This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: neat bug in alias analysis
- To: mark at codesourcery dot com
- Subject: Re: neat bug in alias analysis
- From: Mike Stump <mrs at wrs dot com>
- Date: Wed, 18 Aug 1999 14:46:22 -0700 (PDT)
- Cc: egcs-bugs at egcs dot cygnus dot com
I am sorry Mark, but I don't think that will work, without `solving'
the problem anyway.
The problem is that one can have a pointer to a constant X, and a
pointer to a non-constant X, and these two types _must_ be different,
and yet the _must_ share the same alias set.
So, when that is solved, I don't see much advantage in placing them on
the permanent obstack (I mean, other than the other advantages you
listed).
I will see what I can come up with.