Strip ANSI control codes when writing to the warnings file (#11624)

Co-authored-by: Adam Turner <[email protected]>

CHANGES.rst

@@ -20,6 +20,9 @@ Bugs fixed
   Patch by Vinay Sajip.
 * #11622: Ensure that the order of keys in ``searchindex.js`` is deterministic.
   Patch by Pietro Albini.
+* #11617: ANSI control sequences are stripped from the output when writing to
+  a warnings file with :option:`-w <sphinx-build -w>`.
+  Patch by Bénédikt Tran.
 
 Testing
 -------

doc/man/sphinx-build.rst

@@ -195,6 +195,10 @@ Options
 
    Write warnings (and errors) to the given file, in addition to standard error.
 
+   .. versionchanged:: 7.3
+
+      ANSI control sequences are stripped when writing to *file*.
+
 .. option:: -W
 
    Turn warnings into errors.  This means that the build stops at the first

sphinx/cmd/build.py

@@ -21,7 +21,7 @@
 from sphinx.application import Sphinx
 from sphinx.errors import SphinxError, SphinxParallelError
 from sphinx.locale import __
-from sphinx.util import Tee
+from sphinx.util._io import TeeStripANSI
 from sphinx.util.console import (  # type: ignore[attr-defined]
     color_terminal,
     nocolor,
@@ -34,6 +34,11 @@
 
 if TYPE_CHECKING:
     from collections.abc import Sequence
+    from typing import Protocol
+
+    class SupportsWrite(Protocol):
+        def write(self, text: str, /) -> int | None:
+            ...
 
 
 def handle_exception(
@@ -266,11 +271,12 @@ def _parse_logging(
         try:
             warnfile = path.abspath(warnfile)
             ensuredir(path.dirname(warnfile))
+            # the caller is responsible for closing this file descriptor
             warnfp = open(warnfile, 'w', encoding="utf-8")  # NoQA: SIM115
         except Exception as exc:
             parser.error(__('cannot open warning file %r: %s') % (
                 warnfile, exc))
-        warning = Tee(warning, warnfp)  # type: ignore[assignment]
+        warning = TeeStripANSI(warning, warnfp)  # type: ignore[assignment]
         error = warning
 
     return status, warning, error, warnfp
@@ -334,6 +340,10 @@ def build_main(argv: Sequence[str]) -> int:
     except (Exception, KeyboardInterrupt) as exc:
         handle_exception(app, args, exc, args.error)
         return 2
+    finally:
+        if warnfp is not None:
+            # close the file descriptor for the warnings file opened by Sphinx
+            warnfp.close()
 
 
 def _bug_report_info() -> int:

sphinx/util/_io.py

@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 from typing import TYPE_CHECKING
 
 from sphinx.util.console import _strip_escape_sequences
@@ -6,7 +8,7 @@
     from typing import Protocol
 
     class SupportsWrite(Protocol):
-        def write(self, text: str, /) -> None:
+        def write(self, text: str, /) -> int | None:
             ...
 
 

tests/test_build.py

@@ -2,11 +2,14 @@
 
 import os
 import shutil
+from contextlib import contextmanager
+from pathlib import Path
 from unittest import mock
 
 import pytest
 from docutils import nodes
 
+from sphinx.cmd.build import build_main
 from sphinx.errors import SphinxError
 
 
@@ -133,3 +136,29 @@ def test_image_glob(app, status, warning):
     assert doctree[0][3][0]['candidates'] == {'application/pdf': 'subdir/svgimg.pdf',
                                               'image/svg+xml': 'subdir/svgimg.svg'}
     assert doctree[0][3][0]['uri'] == 'subdir/svgimg.*'
+
+
+@contextmanager
+def force_colors():
+    forcecolor = os.environ.get('FORCE_COLOR', None)
+
+    try:
+        os.environ['FORCE_COLOR'] = '1'
+        yield
+    finally:
+        if forcecolor is None:
+            os.environ.pop('FORCE_COLOR', None)
+        else:
+            os.environ['FORCE_COLOR'] = forcecolor
+
+
+def test_log_no_ansi_colors(tmp_path):
+    with force_colors():
+        wfile = tmp_path / 'warnings.txt'
+        srcdir = Path(__file__).parent / 'roots/test-nitpicky-warnings'
+        argv = list(map(str, ['-b', 'html', srcdir, tmp_path, '-n', '-w', wfile]))
+        retcode = build_main(argv)
+        assert retcode == 0
+
+        content = wfile.read_text(encoding='utf8')
+        assert '\x1b[91m' not in content