On Oct 26, 2009, at 7:37 AM, Douglas Gregor wrote:
> This won't always work, because the deprecated attribute could be
> merged from a previous function declaration (so it never gets attached
> to this declarator).
>
> There are other tricky cases involving template-ids and 'typename'
> types that this also won't handle. I think John was planning to tackle
> these cases as soon as he finishes the TypeLoc work (which is needed
> to get this completely right).

Aha, ok.  I think it is still useful to keep these changes in until  
that happens (it covers a common case that bites people, and is a GCC  
incompatibility), but I'm happy to see John make this "right".  Thanks!

-Chris
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to