On Tue, Jul 26, 2011 at 9:57 PM, John McCall <[email protected]> wrote:
> On Jul 26, 2011, at 6:05 PM, Francois Pichet wrote:
>> Author: fpichet
>> Date: Tue Jul 26 20:05:24 2011
>> New Revision: 136203
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=136203&view=rev
>> Log:
>> In Microsoft mode, if we are within a templated function and we can't 
>> resolve Identifier during BuildCXXNestedNameSpecifier, then extend the SS 
>> with Identifier. This will have  the effect of resolving Identifier during 
>> template instantiation.  The goal is to be able to resolve a function call 
>> whose nested-name-specifier is located inside a dependent base class.
>
> Shouldn't this be gated on DelayedTemplateParsing rather than Microsoft mode?
>

Why should it be? This is really about
http://clang.llvm.org/compatibility.html#dep_lookup_bases.
Maybe it is related but I see it as a somehow different topic.

I am going to have another patch about this topic later for the case
of dependent function calls not prefixed with a nested-name-specifier.

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

Reply via email to