Merge branch 'js/gitweb-path-info-unquote' into maint-1.7.11
authorJunio C Hamano <gitster@pobox.com>
Mon, 10 Sep 2012 22:23:46 +0000 (15:23 -0700)
committerJunio C Hamano <gitster@pobox.com>
Mon, 10 Sep 2012 22:23:46 +0000 (15:23 -0700)
"gitweb" when used with PATH_INFO failed to notice directories with
SP (and other characters that need URL-style quoting) in them.

* js/gitweb-path-info-unquote:
gitweb: URL-decode $my_url/$my_uri when stripping PATH_INFO

1  2 
gitweb/gitweb.perl
diff --combined gitweb/gitweb.perl
index 55e0e9ea38b3080e32467b6faf56f40d45386b96,126d3ef4bd4137e51ed0c4e58445937aa795b966..a40ed0ceb0756d85e26afaabf6d46b7c9105f576
@@@ -54,6 -54,11 +54,11 @@@ sub evaluate_uri 
        # to build the base URL ourselves:
        our $path_info = decode_utf8($ENV{"PATH_INFO"});
        if ($path_info) {
+               # $path_info has already been URL-decoded by the web server, but
+               # $my_url and $my_uri have not. URL-decode them so we can properly
+               # strip $path_info.
+               $my_url = unescape($my_url);
+               $my_uri = unescape($my_uri);
                if ($my_url =~ s,\Q$path_info\E$,, &&
                    $my_uri =~ s,\Q$path_info\E$,, &&
                    defined $ENV{'SCRIPT_NAME'}) {
@@@ -133,12 -138,6 +138,12 @@@ our $default_projects_order = "project"
  # (only effective if this variable evaluates to true)
  our $export_ok = "++GITWEB_EXPORT_OK++";
  
 +# don't generate age column on the projects list page
 +our $omit_age_column = 0;
 +
 +# don't generate information about owners of repositories
 +our $omit_owner=0;
 +
  # show repository only if this subroutine returns true
  # when given the path to the project, for example:
  #    sub { return -e "$_[0]/git-daemon-export-ok"; }
@@@ -1738,29 -1737,20 +1743,29 @@@ sub chop_and_escape_str 
  # '<span class="mark">foo</span>bar'
  sub esc_html_hl_regions {
        my ($str, $css_class, @sel) = @_;
 -      return esc_html($str) unless @sel;
 +      my %opts = grep { ref($_) ne 'ARRAY' } @sel;
 +      @sel     = grep { ref($_) eq 'ARRAY' } @sel;
 +      return esc_html($str, %opts) unless @sel;
  
        my $out = '';
        my $pos = 0;
  
        for my $s (@sel) {
 -              $out .= esc_html(substr($str, $pos, $s->[0] - $pos))
 -                      if ($s->[0] - $pos > 0);
 -              $out .= $cgi->span({-class => $css_class},
 -                                 esc_html(substr($str, $s->[0], $s->[1] - $s->[0])));
 +              my ($begin, $end) = @$s;
 +
 +              # Don't create empty <span> elements.
 +              next if $end <= $begin;
 +
 +              my $escaped = esc_html(substr($str, $begin, $end - $begin),
 +                                     %opts);
  
 -              $pos = $s->[1];
 +              $out .= esc_html(substr($str, $pos, $begin - $pos), %opts)
 +                      if ($begin - $pos > 0);
 +              $out .= $cgi->span({-class => $css_class}, $escaped);
 +
 +              $pos = $end;
        }
 -      $out .= esc_html(substr($str, $pos))
 +      $out .= esc_html(substr($str, $pos), %opts)
                if ($pos < length($str));
  
        return $out;
@@@ -2436,32 -2426,26 +2441,32 @@@ sub format_cc_diff_chunk_header 
  }
  
  # process patch (diff) line (not to be used for diff headers),
 -# returning class and HTML-formatted (but not wrapped) line
 -sub process_diff_line {
 -      my $line = shift;
 -      my ($from, $to) = @_;
 -
 -      my $diff_class = diff_line_class($line, $from, $to);
 -
 -      chomp $line;
 -      $line = untabify($line);
 +# returning HTML-formatted (but not wrapped) line.
 +# If the line is passed as a reference, it is treated as HTML and not
 +# esc_html()'ed.
 +sub format_diff_line {
 +      my ($line, $diff_class, $from, $to) = @_;
 +
 +      if (ref($line)) {
 +              $line = $$line;
 +      } else {
 +              chomp $line;
 +              $line = untabify($line);
  
 -      if ($from && $to && $line =~ m/^\@{2} /) {
 -              $line = format_unidiff_chunk_header($line, $from, $to);
 -              return $diff_class, $line;
 +              if ($from && $to && $line =~ m/^\@{2} /) {
 +                      $line = format_unidiff_chunk_header($line, $from, $to);
 +              } elsif ($from && $to && $line =~ m/^\@{3}/) {
 +                      $line = format_cc_diff_chunk_header($line, $from, $to);
 +              } else {
 +                      $line = esc_html($line, -nbsp=>1);
 +              }
 +      }
  
 -      } elsif ($from && $to && $line =~ m/^\@{3}/) {
 -              $line = format_cc_diff_chunk_header($line, $from, $to);
 -              return $diff_class, $line;
 +      my $diff_classes = "diff";
 +      $diff_classes .= " $diff_class" if ($diff_class);
 +      $line = "<div class=\"$diff_classes\">$line</div>\n";
  
 -      }
 -      return $diff_class, esc_html($line, -nbsp=>1);
 +      return $line;
  }
  
  # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
@@@ -3018,11 -3002,9 +3023,11 @@@ sub git_get_projects_list 
                        }
                        if (check_export_ok("$projectroot/$path")) {
                                my $pr = {
 -                                      path => $path,
 -                                      owner => to_utf8($owner),
 +                                      path => $path
                                };
 +                              if ($owner) {
 +                                      $pr->{'owner'} = to_utf8($owner);
 +                              }
                                push @list, $pr;
                        }
                }
@@@ -3909,7 -3891,6 +3914,7 @@@ sub print_feed_meta 
                                '-type' => "application/$type+xml"
                        );
  
 +                      $href_params{'extra_options'} = undef;
                        $href_params{'action'} = $type;
                        $link_attr{'-href'} = href(%href_params);
                        print "<link ".
@@@ -5017,186 -4998,10 +5022,186 @@@ sub git_difftree_body 
        print "</table>\n";
  }
  
 -sub print_sidebyside_diff_chunk {
 -      my @chunk = @_;
 +# Print context lines and then rem/add lines in a side-by-side manner.
 +sub print_sidebyside_diff_lines {
 +      my ($ctx, $rem, $add) = @_;
 +
 +      # print context block before add/rem block
 +      if (@$ctx) {
 +              print join '',
 +                      '<div class="chunk_block ctx">',
 +                              '<div class="old">',
 +                              @$ctx,
 +                              '</div>',
 +                              '<div class="new">',
 +                              @$ctx,
 +                              '</div>',
 +                      '</div>';
 +      }
 +
 +      if (!@$add) {
 +              # pure removal
 +              print join '',
 +                      '<div class="chunk_block rem">',
 +                              '<div class="old">',
 +                              @$rem,
 +                              '</div>',
 +                      '</div>';
 +      } elsif (!@$rem) {
 +              # pure addition
 +              print join '',
 +                      '<div class="chunk_block add">',
 +                              '<div class="new">',
 +                              @$add,
 +                              '</div>',
 +                      '</div>';
 +      } else {
 +              print join '',
 +                      '<div class="chunk_block chg">',
 +                              '<div class="old">',
 +                              @$rem,
 +                              '</div>',
 +                              '<div class="new">',
 +                              @$add,
 +                              '</div>',
 +                      '</div>';
 +      }
 +}
 +
 +# Print context lines and then rem/add lines in inline manner.
 +sub print_inline_diff_lines {
 +      my ($ctx, $rem, $add) = @_;
 +
 +      print @$ctx, @$rem, @$add;
 +}
 +
 +# Format removed and added line, mark changed part and HTML-format them.
 +# Implementation is based on contrib/diff-highlight
 +sub format_rem_add_lines_pair {
 +      my ($rem, $add, $num_parents) = @_;
 +
 +      # We need to untabify lines before split()'ing them;
 +      # otherwise offsets would be invalid.
 +      chomp $rem;
 +      chomp $add;
 +      $rem = untabify($rem);
 +      $add = untabify($add);
 +
 +      my @rem = split(//, $rem);
 +      my @add = split(//, $add);
 +      my ($esc_rem, $esc_add);
 +      # Ignore leading +/- characters for each parent.
 +      my ($prefix_len, $suffix_len) = ($num_parents, 0);
 +      my ($prefix_has_nonspace, $suffix_has_nonspace);
 +
 +      my $shorter = (@rem < @add) ? @rem : @add;
 +      while ($prefix_len < $shorter) {
 +              last if ($rem[$prefix_len] ne $add[$prefix_len]);
 +
 +              $prefix_has_nonspace = 1 if ($rem[$prefix_len] !~ /\s/);
 +              $prefix_len++;
 +      }
 +
 +      while ($prefix_len + $suffix_len < $shorter) {
 +              last if ($rem[-1 - $suffix_len] ne $add[-1 - $suffix_len]);
 +
 +              $suffix_has_nonspace = 1 if ($rem[-1 - $suffix_len] !~ /\s/);
 +              $suffix_len++;
 +      }
 +
 +      # Mark lines that are different from each other, but have some common
 +      # part that isn't whitespace.  If lines are completely different, don't
 +      # mark them because that would make output unreadable, especially if
 +      # diff consists of multiple lines.
 +      if ($prefix_has_nonspace || $suffix_has_nonspace) {
 +              $esc_rem = esc_html_hl_regions($rem, 'marked',
 +                      [$prefix_len, @rem - $suffix_len], -nbsp=>1);
 +              $esc_add = esc_html_hl_regions($add, 'marked',
 +                      [$prefix_len, @add - $suffix_len], -nbsp=>1);
 +      } else {
 +              $esc_rem = esc_html($rem, -nbsp=>1);
 +              $esc_add = esc_html($add, -nbsp=>1);
 +      }
 +
 +      return format_diff_line(\$esc_rem, 'rem'),
 +             format_diff_line(\$esc_add, 'add');
 +}
 +
 +# HTML-format diff context, removed and added lines.
 +sub format_ctx_rem_add_lines {
 +      my ($ctx, $rem, $add, $num_parents) = @_;
 +      my (@new_ctx, @new_rem, @new_add);
 +      my $can_highlight = 0;
 +      my $is_combined = ($num_parents > 1);
 +
 +      # Highlight if every removed line has a corresponding added line.
 +      if (@$add > 0 && @$add == @$rem) {
 +              $can_highlight = 1;
 +
 +              # Highlight lines in combined diff only if the chunk contains
 +              # diff between the same version, e.g.
 +              #
 +              #    - a
 +              #   -  b
 +              #    + c
 +              #   +  d
 +              #
 +              # Otherwise the highlightling would be confusing.
 +              if ($is_combined) {
 +                      for (my $i = 0; $i < @$add; $i++) {
 +                              my $prefix_rem = substr($rem->[$i], 0, $num_parents);
 +                              my $prefix_add = substr($add->[$i], 0, $num_parents);
 +
 +                              $prefix_rem =~ s/-/+/g;
 +
 +                              if ($prefix_rem ne $prefix_add) {
 +                                      $can_highlight = 0;
 +                                      last;
 +                              }
 +                      }
 +              }
 +      }
 +
 +      if ($can_highlight) {
 +              for (my $i = 0; $i < @$add; $i++) {
 +                      my ($line_rem, $line_add) = format_rem_add_lines_pair(
 +                              $rem->[$i], $add->[$i], $num_parents);
 +                      push @new_rem, $line_rem;
 +                      push @new_add, $line_add;
 +              }
 +      } else {
 +              @new_rem = map { format_diff_line($_, 'rem') } @$rem;
 +              @new_add = map { format_diff_line($_, 'add') } @$add;
 +      }
 +
 +      @new_ctx = map { format_diff_line($_, 'ctx') } @$ctx;
 +
 +      return (\@new_ctx, \@new_rem, \@new_add);
 +}
 +
 +# Print context lines and then rem/add lines.
 +sub print_diff_lines {
 +      my ($ctx, $rem, $add, $diff_style, $num_parents) = @_;
 +      my $is_combined = $num_parents > 1;
 +
 +      ($ctx, $rem, $add) = format_ctx_rem_add_lines($ctx, $rem, $add,
 +              $num_parents);
 +
 +      if ($diff_style eq 'sidebyside' && !$is_combined) {
 +              print_sidebyside_diff_lines($ctx, $rem, $add);
 +      } else {
 +              # default 'inline' style and unknown styles
 +              print_inline_diff_lines($ctx, $rem, $add);
 +      }
 +}
 +
 +sub print_diff_chunk {
 +      my ($diff_style, $num_parents, $from, $to, @chunk) = @_;
        my (@ctx, @rem, @add);
  
 +      # The class of the previous line.
 +      my $prev_class = '';
 +
        return unless @chunk;
  
        # incomplete last line might be among removed or added lines,
  
                # print chunk headers
                if ($class && $class eq 'chunk_header') {
 -                      print $line;
 +                      print format_diff_line($line, $class, $from, $to);
                        next;
                }
  
 -              ## print from accumulator when type of class of lines change
 -              # empty contents block on start rem/add block, or end of chunk
 -              if (@ctx && (!$class || $class eq 'rem' || $class eq 'add')) {
 -                      print join '',
 -                              '<div class="chunk_block ctx">',
 -                                      '<div class="old">',
 -                                      @ctx,
 -                                      '</div>',
 -                                      '<div class="new">',
 -                                      @ctx,
 -                                      '</div>',
 -                              '</div>';
 -                      @ctx = ();
 -              }
 -              # empty add/rem block on start context block, or end of chunk
 -              if ((@rem || @add) && (!$class || $class eq 'ctx')) {
 -                      if (!@add) {
 -                              # pure removal
 -                              print join '',
 -                                      '<div class="chunk_block rem">',
 -                                              '<div class="old">',
 -                                              @rem,
 -                                              '</div>',
 -                                      '</div>';
 -                      } elsif (!@rem) {
 -                              # pure addition
 -                              print join '',
 -                                      '<div class="chunk_block add">',
 -                                              '<div class="new">',
 -                                              @add,
 -                                              '</div>',
 -                                      '</div>';
 -                      } else {
 -                              # assume that it is change
 -                              print join '',
 -                                      '<div class="chunk_block chg">',
 -                                              '<div class="old">',
 -                                              @rem,
 -                                              '</div>',
 -                                              '<div class="new">',
 -                                              @add,
 -                                              '</div>',
 -                                      '</div>';
 -                      }
 -                      @rem = @add = ();
 +              ## print from accumulator when have some add/rem lines or end
 +              # of chunk (flush context lines), or when have add and rem
 +              # lines and new block is reached (otherwise add/rem lines could
 +              # be reordered)
 +              if (!$class || ((@rem || @add) && $class eq 'ctx') ||
 +                  (@rem && @add && $class ne $prev_class)) {
 +                      print_diff_lines(\@ctx, \@rem, \@add,
 +                                       $diff_style, $num_parents);
 +                      @ctx = @rem = @add = ();
                }
  
                ## adding lines to accumulator
                if ($class eq 'ctx') {
                        push @ctx, $line;
                }
 +
 +              $prev_class = $class;
        }
  }
  
@@@ -5366,19 -5205,27 +5371,19 @@@ sub git_patchset_body 
  
                        next PATCH if ($patch_line =~ m/^diff /);
  
 -                      my ($class, $line) = process_diff_line($patch_line, \%from, \%to);
 -                      my $diff_classes = "diff";
 -                      $diff_classes .= " $class" if ($class);
 -                      $line = "<div class=\"$diff_classes\">$line</div>\n";
 +                      my $class = diff_line_class($patch_line, \%from, \%to);
  
 -                      if ($diff_style eq 'sidebyside' && !$is_combined) {
 -                              if ($class eq 'chunk_header') {
 -                                      print_sidebyside_diff_chunk(@chunk);
 -                                      @chunk = ( [ $class, $line ] );
 -                              } else {
 -                                      push @chunk, [ $class, $line ];
 -                              }
 -                      } else {
 -                              # default 'inline' style and unknown styles
 -                              print $line;
 +                      if ($class eq 'chunk_header') {
 +                              print_diff_chunk($diff_style, scalar @hash_parents, \%from, \%to, @chunk);
 +                              @chunk = ();
                        }
 +
 +                      push @chunk, [ $class, $patch_line ];
                }
  
        } continue {
                if (@chunk) {
 -                      print_sidebyside_diff_chunk(@chunk);
 +                      print_diff_chunk($diff_style, scalar @hash_parents, \%from, \%to, @chunk);
                        @chunk = ();
                }
                print "</div>\n"; # class="patch"
@@@ -5618,15 -5465,11 +5623,15 @@@ sub git_project_list_rows 
                                        ? esc_html_match_hl_chopped($pr->{'descr_long'},
                                                                    $pr->{'descr'}, $search_regexp)
                                        : esc_html($pr->{'descr'})) .
 -                    "</td>\n" .
 -                    "<td><i>" . chop_and_escape_str($pr->{'owner'}, 15) . "</i></td>\n";
 -              print "<td class=\"". age_class($pr->{'age'}) . "\">" .
 -                    (defined $pr->{'age_string'} ? $pr->{'age_string'} : "No commits") . "</td>\n" .
 -                    "<td class=\"link\">" .
 +                    "</td>\n";
 +              unless ($omit_owner) {
 +                      print "<td><i>" . chop_and_escape_str($pr->{'owner'}, 15) . "</i></td>\n";
 +              }
 +              unless ($omit_age_column) {
 +                      print "<td class=\"". age_class($pr->{'age'}) . "\">" .
 +                          (defined $pr->{'age_string'} ? $pr->{'age_string'} : "No commits") . "</td>\n";
 +              }
 +              print"<td class=\"link\">" .
                      $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary")}, "summary")   . " | " .
                      $cgi->a({-href => href(project=>$pr->{'path'}, action=>"shortlog")}, "shortlog") . " | " .
                      $cgi->a({-href => href(project=>$pr->{'path'}, action=>"log")}, "log") . " | " .
@@@ -5657,10 -5500,7 +5662,10 @@@ sub git_project_list_body 
                                         'tagfilter'  => $tagfilter)
                if ($tagfilter || $search_regexp);
        # fill the rest
 -      @projects = fill_project_list_info(\@projects);
 +      my @all_fields = ('descr', 'descr_long', 'ctags', 'category');
 +      push @all_fields, ('age', 'age_string') unless($omit_age_column);
 +      push @all_fields, 'owner' unless($omit_owner);
 +      @projects = fill_project_list_info(\@projects, @all_fields);
  
        $order ||= $default_projects_order;
        $from = 0 unless defined $from;
                }
                print_sort_th('project', $order, 'Project');
                print_sort_th('descr', $order, 'Description');
 -              print_sort_th('owner', $order, 'Owner');
 -              print_sort_th('age', $order, 'Last Change');
 +              print_sort_th('owner', $order, 'Owner') unless $omit_owner;
 +              print_sort_th('age', $order, 'Last Change') unless $omit_age_column;
                print "<th></th>\n" . # for links
                      "</tr>\n";
        }
@@@ -6445,10 -6285,8 +6450,10 @@@ sub git_summary 
  
        print "<div class=\"title\">&nbsp;</div>\n";
        print "<table class=\"projects_list\">\n" .
 -            "<tr id=\"metadata_desc\"><td>description</td><td>" . esc_html($descr) . "</td></tr>\n" .
 -            "<tr id=\"metadata_owner\"><td>owner</td><td>" . esc_html($owner) . "</td></tr>\n";
 +            "<tr id=\"metadata_desc\"><td>description</td><td>" . esc_html($descr) . "</td></tr>\n";
 +        unless ($omit_owner) {
 +              print  "<tr id=\"metadata_owner\"><td>owner</td><td>" . esc_html($owner) . "</td></tr>\n";
 +        }
        if (defined $cd{'rfc2822'}) {
                print "<tr id=\"metadata_lchange\"><td>last change</td>" .
                      "<td>".format_timestamp_html(\%cd)."</td></tr>\n";
@@@ -7170,28 -7008,6 +7175,28 @@@ sub snapshot_name 
        return wantarray ? ($name, $name) : $name;
  }
  
 +sub exit_if_unmodified_since {
 +      my ($latest_epoch) = @_;
 +      our $cgi;
 +
 +      my $if_modified = $cgi->http('IF_MODIFIED_SINCE');
 +      if (defined $if_modified) {
 +              my $since;
 +              if (eval { require HTTP::Date; 1; }) {
 +                      $since = HTTP::Date::str2time($if_modified);
 +              } elsif (eval { require Time::ParseDate; 1; }) {
 +                      $since = Time::ParseDate::parsedate($if_modified, GMT => 1);
 +              }
 +              if (defined $since && $latest_epoch <= $since) {
 +                      my %latest_date = parse_date($latest_epoch);
 +                      print $cgi->header(
 +                              -last_modified => $latest_date{'rfc2822'},
 +                              -status => '304 Not Modified');
 +                      goto DONE_GITWEB;
 +              }
 +      }
 +}
 +
  sub git_snapshot {
        my $format = $input_params{'snapshot_format'};
        if (!@snapshot_fmts) {
  
        my ($name, $prefix) = snapshot_name($project, $hash);
        my $filename = "$name$known_snapshot_formats{$format}{'suffix'}";
 +
 +      my %co = parse_commit($hash);
 +      exit_if_unmodified_since($co{'committer_epoch'}) if %co;
 +
        my $cmd = quote_command(
                git_cmd(), 'archive',
                "--format=$known_snapshot_formats{$format}{'format'}",
        }
  
        $filename =~ s/(["\\])/\\$1/g;
 +      my %latest_date;
 +      if (%co) {
 +              %latest_date = parse_date($co{'committer_epoch'}, $co{'committer_tz'});
 +      }
 +
        print $cgi->header(
                -type => $known_snapshot_formats{$format}{'type'},
                -content_disposition => 'inline; filename="' . $filename . '"',
 +              %co ? (-last_modified => $latest_date{'rfc2822'}) : (),
                -status => '200 OK');
  
        open my $fd, "-|", $cmd
@@@ -8019,14 -7825,33 +8024,14 @@@ sub git_feed 
        if (defined($commitlist[0])) {
                %latest_commit = %{$commitlist[0]};
                my $latest_epoch = $latest_commit{'committer_epoch'};
 -              %latest_date   = parse_date($latest_epoch, $latest_commit{'comitter_tz'});
 -              my $if_modified = $cgi->http('IF_MODIFIED_SINCE');
 -              if (defined $if_modified) {
 -                      my $since;
 -                      if (eval { require HTTP::Date; 1; }) {
 -                              $since = HTTP::Date::str2time($if_modified);
 -                      } elsif (eval { require Time::ParseDate; 1; }) {
 -                              $since = Time::ParseDate::parsedate($if_modified, GMT => 1);
 -                      }
 -                      if (defined $since && $latest_epoch <= $since) {
 -                              print $cgi->header(
 -                                      -type => $content_type,
 -                                      -charset => 'utf-8',
 -                                      -last_modified => $latest_date{'rfc2822'},
 -                                      -status => '304 Not Modified');
 -                              return;
 -                      }
 -              }
 -              print $cgi->header(
 -                      -type => $content_type,
 -                      -charset => 'utf-8',
 -                      -last_modified => $latest_date{'rfc2822'});
 -      } else {
 -              print $cgi->header(
 -                      -type => $content_type,
 -                      -charset => 'utf-8');
 +              exit_if_unmodified_since($latest_epoch);
 +              %latest_date = parse_date($latest_epoch, $latest_commit{'comitter_tz'});
        }
 +      print $cgi->header(
 +              -type => $content_type,
 +              -charset => 'utf-8',
 +              %latest_date ? (-last_modified => $latest_date{'rfc2822'}) : (),
 +              -status => '200 OK');
  
        # Optimization: skip generating the body if client asks only
        # for Last-Modified date.