This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: Regressions with "bfd_arch_get_compatible"
- From: Hans-Peter Nilsson <hans-peter dot nilsson at axis dot com>
- To: hjl dot tools at gmail dot com
- Cc: hp at axis dot com, amodra at gmail dot com, binutils at sourceware dot org
- Date: Sun, 5 Jun 2011 02:07:57 +0200
- Subject: Re: Regressions with "bfd_arch_get_compatible"
> Date: Sat, 4 Jun 2011 11:18:33 -0700
> From: "H.J. Lu" <hjl.tools@gmail.com>
> On Sat, Jun 4, 2011 at 5:16 AM, Hans-Peter Nilsson
> <hans-peter.nilsson@axis.com> wrote:
> >> Date: Sat, 4 Jun 2011 13:33:58 +0930
> >> From: Alan Modra <amodra@gmail.com>
> >
> >> ? ? ? * archures.c (bfd_arch_get_compatible): If one arch is unknown,
> >> ? ? ? return the other arch.
> >> ? ? ? * elfcode.h (elf_object_p): Allow explicit match to generic ELF
> >> ? ? ? target.
> >
> > Looks like the target lost leverage in bfd_arch_get_compatible.
> >
> > Regressions for cris-elf:
> >
>
> elf_object_p change also breaks Linux/x86-64. I checked in a patch
> to revert it. Please check if it fixes cris-elf.
The partial reversion did fix the regressions, but I also have
to point out that reverting like that seems like a breach of
protocol as I understand it.
brgds, H-P