Xulrunner in Fedora 8

[Jakub Jelinek]

On Wed, Sep 19, 2007 at 06:14:13PM -0400, Tom spot Callaway wrote:
> 
> On Wed, 2007-09-19 at 22:15 +0100, David Woodhouse wrote:
> > On Wed, 2007-09-19 at 16:23 -0400, Tom "spot" Callaway wrote:
> > > This looks like a bug in -Wp,-D_FORTIFY_SOURCE=2 for C++ code. If I
> > > remove that option, it gets past that point.
> > 
> > Anyway, as with the majority of failures I encounter or investigate on
> > PowerPC or even PPC64, it's not an arch-specific issue.
> 
> It also is triggered on x86_64, and would have been caught there first.

It seems to be some function type sharing bug, minimal testcase:
extern int foo (char *) __attribute__ ((warn_unused_result));
extern int bar (char *) throw () __attribute__ ((warn_unused_result));
extern int bar (char *) throw ();

(or nonnull attribute, regparm attr or some other that affects the
type rather than the decl).
I have reproduced this bug with GCC 4.1.x-RH, 4.2.2 and GCC trunk.

Debugging...

	Jakub