This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: new alias sets from the backend
- To: rth at cygnus dot com
- Subject: Re: new alias sets from the backend
- From: Mark Mitchell <mark at markmitchell dot com>
- Date: Sun, 20 Sep 1998 19:51:57 -0700
- CC: egcs-patches at cygnus dot com
- References: <19980920185800.A9996@dot.cygnus.com>
- Reply-to: mark at markmitchell dot com
>>>>> "Richard" == Richard Henderson <rth@cygnus.com> writes:
Richard> Or does the arithmetic
Richard> invalidate the alias set we might have had before?
That's language dependent. But, in C/C++, it's valid to use the same
alias set after pointer arithmetic; I believe the front ends already
do this.
But, varargs macros do such tricky things that I recently turned off
the use of alias sets completely in varargs functions. There should
be some information in the mail archives around the time of the patch
explaining the situation, I think. If not, Jeff and I should be able
to dig it out of our combined mail files.
Tue Aug 25 12:02:23 1998 Mark Mitchell <mark@markmitchell.com>
* alias.c: Include output.h.
(DIFFERENT_ALIAS_SETS_P): Don't treat alias sets as
different if we're in a varargs function.
* Makefile.in (alias.o): Depend on output.h
--
Mark Mitchell mark@markmitchell.com
Mark Mitchell Consulting http://www.markmitchell.com