Re: [3_0_BRANCH] Backport patch to fix KDE2 blocking bug PR 3387

[Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>]

On Wednesday 12 September 2001 00:10, Jakub Jelinek wrote:
> On Tue, Sep 11, 2001 at 08:47:39PM +0200, Franz Sirl wrote:
> > in an effort to get the "don't use gcc3 with KDE2" label removed I
> > decided to sleep with the devil and debug x86 stuff :-). After a short
> > fight with my rusty x86 assembly knowledge I was successful in tracking
> > down the patch that fixes this PR on the mainline.
> >
> > I would like to commit the appended backported patch (and the original
> > testcase execute/20010403-1.c) to the branch, it fixes the posted
> > testcase and also works in the original context producing a working
> > libxml2 on x86.
>
> Does execute/20010403-1.c fail for you without the GCSE/CSE patch?
> It did crash for me (e.g. at -O2) with gcc-2.96-RH without that patch, but
> current 3.0 branch does not crash. Maybe it would be good to have some
> testcase distilled from libxml2 where it crashes on 3.0 branch without the
> patch...

Try the testcase in PR3387, it got miscompiled on x86 with the branch and is 
fixed with your patch. Should be easy to turn it into an executable testcase 
for gcc.dg.

> (A few days ago I went through all 350 gcc-2.96-RH patches checking whether
> all are either commited, obsoleted etc. and this one was put in
> questionable patches because the testcase did not fail).

350 already? Amazing. Good work.

> > Jakub, do you by chance remember if this patch relies on anything that
> > may not be on the branch yet? Maybe this one:
> >
> > Wed Apr  4 00:45:38 EDT 2001  John Wehle  (john@feith.com)
>
> Nope. I don't remember it relying on anything else (and particularly
> gcc-2.96-RH uses my GCSE/CSE patch but not this one).

Fine, just wanted to make that sure, cause the patch didn't apply 1:1 on the 
branch because of John Wehle's patch missing there.

> Searching gcc-patches
> from April reveals Stan complaining about some Darwin bootstrap failures
> related to this, so it would be good to hear whether Darwin bootstraps the
> trunk now (and better what was the fix which made it bootstrap).

Seems your patch just triggered a latent darwin-specific code generation 
problem:
http://gcc.gnu.org/ml/gcc-patches/2001-05/msg01004.html
http://gcc.gnu.org/ml/gcc-patches/2001-05/msg01319.html

Stan? Can you comment? And maybe close PR 2639 if you can bootstrap on darwin 
now :-)
I checked my build logs and I saw no problems in that timeframe on ppc-linux.

Franz.