Raymond Hettinger <[email protected]> added the comment:
#511 and #512 are false positives. The "kind" variable is indeed uninitialized
in the bytes_writer case:
else if (bytes_writer) {
*bytes_str = _PyBytesWriter_Prepare(bytes_writer, *bytes_str, strlen);
if (*bytes_str == NULL) {
Py_DECREF(scratch);
return -1;
}
}
However, the flagged sections cannot be called in the bytes_writer case:
if (bytes_writer) { <=="kind" not used here
char *p = *bytes_str + strlen;
WRITE_DIGITS(p);
assert(p == *bytes_str);
}
else if (kind == PyUnicode_1BYTE_KIND) { <== bytes_write is false
Py_UCS1 *p;
WRITE_UNICODE_DIGITS(Py_UCS1);
}
else if (kind == PyUnicode_2BYTE_KIND) { <== bytes_write is false
Py_UCS2 *p;
WRITE_UNICODE_DIGITS(Py_UCS2);
}
else {
Py_UCS4 *p;
assert (kind == PyUnicode_4BYTE_KIND);
WRITE_UNICODE_DIGITS(Py_UCS4);
}
----------
_______________________________________
Python tracker <[email protected]>
<https://bugs.python.org/issue46280>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
