Hi Truls

This might sound really odd, but have you tried running it more than
once? I've been experiencing truly weird problems with non-deterministic
segmentation faults at that point in the code i.e. sometimes it crashes,
sometimes it doesn't. I have the feeling there is a race condition or
unitialised memory problem floating about, but I haven't been able to
track down the cause.

Cheers
Chris

On Mon, 25 Oct 2010 23:26 +0200, "Truls Becken" <[email protected]>
wrote:
> On Mon, Oct 25, 2010 at 22:16, Christopher Armstrong wrote:
> 
> > I've experienced this one too on Linux. I was mixing clang and GCC.
> > Having used SVN for everything, recompiled everything with clang,
> > I don't experience it right now.
> 
> Compiling everything with Clang does not help for me. All GCC, all
> Clang, or a mixture gives the same segfault from any tool.
> 
> Looking at the backtrace and the code in NSObject.m, the problem seems
> to be that static autorelease_imp does not get initialized before
> first use.
> 
> According to the backtrace, the call to +[NSZombie class] on line 958
> ends up in -[NSObject autorelease], which calls autorelease_imp. The
> static is initialized in +[NSObject initialize], but after line 958.
> 
> -Truls
> 
> _______________________________________________
> Etoile-discuss mailing list
> [email protected]
> https://mail.gna.org/listinfo/etoile-discuss
> 
-- 
  Christopher Armstrong
  carmstrong ^^AT^ fastmail dOT com /Dot/ au


_______________________________________________
Etoile-discuss mailing list
[email protected]
https://mail.gna.org/listinfo/etoile-discuss

Répondre à