gh-63866: Add the "htmlcharrefreplace" error handler

Doc/library/html.rst

@@ -34,6 +34,25 @@ This module defines utilities to manipulate HTML.
 
    .. versionadded:: 3.4
 
+
+.. function:: htmlcharrefreplace_errors(exception)
+
+   Implements the ``htmlcharrefreplace`` error handling (for encoding only):
+   the unencodable character is replaced by the corresponding HTML named
+   character reference from :data:`html.entities.codepoint2name`, or by a
+   numeric character reference if there is no name for it.
+
+   This error handler is not registered by default, you should register it
+   with :func:`codecs.register_error`::
+
+      >>> import codecs, html
+      >>> codecs.register_error('htmlcharrefreplace',
+      ...                       html.htmlcharrefreplace_errors)
+      >>> '∀ x∈ℜ'.encode('ascii', 'htmlcharrefreplace')
+      b'∀ x∈ℜ'
+
+   .. versionadded:: next
+
 --------------
 
 Submodules in the ``html`` package are:

Lib/html/__init__.py

@@ -3,10 +3,10 @@
 """
 
 import re as _re
-from html.entities import html5 as _html5
+from html.entities import codepoint2name as _codepoint2name, html5 as _html5
 
 
-__all__ = ['escape', 'unescape']
+__all__ = ['escape', 'unescape', 'htmlcharrefreplace_errors']
 
 
 def escape(s, quote=True):
@@ -130,3 +130,20 @@ def unescape(s):
     if '&' not in s:
         return s
     return _charref.sub(_replace_charref, s)
+
+
+def htmlcharrefreplace_errors(exception):
+    """Implements the 'htmlcharrefreplace' error handling.
+
+    Replaces an unencodable character with the corresponding HTML named
+    character reference, or with a numeric character reference if there
+    is no name for it.
+    """
+    if not isinstance(exception, UnicodeEncodeError):
+        raise exception
+    replace = []
+    for c in exception.object[exception.start:exception.end]:
+        n = ord(c)
+        name = _codepoint2name.get(n)
+        replace.append(f'&{name};' if name is not None else f'&#{n};')
+    return ''.join(replace), exception.end

Lib/test/test_html.py

@@ -2,6 +2,7 @@
 Tests for the html module functions.
 """
 
+import codecs
 import html
 import unittest
 
@@ -98,6 +99,27 @@ def check_num(num, expected):
               'ÉricÉric&alphacentauriαcentauri')
         check('&co;', '&co;')
 
+    def test_htmlcharrefreplace_errors(self):
+        codecs.register_error('htmlcharrefreplace',
+                              html.htmlcharrefreplace_errors)
+        self.assertEqual('\u2200 x\u2208\u211c'.encode('ascii',
+                                                       'htmlcharrefreplace'),
+                         b'∀ x∈ℜ')
+        # Characters without a name are replaced with a numeric reference.
+        self.assertEqual('[$\xa5\u20a3\u20ac\U0001d56b]'.encode(
+                             'latin1', 'htmlcharrefreplace'),
+                         b'[$\xa5₣€𝕫]')
+        # Surrogates have no name either.
+        self.assertEqual('\udcff'.encode('ascii', 'htmlcharrefreplace'),
+                         b'�')
+
+    def test_htmlcharrefreplace_errors_bad_exception(self):
+        with self.assertRaises(UnicodeDecodeError):
+            html.htmlcharrefreplace_errors(
+                UnicodeDecodeError('ascii', b'\xff', 0, 1, 'ordinal'))
+        with self.assertRaises(TypeError):
+            html.htmlcharrefreplace_errors(TypeError('spam'))
+
 
 if __name__ == '__main__':
     unittest.main()

Misc/NEWS.d/next/Library/2026-08-05-13-05-00.gh-issue-63866.htmlCR.rst

@@ -0,0 +1,4 @@
+Add :func:`html.htmlcharrefreplace_errors` which implements the
+``htmlcharrefreplace`` error handler: an unencodable character is replaced
+with the corresponding HTML named or numeric character reference.  It is not
+registered by default.