[llvm] r184239 - We want a string representation of the attribute, not the kind as a string.

[Bill Wendling]

On Jun 18, 2013, at 5:10 PM, David Blaikie <dblaikie at gmail.com> wrote:

> On Tue, Jun 18, 2013 at 4:46 PM, Bill Wendling <isanbard at gmail.com> wrote:
>> On Jun 18, 2013, at 3:43 PM, David Blaikie <dblaikie at gmail.com> wrote:
>> 
>>> On Tue, Jun 18, 2013 at 2:52 PM, Bill Wendling <isanbard at gmail.com> wrote:
>>>> On Jun 18, 2013, at 2:39 PM, Bill Wendling <isanbard at gmail.com> wrote:
>>>> 
>>>>> On Jun 18, 2013, at 2:35 PM, David Blaikie <dblaikie at gmail.com> wrote:
>>>>> 
>>>>>> On Tue, Jun 18, 2013 at 2:27 PM, Bill Wendling <isanbard at gmail.com> wrote:
>>>>>>> Author: void
>>>>>>> Date: Tue Jun 18 16:27:00 2013
>>>>>>> New Revision: 184239
>>>>>>> 
>>>>>>> URL: http://llvm.org/viewvc/llvm-project?rev=184239&view=rev
>>>>>>> Log:
>>>>>>> We want a string representation of the attribute, not the kind as a string.
>>>>>> 
>>>>>> Test case? (do we have test cases for the verifier?)
>>>>>> 
>>>>> I'll cobble one together.
>>>>> 
>>>> I don't think I can. This requires that the .bc file be malformed. So the `llvm-as' will reject it before it gets to the code I just fixed...
>>> 
>>> Can we remove this as dead code then? (replace it with an assert)
>> 
>> No. The check is valid for checking bitcode rather than text.
> 
> Be nice to have tests for it, then - even if they're checked in binary
> files. (I'd be OK with them being derived from known good files with a
> comment (in the corresponding 'foo.test' text file that actually has
> the RUN line in it) explaining which bit was twiddled)
> 
> Do we not have any tests for binary-only issues the verifier can diagnose?
> 
No. And the file they sent is rather large...and because of the nature of the bug, I can't reduce it.

I *really* think it's okay that this go in without a test. :-)

-bw