[PATCH] PR c++/10031: Exception lists on builtins

Roger Sayle roger@www.eyesopen.com
Fri Mar 14 05:09:00 GMT 2003


On Thu, 13 Mar 2003, Jason Merrill wrote:
> On Thu, 13 Mar 2003, Roger Sayle <roger@www.eyesopen.com> wrote:
> >> Maybe, but let's hold off on this for a bit.
> >
> > You do of course appreciate that both mainline and the 3.3 branch
> > fail to bootstrap on both powerpc-linux and x86_64-linux due to this?
>
> I wasn't aware of that, no.  Go ahead and put in your patch for the moment.

Oopps!  My mistake.  Andreas' calls to revert the original patch within
48 hours had me thinking this was an urgent problem.  Although it is
serious and must be resolved before 3.3, we probably have a few days
to consider alternate solutions.

I'll take your advice and hold off on this a bit.

I'll also look into a fix along the lines you suggest, i.e. using
the newdecl's type when redeclaring an anticipated built-in for the
first time.  I've also got some other patches to simplify the logic
in duplicate_decls.

Sorry again for the false alarm.  I get confused which of the
bugs I'm investigating, and which patches I have await review,
are for bootstrap failures (v850, ia64-hpux) and which are just
testsuite regressions.

Roger
--



More information about the Gcc-patches mailing list