From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jason Merrill <jason_merrill@redhat.com>
To: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Cc: gcc-patches@gcc.gnu.org
Subject: Re: spec2000 regression
Date: Thu, 16 Aug 2001 06:09:00 -0000
Message-id: <m31ymcdudp.fsf@prospero.cambridge.redhat.com>
References: <10108151621.AA18350@vlsi1.ultra.nyu.edu>
X-SW-Source: 2001-08/msg00969.html

>>>>> "Richard" == Richard Kenner <kenner@vlsi1.ultra.nyu.edu> writes:

>     !   if (DECL_P (t) && ref == DECL_RTL_IF_SET (t))
>     !     /* Avoid returning a bogus alias set from a new MEM.  */
>     !     set_mem_alias_set (ref, get_alias_set (type));
>     !   else
>     !     set_mem_alias_set (ref, get_alias_set (t));

> I suppose, but I think this is a real kludge.

> Instead, I'd abort in the first case and fix all the callers.

Makes sense.

Jason