[patch] Don't let invalid methods mark their RecordDecl invalid

[Nico Weber]

r197848, thanks!


On Fri, Dec 20, 2013 at 4:53 PM, Reid Kleckner <rnk at google.com> wrote:

> On Fri, Dec 20, 2013 at 4:27 PM, Nico Weber <thakis at chromium.org> wrote:
>
>> No, the patch changes things to not mark a class invalid if one of its
>> methods is invalid. This was the only place I could find where a class was
>> marked invalid after its definition is complete. (Hence with this patch,
>> C's superclass A will be valid too, only its method stays invalid, like
>> with the first snippet).
>>
>
> OK, this makes sense, we already have a complete class, and so far as we
> know this is the only place we would mark a record decl invalid after
> completing it.  Go for it.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20131220/1950fb38/attachment.html>