[llvm-testresults] buildbot failure in llvm on llvm-gcc-x86_64-darwin10-selfhost

[Daniel Dunbar]

On Wed, Nov 4, 2009 at 4:35 PM, Evan Cheng <evan.cheng at apple.com> wrote:
> You mean a debug llvm-gcc is taking a long time to compile release llvm?
> Lang, how about if you disable the assertions now? Then let's try again.

Something like that, the llvm-gcc itself isn't actually built with
--enable-checking (I'm not a gcc build guru, but I think this would
give me a "debug" gcc?) but the LLVM its using is debug.

 - Daniel

> Thanks,
>
> Evan
>
> On Nov 4, 2009, at 3:55 PM, Daniel Dunbar wrote:
>
>> On Wed, Nov 4, 2009 at 12:25 PM, Evan Cheng <evan.cheng at apple.com> wrote:
>>>
>>> Yes.
>>
>> And how long did the build take? This is failing very predictably on
>> my home machine (the buildslave), so its obviously a real problem.
>>
>> The steps it takes are:
>> build llvm (debug)
>> build llvm-gcc
>> build llvm (release, using prev llvm-gcc)
>> ... and fails in there, with a >20min compile
>>
>> all on x86_64, the i386 build is working.
>>
>> - Daniel
>>
>>> Evan
>>> On Nov 4, 2009, at 11:33 AM, Daniel Dunbar wrote:
>>>
>>>> On Wed, Nov 4, 2009 at 1:24 AM, Evan Cheng <evan.cheng at apple.com> wrote:
>>>>>
>>>>> Hrm. I just built a i386 / x86_64 fat llvm-gcc and then self-hosted
>>>>> llvm
>>>>> in
>>>>> 64-bit mode. Not problem here.
>>>>
>>>> Did you build a Release version in the second build of llvm?
>>>>
>>>>>
>>>>> Evan
>>>>>
>>>>> On Nov 4, 2009, at 1:19 AM, Daniel Dunbar wrote:
>>>>>
>>>>>> This failure is making me nervous -- it looks like llvm-gcc is timing
>>>>>> out building some part of itself, and I'm definitely seeing cc1plus
>>>>>> executables with CPU times over 20mins on the machine. Eek.
>>>>>>
>>>>>> On the other hand I'm also seeing some definite weirdness with gcc
>>>>>> compile times going through the roof on X86ISelDAGToDag. That seems
>>>>>> like it would be a gcc cliff that doesn't effect llvm-gcc, but I don't
>>>>>> understand whats going on yet.
>>>>>>
>>>>>> - Daniel
>>>>>>
>>>>>> On Tue, Nov 3, 2009 at 6:50 PM,  <buildbot at google1.osuosl.org> wrote:
>>>>>>>
>>>>>>> The Buildbot has detected a new failure of
>>>>>>> llvm-gcc-x86_64-darwin10-selfhost on llvm.
>>>>>>> Full details are available at:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>  http://google1.osuosl.org:8011/builders/llvm-gcc-x86_64-darwin10-selfhost/builds/406
>>>>>>>
>>>>>>> Buildbot URL: http://google1.osuosl.org:8011/
>>>>>>>
>>>>>>> Buildslave for this Build: dunbar-darwin10
>>>>>>>
>>>>>>> Build Reason:
>>>>>>> Build Source Stamp: 85986
>>>>>>> Blamelist: bwilson,evancheng,lattner,lhames
>>>>>>>
>>>>>>> BUILD FAILED: failed compile.llvm_1
>>>>>>>
>>>>>>> sincerely,
>>>>>>>  -The Buildbot
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> llvm-testresults mailing list
>>>>>>> llvm-testresults at cs.uiuc.edu
>>>>>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-testresults
>>>>>>>
>>>>>
>>>>>
>>>
>>>
>
>