You should be able to take the union of flags?

Alternatively, you could add these test cases to the files currently
testing those warning flags, which might be cleaner.

I actually wonder if this test shouldn't just focus on a few "interesting"
tests of complex type diffing, and you should add the basic test that a
particular diagnostic employs type diffing to the existing files in the
test suite that exercise that particular diagnostic.


On Thu, Jul 12, 2012 at 2:22 PM, Richard Trieu <[email protected]> wrote:

> On Thu, Jul 12, 2012 at 2:16 PM, Chandler Carruth <[email protected]>wrote:
>
>> Totally psyched about this going in...
>>
>> Meta comment:
>>
>> Can you fold all of the tests into a single file?
>>
>> Our test running infrastructure scales better with one larger file than
>> with lots of smaller ones.
>>
>
> There's possibly some flag mismatch from going to a single file, but I can
> definitely shrink the number of test files.
>
>>
>>
>> On Thu, Jul 12, 2012 at 2:07 PM, Richard Trieu <[email protected]> wrote:
>>
>>> Change diagnostic messages in DiagnosticsSemaKinds.td to use %diff when
>>> two QualType's are required.  This will enable template type diffing for
>>> all Sema messages.  The default text of the diff matches the old message so
>>> no test cases need to be updated.  Some tests have been added for the
>>> modified messages.
>>>
>>> Patch attached and also available at:
>>> http://llvm-reviews.chandlerc.com/D6
>>>
>>> _______________________________________________
>>> cfe-commits mailing list
>>> [email protected]
>>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>>>
>>>
>>
>
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to