This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix 10219
- From: Mark Mitchell <mark at codesourcery dot com>
- To: Richard Kenner <kenner at vlsi1 dot ultra dot nyu dot edu>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: 02 Jul 2003 10:52:19 -0700
- Subject: Re: [PATCH] Fix 10219
- References: <10307021745.AA02145@vlsi1.ultra.nyu.edu>
On Wed, 2003-07-02 at 10:45, Richard Kenner wrote:
> I think the convention is to always use "== error_mark_node", just
> because there ought never be two nodes with code ERROR_MARK.
>
> I argee we ought to use the same convention everywhere, but testing for
> ERROR_MARK is at least one less memory access on most machines, so there
> is a *tiny* advantage of using it.
We've had this discussion at least twice in the past, and there's never
been a clear outcome. Some machines work one way, some another. PIC
and non-PIC code generation makes a difference.
Please, let's not have this debate again. It's not worth it.
--
Mark Mitchell
CodeSourcery, LLC
mark@codesourcery.com