This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Something else for 1.1.1
- To: mark at markmitchell dot com
- Subject: Re: Something else for 1.1.1
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Mon, 23 Nov 1998 09:43:34 -0700
- cc: root at ihack dot net, egcs-patches at cygnus dot com
- Reply-To: law at cygnus dot com
In message <199811191537.HAA00796@adsl-206-170-148-33.dsl.pacbell.net>you wri
te:
> >>>>> "Jeffrey" == Jeffrey A Law <law@cygnus.com> writes:
>
> Jeffrey> In message
> Jeffrey> <199811150216.VAA23361@zygorthian-space-raiders.mit.edu>you
> Jeffrey> write:
> >> Could this change:
> >>
> >> 1.43 Wed Sep 30 10:09:59 1998 by mmitchel Diffs to 1.42 *
> >> function.c (gen_mem_addressof): If the address REG is
> >> REG_USERVAR_P make the new REG be so also.
> >>
> >> go into 1.1.1? It fixes obscure optimization problems with
> >> `&argument' references in some functions. (This *usually*
> >> loses with loop optimization, which was rather klugily `fixed'
> >> by removing a REG_USERVAR_P() check in loop.c, but I'd rather
> >> have a correct fix in the release.)
>
> Jeffrey> The REG_USERVAR_P check is the right fix. The
> Jeffrey> assumptions made by that REG_USERVAR_P were simply wrong.
> Jeffrey> It was a major bogon.
>
> I think by this you meant "the loop change is correct". I believe
> it's also true that my change (mentioned above) is correct; whether or
> not it should go in 1.1.1 is another question, of course.
Yes. That is precisely what I meant.
jeff