From 72471cc8c6c470c9f355e51585ea6ffe37f6b868 Mon Sep 17 00:00:00 2001
From: harbinresearcher <hsyjy01@gmail.com>
Date: Wed, 7 Oct 2026 10:07:27 +0800
Subject: [PATCH] Accept normalized UTF-8 declarations with a BOM

---
 babel/util.py      |  3 ++-
 tests/test_util.py | 21 +++++++++++++++++++++
 2 files changed, 23 insertions(+), 1 deletion(-)

diff --git a/babel/util.py b/babel/util.py
index a2bf728cc..ca53baba1 100644
--- a/babel/util.py
+++ b/babel/util.py
@@ -91,7 +91,8 @@ def parse_encoding(fp: IO[bytes]) -> str | None:
         if has_bom:
             if m:
                 magic_comment_encoding = m.group(1).decode('latin-1')
-                if magic_comment_encoding != 'utf-8':
+                normalized_encoding = magic_comment_encoding.lower().replace('_', '-')
+                if normalized_encoding != 'utf-8' and not normalized_encoding.startswith('utf-8-'):
                     raise SyntaxError(f"encoding problem: {magic_comment_encoding} with BOM")
             return 'utf-8'
         elif m:
diff --git a/tests/test_util.py b/tests/test_util.py
index f51626e25..07056e938 100644
--- a/tests/test_util.py
+++ b/tests/test_util.py
@@ -12,6 +12,7 @@
 
 import __future__
 
+import codecs
 from io import BytesIO
 
 import pytest
@@ -73,6 +74,26 @@ def test_parse_encoding_non_ascii():
     assert parse_encoding('K\xf6ln') is None
 
 
+@pytest.mark.parametrize('encoding', ['utf-8', 'UTF-8', 'utf_8', 'UTF_8', 'utf-8-sig'])
+@pytest.mark.parametrize('prefix', [b'', b'# source file\n'])
+def test_parse_encoding_bom(encoding, prefix):
+    source = codecs.BOM_UTF8 + prefix + f'# coding: {encoding}\n'.encode('ascii')
+    compile(source, '<test>', 'exec')
+    stream = BytesIO(source)
+    stream.seek(3)
+    assert util.parse_encoding(stream) == 'utf-8'
+    assert stream.tell() == 3
+
+
+@pytest.mark.parametrize('encoding', ['latin-1', 'utf8', 'cp65001'])
+def test_parse_encoding_bom_incompatible(encoding):
+    stream = BytesIO(codecs.BOM_UTF8 + f'# coding: {encoding}\n'.encode('ascii'))
+    stream.seek(3)
+    with pytest.raises(SyntaxError, match=f'encoding problem: {encoding} with BOM'):
+        util.parse_encoding(stream)
+    assert stream.tell() == 3
+
+
 @pytest.mark.parametrize(('source', 'result'), [
     ('''
 from __future__ import print_function,