On 09/12/2017 12:48 PM, Joseph Myers wrote:
I'd be concerned about the possibility of a qsort implementation that
calls the comparison function with two pointers to the same object (as far
as I can tell, it's valid for qsort to do that). That is, I think you
need to check for the two DECLs being the same DECL, before asserting
their names are different.
I suppose we can drop the assert. That does leave it returning +1 in
the case you're concerned about, but I don't really see the need to tell
such a stupid qsort that the things are unordered.
nathan
--
Nathan Sidwell
2017-09-15 Nathan Sidwell <nat...@acm.org>
* c-decl.c (field_decl_cmp): No need to handle NULL or TYPE_DECLs.
Index: c-decl.c
===================================================================
--- c-decl.c (revision 252833)
+++ c-decl.c (working copy)
@@ -7845,19 +7845,10 @@ warn_cxx_compat_finish_struct (tree fiel
static int
field_decl_cmp (const void *x_p, const void *y_p)
{
- const tree *const x = (const tree *) x_p;
- const tree *const y = (const tree *) y_p;
+ const tree x = *(const tree *) x_p;
+ const tree y = *(const tree *) y_p;
- if (DECL_NAME (*x) == DECL_NAME (*y))
- /* A nontype is "greater" than a type. */
- return (TREE_CODE (*y) == TYPE_DECL) - (TREE_CODE (*x) == TYPE_DECL);
- if (DECL_NAME (*x) == NULL_TREE)
- return -1;
- if (DECL_NAME (*y) == NULL_TREE)
- return 1;
- if (DECL_NAME (*x) < DECL_NAME (*y))
- return -1;
- return 1;
+ return DECL_NAME (x) < DECL_NAME (y) ? -1 : +1;
}
/* Fill in the fields of a RECORD_TYPE or UNION_TYPE node, T.
