Date: Sat, 17 Nov 2001 10:05:44 +0200 (WET)
From: Andris Pavenis <pavenis AT lanet DOT lv>
X-Sender: pavenis AT ieva06
To: Andrew Cottrell <acottrel AT ihug DOT com DOT au>
Cc: djgpp-workers AT delorie DOT com
Subject: Re: Test archives of gcc-3.0.3 20011115 (prerelease) and gcc-3.1
 20011113 (experimantal)
In-Reply-To: <004d01c16f13$940e4730$0a02a8c0@acceleron>
Message-ID: <Pine.A41.4.05.10111171000560.69320-100000@ieva06>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk



On Sat, 17 Nov 2001, Andrew Cottrell wrote:

> > Changes in gcc-3.0.3 20011115 (prerelease) against archives of gcc-3.0.2
> > in v2gnu:
> >
> > - fix for Allegro compilation problem. Now newer beta versions
> >   of Allegro can be built (thanks to Brad Kaiser for fix)
> 
> Is this the one where Allegro exe's crash if not built with debug turned on?
> If it is not then does anyone have a link I can check out what the problem
> is to see if it will affect a program I have that uses Allegro.
> 

Yes. gcc-3.X without Brad's patch (really only one line
changed in gcc/reload1.c) miscompiles src/unicode.c in Allegro, when
compiled using -fomit-frame-pointer. The same bug (AFAIK) causes ICE when
compiling 8139too.c in Linux kernel (which is one of most often
reported trouble with gcc-3.0.X)

At least I built Allegro-3.9.40 with my latest binaries of gcc-3.0.3
20011115 without problems (no more -fomit-frame-pointer related crashes)

Andris