Support XID_Start/XID_Continue Unicode chars in C++ identifiers (#14026)

AUTHORS.rst

@@ -105,6 +105,7 @@ Contributors
 * Slawek Figiel -- additional warning suppression
 * Stefan Seefeld -- toctree improvements
 * Stefan van der Walt -- autosummary extension
+* Stefanos Carlström -- minor bug fix
 * Steve Piercy -- documentation improvements
 * Szymon Karpinski -- intersphinx improvements
 * \T. Powers -- HTML output improvements

CHANGES.rst

@@ -4,6 +4,10 @@ Release 8.3.0 (in development)
 Dependencies
 ------------
 
+* #14026: Added `regex >= 2025.7.29`_, to support parsing for Unicode classes.
+
+  .. _regex >= 2025.7.29: _https://github.com/mrabarnett/mrab-regex
+
 * #13786: Support `Docutils 0.22`_. Patch by Adam Turner.
 
   .. _Docutils 0.22: https://docutils.sourceforge.io/RELEASE-NOTES.html#release-0-22-2026-07-29

pyproject.toml

@@ -86,6 +86,7 @@ dependencies = [
     "packaging>=23.0",
     "colorama>=0.4.6; sys_platform == 'win32'",
     "ipython>=9.6.0",
+    "regex>=2025.7.29"
 ]
 dynamic = ["version"]
 

sphinx/util/cfamily.py

@@ -1,15 +1,15 @@
 """Utility functions common to the C and C++ domains."""
 
 from __future__ import annotations
 
-import re
+import regex as re
 from copy import deepcopy
 from typing import TYPE_CHECKING
 
 from docutils import nodes
 
 from sphinx import addnodes
 from sphinx.util import logging
 
 if TYPE_CHECKING:
     from collections.abc import Callable, Sequence
@@ -30,9 +30,13 @@
     (   # This 'extends' _anon_identifier_re with the ordinary identifiers,
         # make sure they are in sync.
         (~?\b[a-zA-Z_])  # ordinary identifiers
+    |   \p{XID_Start}    # Unicode-allowed starting characters for identifiers
     |   (@[a-zA-Z0-9_])  # our extension for names of anonymous entities
     )
-    [a-zA-Z0-9_]*\b
+    (
+        [a-zA-Z0-9_]     # ordinary identifiers
+    |   \p{XID_Continue} # Unicode-allowed continuing characters for identifiers
+    )*\b
     """,
     flags=re.VERBOSE,
 )

tests/test_domains/test_domain_cpp.py

@@ -860,6 +860,10 @@ def test_domain_cpp_ast_function_definitions() -> None:
 
     check('function', 'decltype(auto) f()', {1: 'f', 2: '1fv'})
 
+    # Test derived from https://github.com/sphinx-doc/sphinx/issues/14026
+    # Unicode identifiers
+    check('function', 'void f(int *const ξ)', {1: 'f__iPC', 2: '1fPCi'})
+
     # TODO: make tests for functions in a template, e.g., Test<int&&()>
     # such that the id generation for function type types is correct.