[cfe-commits] r165256 - /cfe/trunk/lib/Sema/SemaDecl.cpp

[Philip Craig]

On Fri, Oct 5, 2012 at 2:40 PM, Abramo Bagnara
<abramo.bagnara at bugseng.com> wrote:
> Il 05/10/2012 00:15, Chandler Carruth ha scritto:
>> On Thu, Oct 4, 2012 at 3:12 PM, Chandler Carruth <chandlerc at google.com
>> <mailto:chandlerc at google.com>> wrote:
>>
>>     Can we get a test case for this?
>>
>>
>> Also, not just for this, but for all of the source location fixes you're
>> making?
>
> This problem has been discussed a few time in past, but after discussion
> on the list we failed to find a nice (and easy) solution for testing.
>
> If things are now changed with tooling infrastructure, this would be
> nice to hear. If someone is so kind to write a test for this patch we
> will be happy to add the others replicating the same mechanism shown.
>
> The testcase if very simple "void f(i) { }", what it should be tested is
> that source range for ParamDecl matches the 'i'.

Here's a patch for the test. I don't know whether you want to keep
these tests from the other RAV tests or not.

>> We've got *lots* of source location bugs (as you're finding!) and we
>> really need to start testing them in addition to fixing them.
>
> FWIW we test massively also for this kind of regressions in our ECLAIR
> huge internal testsuite.
>
> --
> Abramo Bagnara
>
> BUGSENG srl - http://bugseng.com
> mailto:abramo.bagnara at bugseng.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: KNRParamLoc.patch
Type: application/octet-stream
Size: 556 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20121006/4132bbc4/attachment.obj>