[PATCH] PR6677 and explicit specializations of "key functions"

[Rafael Espíndola]

> Especially troubling is that if you remove the explicit instantiation
> declaration, we no longer emit the vtable, so if this *is* necessary for
> some reason, we're not doing it right.

Yes, that has to be a bug.

> So... I think this was originally incorrect, and still works only due to a
> bug. The attached patch fixes the glitch, but I wanted to make sure I'm not
> missing some subtlety here before going ahead with this.

Looks like the test was first added in r99226.  The relevant email
thread seems to be

http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20100322/028636.html

Reading it again it seems that the real bug was the vtable not being
produced on the use,  and me being confused about specializations
having key functions.

LGTM. Thanks for catching this.

Cheers,
Rafael