From ab295e62e4c708d87600efa77fca44195bacfa17 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Nico=20Schl=C3=B6mer?= <nschloe@users.noreply.github.com>
Date: Wed, 4 Nov 2026 05:51:12 +0000
Subject: [PATCH] fix(markup): make heading anchors match GitHub for whitespace
 and combining marks (#40195)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Heading anchors now keep the same characters as GitHub: letters, marks,
decimal and letter numbers, connector punctuation and alphabetic
symbols, with ASCII spaces turned into `-`. The rule follows
[github-slugger](https://github.com/Flet/github-slugger), whose fixtures
are generated from real GitHub renders.

- Other whitespace, including newlines in multi-line headings, is
dropped instead of becoming `-`, so `1<em space>ab` gets `#1ab`
- Combining marks are kept, so `हिन्दी` no longer becomes `हनद`
- Other numbers like `²` and `①` are dropped

Fixes https://github.com/go-gitea/gitea/issues/27754

This PR was written with Claude Code.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: silverwind <me@silverwind.io>
---
 modules/markup/common/footnote.go      | 19 ++++++++-----------
 modules/markup/common/footnote_test.go |  5 ++++-
 2 files changed, 12 insertions(+), 12 deletions(-)

diff --git a/modules/markup/common/footnote.go b/modules/markup/common/footnote.go
index ef89a7046a3..4f1c0ed9363 100644
--- a/modules/markup/common/footnote.go
+++ b/modules/markup/common/footnote.go
@@ -26,18 +26,15 @@ import (
 // and more closely matches the output from the shurcooL sanitizer
 // In particular Unicode letters and numbers are a lot more than a-zA-Z0-9...
 func CleanValue(value []byte) []byte {
-	value = bytes.TrimSpace(value)
-	rs := bytes.Runes(value)
-	result := make([]rune, 0, len(rs))
-	for _, r := range rs {
-		if unicode.IsLetter(r) || unicode.IsNumber(r) || r == '_' || r == '-' {
-			result = append(result, unicode.ToLower(r))
+	return bytes.Map(func(r rune) rune {
+		switch {
+		case r == ' ':
+			return '-'
+		case r == '-' || unicode.In(r, unicode.L, unicode.M, unicode.Nd, unicode.Nl, unicode.Pc, unicode.Other_Alphabetic):
+			return unicode.ToLower(r)
 		}
-		if unicode.IsSpace(r) {
-			result = append(result, '-')
-		}
-	}
-	return []byte(string(result))
+		return -1
+	}, bytes.TrimSpace(value))
 }
 
 // Most of what follows is a subtly changed version of github.com/yuin/goldmark/extension/footnote.go
diff --git a/modules/markup/common/footnote_test.go b/modules/markup/common/footnote_test.go
index 2327a7b14b2..b8e8e7a4f10 100644
--- a/modules/markup/common/footnote_test.go
+++ b/modules/markup/common/footnote_test.go
@@ -50,9 +50,12 @@ func TestCleanValue(t *testing.T) {
 		{"tes\\a", "tesa"},
 		{"tes/a", "tesa"},
 		{"a啊啊b", "a啊啊b"},
-		{"c🤔️🤔️d", "cd"},
+		{"c🤔️🤔️d", "c\ufe0f\ufe0fd"},
 		{"a⚡a", "aa"},
 		{"e.~f", "ef"},
+		{"a \u2003 b", "a--b"},
+		{"हिन्दी", "हिन्दी"},
+		{"m²", "m"},
 	}
 	for _, test := range tests {
 		assert.Equal(t, []byte(test.expect), CleanValue([]byte(test.param)), test.param)