From 61e40bd897da8ab4bf2dffe817d0163e984c1e40 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Martin=20v=2E=20L=C3=B6wis?= Date: Sat, 17 Apr 2004 19:36:48 +0000 Subject: Special case normalization of empty strings. Fixes #924361. Backported to 2.3. --- Lib/test/test_unicodedata.py | 1 + Modules/unicodedata.c | 7 +++++++ 2 files changed, 8 insertions(+) diff --git a/Lib/test/test_unicodedata.py b/Lib/test/test_unicodedata.py index 61b4ffb..8157fb3 100644 --- a/Lib/test/test_unicodedata.py +++ b/Lib/test/test_unicodedata.py @@ -170,6 +170,7 @@ class UnicodeFunctionsTest(UnicodeDatabaseTest): def test_normalize(self): self.assertRaises(TypeError, self.db.normalize) self.assertRaises(ValueError, self.db.normalize, 'unknown', u'xx') + self.assertEqual(self.db.normalize('NFKC', u''), u'') # The rest can be found in test_normalization.py # which requires an external file. diff --git a/Modules/unicodedata.c b/Modules/unicodedata.c index 311db29..ba218a3 100644 --- a/Modules/unicodedata.c +++ b/Modules/unicodedata.c @@ -515,6 +515,13 @@ unicodedata_normalize(PyObject *self, PyObject *args) &form, &PyUnicode_Type, &input)) return NULL; + if (PyUnicode_GetSize(input) == 0) { + /* Special case empty input strings, since resizing + them later would cause internal errors. */ + Py_INCREF(input); + return input; + } + if (strcmp(form, "NFC") == 0) return nfc_nfkc(input, 0); if (strcmp(form, "NFKC") == 0) -- cgit v0.12