r214722 - Highlight the offending function parameter when the format_arg argument refers to an invalid function parameter type.

[Aaron Ballman]

On Mon, Aug 4, 2014 at 1:21 PM, Aaron Ballman <aaron at aaronballman.com> wrote:
> On Mon, Aug 4, 2014 at 12:56 PM, David Blaikie <dblaikie at gmail.com> wrote:
>> On Mon, Aug 4, 2014 at 9:36 AM, Aaron Ballman <aaron at aaronballman.com> wrote:
>>> On Mon, Aug 4, 2014 at 12:34 PM, David Blaikie <dblaikie at gmail.com> wrote:
>>>> Test case? (I know we don't rigorously test the ranges specified in a
>>>> diagnostic, but if we're fixing a bug in it, it might be worthwhile -
>>>> I think they're sometimes tested with the machine readable diagnostic
>>>> format)
>>>
>>> I asked about this last week, but there was no conclusion on whether
>>> this sort of change needs tests or how to effectively test it. Can you
>>> point me to a testcase to model after?
>>
>> Looks like we have no existing test cases for this, nor even any test
>> cases for the feature (that already exists) that can be used to test
>> this. "-fdiagnostics-print-source-range-info" should be able to be
>> used to test this (we have lots of test cases that test /fixits/ using
>> a similar feature "-fdiagnostics-parseable-fixits" - for example
>> test/Sema/warn-main.c)
>
> Okay, I'll see if I can whip something up. Thank you for the suggestion!

Finally had the chance to do something about this -- committed in
r215444. Thank you for the suggestions!

~Aaron