This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [ia64] Cleanup unwind info annotations
On Thu, 2010-09-23 at 13:51 -0700, Richard Henderson wrote:
> On 09/23/2010 09:27 AM, Richard Henderson wrote:
> > On 09/22/2010 09:51 AM, Steve Ellcey wrote:
> >> I don't have that exact version but I built 164472 (no local changes)
> >> and that worked fine with no comparision failures. I did recently have
> >> to update the GCC in my build environment to fix a problem caused by a
> >> bug in GCC 4.3.3.
> >
> > Good to know. I updated binutils and successfully performed a 4-stage
> > build of c,c++ with gcc3.4 as the bootstrap compiler.
> >
> > I'll try again with the unwind info patch.
>
> I've completed testing on ia64-linux.
>
> I've bootstrapped the compiler with and without the patch. I used
> BOOT_CFLAGS='-O2 -g --save-temps' and then compared the .s files
> from stage2 (which gets -gtoggle and thus -g0, thus ignoring the
> debug differences caused by different build paths). There were no
> differences at all, except those expected by the patched ia64.c.
>
> Thus I conclude that the unwind directives are identical with and
> without the patch.
>
> Is the patch ok?
>
>
> r~
I'm confused. Which patch are we talking about here? Do we think
things will work on IA64 HP-UX now? My comments about testing were on
sources without your patch included, I was just confirming that I could
build IA64 Linux (without your patch and without any local changes)
without getting the diff errors that you got during bootstrap.
Steve Ellcey
sje@cup.hp.com