On Feb 29, 2008, at 10:09 AM, Nuno Lopes wrote:
>>> The test currently doesn't pass because attributes aren't being
>>> propagated when merging functions/vars definitions.
>>>
>>> e.g.
>>> void g() __attribute__((deprecated));
>>> void g();
>>>
>>> int a() {
>>> //should produce a warning. it does not as clang discards the old
>>> defnition and its attributes when merging
>>> g();
>>> }
>>
>> Ok, for now, I just disabled these cases from the test case. It
>> would be nice to have this but can happen as a separate step.
>
> Ok. But how should clang handle this attribute merging? Simply
> append the old declaration atrributes to the new ones? Append just
> non-repeating ones? Or..?
It depends on the attributes, but yes they should generally be merged.
> BTW, I just saw how noreturn is handled (it enforces 0 parameters).
> Do you prefer having this enforcement done in deprecated as well as
> other attributes?
Yep, please.
>>> http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080225/004488.html
>
> Thanks. BTW, you forgot to commit the test.
Whoops, fixed. Forgot to svn add it. Thanks,
-Chris
_______________________________________________
cfe-dev mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev
