[LLVMdev] ARM failures

[David Blaikie]

On Tue, Jan 8, 2013 at 5:04 AM, Renato Golin <renato.golin at linaro.org> wrote:
> The following failures are consistent on buildbot (and my local box).
>
> The Clang one I think it's assuming an Intel box, the other two look like
> the FileCheck are not good enough.
>
> http://lab.llvm.org:8011/builders/clang-native-arm-cortex-a9/builds/4305
>
> Clang :: CodeGen/compound-assign-overflow.c
> LLVM :: Transforms/LoopStrengthReduce/post-inc-icmpzero.ll
> LLVM :: Transforms/LoopStrengthReduce/2012-07-18-LimitReassociate.ll

Usually the best way to get traction on such things is to reply to the
commit that caused the regression. Whoever broke things is usually
more invested in making sure the change is solid (& doesn't get
reverted).

[slightly ranting point: Whoever owns this builder should be in a
position of authority/autonomy with the project to be able to maintain
their passing status. That means someone who cares about this bot
should have the rights (both technically & culturally) to revert
patches if it becomes necessary (an unco-operative committer). That's
not to say that patches should be reverted without consideration or
taking steps to help the author reproduce the issue. In the case of a
test that's not platform agnostic & needs a triple, that should be
easy/obvious & once notified the committer should be able to make the
change quickly (or the bot maintainer can make such a commit (simply
adding the triple that was assumed, or even generalizing the test to
be neutral if that looks viable) & leave it to the original committer
to choose an alternative fix when they have time.]

- David