gitweb: Use print_diff_chunk() for both side-by-side and inline diffs
[gitweb.git] / gitweb / gitweb.perl
index 3aeeb8b34e2bd0da386aae6708143af20ac831ee..90836e633c145207ee8a919908feadc47fe5e3ae 100755 (executable)
@@ -1081,7 +1081,16 @@ sub evaluate_and_validate_params {
                if (length($searchtext) < 2) {
                        die_error(403, "At least two characters are required for search parameter");
                }
-               $search_regexp = $search_use_regexp ? $searchtext : quotemeta $searchtext;
+               if ($search_use_regexp) {
+                       $search_regexp = $searchtext;
+                       if (!eval { qr/$search_regexp/; 1; }) {
+                               (my $error = $@) =~ s/ at \S+ line \d+.*\n?//;
+                               die_error(400, "Invalid search regexp '$search_regexp'",
+                                         esc_html($error));
+                       }
+               } else {
+                       $search_regexp = quotemeta $searchtext;
+               }
        }
 }
 
@@ -1715,6 +1724,97 @@ sub chop_and_escape_str {
        }
 }
 
+# Highlight selected fragments of string, using given CSS class,
+# and escape HTML.  It is assumed that fragments do not overlap.
+# Regions are passed as list of pairs (array references).
+#
+# Example: esc_html_hl_regions("foobar", "mark", [ 0, 3 ]) returns
+# '<span class="mark">foo</span>bar'
+sub esc_html_hl_regions {
+       my ($str, $css_class, @sel) = @_;
+       my %opts = grep { ref($_) ne 'ARRAY' } @sel;
+       @sel     = grep { ref($_) eq 'ARRAY' } @sel;
+       return esc_html($str, %opts) unless @sel;
+
+       my $out = '';
+       my $pos = 0;
+
+       for my $s (@sel) {
+               my ($begin, $end) = @$s;
+
+               # Don't create empty <span> elements.
+               next if $end <= $begin;
+
+               my $escaped = esc_html(substr($str, $begin, $end - $begin),
+                                      %opts);
+
+               $out .= esc_html(substr($str, $pos, $begin - $pos), %opts)
+                       if ($begin - $pos > 0);
+               $out .= $cgi->span({-class => $css_class}, $escaped);
+
+               $pos = $end;
+       }
+       $out .= esc_html(substr($str, $pos), %opts)
+               if ($pos < length($str));
+
+       return $out;
+}
+
+# return positions of beginning and end of each match
+sub matchpos_list {
+       my ($str, $regexp) = @_;
+       return unless (defined $str && defined $regexp);
+
+       my @matches;
+       while ($str =~ /$regexp/g) {
+               push @matches, [$-[0], $+[0]];
+       }
+       return @matches;
+}
+
+# highlight match (if any), and escape HTML
+sub esc_html_match_hl {
+       my ($str, $regexp) = @_;
+       return esc_html($str) unless defined $regexp;
+
+       my @matches = matchpos_list($str, $regexp);
+       return esc_html($str) unless @matches;
+
+       return esc_html_hl_regions($str, 'match', @matches);
+}
+
+
+# highlight match (if any) of shortened string, and escape HTML
+sub esc_html_match_hl_chopped {
+       my ($str, $chopped, $regexp) = @_;
+       return esc_html_match_hl($str, $regexp) unless defined $chopped;
+
+       my @matches = matchpos_list($str, $regexp);
+       return esc_html($chopped) unless @matches;
+
+       # filter matches so that we mark chopped string
+       my $tail = "... "; # see chop_str
+       unless ($chopped =~ s/\Q$tail\E$//) {
+               $tail = '';
+       }
+       my $chop_len = length($chopped);
+       my $tail_len = length($tail);
+       my @filtered;
+
+       for my $m (@matches) {
+               if ($m->[0] > $chop_len) {
+                       push @filtered, [ $chop_len, $chop_len + $tail_len ] if ($tail_len > 0);
+                       last;
+               } elsif ($m->[1] > $chop_len) {
+                       push @filtered, [ $m->[0], $chop_len + $tail_len ];
+                       last;
+               }
+               push @filtered, $m;
+       }
+
+       return esc_html_hl_regions($chopped . $tail, 'match', @filtered);
+}
+
 ## ----------------------------------------------------------------------
 ## functions returning short strings
 
@@ -2982,11 +3082,15 @@ sub filter_forks_from_projects_list {
 sub search_projects_list {
        my ($projlist, %opts) = @_;
        my $tagfilter  = $opts{'tagfilter'};
-       my $searchtext = $opts{'searchtext'};
+       my $search_re = $opts{'search_regexp'};
 
        return @$projlist
-               unless ($tagfilter || $searchtext);
+               unless ($tagfilter || $search_re);
 
+       # searching projects require filling to be run before it;
+       fill_project_list_info($projlist,
+                              $tagfilter  ? 'ctags' : (),
+                              $search_re ? ('path', 'descr') : ());
        my @projects;
  PROJECT:
        foreach my $pr (@$projlist) {
@@ -2997,10 +3101,10 @@ sub search_projects_list {
                                grep { lc($_) eq lc($tagfilter) } keys %{$pr->{'ctags'}};
                }
 
-               if ($searchtext) {
+               if ($search_re) {
                        next unless
-                               $pr->{'path'} =~ /$searchtext/ ||
-                               $pr->{'descr_long'} =~ /$searchtext/;
+                               $pr->{'path'} =~ /$search_re/ ||
+                               $pr->{'descr_long'} =~ /$search_re/;
                }
 
                push @projects, $pr;
@@ -4898,10 +5002,78 @@ 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;
+}
+
+# Print context lines and then rem/add lines.
+sub print_diff_lines {
+       my ($ctx, $rem, $add, $diff_style, $is_combined) = @_;
+
+       if ($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, $is_combined, @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,
@@ -4924,51 +5096,15 @@ sub print_sidebyside_diff_chunk {
                        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, $is_combined);
+                       @ctx = @rem = @add = ();
                }
 
                ## adding lines to accumulator
@@ -4984,6 +5120,8 @@ sub print_sidebyside_diff_chunk {
                if ($class eq 'ctx') {
                        push @ctx, $line;
                }
+
+               $prev_class = $class;
        }
 }
 
@@ -5110,22 +5248,17 @@ sub git_patchset_body {
                        $diff_classes .= " $class" if ($class);
                        $line = "<div class=\"$diff_classes\">$line</div>\n";
 
-                       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, $is_combined, @chunk);
+                               @chunk = ();
                        }
+
+                       push @chunk, [ $class, $line ];
                }
 
        } continue {
                if (@chunk) {
-                       print_sidebyside_diff_chunk(@chunk);
+                       print_diff_chunk($diff_style, $is_combined, @chunk);
                        @chunk = ();
                }
                print "</div>\n"; # class="patch"
@@ -5161,7 +5294,7 @@ sub git_patchset_body {
 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
 
 sub git_project_search_form {
-       my ($searchtext, $search_use_regexp);
+       my ($searchtext, $search_use_regexp) = @_;
 
        my $limit = '';
        if ($project_filter) {
@@ -5202,39 +5335,56 @@ sub project_info_needs_filling {
        return;
 }
 
-# fills project list info (age, description, owner, category, forks)
+# fills project list info (age, description, owner, category, forks, etc.)
 # for each project in the list, removing invalid projects from
-# returned list
+# returned list, or fill only specified info.
+#
+# Invalid projects are removed from the returned list if and only if you
+# ask 'age' or 'age_string' to be filled, because they are the only fields
+# that run unconditionally git command that requires repository, and
+# therefore do always check if project repository is invalid.
+#
+# USAGE:
+# * fill_project_list_info(\@project_list, 'descr_long', 'ctags')
+#   ensures that 'descr_long' and 'ctags' fields are filled
+# * @project_list = fill_project_list_info(\@project_list)
+#   ensures that all fields are filled (and invalid projects removed)
+#
 # NOTE: modifies $projlist, but does not remove entries from it
 sub fill_project_list_info {
-       my $projlist = shift;
+       my ($projlist, @wanted_keys) = @_;
        my @projects;
+       my $filter_set = sub { return @_; };
+       if (@wanted_keys) {
+               my %wanted_keys = map { $_ => 1 } @wanted_keys;
+               $filter_set = sub { return grep { $wanted_keys{$_} } @_; };
+       }
 
        my $show_ctags = gitweb_check_feature('ctags');
  PROJECT:
        foreach my $pr (@$projlist) {
-               if (project_info_needs_filling($pr, 'age', 'age_string')) {
+               if (project_info_needs_filling($pr, $filter_set->('age', 'age_string'))) {
                        my (@activity) = git_get_last_activity($pr->{'path'});
                        unless (@activity) {
                                next PROJECT;
                        }
                        ($pr->{'age'}, $pr->{'age_string'}) = @activity;
                }
-               if (project_info_needs_filling($pr, 'descr', 'descr_long')) {
+               if (project_info_needs_filling($pr, $filter_set->('descr', 'descr_long'))) {
                        my $descr = git_get_project_description($pr->{'path'}) || "";
                        $descr = to_utf8($descr);
                        $pr->{'descr_long'} = $descr;
                        $pr->{'descr'} = chop_str($descr, $projects_list_description_width, 5);
                }
-               if (project_info_needs_filling($pr, 'owner')) {
+               if (project_info_needs_filling($pr, $filter_set->('owner'))) {
                        $pr->{'owner'} = git_get_project_owner("$pr->{'path'}") || "";
                }
                if ($show_ctags &&
-                   project_info_needs_filling($pr, 'ctags')) {
+                   project_info_needs_filling($pr, $filter_set->('ctags'))) {
                        $pr->{'ctags'} = git_get_project_ctags($pr->{'path'});
                }
                if ($projects_list_group_categories &&
-                   project_info_needs_filling($pr, 'category')) {
+                   project_info_needs_filling($pr, $filter_set->('category'))) {
                        my $cat = git_get_project_category($pr->{'path'}) ||
                                                           $project_list_default_category;
                        $pr->{'category'} = to_utf8($cat);
@@ -5338,10 +5488,17 @@ sub git_project_list_rows {
                        print "</td>\n";
                }
                print "<td>" . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary"),
-                                       -class => "list"}, esc_html($pr->{'path'})) . "</td>\n" .
+                                       -class => "list"},
+                                      esc_html_match_hl($pr->{'path'}, $search_regexp)) .
+                     "</td>\n" .
                      "<td>" . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary"),
-                                       -class => "list", -title => $pr->{'descr_long'}},
-                                       esc_html($pr->{'descr'})) . "</td>\n" .
+                                       -class => "list",
+                                       -title => $pr->{'descr_long'}},
+                                       $search_regexp
+                                       ? esc_html_match_hl_chopped($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" .
@@ -5365,17 +5522,18 @@ sub git_project_list_body {
        my $show_ctags  = gitweb_check_feature('ctags');
        my $tagfilter = $show_ctags ? $input_params{'ctag'} : undef;
        $check_forks = undef
-               if ($tagfilter || $searchtext);
+               if ($tagfilter || $search_regexp);
 
        # filtering out forks before filling info allows to do less work
        @projects = filter_forks_from_projects_list(\@projects)
                if ($check_forks);
-       @projects = fill_project_list_info(\@projects);
-       # searching projects require filling to be run before it
+       # search_projects_list pre-fills required info
        @projects = search_projects_list(\@projects,
-                                        'searchtext' => $searchtext,
+                                        'search_regexp' => $search_regexp,
                                         'tagfilter'  => $tagfilter)
-               if ($tagfilter || $searchtext);
+               if ($tagfilter || $search_regexp);
+       # fill the rest
+       @projects = fill_project_list_info(\@projects);
 
        $order ||= $default_projects_order;
        $from = 0 unless defined $from;
@@ -5651,7 +5809,7 @@ sub git_tags_body {
 
 sub git_heads_body {
        # uses global variable $project
-       my ($headlist, $head, $from, $to, $extra) = @_;
+       my ($headlist, $head_at, $from, $to, $extra) = @_;
        $from = 0 unless defined $from;
        $to = $#{$headlist} if (!defined $to || $#{$headlist} < $to);
 
@@ -5660,7 +5818,7 @@ sub git_heads_body {
        for (my $i = $from; $i <= $to; $i++) {
                my $entry = $headlist->[$i];
                my %ref = %$entry;
-               my $curr = $ref{'id'} eq $head;
+               my $curr = defined $head_at && $ref{'id'} eq $head_at;
                if ($alternate) {
                        print "<tr class=\"dark\">\n";
                } else {
@@ -5933,9 +6091,10 @@ sub git_search_files {
        my $alternate = 1;
        my $matches = 0;
        my $lastfile = '';
+       my $file_href;
        while (my $line = <$fd>) {
                chomp $line;
-               my ($file, $file_href, $lno, $ltext, $binary);
+               my ($file, $lno, $ltext, $binary);
                last if ($matches++ > 1000);
                if ($line =~ /^Binary file (.+) matches$/) {
                        $file = $1;