This is the mail archive of the
gcc-regression@gcc.gnu.org
mailing list for the GCC project.
Re: 6 GCC regressions, 2 new, with your patch on 2002-07-20T23:26:42Z.
- From: Roger Sayle <roger at eyesopen dot com>
- To: Jan Hubicka <jh at suse dot cz>
- Cc: <gcc-regression at gcc dot gnu dot org>
- Date: Sat, 20 Jul 2002 22:06:27 -0600 (MDT)
- Subject: Re: 6 GCC regressions, 2 new, with your patch on 2002-07-20T23:26:42Z.
Hi Jan,
> There are 2 new failures, and 4 failures that existed before and after
> that patch; 0 failures have been fixed.
>
> The new failures are:
> mips-elf gcc.sum gcc.c-torture/execute/921029-1.c
> native gcc.sum gcc.dg/20010912-1.c
These look like they're yours. I'm guessing its a bug in cselib or
your new cprop code. In 20010912-1.c, the routine foo is miscompiled:
>> int foo (int x, char *y)
>> {
>> int a;
>> char *b = y;
>> a = bar (x, &y);
>> if (a)
>> {
>> y = b;
^
This assignment is incorrectly removed. The call to "bar" changes
the value of y, which isn't correctly noticed by cprop...
>> a = baz (x, &y);
>> }
>> if (a)
>> return a;
>>
>> baz (x, &y);
>> return 0;
>> }
I hope this helps.
Roger
--