This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Fix ICEs in expand_builtin_expect (PR middle-end/31959)


On Wed, Jun 20, 2007 at 03:44:48AM +0900, Andrew Pinski wrote:
> >--- gcc/testsuite/gcc.dg/pr31959-1.c.jj 2007-06-19 18:44:16.000000000 +0200
> >+++ gcc/testsuite/gcc.dg/pr31959-1.c    2007-06-19 18:49:57.000000000 +0200
> >@@ -0,0 +1,26 @@
> >+/* PR middle-end/31959 */
> >+/* { dg-do compile } */
> >+/* { dg-options "-Werror -O2" } */
> >+/* { dg-message "warnings being treated as errors" "" {target "*-*-*"} 0 
> >} */
> >+
> >+struct A { int i; };
> >+
> >+static inline struct A *
> >+baz (struct A *x)
> >+{
> >+  return x;
> >+}
> >+static inline struct A *
> >+bar (const struct A *x)
> >+{
> >+  return baz ((struct A *) x); /* { dg-error "discards qualifiers" } */
> >+}
> >+
> >+void *
> >+foo (struct A *x, int y)
> >+{
> >+  void *p = (void *) 0;
> >+  if (__builtin_expect (y >= 6, 0))
> >+    p = bar (x);
> >+  return p;
> >+}
> 
> This testcase really should not, produce a warning/error at all.
> Please see PR 29478 about that issue.

Ok, I withdraw the pr31959-1.c testcase from the patch,
pr31959-2.c is enough to test it (after renaming it to pr31959.c).

	Jakub


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]