8000 gh-126004: Fix positions handling in `codecs.replace_errors` (#127674) · python/cpython@225296c · GitHub
[go: up one dir, main page]
Skip to content

Commit 225296c

Browse files
picnixz< 8000 /a>picnixz
authored
gh-126004: Fix positions handling in codecs.replace_errors (#127674)
This fixes how `PyCodec_ReplaceErrors` handles the `start` and `end` attributes of `UnicodeError` objects via the `_PyUnicodeError_GetParams` helper.
1 parent 70dcc84 commit 225296c

File tree

3 files changed

+27
-25
lines changed

3 files changed

+27
-25
lines changed
< 8000 div role="region" aria-labelledby="heading-:Rtlab:" class="position-relative" style="contain:layout">

‎Lib/test/test_capi/test_codecs.py‎

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -839,7 +839,8 @@ def test_codec_ignore_errors_handler(self):
839839

840840
def test_codec_replace_errors_handler(self):
841841
handler = _testcapi.codec_replace_errors
842-
self.do_test_codec_errors_handler(handler, self.all_unicode_errors)
842+
self.do_test_codec_errors_handler(handler, self.all_unicode_errors,
843+
safe=True)
843844

844845
def test_codec_xmlcharrefreplace_errors_handler(self):
845846
handler = _testcapi.codec_xmlcharrefreplace_errors

‎Misc/NEWS.d/next/Core_and_Builtins/2024-12-06-11-32-58.gh-issue-126004.CYAwTB.rst‎

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,3 @@
1+
Fix handling of :attr:`UnicodeError.start` and :attr:`UnicodeError.end`
2+
values in the :func:`codecs.replace_errors` error handler. Patch by Bénédikt
3+
Tran.

‎Python/codecs.c‎

Lines changed: 22 additions & 24 deletions
Original file line numberDiff line numberDiff line change
@@ -702,48 +702,46 @@ PyObject *PyCodec_IgnoreErrors(PyObject *exc)
702702

703703
PyObject *PyCodec_ReplaceErrors(PyObject *exc)
704704
{
705-
Py_ssize_t start, end, i, len;
705+
Py_ssize_t start, end, slen;
706706

707707
if (PyObject_TypeCheck(exc, (PyTypeObject *)PyExc_UnicodeEncodeError)) {
708-
PyObject *res;
709-
Py_UCS1 *outp;
710-
if (PyUnicodeEncodeError_GetStart(exc, &start))
708+
if (_PyUnicodeError_GetParams(exc, NULL, NULL,
709+
&start, &end, &slen, false) < 0) {
711710
return NULL;
712-
if (PyUnicodeEncodeError_GetEnd(exc, &end))
713-
return NULL;
714-
len = end - start;
715-
res = PyUnicode_New(len, '?');
716-
if (res == NULL)
711+
}
712+
PyObject *res = PyUnicode_New(slen, '?');
713+
if (res == NULL) {
717714
return NULL;
715+
}
718716
assert(PyUnicode_KIND(res) == PyUnicode_1BYTE_KIND);
719-
outp = PyUnicode_1BYTE_DATA(res);
720-
for (i = 0; i < len; ++i)
721-
outp[i] = '?';
717+
Py_UCS1 *outp = PyUnicode_1BYTE_DATA(res);
718+
memset(outp, '?', sizeof(Py_UCS1) * slen);
722719
assert(_PyUnicode_CheckConsistency(res, 1));
723720
return Py_BuildValue("(Nn)", res, end);
724721
}
725722
else if (PyObject_TypeCheck(exc, (PyTypeObject *)PyExc_UnicodeDecodeError)) {
726-
if (PyUnicodeDecodeError_GetEnd(exc, &end))
723+
if (_PyUnicodeError_GetParams(exc, NULL, NULL,
724+
NULL, &end, NULL, true) < 0) {
727725
return NULL;
726+
}
728727
return Py_BuildValue("(Cn)",
729728
(int)Py_UNICODE_REPLACEMENT_CHARACTER,
730729
end);
731730
}
732731
else if (PyObject_TypeCheck(exc, (PyTypeObject *)PyExc_UnicodeTranslateError)) {
733-
PyObject *res;
734-
Py_UCS2 *outp;
735-
if (PyUnicodeTranslateError_GetStart(exc, &start))
736-
return NULL;
737-
if (PyUnicodeTranslateError_GetEnd(exc, &end))
732+
if (_PyUnicodeError_GetParams(exc, NULL, NULL,
733+
&start, &end, &slen, false) < 0) {
738734
return NULL;
739-
len = end - start;
740-
res = PyUnicode_New(len, Py_UNICODE_REPLACEMENT_CHARACTER);
741-
if (res == NULL)
735+
}
736+
PyObject *res = PyUnicode_New(slen, Py_UNICODE_REPLACEMENT_CHARACTER);
737+
if (res == NULL) {
742738
return NULL;
743-
assert(PyUnicode_KIND(res) == PyUnicode_2BYTE_KIND);
744-
outp = PyUnicode_2BYTE_DATA(res);
745-
for (i = 0; i < len; i++)
739+
}
740+
assert(slen == 0 || PyUnicode_KIND(res) == PyUnicode_2BYTE_KIND);
741+
Py_UCS2 *outp = PyUnicode_2BYTE_DATA(res);
742+
for (Py_ssize_t i = 0; i < slen; ++i) {
746743
outp[i] = Py_UNICODE_REPLACEMENT_CHARACTER;
744+
}
747745
assert(_PyUnicode_CheckConsistency(res, 1));
748746
return Py_BuildValue("(Nn)", res, end);
749747
}

0 commit comments

Comments
 (0)
0