )
19959 || $rOpts_indent_columns <= 0 )
19965 # Handle entab option
19966 elsif ($rOpts_entab_leading_whitespace) {
19968 $leading_space_count % $rOpts_entab_leading_whitespace;
19970 int( $leading_space_count / $rOpts_entab_leading_whitespace );
19971 my $leading_string = "\t" x $tab_count . ' ' x $space_count;
19972 if ( $line =~ /^\s{$leading_space_count,$leading_space_count}/ ) {
19973 substr( $line, 0, $leading_space_count ) = $leading_string;
19977 # REMOVE AFTER TESTING
19978 # shouldn't happen - program error counting whitespace
19979 # we'll skip entabbing
19981 "Error entabbing in entab_and_output: expected count=$leading_space_count\n"
19986 # Handle option of one tab per level
19988 my $leading_string = ( "\t" x $level );
19990 $leading_space_count - $level * $rOpts_indent_columns;
19992 # shouldn't happen:
19993 if ( $space_count < 0 ) {
19995 "Error entabbing in append_line: for level=$group_level count=$leading_space_count\n"
19997 $leading_string = ( ' ' x $leading_space_count );
20000 $leading_string .= ( ' ' x $space_count );
20002 if ( $line =~ /^\s{$leading_space_count,$leading_space_count}/ ) {
20003 substr( $line, 0, $leading_space_count ) = $leading_string;
20007 # REMOVE AFTER TESTING
20008 # shouldn't happen - program error counting whitespace
20009 # we'll skip entabbing
20011 "Error entabbing in entab_and_output: expected count=$leading_space_count\n"
20016 $file_writer_object->write_code_line( $line . "\n" );
20017 if ($seqno_string) {
20018 $last_nonblank_seqno_string = $seqno_string;
20022 { # begin get_leading_string
20024 my @leading_string_cache;
20026 sub get_leading_string {
20028 # define the leading whitespace string for this line..
20029 my $leading_whitespace_count = shift;
20031 # Handle case of zero whitespace, which includes multi-line quotes
20032 # (which may have a finite level; this prevents tab problems)
20033 if ( $leading_whitespace_count <= 0 ) {
20037 # look for previous result
20038 elsif ( $leading_string_cache[$leading_whitespace_count] ) {
20039 return $leading_string_cache[$leading_whitespace_count];
20042 # must compute a string for this number of spaces
20043 my $leading_string;
20045 # Handle simple case of no tabs
20046 if ( !( $rOpts_tabs || $rOpts_entab_leading_whitespace )
20047 || $rOpts_indent_columns <= 0 )
20049 $leading_string = ( ' ' x $leading_whitespace_count );
20052 # Handle entab option
20053 elsif ($rOpts_entab_leading_whitespace) {
20055 $leading_whitespace_count % $rOpts_entab_leading_whitespace;
20056 my $tab_count = int(
20057 $leading_whitespace_count / $rOpts_entab_leading_whitespace );
20058 $leading_string = "\t" x $tab_count . ' ' x $space_count;
20061 # Handle option of one tab per level
20063 $leading_string = ( "\t" x $group_level );
20065 $leading_whitespace_count - $group_level * $rOpts_indent_columns;
20067 # shouldn't happen:
20068 if ( $space_count < 0 ) {
20070 "Error in append_line: for level=$group_level count=$leading_whitespace_count\n"
20072 $leading_string = ( ' ' x $leading_whitespace_count );
20075 $leading_string .= ( ' ' x $space_count );
20078 $leading_string_cache[$leading_whitespace_count] = $leading_string;
20079 return $leading_string;
20081 } # end get_leading_string
20083 sub report_anything_unusual {
20085 if ( $outdented_line_count > 0 ) {
20086 write_logfile_entry(
20087 "$outdented_line_count long lines were outdented:\n");
20088 write_logfile_entry(
20089 " First at output line $first_outdented_line_at\n");
20091 if ( $outdented_line_count > 1 ) {
20092 write_logfile_entry(
20093 " Last at output line $last_outdented_line_at\n");
20095 write_logfile_entry(
20096 " use -noll to prevent outdenting, -l=n to increase line length\n"
20098 write_logfile_entry("\n");
20102 #####################################################################
20104 # the Perl::Tidy::FileWriter class writes the output file
20106 #####################################################################
20108 package Perl::Tidy::FileWriter;
20110 # Maximum number of little messages; probably need not be changed.
20111 use constant MAX_NAG_MESSAGES => 6;
20113 sub write_logfile_entry {
20115 my $logger_object = $self->{_logger_object};
20116 if ($logger_object) {
20117 $logger_object->write_logfile_entry(@_);
20123 my ( $line_sink_object, $rOpts, $logger_object ) = @_;
20126 _line_sink_object => $line_sink_object,
20127 _logger_object => $logger_object,
20129 _output_line_number => 1,
20130 _consecutive_blank_lines => 0,
20131 _consecutive_nonblank_lines => 0,
20132 _first_line_length_error => 0,
20133 _max_line_length_error => 0,
20134 _last_line_length_error => 0,
20135 _first_line_length_error_at => 0,
20136 _max_line_length_error_at => 0,
20137 _last_line_length_error_at => 0,
20138 _line_length_error_count => 0,
20139 _max_output_line_length => 0,
20140 _max_output_line_length_at => 0,
20146 $self->{_line_sink_object}->tee_on();
20151 $self->{_line_sink_object}->tee_off();
20154 sub get_output_line_number {
20156 return $self->{_output_line_number};
20159 sub decrement_output_line_number {
20161 $self->{_output_line_number}--;
20164 sub get_consecutive_nonblank_lines {
20166 return $self->{_consecutive_nonblank_lines};
20169 sub reset_consecutive_blank_lines {
20171 $self->{_consecutive_blank_lines} = 0;
20174 sub want_blank_line {
20176 unless ( $self->{_consecutive_blank_lines} ) {
20177 $self->write_blank_code_line();
20181 sub write_blank_code_line {
20183 my $rOpts = $self->{_rOpts};
20185 if ( $self->{_consecutive_blank_lines} >=
20186 $rOpts->{'maximum-consecutive-blank-lines'} );
20187 $self->{_consecutive_blank_lines}++;
20188 $self->{_consecutive_nonblank_lines} = 0;
20189 $self->write_line("\n");
20192 sub write_code_line {
20196 if ( $a =~ /^\s*$/ ) {
20197 my $rOpts = $self->{_rOpts};
20199 if ( $self->{_consecutive_blank_lines} >=
20200 $rOpts->{'maximum-consecutive-blank-lines'} );
20201 $self->{_consecutive_blank_lines}++;
20202 $self->{_consecutive_nonblank_lines} = 0;
20205 $self->{_consecutive_blank_lines} = 0;
20206 $self->{_consecutive_nonblank_lines}++;
20208 $self->write_line($a);
20215 # TODO: go through and see if the test is necessary here
20216 if ( $a =~ /\n$/ ) { $self->{_output_line_number}++; }
20218 $self->{_line_sink_object}->write_line($a);
20220 # This calculation of excess line length ignores any internal tabs
20221 my $rOpts = $self->{_rOpts};
20222 my $exceed = length($a) - $rOpts->{'maximum-line-length'} - 1;
20223 if ( $a =~ /^\t+/g ) {
20224 $exceed += pos($a) * ( $rOpts->{'indent-columns'} - 1 );
20227 # Note that we just incremented output line number to future value
20228 # so we must subtract 1 for current line number
20229 if ( length($a) > 1 + $self->{_max_output_line_length} ) {
20230 $self->{_max_output_line_length} = length($a) - 1;
20231 $self->{_max_output_line_length_at} = $self->{_output_line_number} - 1;
20234 if ( $exceed > 0 ) {
20235 my $output_line_number = $self->{_output_line_number};
20236 $self->{_last_line_length_error} = $exceed;
20237 $self->{_last_line_length_error_at} = $output_line_number - 1;
20238 if ( $self->{_line_length_error_count} == 0 ) {
20239 $self->{_first_line_length_error} = $exceed;
20240 $self->{_first_line_length_error_at} = $output_line_number - 1;
20244 $self->{_last_line_length_error} > $self->{_max_line_length_error} )
20246 $self->{_max_line_length_error} = $exceed;
20247 $self->{_max_line_length_error_at} = $output_line_number - 1;
20250 if ( $self->{_line_length_error_count} < MAX_NAG_MESSAGES ) {
20251 $self->write_logfile_entry(
20252 "Line length exceeded by $exceed characters\n");
20254 $self->{_line_length_error_count}++;
20259 sub report_line_length_errors {
20261 my $rOpts = $self->{_rOpts};
20262 my $line_length_error_count = $self->{_line_length_error_count};
20263 if ( $line_length_error_count == 0 ) {
20264 $self->write_logfile_entry(
20265 "No lines exceeded $rOpts->{'maximum-line-length'} characters\n");
20266 my $max_output_line_length = $self->{_max_output_line_length};
20267 my $max_output_line_length_at = $self->{_max_output_line_length_at};
20268 $self->write_logfile_entry(
20269 " Maximum output line length was $max_output_line_length at line $max_output_line_length_at\n"
20275 my $word = ( $line_length_error_count > 1 ) ? "s" : "";
20276 $self->write_logfile_entry(
20277 "$line_length_error_count output line$word exceeded $rOpts->{'maximum-line-length'} characters:\n"
20280 $word = ( $line_length_error_count > 1 ) ? "First" : "";
20281 my $first_line_length_error = $self->{_first_line_length_error};
20282 my $first_line_length_error_at = $self->{_first_line_length_error_at};
20283 $self->write_logfile_entry(
20284 " $word at line $first_line_length_error_at by $first_line_length_error characters\n"
20287 if ( $line_length_error_count > 1 ) {
20288 my $max_line_length_error = $self->{_max_line_length_error};
20289 my $max_line_length_error_at = $self->{_max_line_length_error_at};
20290 my $last_line_length_error = $self->{_last_line_length_error};
20291 my $last_line_length_error_at = $self->{_last_line_length_error_at};
20292 $self->write_logfile_entry(
20293 " Maximum at line $max_line_length_error_at by $max_line_length_error characters\n"
20295 $self->write_logfile_entry(
20296 " Last at line $last_line_length_error_at by $last_line_length_error characters\n"
20302 #####################################################################
20304 # The Perl::Tidy::Debugger class shows line tokenization
20306 #####################################################################
20308 package Perl::Tidy::Debugger;
20312 my ( $class, $filename ) = @_;
20315 _debug_file => $filename,
20316 _debug_file_opened => 0,
20321 sub really_open_debug_file {
20324 my $debug_file = $self->{_debug_file};
20326 unless ( $fh = IO::File->new("> $debug_file") ) {
20327 warn("can't open $debug_file: $!\n");
20329 $self->{_debug_file_opened} = 1;
20330 $self->{_fh} = $fh;
20332 "Use -dump-token-types (-dtt) to get a list of token type codes\n";
20335 sub close_debug_file {
20338 my $fh = $self->{_fh};
20339 if ( $self->{_debug_file_opened} ) {
20341 eval { $self->{_fh}->close() };
20345 sub write_debug_entry {
20347 # This is a debug dump routine which may be modified as necessary
20348 # to dump tokens on a line-by-line basis. The output will be written
20349 # to the .DEBUG file when the -D flag is entered.
20351 my $line_of_tokens = shift;
20353 my $input_line = $line_of_tokens->{_line_text};
20354 my $rtoken_type = $line_of_tokens->{_rtoken_type};
20355 my $rtokens = $line_of_tokens->{_rtokens};
20356 my $rlevels = $line_of_tokens->{_rlevels};
20357 my $rslevels = $line_of_tokens->{_rslevels};
20358 my $rblock_type = $line_of_tokens->{_rblock_type};
20359 my $input_line_number = $line_of_tokens->{_line_number};
20360 my $line_type = $line_of_tokens->{_line_type};
20364 my $token_str = "$input_line_number: ";
20365 my $reconstructed_original = "$input_line_number: ";
20366 my $block_str = "$input_line_number: ";
20368 #$token_str .= "$line_type: ";
20369 #$reconstructed_original .= "$line_type: ";
20372 my @next_char = ( '"', '"' );
20374 unless ( $self->{_debug_file_opened} ) { $self->really_open_debug_file() }
20375 my $fh = $self->{_fh};
20377 for ( $j = 0 ; $j < @$rtoken_type ; $j++ ) {
20380 if ( $$rtoken_type[$j] eq 'k' ) {
20381 $pattern .= $$rtokens[$j];
20384 $pattern .= $$rtoken_type[$j];
20386 $reconstructed_original .= $$rtokens[$j];
20387 $block_str .= "($$rblock_type[$j])";
20388 $num = length( $$rtokens[$j] );
20389 my $type_str = $$rtoken_type[$j];
20391 # be sure there are no blank tokens (shouldn't happen)
20392 # This can only happen if a programming error has been made
20393 # because all valid tokens are non-blank
20394 if ( $type_str eq ' ' ) {
20395 print $fh "BLANK TOKEN on the next line\n";
20396 $type_str = $next_char[$i_next];
20397 $i_next = 1 - $i_next;
20400 if ( length($type_str) == 1 ) {
20401 $type_str = $type_str x $num;
20403 $token_str .= $type_str;
20406 # Write what you want here ...
20407 # print $fh "$input_line\n";
20408 # print $fh "$pattern\n";
20409 print $fh "$reconstructed_original\n";
20410 print $fh "$token_str\n";
20412 #print $fh "$block_str\n";
20415 #####################################################################
20417 # The Perl::Tidy::LineBuffer class supplies a 'get_line()'
20418 # method for returning the next line to be parsed, as well as a
20419 # 'peek_ahead()' method
20421 # The input parameter is an object with a 'get_line()' method
20422 # which returns the next line to be parsed
20424 #####################################################################
20426 package Perl::Tidy::LineBuffer;
20431 my $line_source_object = shift;
20434 _line_source_object => $line_source_object,
20435 _rlookahead_buffer => [],
20441 my $buffer_index = shift;
20443 my $line_source_object = $self->{_line_source_object};
20444 my $rlookahead_buffer = $self->{_rlookahead_buffer};
20445 if ( $buffer_index < scalar(@$rlookahead_buffer) ) {
20446 $line = $$rlookahead_buffer[$buffer_index];
20449 $line = $line_source_object->get_line();
20450 push( @$rlookahead_buffer, $line );
20458 my $line_source_object = $self->{_line_source_object};
20459 my $rlookahead_buffer = $self->{_rlookahead_buffer};
20461 if ( scalar(@$rlookahead_buffer) ) {
20462 $line = shift @$rlookahead_buffer;
20465 $line = $line_source_object->get_line();
20470 ########################################################################
20472 # the Perl::Tidy::Tokenizer package is essentially a filter which
20473 # reads lines of perl source code from a source object and provides
20474 # corresponding tokenized lines through its get_line() method. Lines
20475 # flow from the source_object to the caller like this:
20477 # source_object --> LineBuffer_object --> Tokenizer --> calling routine
20478 # get_line() get_line() get_line() line_of_tokens
20480 # The source object can be any object with a get_line() method which
20481 # supplies one line (a character string) perl call.
20482 # The LineBuffer object is created by the Tokenizer.
20483 # The Tokenizer returns a reference to a data structure 'line_of_tokens'
20484 # containing one tokenized line for each call to its get_line() method.
20486 # WARNING: This is not a real class yet. Only one tokenizer my be used.
20488 ########################################################################
20490 package Perl::Tidy::Tokenizer;
20494 # Caution: these debug flags produce a lot of output
20495 # They should all be 0 except when debugging small scripts
20497 use constant TOKENIZER_DEBUG_FLAG_EXPECT => 0;
20498 use constant TOKENIZER_DEBUG_FLAG_NSCAN => 0;
20499 use constant TOKENIZER_DEBUG_FLAG_QUOTE => 0;
20500 use constant TOKENIZER_DEBUG_FLAG_SCAN_ID => 0;
20501 use constant TOKENIZER_DEBUG_FLAG_TOKENIZE => 0;
20503 my $debug_warning = sub {
20504 print "TOKENIZER_DEBUGGING with key $_[0]\n";
20507 TOKENIZER_DEBUG_FLAG_EXPECT && $debug_warning->('EXPECT');
20508 TOKENIZER_DEBUG_FLAG_NSCAN && $debug_warning->('NSCAN');
20509 TOKENIZER_DEBUG_FLAG_QUOTE && $debug_warning->('QUOTE');
20510 TOKENIZER_DEBUG_FLAG_SCAN_ID && $debug_warning->('SCAN_ID');
20511 TOKENIZER_DEBUG_FLAG_TOKENIZE && $debug_warning->('TOKENIZE');
20517 # PACKAGE VARIABLES for for processing an entire FILE.
20521 $last_nonblank_token
20522 $last_nonblank_type
20523 $last_nonblank_block_type
20531 %user_function_prototype
20533 %is_block_list_function
20534 %saw_function_definition
20538 $square_bracket_depth
20543 @nesting_sequence_number
20544 @current_sequence_number
20546 @paren_semicolon_count
20547 @paren_structural_type
20549 @brace_structural_type
20550 @brace_statement_type
20553 @square_bracket_type
20554 @square_bracket_structural_type
20556 @nested_ternary_flag
20557 @starting_line_of_current_depth
20560 # GLOBAL CONSTANTS for routines in this package
20562 %is_indirect_object_taker
20564 %expecting_operator_token
20565 %expecting_operator_types
20566 %expecting_term_types
20567 %expecting_term_token
20569 %is_file_test_operator
20571 %is_valid_token_type
20573 %is_code_block_token
20575 @opening_brace_names
20576 @closing_brace_names
20577 %is_keyword_taking_list
20578 %is_q_qq_qw_qx_qr_s_y_tr_m
20581 # possible values of operator_expected()
20582 use constant TERM => -1;
20583 use constant UNKNOWN => 0;
20584 use constant OPERATOR => 1;
20586 # possible values of context
20587 use constant SCALAR_CONTEXT => -1;
20588 use constant UNKNOWN_CONTEXT => 0;
20589 use constant LIST_CONTEXT => 1;
20591 # Maximum number of little messages; probably need not be changed.
20592 use constant MAX_NAG_MESSAGES => 6;
20596 # methods to count instances
20598 sub get_count { $_count; }
20599 sub _increment_count { ++$_count }
20600 sub _decrement_count { --$_count }
20604 $_[0]->_decrement_count();
20611 # Note: 'tabs' and 'indent_columns' are temporary and should be
20614 source_object => undef,
20615 debugger_object => undef,
20616 diagnostics_object => undef,
20617 logger_object => undef,
20618 starting_level => undef,
20619 indent_columns => 4,
20621 look_for_hash_bang => 0,
20623 look_for_autoloader => 1,
20624 look_for_selfloader => 1,
20625 starting_line_number => 1,
20627 my %args = ( %defaults, @_ );
20629 # we are given an object with a get_line() method to supply source lines
20630 my $source_object = $args{source_object};
20632 # we create another object with a get_line() and peek_ahead() method
20633 my $line_buffer_object = Perl::Tidy::LineBuffer->new($source_object);
20635 # Tokenizer state data is as follows:
20636 # _rhere_target_list reference to list of here-doc targets
20637 # _here_doc_target the target string for a here document
20638 # _here_quote_character the type of here-doc quoting (" ' ` or none)
20639 # to determine if interpolation is done
20640 # _quote_target character we seek if chasing a quote
20641 # _line_start_quote line where we started looking for a long quote
20642 # _in_here_doc flag indicating if we are in a here-doc
20643 # _in_pod flag set if we are in pod documentation
20644 # _in_error flag set if we saw severe error (binary in script)
20645 # _in_data flag set if we are in __DATA__ section
20646 # _in_end flag set if we are in __END__ section
20647 # _in_format flag set if we are in a format description
20648 # _in_attribute_list flag telling if we are looking for attributes
20649 # _in_quote flag telling if we are chasing a quote
20650 # _starting_level indentation level of first line
20651 # _input_tabstr string denoting one indentation level of input file
20652 # _know_input_tabstr flag indicating if we know _input_tabstr
20653 # _line_buffer_object object with get_line() method to supply source code
20654 # _diagnostics_object place to write debugging information
20655 # _unexpected_error_count error count used to limit output
20656 # _lower_case_labels_at line numbers where lower case labels seen
20657 $tokenizer_self = {
20658 _rhere_target_list => [],
20660 _here_doc_target => "",
20661 _here_quote_character => "",
20667 _in_attribute_list => 0,
20669 _quote_target => "",
20670 _line_start_quote => -1,
20671 _starting_level => $args{starting_level},
20672 _know_starting_level => defined( $args{starting_level} ),
20673 _tabs => $args{tabs},
20674 _indent_columns => $args{indent_columns},
20675 _look_for_hash_bang => $args{look_for_hash_bang},
20676 _trim_qw => $args{trim_qw},
20677 _input_tabstr => "",
20678 _know_input_tabstr => -1,
20679 _last_line_number => $args{starting_line_number} - 1,
20680 _saw_perl_dash_P => 0,
20681 _saw_perl_dash_w => 0,
20682 _saw_use_strict => 0,
20683 _saw_v_string => 0,
20684 _look_for_autoloader => $args{look_for_autoloader},
20685 _look_for_selfloader => $args{look_for_selfloader},
20686 _saw_autoloader => 0,
20687 _saw_selfloader => 0,
20688 _saw_hash_bang => 0,
20691 _saw_negative_indentation => 0,
20692 _started_tokenizing => 0,
20693 _line_buffer_object => $line_buffer_object,
20694 _debugger_object => $args{debugger_object},
20695 _diagnostics_object => $args{diagnostics_object},
20696 _logger_object => $args{logger_object},
20697 _unexpected_error_count => 0,
20698 _started_looking_for_here_target_at => 0,
20699 _nearly_matched_here_target_at => undef,
20701 _rlower_case_labels_at => undef,
20704 prepare_for_a_new_file();
20705 find_starting_indentation_level();
20707 bless $tokenizer_self, $class;
20709 # This is not a full class yet, so die if an attempt is made to
20710 # create more than one object.
20712 if ( _increment_count() > 1 ) {
20714 "Attempt to create more than 1 object in $class, which is not a true class yet\n";
20717 return $tokenizer_self;
20721 # interface to Perl::Tidy::Logger routines
20723 my $logger_object = $tokenizer_self->{_logger_object};
20724 if ($logger_object) {
20725 $logger_object->warning(@_);
20730 my $logger_object = $tokenizer_self->{_logger_object};
20731 if ($logger_object) {
20732 $logger_object->complain(@_);
20736 sub write_logfile_entry {
20737 my $logger_object = $tokenizer_self->{_logger_object};
20738 if ($logger_object) {
20739 $logger_object->write_logfile_entry(@_);
20743 sub interrupt_logfile {
20744 my $logger_object = $tokenizer_self->{_logger_object};
20745 if ($logger_object) {
20746 $logger_object->interrupt_logfile();
20750 sub resume_logfile {
20751 my $logger_object = $tokenizer_self->{_logger_object};
20752 if ($logger_object) {
20753 $logger_object->resume_logfile();
20757 sub increment_brace_error {
20758 my $logger_object = $tokenizer_self->{_logger_object};
20759 if ($logger_object) {
20760 $logger_object->increment_brace_error();
20764 sub report_definite_bug {
20765 my $logger_object = $tokenizer_self->{_logger_object};
20766 if ($logger_object) {
20767 $logger_object->report_definite_bug();
20771 sub brace_warning {
20772 my $logger_object = $tokenizer_self->{_logger_object};
20773 if ($logger_object) {
20774 $logger_object->brace_warning(@_);
20778 sub get_saw_brace_error {
20779 my $logger_object = $tokenizer_self->{_logger_object};
20780 if ($logger_object) {
20781 $logger_object->get_saw_brace_error();
20788 # interface to Perl::Tidy::Diagnostics routines
20789 sub write_diagnostics {
20790 if ( $tokenizer_self->{_diagnostics_object} ) {
20791 $tokenizer_self->{_diagnostics_object}->write_diagnostics(@_);
20795 sub report_tokenization_errors {
20799 my $level = get_indentation_level();
20800 if ( $level != $tokenizer_self->{_starting_level} ) {
20801 warning("final indentation level: $level\n");
20804 check_final_nesting_depths();
20806 if ( $tokenizer_self->{_look_for_hash_bang}
20807 && !$tokenizer_self->{_saw_hash_bang} )
20810 "hit EOF without seeing hash-bang line; maybe don't need -x?\n");
20813 if ( $tokenizer_self->{_in_format} ) {
20814 warning("hit EOF while in format description\n");
20817 if ( $tokenizer_self->{_in_pod} ) {
20819 # Just write log entry if this is after __END__ or __DATA__
20820 # because this happens to often, and it is not likely to be
20822 if ( $tokenizer_self->{_saw_data} || $tokenizer_self->{_saw_end} ) {
20823 write_logfile_entry(
20824 "hit eof while in pod documentation (no =cut seen)\n\tthis can cause trouble with some pod utilities\n"
20830 "hit eof while in pod documentation (no =cut seen)\n\tthis can cause trouble with some pod utilities\n"
20836 if ( $tokenizer_self->{_in_here_doc} ) {
20837 my $here_doc_target = $tokenizer_self->{_here_doc_target};
20838 my $started_looking_for_here_target_at =
20839 $tokenizer_self->{_started_looking_for_here_target_at};
20840 if ($here_doc_target) {
20842 "hit EOF in here document starting at line $started_looking_for_here_target_at with target: $here_doc_target\n"
20847 "hit EOF in here document starting at line $started_looking_for_here_target_at with empty target string\n"
20850 my $nearly_matched_here_target_at =
20851 $tokenizer_self->{_nearly_matched_here_target_at};
20852 if ($nearly_matched_here_target_at) {
20854 "NOTE: almost matched at input line $nearly_matched_here_target_at except for whitespace\n"
20859 if ( $tokenizer_self->{_in_quote} ) {
20860 my $line_start_quote = $tokenizer_self->{_line_start_quote};
20861 my $quote_target = $tokenizer_self->{_quote_target};
20863 ( $tokenizer_self->{_in_attribute_list} )
20867 "hit EOF seeking end of $what starting at line $line_start_quote ending in $quote_target\n"
20871 unless ( $tokenizer_self->{_saw_perl_dash_w} ) {
20872 if ( $] < 5.006 ) {
20873 write_logfile_entry("Suggest including '-w parameter'\n");
20876 write_logfile_entry("Suggest including 'use warnings;'\n");
20880 if ( $tokenizer_self->{_saw_perl_dash_P} ) {
20881 write_logfile_entry("Use of -P parameter for defines is discouraged\n");
20884 unless ( $tokenizer_self->{_saw_use_strict} ) {
20885 write_logfile_entry("Suggest including 'use strict;'\n");
20888 # it is suggested that lables have at least one upper case character
20889 # for legibility and to avoid code breakage as new keywords are introduced
20890 if ( $tokenizer_self->{_rlower_case_labels_at} ) {
20891 my @lower_case_labels_at =
20892 @{ $tokenizer_self->{_rlower_case_labels_at} };
20893 write_logfile_entry(
20894 "Suggest using upper case characters in label(s)\n");
20896 write_logfile_entry(" defined at line(s): (@lower_case_labels_at)\n");
20900 sub report_v_string {
20902 # warn if this version can't handle v-strings
20904 unless ( $tokenizer_self->{_saw_v_string} ) {
20905 $tokenizer_self->{_saw_v_string} = $tokenizer_self->{_last_line_number};
20907 if ( $] < 5.006 ) {
20909 "Found v-string '$tok' but v-strings are not implemented in your version of perl; see Camel 3 book ch 2\n"
20914 sub get_input_line_number {
20915 return $tokenizer_self->{_last_line_number};
20918 # returns the next tokenized line
20923 # USES GLOBAL VARIABLES: $tokenizer_self, $brace_depth,
20924 # $square_bracket_depth, $paren_depth
20926 my $input_line = $tokenizer_self->{_line_buffer_object}->get_line();
20927 $tokenizer_self->{_line_text} = $input_line;
20929 return undef unless ($input_line);
20931 my $input_line_number = ++$tokenizer_self->{_last_line_number};
20933 # Find and remove what characters terminate this line, including any
20935 my $input_line_separator = "";
20936 if ( chomp($input_line) ) { $input_line_separator = $/ }
20938 # TODO: what other characters should be included here?
20939 if ( $input_line =~ s/((\r|\035|\032)+)$// ) {
20940 $input_line_separator = $2 . $input_line_separator;
20943 # for backwards compatability we keep the line text terminated with
20944 # a newline character
20945 $input_line .= "\n";
20946 $tokenizer_self->{_line_text} = $input_line; # update
20948 # create a data structure describing this line which will be
20949 # returned to the caller.
20951 # _line_type codes are:
20952 # SYSTEM - system-specific code before hash-bang line
20953 # CODE - line of perl code (including comments)
20954 # POD_START - line starting pod, such as '=head'
20955 # POD - pod documentation text
20956 # POD_END - last line of pod section, '=cut'
20957 # HERE - text of here-document
20958 # HERE_END - last line of here-doc (target word)
20959 # FORMAT - format section
20960 # FORMAT_END - last line of format section, '.'
20961 # DATA_START - __DATA__ line
20962 # DATA - unidentified text following __DATA__
20963 # END_START - __END__ line
20964 # END - unidentified text following __END__
20965 # ERROR - we are in big trouble, probably not a perl script
20968 # _curly_brace_depth - depth of curly braces at start of line
20969 # _square_bracket_depth - depth of square brackets at start of line
20970 # _paren_depth - depth of parens at start of line
20971 # _starting_in_quote - this line continues a multi-line quote
20972 # (so don't trim leading blanks!)
20973 # _ending_in_quote - this line ends in a multi-line quote
20974 # (so don't trim trailing blanks!)
20975 my $line_of_tokens = {
20976 _line_type => 'EOF',
20977 _line_text => $input_line,
20978 _line_number => $input_line_number,
20979 _rtoken_type => undef,
20982 _rslevels => undef,
20983 _rblock_type => undef,
20984 _rcontainer_type => undef,
20985 _rcontainer_environment => undef,
20986 _rtype_sequence => undef,
20987 _rnesting_tokens => undef,
20988 _rci_levels => undef,
20989 _rnesting_blocks => undef,
20990 _python_indentation_level => -1, ## 0,
20991 _starting_in_quote => 0, # to be set by subroutine
20992 _ending_in_quote => 0,
20993 _curly_brace_depth => $brace_depth,
20994 _square_bracket_depth => $square_bracket_depth,
20995 _paren_depth => $paren_depth,
20996 _quote_character => '',
20999 # must print line unchanged if we are in a here document
21000 if ( $tokenizer_self->{_in_here_doc} ) {
21002 $line_of_tokens->{_line_type} = 'HERE';
21003 my $here_doc_target = $tokenizer_self->{_here_doc_target};
21004 my $here_quote_character = $tokenizer_self->{_here_quote_character};
21005 my $candidate_target = $input_line;
21006 chomp $candidate_target;
21007 if ( $candidate_target eq $here_doc_target ) {
21008 $tokenizer_self->{_nearly_matched_here_target_at} = undef;
21009 $line_of_tokens->{_line_type} = 'HERE_END';
21010 write_logfile_entry("Exiting HERE document $here_doc_target\n");
21012 my $rhere_target_list = $tokenizer_self->{_rhere_target_list};
21013 if (@$rhere_target_list) { # there can be multiple here targets
21014 ( $here_doc_target, $here_quote_character ) =
21015 @{ shift @$rhere_target_list };
21016 $tokenizer_self->{_here_doc_target} = $here_doc_target;
21017 $tokenizer_self->{_here_quote_character} =
21018 $here_quote_character;
21019 write_logfile_entry(
21020 "Entering HERE document $here_doc_target\n");
21021 $tokenizer_self->{_nearly_matched_here_target_at} = undef;
21022 $tokenizer_self->{_started_looking_for_here_target_at} =
21023 $input_line_number;
21026 $tokenizer_self->{_in_here_doc} = 0;
21027 $tokenizer_self->{_here_doc_target} = "";
21028 $tokenizer_self->{_here_quote_character} = "";
21032 # check for error of extra whitespace
21033 # note for PERL6: leading whitespace is allowed
21035 $candidate_target =~ s/\s*$//;
21036 $candidate_target =~ s/^\s*//;
21037 if ( $candidate_target eq $here_doc_target ) {
21038 $tokenizer_self->{_nearly_matched_here_target_at} =
21039 $input_line_number;
21042 return $line_of_tokens;
21045 # must print line unchanged if we are in a format section
21046 elsif ( $tokenizer_self->{_in_format} ) {
21048 if ( $input_line =~ /^\.[\s#]*$/ ) {
21049 write_logfile_entry("Exiting format section\n");
21050 $tokenizer_self->{_in_format} = 0;
21051 $line_of_tokens->{_line_type} = 'FORMAT_END';
21054 $line_of_tokens->{_line_type} = 'FORMAT';
21056 return $line_of_tokens;
21059 # must print line unchanged if we are in pod documentation
21060 elsif ( $tokenizer_self->{_in_pod} ) {
21062 $line_of_tokens->{_line_type} = 'POD';
21063 if ( $input_line =~ /^=cut/ ) {
21064 $line_of_tokens->{_line_type} = 'POD_END';
21065 write_logfile_entry("Exiting POD section\n");
21066 $tokenizer_self->{_in_pod} = 0;
21068 if ( $input_line =~ /^\#\!.*perl\b/ ) {
21070 "Hash-bang in pod can cause older versions of perl to fail! \n"
21074 return $line_of_tokens;
21077 # must print line unchanged if we have seen a severe error (i.e., we
21078 # are seeing illegal tokens and connot continue. Syntax errors do
21079 # not pass this route). Calling routine can decide what to do, but
21080 # the default can be to just pass all lines as if they were after __END__
21081 elsif ( $tokenizer_self->{_in_error} ) {
21082 $line_of_tokens->{_line_type} = 'ERROR';
21083 return $line_of_tokens;
21086 # print line unchanged if we are __DATA__ section
21087 elsif ( $tokenizer_self->{_in_data} ) {
21089 # ...but look for POD
21090 # Note that the _in_data and _in_end flags remain set
21091 # so that we return to that state after seeing the
21092 # end of a pod section
21093 if ( $input_line =~ /^=(?!cut)/ ) {
21094 $line_of_tokens->{_line_type} = 'POD_START';
21095 write_logfile_entry("Entering POD section\n");
21096 $tokenizer_self->{_in_pod} = 1;
21097 return $line_of_tokens;
21100 $line_of_tokens->{_line_type} = 'DATA';
21101 return $line_of_tokens;
21105 # print line unchanged if we are in __END__ section
21106 elsif ( $tokenizer_self->{_in_end} ) {
21108 # ...but look for POD
21109 # Note that the _in_data and _in_end flags remain set
21110 # so that we return to that state after seeing the
21111 # end of a pod section
21112 if ( $input_line =~ /^=(?!cut)/ ) {
21113 $line_of_tokens->{_line_type} = 'POD_START';
21114 write_logfile_entry("Entering POD section\n");
21115 $tokenizer_self->{_in_pod} = 1;
21116 return $line_of_tokens;
21119 $line_of_tokens->{_line_type} = 'END';
21120 return $line_of_tokens;
21124 # check for a hash-bang line if we haven't seen one
21125 if ( !$tokenizer_self->{_saw_hash_bang} ) {
21126 if ( $input_line =~ /^\#\!.*perl\b/ ) {
21127 $tokenizer_self->{_saw_hash_bang} = $input_line_number;
21129 # check for -w and -P flags
21130 if ( $input_line =~ /^\#\!.*perl\s.*-.*P/ ) {
21131 $tokenizer_self->{_saw_perl_dash_P} = 1;
21134 if ( $input_line =~ /^\#\!.*perl\s.*-.*w/ ) {
21135 $tokenizer_self->{_saw_perl_dash_w} = 1;
21138 if ( ( $input_line_number > 1 )
21139 && ( !$tokenizer_self->{_look_for_hash_bang} ) )
21142 # this is helpful for VMS systems; we may have accidentally
21143 # tokenized some DCL commands
21144 if ( $tokenizer_self->{_started_tokenizing} ) {
21146 "There seems to be a hash-bang after line 1; do you need to run with -x ?\n"
21150 complain("Useless hash-bang after line 1\n");
21154 # Report the leading hash-bang as a system line
21155 # This will prevent -dac from deleting it
21157 $line_of_tokens->{_line_type} = 'SYSTEM';
21158 return $line_of_tokens;
21163 # wait for a hash-bang before parsing if the user invoked us with -x
21164 if ( $tokenizer_self->{_look_for_hash_bang}
21165 && !$tokenizer_self->{_saw_hash_bang} )
21167 $line_of_tokens->{_line_type} = 'SYSTEM';
21168 return $line_of_tokens;
21171 # a first line of the form ': #' will be marked as SYSTEM
21172 # since lines of this form may be used by tcsh
21173 if ( $input_line_number == 1 && $input_line =~ /^\s*\:\s*\#/ ) {
21174 $line_of_tokens->{_line_type} = 'SYSTEM';
21175 return $line_of_tokens;
21178 # now we know that it is ok to tokenize the line...
21179 # the line tokenizer will modify any of these private variables:
21180 # _rhere_target_list
21187 my $ending_in_quote_last = $tokenizer_self->{_in_quote};
21188 tokenize_this_line($line_of_tokens);
21190 # Now finish defining the return structure and return it
21191 $line_of_tokens->{_ending_in_quote} = $tokenizer_self->{_in_quote};
21193 # handle severe error (binary data in script)
21194 if ( $tokenizer_self->{_in_error} ) {
21195 $tokenizer_self->{_in_quote} = 0; # to avoid any more messages
21196 warning("Giving up after error\n");
21197 $line_of_tokens->{_line_type} = 'ERROR';
21198 reset_indentation_level(0); # avoid error messages
21199 return $line_of_tokens;
21202 # handle start of pod documentation
21203 if ( $tokenizer_self->{_in_pod} ) {
21205 # This gets tricky..above a __DATA__ or __END__ section, perl
21206 # accepts '=cut' as the start of pod section. But afterwards,
21207 # only pod utilities see it and they may ignore an =cut without
21208 # leading =head. In any case, this isn't good.
21209 if ( $input_line =~ /^=cut\b/ ) {
21210 if ( $tokenizer_self->{_saw_data} || $tokenizer_self->{_saw_end} ) {
21211 complain("=cut while not in pod ignored\n");
21212 $tokenizer_self->{_in_pod} = 0;
21213 $line_of_tokens->{_line_type} = 'POD_END';
21216 $line_of_tokens->{_line_type} = 'POD_START';
21218 "=cut starts a pod section .. this can fool pod utilities.\n"
21220 write_logfile_entry("Entering POD section\n");
21225 $line_of_tokens->{_line_type} = 'POD_START';
21226 write_logfile_entry("Entering POD section\n");
21229 return $line_of_tokens;
21232 # update indentation levels for log messages
21233 if ( $input_line !~ /^\s*$/ ) {
21234 my $rlevels = $line_of_tokens->{_rlevels};
21235 my $structural_indentation_level = $$rlevels[0];
21236 my ( $python_indentation_level, $msg ) =
21237 find_indentation_level( $input_line, $structural_indentation_level );
21238 if ($msg) { write_logfile_entry("$msg") }
21239 if ( $tokenizer_self->{_know_input_tabstr} == 1 ) {
21240 $line_of_tokens->{_python_indentation_level} =
21241 $python_indentation_level;
21245 # see if this line contains here doc targets
21246 my $rhere_target_list = $tokenizer_self->{_rhere_target_list};
21247 if (@$rhere_target_list) {
21249 my ( $here_doc_target, $here_quote_character ) =
21250 @{ shift @$rhere_target_list };
21251 $tokenizer_self->{_in_here_doc} = 1;
21252 $tokenizer_self->{_here_doc_target} = $here_doc_target;
21253 $tokenizer_self->{_here_quote_character} = $here_quote_character;
21254 write_logfile_entry("Entering HERE document $here_doc_target\n");
21255 $tokenizer_self->{_started_looking_for_here_target_at} =
21256 $input_line_number;
21259 # NOTE: __END__ and __DATA__ statements are written unformatted
21260 # because they can theoretically contain additional characters
21261 # which are not tokenized (and cannot be read with <DATA> either!).
21262 if ( $tokenizer_self->{_in_data} ) {
21263 $line_of_tokens->{_line_type} = 'DATA_START';
21264 write_logfile_entry("Starting __DATA__ section\n");
21265 $tokenizer_self->{_saw_data} = 1;
21267 # keep parsing after __DATA__ if use SelfLoader was seen
21268 if ( $tokenizer_self->{_saw_selfloader} ) {
21269 $tokenizer_self->{_in_data} = 0;
21270 write_logfile_entry(
21271 "SelfLoader seen, continuing; -nlsl deactivates\n");
21274 return $line_of_tokens;
21277 elsif ( $tokenizer_self->{_in_end} ) {
21278 $line_of_tokens->{_line_type} = 'END_START';
21279 write_logfile_entry("Starting __END__ section\n");
21280 $tokenizer_self->{_saw_end} = 1;
21282 # keep parsing after __END__ if use AutoLoader was seen
21283 if ( $tokenizer_self->{_saw_autoloader} ) {
21284 $tokenizer_self->{_in_end} = 0;
21285 write_logfile_entry(
21286 "AutoLoader seen, continuing; -nlal deactivates\n");
21288 return $line_of_tokens;
21291 # now, finally, we know that this line is type 'CODE'
21292 $line_of_tokens->{_line_type} = 'CODE';
21294 # remember if we have seen any real code
21295 if ( !$tokenizer_self->{_started_tokenizing}
21296 && $input_line !~ /^\s*$/
21297 && $input_line !~ /^\s*#/ )
21299 $tokenizer_self->{_started_tokenizing} = 1;
21302 if ( $tokenizer_self->{_debugger_object} ) {
21303 $tokenizer_self->{_debugger_object}->write_debug_entry($line_of_tokens);
21306 # Note: if keyword 'format' occurs in this line code, it is still CODE
21307 # (keyword 'format' need not start a line)
21308 if ( $tokenizer_self->{_in_format} ) {
21309 write_logfile_entry("Entering format section\n");
21312 if ( $tokenizer_self->{_in_quote}
21313 and ( $tokenizer_self->{_line_start_quote} < 0 ) )
21316 #if ( ( my $quote_target = get_quote_target() ) !~ /^\s*$/ ) {
21318 ( my $quote_target = $tokenizer_self->{_quote_target} ) !~ /^\s*$/ )
21320 $tokenizer_self->{_line_start_quote} = $input_line_number;
21321 write_logfile_entry(
21322 "Start multi-line quote or pattern ending in $quote_target\n");
21325 elsif ( ( $tokenizer_self->{_line_start_quote} >= 0 )
21326 and !$tokenizer_self->{_in_quote} )
21328 $tokenizer_self->{_line_start_quote} = -1;
21329 write_logfile_entry("End of multi-line quote or pattern\n");
21332 # we are returning a line of CODE
21333 return $line_of_tokens;
21336 sub find_starting_indentation_level {
21338 # USES GLOBAL VARIABLES: $tokenizer_self
21339 my $starting_level = 0;
21340 my $know_input_tabstr = -1; # flag for find_indentation_level
21342 # use value if given as parameter
21343 if ( $tokenizer_self->{_know_starting_level} ) {
21344 $starting_level = $tokenizer_self->{_starting_level};
21347 # if we know there is a hash_bang line, the level must be zero
21348 elsif ( $tokenizer_self->{_look_for_hash_bang} ) {
21349 $tokenizer_self->{_know_starting_level} = 1;
21352 # otherwise figure it out from the input file
21356 my $structural_indentation_level = -1; # flag for find_indentation_level
21360 $tokenizer_self->{_line_buffer_object}->peek_ahead( $i++ ) )
21363 # if first line is #! then assume starting level is zero
21364 if ( $i == 1 && $line =~ /^\#\!/ ) {
21365 $starting_level = 0;
21368 next if ( $line =~ /^\s*#/ ); # must not be comment
21369 next if ( $line =~ /^\s*$/ ); # must not be blank
21370 ( $starting_level, $msg ) =
21371 find_indentation_level( $line, $structural_indentation_level );
21372 if ($msg) { write_logfile_entry("$msg") }
21375 $msg = "Line $i implies starting-indentation-level = $starting_level\n";
21377 if ( $starting_level > 0 ) {
21379 my $input_tabstr = $tokenizer_self->{_input_tabstr};
21380 if ( $input_tabstr eq "\t" ) {
21381 $msg .= "by guessing input tabbing uses 1 tab per level\n";
21384 my $cols = length($input_tabstr);
21386 "by guessing input tabbing uses $cols blanks per level\n";
21389 write_logfile_entry("$msg");
21391 $tokenizer_self->{_starting_level} = $starting_level;
21392 reset_indentation_level($starting_level);
21395 # Find indentation level given a input line. At the same time, try to
21396 # figure out the input tabbing scheme.
21398 # There are two types of calls:
21400 # Type 1: $structural_indentation_level < 0
21401 # In this case we have to guess $input_tabstr to figure out the level.
21403 # Type 2: $structural_indentation_level >= 0
21404 # In this case the level of this line is known, and this routine can
21405 # update the tabbing string, if still unknown, to make the level correct.
21407 sub find_indentation_level {
21408 my ( $line, $structural_indentation_level ) = @_;
21410 # USES GLOBAL VARIABLES: $tokenizer_self
21414 my $know_input_tabstr = $tokenizer_self->{_know_input_tabstr};
21415 my $input_tabstr = $tokenizer_self->{_input_tabstr};
21417 # find leading whitespace
21418 my $leading_whitespace = ( $line =~ /^(\s*)/ ) ? $1 : "";
21420 # make first guess at input tabbing scheme if necessary
21421 if ( $know_input_tabstr < 0 ) {
21423 $know_input_tabstr = 0;
21425 if ( $tokenizer_self->{_tabs} ) {
21426 $input_tabstr = "\t";
21427 if ( length($leading_whitespace) > 0 ) {
21428 if ( $leading_whitespace !~ /\t/ ) {
21430 my $cols = $tokenizer_self->{_indent_columns};
21432 if ( length($leading_whitespace) < $cols ) {
21433 $cols = length($leading_whitespace);
21435 $input_tabstr = " " x $cols;
21440 $input_tabstr = " " x $tokenizer_self->{_indent_columns};
21442 if ( length($leading_whitespace) > 0 ) {
21443 if ( $leading_whitespace =~ /^\t/ ) {
21444 $input_tabstr = "\t";
21448 $tokenizer_self->{_know_input_tabstr} = $know_input_tabstr;
21449 $tokenizer_self->{_input_tabstr} = $input_tabstr;
21452 # determine the input tabbing scheme if possible
21453 if ( ( $know_input_tabstr == 0 )
21454 && ( length($leading_whitespace) > 0 )
21455 && ( $structural_indentation_level > 0 ) )
21457 my $saved_input_tabstr = $input_tabstr;
21459 # check for common case of one tab per indentation level
21460 if ( $leading_whitespace eq "\t" x $structural_indentation_level ) {
21461 if ( $leading_whitespace eq "\t" x $structural_indentation_level ) {
21462 $input_tabstr = "\t";
21463 $msg = "Guessing old indentation was tab character\n";
21469 # detab any tabs based on 8 blanks per tab
21471 if ( $leading_whitespace =~ s/^\t+/ /g ) {
21472 $entabbed = "entabbed";
21475 # now compute tabbing from number of spaces
21477 length($leading_whitespace) / $structural_indentation_level;
21478 if ( $columns == int $columns ) {
21480 "Guessing old indentation was $columns $entabbed spaces\n";
21483 $columns = int $columns;
21485 "old indentation is unclear, using $columns $entabbed spaces\n";
21487 $input_tabstr = " " x $columns;
21489 $know_input_tabstr = 1;
21490 $tokenizer_self->{_know_input_tabstr} = $know_input_tabstr;
21491 $tokenizer_self->{_input_tabstr} = $input_tabstr;
21493 # see if mistakes were made
21494 if ( ( $tokenizer_self->{_starting_level} > 0 )
21495 && !$tokenizer_self->{_know_starting_level} )
21498 if ( $input_tabstr ne $saved_input_tabstr ) {
21500 "I made a bad starting level guess; rerun with a value for -sil \n"
21506 # use current guess at input tabbing to get input indentation level
21508 # Patch to handle a common case of entabbed leading whitespace
21509 # If the leading whitespace equals 4 spaces and we also have
21510 # tabs, detab the input whitespace assuming 8 spaces per tab.
21511 if ( length($input_tabstr) == 4 ) {
21512 $leading_whitespace =~ s/^\t+/ /g;
21515 if ( ( my $len_tab = length($input_tabstr) ) > 0 ) {
21518 while ( substr( $leading_whitespace, $pos, $len_tab ) eq $input_tabstr )
21524 return ( $level, $msg );
21527 # This is a currently unused debug routine
21528 sub dump_functions {
21532 foreach $pkg ( keys %is_user_function ) {
21533 print $fh "\nnon-constant subs in package $pkg\n";
21535 foreach $sub ( keys %{ $is_user_function{$pkg} } ) {
21537 if ( $is_block_list_function{$pkg}{$sub} ) {
21538 $msg = 'block_list';
21541 if ( $is_block_function{$pkg}{$sub} ) {
21544 print $fh "$sub $msg\n";
21548 foreach $pkg ( keys %is_constant ) {
21549 print $fh "\nconstants and constant subs in package $pkg\n";
21551 foreach $sub ( keys %{ $is_constant{$pkg} } ) {
21552 print $fh "$sub\n";
21559 # count number of 1's in a string of 1's and 0's
21560 # example: ones_count("010101010101") gives 6
21561 return ( my $cis = $_[0] ) =~ tr/1/0/;
21564 sub prepare_for_a_new_file {
21566 # previous tokens needed to determine what to expect next
21567 $last_nonblank_token = ';'; # the only possible starting state which
21568 $last_nonblank_type = ';'; # will make a leading brace a code block
21569 $last_nonblank_block_type = '';
21571 # scalars for remembering statement types across multiple lines
21572 $statement_type = ''; # '' or 'use' or 'sub..' or 'case..'
21573 $in_attribute_list = 0;
21575 # scalars for remembering where we are in the file
21576 $current_package = "main";
21577 $context = UNKNOWN_CONTEXT;
21579 # hashes used to remember function information
21580 %is_constant = (); # user-defined constants
21581 %is_user_function = (); # user-defined functions
21582 %user_function_prototype = (); # their prototypes
21583 %is_block_function = ();
21584 %is_block_list_function = ();
21585 %saw_function_definition = ();
21587 # variables used to track depths of various containers
21588 # and report nesting errors
21591 $square_bracket_depth = 0;
21592 @current_depth[ 0 .. $#closing_brace_names ] =
21593 (0) x scalar @closing_brace_names;
21596 @nesting_sequence_number[ 0 .. $#closing_brace_names ] =
21597 ( 0 .. $#closing_brace_names );
21598 @current_sequence_number = ();
21599 $paren_type[$paren_depth] = '';
21600 $paren_semicolon_count[$paren_depth] = 0;
21601 $paren_structural_type[$brace_depth] = '';
21602 $brace_type[$brace_depth] = ';'; # identify opening brace as code block
21603 $brace_structural_type[$brace_depth] = '';
21604 $brace_statement_type[$brace_depth] = "";
21605 $brace_context[$brace_depth] = UNKNOWN_CONTEXT;
21606 $brace_package[$paren_depth] = $current_package;
21607 $square_bracket_type[$square_bracket_depth] = '';
21608 $square_bracket_structural_type[$square_bracket_depth] = '';
21610 initialize_tokenizer_state();
21613 { # begin tokenize_this_line
21615 use constant BRACE => 0;
21616 use constant SQUARE_BRACKET => 1;
21617 use constant PAREN => 2;
21618 use constant QUESTION_COLON => 3;
21620 # TV1: scalars for processing one LINE.
21621 # Re-initialized on each entry to sub tokenize_this_line.
21623 $block_type, $container_type, $expecting,
21624 $i, $i_tok, $input_line,
21625 $input_line_number, $last_nonblank_i, $max_token_index,
21626 $next_tok, $next_type, $peeked_ahead,
21627 $prototype, $rhere_target_list, $rtoken_map,
21628 $rtoken_type, $rtokens, $tok,
21629 $type, $type_sequence, $indent_flag,
21632 # TV2: refs to ARRAYS for processing one LINE
21633 # Re-initialized on each call.
21634 my $routput_token_list = []; # stack of output token indexes
21635 my $routput_token_type = []; # token types
21636 my $routput_block_type = []; # types of code block
21637 my $routput_container_type = []; # paren types, such as if, elsif, ..
21638 my $routput_type_sequence = []; # nesting sequential number
21639 my $routput_indent_flag = []; #
21641 # TV3: SCALARS for quote variables. These are initialized with a
21642 # subroutine call and continually updated as lines are processed.
21643 my ( $in_quote, $quote_type, $quote_character, $quote_pos, $quote_depth,
21644 $quoted_string_1, $quoted_string_2, $allowed_quote_modifiers, );
21646 # TV4: SCALARS for multi-line identifiers and
21647 # statements. These are initialized with a subroutine call
21648 # and continually updated as lines are processed.
21649 my ( $id_scan_state, $identifier, $want_paren, $indented_if_level );
21651 # TV5: SCALARS for tracking indentation level.
21652 # Initialized once and continually updated as lines are
21655 $nesting_token_string, $nesting_type_string,
21656 $nesting_block_string, $nesting_block_flag,
21657 $nesting_list_string, $nesting_list_flag,
21658 $ci_string_in_tokenizer, $continuation_string_in_tokenizer,
21659 $in_statement_continuation, $level_in_tokenizer,
21660 $slevel_in_tokenizer, $rslevel_stack,
21663 # TV6: SCALARS for remembering several previous
21664 # tokens. Initialized once and continually updated as
21665 # lines are processed.
21667 $last_nonblank_container_type, $last_nonblank_type_sequence,
21668 $last_last_nonblank_token, $last_last_nonblank_type,
21669 $last_last_nonblank_block_type, $last_last_nonblank_container_type,
21670 $last_last_nonblank_type_sequence, $last_nonblank_prototype,
21673 # ----------------------------------------------------------------
21674 # beginning of tokenizer variable access and manipulation routines
21675 # ----------------------------------------------------------------
21677 sub initialize_tokenizer_state {
21679 # TV1: initialized on each call
21680 # TV2: initialized on each call
21684 $quote_character = "";
21687 $quoted_string_1 = "";
21688 $quoted_string_2 = "";
21689 $allowed_quote_modifiers = "";
21692 $id_scan_state = '';
21695 $indented_if_level = 0;
21698 $nesting_token_string = "";
21699 $nesting_type_string = "";
21700 $nesting_block_string = '1'; # initially in a block
21701 $nesting_block_flag = 1;
21702 $nesting_list_string = '0'; # initially not in a list
21703 $nesting_list_flag = 0; # initially not in a list
21704 $ci_string_in_tokenizer = "";
21705 $continuation_string_in_tokenizer = "0";
21706 $in_statement_continuation = 0;
21707 $level_in_tokenizer = 0;
21708 $slevel_in_tokenizer = 0;
21709 $rslevel_stack = [];
21712 $last_nonblank_container_type = '';
21713 $last_nonblank_type_sequence = '';
21714 $last_last_nonblank_token = ';';
21715 $last_last_nonblank_type = ';';
21716 $last_last_nonblank_block_type = '';
21717 $last_last_nonblank_container_type = '';
21718 $last_last_nonblank_type_sequence = '';
21719 $last_nonblank_prototype = "";
21722 sub save_tokenizer_state {
21725 $block_type, $container_type, $expecting,
21726 $i, $i_tok, $input_line,
21727 $input_line_number, $last_nonblank_i, $max_token_index,
21728 $next_tok, $next_type, $peeked_ahead,
21729 $prototype, $rhere_target_list, $rtoken_map,
21730 $rtoken_type, $rtokens, $tok,
21731 $type, $type_sequence, $indent_flag,
21735 $routput_token_list, $routput_token_type,
21736 $routput_block_type, $routput_container_type,
21737 $routput_type_sequence, $routput_indent_flag,
21741 $in_quote, $quote_type,
21742 $quote_character, $quote_pos,
21743 $quote_depth, $quoted_string_1,
21744 $quoted_string_2, $allowed_quote_modifiers,
21748 [ $id_scan_state, $identifier, $want_paren, $indented_if_level ];
21751 $nesting_token_string, $nesting_type_string,
21752 $nesting_block_string, $nesting_block_flag,
21753 $nesting_list_string, $nesting_list_flag,
21754 $ci_string_in_tokenizer, $continuation_string_in_tokenizer,
21755 $in_statement_continuation, $level_in_tokenizer,
21756 $slevel_in_tokenizer, $rslevel_stack,
21760 $last_nonblank_container_type,
21761 $last_nonblank_type_sequence,
21762 $last_last_nonblank_token,
21763 $last_last_nonblank_type,
21764 $last_last_nonblank_block_type,
21765 $last_last_nonblank_container_type,
21766 $last_last_nonblank_type_sequence,
21767 $last_nonblank_prototype,
21769 return [ $rTV1, $rTV2, $rTV3, $rTV4, $rTV5, $rTV6 ];
21772 sub restore_tokenizer_state {
21774 my ( $rTV1, $rTV2, $rTV3, $rTV4, $rTV5, $rTV6 ) = @{$rstate};
21776 $block_type, $container_type, $expecting,
21777 $i, $i_tok, $input_line,
21778 $input_line_number, $last_nonblank_i, $max_token_index,
21779 $next_tok, $next_type, $peeked_ahead,
21780 $prototype, $rhere_target_list, $rtoken_map,
21781 $rtoken_type, $rtokens, $tok,
21782 $type, $type_sequence, $indent_flag,
21786 $routput_token_list, $routput_token_type,
21787 $routput_block_type, $routput_container_type,
21788 $routput_type_sequence, $routput_type_sequence,
21792 $in_quote, $quote_type, $quote_character, $quote_pos, $quote_depth,
21793 $quoted_string_1, $quoted_string_2, $allowed_quote_modifiers,
21796 ( $id_scan_state, $identifier, $want_paren, $indented_if_level ) =
21800 $nesting_token_string, $nesting_type_string,
21801 $nesting_block_string, $nesting_block_flag,
21802 $nesting_list_string, $nesting_list_flag,
21803 $ci_string_in_tokenizer, $continuation_string_in_tokenizer,
21804 $in_statement_continuation, $level_in_tokenizer,
21805 $slevel_in_tokenizer, $rslevel_stack,
21809 $last_nonblank_container_type,
21810 $last_nonblank_type_sequence,
21811 $last_last_nonblank_token,
21812 $last_last_nonblank_type,
21813 $last_last_nonblank_block_type,
21814 $last_last_nonblank_container_type,
21815 $last_last_nonblank_type_sequence,
21816 $last_nonblank_prototype,
21820 sub get_indentation_level {
21822 # patch to avoid reporting error if indented if is not terminated
21823 if ($indented_if_level) { return $level_in_tokenizer - 1 }
21824 return $level_in_tokenizer;
21827 sub reset_indentation_level {
21828 $level_in_tokenizer = $_[0];
21829 $slevel_in_tokenizer = $_[0];
21830 push @{$rslevel_stack}, $slevel_in_tokenizer;
21834 $peeked_ahead = defined( $_[0] ) ? $_[0] : $peeked_ahead;
21837 # ------------------------------------------------------------
21838 # end of tokenizer variable access and manipulation routines
21839 # ------------------------------------------------------------
21841 # ------------------------------------------------------------
21842 # beginning of various scanner interface routines
21843 # ------------------------------------------------------------
21844 sub scan_replacement_text {
21846 # check for here-docs in replacement text invoked by
21847 # a substitution operator with executable modifier 'e'.
21850 # $replacement_text
21852 # $rht = reference to any here-doc targets
21853 my ($replacement_text) = @_;
21856 return undef unless ( $replacement_text =~ /<</ );
21858 write_logfile_entry("scanning replacement text for here-doc targets\n");
21860 # save the logger object for error messages
21861 my $logger_object = $tokenizer_self->{_logger_object};
21863 # localize all package variables
21865 $tokenizer_self, $last_nonblank_token,
21866 $last_nonblank_type, $last_nonblank_block_type,
21867 $statement_type, $in_attribute_list,
21868 $current_package, $context,
21869 %is_constant, %is_user_function,
21870 %user_function_prototype, %is_block_function,
21871 %is_block_list_function, %saw_function_definition,
21872 $brace_depth, $paren_depth,
21873 $square_bracket_depth, @current_depth,
21874 @total_depth, $total_depth,
21875 @nesting_sequence_number, @current_sequence_number,
21876 @paren_type, @paren_semicolon_count,
21877 @paren_structural_type, @brace_type,
21878 @brace_structural_type, @brace_statement_type,
21879 @brace_context, @brace_package,
21880 @square_bracket_type, @square_bracket_structural_type,
21881 @depth_array, @starting_line_of_current_depth,
21882 @nested_ternary_flag,
21885 # save all lexical variables
21886 my $rstate = save_tokenizer_state();
21887 _decrement_count(); # avoid error check for multiple tokenizers
21889 # make a new tokenizer
21891 my $rpending_logfile_message;
21892 my $source_object =
21893 Perl::Tidy::LineSource->new( \$replacement_text, $rOpts,
21894 $rpending_logfile_message );
21895 my $tokenizer = Perl::Tidy::Tokenizer->new(
21896 source_object => $source_object,
21897 logger_object => $logger_object,
21898 starting_line_number => $input_line_number,
21901 # scan the replacement text
21902 1 while ( $tokenizer->get_line() );
21904 # remove any here doc targets
21906 if ( $tokenizer_self->{_in_here_doc} ) {
21910 $tokenizer_self->{_here_doc_target},
21911 $tokenizer_self->{_here_quote_character}
21913 if ( $tokenizer_self->{_rhere_target_list} ) {
21914 push @{$rht}, @{ $tokenizer_self->{_rhere_target_list} };
21915 $tokenizer_self->{_rhere_target_list} = undef;
21917 $tokenizer_self->{_in_here_doc} = undef;
21920 # now its safe to report errors
21921 $tokenizer->report_tokenization_errors();
21923 # restore all tokenizer lexical variables
21924 restore_tokenizer_state($rstate);
21926 # return the here doc targets
21930 sub scan_bare_identifier {
21931 ( $i, $tok, $type, $prototype ) =
21932 scan_bare_identifier_do( $input_line, $i, $tok, $type, $prototype,
21933 $rtoken_map, $max_token_index );
21936 sub scan_identifier {
21937 ( $i, $tok, $type, $id_scan_state, $identifier ) =
21938 scan_identifier_do( $i, $id_scan_state, $identifier, $rtokens,
21939 $max_token_index, $expecting );
21943 ( $i, $tok, $type, $id_scan_state ) =
21944 scan_id_do( $input_line, $i, $tok, $rtokens, $rtoken_map,
21945 $id_scan_state, $max_token_index );
21950 ( $i, $type, $number ) =
21951 scan_number_do( $input_line, $i, $rtoken_map, $type,
21952 $max_token_index );
21956 # a sub to warn if token found where term expected
21957 sub error_if_expecting_TERM {
21958 if ( $expecting == TERM ) {
21959 if ( $really_want_term{$last_nonblank_type} ) {
21960 unexpected( $tok, "term", $i_tok, $last_nonblank_i, $rtoken_map,
21961 $rtoken_type, $input_line );
21967 # a sub to warn if token found where operator expected
21968 sub error_if_expecting_OPERATOR {
21969 if ( $expecting == OPERATOR ) {
21970 my $thing = defined $_[0] ? $_[0] : $tok;
21971 unexpected( $thing, "operator", $i_tok, $last_nonblank_i,
21972 $rtoken_map, $rtoken_type, $input_line );
21973 if ( $i_tok == 0 ) {
21974 interrupt_logfile();
21975 warning("Missing ';' above?\n");
21982 # ------------------------------------------------------------
21983 # end scanner interfaces
21984 # ------------------------------------------------------------
21986 my %is_for_foreach;
21987 @_ = qw(for foreach);
21988 @is_for_foreach{@_} = (1) x scalar(@_);
21992 @is_my_our{@_} = (1) x scalar(@_);
21994 # These keywords may introduce blocks after parenthesized expressions,
21996 # keyword ( .... ) { BLOCK }
21997 # patch for SWITCH/CASE: added 'switch' 'case' 'given' 'when'
21998 my %is_blocktype_with_paren;
21999 @_ = qw(if elsif unless while until for foreach switch case given when);
22000 @is_blocktype_with_paren{@_} = (1) x scalar(@_);
22002 # ------------------------------------------------------------
22003 # begin hash of code for handling most token types
22004 # ------------------------------------------------------------
22005 my $tokenization_code = {
22007 # no special code for these types yet, but syntax checks
22042 error_if_expecting_TERM()
22043 if ( $expecting == TERM );
22046 error_if_expecting_TERM()
22047 if ( $expecting == TERM );
22051 # start looking for a scalar
22052 error_if_expecting_OPERATOR("Scalar")
22053 if ( $expecting == OPERATOR );
22056 if ( $identifier eq '$^W' ) {
22057 $tokenizer_self->{_saw_perl_dash_w} = 1;
22060 # Check for indentifier in indirect object slot
22061 # (vorboard.pl, sort.t). Something like:
22062 # /^(print|printf|sort|exec|system)$/
22064 $is_indirect_object_taker{$last_nonblank_token}
22066 || ( ( $last_nonblank_token eq '(' )
22067 && $is_indirect_object_taker{ $paren_type[$paren_depth] } )
22068 || ( $last_nonblank_type =~ /^[Uw]$/ ) # possible object
22077 $paren_semicolon_count[$paren_depth] = 0;
22079 $container_type = $want_paren;
22083 $container_type = $last_nonblank_token;
22085 # We can check for a syntax error here of unexpected '(',
22086 # but this is going to get messy...
22088 $expecting == OPERATOR
22090 # be sure this is not a method call of the form
22091 # &method(...), $method->(..), &{method}(...),
22092 # $ref[2](list) is ok & short for $ref[2]->(list)
22093 # NOTE: at present, braces in something like &{ xxx }
22094 # are not marked as a block, we might have a method call
22095 && $last_nonblank_token !~ /^([\]\}\&]|\-\>)/
22100 # ref: camel 3 p 703.
22101 if ( $last_last_nonblank_token eq 'do' ) {
22103 "do SUBROUTINE is deprecated; consider & or -> notation\n"
22108 # if this is an empty list, (), then it is not an
22109 # error; for example, we might have a constant pi and
22110 # invoke it with pi() or just pi;
22111 my ( $next_nonblank_token, $i_next ) =
22112 find_next_nonblank_token( $i, $rtokens,
22113 $max_token_index );
22114 if ( $next_nonblank_token ne ')' ) {
22116 error_if_expecting_OPERATOR('(');
22118 if ( $last_nonblank_type eq 'C' ) {
22120 "$last_nonblank_token has a void prototype\n";
22122 elsif ( $last_nonblank_type eq 'i' ) {
22124 && $last_nonblank_token =~ /^\$/ )
22127 "Do you mean '$last_nonblank_token->(' ?\n";
22131 interrupt_logfile();
22135 } ## end if ( $next_nonblank_token...
22136 } ## end else [ if ( $last_last_nonblank_token...
22137 } ## end if ( $expecting == OPERATOR...
22139 $paren_type[$paren_depth] = $container_type;
22140 ( $type_sequence, $indent_flag ) =
22141 increase_nesting_depth( PAREN, $$rtoken_map[$i_tok] );
22143 # propagate types down through nested parens
22144 # for example: the second paren in 'if ((' would be structural
22145 # since the first is.
22147 if ( $last_nonblank_token eq '(' ) {
22148 $type = $last_nonblank_type;
22151 # We exclude parens as structural after a ',' because it
22152 # causes subtle problems with continuation indentation for
22153 # something like this, where the first 'or' will not get
22158 # ( not defined $check )
22160 # or $check eq "new"
22161 # or $check eq "old",
22164 # Likewise, we exclude parens where a statement can start
22165 # because of problems with continuation indentation, like
22168 # ($firstline =~ /^#\!.*perl/)
22169 # and (print $File::Find::name, "\n")
22172 # (ref($usage_fref) =~ /CODE/)
22174 # : (&blast_usage, &blast_params, &blast_general_params);
22180 if ( $last_nonblank_type eq ')' ) {
22182 "Syntax error? found token '$last_nonblank_type' then '('\n"
22185 $paren_structural_type[$paren_depth] = $type;
22189 ( $type_sequence, $indent_flag ) =
22190 decrease_nesting_depth( PAREN, $$rtoken_map[$i_tok] );
22192 if ( $paren_structural_type[$paren_depth] eq '{' ) {
22196 $container_type = $paren_type[$paren_depth];
22198 # /^(for|foreach)$/
22199 if ( $is_for_foreach{ $paren_type[$paren_depth] } ) {
22200 my $num_sc = $paren_semicolon_count[$paren_depth];
22201 if ( $num_sc > 0 && $num_sc != 2 ) {
22202 warning("Expected 2 ';' in 'for(;;)' but saw $num_sc\n");
22206 if ( $paren_depth > 0 ) { $paren_depth-- }
22209 if ( $last_nonblank_type eq ',' ) {
22210 complain("Repeated ','s \n");
22213 # patch for operator_expected: note if we are in the list (use.t)
22214 if ( $statement_type eq 'use' ) { $statement_type = '_use' }
22215 ## FIXME: need to move this elsewhere, perhaps check after a '('
22216 ## elsif ($last_nonblank_token eq '(') {
22217 ## warning("Leading ','s illegal in some versions of perl\n");
22221 $context = UNKNOWN_CONTEXT;
22222 $statement_type = '';
22224 # /^(for|foreach)$/
22225 if ( $is_for_foreach{ $paren_type[$paren_depth] } )
22226 { # mark ; in for loop
22228 # Be careful: we do not want a semicolon such as the
22229 # following to be included:
22231 # for (sort {strcoll($a,$b);} keys %investments) {
22233 if ( $brace_depth == $depth_array[PAREN][BRACE][$paren_depth]
22234 && $square_bracket_depth ==
22235 $depth_array[PAREN][SQUARE_BRACKET][$paren_depth] )
22239 $paren_semicolon_count[$paren_depth]++;
22245 error_if_expecting_OPERATOR("String")
22246 if ( $expecting == OPERATOR );
22249 $allowed_quote_modifiers = "";
22252 error_if_expecting_OPERATOR("String")
22253 if ( $expecting == OPERATOR );
22256 $allowed_quote_modifiers = "";
22259 error_if_expecting_OPERATOR("String")
22260 if ( $expecting == OPERATOR );
22263 $allowed_quote_modifiers = "";
22268 if ( $expecting == UNKNOWN ) { # indeterminte, must guess..
22270 ( $is_pattern, $msg ) =
22271 guess_if_pattern_or_division( $i, $rtokens, $rtoken_map,
22272 $max_token_index );
22275 write_diagnostics("DIVIDE:$msg\n");
22276 write_logfile_entry($msg);
22279 else { $is_pattern = ( $expecting == TERM ) }
22284 $allowed_quote_modifiers = '[cgimosxp]';
22286 else { # not a pattern; check for a /= token
22288 if ( $$rtokens[ $i + 1 ] eq '=' ) { # form token /=
22294 #DEBUG - collecting info on what tokens follow a divide
22295 # for development of guessing algorithm
22296 #if ( numerator_expected( $i, $rtokens, $max_token_index ) < 0 ) {
22297 # #write_diagnostics( "DIVIDE? $input_line\n" );
22303 # if we just saw a ')', we will label this block with
22304 # its type. We need to do this to allow sub
22305 # code_block_type to determine if this brace starts a
22306 # code block or anonymous hash. (The type of a paren
22307 # pair is the preceding token, such as 'if', 'else',
22309 $container_type = "";
22311 # ATTRS: for a '{' following an attribute list, reset
22312 # things to look like we just saw the sub name
22313 if ( $statement_type =~ /^sub/ ) {
22314 $last_nonblank_token = $statement_type;
22315 $last_nonblank_type = 'i';
22316 $statement_type = "";
22319 # patch for SWITCH/CASE: hide these keywords from an immediately
22320 # following opening brace
22321 elsif ( ( $statement_type eq 'case' || $statement_type eq 'when' )
22322 && $statement_type eq $last_nonblank_token )
22324 $last_nonblank_token = ";";
22327 elsif ( $last_nonblank_token eq ')' ) {
22328 $last_nonblank_token = $paren_type[ $paren_depth + 1 ];
22330 # defensive move in case of a nesting error (pbug.t)
22331 # in which this ')' had no previous '('
22332 # this nesting error will have been caught
22333 if ( !defined($last_nonblank_token) ) {
22334 $last_nonblank_token = 'if';
22337 # check for syntax error here;
22338 unless ( $is_blocktype_with_paren{$last_nonblank_token} ) {
22339 my $list = join( ' ', sort keys %is_blocktype_with_paren );
22341 "syntax error at ') {', didn't see one of: $list\n");
22345 # patch for paren-less for/foreach glitch, part 2.
22346 # see note below under 'qw'
22347 elsif ($last_nonblank_token eq 'qw'
22348 && $is_for_foreach{$want_paren} )
22350 $last_nonblank_token = $want_paren;
22351 if ( $last_last_nonblank_token eq $want_paren ) {
22353 "syntax error at '$want_paren .. {' -- missing \$ loop variable\n"
22360 # now identify which of the three possible types of
22361 # curly braces we have: hash index container, anonymous
22362 # hash reference, or code block.
22364 # non-structural (hash index) curly brace pair
22365 # get marked 'L' and 'R'
22366 if ( is_non_structural_brace() ) {
22369 # patch for SWITCH/CASE:
22370 # allow paren-less identifier after 'when'
22371 # if the brace is preceded by a space
22372 if ( $statement_type eq 'when'
22373 && $last_nonblank_type eq 'i'
22374 && $last_last_nonblank_type eq 'k'
22375 && ( $i_tok == 0 || $rtoken_type->[ $i_tok - 1 ] eq 'b' ) )
22378 $block_type = $statement_type;
22382 # code and anonymous hash have the same type, '{', but are
22383 # distinguished by 'block_type',
22384 # which will be blank for an anonymous hash
22387 $block_type = code_block_type( $i_tok, $rtokens, $rtoken_type,
22388 $max_token_index );
22390 # patch to promote bareword type to function taking block
22392 && $last_nonblank_type eq 'w'
22393 && $last_nonblank_i >= 0 )
22395 if ( $routput_token_type->[$last_nonblank_i] eq 'w' ) {
22396 $routput_token_type->[$last_nonblank_i] = 'G';
22400 # patch for SWITCH/CASE: if we find a stray opening block brace
22401 # where we might accept a 'case' or 'when' block, then take it
22402 if ( $statement_type eq 'case'
22403 || $statement_type eq 'when' )
22405 if ( !$block_type || $block_type eq '}' ) {
22406 $block_type = $statement_type;
22410 $brace_type[ ++$brace_depth ] = $block_type;
22411 $brace_package[$brace_depth] = $current_package;
22412 ( $type_sequence, $indent_flag ) =
22413 increase_nesting_depth( BRACE, $$rtoken_map[$i_tok] );
22414 $brace_structural_type[$brace_depth] = $type;
22415 $brace_context[$brace_depth] = $context;
22416 $brace_statement_type[$brace_depth] = $statement_type;
22419 $block_type = $brace_type[$brace_depth];
22420 if ($block_type) { $statement_type = '' }
22421 if ( defined( $brace_package[$brace_depth] ) ) {
22422 $current_package = $brace_package[$brace_depth];
22425 # can happen on brace error (caught elsewhere)
22428 ( $type_sequence, $indent_flag ) =
22429 decrease_nesting_depth( BRACE, $$rtoken_map[$i_tok] );
22431 if ( $brace_structural_type[$brace_depth] eq 'L' ) {
22435 # propagate type information for 'do' and 'eval' blocks.
22436 # This is necessary to enable us to know if an operator
22437 # or term is expected next
22438 if ( $is_block_operator{ $brace_type[$brace_depth] } ) {
22439 $tok = $brace_type[$brace_depth];
22442 $context = $brace_context[$brace_depth];
22443 $statement_type = $brace_statement_type[$brace_depth];
22444 if ( $brace_depth > 0 ) { $brace_depth--; }
22446 '&' => sub { # maybe sub call? start looking
22448 # We have to check for sub call unless we are sure we
22449 # are expecting an operator. This example from s2p
22450 # got mistaken as a q operator in an early version:
22451 # print BODY &q(<<'EOT');
22452 if ( $expecting != OPERATOR ) {
22458 '<' => sub { # angle operator or less than?
22460 if ( $expecting != OPERATOR ) {
22462 find_angle_operator_termination( $input_line, $i, $rtoken_map,
22463 $expecting, $max_token_index );
22469 '?' => sub { # ?: conditional or starting pattern?
22473 if ( $expecting == UNKNOWN ) {
22476 ( $is_pattern, $msg ) =
22477 guess_if_pattern_or_conditional( $i, $rtokens, $rtoken_map,
22478 $max_token_index );
22480 if ($msg) { write_logfile_entry($msg) }
22482 else { $is_pattern = ( $expecting == TERM ) }
22487 $allowed_quote_modifiers = '[cgimosxp]';
22490 ( $type_sequence, $indent_flag ) =
22491 increase_nesting_depth( QUESTION_COLON,
22492 $$rtoken_map[$i_tok] );
22495 '*' => sub { # typeglob, or multiply?
22497 if ( $expecting == TERM ) {
22502 if ( $$rtokens[ $i + 1 ] eq '=' ) {
22507 elsif ( $$rtokens[ $i + 1 ] eq '*' ) {
22511 if ( $$rtokens[ $i + 1 ] eq '=' ) {
22519 '.' => sub { # what kind of . ?
22521 if ( $expecting != OPERATOR ) {
22523 if ( $type eq '.' ) {
22524 error_if_expecting_TERM()
22525 if ( $expecting == TERM );
22533 # if this is the first nonblank character, call it a label
22534 # since perl seems to just swallow it
22535 if ( $input_line_number == 1 && $last_nonblank_i == -1 ) {
22539 # ATTRS: check for a ':' which introduces an attribute list
22540 # (this might eventually get its own token type)
22541 elsif ( $statement_type =~ /^sub/ ) {
22543 $in_attribute_list = 1;
22546 # check for scalar attribute, such as
22547 # my $foo : shared = 1;
22548 elsif ($is_my_our{$statement_type}
22549 && $current_depth[QUESTION_COLON] == 0 )
22552 $in_attribute_list = 1;
22555 # otherwise, it should be part of a ?/: operator
22557 ( $type_sequence, $indent_flag ) =
22558 decrease_nesting_depth( QUESTION_COLON,
22559 $$rtoken_map[$i_tok] );
22560 if ( $last_nonblank_token eq '?' ) {
22561 warning("Syntax error near ? :\n");
22565 '+' => sub { # what kind of plus?
22567 if ( $expecting == TERM ) {
22568 my $number = scan_number();
22570 # unary plus is safest assumption if not a number
22571 if ( !defined($number) ) { $type = 'p'; }
22573 elsif ( $expecting == OPERATOR ) {
22576 if ( $next_type eq 'w' ) { $type = 'p' }
22581 error_if_expecting_OPERATOR("Array")
22582 if ( $expecting == OPERATOR );
22585 '%' => sub { # hash or modulo?
22587 # first guess is hash if no following blank
22588 if ( $expecting == UNKNOWN ) {
22589 if ( $next_type ne 'b' ) { $expecting = TERM }
22591 if ( $expecting == TERM ) {
22596 $square_bracket_type[ ++$square_bracket_depth ] =
22597 $last_nonblank_token;
22598 ( $type_sequence, $indent_flag ) =
22599 increase_nesting_depth( SQUARE_BRACKET, $$rtoken_map[$i_tok] );
22601 # It may seem odd, but structural square brackets have
22602 # type '{' and '}'. This simplifies the indentation logic.
22603 if ( !is_non_structural_brace() ) {
22606 $square_bracket_structural_type[$square_bracket_depth] = $type;
22609 ( $type_sequence, $indent_flag ) =
22610 decrease_nesting_depth( SQUARE_BRACKET, $$rtoken_map[$i_tok] );
22612 if ( $square_bracket_structural_type[$square_bracket_depth] eq '{' )
22616 if ( $square_bracket_depth > 0 ) { $square_bracket_depth--; }
22618 '-' => sub { # what kind of minus?
22620 if ( ( $expecting != OPERATOR )
22621 && $is_file_test_operator{$next_tok} )
22623 my ( $next_nonblank_token, $i_next ) =
22624 find_next_nonblank_token( $i + 1, $rtokens,
22625 $max_token_index );
22627 # check for a quoted word like "-w=>xx";
22628 # it is sufficient to just check for a following '='
22629 if ( $next_nonblank_token eq '=' ) {
22638 elsif ( $expecting == TERM ) {
22639 my $number = scan_number();
22641 # maybe part of bareword token? unary is safest
22642 if ( !defined($number) ) { $type = 'm'; }
22645 elsif ( $expecting == OPERATOR ) {
22649 if ( $next_type eq 'w' ) {
22657 # check for special variables like ${^WARNING_BITS}
22658 if ( $expecting == TERM ) {
22660 # FIXME: this should work but will not catch errors
22661 # because we also have to be sure that previous token is
22662 # a type character ($,@,%).
22663 if ( $last_nonblank_token eq '{'
22664 && ( $next_tok =~ /^[A-Za-z_]/ ) )
22667 if ( $next_tok eq 'W' ) {
22668 $tokenizer_self->{_saw_perl_dash_w} = 1;
22670 $tok = $tok . $next_tok;
22676 unless ( error_if_expecting_TERM() ) {
22678 # Something like this is valid but strange:
22680 complain("The '^' seems unusual here\n");
22686 '::' => sub { # probably a sub call
22687 scan_bare_identifier();
22689 '<<' => sub { # maybe a here-doc?
22691 unless ( $i < $max_token_index )
22692 ; # here-doc not possible if end of line
22694 if ( $expecting != OPERATOR ) {
22695 my ( $found_target, $here_doc_target, $here_quote_character,
22698 $found_target, $here_doc_target, $here_quote_character, $i,
22701 = find_here_doc( $expecting, $i, $rtokens, $rtoken_map,
22702 $max_token_index );
22704 if ($found_target) {
22705 push @{$rhere_target_list},
22706 [ $here_doc_target, $here_quote_character ];
22708 if ( length($here_doc_target) > 80 ) {
22709 my $truncated = substr( $here_doc_target, 0, 80 );
22710 complain("Long here-target: '$truncated' ...\n");
22712 elsif ( $here_doc_target !~ /^[A-Z_]\w+$/ ) {
22714 "Unconventional here-target: '$here_doc_target'\n"
22718 elsif ( $expecting == TERM ) {
22719 unless ($saw_error) {
22721 # shouldn't happen..
22722 warning("Program bug; didn't find here doc target\n");
22723 report_definite_bug();
22732 # if -> points to a bare word, we must scan for an identifier,
22733 # otherwise something like ->y would look like the y operator
22737 # type = 'pp' for pre-increment, '++' for post-increment
22739 if ( $expecting == TERM ) { $type = 'pp' }
22740 elsif ( $expecting == UNKNOWN ) {
22741 my ( $next_nonblank_token, $i_next ) =
22742 find_next_nonblank_token( $i, $rtokens, $max_token_index );
22743 if ( $next_nonblank_token eq '$' ) { $type = 'pp' }
22748 if ( $last_nonblank_type eq $tok ) {
22749 complain("Repeated '=>'s \n");
22752 # patch for operator_expected: note if we are in the list (use.t)
22753 # TODO: make version numbers a new token type
22754 if ( $statement_type eq 'use' ) { $statement_type = '_use' }
22757 # type = 'mm' for pre-decrement, '--' for post-decrement
22760 if ( $expecting == TERM ) { $type = 'mm' }
22761 elsif ( $expecting == UNKNOWN ) {
22762 my ( $next_nonblank_token, $i_next ) =
22763 find_next_nonblank_token( $i, $rtokens, $max_token_index );
22764 if ( $next_nonblank_token eq '$' ) { $type = 'mm' }
22769 error_if_expecting_TERM()
22770 if ( $expecting == TERM );
22774 error_if_expecting_TERM()
22775 if ( $expecting == TERM );
22779 error_if_expecting_TERM()
22780 if ( $expecting == TERM );
22784 # ------------------------------------------------------------
22785 # end hash of code for handling individual token types
22786 # ------------------------------------------------------------
22788 my %matching_start_token = ( '}' => '{', ']' => '[', ')' => '(' );
22790 # These block types terminate statements and do not need a trailing
22792 # patched for SWITCH/CASE:
22793 my %is_zero_continuation_block_type;
22794 @_ = qw( } { BEGIN END CHECK INIT AUTOLOAD DESTROY UNITCHECK continue ;
22795 if elsif else unless while until for foreach switch case given when);
22796 @is_zero_continuation_block_type{@_} = (1) x scalar(@_);
22798 my %is_not_zero_continuation_block_type;
22799 @_ = qw(sort grep map do eval);
22800 @is_not_zero_continuation_block_type{@_} = (1) x scalar(@_);
22802 my %is_logical_container;
22803 @_ = qw(if elsif unless while and or err not && ! || for foreach);
22804 @is_logical_container{@_} = (1) x scalar(@_);
22806 my %is_binary_type;
22808 @is_binary_type{@_} = (1) x scalar(@_);
22810 my %is_binary_keyword;
22811 @_ = qw(and or err eq ne cmp);
22812 @is_binary_keyword{@_} = (1) x scalar(@_);
22814 # 'L' is token for opening { at hash key
22815 my %is_opening_type;
22816 @_ = qw" L { ( [ ";
22817 @is_opening_type{@_} = (1) x scalar(@_);
22819 # 'R' is token for closing } at hash key
22820 my %is_closing_type;
22821 @_ = qw" R } ) ] ";
22822 @is_closing_type{@_} = (1) x scalar(@_);
22824 my %is_redo_last_next_goto;
22825 @_ = qw(redo last next goto);
22826 @is_redo_last_next_goto{@_} = (1) x scalar(@_);
22828 my %is_use_require;
22829 @_ = qw(use require);
22830 @is_use_require{@_} = (1) x scalar(@_);
22832 my %is_sub_package;
22833 @_ = qw(sub package);
22834 @is_sub_package{@_} = (1) x scalar(@_);
22836 # This hash holds the hash key in $tokenizer_self for these keywords:
22837 my %is_format_END_DATA = (
22838 'format' => '_in_format',
22839 '__END__' => '_in_end',
22840 '__DATA__' => '_in_data',
22843 # ref: camel 3 p 147,
22844 # but perl may accept undocumented flags
22845 # perl 5.10 adds 'p' (preserve)
22846 my %quote_modifiers = (
22847 's' => '[cegimosxp]',
22850 'm' => '[cgimosxp]',
22851 'qr' => '[imosxp]',
22858 # table showing how many quoted things to look for after quote operator..
22859 # s, y, tr have 2 (pattern and replacement)
22860 # others have 1 (pattern only)
22861 my %quote_items = (
22873 sub tokenize_this_line {
22875 # This routine breaks a line of perl code into tokens which are of use in
22876 # indentation and reformatting. One of my goals has been to define tokens
22877 # such that a newline may be inserted between any pair of tokens without
22878 # changing or invalidating the program. This version comes close to this,
22879 # although there are necessarily a few exceptions which must be caught by
22880 # the formatter. Many of these involve the treatment of bare words.
22882 # The tokens and their types are returned in arrays. See previous
22883 # routine for their names.
22885 # See also the array "valid_token_types" in the BEGIN section for an
22888 # To simplify things, token types are either a single character, or they
22889 # are identical to the tokens themselves.
22891 # As a debugging aid, the -D flag creates a file containing a side-by-side
22892 # comparison of the input string and its tokenization for each line of a file.
22893 # This is an invaluable debugging aid.
22895 # In addition to tokens, and some associated quantities, the tokenizer
22896 # also returns flags indication any special line types. These include
22897 # quotes, here_docs, formats.
22899 # -----------------------------------------------------------------------
22901 # How to add NEW_TOKENS:
22903 # New token types will undoubtedly be needed in the future both to keep up
22904 # with changes in perl and to help adapt the tokenizer to other applications.
22906 # Here are some notes on the minimal steps. I wrote these notes while
22907 # adding the 'v' token type for v-strings, which are things like version
22908 # numbers 5.6.0, and ip addresses, and will use that as an example. ( You
22909 # can use your editor to search for the string "NEW_TOKENS" to find the
22910 # appropriate sections to change):
22912 # *. Try to talk somebody else into doing it! If not, ..
22914 # *. Make a backup of your current version in case things don't work out!
22916 # *. Think of a new, unused character for the token type, and add to
22917 # the array @valid_token_types in the BEGIN section of this package.
22918 # For example, I used 'v' for v-strings.
22920 # *. Implement coding to recognize the $type of the token in this routine.
22921 # This is the hardest part, and is best done by immitating or modifying
22922 # some of the existing coding. For example, to recognize v-strings, I
22923 # patched 'sub scan_bare_identifier' to recognize v-strings beginning with
22924 # 'v' and 'sub scan_number' to recognize v-strings without the leading 'v'.
22926 # *. Update sub operator_expected. This update is critically important but
22927 # the coding is trivial. Look at the comments in that routine for help.
22928 # For v-strings, which should behave like numbers, I just added 'v' to the
22929 # regex used to handle numbers and strings (types 'n' and 'Q').
22931 # *. Implement a 'bond strength' rule in sub set_bond_strengths in
22932 # Perl::Tidy::Formatter for breaking lines around this token type. You can
22933 # skip this step and take the default at first, then adjust later to get
22934 # desired results. For adding type 'v', I looked at sub bond_strength and
22935 # saw that number type 'n' was using default strengths, so I didn't do
22936 # anything. I may tune it up someday if I don't like the way line
22937 # breaks with v-strings look.
22939 # *. Implement a 'whitespace' rule in sub set_white_space_flag in
22940 # Perl::Tidy::Formatter. For adding type 'v', I looked at this routine
22941 # and saw that type 'n' used spaces on both sides, so I just added 'v'
22942 # to the array @spaces_both_sides.
22944 # *. Update HtmlWriter package so that users can colorize the token as
22945 # desired. This is quite easy; see comments identified by 'NEW_TOKENS' in
22946 # that package. For v-strings, I initially chose to use a default color
22947 # equal to the default for numbers, but it might be nice to change that
22950 # *. Update comments in Perl::Tidy::Tokenizer::dump_token_types.
22952 # *. Run lots and lots of debug tests. Start with special files designed
22953 # to test the new token type. Run with the -D flag to create a .DEBUG
22954 # file which shows the tokenization. When these work ok, test as many old
22955 # scripts as possible. Start with all of the '.t' files in the 'test'
22956 # directory of the distribution file. Compare .tdy output with previous
22957 # version and updated version to see the differences. Then include as
22958 # many more files as possible. My own technique has been to collect a huge
22959 # number of perl scripts (thousands!) into one directory and run perltidy
22960 # *, then run diff between the output of the previous version and the
22963 # *. For another example, search for the smartmatch operator '~~'
22964 # with your editor to see where updates were made for it.
22966 # -----------------------------------------------------------------------
22968 my $line_of_tokens = shift;
22969 my ($untrimmed_input_line) = $line_of_tokens->{_line_text};
22971 # patch while coding change is underway
22972 # make callers private data to allow access
22973 # $tokenizer_self = $caller_tokenizer_self;
22975 # extract line number for use in error messages
22976 $input_line_number = $line_of_tokens->{_line_number};
22978 # reinitialize for multi-line quote
22979 $line_of_tokens->{_starting_in_quote} = $in_quote && $quote_type eq 'Q';
22981 # check for pod documentation
22982 if ( ( $untrimmed_input_line =~ /^=[A-Za-z_]/ ) ) {
22984 # must not be in multi-line quote
22985 # and must not be in an eqn
22986 if ( !$in_quote and ( operator_expected( 'b', '=', 'b' ) == TERM ) )
22988 $tokenizer_self->{_in_pod} = 1;
22993 $input_line = $untrimmed_input_line;
22997 # trim start of this line unless we are continuing a quoted line
22998 # do not trim end because we might end in a quote (test: deken4.pl)
22999 # Perl::Tidy::Formatter will delete needless trailing blanks
23000 unless ( $in_quote && ( $quote_type eq 'Q' ) ) {
23001 $input_line =~ s/^\s*//; # trim left end
23004 # update the copy of the line for use in error messages
23005 # This must be exactly what we give the pre_tokenizer
23006 $tokenizer_self->{_line_text} = $input_line;
23008 # re-initialize for the main loop
23009 $routput_token_list = []; # stack of output token indexes
23010 $routput_token_type = []; # token types
23011 $routput_block_type = []; # types of code block
23012 $routput_container_type = []; # paren types, such as if, elsif, ..
23013 $routput_type_sequence = []; # nesting sequential number
23015 $rhere_target_list = [];
23017 $tok = $last_nonblank_token;
23018 $type = $last_nonblank_type;
23019 $prototype = $last_nonblank_prototype;
23020 $last_nonblank_i = -1;
23021 $block_type = $last_nonblank_block_type;
23022 $container_type = $last_nonblank_container_type;
23023 $type_sequence = $last_nonblank_type_sequence;
23027 # tokenization is done in two stages..
23028 # stage 1 is a very simple pre-tokenization
23029 my $max_tokens_wanted = 0; # this signals pre_tokenize to get all tokens
23031 # a little optimization for a full-line comment
23032 if ( !$in_quote && ( $input_line =~ /^#/ ) ) {
23033 $max_tokens_wanted = 1 # no use tokenizing a comment
23036 # start by breaking the line into pre-tokens
23037 ( $rtokens, $rtoken_map, $rtoken_type ) =
23038 pre_tokenize( $input_line, $max_tokens_wanted );
23040 $max_token_index = scalar(@$rtokens) - 1;
23041 push( @$rtokens, ' ', ' ', ' ' ); # extra whitespace simplifies logic
23042 push( @$rtoken_map, 0, 0, 0 ); # shouldn't be referenced
23043 push( @$rtoken_type, 'b', 'b', 'b' );
23045 # initialize for main loop
23046 for $i ( 0 .. $max_token_index + 3 ) {
23047 $routput_token_type->[$i] = "";
23048 $routput_block_type->[$i] = "";
23049 $routput_container_type->[$i] = "";
23050 $routput_type_sequence->[$i] = "";
23051 $routput_indent_flag->[$i] = 0;
23056 # ------------------------------------------------------------
23057 # begin main tokenization loop
23058 # ------------------------------------------------------------
23060 # we are looking at each pre-token of one line and combining them
23062 while ( ++$i <= $max_token_index ) {
23064 if ($in_quote) { # continue looking for end of a quote
23065 $type = $quote_type;
23067 unless ( @{$routput_token_list} )
23068 { # initialize if continuation line
23069 push( @{$routput_token_list}, $i );
23070 $routput_token_type->[$i] = $type;
23073 $tok = $quote_character unless ( $quote_character =~ /^\s*$/ );
23075 # scan for the end of the quote or pattern
23077 $i, $in_quote, $quote_character, $quote_pos, $quote_depth,
23078 $quoted_string_1, $quoted_string_2
23081 $i, $in_quote, $quote_character,
23082 $quote_pos, $quote_depth, $quoted_string_1,
23083 $quoted_string_2, $rtokens, $rtoken_map,
23087 # all done if we didn't find it
23088 last if ($in_quote);
23090 # save pattern and replacement text for rescanning
23091 my $qs1 = $quoted_string_1;
23092 my $qs2 = $quoted_string_2;
23094 # re-initialize for next search
23095 $quote_character = '';
23098 $quoted_string_1 = "";
23099 $quoted_string_2 = "";
23100 last if ( ++$i > $max_token_index );
23102 # look for any modifiers
23103 if ($allowed_quote_modifiers) {
23105 # check for exact quote modifiers
23106 if ( $$rtokens[$i] =~ /^[A-Za-z_]/ ) {
23107 my $str = $$rtokens[$i];
23108 my $saw_modifier_e;
23109 while ( $str =~ /\G$allowed_quote_modifiers/gc ) {
23110 my $pos = pos($str);
23111 my $char = substr( $str, $pos - 1, 1 );
23112 $saw_modifier_e ||= ( $char eq 'e' );
23115 # For an 'e' quote modifier we must scan the replacement
23116 # text for here-doc targets.
23117 if ($saw_modifier_e) {
23119 my $rht = scan_replacement_text($qs1);
23121 # Change type from 'Q' to 'h' for quotes with
23122 # here-doc targets so that the formatter (see sub
23123 # print_line_of_tokens) will not make any line
23124 # breaks after this point.
23126 push @{$rhere_target_list}, @{$rht};
23128 if ( $i_tok < 0 ) {
23129 my $ilast = $routput_token_list->[-1];
23130 $routput_token_type->[$ilast] = $type;
23135 if ( defined( pos($str) ) ) {
23138 if ( pos($str) == length($str) ) {
23139 last if ( ++$i > $max_token_index );
23142 # Looks like a joined quote modifier
23143 # and keyword, maybe something like
23144 # s/xxx/yyy/gefor @k=...
23145 # Example is "galgen.pl". Would have to split
23146 # the word and insert a new token in the
23147 # pre-token list. This is so rare that I haven't
23148 # done it. Will just issue a warning citation.
23150 # This error might also be triggered if my quote
23151 # modifier characters are incomplete
23155 Partial match to quote modifier $allowed_quote_modifiers at word: '$str'
23156 Please put a space between quote modifiers and trailing keywords.
23159 # print "token $$rtokens[$i]\n";
23160 # my $num = length($str) - pos($str);
23161 # $$rtokens[$i]=substr($$rtokens[$i],pos($str),$num);
23162 # print "continuing with new token $$rtokens[$i]\n";
23164 # skipping past this token does least damage
23165 last if ( ++$i > $max_token_index );
23170 # example file: rokicki4.pl
23171 # This error might also be triggered if my quote
23172 # modifier characters are incomplete
23173 write_logfile_entry(
23174 "Note: found word $str at quote modifier location\n"
23180 $allowed_quote_modifiers = "";
23184 unless ( $tok =~ /^\s*$/ ) {
23186 # try to catch some common errors
23187 if ( ( $type eq 'n' ) && ( $tok ne '0' ) ) {
23189 if ( $last_nonblank_token eq 'eq' ) {
23190 complain("Should 'eq' be '==' here ?\n");
23192 elsif ( $last_nonblank_token eq 'ne' ) {
23193 complain("Should 'ne' be '!=' here ?\n");
23197 $last_last_nonblank_token = $last_nonblank_token;
23198 $last_last_nonblank_type = $last_nonblank_type;
23199 $last_last_nonblank_block_type = $last_nonblank_block_type;
23200 $last_last_nonblank_container_type =
23201 $last_nonblank_container_type;
23202 $last_last_nonblank_type_sequence =
23203 $last_nonblank_type_sequence;
23204 $last_nonblank_token = $tok;
23205 $last_nonblank_type = $type;
23206 $last_nonblank_prototype = $prototype;
23207 $last_nonblank_block_type = $block_type;
23208 $last_nonblank_container_type = $container_type;
23209 $last_nonblank_type_sequence = $type_sequence;
23210 $last_nonblank_i = $i_tok;
23213 # store previous token type
23214 if ( $i_tok >= 0 ) {
23215 $routput_token_type->[$i_tok] = $type;
23216 $routput_block_type->[$i_tok] = $block_type;
23217 $routput_container_type->[$i_tok] = $container_type;
23218 $routput_type_sequence->[$i_tok] = $type_sequence;
23219 $routput_indent_flag->[$i_tok] = $indent_flag;
23221 my $pre_tok = $$rtokens[$i]; # get the next pre-token
23222 my $pre_type = $$rtoken_type[$i]; # and type
23224 $type = $pre_type; # to be modified as necessary
23225 $block_type = ""; # blank for all tokens except code block braces
23226 $container_type = ""; # blank for all tokens except some parens
23227 $type_sequence = ""; # blank for all tokens except ?/:
23229 $prototype = ""; # blank for all tokens except user defined subs
23232 # this pre-token will start an output token
23233 push( @{$routput_token_list}, $i_tok );
23235 # continue gathering identifier if necessary
23236 # but do not start on blanks and comments
23237 if ( $id_scan_state && $pre_type !~ /[b#]/ ) {
23239 if ( $id_scan_state =~ /^(sub|package)/ ) {
23246 last if ($id_scan_state);
23247 next if ( ( $i > 0 ) || $type );
23249 # didn't find any token; start over
23254 # handle whitespace tokens..
23255 next if ( $type eq 'b' );
23256 my $prev_tok = $i > 0 ? $$rtokens[ $i - 1 ] : ' ';
23257 my $prev_type = $i > 0 ? $$rtoken_type[ $i - 1 ] : 'b';
23259 # Build larger tokens where possible, since we are not in a quote.
23261 # First try to assemble digraphs. The following tokens are
23262 # excluded and handled specially:
23263 # '/=' is excluded because the / might start a pattern.
23264 # 'x=' is excluded since it might be $x=, with $ on previous line
23265 # '**' and *= might be typeglobs of punctuation variables
23266 # I have allowed tokens starting with <, such as <=,
23267 # because I don't think these could be valid angle operators.
23268 # test file: storrs4.pl
23269 my $test_tok = $tok . $$rtokens[ $i + 1 ];
23270 my $combine_ok = $is_digraph{$test_tok};
23272 # check for special cases which cannot be combined
23275 # '//' must be defined_or operator if an operator is expected.
23276 # TODO: Code for other ambiguous digraphs (/=, x=, **, *=)
23277 # could be migrated here for clarity
23278 if ( $test_tok eq '//' ) {
23279 my $next_type = $$rtokens[ $i + 1 ];
23281 operator_expected( $prev_type, $tok, $next_type );
23282 $combine_ok = 0 unless ( $expecting == OPERATOR );
23288 && ( $test_tok ne '/=' ) # might be pattern
23289 && ( $test_tok ne 'x=' ) # might be $x
23290 && ( $test_tok ne '**' ) # typeglob?
23291 && ( $test_tok ne '*=' ) # typeglob?
23297 # Now try to assemble trigraphs. Note that all possible
23298 # perl trigraphs can be constructed by appending a character
23300 $test_tok = $tok . $$rtokens[ $i + 1 ];
23302 if ( $is_trigraph{$test_tok} ) {
23309 $next_tok = $$rtokens[ $i + 1 ];
23310 $next_type = $$rtoken_type[ $i + 1 ];
23312 TOKENIZER_DEBUG_FLAG_TOKENIZE && do {
23315 $last_nonblank_token, $tok,
23316 $next_tok, $brace_depth,
23317 $brace_type[$brace_depth], $paren_depth,
23318 $paren_type[$paren_depth]
23320 print "TOKENIZE:(@debug_list)\n";
23323 # turn off attribute list on first non-blank, non-bareword
23324 if ( $pre_type ne 'w' ) { $in_attribute_list = 0 }
23326 ###############################################################
23327 # We have the next token, $tok.
23328 # Now we have to examine this token and decide what it is
23329 # and define its $type
23331 # section 1: bare words
23332 ###############################################################
23334 if ( $pre_type eq 'w' ) {
23335 $expecting = operator_expected( $prev_type, $tok, $next_type );
23336 my ( $next_nonblank_token, $i_next ) =
23337 find_next_nonblank_token( $i, $rtokens, $max_token_index );
23339 # ATTRS: handle sub and variable attributes
23340 if ($in_attribute_list) {
23342 # treat bare word followed by open paren like qw(
23343 if ( $next_nonblank_token eq '(' ) {
23344 $in_quote = $quote_items{'q'};
23345 $allowed_quote_modifiers = $quote_modifiers{'q'};
23351 # handle bareword not followed by open paren
23358 # quote a word followed by => operator
23359 if ( $next_nonblank_token eq '=' ) {
23361 if ( $$rtokens[ $i_next + 1 ] eq '>' ) {
23362 if ( $is_constant{$current_package}{$tok} ) {
23365 elsif ( $is_user_function{$current_package}{$tok} ) {
23368 $user_function_prototype{$current_package}{$tok};
23370 elsif ( $tok =~ /^v\d+$/ ) {
23372 report_v_string($tok);
23374 else { $type = 'w' }
23380 # quote a bare word within braces..like xxx->{s}; note that we
23381 # must be sure this is not a structural brace, to avoid
23382 # mistaking {s} in the following for a quoted bare word:
23383 # for(@[){s}bla}BLA}
23384 # Also treat q in something like var{-q} as a bare word, not qoute operator
23385 ##if ( ( $last_nonblank_type eq 'L' )
23386 ## && ( $next_nonblank_token eq '}' ) )
23388 $next_nonblank_token eq '}'
23390 $last_nonblank_type eq 'L'
23391 || ( $last_nonblank_type eq 'm'
23392 && $last_last_nonblank_type eq 'L' )
23400 # a bare word immediately followed by :: is not a keyword;
23401 # use $tok_kw when testing for keywords to avoid a mistake
23403 if ( $$rtokens[ $i + 1 ] eq ':' && $$rtokens[ $i + 2 ] eq ':' )
23408 # handle operator x (now we know it isn't $x=)
23409 if ( ( $tok =~ /^x\d*$/ ) && ( $expecting == OPERATOR ) ) {
23410 if ( $tok eq 'x' ) {
23412 if ( $$rtokens[ $i + 1 ] eq '=' ) { # x=
23422 # FIXME: Patch: mark something like x4 as an integer for now
23423 # It gets fixed downstream. This is easier than
23424 # splitting the pretoken.
23430 elsif ( ( $tok eq 'strict' )
23431 and ( $last_nonblank_token eq 'use' ) )
23433 $tokenizer_self->{_saw_use_strict} = 1;
23434 scan_bare_identifier();
23437 elsif ( ( $tok eq 'warnings' )
23438 and ( $last_nonblank_token eq 'use' ) )
23440 $tokenizer_self->{_saw_perl_dash_w} = 1;
23442 # scan as identifier, so that we pick up something like:
23443 # use warnings::register
23444 scan_bare_identifier();
23448 $tok eq 'AutoLoader'
23449 && $tokenizer_self->{_look_for_autoloader}
23451 $last_nonblank_token eq 'use'
23453 # these regexes are from AutoSplit.pm, which we want
23455 || $input_line =~ /^\s*(use|require)\s+AutoLoader\b/
23456 || $input_line =~ /\bISA\s*=.*\bAutoLoader\b/
23460 write_logfile_entry("AutoLoader seen, -nlal deactivates\n");
23461 $tokenizer_self->{_saw_autoloader} = 1;
23462 $tokenizer_self->{_look_for_autoloader} = 0;
23463 scan_bare_identifier();
23467 $tok eq 'SelfLoader'
23468 && $tokenizer_self->{_look_for_selfloader}
23469 && ( $last_nonblank_token eq 'use'
23470 || $input_line =~ /^\s*(use|require)\s+SelfLoader\b/
23471 || $input_line =~ /\bISA\s*=.*\bSelfLoader\b/ )
23474 write_logfile_entry("SelfLoader seen, -nlsl deactivates\n");
23475 $tokenizer_self->{_saw_selfloader} = 1;
23476 $tokenizer_self->{_look_for_selfloader} = 0;
23477 scan_bare_identifier();
23480 elsif ( ( $tok eq 'constant' )
23481 and ( $last_nonblank_token eq 'use' ) )
23483 scan_bare_identifier();
23484 my ( $next_nonblank_token, $i_next ) =
23485 find_next_nonblank_token( $i, $rtokens,
23486 $max_token_index );
23488 if ($next_nonblank_token) {
23490 if ( $is_keyword{$next_nonblank_token} ) {
23492 "Attempting to define constant '$next_nonblank_token' which is a perl keyword\n"
23496 # FIXME: could check for error in which next token is
23497 # not a word (number, punctuation, ..)
23499 $is_constant{$current_package}
23500 {$next_nonblank_token} = 1;
23505 # various quote operators
23506 elsif ( $is_q_qq_qw_qx_qr_s_y_tr_m{$tok} ) {
23507 if ( $expecting == OPERATOR ) {
23509 # patch for paren-less for/foreach glitch, part 1
23510 # perl will accept this construct as valid:
23512 # foreach my $key qw\Uno Due Tres Quadro\ {
23513 # print "Set $key\n";
23515 unless ( $tok eq 'qw' && $is_for_foreach{$want_paren} )
23517 error_if_expecting_OPERATOR();
23520 $in_quote = $quote_items{$tok};
23521 $allowed_quote_modifiers = $quote_modifiers{$tok};
23523 # All quote types are 'Q' except possibly qw quotes.
23524 # qw quotes are special in that they may generally be trimmed
23525 # of leading and trailing whitespace. So they are given a
23526 # separate type, 'q', unless requested otherwise.
23528 ( $tok eq 'qw' && $tokenizer_self->{_trim_qw} )
23531 $quote_type = $type;
23534 # check for a statement label
23536 ( $next_nonblank_token eq ':' )
23537 && ( $$rtokens[ $i_next + 1 ] ne ':' )
23538 && ( $i_next <= $max_token_index ) # colon on same line
23542 if ( $tok !~ /[A-Z]/ ) {
23543 push @{ $tokenizer_self->{_rlower_case_labels_at} },
23544 $input_line_number;
23552 # 'sub' || 'package'
23553 elsif ( $is_sub_package{$tok_kw} ) {
23554 error_if_expecting_OPERATOR()
23555 if ( $expecting == OPERATOR );
23559 # Note on token types for format, __DATA__, __END__:
23560 # It simplifies things to give these type ';', so that when we
23561 # start rescanning we will be expecting a token of type TERM.
23562 # We will switch to type 'k' before outputting the tokens.
23563 elsif ( $is_format_END_DATA{$tok_kw} ) {
23564 $type = ';'; # make tokenizer look for TERM next
23565 $tokenizer_self->{ $is_format_END_DATA{$tok_kw} } = 1;
23569 elsif ( $is_keyword{$tok_kw} ) {
23572 # Since for and foreach may not be followed immediately
23573 # by an opening paren, we have to remember which keyword
23574 # is associated with the next '('
23575 if ( $is_for_foreach{$tok} ) {
23576 if ( new_statement_ok() ) {
23577 $want_paren = $tok;
23581 # recognize 'use' statements, which are special
23582 elsif ( $is_use_require{$tok} ) {
23583 $statement_type = $tok;
23584 error_if_expecting_OPERATOR()
23585 if ( $expecting == OPERATOR );
23588 # remember my and our to check for trailing ": shared"
23589 elsif ( $is_my_our{$tok} ) {
23590 $statement_type = $tok;
23593 # Check for misplaced 'elsif' and 'else', but allow isolated
23594 # else or elsif blocks to be formatted. This is indicated
23595 # by a last noblank token of ';'
23596 elsif ( $tok eq 'elsif' ) {
23597 if ( $last_nonblank_token ne ';'
23598 && $last_nonblank_block_type !~
23599 /^(if|elsif|unless)$/ )
23602 "expecting '$tok' to follow one of 'if|elsif|unless'\n"
23606 elsif ( $tok eq 'else' ) {
23608 # patched for SWITCH/CASE
23609 if ( $last_nonblank_token ne ';'
23610 && $last_nonblank_block_type !~
23611 /^(if|elsif|unless|case|when)$/ )
23614 "expecting '$tok' to follow one of 'if|elsif|unless|case|when'\n"
23618 elsif ( $tok eq 'continue' ) {
23619 if ( $last_nonblank_token ne ';'
23620 && $last_nonblank_block_type !~
23621 /(^(\{|\}|;|while|until|for|foreach)|:$)/ )
23624 # note: ';' '{' and '}' in list above
23625 # because continues can follow bare blocks;
23626 # ':' is labeled block
23628 ############################################
23629 # NOTE: This check has been deactivated because
23630 # continue has an alternative usage for given/when
23631 # blocks in perl 5.10
23632 ## warning("'$tok' should follow a block\n");
23633 ############################################
23637 # patch for SWITCH/CASE if 'case' and 'when are
23638 # treated as keywords.
23639 elsif ( $tok eq 'when' || $tok eq 'case' ) {
23640 $statement_type = $tok; # next '{' is block
23643 # indent trailing if/unless/while/until
23644 # outdenting will be handled by later indentation loop
23645 if ( $tok =~ /^(if|unless|while|until)$/
23646 && $next_nonblank_token ne '(' )
23652 # check for inline label following
23653 # /^(redo|last|next|goto)$/
23654 elsif (( $last_nonblank_type eq 'k' )
23655 && ( $is_redo_last_next_goto{$last_nonblank_token} ) )
23661 # something else --
23664 scan_bare_identifier();
23665 if ( $type eq 'w' ) {
23667 if ( $expecting == OPERATOR ) {
23669 # don't complain about possible indirect object
23673 # sub new($) { ... }
23674 # $b = new A::; # calls A::new
23675 # $c = new A; # same thing but suspicious
23676 # This will call A::new but we have a 'new' in
23677 # main:: which looks like a constant.
23679 if ( $last_nonblank_type eq 'C' ) {
23680 if ( $tok !~ /::$/ ) {
23682 Expecting operator after '$last_nonblank_token' but found bare word '$tok'
23683 Maybe indirectet object notation?
23688 error_if_expecting_OPERATOR("bareword");
23692 # mark bare words immediately followed by a paren as
23694 $next_tok = $$rtokens[ $i + 1 ];
23695 if ( $next_tok eq '(' ) {
23699 # underscore after file test operator is file handle
23700 if ( $tok eq '_' && $last_nonblank_type eq 'F' ) {
23704 # patch for SWITCH/CASE if 'case' and 'when are
23705 # not treated as keywords:
23709 && $brace_type[$brace_depth] eq 'switch'
23711 || ( $tok eq 'when'
23712 && $brace_type[$brace_depth] eq 'given' )
23715 $statement_type = $tok; # next '{' is block
23716 $type = 'k'; # for keyword syntax coloring
23719 # patch for SWITCH/CASE if switch and given not keywords
23720 # Switch is not a perl 5 keyword, but we will gamble
23721 # and mark switch followed by paren as a keyword. This
23722 # is only necessary to get html syntax coloring nice,
23723 # and does not commit this as being a switch/case.
23724 if ( $next_nonblank_token eq '('
23725 && ( $tok eq 'switch' || $tok eq 'given' ) )
23727 $type = 'k'; # for keyword syntax coloring
23733 ###############################################################
23734 # section 2: strings of digits
23735 ###############################################################
23736 elsif ( $pre_type eq 'd' ) {
23737 $expecting = operator_expected( $prev_type, $tok, $next_type );
23738 error_if_expecting_OPERATOR("Number")
23739 if ( $expecting == OPERATOR );
23740 my $number = scan_number();
23741 if ( !defined($number) ) {
23743 # shouldn't happen - we should always get a number
23744 warning("non-number beginning with digit--program bug\n");
23745 report_definite_bug();
23749 ###############################################################
23750 # section 3: all other tokens
23751 ###############################################################
23754 last if ( $tok eq '#' );
23755 my $code = $tokenization_code->{$tok};
23758 operator_expected( $prev_type, $tok, $next_type );
23765 # -----------------------------
23766 # end of main tokenization loop
23767 # -----------------------------
23769 if ( $i_tok >= 0 ) {
23770 $routput_token_type->[$i_tok] = $type;
23771 $routput_block_type->[$i_tok] = $block_type;
23772 $routput_container_type->[$i_tok] = $container_type;
23773 $routput_type_sequence->[$i_tok] = $type_sequence;
23774 $routput_indent_flag->[$i_tok] = $indent_flag;
23777 unless ( ( $type eq 'b' ) || ( $type eq '#' ) ) {
23778 $last_last_nonblank_token = $last_nonblank_token;
23779 $last_last_nonblank_type = $last_nonblank_type;
23780 $last_last_nonblank_block_type = $last_nonblank_block_type;
23781 $last_last_nonblank_container_type = $last_nonblank_container_type;
23782 $last_last_nonblank_type_sequence = $last_nonblank_type_sequence;
23783 $last_nonblank_token = $tok;
23784 $last_nonblank_type = $type;
23785 $last_nonblank_block_type = $block_type;
23786 $last_nonblank_container_type = $container_type;
23787 $last_nonblank_type_sequence = $type_sequence;
23788 $last_nonblank_prototype = $prototype;
23791 # reset indentation level if necessary at a sub or package
23792 # in an attempt to recover from a nesting error
23793 if ( $level_in_tokenizer < 0 ) {
23794 if ( $input_line =~ /^\s*(sub|package)\s+(\w+)/ ) {
23795 reset_indentation_level(0);
23796 brace_warning("resetting level to 0 at $1 $2\n");
23800 # all done tokenizing this line ...
23801 # now prepare the final list of tokens and types
23803 my @token_type = (); # stack of output token types
23804 my @block_type = (); # stack of output code block types
23805 my @container_type = (); # stack of output code container types
23806 my @type_sequence = (); # stack of output type sequence numbers
23807 my @tokens = (); # output tokens
23808 my @levels = (); # structural brace levels of output tokens
23809 my @slevels = (); # secondary nesting levels of output tokens
23810 my @nesting_tokens = (); # string of tokens leading to this depth
23811 my @nesting_types = (); # string of token types leading to this depth
23812 my @nesting_blocks = (); # string of block types leading to this depth
23813 my @nesting_lists = (); # string of list types leading to this depth
23814 my @ci_string = (); # string needed to compute continuation indentation
23815 my @container_environment = (); # BLOCK or LIST
23816 my $container_environment = '';
23817 my $im = -1; # previous $i value
23819 my $ci_string_sum = ones_count($ci_string_in_tokenizer);
23821 # Computing Token Indentation
23823 # The final section of the tokenizer forms tokens and also computes
23824 # parameters needed to find indentation. It is much easier to do it
23825 # in the tokenizer than elsewhere. Here is a brief description of how
23826 # indentation is computed. Perl::Tidy computes indentation as the sum
23829 # (1) structural indentation, such as if/else/elsif blocks
23830 # (2) continuation indentation, such as long parameter call lists.
23832 # These are occasionally called primary and secondary indentation.
23834 # Structural indentation is introduced by tokens of type '{', although
23835 # the actual tokens might be '{', '(', or '['. Structural indentation
23836 # is of two types: BLOCK and non-BLOCK. Default structural indentation
23837 # is 4 characters if the standard indentation scheme is used.
23839 # Continuation indentation is introduced whenever a line at BLOCK level
23840 # is broken before its termination. Default continuation indentation
23841 # is 2 characters in the standard indentation scheme.
23843 # Both types of indentation may be nested arbitrarily deep and
23844 # interlaced. The distinction between the two is somewhat arbitrary.
23846 # For each token, we will define two variables which would apply if
23847 # the current statement were broken just before that token, so that
23848 # that token started a new line:
23850 # $level = the structural indentation level,
23851 # $ci_level = the continuation indentation level
23853 # The total indentation will be $level * (4 spaces) + $ci_level * (2 spaces),
23854 # assuming defaults. However, in some special cases it is customary
23855 # to modify $ci_level from this strict value.
23857 # The total structural indentation is easy to compute by adding and
23858 # subtracting 1 from a saved value as types '{' and '}' are seen. The
23859 # running value of this variable is $level_in_tokenizer.
23861 # The total continuation is much more difficult to compute, and requires
23862 # several variables. These veriables are:
23864 # $ci_string_in_tokenizer = a string of 1's and 0's indicating, for
23865 # each indentation level, if there are intervening open secondary
23866 # structures just prior to that level.
23867 # $continuation_string_in_tokenizer = a string of 1's and 0's indicating
23868 # if the last token at that level is "continued", meaning that it
23869 # is not the first token of an expression.
23870 # $nesting_block_string = a string of 1's and 0's indicating, for each
23871 # indentation level, if the level is of type BLOCK or not.
23872 # $nesting_block_flag = the most recent 1 or 0 of $nesting_block_string
23873 # $nesting_list_string = a string of 1's and 0's indicating, for each
23874 # indentation level, if it is is appropriate for list formatting.
23875 # If so, continuation indentation is used to indent long list items.
23876 # $nesting_list_flag = the most recent 1 or 0 of $nesting_list_string
23877 # @{$rslevel_stack} = a stack of total nesting depths at each
23878 # structural indentation level, where "total nesting depth" means
23879 # the nesting depth that would occur if every nesting token -- '{', '[',
23880 # and '(' -- , regardless of context, is used to compute a nesting
23883 #my $nesting_block_flag = ($nesting_block_string =~ /1$/);
23884 #my $nesting_list_flag = ($nesting_list_string =~ /1$/);
23886 my ( $ci_string_i, $level_i, $nesting_block_string_i,
23887 $nesting_list_string_i, $nesting_token_string_i,
23888 $nesting_type_string_i, );
23890 foreach $i ( @{$routput_token_list} )
23891 { # scan the list of pre-tokens indexes
23893 # self-checking for valid token types
23894 my $type = $routput_token_type->[$i];
23895 my $forced_indentation_flag = $routput_indent_flag->[$i];
23897 # See if we should undo the $forced_indentation_flag.
23898 # Forced indentation after 'if', 'unless', 'while' and 'until'
23899 # expressions without trailing parens is optional and doesn't
23900 # always look good. It is usually okay for a trailing logical
23901 # expression, but if the expression is a function call, code block,
23902 # or some kind of list it puts in an unwanted extra indentation
23903 # level which is hard to remove.
23905 # Example where extra indentation looks ok:
23907 # if $det_a < 0 and $det_b > 0
23908 # or $det_a > 0 and $det_b < 0;
23910 # Example where extra indentation is not needed because
23911 # the eval brace also provides indentation:
23912 # print "not " if defined eval {
23913 # reduce { die if $b > 2; $a + $b } 0, 1, 2, 3, 4;
23916 # The following rule works fairly well:
23917 # Undo the flag if the end of this line, or start of the next
23918 # line, is an opening container token or a comma.
23919 # This almost always works, but if not after another pass it will
23921 if ( $forced_indentation_flag && $type eq 'k' ) {
23923 my $ilast = $routput_token_list->[$ixlast];
23924 my $toklast = $routput_token_type->[$ilast];
23925 if ( $toklast eq '#' ) {
23927 $ilast = $routput_token_list->[$ixlast];
23928 $toklast = $routput_token_type->[$ilast];
23930 if ( $toklast eq 'b' ) {
23932 $ilast = $routput_token_list->[$ixlast];
23933 $toklast = $routput_token_type->[$ilast];
23935 if ( $toklast =~ /^[\{,]$/ ) {
23936 $forced_indentation_flag = 0;
23939 ( $toklast, my $i_next ) =
23940 find_next_nonblank_token( $max_token_index, $rtokens,
23941 $max_token_index );
23942 if ( $toklast =~ /^[\{,]$/ ) {
23943 $forced_indentation_flag = 0;
23948 # if we are already in an indented if, see if we should outdent
23949 if ($indented_if_level) {
23951 # don't try to nest trailing if's - shouldn't happen
23952 if ( $type eq 'k' ) {
23953 $forced_indentation_flag = 0;
23956 # check for the normal case - outdenting at next ';'
23957 elsif ( $type eq ';' ) {
23958 if ( $level_in_tokenizer == $indented_if_level ) {
23959 $forced_indentation_flag = -1;
23960 $indented_if_level = 0;
23964 # handle case of missing semicolon
23965 elsif ( $type eq '}' ) {
23966 if ( $level_in_tokenizer == $indented_if_level ) {
23967 $indented_if_level = 0;
23969 # TBD: This could be a subroutine call
23970 $level_in_tokenizer--;
23971 if ( @{$rslevel_stack} > 1 ) {
23972 pop( @{$rslevel_stack} );
23974 if ( length($nesting_block_string) > 1 )
23975 { # true for valid script
23976 chop $nesting_block_string;
23977 chop $nesting_list_string;
23984 my $tok = $$rtokens[$i]; # the token, but ONLY if same as pretoken
23985 $level_i = $level_in_tokenizer;
23987 # This can happen by running perltidy on non-scripts
23988 # although it could also be bug introduced by programming change.
23989 # Perl silently accepts a 032 (^Z) and takes it as the end
23990 if ( !$is_valid_token_type{$type} ) {
23991 my $val = ord($type);
23993 "unexpected character decimal $val ($type) in script\n");
23994 $tokenizer_self->{_in_error} = 1;
23997 # ----------------------------------------------------------------
23998 # TOKEN TYPE PATCHES
23999 # output __END__, __DATA__, and format as type 'k' instead of ';'
24000 # to make html colors correct, etc.
24001 my $fix_type = $type;
24002 if ( $type eq ';' && $tok =~ /\w/ ) { $fix_type = 'k' }
24004 # output anonymous 'sub' as keyword
24005 if ( $type eq 't' && $tok eq 'sub' ) { $fix_type = 'k' }
24007 # -----------------------------------------------------------------
24009 $nesting_token_string_i = $nesting_token_string;
24010 $nesting_type_string_i = $nesting_type_string;
24011 $nesting_block_string_i = $nesting_block_string;
24012 $nesting_list_string_i = $nesting_list_string;
24014 # set primary indentation levels based on structural braces
24015 # Note: these are set so that the leading braces have a HIGHER
24016 # level than their CONTENTS, which is convenient for indentation
24017 # Also, define continuation indentation for each token.
24018 if ( $type eq '{' || $type eq 'L' || $forced_indentation_flag > 0 )
24021 # use environment before updating
24022 $container_environment =
24023 $nesting_block_flag ? 'BLOCK'
24024 : $nesting_list_flag ? 'LIST'
24027 # if the difference between total nesting levels is not 1,
24028 # there are intervening non-structural nesting types between
24029 # this '{' and the previous unclosed '{'
24030 my $intervening_secondary_structure = 0;
24031 if ( @{$rslevel_stack} ) {
24032 $intervening_secondary_structure =
24033 $slevel_in_tokenizer - $rslevel_stack->[-1];
24036 # Continuation Indentation
24038 # Having tried setting continuation indentation both in the formatter and
24039 # in the tokenizer, I can say that setting it in the tokenizer is much,
24040 # much easier. The formatter already has too much to do, and can't
24041 # make decisions on line breaks without knowing what 'ci' will be at
24042 # arbitrary locations.
24044 # But a problem with setting the continuation indentation (ci) here
24045 # in the tokenizer is that we do not know where line breaks will actually
24046 # be. As a result, we don't know if we should propagate continuation
24047 # indentation to higher levels of structure.
24049 # For nesting of only structural indentation, we never need to do this.
24050 # For example, in a long if statement, like this
24052 # if ( !$output_block_type[$i]
24053 # && ($in_statement_continuation) )
24058 # the second line has ci but we do normally give the lines within the BLOCK
24059 # any ci. This would be true if we had blocks nested arbitrarily deeply.
24061 # But consider something like this, where we have created a break after
24062 # an opening paren on line 1, and the paren is not (currently) a
24063 # structural indentation token:
24065 # my $file = $menubar->Menubutton(
24066 # qw/-text File -underline 0 -menuitems/ => [
24068 # Cascade => '~View',
24072 # The second line has ci, so it would seem reasonable to propagate it
24073 # down, giving the third line 1 ci + 1 indentation. This suggests the
24074 # following rule, which is currently used to propagating ci down: if there
24075 # are any non-structural opening parens (or brackets, or braces), before
24076 # an opening structural brace, then ci is propagated down, and otherwise
24077 # not. The variable $intervening_secondary_structure contains this
24078 # information for the current token, and the string
24079 # "$ci_string_in_tokenizer" is a stack of previous values of this
24082 # save the current states
24083 push( @{$rslevel_stack}, 1 + $slevel_in_tokenizer );
24084 $level_in_tokenizer++;
24086 if ($forced_indentation_flag) {
24088 # break BEFORE '?' when there is forced indentation
24089 if ( $type eq '?' ) { $level_i = $level_in_tokenizer; }
24090 if ( $type eq 'k' ) {
24091 $indented_if_level = $level_in_tokenizer;
24095 if ( $routput_block_type->[$i] ) {
24096 $nesting_block_flag = 1;
24097 $nesting_block_string .= '1';
24100 $nesting_block_flag = 0;
24101 $nesting_block_string .= '0';
24104 # we will use continuation indentation within containers
24105 # which are not blocks and not logical expressions
24107 if ( !$routput_block_type->[$i] ) {
24109 # propagate flag down at nested open parens
24110 if ( $routput_container_type->[$i] eq '(' ) {
24111 $bit = 1 if $nesting_list_flag;
24114 # use list continuation if not a logical grouping
24115 # /^(if|elsif|unless|while|and|or|not|&&|!|\|\||for|foreach)$/
24119 $is_logical_container{ $routput_container_type->[$i]
24123 $nesting_list_string .= $bit;
24124 $nesting_list_flag = $bit;
24126 $ci_string_in_tokenizer .=
24127 ( $intervening_secondary_structure != 0 ) ? '1' : '0';
24128 $ci_string_sum = ones_count($ci_string_in_tokenizer);
24129 $continuation_string_in_tokenizer .=
24130 ( $in_statement_continuation > 0 ) ? '1' : '0';
24132 # Sometimes we want to give an opening brace continuation indentation,
24133 # and sometimes not. For code blocks, we don't do it, so that the leading
24134 # '{' gets outdented, like this:
24136 # if ( !$output_block_type[$i]
24137 # && ($in_statement_continuation) )
24140 # For other types, we will give them continuation indentation. For example,
24141 # here is how a list looks with the opening paren indented:
24144 # ( [ "fred", "barney" ], [ "george", "jane", "elroy" ],
24145 # [ "homer", "marge", "bart" ], );
24147 # This looks best when 'ci' is one-half of the indentation (i.e., 2 and 4)
24149 my $total_ci = $ci_string_sum;
24151 !$routput_block_type->[$i] # patch: skip for BLOCK
24152 && ($in_statement_continuation)
24153 && !( $forced_indentation_flag && $type eq ':' )
24156 $total_ci += $in_statement_continuation
24157 unless ( $ci_string_in_tokenizer =~ /1$/ );
24160 $ci_string_i = $total_ci;
24161 $in_statement_continuation = 0;
24164 elsif ($type eq '}'
24166 || $forced_indentation_flag < 0 )
24169 # only a nesting error in the script would prevent popping here
24170 if ( @{$rslevel_stack} > 1 ) { pop( @{$rslevel_stack} ); }
24172 $level_i = --$level_in_tokenizer;
24174 # restore previous level values
24175 if ( length($nesting_block_string) > 1 )
24176 { # true for valid script
24177 chop $nesting_block_string;
24178 $nesting_block_flag = ( $nesting_block_string =~ /1$/ );
24179 chop $nesting_list_string;
24180 $nesting_list_flag = ( $nesting_list_string =~ /1$/ );
24182 chop $ci_string_in_tokenizer;
24183 $ci_string_sum = ones_count($ci_string_in_tokenizer);
24185 $in_statement_continuation =
24186 chop $continuation_string_in_tokenizer;
24188 # zero continuation flag at terminal BLOCK '}' which
24189 # ends a statement.
24190 if ( $routput_block_type->[$i] ) {
24192 # ...These include non-anonymous subs
24193 # note: could be sub ::abc { or sub 'abc
24194 if ( $routput_block_type->[$i] =~ m/^sub\s*/gc ) {
24196 # note: older versions of perl require the /gc modifier
24197 # here or else the \G does not work.
24198 if ( $routput_block_type->[$i] =~ /\G('|::|\w)/gc )
24200 $in_statement_continuation = 0;
24204 # ...and include all block types except user subs with
24205 # block prototypes and these: (sort|grep|map|do|eval)
24206 # /^(\}|\{|BEGIN|END|CHECK|INIT|AUTOLOAD|DESTROY|UNITCHECK|continue|;|if|elsif|else|unless|while|until|for|foreach)$/
24208 $is_zero_continuation_block_type{
24209 $routput_block_type->[$i] } )
24211 $in_statement_continuation = 0;
24214 # ..but these are not terminal types:
24215 # /^(sort|grep|map|do|eval)$/ )
24217 $is_not_zero_continuation_block_type{
24218 $routput_block_type->[$i] } )
24222 # ..and a block introduced by a label
24223 # /^\w+\s*:$/gc ) {
24224 elsif ( $routput_block_type->[$i] =~ /:$/ ) {
24225 $in_statement_continuation = 0;
24228 # user function with block prototype
24230 $in_statement_continuation = 0;
24234 # If we are in a list, then
24235 # we must set continuatoin indentation at the closing
24236 # paren of something like this (paren after $check):
24239 # ( not defined $check )
24241 # or $check eq "new"
24242 # or $check eq "old",
24244 elsif ( $tok eq ')' ) {
24245 $in_statement_continuation = 1
24246 if $routput_container_type->[$i] =~ /^[;,\{\}]$/;
24249 elsif ( $tok eq ';' ) { $in_statement_continuation = 0 }
24252 # use environment after updating
24253 $container_environment =
24254 $nesting_block_flag ? 'BLOCK'
24255 : $nesting_list_flag ? 'LIST'
24257 $ci_string_i = $ci_string_sum + $in_statement_continuation;
24258 $nesting_block_string_i = $nesting_block_string;
24259 $nesting_list_string_i = $nesting_list_string;
24262 # not a structural indentation type..
24265 $container_environment =
24266 $nesting_block_flag ? 'BLOCK'
24267 : $nesting_list_flag ? 'LIST'
24270 # zero the continuation indentation at certain tokens so
24271 # that they will be at the same level as its container. For
24272 # commas, this simplifies the -lp indentation logic, which
24273 # counts commas. For ?: it makes them stand out.
24274 if ($nesting_list_flag) {
24275 if ( $type =~ /^[,\?\:]$/ ) {
24276 $in_statement_continuation = 0;
24280 # be sure binary operators get continuation indentation
24282 $container_environment
24283 && ( $type eq 'k' && $is_binary_keyword{$tok}
24284 || $is_binary_type{$type} )
24287 $in_statement_continuation = 1;
24290 # continuation indentation is sum of any open ci from previous
24291 # levels plus the current level
24292 $ci_string_i = $ci_string_sum + $in_statement_continuation;
24294 # update continuation flag ...
24295 # if this isn't a blank or comment..
24296 if ( $type ne 'b' && $type ne '#' ) {
24298 # and we are in a BLOCK
24299 if ($nesting_block_flag) {
24301 # the next token after a ';' and label starts a new stmt
24302 if ( $type eq ';' || $type eq 'J' ) {
24303 $in_statement_continuation = 0;
24306 # otherwise, we are continuing the current statement
24308 $in_statement_continuation = 1;
24312 # if we are not in a BLOCK..
24315 # do not use continuation indentation if not list
24316 # environment (could be within if/elsif clause)
24317 if ( !$nesting_list_flag ) {
24318 $in_statement_continuation = 0;
24321 # otherwise, the next token after a ',' starts a new term
24322 elsif ( $type eq ',' ) {
24323 $in_statement_continuation = 0;
24326 # otherwise, we are continuing the current term
24328 $in_statement_continuation = 1;
24334 if ( $level_in_tokenizer < 0 ) {
24335 unless ( $tokenizer_self->{_saw_negative_indentation} ) {
24336 $tokenizer_self->{_saw_negative_indentation} = 1;
24337 warning("Starting negative indentation\n");
24341 # set secondary nesting levels based on all continment token types
24342 # Note: these are set so that the nesting depth is the depth
24343 # of the PREVIOUS TOKEN, which is convenient for setting
24344 # the stength of token bonds
24345 my $slevel_i = $slevel_in_tokenizer;
24348 if ( $is_opening_type{$type} ) {
24349 $slevel_in_tokenizer++;
24350 $nesting_token_string .= $tok;
24351 $nesting_type_string .= $type;
24355 elsif ( $is_closing_type{$type} ) {
24356 $slevel_in_tokenizer--;
24357 my $char = chop $nesting_token_string;
24359 if ( $char ne $matching_start_token{$tok} ) {
24360 $nesting_token_string .= $char . $tok;
24361 $nesting_type_string .= $type;
24364 chop $nesting_type_string;
24368 push( @block_type, $routput_block_type->[$i] );
24369 push( @ci_string, $ci_string_i );
24370 push( @container_environment, $container_environment );
24371 push( @container_type, $routput_container_type->[$i] );
24372 push( @levels, $level_i );
24373 push( @nesting_tokens, $nesting_token_string_i );
24374 push( @nesting_types, $nesting_type_string_i );
24375 push( @slevels, $slevel_i );
24376 push( @token_type, $fix_type );
24377 push( @type_sequence, $routput_type_sequence->[$i] );
24378 push( @nesting_blocks, $nesting_block_string );
24379 push( @nesting_lists, $nesting_list_string );
24381 # now form the previous token
24384 $$rtoken_map[$i] - $$rtoken_map[$im]; # how many characters
24388 substr( $input_line, $$rtoken_map[$im], $num ) );
24394 $num = length($input_line) - $$rtoken_map[$im]; # make the last token
24396 push( @tokens, substr( $input_line, $$rtoken_map[$im], $num ) );
24399 $tokenizer_self->{_in_attribute_list} = $in_attribute_list;
24400 $tokenizer_self->{_in_quote} = $in_quote;
24401 $tokenizer_self->{_quote_target} =
24402 $in_quote ? matching_end_token($quote_character) : "";
24403 $tokenizer_self->{_rhere_target_list} = $rhere_target_list;
24405 $line_of_tokens->{_rtoken_type} = \@token_type;
24406 $line_of_tokens->{_rtokens} = \@tokens;
24407 $line_of_tokens->{_rblock_type} = \@block_type;
24408 $line_of_tokens->{_rcontainer_type} = \@container_type;
24409 $line_of_tokens->{_rcontainer_environment} = \@container_environment;
24410 $line_of_tokens->{_rtype_sequence} = \@type_sequence;
24411 $line_of_tokens->{_rlevels} = \@levels;
24412 $line_of_tokens->{_rslevels} = \@slevels;
24413 $line_of_tokens->{_rnesting_tokens} = \@nesting_tokens;
24414 $line_of_tokens->{_rci_levels} = \@ci_string;
24415 $line_of_tokens->{_rnesting_blocks} = \@nesting_blocks;
24419 } # end tokenize_this_line
24421 #########i#############################################################
24422 # Tokenizer routines which assist in identifying token types
24423 #######################################################################
24425 sub operator_expected {
24427 # Many perl symbols have two or more meanings. For example, '<<'
24428 # can be a shift operator or a here-doc operator. The
24429 # interpretation of these symbols depends on the current state of
24430 # the tokenizer, which may either be expecting a term or an
24431 # operator. For this example, a << would be a shift if an operator
24432 # is expected, and a here-doc if a term is expected. This routine
24433 # is called to make this decision for any current token. It returns
24434 # one of three possible values:
24436 # OPERATOR - operator expected (or at least, not a term)
24437 # UNKNOWN - can't tell
24438 # TERM - a term is expected (or at least, not an operator)
24440 # The decision is based on what has been seen so far. This
24441 # information is stored in the "$last_nonblank_type" and
24442 # "$last_nonblank_token" variables. For example, if the
24443 # $last_nonblank_type is '=~', then we are expecting a TERM, whereas
24444 # if $last_nonblank_type is 'n' (numeric), we are expecting an
24447 # If a UNKNOWN is returned, the calling routine must guess. A major
24448 # goal of this tokenizer is to minimize the possiblity of returning
24449 # UNKNOWN, because a wrong guess can spoil the formatting of a
24452 # adding NEW_TOKENS: it is critically important that this routine be
24453 # updated to allow it to determine if an operator or term is to be
24454 # expected after the new token. Doing this simply involves adding
24455 # the new token character to one of the regexes in this routine or
24456 # to one of the hash lists
24457 # that it uses, which are initialized in the BEGIN section.
24458 # USES GLOBAL VARIABLES: $last_nonblank_type, $last_nonblank_token,
24461 my ( $prev_type, $tok, $next_type ) = @_;
24463 my $op_expected = UNKNOWN;
24465 #print "tok=$tok last type=$last_nonblank_type last tok=$last_nonblank_token\n";
24467 # Note: function prototype is available for token type 'U' for future
24468 # program development. It contains the leading and trailing parens,
24469 # and no blanks. It might be used to eliminate token type 'C', for
24470 # example (prototype = '()'). Thus:
24471 # if ($last_nonblank_type eq 'U') {
24472 # print "previous token=$last_nonblank_token type=$last_nonblank_type prototype=$last_nonblank_prototype\n";
24475 # A possible filehandle (or object) requires some care...
24476 if ( $last_nonblank_type eq 'Z' ) {
24479 if ( $last_nonblank_token =~ /^[A-Za-z_]/ ) {
24480 $op_expected = UNKNOWN;
24483 # For possible file handle like "$a", Perl uses weird parsing rules.
24485 # print $a/2,"/hi"; - division
24486 # print $a / 2,"/hi"; - division
24487 # print $a/ 2,"/hi"; - division
24488 # print $a /2,"/hi"; - pattern (and error)!
24489 elsif ( ( $prev_type eq 'b' ) && ( $next_type ne 'b' ) ) {
24490 $op_expected = TERM;
24493 # Note when an operation is being done where a
24494 # filehandle might be expected, since a change in whitespace
24495 # could change the interpretation of the statement.
24497 if ( $tok =~ /^([x\/\+\-\*\%\&\.\?\<]|\>\>)$/ ) {
24498 complain("operator in print statement not recommended\n");
24499 $op_expected = OPERATOR;
24504 # handle something after 'do' and 'eval'
24505 elsif ( $is_block_operator{$last_nonblank_token} ) {
24507 # something like $a = eval "expression";
24509 if ( $last_nonblank_type eq 'k' ) {
24510 $op_expected = TERM; # expression or list mode following keyword
24513 # something like $a = do { BLOCK } / 2;
24516 $op_expected = OPERATOR; # block mode following }
24520 # handle bare word..
24521 elsif ( $last_nonblank_type eq 'w' ) {
24523 # unfortunately, we can't tell what type of token to expect next
24524 # after most bare words
24525 $op_expected = UNKNOWN;
24528 # operator, but not term possible after these types
24529 # Note: moved ')' from type to token because parens in list context
24530 # get marked as '{' '}' now. This is a minor glitch in the following:
24531 # my %opts = (ref $_[0] eq 'HASH') ? %{shift()} : ();
24533 elsif (( $last_nonblank_type =~ /^[\]RnviQh]$/ )
24534 || ( $last_nonblank_token =~ /^(\)|\$|\-\>)/ ) )
24536 $op_expected = OPERATOR;
24538 # in a 'use' statement, numbers and v-strings are not true
24539 # numbers, so to avoid incorrect error messages, we will
24540 # mark them as unknown for now (use.t)
24541 # TODO: it would be much nicer to create a new token V for VERSION
24542 # number in a use statement. Then this could be a check on type V
24543 # and related patches which change $statement_type for '=>'
24544 # and ',' could be removed. Further, it would clean things up to
24545 # scan the 'use' statement with a separate subroutine.
24546 if ( ( $statement_type eq 'use' )
24547 && ( $last_nonblank_type =~ /^[nv]$/ ) )
24549 $op_expected = UNKNOWN;
24553 # no operator after many keywords, such as "die", "warn", etc
24554 elsif ( $expecting_term_token{$last_nonblank_token} ) {
24556 # patch for dor.t (defined or).
24557 # perl functions which may be unary operators
24558 # TODO: This list is incomplete, and these should be put
24561 && $next_type eq '/'
24562 && $last_nonblank_type eq 'k'
24563 && $last_nonblank_token =~ /^eof|undef|shift|pop$/ )
24565 $op_expected = OPERATOR;
24568 $op_expected = TERM;
24572 # no operator after things like + - ** (i.e., other operators)
24573 elsif ( $expecting_term_types{$last_nonblank_type} ) {
24574 $op_expected = TERM;
24577 # a few operators, like "time", have an empty prototype () and so
24578 # take no parameters but produce a value to operate on
24579 elsif ( $expecting_operator_token{$last_nonblank_token} ) {
24580 $op_expected = OPERATOR;
24583 # post-increment and decrement produce values to be operated on
24584 elsif ( $expecting_operator_types{$last_nonblank_type} ) {
24585 $op_expected = OPERATOR;
24588 # no value to operate on after sub block
24589 elsif ( $last_nonblank_token =~ /^sub\s/ ) { $op_expected = TERM; }
24591 # a right brace here indicates the end of a simple block.
24592 # all non-structural right braces have type 'R'
24593 # all braces associated with block operator keywords have been given those
24594 # keywords as "last_nonblank_token" and caught above.
24595 # (This statement is order dependent, and must come after checking
24596 # $last_nonblank_token).
24597 elsif ( $last_nonblank_type eq '}' ) {
24599 # patch for dor.t (defined or).
24601 && $next_type eq '/'
24602 && $last_nonblank_token eq ']' )
24604 $op_expected = OPERATOR;
24607 $op_expected = TERM;
24611 # something else..what did I forget?
24614 # collecting diagnostics on unknown operator types..see what was missed
24615 $op_expected = UNKNOWN;
24617 "OP: unknown after type=$last_nonblank_type token=$last_nonblank_token\n"
24621 TOKENIZER_DEBUG_FLAG_EXPECT && do {
24623 "EXPECT: returns $op_expected for last type $last_nonblank_type token $last_nonblank_token\n";
24625 return $op_expected;
24628 sub new_statement_ok {
24630 # return true if the current token can start a new statement
24631 # USES GLOBAL VARIABLES: $last_nonblank_type
24633 return label_ok() # a label would be ok here
24635 || $last_nonblank_type eq 'J'; # or we follow a label
24641 # Decide if a bare word followed by a colon here is a label
24642 # USES GLOBAL VARIABLES: $last_nonblank_token, $last_nonblank_type,
24643 # $brace_depth, @brace_type
24645 # if it follows an opening or closing code block curly brace..
24646 if ( ( $last_nonblank_token eq '{' || $last_nonblank_token eq '}' )
24647 && $last_nonblank_type eq $last_nonblank_token )
24650 # it is a label if and only if the curly encloses a code block
24651 return $brace_type[$brace_depth];
24654 # otherwise, it is a label if and only if it follows a ';'
24657 return ( $last_nonblank_type eq ';' );
24661 sub code_block_type {
24663 # Decide if this is a block of code, and its type.
24664 # Must be called only when $type = $token = '{'
24665 # The problem is to distinguish between the start of a block of code
24666 # and the start of an anonymous hash reference
24667 # Returns "" if not code block, otherwise returns 'last_nonblank_token'
24668 # to indicate the type of code block. (For example, 'last_nonblank_token'
24669 # might be 'if' for an if block, 'else' for an else block, etc).
24670 # USES GLOBAL VARIABLES: $last_nonblank_token, $last_nonblank_type,
24671 # $last_nonblank_block_type, $brace_depth, @brace_type
24673 # handle case of multiple '{'s
24675 # print "BLOCK_TYPE EXAMINING: type=$last_nonblank_type tok=$last_nonblank_token\n";
24677 my ( $i, $rtokens, $rtoken_type, $max_token_index ) = @_;
24678 if ( $last_nonblank_token eq '{'
24679 && $last_nonblank_type eq $last_nonblank_token )
24682 # opening brace where a statement may appear is probably
24683 # a code block but might be and anonymous hash reference
24684 if ( $brace_type[$brace_depth] ) {
24685 return decide_if_code_block( $i, $rtokens, $rtoken_type,
24686 $max_token_index );
24689 # cannot start a code block within an anonymous hash
24695 elsif ( $last_nonblank_token eq ';' ) {
24697 # an opening brace where a statement may appear is probably
24698 # a code block but might be and anonymous hash reference
24699 return decide_if_code_block( $i, $rtokens, $rtoken_type,
24700 $max_token_index );
24703 # handle case of '}{'
24704 elsif ($last_nonblank_token eq '}'
24705 && $last_nonblank_type eq $last_nonblank_token )
24708 # a } { situation ...
24709 # could be hash reference after code block..(blktype1.t)
24710 if ($last_nonblank_block_type) {
24711 return decide_if_code_block( $i, $rtokens, $rtoken_type,
24712 $max_token_index );
24715 # must be a block if it follows a closing hash reference
24717 return $last_nonblank_token;
24721 # NOTE: braces after type characters start code blocks, but for
24722 # simplicity these are not identified as such. See also
24723 # sub is_non_structural_brace.
24724 # elsif ( $last_nonblank_type eq 't' ) {
24725 # return $last_nonblank_token;
24728 # brace after label:
24729 elsif ( $last_nonblank_type eq 'J' ) {
24730 return $last_nonblank_token;
24733 # otherwise, look at previous token. This must be a code block if
24734 # it follows any of these:
24735 # /^(BEGIN|END|CHECK|INIT|AUTOLOAD|DESTROY|UNITCHECK|continue|if|elsif|else|unless|do|while|until|eval|for|foreach|map|grep|sort)$/
24736 elsif ( $is_code_block_token{$last_nonblank_token} ) {
24737 return $last_nonblank_token;
24740 # or a sub definition
24741 elsif ( ( $last_nonblank_type eq 'i' || $last_nonblank_type eq 't' )
24742 && $last_nonblank_token =~ /^sub\b/ )
24744 return $last_nonblank_token;
24747 # user-defined subs with block parameters (like grep/map/eval)
24748 elsif ( $last_nonblank_type eq 'G' ) {
24749 return $last_nonblank_token;
24753 elsif ( $last_nonblank_type eq 'w' ) {
24754 return decide_if_code_block( $i, $rtokens, $rtoken_type,
24755 $max_token_index );
24758 # anything else must be anonymous hash reference
24764 sub decide_if_code_block {
24766 # USES GLOBAL VARIABLES: $last_nonblank_token
24767 my ( $i, $rtokens, $rtoken_type, $max_token_index ) = @_;
24768 my ( $next_nonblank_token, $i_next ) =
24769 find_next_nonblank_token( $i, $rtokens, $max_token_index );
24771 # we are at a '{' where a statement may appear.
24772 # We must decide if this brace starts an anonymous hash or a code
24774 # return "" if anonymous hash, and $last_nonblank_token otherwise
24776 # initialize to be code BLOCK
24777 my $code_block_type = $last_nonblank_token;
24779 # Check for the common case of an empty anonymous hash reference:
24780 # Maybe something like sub { { } }
24781 if ( $next_nonblank_token eq '}' ) {
24782 $code_block_type = "";
24787 # To guess if this '{' is an anonymous hash reference, look ahead
24788 # and test as follows:
24790 # it is a hash reference if next come:
24791 # - a string or digit followed by a comma or =>
24792 # - bareword followed by =>
24793 # otherwise it is a code block
24795 # Examples of anonymous hash ref:
24799 # Examples of code blocks:
24800 # {1; print "hello\n", 1;}
24803 # We are only going to look ahead one more (nonblank/comment) line.
24804 # Strange formatting could cause a bad guess, but that's unlikely.
24805 my @pre_types = @$rtoken_type[ $i + 1 .. $max_token_index ];
24806 my @pre_tokens = @$rtokens[ $i + 1 .. $max_token_index ];
24807 my ( $rpre_tokens, $rpre_types ) =
24808 peek_ahead_for_n_nonblank_pre_tokens(20); # 20 is arbitrary but
24809 # generous, and prevents
24811 # time in mangled files
24812 if ( defined($rpre_types) && @$rpre_types ) {
24813 push @pre_types, @$rpre_types;
24814 push @pre_tokens, @$rpre_tokens;
24817 # put a sentinal token to simplify stopping the search
24818 push @pre_types, '}';
24821 $jbeg = 1 if $pre_types[0] eq 'b';
24823 # first look for one of these
24825 # - bareword with leading -
24829 if ( $pre_types[$j] =~ /^[\'\"]/ ) {
24831 # find the closing quote; don't worry about escapes
24832 my $quote_mark = $pre_types[$j];
24833 for ( my $k = $j + 1 ; $k < $#pre_types ; $k++ ) {
24834 if ( $pre_types[$k] eq $quote_mark ) {
24836 my $next = $pre_types[$j];
24841 elsif ( $pre_types[$j] eq 'd' ) {
24844 elsif ( $pre_types[$j] eq 'w' ) {
24845 unless ( $is_keyword{ $pre_tokens[$j] } ) {
24849 elsif ( $pre_types[$j] eq '-' && $pre_types[ ++$j ] eq 'w' ) {
24852 if ( $j > $jbeg ) {
24854 $j++ if $pre_types[$j] eq 'b';
24856 # it's a hash ref if a comma or => follow next
24857 if ( $pre_types[$j] eq ','
24858 || ( $pre_types[$j] eq '=' && $pre_types[ ++$j ] eq '>' ) )
24860 $code_block_type = "";
24865 return $code_block_type;
24870 # report unexpected token type and show where it is
24871 # USES GLOBAL VARIABLES: $tokenizer_self
24872 my ( $found, $expecting, $i_tok, $last_nonblank_i, $rpretoken_map,
24873 $rpretoken_type, $input_line )
24876 if ( ++$tokenizer_self->{_unexpected_error_count} <= MAX_NAG_MESSAGES ) {
24877 my $msg = "found $found where $expecting expected";
24878 my $pos = $$rpretoken_map[$i_tok];
24879 interrupt_logfile();
24880 my $input_line_number = $tokenizer_self->{_last_line_number};
24881 my ( $offset, $numbered_line, $underline ) =
24882 make_numbered_line( $input_line_number, $input_line, $pos );
24883 $underline = write_on_underline( $underline, $pos - $offset, '^' );
24886 if ( ( $i_tok > 0 ) && ( $last_nonblank_i >= 0 ) ) {
24887 my $pos_prev = $$rpretoken_map[$last_nonblank_i];
24889 if ( $$rpretoken_type[ $i_tok - 1 ] eq 'b' ) {
24890 $num = $$rpretoken_map[ $i_tok - 1 ] - $pos_prev;
24893 $num = $pos - $pos_prev;
24895 if ( $num > 40 ) { $num = 40; $pos_prev = $pos - 40; }
24898 write_on_underline( $underline, $pos_prev - $offset, '-' x $num );
24899 $trailer = " (previous token underlined)";
24901 warning( $numbered_line . "\n" );
24902 warning( $underline . "\n" );
24903 warning( $msg . $trailer . "\n" );
24908 sub is_non_structural_brace {
24910 # Decide if a brace or bracket is structural or non-structural
24911 # by looking at the previous token and type
24912 # USES GLOBAL VARIABLES: $last_nonblank_type, $last_nonblank_token
24914 # EXPERIMENTAL: Mark slices as structural; idea was to improve formatting.
24915 # Tentatively deactivated because it caused the wrong operator expectation
24917 # $user = @vars[1] / 100;
24918 # Must update sub operator_expected before re-implementing.
24919 # if ( $last_nonblank_type eq 'i' && $last_nonblank_token =~ /^@/ ) {
24923 # NOTE: braces after type characters start code blocks, but for
24924 # simplicity these are not identified as such. See also
24925 # sub code_block_type
24926 # if ($last_nonblank_type eq 't') {return 0}
24928 # otherwise, it is non-structural if it is decorated
24929 # by type information.
24930 # For example, the '{' here is non-structural: ${xxx}
24932 $last_nonblank_token =~ /^([\$\@\*\&\%\)]|->|::)/
24934 # or if we follow a hash or array closing curly brace or bracket
24935 # For example, the second '{' in this is non-structural: $a{'x'}{'y'}
24936 # because the first '}' would have been given type 'R'
24937 || $last_nonblank_type =~ /^([R\]])$/
24941 #########i#############################################################
24942 # Tokenizer routines for tracking container nesting depths
24943 #######################################################################
24945 # The following routines keep track of nesting depths of the nesting
24946 # types, ( [ { and ?. This is necessary for determining the indentation
24947 # level, and also for debugging programs. Not only do they keep track of
24948 # nesting depths of the individual brace types, but they check that each
24949 # of the other brace types is balanced within matching pairs. For
24950 # example, if the program sees this sequence:
24954 # then it can determine that there is an extra left paren somewhere
24955 # between the { and the }. And so on with every other possible
24956 # combination of outer and inner brace types. For another
24961 # which has an extra ] within the parens.
24963 # The brace types have indexes 0 .. 3 which are indexes into
24966 # The pair ? : are treated as just another nesting type, with ? acting
24967 # as the opening brace and : acting as the closing brace.
24971 # $depth_array[$a][$b][ $current_depth[$a] ] = $current_depth[$b];
24973 # saves the nesting depth of brace type $b (where $b is either of the other
24974 # nesting types) when brace type $a enters a new depth. When this depth
24975 # decreases, a check is made that the current depth of brace types $b is
24976 # unchanged, or otherwise there must have been an error. This can
24977 # be very useful for localizing errors, particularly when perl runs to
24978 # the end of a large file (such as this one) and announces that there
24979 # is a problem somewhere.
24981 # A numerical sequence number is maintained for every nesting type,
24982 # so that each matching pair can be uniquely identified in a simple
24985 sub increase_nesting_depth {
24986 my ( $aa, $pos ) = @_;
24988 # USES GLOBAL VARIABLES: $tokenizer_self, @current_depth,
24989 # @current_sequence_number, @depth_array, @starting_line_of_current_depth
24991 $current_depth[$aa]++;
24993 $total_depth[$aa][ $current_depth[$aa] ] = $total_depth;
24994 my $input_line_number = $tokenizer_self->{_last_line_number};
24995 my $input_line = $tokenizer_self->{_line_text};
24997 # Sequence numbers increment by number of items. This keeps
24998 # a unique set of numbers but still allows the relative location
24999 # of any type to be determined.
25000 $nesting_sequence_number[$aa] += scalar(@closing_brace_names);
25001 my $seqno = $nesting_sequence_number[$aa];
25002 $current_sequence_number[$aa][ $current_depth[$aa] ] = $seqno;
25004 $starting_line_of_current_depth[$aa][ $current_depth[$aa] ] =
25005 [ $input_line_number, $input_line, $pos ];
25007 for $bb ( 0 .. $#closing_brace_names ) {
25008 next if ( $bb == $aa );
25009 $depth_array[$aa][$bb][ $current_depth[$aa] ] = $current_depth[$bb];
25012 # set a flag for indenting a nested ternary statement
25014 if ( $aa == QUESTION_COLON ) {
25015 $nested_ternary_flag[ $current_depth[$aa] ] = 0;
25016 if ( $current_depth[$aa] > 1 ) {
25017 if ( $nested_ternary_flag[ $current_depth[$aa] - 1 ] == 0 ) {
25018 my $pdepth = $total_depth[$aa][ $current_depth[$aa] - 1 ];
25019 if ( $pdepth == $total_depth - 1 ) {
25021 $nested_ternary_flag[ $current_depth[$aa] - 1 ] = -1;
25026 return ( $seqno, $indent );
25029 sub decrease_nesting_depth {
25031 my ( $aa, $pos ) = @_;
25033 # USES GLOBAL VARIABLES: $tokenizer_self, @current_depth,
25034 # @current_sequence_number, @depth_array, @starting_line_of_current_depth
25037 my $input_line_number = $tokenizer_self->{_last_line_number};
25038 my $input_line = $tokenizer_self->{_line_text};
25042 if ( $current_depth[$aa] > 0 ) {
25044 # set a flag for un-indenting after seeing a nested ternary statement
25045 $seqno = $current_sequence_number[$aa][ $current_depth[$aa] ];
25046 if ( $aa == QUESTION_COLON ) {
25047 $outdent = $nested_ternary_flag[ $current_depth[$aa] ];
25050 # check that any brace types $bb contained within are balanced
25051 for $bb ( 0 .. $#closing_brace_names ) {
25052 next if ( $bb == $aa );
25054 unless ( $depth_array[$aa][$bb][ $current_depth[$aa] ] ==
25055 $current_depth[$bb] )
25058 $current_depth[$bb] -
25059 $depth_array[$aa][$bb][ $current_depth[$aa] ];
25061 # don't whine too many times
25062 my $saw_brace_error = get_saw_brace_error();
25064 $saw_brace_error <= MAX_NAG_MESSAGES
25066 # if too many closing types have occured, we probably
25067 # already caught this error
25068 && ( ( $diff > 0 ) || ( $saw_brace_error <= 0 ) )
25071 interrupt_logfile();
25073 $starting_line_of_current_depth[$aa]
25074 [ $current_depth[$aa] ];
25076 my $rel = [ $input_line_number, $input_line, $pos ];
25080 if ( $diff == 1 || $diff == -1 ) {
25088 ? $opening_brace_names[$bb]
25089 : $closing_brace_names[$bb];
25090 write_error_indicator_pair( @$rsl, '^' );
25092 Found $diff extra $bname$ess between $opening_brace_names[$aa] on line $sl and $closing_brace_names[$aa] on line $el
25097 $starting_line_of_current_depth[$bb]
25098 [ $current_depth[$bb] ];
25101 " The most recent un-matched $bname is on line $ml\n";
25102 write_error_indicator_pair( @$rml, '^' );
25104 write_error_indicator_pair( @$rel, '^' );
25108 increment_brace_error();
25111 $current_depth[$aa]--;
25115 my $saw_brace_error = get_saw_brace_error();
25116 if ( $saw_brace_error <= MAX_NAG_MESSAGES ) {
25118 There is no previous $opening_brace_names[$aa] to match a $closing_brace_names[$aa] on line $input_line_number
25120 indicate_error( $msg, $input_line_number, $input_line, $pos, '^' );
25122 increment_brace_error();
25124 return ( $seqno, $outdent );
25127 sub check_final_nesting_depths {
25130 # USES GLOBAL VARIABLES: @current_depth, @starting_line_of_current_depth
25132 for $aa ( 0 .. $#closing_brace_names ) {
25134 if ( $current_depth[$aa] ) {
25136 $starting_line_of_current_depth[$aa][ $current_depth[$aa] ];
25139 Final nesting depth of $opening_brace_names[$aa]s is $current_depth[$aa]
25140 The most recent un-matched $opening_brace_names[$aa] is on line $sl
25142 indicate_error( $msg, @$rsl, '^' );
25143 increment_brace_error();
25148 #########i#############################################################
25149 # Tokenizer routines for looking ahead in input stream
25150 #######################################################################
25152 sub peek_ahead_for_n_nonblank_pre_tokens {
25154 # returns next n pretokens if they exist
25155 # returns undef's if hits eof without seeing any pretokens
25156 # USES GLOBAL VARIABLES: $tokenizer_self
25157 my $max_pretokens = shift;
25160 my ( $rpre_tokens, $rmap, $rpre_types );
25162 while ( $line = $tokenizer_self->{_line_buffer_object}->peek_ahead( $i++ ) )
25164 $line =~ s/^\s*//; # trim leading blanks
25165 next if ( length($line) <= 0 ); # skip blank
25166 next if ( $line =~ /^#/ ); # skip comment
25167 ( $rpre_tokens, $rmap, $rpre_types ) =
25168 pre_tokenize( $line, $max_pretokens );
25171 return ( $rpre_tokens, $rpre_types );
25174 # look ahead for next non-blank, non-comment line of code
25175 sub peek_ahead_for_nonblank_token {
25177 # USES GLOBAL VARIABLES: $tokenizer_self
25178 my ( $rtokens, $max_token_index ) = @_;
25182 while ( $line = $tokenizer_self->{_line_buffer_object}->peek_ahead( $i++ ) )
25184 $line =~ s/^\s*//; # trim leading blanks
25185 next if ( length($line) <= 0 ); # skip blank
25186 next if ( $line =~ /^#/ ); # skip comment
25187 my ( $rtok, $rmap, $rtype ) =
25188 pre_tokenize( $line, 2 ); # only need 2 pre-tokens
25189 my $j = $max_token_index + 1;
25192 foreach $tok (@$rtok) {
25193 last if ( $tok =~ "\n" );
25194 $$rtokens[ ++$j ] = $tok;
25201 #########i#############################################################
25202 # Tokenizer guessing routines for ambiguous situations
25203 #######################################################################
25205 sub guess_if_pattern_or_conditional {
25207 # this routine is called when we have encountered a ? following an
25208 # unknown bareword, and we must decide if it starts a pattern or not
25209 # input parameters:
25210 # $i - token index of the ? starting possible pattern
25211 # output parameters:
25212 # $is_pattern = 0 if probably not pattern, =1 if probably a pattern
25213 # msg = a warning or diagnostic message
25214 # USES GLOBAL VARIABLES: $last_nonblank_token
25215 my ( $i, $rtokens, $rtoken_map, $max_token_index ) = @_;
25216 my $is_pattern = 0;
25217 my $msg = "guessing that ? after $last_nonblank_token starts a ";
25219 if ( $i >= $max_token_index ) {
25220 $msg .= "conditional (no end to pattern found on the line)\n";
25225 my $next_token = $$rtokens[$i]; # first token after ?
25227 # look for a possible ending ? on this line..
25229 my $quote_depth = 0;
25230 my $quote_character = '';
25234 $i, $in_quote, $quote_character, $quote_pos, $quote_depth,
25237 = follow_quoted_string( $ibeg, $in_quote, $rtokens, $quote_character,
25238 $quote_pos, $quote_depth, $max_token_index );
25242 # we didn't find an ending ? on this line,
25243 # so we bias towards conditional
25245 $msg .= "conditional (no ending ? on this line)\n";
25247 # we found an ending ?, so we bias towards a pattern
25251 if ( pattern_expected( $i, $rtokens, $max_token_index ) >= 0 ) {
25253 $msg .= "pattern (found ending ? and pattern expected)\n";
25256 $msg .= "pattern (uncertain, but found ending ?)\n";
25260 return ( $is_pattern, $msg );
25263 sub guess_if_pattern_or_division {
25265 # this routine is called when we have encountered a / following an
25266 # unknown bareword, and we must decide if it starts a pattern or is a
25268 # input parameters:
25269 # $i - token index of the / starting possible pattern
25270 # output parameters:
25271 # $is_pattern = 0 if probably division, =1 if probably a pattern
25272 # msg = a warning or diagnostic message
25273 # USES GLOBAL VARIABLES: $last_nonblank_token
25274 my ( $i, $rtokens, $rtoken_map, $max_token_index ) = @_;
25275 my $is_pattern = 0;
25276 my $msg = "guessing that / after $last_nonblank_token starts a ";
25278 if ( $i >= $max_token_index ) {
25279 "division (no end to pattern found on the line)\n";
25283 my $divide_expected =
25284 numerator_expected( $i, $rtokens, $max_token_index );
25286 my $next_token = $$rtokens[$i]; # first token after slash
25288 # look for a possible ending / on this line..
25290 my $quote_depth = 0;
25291 my $quote_character = '';
25295 $i, $in_quote, $quote_character, $quote_pos, $quote_depth,
25298 = follow_quoted_string( $ibeg, $in_quote, $rtokens, $quote_character,
25299 $quote_pos, $quote_depth, $max_token_index );
25303 # we didn't find an ending / on this line,
25304 # so we bias towards division
25305 if ( $divide_expected >= 0 ) {
25307 $msg .= "division (no ending / on this line)\n";
25310 $msg = "multi-line pattern (division not possible)\n";
25316 # we found an ending /, so we bias towards a pattern
25319 if ( pattern_expected( $i, $rtokens, $max_token_index ) >= 0 ) {
25321 if ( $divide_expected >= 0 ) {
25323 if ( $i - $ibeg > 60 ) {
25324 $msg .= "division (matching / too distant)\n";
25328 $msg .= "pattern (but division possible too)\n";
25334 $msg .= "pattern (division not possible)\n";
25339 if ( $divide_expected >= 0 ) {
25341 $msg .= "division (pattern not possible)\n";
25346 "pattern (uncertain, but division would not work here)\n";
25351 return ( $is_pattern, $msg );
25354 # try to resolve here-doc vs. shift by looking ahead for
25355 # non-code or the end token (currently only looks for end token)
25356 # returns 1 if it is probably a here doc, 0 if not
25357 sub guess_if_here_doc {
25359 # This is how many lines we will search for a target as part of the
25360 # guessing strategy. It is a constant because there is probably
25361 # little reason to change it.
25362 # USES GLOBAL VARIABLES: $tokenizer_self, $current_package
25364 use constant HERE_DOC_WINDOW => 40;
25366 my $next_token = shift;
25367 my $here_doc_expected = 0;
25370 my $msg = "checking <<";
25372 while ( $line = $tokenizer_self->{_line_buffer_object}->peek_ahead( $k++ ) )
25376 if ( $line =~ /^$next_token$/ ) {
25377 $msg .= " -- found target $next_token ahead $k lines\n";
25378 $here_doc_expected = 1; # got it
25381 last if ( $k >= HERE_DOC_WINDOW );
25384 unless ($here_doc_expected) {
25386 if ( !defined($line) ) {
25387 $here_doc_expected = -1; # hit eof without seeing target
25388 $msg .= " -- must be shift; target $next_token not in file\n";
25391 else { # still unsure..taking a wild guess
25393 if ( !$is_constant{$current_package}{$next_token} ) {
25394 $here_doc_expected = 1;
25396 " -- guessing it's a here-doc ($next_token not a constant)\n";
25400 " -- guessing it's a shift ($next_token is a constant)\n";
25404 write_logfile_entry($msg);
25405 return $here_doc_expected;
25408 #########i#############################################################
25409 # Tokenizer Routines for scanning identifiers and related items
25410 #######################################################################
25412 sub scan_bare_identifier_do {
25414 # this routine is called to scan a token starting with an alphanumeric
25415 # variable or package separator, :: or '.
25416 # USES GLOBAL VARIABLES: $current_package, $last_nonblank_token,
25417 # $last_nonblank_type,@paren_type, $paren_depth
25419 my ( $input_line, $i, $tok, $type, $prototype, $rtoken_map,
25423 my $package = undef;
25427 # we have to back up one pretoken at a :: since each : is one pretoken
25428 if ( $tok eq '::' ) { $i_beg-- }
25429 if ( $tok eq '->' ) { $i_beg-- }
25430 my $pos_beg = $$rtoken_map[$i_beg];
25431 pos($input_line) = $pos_beg;
25438 if ( $input_line =~ m/\G\s*((?:\w*(?:'|::)))*(?:(?:->)?(\w+))?/gc ) {
25440 my $pos = pos($input_line);
25441 my $numc = $pos - $pos_beg;
25442 $tok = substr( $input_line, $pos_beg, $numc );
25444 # type 'w' includes anything without leading type info
25445 # ($,%,@,*) including something like abc::def::ghi
25449 if ( defined($2) ) { $sub_name = $2; }
25450 if ( defined($1) ) {
25453 # patch: don't allow isolated package name which just ends
25454 # in the old style package separator (single quote). Example:
25456 if ( !($sub_name) && substr( $package, -1, 1 ) eq '\'' ) {
25460 $package =~ s/\'/::/g;
25461 if ( $package =~ /^\:/ ) { $package = 'main' . $package }
25462 $package =~ s/::$//;
25465 $package = $current_package;
25467 if ( $is_keyword{$tok} ) {
25472 # if it is a bareword..
25473 if ( $type eq 'w' ) {
25475 # check for v-string with leading 'v' type character
25476 # (This seems to have presidence over filehandle, type 'Y')
25477 if ( $tok =~ /^v\d[_\d]*$/ ) {
25479 # we only have the first part - something like 'v101' -
25481 if ( $input_line =~ m/\G(\.\d[_\d]*)+/gc ) {
25482 $pos = pos($input_line);
25483 $numc = $pos - $pos_beg;
25484 $tok = substr( $input_line, $pos_beg, $numc );
25488 # warn if this version can't handle v-strings
25489 report_v_string($tok);
25492 elsif ( $is_constant{$package}{$sub_name} ) {
25496 # bareword after sort has implied empty prototype; for example:
25497 # @sorted = sort numerically ( 53, 29, 11, 32, 7 );
25498 # This has priority over whatever the user has specified.
25499 elsif ($last_nonblank_token eq 'sort'
25500 && $last_nonblank_type eq 'k' )
25505 # Note: strangely, perl does not seem to really let you create
25506 # functions which act like eval and do, in the sense that eval
25507 # and do may have operators following the final }, but any operators
25508 # that you create with prototype (&) apparently do not allow
25509 # trailing operators, only terms. This seems strange.
25510 # If this ever changes, here is the update
25511 # to make perltidy behave accordingly:
25513 # elsif ( $is_block_function{$package}{$tok} ) {
25514 # $tok='eval'; # patch to do braces like eval - doesn't work
25517 # FIXME: This could become a separate type to allow for different
25519 elsif ( $is_block_function{$package}{$sub_name} ) {
25523 elsif ( $is_block_list_function{$package}{$sub_name} ) {
25526 elsif ( $is_user_function{$package}{$sub_name} ) {
25528 $prototype = $user_function_prototype{$package}{$sub_name};
25531 # check for indirect object
25534 # added 2001-03-27: must not be followed immediately by '('
25536 ( $input_line !~ m/\G\(/gc )
25541 # preceded by keyword like 'print', 'printf' and friends
25542 $is_indirect_object_taker{$last_nonblank_token}
25544 # or preceded by something like 'print(' or 'printf('
25546 ( $last_nonblank_token eq '(' )
25547 && $is_indirect_object_taker{ $paren_type[$paren_depth]
25555 # may not be indirect object unless followed by a space
25556 if ( $input_line =~ m/\G\s+/gc ) {
25560 # Perl's indirect object notation is a very bad
25561 # thing and can cause subtle bugs, especially for
25562 # beginning programmers. And I haven't even been
25563 # able to figure out a sane warning scheme which
25564 # doesn't get in the way of good scripts.
25566 # Complain if a filehandle has any lower case
25567 # letters. This is suggested good practice.
25568 # Use 'sub_name' because something like
25569 # main::MYHANDLE is ok for filehandle
25570 if ( $sub_name =~ /[a-z]/ ) {
25572 # could be bug caused by older perltidy if
25574 if ( $input_line =~ m/\G\s*\(/gc ) {
25576 "Caution: unknown word '$tok' in indirect object slot\n"
25582 # bareword not followed by a space -- may not be filehandle
25583 # (may be function call defined in a 'use' statement)
25590 # Now we must convert back from character position
25591 # to pre_token index.
25592 # I don't think an error flag can occur here ..but who knows
25595 inverse_pretoken_map( $i, $pos, $rtoken_map, $max_token_index );
25597 warning("scan_bare_identifier: Possibly invalid tokenization\n");
25601 # no match but line not blank - could be syntax error
25602 # perl will take '::' alone without complaint
25606 # change this warning to log message if it becomes annoying
25607 warning("didn't find identifier after leading ::\n");
25609 return ( $i, $tok, $type, $prototype );
25614 # This is the new scanner and will eventually replace scan_identifier.
25615 # Only type 'sub' and 'package' are implemented.
25616 # Token types $ * % @ & -> are not yet implemented.
25618 # Scan identifier following a type token.
25619 # The type of call depends on $id_scan_state: $id_scan_state = ''
25620 # for starting call, in which case $tok must be the token defining
25623 # If the type token is the last nonblank token on the line, a value
25624 # of $id_scan_state = $tok is returned, indicating that further
25625 # calls must be made to get the identifier. If the type token is
25626 # not the last nonblank token on the line, the identifier is
25627 # scanned and handled and a value of '' is returned.
25628 # USES GLOBAL VARIABLES: $current_package, $last_nonblank_token, $in_attribute_list,
25629 # $statement_type, $tokenizer_self
25631 my ( $input_line, $i, $tok, $rtokens, $rtoken_map, $id_scan_state,
25635 my ( $i_beg, $pos_beg );
25637 #print "NSCAN:entering i=$i, tok=$tok, type=$type, state=$id_scan_state\n";
25638 #my ($a,$b,$c) = caller;
25639 #print "NSCAN: scan_id called with tok=$tok $a $b $c\n";
25641 # on re-entry, start scanning at first token on the line
25642 if ($id_scan_state) {
25647 # on initial entry, start scanning just after type token
25650 $id_scan_state = $tok;
25654 # find $i_beg = index of next nonblank token,
25655 # and handle empty lines
25656 my $blank_line = 0;
25657 my $next_nonblank_token = $$rtokens[$i_beg];
25658 if ( $i_beg > $max_token_index ) {
25663 # only a '#' immediately after a '$' is not a comment
25664 if ( $next_nonblank_token eq '#' ) {
25665 unless ( $tok eq '$' ) {
25670 if ( $next_nonblank_token =~ /^\s/ ) {
25671 ( $next_nonblank_token, $i_beg ) =
25672 find_next_nonblank_token_on_this_line( $i_beg, $rtokens,
25673 $max_token_index );
25674 if ( $next_nonblank_token =~ /(^#|^\s*$)/ ) {
25680 # handle non-blank line; identifier, if any, must follow
25681 unless ($blank_line) {
25683 if ( $id_scan_state eq 'sub' ) {
25684 ( $i, $tok, $type, $id_scan_state ) = do_scan_sub(
25685 $input_line, $i, $i_beg,
25686 $tok, $type, $rtokens,
25687 $rtoken_map, $id_scan_state, $max_token_index
25691 elsif ( $id_scan_state eq 'package' ) {
25692 ( $i, $tok, $type ) =
25693 do_scan_package( $input_line, $i, $i_beg, $tok, $type, $rtokens,
25694 $rtoken_map, $max_token_index );
25695 $id_scan_state = '';
25699 warning("invalid token in scan_id: $tok\n");
25700 $id_scan_state = '';
25704 if ( $id_scan_state && ( !defined($type) || !$type ) ) {
25706 # shouldn't happen:
25708 "Program bug in scan_id: undefined type but scan_state=$id_scan_state\n"
25710 report_definite_bug();
25713 TOKENIZER_DEBUG_FLAG_NSCAN && do {
25715 "NSCAN: returns i=$i, tok=$tok, type=$type, state=$id_scan_state\n";
25717 return ( $i, $tok, $type, $id_scan_state );
25720 sub check_prototype {
25721 my ( $proto, $package, $subname ) = @_;
25722 return unless ( defined($package) && defined($subname) );
25723 if ( defined($proto) ) {
25724 $proto =~ s/^\s*\(\s*//;
25725 $proto =~ s/\s*\)$//;
25727 $is_user_function{$package}{$subname} = 1;
25728 $user_function_prototype{$package}{$subname} = "($proto)";
25730 # prototypes containing '&' must be treated specially..
25731 if ( $proto =~ /\&/ ) {
25733 # right curly braces of prototypes ending in
25734 # '&' may be followed by an operator
25735 if ( $proto =~ /\&$/ ) {
25736 $is_block_function{$package}{$subname} = 1;
25739 # right curly braces of prototypes NOT ending in
25740 # '&' may NOT be followed by an operator
25741 elsif ( $proto !~ /\&$/ ) {
25742 $is_block_list_function{$package}{$subname} = 1;
25747 $is_constant{$package}{$subname} = 1;
25751 $is_user_function{$package}{$subname} = 1;
25755 sub do_scan_package {
25757 # do_scan_package parses a package name
25758 # it is called with $i_beg equal to the index of the first nonblank
25759 # token following a 'package' token.
25760 # USES GLOBAL VARIABLES: $current_package,
25762 my ( $input_line, $i, $i_beg, $tok, $type, $rtokens, $rtoken_map,
25765 my $package = undef;
25766 my $pos_beg = $$rtoken_map[$i_beg];
25767 pos($input_line) = $pos_beg;
25769 # handle non-blank line; package name, if any, must follow
25770 if ( $input_line =~ m/\G\s*((?:\w*(?:'|::))*\w+)/gc ) {
25772 $package = ( defined($1) && $1 ) ? $1 : 'main';
25773 $package =~ s/\'/::/g;
25774 if ( $package =~ /^\:/ ) { $package = 'main' . $package }
25775 $package =~ s/::$//;
25776 my $pos = pos($input_line);
25777 my $numc = $pos - $pos_beg;
25778 $tok = 'package ' . substr( $input_line, $pos_beg, $numc );
25781 # Now we must convert back from character position
25782 # to pre_token index.
25783 # I don't think an error flag can occur here ..but ?
25786 inverse_pretoken_map( $i, $pos, $rtoken_map, $max_token_index );
25787 if ($error) { warning("Possibly invalid package\n") }
25788 $current_package = $package;
25791 my ( $next_nonblank_token, $i_next ) =
25792 find_next_nonblank_token( $i, $rtokens, $max_token_index );
25793 if ( $next_nonblank_token !~ /^[;\}]$/ ) {
25795 "Unexpected '$next_nonblank_token' after package name '$tok'\n"
25800 # no match but line not blank --
25801 # could be a label with name package, like package: , for example.
25806 return ( $i, $tok, $type );
25809 sub scan_identifier_do {
25811 # This routine assembles tokens into identifiers. It maintains a
25812 # scan state, id_scan_state. It updates id_scan_state based upon
25813 # current id_scan_state and token, and returns an updated
25814 # id_scan_state and the next index after the identifier.
25815 # USES GLOBAL VARIABLES: $context, $last_nonblank_token,
25816 # $last_nonblank_type
25818 my ( $i, $id_scan_state, $identifier, $rtokens, $max_token_index,
25823 my $tok_begin = $$rtokens[$i_begin];
25824 if ( $tok_begin eq ':' ) { $tok_begin = '::' }
25825 my $id_scan_state_begin = $id_scan_state;
25826 my $identifier_begin = $identifier;
25827 my $tok = $tok_begin;
25830 # these flags will be used to help figure out the type:
25831 my $saw_alpha = ( $tok =~ /^[A-Za-z_]/ );
25834 # allow old package separator (') except in 'use' statement
25835 my $allow_tick = ( $last_nonblank_token ne 'use' );
25837 # get started by defining a type and a state if necessary
25838 unless ($id_scan_state) {
25839 $context = UNKNOWN_CONTEXT;
25841 # fixup for digraph
25842 if ( $tok eq '>' ) {
25846 $identifier = $tok;
25848 if ( $tok eq '$' || $tok eq '*' ) {
25849 $id_scan_state = '$';
25850 $context = SCALAR_CONTEXT;
25852 elsif ( $tok eq '%' || $tok eq '@' ) {
25853 $id_scan_state = '$';
25854 $context = LIST_CONTEXT;
25856 elsif ( $tok eq '&' ) {
25857 $id_scan_state = '&';
25859 elsif ( $tok eq 'sub' or $tok eq 'package' ) {
25860 $saw_alpha = 0; # 'sub' is considered type info here
25861 $id_scan_state = '$';
25862 $identifier .= ' '; # need a space to separate sub from sub name
25864 elsif ( $tok eq '::' ) {
25865 $id_scan_state = 'A';
25867 elsif ( $tok =~ /^[A-Za-z_]/ ) {
25868 $id_scan_state = ':';
25870 elsif ( $tok eq '->' ) {
25871 $id_scan_state = '$';
25876 my ( $a, $b, $c ) = caller;
25877 warning("Program Bug: scan_identifier given bad token = $tok \n");
25878 warning(" called from sub $a line: $c\n");
25879 report_definite_bug();
25881 $saw_type = !$saw_alpha;
25885 $saw_type = ( $tok =~ /([\$\%\@\*\&])/ );
25888 # now loop to gather the identifier
25891 while ( $i < $max_token_index ) {
25892 $i_save = $i unless ( $tok =~ /^\s*$/ );
25893 $tok = $$rtokens[ ++$i ];
25895 if ( ( $tok eq ':' ) && ( $$rtokens[ $i + 1 ] eq ':' ) ) {
25900 if ( $id_scan_state eq '$' ) { # starting variable name
25902 if ( $tok eq '$' ) {
25904 $identifier .= $tok;
25906 # we've got a punctuation variable if end of line (punct.t)
25907 if ( $i == $max_token_index ) {
25909 $id_scan_state = '';
25913 elsif ( $tok =~ /^[A-Za-z_]/ ) { # alphanumeric ..
25915 $id_scan_state = ':'; # now need ::
25916 $identifier .= $tok;
25918 elsif ( $tok eq "'" && $allow_tick ) { # alphanumeric ..
25920 $id_scan_state = ':'; # now need ::
25921 $identifier .= $tok;
25923 # Perl will accept leading digits in identifiers,
25924 # although they may not always produce useful results.
25925 # Something like $main::0 is ok. But this also works:
25927 # sub howdy::123::bubba{ print "bubba $54321!\n" }
25928 # howdy::123::bubba();
25931 elsif ( $tok =~ /^[0-9]/ ) { # numeric
25933 $id_scan_state = ':'; # now need ::
25934 $identifier .= $tok;
25936 elsif ( $tok eq '::' ) {
25937 $id_scan_state = 'A';
25938 $identifier .= $tok;
25940 elsif ( ( $tok eq '#' ) && ( $identifier eq '$' ) ) { # $#array
25941 $identifier .= $tok; # keep same state, a $ could follow
25943 elsif ( $tok eq '{' ) {
25945 # check for something like ${#} or ${©}
25946 if ( $identifier eq '$'
25947 && $i + 2 <= $max_token_index
25948 && $$rtokens[ $i + 2 ] eq '}'
25949 && $$rtokens[ $i + 1 ] !~ /[\s\w]/ )
25951 my $next2 = $$rtokens[ $i + 2 ];
25952 my $next1 = $$rtokens[ $i + 1 ];
25953 $identifier .= $tok . $next1 . $next2;
25955 $id_scan_state = '';
25959 # skip something like ${xxx} or ->{
25960 $id_scan_state = '';
25962 # if this is the first token of a line, any tokens for this
25963 # identifier have already been accumulated
25964 if ( $identifier eq '$' || $i == 0 ) { $identifier = ''; }
25969 # space ok after leading $ % * & @
25970 elsif ( $tok =~ /^\s*$/ ) {
25972 if ( $identifier =~ /^[\$\%\*\&\@]/ ) {
25974 if ( length($identifier) > 1 ) {
25975 $id_scan_state = '';
25977 $type = 'i'; # probably punctuation variable
25982 # spaces after $'s are common, and space after @
25983 # is harmless, so only complain about space
25984 # after other type characters. Space after $ and
25985 # @ will be removed in formatting. Report space
25986 # after % and * because they might indicate a
25987 # parsing error. In other words '% ' might be a
25988 # modulo operator. Delete this warning if it
25990 if ( $identifier !~ /^[\@\$]$/ ) {
25992 "Space in identifier, following $identifier\n";
25998 # space after '->' is ok
26000 elsif ( $tok eq '^' ) {
26002 # check for some special variables like $^W
26003 if ( $identifier =~ /^[\$\*\@\%]$/ ) {
26004 $identifier .= $tok;
26005 $id_scan_state = 'A';
26007 # Perl accepts '$^]' or '@^]', but
26008 # there must not be a space before the ']'.
26009 my $next1 = $$rtokens[ $i + 1 ];
26010 if ( $next1 eq ']' ) {
26012 $identifier .= $next1;
26013 $id_scan_state = "";
26018 $id_scan_state = '';
26021 else { # something else
26023 # check for various punctuation variables
26024 if ( $identifier =~ /^[\$\*\@\%]$/ ) {
26025 $identifier .= $tok;
26028 elsif ( $identifier eq '$#' ) {
26030 if ( $tok eq '{' ) { $type = 'i'; $i = $i_save }
26032 # perl seems to allow just these: $#: $#- $#+
26033 elsif ( $tok =~ /^[\:\-\+]$/ ) {
26035 $identifier .= $tok;
26039 write_logfile_entry( 'Use of $# is deprecated' . "\n" );
26042 elsif ( $identifier eq '$$' ) {
26044 # perl does not allow references to punctuation
26045 # variables without braces. For example, this
26049 # You would have to use
26053 if ( $tok eq '{' ) { $type = 't' }
26054 else { $type = 'i' }
26056 elsif ( $identifier eq '->' ) {
26061 if ( length($identifier) == 1 ) { $identifier = ''; }
26063 $id_scan_state = '';
26067 elsif ( $id_scan_state eq '&' ) { # starting sub call?
26069 if ( $tok =~ /^[\$A-Za-z_]/ ) { # alphanumeric ..
26070 $id_scan_state = ':'; # now need ::
26072 $identifier .= $tok;
26074 elsif ( $tok eq "'" && $allow_tick ) { # alphanumeric ..
26075 $id_scan_state = ':'; # now need ::
26077 $identifier .= $tok;
26079 elsif ( $tok =~ /^[0-9]/ ) { # numeric..see comments above
26080 $id_scan_state = ':'; # now need ::
26082 $identifier .= $tok;
26084 elsif ( $tok =~ /^\s*$/ ) { # allow space
26086 elsif ( $tok eq '::' ) { # leading ::
26087 $id_scan_state = 'A'; # accept alpha next
26088 $identifier .= $tok;
26090 elsif ( $tok eq '{' ) {
26091 if ( $identifier eq '&' || $i == 0 ) { $identifier = ''; }
26093 $id_scan_state = '';
26098 # punctuation variable?
26099 # testfile: cunningham4.pl
26101 # We have to be careful here. If we are in an unknown state,
26102 # we will reject the punctuation variable. In the following
26103 # example the '&' is a binary opeator but we are in an unknown
26104 # state because there is no sigil on 'Prima', so we don't
26105 # know what it is. But it is a bad guess that
26106 # '&~' is a punction variable.
26107 # $self->{text}->{colorMap}->[
26108 # Prima::PodView::COLOR_CODE_FOREGROUND
26109 # & ~tb::COLOR_INDEX ] =
26110 # $sec->{ColorCode}
26111 if ( $identifier eq '&' && $expecting ) {
26112 $identifier .= $tok;
26119 $id_scan_state = '';
26123 elsif ( $id_scan_state eq 'A' ) { # looking for alpha (after ::)
26125 if ( $tok =~ /^[A-Za-z_]/ ) { # found it
26126 $identifier .= $tok;
26127 $id_scan_state = ':'; # now need ::
26130 elsif ( $tok eq "'" && $allow_tick ) {
26131 $identifier .= $tok;
26132 $id_scan_state = ':'; # now need ::
26135 elsif ( $tok =~ /^[0-9]/ ) { # numeric..see comments above
26136 $identifier .= $tok;
26137 $id_scan_state = ':'; # now need ::
26140 elsif ( ( $identifier =~ /^sub / ) && ( $tok =~ /^\s*$/ ) ) {
26141 $id_scan_state = '(';
26142 $identifier .= $tok;
26144 elsif ( ( $identifier =~ /^sub / ) && ( $tok eq '(' ) ) {
26145 $id_scan_state = ')';
26146 $identifier .= $tok;
26149 $id_scan_state = '';
26154 elsif ( $id_scan_state eq ':' ) { # looking for :: after alpha
26156 if ( $tok eq '::' ) { # got it
26157 $identifier .= $tok;
26158 $id_scan_state = 'A'; # now require alpha
26160 elsif ( $tok =~ /^[A-Za-z_]/ ) { # more alphanumeric is ok here
26161 $identifier .= $tok;
26162 $id_scan_state = ':'; # now need ::
26165 elsif ( $tok =~ /^[0-9]/ ) { # numeric..see comments above
26166 $identifier .= $tok;
26167 $id_scan_state = ':'; # now need ::
26170 elsif ( $tok eq "'" && $allow_tick ) { # tick
26172 if ( $is_keyword{$identifier} ) {
26173 $id_scan_state = ''; # that's all
26177 $identifier .= $tok;
26180 elsif ( ( $identifier =~ /^sub / ) && ( $tok =~ /^\s*$/ ) ) {
26181 $id_scan_state = '(';
26182 $identifier .= $tok;
26184 elsif ( ( $identifier =~ /^sub / ) && ( $tok eq '(' ) ) {
26185 $id_scan_state = ')';
26186 $identifier .= $tok;
26189 $id_scan_state = ''; # that's all
26194 elsif ( $id_scan_state eq '(' ) { # looking for ( of prototype
26196 if ( $tok eq '(' ) { # got it
26197 $identifier .= $tok;
26198 $id_scan_state = ')'; # now find the end of it
26200 elsif ( $tok =~ /^\s*$/ ) { # blank - keep going
26201 $identifier .= $tok;
26204 $id_scan_state = ''; # that's all - no prototype
26209 elsif ( $id_scan_state eq ')' ) { # looking for ) to end
26211 if ( $tok eq ')' ) { # got it
26212 $identifier .= $tok;
26213 $id_scan_state = ''; # all done
26216 elsif ( $tok =~ /^[\s\$\%\\\*\@\&\;]/ ) {
26217 $identifier .= $tok;
26219 else { # probable error in script, but keep going
26220 warning("Unexpected '$tok' while seeking end of prototype\n");
26221 $identifier .= $tok;
26224 else { # can get here due to error in initialization
26225 $id_scan_state = '';
26231 if ( $id_scan_state eq ')' ) {
26232 warning("Hit end of line while seeking ) to end prototype\n");
26235 # once we enter the actual identifier, it may not extend beyond
26236 # the end of the current line
26237 if ( $id_scan_state =~ /^[A\:\(\)]/ ) {
26238 $id_scan_state = '';
26240 if ( $i < 0 ) { $i = 0 }
26247 if ( $identifier =~ /^->/ && $last_nonblank_type eq 'w' ) {
26250 else { $type = 'i' }
26252 elsif ( $identifier eq '->' ) {
26256 ( length($identifier) > 1 )
26258 # In something like '@$=' we have an identifier '@$'
26259 # In something like '$${' we have type '$$' (and only
26260 # part of an identifier)
26261 && !( $identifier =~ /\$$/ && $tok eq '{' )
26262 && ( $identifier !~ /^(sub |package )$/ )
26267 else { $type = 't' }
26269 elsif ($saw_alpha) {
26271 # type 'w' includes anything without leading type info
26272 # ($,%,@,*) including something like abc::def::ghi
26277 } # this can happen on a restart
26281 $tok = $identifier;
26282 if ($message) { write_logfile_entry($message) }
26289 TOKENIZER_DEBUG_FLAG_SCAN_ID && do {
26290 my ( $a, $b, $c ) = caller;
26292 "SCANID: called from $a $b $c with tok, i, state, identifier =$tok_begin, $i_begin, $id_scan_state_begin, $identifier_begin\n";
26294 "SCANID: returned with tok, i, state, identifier =$tok, $i, $id_scan_state, $identifier\n";
26296 return ( $i, $tok, $type, $id_scan_state, $identifier );
26301 # saved package and subnames in case prototype is on separate line
26302 my ( $package_saved, $subname_saved );
26306 # do_scan_sub parses a sub name and prototype
26307 # it is called with $i_beg equal to the index of the first nonblank
26308 # token following a 'sub' token.
26310 # TODO: add future error checks to be sure we have a valid
26311 # sub name. For example, 'sub &doit' is wrong. Also, be sure
26312 # a name is given if and only if a non-anonymous sub is
26314 # USES GLOBAL VARS: $current_package, $last_nonblank_token,
26315 # $in_attribute_list, %saw_function_definition,
26319 $input_line, $i, $i_beg,
26320 $tok, $type, $rtokens,
26321 $rtoken_map, $id_scan_state, $max_token_index
26323 $id_scan_state = ""; # normally we get everything in one call
26324 my $subname = undef;
26325 my $package = undef;
26330 my $pos_beg = $$rtoken_map[$i_beg];
26331 pos($input_line) = $pos_beg;
26333 # sub NAME PROTO ATTRS
26335 $input_line =~ m/\G\s*
26336 ((?:\w*(?:'|::))*) # package - something that ends in :: or '
26337 (\w+) # NAME - required
26338 (\s*\([^){]*\))? # PROTO - something in parens
26339 (\s*:)? # ATTRS - leading : of attribute list
26348 $package = ( defined($1) && $1 ) ? $1 : $current_package;
26349 $package =~ s/\'/::/g;
26350 if ( $package =~ /^\:/ ) { $package = 'main' . $package }
26351 $package =~ s/::$//;
26352 my $pos = pos($input_line);
26353 my $numc = $pos - $pos_beg;
26354 $tok = 'sub ' . substr( $input_line, $pos_beg, $numc );
26358 # Look for prototype/attributes not preceded on this line by subname;
26359 # This might be an anonymous sub with attributes,
26360 # or a prototype on a separate line from its sub name
26362 $input_line =~ m/\G(\s*\([^){]*\))? # PROTO
26363 (\s*:)? # ATTRS leading ':'
26372 # Handle prototype on separate line from subname
26373 if ($subname_saved) {
26374 $package = $package_saved;
26375 $subname = $subname_saved;
26376 $tok = $last_nonblank_token;
26383 # ATTRS: if there are attributes, back up and let the ':' be
26384 # found later by the scanner.
26385 my $pos = pos($input_line);
26387 $pos -= length($attrs);
26390 my $next_nonblank_token = $tok;
26392 # catch case of line with leading ATTR ':' after anonymous sub
26393 if ( $pos == $pos_beg && $tok eq ':' ) {
26395 $in_attribute_list = 1;
26398 # We must convert back from character position
26399 # to pre_token index.
26402 # I don't think an error flag can occur here ..but ?
26404 ( $i, $error ) = inverse_pretoken_map( $i, $pos, $rtoken_map,
26405 $max_token_index );
26406 if ($error) { warning("Possibly invalid sub\n") }
26408 # check for multiple definitions of a sub
26409 ( $next_nonblank_token, my $i_next ) =
26410 find_next_nonblank_token_on_this_line( $i, $rtokens,
26411 $max_token_index );
26414 if ( $next_nonblank_token =~ /^(\s*|#)$/ )
26415 { # skip blank or side comment
26416 my ( $rpre_tokens, $rpre_types ) =
26417 peek_ahead_for_n_nonblank_pre_tokens(1);
26418 if ( defined($rpre_tokens) && @$rpre_tokens ) {
26419 $next_nonblank_token = $rpre_tokens->[0];
26422 $next_nonblank_token = '}';
26425 $package_saved = "";
26426 $subname_saved = "";
26427 if ( $next_nonblank_token eq '{' ) {
26430 # Check for multiple definitions of a sub, but
26431 # it is ok to have multiple sub BEGIN, etc,
26432 # so we do not complain if name is all caps
26433 if ( $saw_function_definition{$package}{$subname}
26434 && $subname !~ /^[A-Z]+$/ )
26436 my $lno = $saw_function_definition{$package}{$subname};
26438 "already saw definition of 'sub $subname' in package '$package' at line $lno\n"
26441 $saw_function_definition{$package}{$subname} =
26442 $tokenizer_self->{_last_line_number};
26445 elsif ( $next_nonblank_token eq ';' ) {
26447 elsif ( $next_nonblank_token eq '}' ) {
26450 # ATTRS - if an attribute list follows, remember the name
26451 # of the sub so the next opening brace can be labeled.
26452 # Setting 'statement_type' causes any ':'s to introduce
26454 elsif ( $next_nonblank_token eq ':' ) {
26455 $statement_type = $tok;
26458 # see if PROTO follows on another line:
26459 elsif ( $next_nonblank_token eq '(' ) {
26460 if ( $attrs || $proto ) {
26462 "unexpected '(' after definition or declaration of sub '$subname'\n"
26466 $id_scan_state = 'sub'; # we must come back to get proto
26467 $statement_type = $tok;
26468 $package_saved = $package;
26469 $subname_saved = $subname;
26472 elsif ($next_nonblank_token) { # EOF technically ok
26474 "expecting ':' or ';' or '{' after definition or declaration of sub '$subname' but saw '$next_nonblank_token'\n"
26477 check_prototype( $proto, $package, $subname );
26480 # no match but line not blank
26483 return ( $i, $tok, $type, $id_scan_state );
26487 #########i###############################################################
26488 # Tokenizer utility routines which may use CONSTANTS but no other GLOBALS
26489 #########################################################################
26491 sub find_next_nonblank_token {
26492 my ( $i, $rtokens, $max_token_index ) = @_;
26494 if ( $i >= $max_token_index ) {
26495 if ( !peeked_ahead() ) {
26498 peek_ahead_for_nonblank_token( $rtokens, $max_token_index );
26501 my $next_nonblank_token = $$rtokens[ ++$i ];
26503 if ( $next_nonblank_token =~ /^\s*$/ ) {
26504 $next_nonblank_token = $$rtokens[ ++$i ];
26506 return ( $next_nonblank_token, $i );
26509 sub numerator_expected {
26511 # this is a filter for a possible numerator, in support of guessing
26512 # for the / pattern delimiter token.
26517 # Note: I am using the convention that variables ending in
26518 # _expected have these 3 possible values.
26519 my ( $i, $rtokens, $max_token_index ) = @_;
26520 my $next_token = $$rtokens[ $i + 1 ];
26521 if ( $next_token eq '=' ) { $i++; } # handle /=
26522 my ( $next_nonblank_token, $i_next ) =
26523 find_next_nonblank_token( $i, $rtokens, $max_token_index );
26525 if ( $next_nonblank_token =~ /(\(|\$|\w|\.|\@)/ ) {
26530 if ( $next_nonblank_token =~ /^\s*$/ ) {
26539 sub pattern_expected {
26541 # This is the start of a filter for a possible pattern.
26542 # It looks at the token after a possbible pattern and tries to
26543 # determine if that token could end a pattern.
26548 my ( $i, $rtokens, $max_token_index ) = @_;
26549 my $next_token = $$rtokens[ $i + 1 ];
26550 if ( $next_token =~ /^[cgimosxp]/ ) { $i++; } # skip possible modifier
26551 my ( $next_nonblank_token, $i_next ) =
26552 find_next_nonblank_token( $i, $rtokens, $max_token_index );
26554 # list of tokens which may follow a pattern
26555 # (can probably be expanded)
26556 if ( $next_nonblank_token =~ /(\)|\}|\;|\&\&|\|\||and|or|while|if|unless)/ )
26562 if ( $next_nonblank_token =~ /^\s*$/ ) {
26571 sub find_next_nonblank_token_on_this_line {
26572 my ( $i, $rtokens, $max_token_index ) = @_;
26573 my $next_nonblank_token;
26575 if ( $i < $max_token_index ) {
26576 $next_nonblank_token = $$rtokens[ ++$i ];
26578 if ( $next_nonblank_token =~ /^\s*$/ ) {
26580 if ( $i < $max_token_index ) {
26581 $next_nonblank_token = $$rtokens[ ++$i ];
26586 $next_nonblank_token = "";
26588 return ( $next_nonblank_token, $i );
26591 sub find_angle_operator_termination {
26593 # We are looking at a '<' and want to know if it is an angle operator.
26594 # We are to return:
26595 # $i = pretoken index of ending '>' if found, current $i otherwise
26596 # $type = 'Q' if found, '>' otherwise
26597 my ( $input_line, $i_beg, $rtoken_map, $expecting, $max_token_index ) = @_;
26600 pos($input_line) = 1 + $$rtoken_map[$i];
26604 # we just have to find the next '>' if a term is expected
26605 if ( $expecting == TERM ) { $filter = '[\>]' }
26607 # we have to guess if we don't know what is expected
26608 elsif ( $expecting == UNKNOWN ) { $filter = '[\>\;\=\#\|\<]' }
26610 # shouldn't happen - we shouldn't be here if operator is expected
26611 else { warning("Program Bug in find_angle_operator_termination\n") }
26613 # To illustrate what we might be looking at, in case we are
26614 # guessing, here are some examples of valid angle operators
26621 # <jskdfjskdfj* op/* jskdjfjkosvk*> ( glob.t)
26622 # <${PREFIX}*img*.$IMAGE_TYPE>
26623 # <img*.$IMAGE_TYPE>
26624 # <Timg*.$IMAGE_TYPE>
26625 # <$LATEX2HTMLVERSIONS${dd}html[1-9].[0-9].pl>
26627 # Here are some examples of lines which do not have angle operators:
26628 # return undef unless $self->[2]++ < $#{$self->[1]};
26631 # the following line from dlister.pl caused trouble:
26632 # print'~'x79,"\n",$D<1024?"0.$D":$D>>10,"K, $C files\n\n\n";
26634 # If the '<' starts an angle operator, it must end on this line and
26635 # it must not have certain characters like ';' and '=' in it. I use
26636 # this to limit the testing. This filter should be improved if
26639 if ( $input_line =~ /($filter)/g ) {
26643 # We MAY have found an angle operator termination if we get
26644 # here, but we need to do more to be sure we haven't been
26646 my $pos = pos($input_line);
26648 my $pos_beg = $$rtoken_map[$i];
26649 my $str = substr( $input_line, $pos_beg, ( $pos - $pos_beg ) );
26651 # Reject if the closing '>' follows a '-' as in:
26652 # if ( VERSION < 5.009 && $op-> name eq 'aassign' ) { }
26653 if ( $expecting eq UNKNOWN ) {
26654 my $check = substr( $input_line, $pos - 2, 1 );
26655 if ( $check eq '-' ) {
26656 return ( $i, $type );
26660 ######################################debug#####
26661 #write_diagnostics( "ANGLE? :$str\n");
26662 #print "ANGLE: found $1 at pos=$pos str=$str check=$check\n";
26663 ######################################debug#####
26667 inverse_pretoken_map( $i, $pos, $rtoken_map, $max_token_index );
26669 # It may be possible that a quote ends midway in a pretoken.
26670 # If this happens, it may be necessary to split the pretoken.
26673 "Possible tokinization error..please check this line\n");
26674 report_possible_bug();
26677 # Now let's see where we stand....
26678 # OK if math op not possible
26679 if ( $expecting == TERM ) {
26682 # OK if there are no more than 2 pre-tokens inside
26683 # (not possible to write 2 token math between < and >)
26684 # This catches most common cases
26685 elsif ( $i <= $i_beg + 3 ) {
26686 write_diagnostics("ANGLE(1 or 2 tokens): $str\n");
26692 # Let's try a Brace Test: any braces inside must balance
26694 while ( $str =~ /\{/g ) { $br++ }
26695 while ( $str =~ /\}/g ) { $br-- }
26697 while ( $str =~ /\[/g ) { $sb++ }
26698 while ( $str =~ /\]/g ) { $sb-- }
26700 while ( $str =~ /\(/g ) { $pr++ }
26701 while ( $str =~ /\)/g ) { $pr-- }
26703 # if braces do not balance - not angle operator
26704 if ( $br || $sb || $pr ) {
26708 "NOT ANGLE (BRACE={$br ($pr [$sb ):$str\n");
26711 # we should keep doing more checks here...to be continued
26712 # Tentatively accepting this as a valid angle operator.
26713 # There are lots more things that can be checked.
26716 "ANGLE-Guessing yes: $str expecting=$expecting\n");
26717 write_logfile_entry("Guessing angle operator here: $str\n");
26722 # didn't find ending >
26724 if ( $expecting == TERM ) {
26725 warning("No ending > for angle operator\n");
26729 return ( $i, $type );
26732 sub scan_number_do {
26734 # scan a number in any of the formats that Perl accepts
26735 # Underbars (_) are allowed in decimal numbers.
26736 # input parameters -
26737 # $input_line - the string to scan
26738 # $i - pre_token index to start scanning
26739 # $rtoken_map - reference to the pre_token map giving starting
26740 # character position in $input_line of token $i
26741 # output parameters -
26742 # $i - last pre_token index of the number just scanned
26743 # number - the number (characters); or undef if not a number
26745 my ( $input_line, $i, $rtoken_map, $input_type, $max_token_index ) = @_;
26746 my $pos_beg = $$rtoken_map[$i];
26749 my $number = undef;
26750 my $type = $input_type;
26752 my $first_char = substr( $input_line, $pos_beg, 1 );
26754 # Look for bad starting characters; Shouldn't happen..
26755 if ( $first_char !~ /[\d\.\+\-Ee]/ ) {
26756 warning("Program bug - scan_number given character $first_char\n");
26757 report_definite_bug();
26758 return ( $i, $type, $number );
26761 # handle v-string without leading 'v' character ('Two Dot' rule)
26763 # TODO: v-strings may contain underscores
26764 pos($input_line) = $pos_beg;
26765 if ( $input_line =~ /\G((\d+)?\.\d+(\.\d+)+)/g ) {
26766 $pos = pos($input_line);
26767 my $numc = $pos - $pos_beg;
26768 $number = substr( $input_line, $pos_beg, $numc );
26770 report_v_string($number);
26773 # handle octal, hex, binary
26774 if ( !defined($number) ) {
26775 pos($input_line) = $pos_beg;
26776 if ( $input_line =~ /\G[+-]?0((x[0-9a-fA-F_]+)|([0-7_]+)|(b[01_]+))/g )
26778 $pos = pos($input_line);
26779 my $numc = $pos - $pos_beg;
26780 $number = substr( $input_line, $pos_beg, $numc );
26786 if ( !defined($number) ) {
26787 pos($input_line) = $pos_beg;
26789 if ( $input_line =~ /\G([+-]?[\d_]*(\.[\d_]*)?([Ee][+-]?(\d+))?)/g ) {
26790 $pos = pos($input_line);
26792 # watch out for things like 0..40 which would give 0. by this;
26793 if ( ( substr( $input_line, $pos - 1, 1 ) eq '.' )
26794 && ( substr( $input_line, $pos, 1 ) eq '.' ) )
26798 my $numc = $pos - $pos_beg;
26799 $number = substr( $input_line, $pos_beg, $numc );
26804 # filter out non-numbers like e + - . e2 .e3 +e6
26805 # the rule: at least one digit, and any 'e' must be preceded by a digit
26807 $number !~ /\d/ # no digits
26808 || ( $number =~ /^(.*)[eE]/
26809 && $1 !~ /\d/ ) # or no digits before the 'e'
26813 $type = $input_type;
26814 return ( $i, $type, $number );
26817 # Found a number; now we must convert back from character position
26818 # to pre_token index. An error here implies user syntax error.
26819 # An example would be an invalid octal number like '009'.
26822 inverse_pretoken_map( $i, $pos, $rtoken_map, $max_token_index );
26823 if ($error) { warning("Possibly invalid number\n") }
26825 return ( $i, $type, $number );
26828 sub inverse_pretoken_map {
26830 # Starting with the current pre_token index $i, scan forward until
26831 # finding the index of the next pre_token whose position is $pos.
26832 my ( $i, $pos, $rtoken_map, $max_token_index ) = @_;
26835 while ( ++$i <= $max_token_index ) {
26837 if ( $pos <= $$rtoken_map[$i] ) {
26839 # Let the calling routine handle errors in which we do not
26840 # land on a pre-token boundary. It can happen by running
26841 # perltidy on some non-perl scripts, for example.
26842 if ( $pos < $$rtoken_map[$i] ) { $error = 1 }
26847 return ( $i, $error );
26850 sub find_here_doc {
26852 # find the target of a here document, if any
26853 # input parameters:
26854 # $i - token index of the second < of <<
26855 # ($i must be less than the last token index if this is called)
26856 # output parameters:
26857 # $found_target = 0 didn't find target; =1 found target
26858 # HERE_TARGET - the target string (may be empty string)
26859 # $i - unchanged if not here doc,
26860 # or index of the last token of the here target
26861 # $saw_error - flag noting unbalanced quote on here target
26862 my ( $expecting, $i, $rtokens, $rtoken_map, $max_token_index ) = @_;
26864 my $found_target = 0;
26865 my $here_doc_target = '';
26866 my $here_quote_character = '';
26868 my ( $next_nonblank_token, $i_next_nonblank, $next_token );
26869 $next_token = $$rtokens[ $i + 1 ];
26871 # perl allows a backslash before the target string (heredoc.t)
26873 if ( $next_token eq '\\' ) {
26875 $next_token = $$rtokens[ $i + 2 ];
26878 ( $next_nonblank_token, $i_next_nonblank ) =
26879 find_next_nonblank_token_on_this_line( $i, $rtokens, $max_token_index );
26881 if ( $next_nonblank_token =~ /[\'\"\`]/ ) {
26884 my $quote_depth = 0;
26889 $i, $in_quote, $here_quote_character, $quote_pos, $quote_depth,
26892 = follow_quoted_string( $i_next_nonblank, $in_quote, $rtokens,
26893 $here_quote_character, $quote_pos, $quote_depth, $max_token_index );
26895 if ($in_quote) { # didn't find end of quote, so no target found
26897 if ( $expecting == TERM ) {
26899 "Did not find here-doc string terminator ($here_quote_character) before end of line \n"
26904 else { # found ending quote
26909 for ( $j = $i_next_nonblank + 1 ; $j < $i ; $j++ ) {
26910 $tokj = $$rtokens[$j];
26912 # we have to remove any backslash before the quote character
26913 # so that the here-doc-target exactly matches this string
26917 && $$rtokens[ $j + 1 ] eq $here_quote_character );
26918 $here_doc_target .= $tokj;
26923 elsif ( ( $next_token =~ /^\s*$/ ) and ( $expecting == TERM ) ) {
26925 write_logfile_entry(
26926 "found blank here-target after <<; suggest using \"\"\n");
26929 elsif ( $next_token =~ /^\w/ ) { # simple bareword or integer after <<
26931 my $here_doc_expected;
26932 if ( $expecting == UNKNOWN ) {
26933 $here_doc_expected = guess_if_here_doc($next_token);
26936 $here_doc_expected = 1;
26939 if ($here_doc_expected) {
26941 $here_doc_target = $next_token;
26948 if ( $expecting == TERM ) {
26950 write_logfile_entry("Note: bare here-doc operator <<\n");
26957 # patch to neglect any prepended backslash
26958 if ( $found_target && $backslash ) { $i++ }
26960 return ( $found_target, $here_doc_target, $here_quote_character, $i,
26966 # follow (or continue following) quoted string(s)
26967 # $in_quote return code:
26968 # 0 - ok, found end
26969 # 1 - still must find end of quote whose target is $quote_character
26970 # 2 - still looking for end of first of two quotes
26972 # Returns updated strings:
26973 # $quoted_string_1 = quoted string seen while in_quote=1
26974 # $quoted_string_2 = quoted string seen while in_quote=2
26976 $i, $in_quote, $quote_character,
26977 $quote_pos, $quote_depth, $quoted_string_1,
26978 $quoted_string_2, $rtokens, $rtoken_map,
26982 my $in_quote_starting = $in_quote;
26985 if ( $in_quote == 2 ) { # two quotes/quoted_string_1s to follow
26988 $i, $in_quote, $quote_character, $quote_pos, $quote_depth,
26991 = follow_quoted_string( $i, $in_quote, $rtokens, $quote_character,
26992 $quote_pos, $quote_depth, $max_token_index );
26993 $quoted_string_2 .= $quoted_string;
26994 if ( $in_quote == 1 ) {
26995 if ( $quote_character =~ /[\{\[\<\(]/ ) { $i++; }
26996 $quote_character = '';
26999 $quoted_string_2 .= "\n";
27003 if ( $in_quote == 1 ) { # one (more) quote to follow
27006 $i, $in_quote, $quote_character, $quote_pos, $quote_depth,
27009 = follow_quoted_string( $ibeg, $in_quote, $rtokens, $quote_character,
27010 $quote_pos, $quote_depth, $max_token_index );
27011 $quoted_string_1 .= $quoted_string;
27012 if ( $in_quote == 1 ) {
27013 $quoted_string_1 .= "\n";
27016 return ( $i, $in_quote, $quote_character, $quote_pos, $quote_depth,
27017 $quoted_string_1, $quoted_string_2 );
27020 sub follow_quoted_string {
27022 # scan for a specific token, skipping escaped characters
27023 # if the quote character is blank, use the first non-blank character
27024 # input parameters:
27025 # $rtokens = reference to the array of tokens
27026 # $i = the token index of the first character to search
27027 # $in_quote = number of quoted strings being followed
27028 # $beginning_tok = the starting quote character
27029 # $quote_pos = index to check next for alphanumeric delimiter
27030 # output parameters:
27031 # $i = the token index of the ending quote character
27032 # $in_quote = decremented if found end, unchanged if not
27033 # $beginning_tok = the starting quote character
27034 # $quote_pos = index to check next for alphanumeric delimiter
27035 # $quote_depth = nesting depth, since delimiters '{ ( [ <' can be nested.
27036 # $quoted_string = the text of the quote (without quotation tokens)
27037 my ( $i_beg, $in_quote, $rtokens, $beginning_tok, $quote_pos, $quote_depth,
27040 my ( $tok, $end_tok );
27041 my $i = $i_beg - 1;
27042 my $quoted_string = "";
27044 TOKENIZER_DEBUG_FLAG_QUOTE && do {
27046 "QUOTE entering with quote_pos = $quote_pos i=$i beginning_tok =$beginning_tok\n";
27049 # get the corresponding end token
27050 if ( $beginning_tok !~ /^\s*$/ ) {
27051 $end_tok = matching_end_token($beginning_tok);
27054 # a blank token means we must find and use the first non-blank one
27056 my $allow_quote_comments = ( $i < 0 ) ? 1 : 0; # i<0 means we saw a <cr>
27058 while ( $i < $max_token_index ) {
27059 $tok = $$rtokens[ ++$i ];
27061 if ( $tok !~ /^\s*$/ ) {
27063 if ( ( $tok eq '#' ) && ($allow_quote_comments) ) {
27064 $i = $max_token_index;
27068 if ( length($tok) > 1 ) {
27069 if ( $quote_pos <= 0 ) { $quote_pos = 1 }
27070 $beginning_tok = substr( $tok, $quote_pos - 1, 1 );
27073 $beginning_tok = $tok;
27076 $end_tok = matching_end_token($beginning_tok);
27082 $allow_quote_comments = 1;
27087 # There are two different loops which search for the ending quote
27088 # character. In the rare case of an alphanumeric quote delimiter, we
27089 # have to look through alphanumeric tokens character-by-character, since
27090 # the pre-tokenization process combines multiple alphanumeric
27091 # characters, whereas for a non-alphanumeric delimiter, only tokens of
27092 # length 1 can match.
27094 ###################################################################
27095 # Case 1 (rare): loop for case of alphanumeric quote delimiter..
27096 # "quote_pos" is the position the current word to begin searching
27097 ###################################################################
27098 if ( $beginning_tok =~ /\w/ ) {
27100 # Note this because it is not recommended practice except
27101 # for obfuscated perl contests
27102 if ( $in_quote == 1 ) {
27103 write_logfile_entry(
27104 "Note: alphanumeric quote delimiter ($beginning_tok) \n");
27107 while ( $i < $max_token_index ) {
27109 if ( $quote_pos == 0 || ( $i < 0 ) ) {
27110 $tok = $$rtokens[ ++$i ];
27112 if ( $tok eq '\\' ) {
27114 # retain backslash unless it hides the end token
27115 $quoted_string .= $tok
27116 unless $$rtokens[ $i + 1 ] eq $end_tok;
27118 last if ( $i >= $max_token_index );
27119 $tok = $$rtokens[ ++$i ];
27122 my $old_pos = $quote_pos;
27124 unless ( defined($tok) && defined($end_tok) && defined($quote_pos) )
27128 $quote_pos = 1 + index( $tok, $end_tok, $quote_pos );
27130 if ( $quote_pos > 0 ) {
27133 substr( $tok, $old_pos, $quote_pos - $old_pos - 1 );
27137 if ( $quote_depth == 0 ) {
27143 $quoted_string .= substr( $tok, $old_pos );
27148 ########################################################################
27149 # Case 2 (normal): loop for case of a non-alphanumeric quote delimiter..
27150 ########################################################################
27153 while ( $i < $max_token_index ) {
27154 $tok = $$rtokens[ ++$i ];
27156 if ( $tok eq $end_tok ) {
27159 if ( $quote_depth == 0 ) {
27164 elsif ( $tok eq $beginning_tok ) {
27167 elsif ( $tok eq '\\' ) {
27169 # retain backslash unless it hides the beginning or end token
27170 $tok = $$rtokens[ ++$i ];
27171 $quoted_string .= '\\'
27172 unless ( $tok eq $end_tok || $tok eq $beginning_tok );
27174 $quoted_string .= $tok;
27177 if ( $i > $max_token_index ) { $i = $max_token_index }
27178 return ( $i, $in_quote, $beginning_tok, $quote_pos, $quote_depth,
27182 sub indicate_error {
27183 my ( $msg, $line_number, $input_line, $pos, $carrat ) = @_;
27184 interrupt_logfile();
27186 write_error_indicator_pair( $line_number, $input_line, $pos, $carrat );
27190 sub write_error_indicator_pair {
27191 my ( $line_number, $input_line, $pos, $carrat ) = @_;
27192 my ( $offset, $numbered_line, $underline ) =
27193 make_numbered_line( $line_number, $input_line, $pos );
27194 $underline = write_on_underline( $underline, $pos - $offset, $carrat );
27195 warning( $numbered_line . "\n" );
27196 $underline =~ s/\s*$//;
27197 warning( $underline . "\n" );
27200 sub make_numbered_line {
27202 # Given an input line, its line number, and a character position of
27203 # interest, create a string not longer than 80 characters of the form
27204 # $lineno: sub_string
27205 # such that the sub_string of $str contains the position of interest
27207 # Here is an example of what we want, in this case we add trailing
27208 # '...' because the line is long.
27210 # 2: (One of QAML 2.0's authors is a member of the World Wide Web Con ...
27212 # Here is another example, this time in which we used leading '...'
27213 # because of excessive length:
27215 # 2: ... er of the World Wide Web Consortium's
27217 # input parameters are:
27218 # $lineno = line number
27219 # $str = the text of the line
27220 # $pos = position of interest (the error) : 0 = first character
27223 # - $offset = an offset which corrects the position in case we only
27224 # display part of a line, such that $pos-$offset is the effective
27225 # position from the start of the displayed line.
27226 # - $numbered_line = the numbered line as above,
27227 # - $underline = a blank 'underline' which is all spaces with the same
27228 # number of characters as the numbered line.
27230 my ( $lineno, $str, $pos ) = @_;
27231 my $offset = ( $pos < 60 ) ? 0 : $pos - 40;
27232 my $excess = length($str) - $offset - 68;
27233 my $numc = ( $excess > 0 ) ? 68 : undef;
27235 if ( defined($numc) ) {
27236 if ( $offset == 0 ) {
27237 $str = substr( $str, $offset, $numc - 4 ) . " ...";
27240 $str = "... " . substr( $str, $offset + 4, $numc - 4 ) . " ...";
27245 if ( $offset == 0 ) {
27248 $str = "... " . substr( $str, $offset + 4 );
27252 my $numbered_line = sprintf( "%d: ", $lineno );
27253 $offset -= length($numbered_line);
27254 $numbered_line .= $str;
27255 my $underline = " " x length($numbered_line);
27256 return ( $offset, $numbered_line, $underline );
27259 sub write_on_underline {
27261 # The "underline" is a string that shows where an error is; it starts
27262 # out as a string of blanks with the same length as the numbered line of
27263 # code above it, and we have to add marking to show where an error is.
27264 # In the example below, we want to write the string '--^' just below
27265 # the line of bad code:
27267 # 2: (One of QAML 2.0's authors is a member of the World Wide Web Con ...
27269 # We are given the current underline string, plus a position and a
27270 # string to write on it.
27272 # In the above example, there will be 2 calls to do this:
27273 # First call: $pos=19, pos_chr=^
27274 # Second call: $pos=16, pos_chr=---
27276 # This is a trivial thing to do with substr, but there is some
27279 my ( $underline, $pos, $pos_chr ) = @_;
27281 # check for error..shouldn't happen
27282 unless ( ( $pos >= 0 ) && ( $pos <= length($underline) ) ) {
27285 my $excess = length($pos_chr) + $pos - length($underline);
27286 if ( $excess > 0 ) {
27287 $pos_chr = substr( $pos_chr, 0, length($pos_chr) - $excess );
27289 substr( $underline, $pos, length($pos_chr) ) = $pos_chr;
27290 return ($underline);
27295 # Break a string, $str, into a sequence of preliminary tokens. We
27296 # are interested in these types of tokens:
27297 # words (type='w'), example: 'max_tokens_wanted'
27298 # digits (type = 'd'), example: '0755'
27299 # whitespace (type = 'b'), example: ' '
27300 # any other single character (i.e. punct; type = the character itself).
27301 # We cannot do better than this yet because we might be in a quoted
27302 # string or pattern. Caller sets $max_tokens_wanted to 0 to get all
27304 my ( $str, $max_tokens_wanted ) = @_;
27306 # we return references to these 3 arrays:
27307 my @tokens = (); # array of the tokens themselves
27308 my @token_map = (0); # string position of start of each token
27309 my @type = (); # 'b'=whitespace, 'd'=digits, 'w'=alpha, or punct
27314 if ( $str =~ /\G(\s+)/gc ) { push @type, 'b'; }
27317 # note that this must come before words!
27318 elsif ( $str =~ /\G(\d+)/gc ) { push @type, 'd'; }
27321 elsif ( $str =~ /\G(\w+)/gc ) { push @type, 'w'; }
27323 # single-character punctuation
27324 elsif ( $str =~ /\G(\W)/gc ) { push @type, $1; }
27328 return ( \@tokens, \@token_map, \@type );
27332 push @token_map, pos($str);
27334 } while ( --$max_tokens_wanted != 0 );
27336 return ( \@tokens, \@token_map, \@type );
27341 # this is an old debug routine
27342 my ( $rtokens, $rtoken_map ) = @_;
27343 my $num = scalar(@$rtokens);
27346 for ( $i = 0 ; $i < $num ; $i++ ) {
27347 my $len = length( $$rtokens[$i] );
27348 print "$i:$len:$$rtoken_map[$i]:$$rtokens[$i]:\n";
27352 sub matching_end_token {
27354 # find closing character for a pattern
27355 my $beginning_token = shift;
27357 if ( $beginning_token eq '{' ) {
27360 elsif ( $beginning_token eq '[' ) {
27363 elsif ( $beginning_token eq '<' ) {
27366 elsif ( $beginning_token eq '(' ) {
27374 sub dump_token_types {
27378 # This should be the latest list of token types in use
27379 # adding NEW_TOKENS: add a comment here
27380 print $fh <<'END_OF_LIST';
27382 Here is a list of the token types currently used for lines of type 'CODE'.
27383 For the following tokens, the "type" of a token is just the token itself.
27385 .. :: << >> ** && .. || // -> => += -= .= %= &= |= ^= *= <>
27386 ( ) <= >= == =~ !~ != ++ -- /= x=
27387 ... **= <<= >>= &&= ||= //= <=>
27388 , + - / * | % ! x ~ = \ ? : . < > ^ &
27390 The following additional token types are defined:
27393 b blank (white space)
27394 { indent: opening structural curly brace or square bracket or paren
27395 (code block, anonymous hash reference, or anonymous array reference)
27396 } outdent: right structural curly brace or square bracket or paren
27397 [ left non-structural square bracket (enclosing an array index)
27398 ] right non-structural square bracket
27399 ( left non-structural paren (all but a list right of an =)
27400 ) right non-structural parena
27401 L left non-structural curly brace (enclosing a key)
27402 R right non-structural curly brace
27403 ; terminal semicolon
27404 f indicates a semicolon in a "for" statement
27405 h here_doc operator <<
27407 Q indicates a quote or pattern
27408 q indicates a qw quote block
27410 C user-defined constant or constant function (with void prototype = ())
27411 U user-defined function taking parameters
27412 G user-defined function taking block parameter (like grep/map/eval)
27413 M (unused, but reserved for subroutine definition name)
27414 P (unused, but -html uses it to label pod text)
27415 t type indicater such as %,$,@,*,&,sub
27416 w bare word (perhaps a subroutine call)
27417 i identifier of some type (with leading %, $, @, *, &, sub, -> )
27420 F a file test operator (like -e)
27422 Z identifier in indirect object slot: may be file handle, object
27423 J LABEL: code block label
27424 j LABEL after next, last, redo, goto
27427 pp pre-increment operator ++
27428 mm pre-decrement operator --
27429 A : used as attribute separator
27431 Here are the '_line_type' codes used internally:
27432 SYSTEM - system-specific code before hash-bang line
27433 CODE - line of perl code (including comments)
27434 POD_START - line starting pod, such as '=head'
27435 POD - pod documentation text
27436 POD_END - last line of pod section, '=cut'
27437 HERE - text of here-document
27438 HERE_END - last line of here-doc (target word)
27439 FORMAT - format section
27440 FORMAT_END - last line of format section, '.'
27441 DATA_START - __DATA__ line
27442 DATA - unidentified text following __DATA__
27443 END_START - __END__ line
27444 END - unidentified text following __END__
27445 ERROR - we are in big trouble, probably not a perl script
27451 # These names are used in error messages
27452 @opening_brace_names = qw# '{' '[' '(' '?' #;
27453 @closing_brace_names = qw# '}' ']' ')' ':' #;
27456 .. :: << >> ** && .. || // -> => += -= .= %= &= |= ^= *= <>
27457 <= >= == =~ !~ != ++ -- /= x= ~~
27459 @is_digraph{@digraphs} = (1) x scalar(@digraphs);
27461 my @trigraphs = qw( ... **= <<= >>= &&= ||= //= <=> !~~ );
27462 @is_trigraph{@trigraphs} = (1) x scalar(@trigraphs);
27464 # make a hash of all valid token types for self-checking the tokenizer
27465 # (adding NEW_TOKENS : select a new character and add to this list)
27466 my @valid_token_types = qw#
27467 A b C G L R f h Q k t w i q n p m F pp mm U j J Y Z v
27468 { } ( ) [ ] ; + - / * | % ! x ~ = \ ? : . < > ^ &
27470 push( @valid_token_types, @digraphs );
27471 push( @valid_token_types, @trigraphs );
27472 push( @valid_token_types, '#' );
27473 push( @valid_token_types, ',' );
27474 @is_valid_token_type{@valid_token_types} = (1) x scalar(@valid_token_types);
27476 # a list of file test letters, as in -e (Table 3-4 of 'camel 3')
27477 my @file_test_operators =
27478 qw( A B C M O R S T W X b c d e f g k l o p r s t u w x z);
27479 @is_file_test_operator{@file_test_operators} =
27480 (1) x scalar(@file_test_operators);
27482 # these functions have prototypes of the form (&), so when they are
27483 # followed by a block, that block MAY BE followed by an operator.
27484 @_ = qw( do eval );
27485 @is_block_operator{@_} = (1) x scalar(@_);
27487 # these functions allow an identifier in the indirect object slot
27488 @_ = qw( print printf sort exec system say);
27489 @is_indirect_object_taker{@_} = (1) x scalar(@_);
27491 # These tokens may precede a code block
27492 # patched for SWITCH/CASE
27494 qw( BEGIN END CHECK INIT AUTOLOAD DESTROY UNITCHECK continue if elsif else
27495 unless do while until eval for foreach map grep sort
27496 switch case given when);
27497 @is_code_block_token{@_} = (1) x scalar(@_);
27499 # I'll build the list of keywords incrementally
27502 # keywords and tokens after which a value or pattern is expected,
27503 # but not an operator. In other words, these should consume terms
27504 # to their right, or at least they are not expected to be followed
27505 # immediately by operators.
27506 my @value_requestor = qw(
27727 # patched above for SWITCH/CASE given/when err say
27728 # 'err' is a fairly safe addition.
27729 # TODO: 'default' still needed if appropriate
27730 # 'use feature' seen, but perltidy works ok without it.
27731 # Concerned that 'default' could break code.
27732 push( @Keywords, @value_requestor );
27734 # These are treated the same but are not keywords:
27739 push( @value_requestor, @extra_vr );
27741 @expecting_term_token{@value_requestor} = (1) x scalar(@value_requestor);
27743 # this list contains keywords which do not look for arguments,
27744 # so that they might be followed by an operator, or at least
27746 my @operator_requestor = qw(
27770 push( @Keywords, @operator_requestor );
27772 # These are treated the same but are not considered keywords:
27779 push( @operator_requestor, @extra_or );
27781 @expecting_operator_token{@operator_requestor} =
27782 (1) x scalar(@operator_requestor);
27784 # these token TYPES expect trailing operator but not a term
27785 # note: ++ and -- are post-increment and decrement, 'C' = constant
27786 my @operator_requestor_types = qw( ++ -- C <> q );
27787 @expecting_operator_types{@operator_requestor_types} =
27788 (1) x scalar(@operator_requestor_types);
27790 # these token TYPES consume values (terms)
27791 # note: pp and mm are pre-increment and decrement
27792 # f=semicolon in for, F=file test operator
27793 my @value_requestor_type = qw#
27794 L { ( [ ~ !~ =~ ; . .. ... A : && ! || // = + - x
27795 **= += -= .= /= *= %= x= &= |= ^= <<= >>= &&= ||= //=
27796 <= >= == != => \ > < % * / ? & | ** <=> ~~ !~~
27797 f F pp mm Y p m U J G j >> << ^ t
27799 push( @value_requestor_type, ',' )
27800 ; # (perl doesn't like a ',' in a qw block)
27801 @expecting_term_types{@value_requestor_type} =
27802 (1) x scalar(@value_requestor_type);
27804 # Note: the following valid token types are not assigned here to
27805 # hashes requesting to be followed by values or terms, but are
27806 # instead currently hard-coded into sub operator_expected:
27807 # ) -> :: Q R Z ] b h i k n v w } #
27809 # For simple syntax checking, it is nice to have a list of operators which
27810 # will really be unhappy if not followed by a term. This includes most
27812 %really_want_term = %expecting_term_types;
27814 # with these exceptions...
27815 delete $really_want_term{'U'}; # user sub, depends on prototype
27816 delete $really_want_term{'F'}; # file test works on $_ if no following term
27817 delete $really_want_term{'Y'}; # indirect object, too risky to check syntax;
27820 @_ = qw(q qq qw qx qr s y tr m);
27821 @is_q_qq_qw_qx_qr_s_y_tr_m{@_} = (1) x scalar(@_);
27823 # These keywords are handled specially in the tokenizer code:
27824 my @special_keywords = qw(
27840 push( @Keywords, @special_keywords );
27842 # Keywords after which list formatting may be used
27843 # WARNING: do not include |map|grep|eval or perl may die on
27844 # syntax errors (map1.t).
27845 my @keyword_taking_list = qw(
27917 @is_keyword_taking_list{@keyword_taking_list} =
27918 (1) x scalar(@keyword_taking_list);
27920 # These are not used in any way yet
27921 # my @unused_keywords = qw(
27928 # The list of keywords was extracted from function 'keyword' in
27929 # perl file toke.c version 5.005.03, using this utility, plus a
27930 # little editing: (file getkwd.pl):
27931 # while (<>) { while (/\"(.*)\"/g) { print "$1\n"; } }
27932 # Add 'get' prefix where necessary, then split into the above lists.
27933 # This list should be updated as necessary.
27934 # The list should not contain these special variables:
27935 # ARGV DATA ENV SIG STDERR STDIN STDOUT
27938 @is_keyword{@Keywords} = (1) x scalar(@Keywords);
27945 Perl::Tidy - Parses and beautifies perl source
27951 Perl::Tidy::perltidy(
27953 destination => $destination,
27956 perltidyrc => $perltidyrc,
27957 logfile => $logfile,
27958 errorfile => $errorfile,
27959 formatter => $formatter, # callback object (see below)
27960 dump_options => $dump_options,
27961 dump_options_type => $dump_options_type,
27966 This module makes the functionality of the perltidy utility available to perl
27967 scripts. Any or all of the input parameters may be omitted, in which case the
27968 @ARGV array will be used to provide input parameters as described
27969 in the perltidy(1) man page.
27971 For example, the perltidy script is basically just this:
27974 Perl::Tidy::perltidy();
27976 The module accepts input and output streams by a variety of methods.
27977 The following list of parameters may be any of a the following: a
27978 filename, an ARRAY reference, a SCALAR reference, or an object with
27979 either a B<getline> or B<print> method, as appropriate.
27981 source - the source of the script to be formatted
27982 destination - the destination of the formatted output
27983 stderr - standard error output
27984 perltidyrc - the .perltidyrc file
27985 logfile - the .LOG file stream, if any
27986 errorfile - the .ERR file stream, if any
27987 dump_options - ref to a hash to receive parameters (see below),
27988 dump_options_type - controls contents of dump_options
27989 dump_getopt_flags - ref to a hash to receive Getopt flags
27990 dump_options_category - ref to a hash giving category of options
27991 dump_abbreviations - ref to a hash giving all abbreviations
27993 The following chart illustrates the logic used to decide how to
27996 ref($param) $param is assumed to be:
27997 ----------- ---------------------
27999 SCALAR ref to string
28001 (other) object with getline (if source) or print method
28003 If the parameter is an object, and the object has a B<close> method, that
28004 close method will be called at the end of the stream.
28010 If the B<source> parameter is given, it defines the source of the
28015 If the B<destination> parameter is given, it will be used to define the
28016 file or memory location to receive output of perltidy.
28020 The B<stderr> parameter allows the calling program to capture the output
28021 to what would otherwise go to the standard error output device.
28025 If the B<perltidyrc> file is given, it will be used instead of any
28026 F<.perltidyrc> configuration file that would otherwise be used.
28030 If the B<argv> parameter is given, it will be used instead of the
28031 B<@ARGV> array. The B<argv> parameter may be a string, a reference to a
28032 string, or a reference to an array. If it is a string or reference to a
28033 string, it will be parsed into an array of items just as if it were a
28034 command line string.
28038 If the B<dump_options> parameter is given, it must be the reference to a hash.
28039 In this case, the parameters contained in any perltidyrc configuration file
28040 will be placed in this hash and perltidy will return immediately. This is
28041 equivalent to running perltidy with --dump-options, except that the perameters
28042 are returned in a hash rather than dumped to standard output. Also, by default
28043 only the parameters in the perltidyrc file are returned, but this can be
28044 changed (see the next parameter). This parameter provides a convenient method
28045 for external programs to read a perltidyrc file. An example program using
28046 this feature, F<perltidyrc_dump.pl>, is included in the distribution.
28048 Any combination of the B<dump_> parameters may be used together.
28050 =item dump_options_type
28052 This parameter is a string which can be used to control the parameters placed
28053 in the hash reference supplied by B<dump_options>. The possible values are
28054 'perltidyrc' (default) and 'full'. The 'full' parameter causes both the
28055 default options plus any options found in a perltidyrc file to be returned.
28057 =item dump_getopt_flags
28059 If the B<dump_getopt_flags> parameter is given, it must be the reference to a
28060 hash. This hash will receive all of the parameters that perltidy understands
28061 and flags that are passed to Getopt::Long. This parameter may be
28062 used alone or with the B<dump_options> flag. Perltidy will
28063 exit immediately after filling this hash. See the demo program
28064 F<perltidyrc_dump.pl> for example usage.
28066 =item dump_options_category
28068 If the B<dump_options_category> parameter is given, it must be the reference to a
28069 hash. This hash will receive a hash with keys equal to all long parameter names
28070 and values equal to the title of the corresponding section of the perltidy manual.
28071 See the demo program F<perltidyrc_dump.pl> for example usage.
28073 =item dump_abbreviations
28075 If the B<dump_abbreviations> parameter is given, it must be the reference to a
28076 hash. This hash will receive all abbreviations used by Perl::Tidy. See the
28077 demo program F<perltidyrc_dump.pl> for example usage.
28083 The following example passes perltidy a snippet as a reference
28084 to a string and receives the result back in a reference to
28089 # some messy source code to format
28090 my $source = <<'EOM';
28092 my @editors=('Emacs', 'Vi '); my $rand = rand();
28093 print "A poll of 10 random programmers gave these results:\n";
28095 my $i=int ($rand+rand());
28096 print " $editors[$i] users are from Venus" . ", " .
28097 "$editors[1-$i] users are from Mars" .
28102 # We'll pass it as ref to SCALAR and receive it in a ref to ARRAY
28104 perltidy( source => \$source, destination => \@dest );
28105 foreach (@dest) {print}
28107 =head1 Using the B<formatter> Callback Object
28109 The B<formatter> parameter is an optional callback object which allows
28110 the calling program to receive tokenized lines directly from perltidy for
28111 further specialized processing. When this parameter is used, the two
28112 formatting options which are built into perltidy (beautification or
28113 html) are ignored. The following diagram illustrates the logical flow:
28115 |-- (normal route) -> code beautification
28116 caller->perltidy->|-- (-html flag ) -> create html
28117 |-- (formatter given)-> callback to write_line
28119 This can be useful for processing perl scripts in some way. The
28120 parameter C<$formatter> in the perltidy call,
28122 formatter => $formatter,
28124 is an object created by the caller with a C<write_line> method which
28125 will accept and process tokenized lines, one line per call. Here is
28126 a simple example of a C<write_line> which merely prints the line number,
28127 the line type (as determined by perltidy), and the text of the line:
28131 # This is called from perltidy line-by-line
28133 my $line_of_tokens = shift;
28134 my $line_type = $line_of_tokens->{_line_type};
28135 my $input_line_number = $line_of_tokens->{_line_number};
28136 my $input_line = $line_of_tokens->{_line_text};
28137 print "$input_line_number:$line_type:$input_line";
28140 The complete program, B<perllinetype>, is contained in the examples section of
28141 the source distribution. As this example shows, the callback method
28142 receives a parameter B<$line_of_tokens>, which is a reference to a hash
28143 of other useful information. This example uses these hash entries:
28145 $line_of_tokens->{_line_number} - the line number (1,2,...)
28146 $line_of_tokens->{_line_text} - the text of the line
28147 $line_of_tokens->{_line_type} - the type of the line, one of:
28149 SYSTEM - system-specific code before hash-bang line
28150 CODE - line of perl code (including comments)
28151 POD_START - line starting pod, such as '=head'
28152 POD - pod documentation text
28153 POD_END - last line of pod section, '=cut'
28154 HERE - text of here-document
28155 HERE_END - last line of here-doc (target word)
28156 FORMAT - format section
28157 FORMAT_END - last line of format section, '.'
28158 DATA_START - __DATA__ line
28159 DATA - unidentified text following __DATA__
28160 END_START - __END__ line
28161 END - unidentified text following __END__
28162 ERROR - we are in big trouble, probably not a perl script
28164 Most applications will be only interested in lines of type B<CODE>. For
28165 another example, let's write a program which checks for one of the
28166 so-called I<naughty matching variables> C<&`>, C<$&>, and C<$'>, which
28167 can slow down processing. Here is a B<write_line>, from the example
28168 program B<find_naughty.pl>, which does that:
28172 # This is called back from perltidy line-by-line
28173 # We're looking for $`, $&, and $'
28174 my ( $self, $line_of_tokens ) = @_;
28176 # pull out some stuff we might need
28177 my $line_type = $line_of_tokens->{_line_type};
28178 my $input_line_number = $line_of_tokens->{_line_number};
28179 my $input_line = $line_of_tokens->{_line_text};
28180 my $rtoken_type = $line_of_tokens->{_rtoken_type};
28181 my $rtokens = $line_of_tokens->{_rtokens};
28184 # skip comments, pod, etc
28185 return if ( $line_type ne 'CODE' );
28187 # loop over tokens looking for $`, $&, and $'
28188 for ( my $j = 0 ; $j < @$rtoken_type ; $j++ ) {
28190 # we only want to examine token types 'i' (identifier)
28191 next unless $$rtoken_type[$j] eq 'i';
28193 # pull out the actual token text
28194 my $token = $$rtokens[$j];
28197 if ( $token =~ /^\$[\`\&\']$/ ) {
28199 "$input_line_number: $token\n";
28204 This example pulls out these tokenization variables from the $line_of_tokens
28207 $rtoken_type = $line_of_tokens->{_rtoken_type};
28208 $rtokens = $line_of_tokens->{_rtokens};
28210 The variable C<$rtoken_type> is a reference to an array of token type codes,
28211 and C<$rtokens> is a reference to a corresponding array of token text.
28212 These are obviously only defined for lines of type B<CODE>.
28213 Perltidy classifies tokens into types, and has a brief code for each type.
28214 You can get a complete list at any time by running perltidy from the
28217 perltidy --dump-token-types
28219 In the present example, we are only looking for tokens of type B<i>
28220 (identifiers), so the for loop skips past all other types. When an
28221 identifier is found, its actual text is checked to see if it is one
28222 being sought. If so, the above write_line prints the token and its
28225 The B<formatter> feature is relatively new in perltidy, and further
28226 documentation needs to be written to complete its description. However,
28227 several example programs have been written and can be found in the
28228 B<examples> section of the source distribution. Probably the best way
28229 to get started is to find one of the examples which most closely matches
28230 your application and start modifying it.
28232 For help with perltidy's pecular way of breaking lines into tokens, you
28233 might run, from the command line,
28235 perltidy -D filename
28237 where F<filename> is a short script of interest. This will produce
28238 F<filename.DEBUG> with interleaved lines of text and their token types.
28239 The B<-D> flag has been in perltidy from the beginning for this purpose.
28240 If you want to see the code which creates this file, it is
28241 C<write_debug_entry> in Tidy.pm.
28249 Thanks to Hugh Myers who developed the initial modular interface
28254 This man page documents Perl::Tidy version 20071205.
28259 perltidy at users.sourceforge.net
28263 The perltidy(1) man page describes all of the features of perltidy. It
28264 can be found at http://perltidy.sourceforge.net.