[gfortran, committed] Fix regressions introduced by patch for PR 25045

[Steve Kargl]

On Wed, Feb 15, 2006 at 01:05:40PM -0800, Steve Kargl wrote:
> On Wed, Feb 15, 2006 at 12:58:23PM -0800, H. J. Lu wrote:
>> On Wed, Feb 15, 2006 at 02:47:29PM -0500, Jakub Jelinek wrote:
>>> On line 825 I think there is nothing that can segfault, list is an
>>> automatic int variable.  I certainly can't reproduce what you are seeing,
>>> on at least {i386,x86_64,ppc,ppc64,s390,s390x,ia64}-linux, so I'd appreciate
>>> if you could find out what exactly segfaulted.
>>> 
>> 
>> Do you have this
>> 
>> http://gcc.gnu.org/ml/gcc-patches/2006-02/msg01249.html
> 
> Jakub just committed this patch to trunk.  I'm recompiling now
> to check if the segfault goes away.
> 

I have the above patch installed on i386-*-freebsd and amd64-*-freebsd.
The segfaults, reported by HJ Lu, are not present on i386-*-freebsd, but
the same tests fail on amd64-*-freebsd.  This appears to be a sizeof(int)
!= sizeof(long) type problemi, which probably means some memory is 
getting modified that shouldn't. ;(

-- 
Steve