SF patch #1467512, fix double free with triple quoted string in standard build.
This was the result of inconsistent use of PyMem_* and PyObject_* allocators.
By changing to use PyObject_* allocator almost everywhere, this removes
the inconsistency.
diff --git a/Parser/tokenizer.c b/Parser/tokenizer.c
index 001d31a..469de27 100644
--- a/Parser/tokenizer.c
+++ b/Parser/tokenizer.c
@@ -105,7 +105,7 @@
static struct tok_state *
tok_new(void)
{
- struct tok_state *tok = PyMem_NEW(struct tok_state, 1);
+ struct tok_state *tok = PyMem_MALLOC(sizeof(struct tok_state));
if (tok == NULL)
return NULL;
tok->buf = tok->cur = tok->end = tok->inp = tok->start = NULL;
@@ -721,7 +721,7 @@
if (converted == NULL)
goto error_nomem;
- PyMem_FREE(*inp);
+ PyObject_FREE(*inp);
*inp = converted;
if (tok->encoding != NULL)
PyObject_FREE(tok->encoding);
@@ -781,12 +781,12 @@
if (new == NULL)
tok->done = E_INTR;
else if (*new == '\0') {
- PyMem_FREE(new);
+ PyObject_FREE(new);
tok->done = E_EOF;
}
#if !defined(PGEN) && defined(Py_USING_UNICODE)
else if (tok_stdin_decode(tok, &new) != 0)
- PyMem_FREE(new);
+ PyObject_FREE(new);
#endif
else if (tok->start != NULL) {
size_t start = tok->start - tok->buf;
@@ -798,7 +798,7 @@
if (buf == NULL) {
PyObject_FREE(tok->buf);
tok->buf = NULL;
- PyMem_FREE(new);
+ PyObject_FREE(new);
tok->done = E_NOMEM;
return EOF;
}
@@ -806,7 +806,7 @@
tok->cur = tok->buf + oldlen;
tok->line_start = tok->cur;
strcpy(tok->buf + oldlen, new);
- PyMem_FREE(new);
+ PyObject_FREE(new);
tok->inp = tok->buf + newlen;
tok->end = tok->inp + 1;
tok->start = tok->buf + start;