1 # This module is part of debbugs, and is released
2 # under the terms of the GPL version 2, or any later
3 # version at your option.
4 # See the file README and COPYING for more information.
6 # Copyright 2007 by Don Armstrong <don@donarmstrong.com>.
12 Debbugs::Bugs -- Bug selection routines for debbugs
16 use Debbugs::Bugs qw(get_bugs);
21 This module is a replacement for all of the various methods of
22 selecting different types of bugs.
24 It implements a single function, get_bugs, which defines the master
25 interface for selecting bugs.
27 It attempts to use subsidiary functions to actually do the selection,
28 in the order specified in the configuration files. [Unless you're
29 insane, they should be in order from fastest (and often most
30 incomplete) to slowest (and most complete).]
40 use vars qw($VERSION $DEBUG %EXPORT_TAGS @EXPORT_OK @EXPORT);
41 use Exporter qw(import);
45 $DEBUG = 0 unless defined $DEBUG;
49 @EXPORT_OK = (qw(get_bugs count_bugs newest_bug bug_filter));
50 $EXPORT_TAGS{all} = [@EXPORT_OK];
53 use Debbugs::Config qw(:config);
54 use Params::Validate qw(validate_with :types);
56 use Debbugs::Status qw(splitpackages get_bug_status);
57 use Debbugs::Packages qw(getsrcpkgs getpkgsrc);
58 use Debbugs::Common qw(getparsedaddrs package_maintainer getmaintainers make_list hash_slice);
59 use Fcntl qw(O_RDONLY);
60 use MLDBM qw(DB_File Storable);
61 use List::AllUtils qw(first);
70 The following parameters can either be a single scalar or a reference
71 to an array. The parameters are ANDed together, and the elements of
72 arrayrefs are a parameter are ORed. Future versions of this may allow
73 for limited regular expressions, and/or more complex expressions.
77 =item package -- name of the binary package
79 =item src -- name of the source package
81 =item maint -- address of the maintainer
83 =item submitter -- address of the submitter
85 =item severity -- severity of the bug
87 =item status -- status of the bug
91 =item owner -- owner of the bug
93 =item correspondent -- address of someone who sent mail to the log
95 =item affects -- bugs which affect this package
97 =item dist -- distribution (I don't know about this one yet)
99 =item bugs -- list of bugs to search within
101 =item function -- see description below
105 =head3 Special options
107 The following options are special options used to modulate how the
108 searches are performed.
112 =item archive -- whether to search archived bugs or normal bugs;
113 defaults to false. As a special case, if archive is 'both', but
114 archived and unarchived bugs are returned.
116 =item usertags -- set of usertags and the bugs they are applied to
121 =head3 Subsidiary routines
123 All subsidiary routines get passed exactly the same set of options as
124 get_bugs. If for some reason they are unable to handle the options
125 passed (for example, they don't have the right type of index for the
126 type of selection) they should die as early as possible. [Using
127 Params::Validate and/or die when files don't exist makes this fairly
130 This function will then immediately move on to the next subroutine,
131 giving it the same arguments.
135 This option allows you to provide an arbitrary function which will be
136 given the information in the index.db file. This will be super, super
137 slow, so only do this if there's no other way to write the search.
139 You'll be given a list (which you can turn into a hash) like the
142 (pkg => ['a','b'], # may be a scalar (most common)
145 submitter => 'boo@baz.com',
146 severity => 'serious',
147 tags => ['a','b','c'], # may be an empty arrayref
150 The function should return 1 if the bug should be included; 0 if the
155 my $_non_search_key_regex = qr/^(bugs|archive|usertags|schema)$/;
157 my %_get_bugs_common_options =
158 (package => {type => SCALAR|ARRAYREF,
161 src => {type => SCALAR|ARRAYREF,
164 maint => {type => SCALAR|ARRAYREF,
167 submitter => {type => SCALAR|ARRAYREF,
170 severity => {type => SCALAR|ARRAYREF,
173 status => {type => SCALAR|ARRAYREF,
176 tag => {type => SCALAR|ARRAYREF,
179 owner => {type => SCALAR|ARRAYREF,
182 dist => {type => SCALAR|ARRAYREF,
185 correspondent => {type => SCALAR|ARRAYREF,
188 affects => {type => SCALAR|ARRAYREF,
191 function => {type => CODEREF,
194 bugs => {type => SCALAR|ARRAYREF,
197 archive => {type => BOOLEAN|SCALAR,
200 usertags => {type => HASHREF,
203 schema => {type => OBJECT,
209 my $_get_bugs_options = {%_get_bugs_common_options};
211 my %param = validate_with(params => \@_,
212 spec => $_get_bugs_options,
216 my %options = %param;
218 if ($options{archive} eq 'both') {
219 push @bugs, get_bugs(%options,archive=>0);
220 push @bugs, get_bugs(%options,archive=>1);
222 @bugs{@bugs} = @bugs;
225 # A configuration option will set an array that we'll use here instead.
226 for my $routine (qw(Debbugs::Bugs::get_bugs_by_db Debbugs::Bugs::get_bugs_by_idx Debbugs::Bugs::get_bugs_flatfile)) {
227 my ($package) = $routine =~ m/^(.+)\:\:/;
228 eval "use $package;";
230 # We output errors here because using an invalid function
231 # in the configuration file isn't something that should
233 warn "use $package failed with $@";
236 @bugs = eval "${routine}(\%options)";
239 # We don't output errors here, because failure here
240 # via die may be a perfectly normal thing.
241 print STDERR "$@" if $DEBUG;
246 # If no one succeeded, die
255 count_bugs(function => sub {...})
257 Uses a subroutine to classify bugs into categories and return the
258 number of bugs which fall into those categories
263 my %param = validate_with(params => \@_,
264 spec => {function => {type => CODEREF,
266 archive => {type => BOOLEAN,
272 if ($param{archive}) {
273 $flatfile = IO::File->new("$config{spool_dir}/index.archive", 'r')
274 or die "Unable to open $config{spool_dir}/index.archive for reading: $!";
277 $flatfile = IO::File->new("$config{spool_dir}/index.db", 'r')
278 or die "Unable to open $config{spool_dir}/index.db for reading: $!";
282 if (m/^(\S+)\s+(\d+)\s+(\d+)\s+(\S+)\s+\[\s*([^]]*)\s*\]\s+(\w+)\s+(.*)$/) {
283 my @x = $param{function}->(pkg => $1,
291 $count{$_}++ foreach @x;
300 my $bug = newest_bug();
302 Returns the bug number of the newest bug, which is nextnumber-1.
307 my $nn_fh = IO::File->new("$config{spool_dir}/nextnumber",'r')
308 or die "Unable to open $config{spool_dir}nextnumber for reading: $!";
310 my $next_number = <$nn_fh>;
313 return $next_number-1;
320 Allows filtering bugs on commonly used criteria
327 my %param = validate_with(params => \@_,
328 spec => {bug => {type => ARRAYREF|SCALAR,
331 status => {type => HASHREF|ARRAYREF,
334 seen_merged => {type => HASHREF,
337 repeat_merged => {type => BOOLEAN,
340 include => {type => HASHREF,
343 exclude => {type => HASHREF,
346 min_days => {type => SCALAR,
349 max_days => {type => SCALAR,
354 if (exists $param{repeat_merged} and
355 not $param{repeat_merged} and
356 not defined $param{seen_merged}) {
357 croak "repeat_merged false requires seen_merged to be passed";
359 if (not exists $param{bug} and not exists $param{status}) {
360 croak "one of bug or status must be passed";
363 if (not exists $param{status}) {
364 my $location = getbuglocation($param{bug}, 'summary');
365 return 0 if not defined $location or not length $location;
366 $param{status} = readbug( $param{bug}, $location );
367 return 0 if not defined $param{status};
370 if (exists $param{include}) {
371 return 1 if (!__bug_matches($param{include}, $param{status}));
373 if (exists $param{exclude}) {
374 return 1 if (__bug_matches($param{exclude}, $param{status}));
376 if (exists $param{repeat_merged} and not $param{repeat_merged}) {
377 my @merged = sort {$a<=>$b} $param{bug}, split(/ /, $param{status}{mergedwith});
378 return 1 if first {defined $_} @{$param{seen_merged}}{@merged};
379 @{$param{seen_merged}}{@merged} = (1) x @merged;
381 my $daysold = int((time - $param{status}{date}) / 86400); # seconds to days
382 if (exists $param{min_days}) {
383 return 1 unless $param{min_days} <= $daysold;
385 if (exists $param{max_days}) {
386 return 1 unless $param{max_days} == -1 or
387 $param{max_days} >= $daysold;
393 =head2 get_bugs_by_idx
395 This routine uses the by-$index.idx indicies to try to speed up
402 my $_get_bugs_by_idx_options =
403 {hash_slice(%_get_bugs_common_options,
404 (qw(package submitter severity tag archive),
405 qw(owner src maint bugs correspondent),
406 qw(affects usertags))
410 my %param = validate_with(params => \@_,
411 spec => $_get_bugs_by_idx_options
415 # If we're given an empty maint (unmaintained packages), we can't
416 # handle it, so bail out here
417 for my $maint (make_list(exists $param{maint}?$param{maint}:[])) {
418 if (defined $maint and $maint eq '') {
419 die "Can't handle empty maint (unmaintained packages) in get_bugs_by_idx";
423 # We handle src packages, maint and maintenc by mapping to the
424 # appropriate binary packages, then removing all packages which
425 # don't match all queries
426 my @packages = __handle_pkg_src_and_maint(map {exists $param{$_}?($_,$param{$_}):()}
427 qw(package src maint)
429 if (exists $param{package} or
430 exists $param{src} or
431 exists $param{maint}) {
432 delete @param{qw(maint src)};
433 $param{package} = [@packages];
435 my $keys = grep {$_ !~ $_non_search_key_regex} keys(%param);
436 die "Need at least 1 key to search by" unless $keys;
437 my $arc = $param{archive} ? '-arc':'';
439 for my $key (grep {$_ !~ $_non_search_key_regex} keys %param) {
441 $index = 'submitter-email' if $key eq 'submitter';
442 $index = "$config{spool_dir}/by-${index}${arc}.idx";
443 tie(%idx, MLDBM => $index, O_RDONLY)
444 or die "Unable to open $index: $!";
445 my %bug_matching = ();
446 for my $search (make_list($param{$key})) {
447 for my $bug (keys %{$idx{$search}||{}}) {
448 next if $bug_matching{$bug};
449 # increment the number of searches that this bug matched
451 $bug_matching{$bug}=1;
453 if ($search ne lc($search)) {
454 for my $bug (keys %{$idx{lc($search)}||{}}) {
455 next if $bug_matching{$bug};
456 # increment the number of searches that this bug matched
458 $bug_matching{$bug}=1;
462 if ($key eq 'tag' and exists $param{usertags}) {
463 for my $bug (make_list(grep {defined $_ } @{$param{usertags}}{make_list($param{tag})})) {
464 next if $bug_matching{$bug};
466 $bug_matching{$bug}=1;
469 untie %idx or die 'Unable to untie %idx';
473 for my $bug (make_list($param{bugs})) {
477 # Throw out results that do not match all of the search specifications
478 return map {$keys <= $bugs{$_}?($_):()} keys %bugs;
482 =head2 get_bugs_by_db
484 This routine uses the database to try to speed up
490 my $_get_bugs_by_db_options =
491 {hash_slice(%_get_bugs_common_options,
492 (qw(package submitter severity tag archive),
493 qw(owner src maint bugs correspondent),
494 qw(affects usertags))
496 schema => {type => OBJECT,
500 my %param = validate_with(params => \@_,
501 spec => $_get_bugs_by_db_options,
505 my $keys = grep {$_ !~ $_non_search_key_regex} keys(%param);
506 die "Need at least 1 key to search by" unless $keys;
507 my $rs = $param{schema}->resultset('Bug');
508 if (exists $param{package}) {
509 $rs = $rs->search({-or => {map 'bin_package.'}})
511 if (exists $param{severity}) {
512 $rs = $rs->search({-or => {map {('severity.severity' => $_)}
513 make_list($param{severity})},
515 {join => 'severity'},
518 for my $key (qw(owner submitter done)) {
519 if (exists $param{$key}) {
520 $rs = $rs->search({-or => {map {("${key}.addr" => $_)}
521 make_list($param{$key})},
527 if (exists $param{correspondent}) {
528 $rs = $rs->search({-or => {map {('message_correspondents.addr' => $_)}
529 make_list($param{correspondent})},
531 {join => {correspondent =>
533 {message => 'message_correspondents'}}}},
536 if (exists $param{affects}) {
537 $rs = $rs->search({-or => {map {('bin_pkg.pkg' => $_,
540 make_list($param{affects}),
543 {join => [{bug_affects_binpackages => 'bin_pkg'},
544 {bug_affects_srcpackages => 'src_pkg'},
549 if (exists $param{package}) {
550 $rs = $rs->search({-or => {map {('bin_pkg.pkg' => $_)}
551 make_list($param{package})},
553 {join => {bug_binpackages => 'bin_pkg'}});
555 if (exists $param{maintainer}) {
556 $rs = $rs->search({-or => {map {(correspondent => $_ eq '' ? undef : $_,
557 correspondent2 => $_ eq '' ? undef : $_,
559 make_list($param{maintainer})
562 {join => {bug_affects_binpackage =>
566 {maintainer => 'correspondent'}
568 {bug_affects_srcpackage =>
571 {maintainer => 'correspondent'}
576 if (exists $param{src}) {
577 $rs = $rs->search({-or => {map {('src_pkg.pkg' => $_)}
578 make_list($param{src})},
580 {join => {bug_srcpackages => 'src_pkg'}});
582 # tags are very odd, because we must handle usertags.
583 if (exists $param{tag}) {
584 # bugs from usertags which matter
585 my %bugs_matching_usertags;
586 for my $bug (make_list(grep {defined $_ }
587 @{$param{usertags}}{make_list($param{tag})})) {
588 $bugs_matching_usertags{$bug} = 1;
590 # we want all bugs which either match the tag name given in
591 # param, or have a usertag set which matches one of the tag
592 # names given in param.
593 $rs = $rs->search({-or => {map {('tag.tag' => $_)}
594 make_list($param{tag}),
595 map {('me.id' => $_)}
596 keys %bugs_matching_usertags
599 {join => {bug_tags => 'tag'}});
601 if (exists $param{bugs}) {
602 $rs = $rs->search({-or => {map {('me.id' => $_)}
603 make_list($param{bugs})}
607 if (defined $param{archive} and $param{archive} ne 'both') {
608 $rs = $rs->search({'me.archived' => $param{archive}});
610 return $rs->get_column('id')->all();
614 =head2 get_bugs_flatfile
616 This is the fallback search routine. It should be able to complete all
617 searches. [Or at least, that's the idea.]
621 my $_get_bugs_flatfile_options =
622 {hash_slice(%_get_bugs_common_options,
623 map {$_ eq 'dist'?():($_)} keys %_get_bugs_common_options
627 sub get_bugs_flatfile{
628 my %param = validate_with(params => \@_,
629 spec => $_get_bugs_flatfile_options
632 if ($param{archive}) {
633 $flatfile = IO::File->new("$config{spool_dir}/index.archive", 'r')
634 or die "Unable to open $config{spool_dir}/index.archive for reading: $!";
637 $flatfile = IO::File->new("$config{spool_dir}/index.db", 'r')
638 or die "Unable to open $config{spool_dir}/index.db for reading: $!";
641 if (exists $param{tag} and exists $param{usertags}) {
642 # This complex slice makes a hash with the bugs which have the
643 # usertags passed in $param{tag} set.
644 @usertag_bugs{make_list(@{$param{usertags}}{make_list($param{tag})})
645 } = (1) x make_list(@{$param{usertags}}{make_list($param{tag})});
647 my $unmaintained_packages = 0;
648 # unmaintained packages is a special case
649 my @maints = make_list(exists $param{maint}?$param{maint}:[]);
651 for my $maint (@maints) {
652 if (defined $maint and $maint eq '' and not $unmaintained_packages) {
653 $unmaintained_packages = 1;
654 our %maintainers = %{getmaintainers()};
655 $param{function} = [(exists $param{function}?
656 (ref $param{function}?@{$param{function}}:$param{function}):()),
658 foreach my $try (make_list($d{"pkg"})) {
659 next unless length $try;
660 ($try) = $try =~ m/^(?:src:)?(.+)/;
661 return 1 if not exists $maintainers{$try};
667 elsif (defined $maint and $maint ne '') {
668 push @{$param{maint}},$maint;
671 # We handle src packages, maint and maintenc by mapping to the
672 # appropriate binary packages, then removing all packages which
673 # don't match all queries
674 my @packages = __handle_pkg_src_and_maint(map {exists $param{$_}?($_,$param{$_}):()}
675 qw(package src maint)
677 if (exists $param{package} or
678 exists $param{src} or
679 exists $param{maint}) {
680 delete @param{qw(maint src)};
681 $param{package} = [@packages] if @packages;
685 if (exists $param{bugs}) {
686 $bugs{$_} = 1 for make_list($param{bugs});
689 # These queries have to be handled by get_bugs_by_idx
690 if (exists $param{owner}
691 or exists $param{correspondent}
692 or exists $param{affects}) {
693 $bugs{$_} = 1 for get_bugs_by_idx(map {exists $param{$_}?($_,$param{$_}):()}
694 qw(owner correspondent affects),
699 BUG: while (<$flatfile>) {
700 next unless m/^(\S+)\s+(\d+)\s+(\d+)\s+(\S+)\s+\[\s*(.*)\s*\]\s+(\w+)\s+(.*)$/;
701 my ($pkg,$bug,$time,$status,$submitter,$severity,$tags) = ($1,$2,$3,$4,$5,$6,$7);
702 next if $grep_bugs and not exists $bugs{$bug};
703 if (exists $param{package}) {
704 my @packages = splitpackages($pkg);
705 next unless grep { my $pkg_list = $_;
706 grep {$pkg_list eq $_} make_list($param{package})
709 if (exists $param{src}) {
710 my @src_packages = map { getsrcpkgs($_)} make_list($param{src});
711 my @packages = splitpackages($pkg);
712 next unless grep { my $pkg_list = $_;
713 grep {$pkg_list eq $_} @packages
716 if (exists $param{submitter}) {
717 my @p_addrs = map {lc($_->address)}
718 map {getparsedaddrs($_)}
719 make_list($param{submitter});
720 my @f_addrs = map {$_->address}
721 getparsedaddrs($submitter||'');
722 next unless grep { my $f_addr = $_;
723 grep {$f_addr eq $_} @p_addrs
726 next if exists $param{severity} and not grep {$severity eq $_} make_list($param{severity});
727 next if exists $param{status} and not grep {$status eq $_} make_list($param{status});
728 if (exists $param{tag}) {
730 # either a normal tag, or a usertag must be set
731 $bug_ok = 1 if exists $param{usertags} and $usertag_bugs{$bug};
732 my @bug_tags = split ' ', $tags;
733 $bug_ok = 1 if grep {my $bug_tag = $_;
734 grep {$bug_tag eq $_} make_list($param{tag});
738 # We do this last, because a function may be slow...
739 if (exists $param{function}) {
740 my @bug_tags = split ' ', $tags;
741 my @packages = splitpackages($pkg);
742 my $package = (@packages > 1)?\@packages:$packages[0];
743 for my $function (make_list($param{function})) {
745 $function->(pkg => $package,
748 submitter => $submitter,
749 severity => $severity,
759 =head1 PRIVATE FUNCTIONS
761 =head2 __handle_pkg_src_and_maint
763 my @packages = __handle_pkg_src_and_maint(map {exists $param{$_}?($_,$param{$_}):()}
764 qw(package src maint)
767 Turn package/src/maint into a list of packages
771 sub __handle_pkg_src_and_maint{
772 my %param = validate_with(params => \@_,
773 spec => {package => {type => SCALAR|ARRAYREF,
776 src => {type => SCALAR|ARRAYREF,
779 maint => {type => SCALAR|ARRAYREF,
787 @packages = make_list($param{package}) if exists $param{package};
788 my $package_keys = @packages?1:0;
790 @packages{@packages} = (1) x @packages;
791 if (exists $param{src}) {
792 # We only want to increment the number of keys if there is
795 # in case there are binaries with the same name as the
798 for my $package ((map {getsrcpkgs($_)} make_list($param{src}))) {
799 $packages{$package}++ unless exists $_temp_p{$package};
800 $_temp_p{$package} = 1;
803 for my $package (make_list($param{src})) {
804 $packages{"src:$package"}++ unless exists $_temp_p{"src:$package"};
805 $_temp_p{"src:$package"} = 1;
807 # As a temporary hack, we will also include $param{src}
808 # in this list for packages passed which do not have a
809 # corresponding binary package
810 if (not exists getpkgsrc()->{$package}) {
811 $packages{$package}++ unless exists $_temp_p{$package};
812 $_temp_p{$package} = 1;
815 $package_keys += $key_inc;
817 if (exists $param{maint}) {
820 for my $package (package_maintainer(maintainer=>$param{maint})) {
821 $packages{$package}++ unless exists $_temp_p{$package};
822 $_temp_p{$package} = 1;
825 $package_keys += $key_inc;
827 return grep {$packages{$_} >= $package_keys} keys %packages;
831 'subject' => \&__contains_field_match,
833 my ($field, $values, $status) = @_;
834 my %values = map {$_=>1} @$values;
835 foreach my $t (split /\s+/, $status->{$field}) {
836 return 1 if (defined $values{$t});
840 'severity' => \&__exact_field_match,
841 'pending' => \&__exact_field_match,
842 'package' => \&__exact_field_match,
843 'originator' => \&__contains_field_match,
844 'forwarded' => \&__contains_field_match,
845 'owner' => \&__contains_field_match,
849 my ($hash, $status) = @_;
850 foreach my $key( keys( %$hash ) ) {
851 my $value = $hash->{$key};
852 next unless exists $field_match{$key};
853 my $sub = $field_match{$key};
854 if (not defined $sub) {
855 die "No defined subroutine for key: $key";
857 return 1 if ($sub->($key, $value, $status));
862 sub __exact_field_match {
863 my ($field, $values, $status) = @_;
864 my @values = @$values;
865 my @ret = grep {$_ eq $status->{$field} } @values;
869 sub __contains_field_match {
870 my ($field, $values, $status) = @_;
871 foreach my $data (@$values) {
872 return 1 if (index($status->{$field}, $data) > -1);