[vtk-developers] VTK Modular: Coverage and Valgrind process breakdown

[David Cole]

No problem... I did think "useless" was a little bit strong, though. ;-)


On Tue, May 1, 2012 at 4:53 PM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
> Dave,
>
> I'm sorry. I did not mean the links were generally useless. I meant
> that don't address my concerns about the process breakdown.
>
> Bill
>
> On Tue, May 1, 2012 at 4:47 PM, David Cole <david.cole at kitware.com> wrote:
>> I disagree. It gives you a list of coverage (or valgrind) errors in a
>> list that you can scan to evaluate what's happening over time. Could
>> be useful to somebody who wants to see what day something changed
>> drastically.
>>
>> I agree that coverage and valgrind results are important.
>>
>>
>> On Tue, May 1, 2012 at 4:42 PM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
>>> Dave,
>>>
>>> Those links are pretty much useless. I'm looking forward to the links
>>> we see on the daily dashboards.
>>>
>>> I understand that nothing is free, but, coverage and valgrind are an
>>> important part of our process.
>>>
>>> Bill
>>>
>>> On Tue, May 1, 2012 at 4:28 PM, David Cole <david.cole at kitware.com> wrote:
>>>> Sorry... those links were both the same. Here is the list of coverage
>>>> dashboards:
>>>>
>>>>  http://open.cdash.org/index.php?project=VTK&filtercount=2&showfilters=1&filtercombine=and&field1=hascoverage/bool&compare1=1&value1=&field2=buildstarttime/date&compare2=84&value2=now
>>>>
>>>>
>>>> On Tue, May 1, 2012 at 4:17 PM, David Cole <david.cole at kitware.com> wrote:
>>>>> The AIX machine is not the coverage dashboard. It used to be
>>>>> londinium. Unfortunately, there's a not-yet-fixed issue with gcov from
>>>>> gcc 4.7 not working well with ctest. We have to get the fix into ctest
>>>>> first to run it on a machine with gcc 4.7 -- if somebody else could
>>>>> volunteer a machine using an earlier gcc to report VTK coverage, that
>>>>> would be a good thing.
>>>>>
>>>>> Here's a filtered view of coverage dashboards:
>>>>>
>>>>> http://open.cdash.org/index.php?project=VTK&filtercount=2&showfilters=1&filtercombine=and&field1=hasdynamicanalysis/bool&compare1=1&value1=&field2=buildstarttime/date&compare2=84&value2=now
>>>>>
>>>>> And of dynamic analysis dashboards:
>>>>>
>>>>> http://open.cdash.org/index.php?project=VTK&filtercount=2&showfilters=1&filtercombine=and&field1=hasdynamicanalysis/bool&compare1=1&value1=&field2=buildstarttime/date&compare2=84&value2=now
>>>>>
>>>>> You can scroll through those to see what changed from day to day,
>>>>> going as far back as we have CDash data.
>>>>>
>>>>>
>>>>> HTH,
>>>>> David C.
>>>>>
>>>>>
>>>>> On Tue, May 1, 2012 at 3:15 PM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
>>>>>> Folks,
>>>>>>
>>>>>> Now that VTK modular is settling in, two important software process
>>>>>> components need attention.
>>>>>>
>>>>>> 1) Coverage:
>>>>>> p90n03.pbm.ihost.com - AIX00F614-xlC is the machine that produces
>>>>>> coverage. It has not been building for 22 days.
>>>>>>
>>>>>> 2) Valgrind has over 12000 defects. Prior to modularization it had 0 defects.
>>>>>> http://open.cdash.org/viewDynamicAnalysis.php?buildid=2238424
>>>>>> Seems that many of the defects are python related and need some
>>>>>> valgrind suppressions. Once that is done, we can fix the real defects.
>>>>>>
>>>>>> I'm willing to help improve coverage and real valgrind defects, but
>>>>>> dashboard maintainers need to address the compilation failures and
>>>>>> valgrdin suppressions.
>>>>>>
>>>>>> Bill
>>>>>> --
>>>>>> Unpaid intern in BillsBasement at noware dot com
>>>>>> _______________________________________________
>>>>>> Powered by www.kitware.com
>>>>>>
>>>>>> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
>>>>>>
>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>> http://www.vtk.org/mailman/listinfo/vtk-developers
>>>>>>
>>>
>>>
>>>
>>> --
>>> Unpaid intern in BillsBasement at noware dot com
>
>
>
> --
> Unpaid intern in BillsBasement at noware dot com