This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] fix Objective-C, further improve error recovery for unknown type names

On 14 Nov 2010, at 11:08, Nicola Pero wrote:

indeed, we should not be doing that....
... clang is the de-facto 'standard'  I guess ...

Yes, but we don't really want to be bug compatible with them, if we
think they got it wrong.  One can always file a bug report against
clang, if they want to know for sure.

In this instance, it seems to me that clang is doing it right ...

I agree that clang is doing it right in this instance [and our compiler
needs more work ;-)]; I think Mike was not commenting on this particular case,
but just giving general guidelines on how to approach differences.

I read it that way - but also am interested in his view on this specific case, in case we missed something...

And I agree with his guidelines; we obviously want to be very compatible,
but if they get something wrong (not in this case) we don't necessarily need
to be faithfully reproducing obvious bugs. :-)

completely agreed ... (although Darwin might end up needing a '- mcompat=apple' :/ )

Where does this leave us with the 'fix' to the current FE?

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]