changeset: 99487:376b100107ba branch: 3.5 user: Serhiy Storchaka date: Thu Dec 03 01:02:03 2015 +0200 files: Lib/test/test_unicode.py Misc/NEWS Objects/unicodeobject.c description: Issue #25709: Fixed problem with in-place string concatenation and utf-8 cache. diff -r f668f41b378c -r 376b100107ba Lib/test/test_unicode.py --- a/Lib/test/test_unicode.py Sat Dec 05 16:49:19 2015 -0800 +++ b/Lib/test/test_unicode.py Thu Dec 03 01:02:03 2015 +0200 @@ -2699,6 +2699,23 @@ self.assertTrue(astral >= bmp2) self.assertFalse(astral >= astral2) + @support.cpython_only + def test_pep393_utf8_caching_bug(self): + # Issue #25709: Problem with string concatenation and utf-8 cache + from _testcapi import getargs_s_hash + for k in 0x24, 0xa4, 0x20ac, 0x1f40d: + s = '' + for i in range(5): + # Due to CPython specific optimization the 's' string can be + # resized in-place. + s += chr(k) + # Parsing with the "s#" format code calls indirectly + # PyUnicode_AsUTF8AndSize() which creates the UTF-8 + # encoded string cached in the Unicode object. + self.assertEqual(getargs_s_hash(s), chr(k).encode() * (i + 1)) + # Check that the second call returns the same result + self.assertEqual(getargs_s_hash(s), chr(k).encode() * (i + 1)) + class StringModuleTest(unittest.TestCase): def test_formatter_parser(self): diff -r f668f41b378c -r 376b100107ba Misc/NEWS --- a/Misc/NEWS Sat Dec 05 16:49:19 2015 -0800 +++ b/Misc/NEWS Thu Dec 03 01:02:03 2015 +0200 @@ -10,9 +10,14 @@ Core and Builtins ----------------- -Library +- Issue #25709: Fixed problem with in-place string concatenation and utf-8 cache. + +Windows ------- +- Issue #25715: Python 3.5.1 installer shows wrong upgrade path and incorrect + logic for launcher detection. + What's New in Python 3.5.1 release candidate 1? =============================================== diff -r f668f41b378c -r 376b100107ba Objects/unicodeobject.c --- a/Objects/unicodeobject.c Sat Dec 05 16:49:19 2015 -0800 +++ b/Objects/unicodeobject.c Thu Dec 03 01:02:03 2015 +0200 @@ -722,6 +722,11 @@ } new_size = (struct_size + (length + 1) * char_size); + if (_PyUnicode_HAS_UTF8_MEMORY(unicode)) { + PyObject_DEL(_PyUnicode_UTF8(unicode)); + _PyUnicode_UTF8(unicode) = NULL; + _PyUnicode_UTF8_LENGTH(unicode) = 0; + } _Py_DEC_REFTOTAL; _Py_ForgetReference(unicode);