=== modified file 'Lib/test/test_pep263.py' --- Lib/test/test_pep263.py 2009-03-02 23:31:26 +0000 +++ Lib/test/test_pep263.py 2009-10-04 22:37:18 +0000 @@ -36,6 +36,14 @@ exec(c, d) self.assertEquals(d['\xc6'], '\xc6') + def test_issue3297(self): + c = compile("a, b = '\U0001010F', '\\U0001010F'", "dummy", "exec") + d = {} + exec(c, d) + self.assertEqual(d['a'], d['b']) + self.assertEqual(len(d['a']), len(d['b'])) + self.assertEqual(ascii(d['a']), ascii(d['b'])) + def test_main(): support.run_unittest(PEP263Test) === modified file 'Python/ast.c' --- Python/ast.c 2009-09-11 21:36:20 +0000 +++ Python/ast.c 2009-10-04 20:32:43 +0000 @@ -3240,10 +3240,11 @@ u = NULL; } else { /* check for integer overflow */ - if (len > PY_SIZE_MAX / 4) + if (len > PY_SIZE_MAX / 6) return NULL; - /* "\XX" may become "\u005c\uHHLL" (12 bytes) */ - u = PyBytes_FromStringAndSize((char *)NULL, len * 4); + /* "ä" (2 bytes) may become "\U000000E4" (10 bytes), or 1:5 + "\ä" (3 bytes) may become "\u005c\U000000E4" (16 bytes), or ~1:6 */ + u = PyBytes_FromStringAndSize((char *)NULL, len * 6); if (u == NULL) return NULL; p = buf = PyBytes_AsString(u); @@ -3260,20 +3261,24 @@ PyObject *w; char *r; Py_ssize_t rn, i; - w = decode_utf8(c, &s, end, "utf-16-be"); + w = decode_utf8(c, &s, end, "utf-32-be"); if (w == NULL) { Py_DECREF(u); return NULL; } r = PyBytes_AS_STRING(w); rn = Py_SIZE(w); - assert(rn % 2 == 0); - for (i = 0; i < rn; i += 2) { - sprintf(p, "\\u%02x%02x", + assert(rn % 4 == 0); + for (i = 0; i < rn; i += 4) { + sprintf(p, "\\U%02x%02x%02x%02x", r[i + 0] & 0xFF, - r[i + 1] & 0xFF); - p += 6; + r[i + 1] & 0xFF, + r[i + 2] & 0xFF, + r[i + 3] & 0xFF); + p += 10; } + /* Should be impossible to overflow */ + assert(p - buf <= Py_SIZE(u)); Py_DECREF(w); } else { *p++ = *s++;