https://github.com/python/cpython/commit/a64b7ecba4db5480504a1a83ae6e80c81472fe16
commit: a64b7ecba4db5480504a1a83ae6e80c81472fe16
branch: main
author: Victor Stinner <[email protected]>
committer: vstinner <[email protected]>
date: 2026-09-20T16:25:20+02:00
summary:
gh-157710: Avoid resize in PyUnicodeWriter_Finish() for singleton (#157862)
PyBytesWriter_FinishWithSize() and PyUnicodeWriter_Finish() now
discard the output string instead of resizing it if the single
byte/character singleton is used.
files:
M Lib/test/test_capi/test_unicode.py
M Objects/bytesobject.c
M Objects/unicode_writer.c
diff --git a/Lib/test/test_capi/test_unicode.py
b/Lib/test/test_capi/test_unicode.py
index f9e7b5ab041962..f13ad6f428ec09 100644
--- a/Lib/test/test_capi/test_unicode.py
+++ b/Lib/test/test_capi/test_unicode.py
@@ -2050,12 +2050,19 @@ def test_singletons(self):
for size in (0, 123):
with self.subTest(size=size):
writer = self.create_writer(size)
+ writer.write_utf8(b'utf8', 0)
+ writer.write_ascii(b'ascii', 0)
+ writer.write_widechar(b'wstr', 0)
+ writer.write_ucs4(b'ucs4', 0)
+ writer.write_substring('text', 0, 0)
self.assertIs(writer.finish(), '')
for ch in range(256):
with self.subTest(ch=ch):
ch = chr(ch)
writer = self.create_writer(0)
+ # Use PyUnicodeWriter_WriteSubstring() to avoid the read-only
+ # buffer optimization
writer.write_substring(ch + 'xxx', 0, 1)
self.assertIs(writer.finish(), ch)
diff --git a/Objects/bytesobject.c b/Objects/bytesobject.c
index 7fefd64eefb2ac..91cbfa23e30b20 100644
--- a/Objects/bytesobject.c
+++ b/Objects/bytesobject.c
@@ -3938,12 +3938,6 @@ PyBytesWriter_FinishWithSize(PyBytesWriter *writer,
Py_ssize_t size)
}
}
else {
- if (size != PyBytes_GET_SIZE(writer->obj)) {
- if (_PyBytes_Resize(&writer->obj, size)) {
- goto error;
- }
- }
-
if (size == 1) {
// Get the single byte singleton
unsigned char ch = PyBytes_AS_STRING(writer->obj)[0];
@@ -3951,6 +3945,11 @@ PyBytesWriter_FinishWithSize(PyBytesWriter *writer,
Py_ssize_t size)
assert(_Py_IsImmortal(op));
Py_SETREF(writer->obj, op);
}
+ else if (size != PyBytes_GET_SIZE(writer->obj)) {
+ if (_PyBytes_Resize(&writer->obj, size)) {
+ goto error;
+ }
+ }
}
result = writer->obj;
diff --git a/Objects/unicode_writer.c b/Objects/unicode_writer.c
index 92e0db08b9de00..751fca9948598f 100644
--- a/Objects/unicode_writer.c
+++ b/Objects/unicode_writer.c
@@ -609,8 +609,6 @@ _PyUnicodeWriter_WriteLatin1String(_PyUnicodeWriter *writer,
PyObject *
_PyUnicodeWriter_Finish(_PyUnicodeWriter *writer)
{
- PyObject *str;
-
#ifdef Py_DEBUG
// Check for buffer overflow
if (writer->buffer != NULL) {
@@ -625,23 +623,28 @@ _PyUnicodeWriter_Finish(_PyUnicodeWriter *writer)
}
#endif
- if (writer->pos == 0) {
- Py_CLEAR(writer->buffer);
- return _PyUnicode_GetEmpty();
- }
-
- str = writer->buffer;
+ PyObject *str = writer->buffer;
writer->buffer = NULL;
- if (writer->readonly) {
- assert(PyUnicode_GET_LENGTH(str) == writer->pos);
- assert(_PyUnicode_CheckConsistency(str, 1));
- return str;
+ Py_ssize_t final_size = writer->pos;
+ if (final_size == 0) {
+ PyObject *empty = _PyUnicode_GetEmpty();
+ Py_XDECREF(str); // writer->buffer can be NULL if the position is 0
+ return empty;
}
- if (PyUnicode_GET_LENGTH(str) != writer->pos) {
- PyObject *str2;
- str2 = _PyUnicode_ResizeCompact(str, writer->pos);
+ Py_ssize_t length = PyUnicode_GET_LENGTH(str);
+ if (final_size == 1 && PyUnicode_KIND(str) == PyUnicode_1BYTE_KIND) {
+ assert(length >= 1);
+ const Py_UCS1 *data = PyUnicode_1BYTE_DATA(str);
+ Py_UCS1 ch = data[0];
+ PyObject *latin1_char = _Py_LATIN1_CHR(ch);
+ Py_DECREF(str);
+ return latin1_char;
+ }
+
+ if (!writer->readonly && length != final_size) {
+ PyObject *str2 = _PyUnicode_ResizeCompact(str, final_size);
if (str2 == NULL) {
Py_DECREF(str);
return NULL;
@@ -650,7 +653,7 @@ _PyUnicodeWriter_Finish(_PyUnicodeWriter *writer)
}
assert(_PyUnicode_CheckConsistency(str, 1));
- return _PyUnicode_Result(str);
+ return str;
}
_______________________________________________
Python-checkins mailing list -- [email protected]
To unsubscribe send an email to [email protected]
https://mail.python.org/mailman3//lists/python-checkins.python.org
Member address: [email protected]
