]> git.donarmstrong.com Git - debbugs.git/blobdiff - Debbugs/Common.pm
reuse the same null handle in globify_scalar
[debbugs.git] / Debbugs / Common.pm
index 732ac2e5bf2f6c02fcaa4b718cb2a7780e3bfc2b..b135c42cd6c263cc37e4fd997b41b7ec1be9c13d 100644 (file)
@@ -31,7 +31,8 @@ with equivalent (or better) functionality here.
 use warnings;
 use strict;
 use vars qw($VERSION $DEBUG %EXPORT_TAGS @EXPORT_OK @EXPORT);
-use base qw(Exporter);
+use Exporter qw(import);
+use v5.10;
 
 BEGIN{
      $VERSION = 1.00;
@@ -40,11 +41,14 @@ BEGIN{
      @EXPORT = ();
      %EXPORT_TAGS = (util   => [qw(getbugcomponent getbuglocation getlocationpath get_hashname),
                                qw(appendfile overwritefile buglog getparsedaddrs getmaintainers),
+                                qw(getsourcemaintainers getsourcemaintainers_reverse),
                                qw(bug_status),
                                qw(getmaintainers_reverse),
                                qw(getpseudodesc),
                                qw(package_maintainer),
                                qw(sort_versions),
+                               qw(open_compressed_file),
+                               qw(walk_bugs),
                               ],
                     misc   => [qw(make_list globify_scalar english_join checkpid),
                                qw(cleanup_eval_fail),
@@ -72,6 +76,11 @@ use Mail::Address;
 use Cwd qw(cwd);
 use Storable qw(dclone);
 use Time::HiRes qw(usleep);
+use File::Path qw(mkpath);
+use File::Basename qw(dirname);
+use MLDBM qw(DB_File Storable);
+$MLDBM::DumpMeth='portable';
+use List::AllUtils qw(natatime);
 
 use Params::Validate qw(validate_with :types);
 
@@ -239,9 +248,153 @@ sub overwritefile {
            die "Unable to rename ${file}.new to $file: $!";
 }
 
+=head2 open_compressed_file
 
+     my $fh = open_compressed_file('foo.gz') or
+          die "Unable to open compressed file: $!";
 
 
+Opens a file; if the file ends in .gz, .xz, or .bz2, the appropriate
+decompression program is forked and output from it is read.
+
+This routine by default opens the file with UTF-8 encoding; if you want some
+other encoding, specify it with the second option.
+
+=cut
+sub open_compressed_file {
+    my ($file,$encoding) = @_;
+    $encoding //= ':encoding(UTF-8)';
+    my $fh;
+    my $mode = "<$encoding";
+    my @opts;
+    if ($file =~ /\.gz$/) {
+       $mode = "-|$encoding";
+       push @opts,'gzip','-dc';
+    }
+    if ($file =~ /\.xz$/) {
+       $mode = "-|$encoding";
+       push @opts,'xz','-dc';
+    }
+    if ($file =~ /\.bz2$/) {
+       $mode = "-|$encoding";
+       push @opts,'bzip2','-dc';
+    }
+    open($fh,$mode,@opts,$file);
+    return $fh;
+}
+
+=head2 walk_bugs
+
+Walk through directories of bugs, calling a subroutine with a list of bugs
+found.
+
+C<walk_bugs(callback => sub {print map {qq($_\n)} @_},dirs => [qw(db-h)];>
+
+=over
+
+=item callback -- CODEREF of a subroutine to call with a list of bugs
+
+=item dirs -- ARRAYREF of directories to get bugs from. Like C<[qw(db-h archive)]>.
+
+=item bugs -- ARRAYREF of bugs to walk through. If both C<dirs> and C<bugs> are
+provided, both are walked through.
+
+=item bugs_per_call -- maximum number of bugs to provide to callback
+
+=item progress_bar -- optional L<Term::ProgressBar>
+
+=item bug_file -- bug file to look for (generally C<summary>)
+
+=item logging -- optional filehandle to output logging information
+
+=back
+
+=cut
+
+sub walk_bugs {
+    state $spec =
+       {dirs => {type => ARRAYREF,
+                default => [],
+               },
+       bugs => {type => ARRAYREF,
+                default => [],
+               },
+       progress_bar => {type => OBJECT|UNDEF,
+                        optional => 1,
+                       },
+       bug_file => {type => SCALAR,
+                    default => 'summary',
+                   },
+       logging => {type => HANDLE,
+                   optional => 1,
+                  },
+       callback => {type => CODEREF,
+                   },
+       bugs_per_call => {type => SCALAR,
+                         default => 1,
+                        },
+       };
+    my %param = validate_with(params => \@_,
+                             spec => $spec
+                            );
+    my @dirs = @{$param{dirs}};
+    my @initial_bugs = ();
+    if (@{$param{bugs}}) {
+       unshift @dirs,'';
+       @initial_bugs = @{$param{bugs}};
+    }
+    my $tot_dirs = @dirs;
+    my $done_dirs = 0;
+    my $avg_subfiles = 0;
+    my $completed_files = 0;
+    my $dir;
+    while ($dir = shift @dirs or defined $dir) {
+       my @list;
+       my @subdirs;
+       if (not length $dir and @initial_bugs) {
+           push @list,@initial_bugs;
+           @initial_bugs = ();
+       } else {
+           printf {$param{verbose}} "Doing dir %s ...\n", $dir
+               if defined $param{verbose};
+           opendir(my $DIR, "$dir/.") or
+               die "opendir $dir: $!";
+           @subdirs = readdir($DIR) or
+               die "Unable to readdir $dir: $!";
+           closedir($DIR) or
+               die "Unable to closedir $dir: $!";
+
+           @list = map { m/^(\d+)\.$param{bug_file}$/?($1):() } @subdirs;
+       }
+        $tot_dirs -= @dirs;
+        push @dirs, map { m/^(\d+)$/ && -d "$dir/$1"?("$dir/$1"):() } @subdirs;
+        $tot_dirs += @dirs;
+       if ($param{progress_bar}) {
+           if ($avg_subfiles == 0) {
+               $avg_subfiles = @list;
+           }
+           $param{progress_bar}->
+               target($avg_subfiles*($tot_dirs-$done_dirs)+$completed_files+@list);
+           $avg_subfiles = ($avg_subfiles * $done_dirs + @list) / ($done_dirs+1);
+           $done_dirs += 1;
+       }
+
+       my $it = natatime $param{bugs_per_call},@list;
+       while (my @bugs = $it->()) {
+           $param{callback}->(@bugs);
+           $completed_files += scalar @bugs;
+           if ($param{progress_bar}) {
+               $param{progress_bar}->update($completed_files) if $param{progress_bar};
+           }
+           if ($completed_files % 100 == 0 and
+               defined $param{verbose}) {
+               print {$param{verbose}} "Up to $completed_files bugs...\n"
+           }
+        }
+    }
+    $param{progress_bar}->remove() if $param{progress_bar};
+}
+
 
 =head2 getparsedaddrs
 
@@ -299,6 +452,36 @@ sub getmaintainers_reverse{
      return $_maintainer_rev;
 }
 
+=head2 getsourcemaintainers
+
+     my $maintainer = getsourcemaintainers()->{debbugs}
+
+Returns a hashref of src_package => maintainer pairs.
+
+=cut
+
+our $_source_maintainer = undef;
+our $_source_maintainer_rev = undef;
+sub getsourcemaintainers {
+    return $_source_maintainer if defined $_source_maintainer;
+    package_maintainer(rehash => 1);
+    return $_source_maintainer;
+}
+
+=head2 getsourcemaintainers_reverse
+
+     my @src_packages = @{getsourcemaintainers_reverse->{'don@debian.org'}||[]};
+
+Returns a hashref of maintainer => [qw(list of source packages)] pairs.
+
+=cut
+
+sub getsourcemaintainers_reverse{
+     return $_source_maintainer_rev if defined $_source_maintainer_rev;
+     package_maintainer(rehash => 1);
+     return $_source_maintainer_rev;
+}
+
 =head2 package_maintainer
 
      my @s = package_maintainer(source => [qw(foo bar baz)],
@@ -314,18 +497,19 @@ maintainers for, defaults to the empty arrayref.
 maintainers for; automatically returns source package maintainer if
 the package name starts with 'src:', defaults to the empty arrayref.
 
-=item reverse -- whether to return the source/binary packages a
-maintainer maintains instead
+=item maintainer -- scalar or arrayref of maintainers to return source packages
+for. If given, binary and source cannot be given.
 
 =item rehash -- whether to reread the maintainer and source maintainer
 files; defaults to 0
 
+=item schema -- Debbugs::DB schema. If set, uses the database for maintainer
+information.
+
 =back
 
 =cut
 
-our $_source_maintainer = undef;
-our $_source_maintainer_rev = undef;
 sub package_maintainer {
     my %param = validate_with(params => \@_,
                              spec   => {source => {type => SCALAR|ARRAYREF,
@@ -343,6 +527,9 @@ sub package_maintainer {
                                         reverse => {type => BOOLEAN,
                                                     default => 0,
                                                    },
+                                        schema => {type => OBJECT,
+                                                   optional => 1,
+                                                  }
                                        },
                             );
     my @binary = make_list($param{binary});
@@ -351,6 +538,53 @@ sub package_maintainer {
     if ((@binary or @source) and @maintainers) {
        croak "It is nonsensical to pass both maintainers and source or binary";
     }
+    if (@binary) {
+       @source = grep {/^src:/} @binary;
+       @binary = grep {!/^src:/} @binary;
+    }
+    # remove leading src: from source package names
+    s/^src:// foreach @source;
+    if ($param{schema}) {
+       my $s = $param{schema};
+       if (@maintainers) {
+           my $m_rs = $s->resultset('SrcPkg')->
+               search({'correspondent.addr' => [@maintainers]},
+                     {join => {src_vers =>
+                              {maintainer =>
+                               'correspondent'},
+                              },
+                      columns => ['pkg'],
+                      group_by => [qw(me.pkg)],
+                      });
+           return $m_rs->get_column('pkg')->all();
+       } elsif (@binary or @source) {
+           my $rs = $s->resultset('Maintainer');
+           if (@binary) {
+               $rs =
+                   $rs->search({'bin_pkg.pkg' => [@binary]},
+                              {join => {src_vers =>
+                                       {bin_vers => 'bin_pkg'},
+                                       },
+                               columns => ['name'],
+                               group_by => [qw(me.name)],
+                              }
+                              );
+           }
+           if (@source) {
+               $rs =
+                   $rs->search({'src_pkg.pkg' => [@source]},
+                              {join => {src_vers =>
+                                        'src_pkg',
+                                       },
+                               columns => ['name'],
+                               group_by => [qw(me.name)],
+                              }
+                              );
+           }
+           return $rs->get_column('name')->all();
+       }
+       return ();
+    }
     if ($param{rehash}) {
        $_source_maintainer = undef;
        $_source_maintainer_rev = undef;
@@ -361,37 +595,63 @@ sub package_maintainer {
        not defined $_source_maintainer_rev) {
        $_source_maintainer = {};
        $_source_maintainer_rev = {};
-       for my $fn (@config{('source_maintainer_file',
-                            'source_maintainer_file_override',
-                            'pseudo_maint_file')}) {
-           next unless defined $fn;
-           if (not -e $fn) {
-               warn "Missing source maintainer file '$fn'";
-               next;
+       if (-e $config{spool_dir}.'/source_maintainers.idx' and
+           -e $config{spool_dir}.'/source_maintainers_reverse.idx'
+          ) {
+           tie %{$_source_maintainer},
+               MLDBM => $config{spool_dir}.'/source_maintainers.idx',
+               O_RDONLY or
+               die "Unable to tie source maintainers: $!";
+           tie %{$_source_maintainer_rev},
+               MLDBM => $config{spool_dir}.'/source_maintainers_reverse.idx',
+               O_RDONLY or
+               die "Unable to tie source maintainers reverse: $!";
+       } else {
+           for my $fn (@config{('source_maintainer_file',
+                                'source_maintainer_file_override',
+                                'pseudo_maint_file')}) {
+               next unless defined $fn and length $fn;
+               if (not -e $fn) {
+                   warn "Missing source maintainer file '$fn'";
+                   next;
+               }
+               __add_to_hash($fn,$_source_maintainer,
+                             $_source_maintainer_rev);
            }
-           __add_to_hash($fn,$_source_maintainer,
-                         $_source_maintainer_rev);
        }
     }
     if (not defined $_maintainer or
        not defined $_maintainer_rev) {
        $_maintainer = {};
        $_maintainer_rev = {};
-       for my $fn (@config{('maintainer_file',
-                            'maintainer_file_override',
-                            'pseudo_maint_file')}) {
-           next unless defined $fn;
-           if (not -e $fn) {
-               warn "Missing maintainer file '$fn'";
-               next;
-           }
-           __add_to_hash($fn,$_maintainer,
+       if (-e $config{spool_dir}.'/maintainers.idx' and
+           -e $config{spool_dir}.'/maintainers_reverse.idx'
+          ) {
+           tie %{$_maintainer},
+               MLDBM => $config{spool_dir}.'/binary_maintainers.idx',
+               O_RDONLY or
+               die "Unable to tie binary maintainers: $!";
+           tie %{$_maintainer_rev},
+               MLDBM => $config{spool_dir}.'/binary_maintainers_reverse.idx',
+               O_RDONLY or
+               die "Unable to binary maintainers reverse: $!";
+       } else {
+           for my $fn (@config{('maintainer_file',
+                                'maintainer_file_override',
+                                'pseudo_maint_file')}) {
+               next unless defined $fn and length $fn;
+               if (not -e $fn) {
+                   warn "Missing maintainer file '$fn'";
+                   next;
+               }
+               __add_to_hash($fn,$_maintainer,
                              $_maintainer_rev);
+           }
        }
     }
     my @return;
     for my $binary (@binary) {
-       if (not $param{reverse} and $binary =~ /^src:/) {
+       if ($binary =~ /^src:/) {
            push @source,$binary;
            next;
        }
@@ -431,12 +691,12 @@ sub __add_to_hash {
     }
     $type //= 'address';
     my $fh = IO::File->new($fn,'r') or
-       die "Unable to open $fn for reading: $!";
+       croak "Unable to open $fn for reading: $!";
     binmode($fh,':encoding(UTF-8)');
     while (<$fh>) {
        chomp;
-       next unless m/^(\S+)\s+(\S.*\S)\s*$/;
-       my ($key,$value)=($1,$2);
+        next unless m/^(\S+)\s+(\S.*\S)\s*$/;
+        my ($key,$value)=($1,$2);
        $key = lc $key;
        $forward->{$key}= $value;
        if (defined $reverse) {
@@ -471,7 +731,8 @@ sub getpseudodesc {
     return $_pseudodesc if defined $_pseudodesc;
     $_pseudodesc = {};
     __add_to_hash($config{pseudo_desc_file},$_pseudodesc) if
-       defined $config{pseudo_desc_file};
+       defined $config{pseudo_desc_file} and
+       length $config{pseudo_desc_file};
     return $_pseudodesc;
 }
 
@@ -728,6 +989,7 @@ sub lockpid {
          unlink $pidfile or
               die "Unable to unlink stale pidfile $pidfile $!";
      }
+     mkpath(dirname($pidfile));
      my $pidfh = IO::File->new($pidfile,O_CREAT|O_EXCL|O_WRONLY) or
          die "Unable to open $pidfile for writing: $!";
      print {$pidfh} $$ or die "Unable to write to $pidfile $!";
@@ -878,6 +1140,8 @@ before. This appears to be a bug in the underlying modules.
 
 =cut
 
+our $_NULL_HANDLE;
+
 sub globify_scalar {
      my ($scalar) = @_;
      my $handle;
@@ -902,8 +1166,15 @@ sub globify_scalar {
          else {
               carp "Given a non-scalar reference, non-glob to globify_scalar; returning /dev/null handle";
          }
+      }
+     if (not defined $_NULL_HANDLE or
+        not $_NULL_HANDLE->opened()
+       ) {
+        $_NULL_HANDLE =
+            IO::File->new('/dev/null','>:encoding(UTF-8)') or
+                die "Unable to open /dev/null for writing: $!";
      }
-     return IO::File->new('/dev/null','>:encoding(UTF-8)');
+     return $_NULL_HANDLE;
 }
 
 =head2 cleanup_eval_fail()
@@ -937,7 +1208,7 @@ sub cleanup_eval_fail {
     # ditch the "at foo/bar/baz.pm line 5"
     $error =~ s/\sat\s\S+\sline\s\d+//;
     # ditch croak messages
-    $error =~ s/^\t+.+\n?//g;
+    $error =~ s/^\t+.+\n?//mg;
     # ditch trailing multiple periods in case there was a cascade of
     # die messages.
     $error =~ s/\.+$/\./;