Richard Guenther wrote:
> On Sun, Jul 12, 2009 at 9:41 PM, Jack Howarth<howa...@bromo.med.uc.edu> wrote:
>> I am seeing a new failure on x86_64-apple-darwin10 in current gcc 4.4
>> branch
>> that wasn't present in the gcc 4.4.0 release. At -m32, the testsuite
>> failure...
>>
>> FAIL: gcc.c-torture/compile/20000804-1.c -O0 (test for excess errors)
Do you know what this test is supposed to be doing?
> $ cat /gnu/gcc/gcc/gcc/testsuite/gcc.c-torture/compile/20000804-1.c
> /* This does not work on m68hc11 or h8300 due to the use of an asm
> statement to force a 'long long' (64-bits) to go in a register. */
> /* { dg-do assemble } */
> /* { dg-skip-if "" { { i?86-*-* x86_64-*-* } && ilp32 } { "-fpic" "-fPIC" } {
> ""
> } } */
> /* { dg-skip-if "PIC default" { i?86-*-darwin* } { "*" } { "" } } */
> /* { dg-skip-if "No 64-bit registers" { m32c-*-* } { "*" } { "" } } */
> /* { dg-xfail-if "" { m6811-*-* m6812-*-* h8300-*-* } { "*" } { "" } } */
>
> /* Copyright (C) 2000, 2003 Free Software Foundation */
> __complex__ long long f ()
> {
> int i[99];
> __complex__ long long v;
>
> v += f ();
Undefined behaviour, no?
cheers,
DaveK
