[kaffe] build failure if old version present, ServletContext.setAttribute

[Kevin D. Kissell]

> On Thu, 29 May 2003 13:42:07 +0200
> "Kevin D. Kissell" <kevink at mips.com> wrote:
> 
> > > > > From the traces you've sent in this thread, I would assume
> > > > > that this assertion failure only occurrs when you get a
> > > > > NullPointerException, is that correct?
> > > > 
> > > > The past two times, that seemed to be the case.  However, it's
> > > > different today, as shown below.
> > > 
> > > There are actually two different assertions that fail:
> > > 
> > > * The !INTS_DISABLED one somewhere in exception.c which seems to
> > >   fail only after a NullPointerException. If that is correct, I
> > >   would suspect that kaffe's signal handling doesn't work 100%
> > >   correct on OpenBSD
> > 
> > One sees (or used to see) the same failure mode under Linux if the stack
> > frame or (more often) the frame pointer register is corrupted.
> 
> With both, mips and i386?

While the corruption problem that I was looking at was specific
to MIPS, there is nothing architecture-dependent about that assert
failing once things are corrupted.  I don't know about i386, but I 
remember reading on this list that others have seen it on non-MIPS 
platforms over the past year.