C identifier lookup speedups, 2/2

[Mark Mitchell]

On Mon, 2003-04-14 at 09:57, David Edelsohn wrote:
> >>>>> Mark Mitchell writes:
> 
> Mark> I am looking into this on Zack's behalf.
> 
> 	If we want to agree that this is an expected failure on all
> platforms, that's fine.  I assume it previously was expected to succeed
> for a reason.

It's an extension accepted by a lot of other compilers.  

With Zack's changes, we're silently accepting the extension, but giving
a different interpretation than we used to.  Here's a patch which makes
us pedwarn (since it's not valid ISO C), and then do what we used to do.

I think.

David, I've confirmed that this changes the .s file on AIX in a way that
I'd expect it to change; would you confirm that it fixes the test case?

One you do that, I'll check it in.

Thanks,

-- 
Mark Mitchell
CodeSourcery, LLC
mark@codesourcery.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: diffs
Type: text/x-patch
Size: 2118 bytes
Desc: 
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20030414/fb530112/attachment.bin>