develooper Front page | perl.perl5.porters | Postings from August 2006

Re: Solaris gcc optimisation failures; was RE: [PATCH pp_hot.c] test failure in op/dor.t

Thread Previous
From:
Steve Peters
Date:
August 23, 2006 06:58
Subject:
Re: Solaris gcc optimisation failures; was RE: [PATCH pp_hot.c] test failure in op/dor.t
Message ID:
20060823135758.GA3097@kirk.peters.homeunix.org
On Wed, Aug 23, 2006 at 11:49:36AM +0100, Robin Barker wrote:
> > From: Rafael Garcia-Suarez [mailto:rgarciasuarez@gmail.com]
> > Sent: 18 August 2006 16:48
> > Subject: Re: [PATCH pp_hot.c] test failure in op/dor.t
> 
> > On 18/08/06, Robin Barker <Robin.Barker@npl.co.uk> wrote:
> > > When testing 5.9.4 on solaris, I had a failure in t/op/dor.t
> 
> > I end up suspecting the optimizer of gcc 4.1.1 to be really bad. 
> > I saw other errors. Can you reproduce it without -O in "optimize" ?
> > I applied your patch as change #28735, thanks.
> 
> I tried building and testing with different -O<n>
> 
> The op/dor.t failure occurs with -O1 but not -O0 or -O2 !
> 
> I see a failure in ext/POSIX/t/sysconf.t 
> (a segmentation fault after "ok 39")
> this one occurs with -O2 but not -O0 or -O1.

I was also getting the coredump in ext/POSIX/t/sysconf.t until change
#28731.  Can you please try again with a bleadperl more recent than that?

Steve Peters
steve@fisharerojo.org



Thread Previous


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About