FYI, I re-applied a greatly updated version of this patch in r130628. So
far, it's passed my self-host, and I'm hopeful the problems have been
resolved.

Comments on that revision would of course be appreciated.

On Wed, Jun 23, 2010 at 11:30 PM, Enea Zaffanella <[email protected]>wrote:

> Ping.
>
> Any hint about resolving the problem below?
>
> Thanks in advance,
> Enea.
>
> Abramo Bagnara wrote:
> > Il 16/06/2010 17:23, Douglas Gregor ha scritto:
> >> It looks like this broke self-host with an assertion, so I've reverted
> it in r106100.
> >
> > This is a reduced testcase:
> >
> > template <typename C>
> > struct v {
> >   template <typename T, typename U = T>
> >   struct s { };
> >   s<C> v;
> > };
> >
> > The failing assertion happens when default template argument is
> > substituted due to call of Sema::SubstDefaultTemplateArgument from
> > Sema::CheckTemplateArgumentList.
> >
> > Sema::LocalInstantiationScope::getInstantiationOf is called on a null
> > LocalInstantiationScope.
> >
> > We've tried to add a Sema::LocalInstantiationScope just before the
> > related Sema::InstantiatingTemplate to have a valid
> > LocalInstantiationScope but this is has no effects (apart the obvious
> > one to reach getInstantiationOf with a non null LocalInstantiationScope).
> >
> > I guess that the problem is that Sema::CheckTemplateArgumentList is not
> > called in an appropriately prepared framework, but I'm rather confused
> > now about why and what is needed to make it work properly.
> >
> > Can you shed some light on this failure?
> >
> > BTW I've also noted that (unexpectedly, at least for me) the declaration
> > context of T is the translation unit.
> > _______________________________________________
> > cfe-commits mailing list
> > [email protected]
> > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
> >
> >
>
> _______________________________________________
> cfe-commits mailing list
> [email protected]
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to