On 6/10/07, martin.v.loewis <[email protected]> wrote: > Author: martin.v.loewis > Date: Sun Jun 10 11:51:05 2007 > New Revision: 55850 > > Modified: python/branches/py3k-struni/Modules/unicodedata.c > ============================================================================== > --- python/branches/py3k-struni/Modules/unicodedata.c (original) > +++ python/branches/py3k-struni/Modules/unicodedata.c Sun Jun 10 11:51:05 > 2007 > @@ -515,7 +515,7 @@ > /* Hangul Decomposition adds three characters in > a single step, so we need atleast that much room. */ > if (space < 3) { > - Py_ssize_t newsize = PyString_GET_SIZE(result) + 10; > + Py_ssize_t newsize = PyUnicode_GET_SIZE(result) + 10; > space += 10; > if (PyUnicode_Resize(&result, newsize) == -1) > return NULL;
Martin, Is it a problem to add a constant value (10 in this case) to the new unicode object? Does the size of unicode (UCS-2 vs 4) matter? Does it matter if there are unicode escapes U+000...? Even if it doesn't matter in this specific example what are the cases where it does matter? (I'm remembering the block of code where we didn't allocate enough for U+000.... for ucs-4.) Thanks, n _______________________________________________ Python-3000-checkins mailing list [email protected] http://mail.python.org/mailman/listinfo/python-3000-checkins
