[PATCH] Fix darwin/25908, key functions vs non weak vtables

[Mike Stump]

On Jan 26, 2006, at 7:32 AM, Andrew Pinski wrote:
> The problem here is that Darwin has non weak vtables which means that
> the key function for the following source can no longer be the  
> inline function:

> OK?

Ok.

Ok for all release branches that this is applicable to, I don't think  
it made its way onto any (yet), but just in case it has or does.

I was hoping that Geoff would comment on this, as it looks like he  
was playing in this area and because he has a couple of radars that  
relate to this (4263752).  Geoff, the short story is that this fixes  
gcc to be more like it was in our 3.3.

Anyway, self-consistency is a virtue, so I'll approve it.