This is the mail archive of the
mailing list for the GCC project.
Re: [RFA] make bootstrap-debug default, add broader -fcompare-debug testing options
- From: Richard Guenther <richard dot guenther at gmail dot com>
- To: Alexandre Oliva <aoliva at redhat dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Wed, 17 Jun 2009 13:16:57 +0200
- Subject: Re: [RFA] make bootstrap-debug default, add broader -fcompare-debug testing options
- References: <email@example.com>
On Wed, Jun 17, 2009 at 9:59 AM, Alexandre Oliva<firstname.lastname@example.org> wrote:
> This patch implements richi's and RTH's suggestions, and it requires
> review to some command-line flag processing changes.
> It makes bootstrap-debug by default, building stage2 without -g and
> stage3 with it, and using compare-debug to compare.
Would have been nice if the above was posted as a separate patch ;)
> It also introduces bootstrap-debug-big and bootstrap-debug-lean to
> perform -fcompare-debug-like testing.
> -big outputs the final insn dumps that -fcompare-debug uses during
> stage2 and stage3, so that the compare-debug script compares them.
> -lean saves the disk space to hold all such dumps at the same time (it's
> a lot) and uses -fcompare-debug proper, generating dumps that should be
> equivalent to stage2's during stage3.
> In theory, the former might be faster, but the extra disk space and the
> fact that stage2 is compiled without optimization might actually make it
> slower. ?I haven't compared them myself.
> Implementing -big required some ugly changes to gcc.c and toplev.c: we
> want the final insn dump to be named after the compiler output, so that
> compare-debug can easily find it, but adding -fdump-final-insns=$@.gkd
> to BOOT_CFLAGS feels too risky. ?So I tweaked the compiler driver to
> accept -fdump-final-insns and turn that into what
> -fdump-final-insns=$@.gkd would have expanded to (assuming -o $@ or -c
> without -o).
> Much of this patch has been in the VTA branch for a few days already.
> This patch combines that recent patch with some other pieces of
> documentation and build configurations that were still maintained in the
> VTA branch, but that did not have any dependencies on VTA changes, or
> that were changed so as to not have them.
> Are the gcc.c and toplev.c changes ok for the trunk?
> Is the change to the default bootstrap behavior, so that it's faster and
> gets better -g/-g0 coverage, ok?
> Alexandre Oliva, freedom fighter ? ?http://FSFLA.org/~lxoliva/
> You must be the change you wish to see in the world. -- Gandhi
> Be Free! -- http://FSFLA.org/ ? FSF Latin America board member
> Free Software Evangelist ? ? ?Red Hat Brazil Compiler Engineer