Switch Markdown rendering from Text::MultiMarkdown to CommonMark

Custom rendering function to add `id` attribute to headings.

Fixes #2312

Author: ilmari

.github/workflows/test.yml

@@ -33,6 +33,7 @@ jobs:
         with:
           node-version: "18"
       - run: npm install -g yarn && yarn install
+      - run: apt-get update && apt-get -y install libcmark-dev
       - name: Install Carton
         uses: perl-actions/install-with-cpm@stable
         with:

cpanfile

@@ -14,6 +14,7 @@ requires 'Catalyst::TraitFor::Request::REST::ForBrowsers';
 requires 'Catalyst::View::JSON';
 requires 'Catalyst::View::Xslate';
 requires 'CatalystX::Fastly::Role::Response', '0.06';
+requires 'CommonMark';
 requires 'Config::General';
 requires 'Config::ZOMG', '1.000000';
 requires 'Cpanel::JSON::XS';
@@ -69,7 +70,6 @@ requires 'Plack::Test';
 requires 'Ref::Util', '>= 0.008';
 requires 'Router::Simple';
 requires 'Term::Size::Any';
-requires 'Text::MultiMarkdown';
 requires 'Text::Pluralize';
 requires 'Text::Xslate::Bridge';
 requires 'Text::Xslate::Bridge::Star';

cpanfile.snapshot

@@ -809,6 +809,15 @@ DISTRIBUTIONS
       perl 5.006
       strict 0
       warnings 0
+  CommonMark-0.290000
+    pathname: N/NW/NWELLNHOF/CommonMark-0.290000.tar.gz
+    provides:
+      CommonMark 0.290000
+      CommonMark::Node 0.290000
+    requirements:
+      Devel::CheckLib 0
+      ExtUtils::MakeMaker 0
+      perl 5.008
   Config-Any-0.33
     pathname: H/HA/HAARG/Config-Any-0.33.tar.gz
     provides:
@@ -1481,6 +1490,16 @@ DISTRIBUTIONS
       Test::More 0.98
       parent 0
       perl 5.008001
+  Devel-CheckLib-1.16
+    pathname: M/MA/MATTN/Devel-CheckLib-1.16.tar.gz
+    provides:
+      Devel::CheckLib 1.16
+    requirements:
+      Exporter 0
+      ExtUtils::MakeMaker 0
+      File::Spec 0
+      File::Temp 0.16
+      perl 5.004050
   Devel-Confess-0.009004
     pathname: H/HA/HAARG/Devel-Confess-0.009004.tar.gz
     provides:
@@ -5457,35 +5476,6 @@ DISTRIBUTIONS
       Exporter 0
       ExtUtils::MakeMaker 0
       perl 5.006
-  Text-Markdown-1.000031
-    pathname: B/BO/BOBTFISH/Text-Markdown-1.000031.tar.gz
-    provides:
-      Text::Markdown 1.000031
-    requirements:
-      Digest::MD5 0
-      Encode 0
-      ExtUtils::MakeMaker 6.42
-      FindBin 0
-      List::MoreUtils 0
-      Test::Differences 0
-      Test::Exception 0
-      Test::More 0.42
-      Text::Balanced 0
-  Text-MultiMarkdown-1.000035
-    pathname: B/BO/BOBTFISH/Text-MultiMarkdown-1.000035.tar.gz
-    provides:
-      Text::MultiMarkdown 1.000035
-    requirements:
-      Digest::MD5 0
-      Encode 0
-      ExtUtils::MakeMaker 6.59
-      FindBin 0
-      HTML::Entities 0
-      List::MoreUtils 0
-      Test::Exception 0
-      Test::More 0.42
-      Text::Markdown v1.0.26
-      perl 5.008
   Text-Pluralize-1.1
     pathname: K/KV/KVAIL/Text-Pluralize-1.1.tar.gz
     provides:

lib/MetaCPAN/Web/Controller/Feed.pm

@@ -9,7 +9,7 @@ use HTML::Escape               qw( escape_html );
 use MetaCPAN::Web::Types       qw( ArrayRef Enum HashRef Str Undef Uri );
 use Params::ValidationCompiler qw( validation_for );
 use Path::Tiny                 qw( path );
-use Text::MultiMarkdown        qw( markdown );
+use MetaCPAN::Web::RenderUtil  qw( render_markdown );
 use XML::FeedPP                ();
 
 sub recent_rss : Path('/recent.rss') Args(0) {
@@ -91,7 +91,7 @@ sub news : Private {
 
         #$str =~ s{\[([^]]+)\]\(([^)]+)\)}{<a href="$2">$1</a>}g;
         $e{abstract} = $str;
-        $e{abstract} = markdown($str);
+        $e{abstract} = render_markdown($str);
 
         push @entries, \%e;
     }

lib/MetaCPAN/Web/RenderUtil.pm

@@ -8,10 +8,12 @@ use HTML::Escape   qw( escape_html );
 use HTML::Restrict ();
 use URI            ();
 use Digest::MD5    ();
+use CommonMark     qw( :node :event );
 
 our @EXPORT_OK = qw(
     filter_html
     gravatar_image
+    render_markdown
 );
 
 sub filter_html {
@@ -146,4 +148,66 @@ sub gravatar_image {
         $grav_id, $size // 80;
 }
 
+my @is_leaf;
+$is_leaf[$_] = 1
+    for (
+    NODE_HTML,      NODE_HRULE,     NODE_CODE_BLOCK, NODE_TEXT,
+    NODE_SOFTBREAK, NODE_LINEBREAK, NODE_CODE,       NODE_INLINE_HTML,
+    );
+
+sub render_markdown {
+    my ($markdown) = @_;
+
+    my $doc = CommonMark->parse_document($markdown);
+
+    my ( $html, $header_content, %seen_header );
+
+    my $iter = $doc->iterator;
+    while ( my ( $ev_type, $node ) = $iter->next ) {
+        my $node_type = $node->get_type;
+
+        if ( $node_type == NODE_DOCUMENT ) {
+            next;
+        }
+
+        if ( $node_type == NODE_HEADER ) {
+            if ( $ev_type == EVENT_ENTER ) {
+                $header_content = '';
+            }
+            if ( $ev_type == EVENT_EXIT ) {
+                $header_content =~ s{(?:-(\d+))?$}{'-' . (($1 // 1) + 1)}e
+                    while $seen_header{$header_content}++;
+
+                my $header_html = $node->render_html;
+                $header_html
+                    =~ s/^<h[0-9]+\b\K/' id="'.escape_html($header_content).'"'/e;
+                $html .= $header_html;
+
+                undef $header_content;
+            }
+        }
+        elsif ($ev_type == EVENT_ENTER
+            && $node->parent->get_type == NODE_DOCUMENT )
+        {
+            $html .= $node->render_html;
+        }
+
+        if ( defined $header_content ) {
+            if ( $is_leaf[$node_type] ) {
+                my $content = lc( $node->get_literal );
+                $content =~ s/\A\s+//;
+                $content =~ s/\s+\z//;
+                $content =~ s/\s+/-/g;
+
+                if ( length $content ) {
+                    $header_content .= '-' if length $header_content;
+                    $header_content .= $content;
+                }
+            }
+        }
+    }
+
+    return $html;
+}
+
 1;

lib/MetaCPAN/Web/View/Xslate/Bridge.pm

@@ -6,12 +6,12 @@ use parent qw(Text::Xslate::Bridge);
 use Text::Xslate::Util        qw( mark_raw );
 use Number::Format            ();
 use Ref::Util                 qw( is_coderef is_regexpref );
-use Text::MultiMarkdown       ();
 use List::Util                ();
 use DateTime                  ();
 use With::Roles               ();
 use Text::Pluralize           ();
 use MetaCPAN::Web::RenderUtil qw( gravatar_image );          ## no perlimports
+use MetaCPAN::Web::RenderUtil qw( render_markdown );
 use overload                  ();
 
 my $num_formatter = Number::Format->new;
@@ -26,11 +26,9 @@ sub format_bytes {
     $num_formatter->format_bytes($number);
 }
 
-my $md = Text::MultiMarkdown->new( heading_ids => 1 );
-
 sub markdown {
     my ($text) = @_;
-    mark_raw( $md->markdown($text) );
+    mark_raw( render_markdown($text) );
 }
 
 sub filter_html {

t/markdown.t

@@ -0,0 +1,32 @@
+use strict;
+use warnings;
+use Test::More;
+
+use MetaCPAN::Web::RenderUtil qw( render_markdown );
+
+my $html = render_markdown(<<'EOM');
+# Heading
+
+Some body text
+
+## Heading
+
+More stuff
+
+## Heading
+
+more stuff
+
+## Heading **with** _markup_
+
+Content
+
+EOM
+
+like $html, qr{<h1 id="heading">Heading</h1>},   'first heading';
+like $html, qr{<h2 id="heading-2">Heading</h2>}, 'second heading';
+like $html, qr{<h2 id="heading-3">Heading</h2>}, 'third heading';
+
+like $html, qr{<h2 id="heading-with-markup">Heading <}, 'heading with markup';
+
+done_testing();