[llvm-commits] [llvm] r172534 - /llvm/trunk/test/Transforms/LoopStrengthReduce/post-inc-icmpzero.ll

[Andrew Trick]

On Jan 15, 2013, at 9:49 AM, Dmitri Gribenko <gribozavr at gmail.com> wrote:

> On Tue, Jan 15, 2013 at 7:44 PM, Andrew Trick <atrick at apple.com> wrote:
>> On Jan 15, 2013, at 9:21 AM, Dmitri Gribenko <gribozavr at gmail.com> wrote:
>> 
>>> On Tue, Jan 15, 2013 at 6:52 PM, Renato Golin <rengolin at systemcall.org> wrote:
>>>> The output is not identical because the command doesn't specify the
>>>> triple, but the relationship between the constructs is still the same.
>>> 
>>> I might be misunderstanding something, but if the triple is not
>>> specified, host system type should have zero influence on how 'opt'
>>> works.
>> 
>> The test case specifies a triple. I agree it would be nice to know how this test was affected by the host. I'd also like to know if this broke after some recent checkin. This may just be a symptom of a real issue with cross-compilation.
> 
> This failed somewhere in between:
> http://lab.llvm.org:8011/builders/clang-native-arm-cortex-a9/builds/4294
> http://lab.llvm.org:8011/builders/clang-native-arm-cortex-a9/builds/4299
> 
> That translates to r171697 .. r171735.


There were some issues with LSR @r171735. AFAIK, they've all been fixed. If anyone sees behavior on latest trunk that differs from r171697, I'd like to know about that!

Thanks!
-Andy