This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH: Fix for ObjC regressions
- To: Stan Shebs <shebs at apple dot com>
- Subject: Re: PATCH: Fix for ObjC regressions
- From: Ziemowit Laski <zlaski at apple dot com>
- Date: Fri, 12 Oct 2001 13:18:11 -0700
- Cc: gcc-patches at gcc dot gnu dot org, Nicola Pero <nicola at brainstorm dot co dot uk>
On Friday, October 12, 2001, at 10:06 , Stan Shebs wrote:
> Ziemowit Laski wrote:
>>
>> Many thanks to Richard Henderson for finding the trouble spot.
>>
>> Also tried (and failed) to fix the objc.dg/method-2.m failure.
>> Better luck next time.
>
> What's this all about?
When compiliing for the GNU runtime, the compiler does not give
a warning (in one case) where it should. I'm assuming that even
though the runtimes differ somewhat, the semantics of the language
are the same (please say yes!).
I've recently added code to the objc-act.c:receiver_is_class_object
routine to detect the usage of 'self' in the context of a class
method, but only did it for the NeXT case. I tried tinkering with
the GNU case last night before dropping dead from exhaustion. :)
The idea is this: If we're attempting to invoke a class method that
the compiler cannot find, the compiler should look for an instance
method with the same name in the root class. In objc.dg/method-2.m
there are 4 message sends; in two of them, the compiler should
successfully locate the instance method in the root class, and in
the other 2 it should complain. Nicola, since I see you've done
some work on the GNU side of things, perhaps you could take a look
at this when you have time? :) :)
--Zem
--------------------------------------------------------------
Ziemowit Laski Apple Computer, Inc.
zlaski@apple.com 2 Infinite Loop, MS 302-4SN
+1.408.974.6229 Fax .1344 Cupertino, CA 95014-2085