From 6783ecfd03bb754510c950ebc473f5aa12341c01 Mon Sep 17 00:00:00 2001 From: Parker Lougheed Date: Tue, 30 Sep 2025 15:48:05 +0200 Subject: [PATCH 1/2] [markdown] Preserve metadata passed to fenced code blocks --- pkgs/markdown/CHANGELOG.md | 2 + .../fenced_code_block_syntax.dart | 53 ++++++++++++++----- .../test/common_mark/fenced_code_blocks.unit | 4 +- .../markdown/test/gfm/fenced_code_blocks.unit | 4 +- .../test/original/fenced_code_block.unit | 24 +++++++++ 5 files changed, 71 insertions(+), 16 deletions(-) diff --git a/pkgs/markdown/CHANGELOG.md b/pkgs/markdown/CHANGELOG.md index 60a2508877..49b9e358ec 100644 --- a/pkgs/markdown/CHANGELOG.md +++ b/pkgs/markdown/CHANGELOG.md @@ -1,5 +1,7 @@ ## 7.3.1-wip +* Preserve metadata passed to fenced code blocks as + `data-metadata` on the created `pre` element. * Update the README link to the markdown playground (https://dart-lang.github.io/tools). * Update `package:web` API references in the example. diff --git a/pkgs/markdown/lib/src/block_syntaxes/fenced_code_block_syntax.dart b/pkgs/markdown/lib/src/block_syntaxes/fenced_code_block_syntax.dart index cea110665f..ed0864e2ef 100644 --- a/pkgs/markdown/lib/src/block_syntaxes/fenced_code_block_syntax.dart +++ b/pkgs/markdown/lib/src/block_syntaxes/fenced_code_block_syntax.dart @@ -21,9 +21,9 @@ class FencedCodeBlockSyntax extends BlockSyntax { @override Node parse(BlockParser parser) { - final openingFence = _FenceMatch.fromMatch(pattern.firstMatch( - escapePunctuation(parser.current.content), - )!); + final openingFence = _FenceMatch.fromMatch( + pattern.firstMatch(escapePunctuation(parser.current.content))!, + ); var text = parseChildLines( parser, @@ -38,16 +38,27 @@ class FencedCodeBlockSyntax extends BlockSyntax { text = '$text\n'; } + final (languageString, metadataString) = openingFence.languageAndMetadata; + final code = Element.text('code', text); - if (openingFence.hasLanguage) { - var language = decodeHtmlCharacters(openingFence.language); + if (languageString != null) { + var language = decodeHtmlCharacters(languageString); if (parser.document.encodeHtml) { language = escapeHtmlAttribute(language); } code.attributes['class'] = 'language-$language'; } - return Element('pre', [code]); + final pre = Element('pre', [code]); + if (metadataString != null) { + var metadata = decodeHtmlCharacters(metadataString); + if (parser.document.encodeHtml) { + metadata = escapeHtmlAttribute(metadata); + } + pre.attributes['data-metadata'] = metadata; + } + + return pre; } String _removeIndentation(String content, int length) { @@ -130,12 +141,30 @@ class _FenceMatch { // https://spec.commonmark.org/0.30/#info-string. final String info; - // The first word of the info string is typically used to specify the language - // of the code sample, - // https://spec.commonmark.org/0.30/#example-143. - String get language => info.split(' ').first; + /// Returns the language and remaining metadata from the [info] string. + /// + /// The language is the first word of the info string, + /// to match the (unspecified, but typical) behavior of CommonMark parsers, + /// as suggested in https://spec.commonmark.org/0.30/#example-143. + /// + /// The metadata is any remaining part of the info string after the language. + (String? language, String? metadata) get languageAndMetadata { + if (info.isEmpty) { + return (null, null); + } - bool get hasInfo => info.isNotEmpty; + // We assume the info string is trimmed already. + final firstSpaceIndex = info.indexOf(' '); + if (firstSpaceIndex == -1) { + // If there is no space, the whole string is the language. + return (info, null); + } + + return ( + info.substring(0, firstSpaceIndex), + info.substring(firstSpaceIndex + 1), + ); + } - bool get hasLanguage => language.isNotEmpty; + bool get hasInfo => info.isNotEmpty; } diff --git a/pkgs/markdown/test/common_mark/fenced_code_blocks.unit b/pkgs/markdown/test/common_mark/fenced_code_blocks.unit index 06dec5865c..5854bc370f 100644 --- a/pkgs/markdown/test/common_mark/fenced_code_blocks.unit +++ b/pkgs/markdown/test/common_mark/fenced_code_blocks.unit @@ -214,7 +214,7 @@ def foo(x) end ~~~~~~~ <<< -
def foo(x)
+
def foo(x)
   return 3
 end
 

@@ -234,7 +234,7 @@ foo

 foo
 ~~~
 <<<
-
foo
+
foo
 

 >>> Fenced code blocks - 147
 ```
diff --git a/pkgs/markdown/test/gfm/fenced_code_blocks.unit b/pkgs/markdown/test/gfm/fenced_code_blocks.unit
index f6fa8432a7..52f217ae18 100644
--- a/pkgs/markdown/test/gfm/fenced_code_blocks.unit
+++ b/pkgs/markdown/test/gfm/fenced_code_blocks.unit
@@ -214,7 +214,7 @@ def foo(x)
 end
 ~~~~~~~
 <<<
-
def foo(x)
+
def foo(x)
   return 3
 end
 

@@ -234,7 +234,7 @@ foo

 foo
 ~~~
 <<<
-
foo
+
foo
 

 >>> Fenced code blocks - 117
 ```
diff --git a/pkgs/markdown/test/original/fenced_code_block.unit b/pkgs/markdown/test/original/fenced_code_block.unit
index 779e747c3c..3492dfca6a 100644
--- a/pkgs/markdown/test/original/fenced_code_block.unit
+++ b/pkgs/markdown/test/original/fenced_code_block.unit
@@ -5,3 +5,27 @@
 <<<
 
'foo'
 

+>>> with basic metadata string
+```dart meta
+code
+```
+
+<<<
+
code
+

+>>> with characters to escape
+```dart title="main.dart"
+code
+```
+
+<<<
+
code
+

+>>> with HTML character reference
+```dart |
+code
+```
+
+<<<
+
code
+

\ No newline at end of file

From df5f03f68266a2ca25e8b7bec39c2aaad845fab7 Mon Sep 17 00:00:00 2001
From: Parker Lougheed 
Date: Tue, 30 Sep 2025 16:28:23 +0200
Subject: [PATCH 2/2] Extract shared logic as suggested by Gemini

---
 .../fenced_code_block_syntax.dart             | 24 +++++++++++--------
 1 file changed, 14 insertions(+), 10 deletions(-)

diff --git a/pkgs/markdown/lib/src/block_syntaxes/fenced_code_block_syntax.dart b/pkgs/markdown/lib/src/block_syntaxes/fenced_code_block_syntax.dart
index ed0864e2ef..158aba9873 100644
--- a/pkgs/markdown/lib/src/block_syntaxes/fenced_code_block_syntax.dart
+++ b/pkgs/markdown/lib/src/block_syntaxes/fenced_code_block_syntax.dart
@@ -42,25 +42,29 @@ class FencedCodeBlockSyntax extends BlockSyntax {
 
     final code = Element.text('code', text);
     if (languageString != null) {
-      var language = decodeHtmlCharacters(languageString);
-      if (parser.document.encodeHtml) {
-        language = escapeHtmlAttribute(language);
-      }
-      code.attributes['class'] = 'language-$language';
+      final processedLanguage = _processAttribute(languageString,
+          encodeHtml: parser.document.encodeHtml);
+      code.attributes['class'] = 'language-$processedLanguage';
     }
 
     final pre = Element('pre', [code]);
     if (metadataString != null) {
-      var metadata = decodeHtmlCharacters(metadataString);
-      if (parser.document.encodeHtml) {
-        metadata = escapeHtmlAttribute(metadata);
-      }
-      pre.attributes['data-metadata'] = metadata;
+      final processedMetadata = _processAttribute(metadataString,
+          encodeHtml: parser.document.encodeHtml);
+      pre.attributes['data-metadata'] = processedMetadata;
     }
 
     return pre;
   }
 
+  static String _processAttribute(String value, {bool encodeHtml = false}) {
+    final decodedValue = decodeHtmlCharacters(value);
+    if (encodeHtml) {
+      return escapeHtmlAttribute(decodedValue);
+    }
+    return decodedValue;
+  }
+
   String _removeIndentation(String content, int length) {
     final text = content.replaceFirst(RegExp('^\\s{0,$length}'), '');
     return content.substring(content.length - text.length);