Refactor ATX headings (#486)
* Refactor ATX headings
* Update a comment
* A small fix
* Update stats
diff --git a/lib/src/block_syntaxes/header_syntax.dart b/lib/src/block_syntaxes/header_syntax.dart
index 7d2b94e..b68268d 100644
--- a/lib/src/block_syntaxes/header_syntax.dart
+++ b/lib/src/block_syntaxes/header_syntax.dart
@@ -17,9 +17,28 @@
@override
Node parse(BlockParser parser) {
final match = pattern.firstMatch(parser.current)!;
+ final matchedText = match[0]!;
+ final openMarker = match[1]!;
+ final closeMarker = match[2];
+ final level = openMarker.length;
+ final openMarkerStart = matchedText.indexOf(openMarker);
+ final openMarkerEnd = openMarkerStart + level;
+
+ String? content;
+ if (closeMarker == null) {
+ content = parser.current.substring(openMarkerEnd);
+ } else {
+ final closeMarkerStart = matchedText.lastIndexOf(closeMarker);
+ content = parser.current.substring(openMarkerEnd, closeMarkerStart);
+ }
+ content = content.trim();
+
+ // https://spec.commonmark.org/0.30/#example-79
+ if (closeMarker == null && RegExp(r'^#+$').hasMatch(content)) {
+ content = null;
+ }
+
parser.advance();
- final level = match[1]!.length;
- final contents = UnparsedContent(match[2]!.trim());
- return Element('h$level', [contents]);
+ return Element('h$level', [if (content != null) UnparsedContent(content)]);
}
}
diff --git a/lib/src/patterns.dart b/lib/src/patterns.dart
index 269689f..2bd696b 100644
--- a/lib/src/patterns.dart
+++ b/lib/src/patterns.dart
@@ -12,7 +12,8 @@
///
/// Starts with 1-6 unescaped `#` characters which must not be followed by a
/// non-space character. Line may end with any number of `#` characters,.
-final headerPattern = RegExp(r'^ {0,3}(#{1,6})[ \x09\x0b\x0c](.*?)#*$');
+final headerPattern =
+ RegExp(r'^ {0,3}(#{1,6})(?:[ \x09\x0b\x0c].*?)?(?:\s(#*)\s*)?$');
/// The line starts with `>` with one optional space after.
final blockquotePattern = RegExp(r'^[ ]{0,3}>[ \t]?.*$');
diff --git a/test/common_mark/atx_headings.unit b/test/common_mark/atx_headings.unit
index 1bb50e7..1cb7e98 100644
--- a/test/common_mark/atx_headings.unit
+++ b/test/common_mark/atx_headings.unit
@@ -69,7 +69,7 @@
>>> ATX headings - 73
### foo ###
<<<
-<h3>foo ###</h3>
+<h3>foo</h3>
>>> ATX headings - 74
### foo ### b
<<<
@@ -77,15 +77,15 @@
>>> ATX headings - 75
# foo#
<<<
-<h1>foo</h1>
+<h1>foo#</h1>
>>> ATX headings - 76
### foo \###
## foo #\##
# foo \#
<<<
-<h3>foo \</h3>
-<h2>foo #\</h2>
-<h1>foo \</h1>
+<h3>foo ###</h3>
+<h2>foo ###</h2>
+<h1>foo #</h1>
>>> ATX headings - 77
****
## foo
@@ -108,5 +108,5 @@
### ###
<<<
<h2></h2>
-<p>#</p>
+<h1></h1>
<h3></h3>
diff --git a/test/gfm/atx_headings.unit b/test/gfm/atx_headings.unit
index 04e8d9f..5a60cf4 100644
--- a/test/gfm/atx_headings.unit
+++ b/test/gfm/atx_headings.unit
@@ -69,7 +69,7 @@
>>> ATX headings - 43
### foo ###
<<<
-<h3>foo ###</h3>
+<h3>foo</h3>
>>> ATX headings - 44
### foo ### b
<<<
@@ -77,15 +77,15 @@
>>> ATX headings - 45
# foo#
<<<
-<h1>foo</h1>
+<h1>foo#</h1>
>>> ATX headings - 46
### foo \###
## foo #\##
# foo \#
<<<
-<h3>foo \</h3>
-<h2>foo #\</h2>
-<h1>foo \</h1>
+<h3>foo ###</h3>
+<h2>foo ###</h2>
+<h1>foo #</h1>
>>> ATX headings - 47
****
## foo
@@ -108,5 +108,5 @@
### ###
<<<
<h2></h2>
-<p>#</p>
+<h1></h1>
<h3></h3>
diff --git a/tool/common_mark_stats.json b/tool/common_mark_stats.json
index 1cfe950..daaa718 100644
--- a/tool/common_mark_stats.json
+++ b/tool/common_mark_stats.json
@@ -11,13 +11,13 @@
"70": "strict",
"71": "strict",
"72": "strict",
- "73": "loose",
+ "73": "strict",
"74": "strict",
- "75": "loose",
- "76": "loose",
+ "75": "strict",
+ "76": "strict",
"77": "strict",
"78": "strict",
- "79": "fail"
+ "79": "strict"
},
"Autolinks": {
"593": "strict",
diff --git a/tool/common_mark_stats.txt b/tool/common_mark_stats.txt
index a42dfb3..337d16f 100644
--- a/tool/common_mark_stats.txt
+++ b/tool/common_mark_stats.txt
@@ -1,4 +1,4 @@
- 17 of 18 – 94.4% ATX headings
+ 18 of 18 – 100.0% ATX headings
19 of 19 – 100.0% Autolinks
13 of 13 – 100.0% Backslash escapes
1 of 1 – 100.0% Blank lines
@@ -24,5 +24,5 @@
11 of 11 – 100.0% Tabs
3 of 3 – 100.0% Textual content
19 of 19 – 100.0% Thematic breaks
- 633 of 652 – 97.1% TOTAL
- 610 of 633 – 96.4% TOTAL Strict
+ 634 of 652 – 97.2% TOTAL
+ 614 of 634 – 96.8% TOTAL Strict
diff --git a/tool/gfm_stats.json b/tool/gfm_stats.json
index 5e5aae3..91de967 100644
--- a/tool/gfm_stats.json
+++ b/tool/gfm_stats.json
@@ -11,13 +11,13 @@
"40": "strict",
"41": "strict",
"42": "strict",
- "43": "loose",
+ "43": "strict",
"44": "strict",
- "45": "loose",
- "46": "loose",
+ "45": "strict",
+ "46": "strict",
"47": "strict",
"48": "strict",
- "49": "fail"
+ "49": "strict"
},
"Autolinks": {
"602": "strict",
diff --git a/tool/gfm_stats.txt b/tool/gfm_stats.txt
index cc1d67f..b7598f9 100644
--- a/tool/gfm_stats.txt
+++ b/tool/gfm_stats.txt
@@ -1,4 +1,4 @@
- 17 of 18 – 94.4% ATX headings
+ 18 of 18 – 100.0% ATX headings
19 of 19 – 100.0% Autolinks
11 of 11 – 100.0% Autolinks (extension)
13 of 13 – 100.0% Backslash escapes
@@ -28,5 +28,5 @@
11 of 11 – 100.0% Tabs
3 of 3 – 100.0% Textual content
19 of 19 – 100.0% Thematic breaks
- 651 of 671 – 97.0% TOTAL
- 627 of 651 – 96.3% TOTAL Strict
+ 652 of 671 – 97.2% TOTAL
+ 631 of 652 – 96.8% TOTAL Strict
