[vtk-developers] volTM2DRotateClip recurring test failures

[David Gobbi]

The very fact that kamino is failing a test that passes on other machines
demonstrates its value as a dashboard machine.

 - David


On Mon, Feb 18, 2013 at 9:28 AM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
> I'm not sure why we are running those machines. Looks like the
> RogueResearch machines cover those compilers and they have tcl and
> python enabled.
>
> On Mon, Feb 18, 2013 at 11:15 AM, David Gobbi <david.gobbi at gmail.com> wrote:
>> The only machine it fails on is kamino, and on that machine it fails
>> for clang and gcc.  Is it possible to enlist the help of Chris Harris,
>> who owns machine?  I'd be interested to see if the Tcl version of the
>> test fails in the same way as the Python one, but kamino doesn't build
>> the Tcl wrappers.
>>
>>  - David
>>
>> On Mon, Feb 18, 2013 at 9:06 AM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
>>> How should we proceed? The test has been failing on those two clang
>>> platforms since at least October. Did it ever pass on those platforms?
>>> I can't help because the test never fails on my Mac with clang.
>>>
>>> On Mon, Feb 18, 2013 at 10:53 AM, David Gobbi <david.gobbi at gmail.com> wrote:
>>>> Bill, the incorrectness of the image is not just a matter of tolerance...
>>>> that machine is rendering the image _wrong_ and we don't know why.
>>>> Ditto for TestSplitViewportHorizontal.  These things have to be fixed,
>>>> not swept under the rug.
>>>>
>>>>  - David
>>>>
>>>>
>>>> On Mon, Feb 18, 2013 at 8:27 AM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
>>>>> David G,
>>>>>
>>>>> Looks like the delay did not help. Why not add an alternate Baseline?
>>>>>
>>>>> Bill
>>>>>
>>>>> On Wed, Feb 13, 2013 at 1:44 PM, David Cole <dlrdave at aol.com> wrote:
>>>>>> Is there a call we can add at screen grab time to make sure we're properly
>>>>>> waiting for all pending/queued output? (Or are we making it already, and
>>>>>> this is really a symptom of a bad graphics driver...?)
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> -----Original Message-----
>>>>>> From: David Gobbi <david.gobbi at gmail.com>
>>>>>> To: Bill Lorensen <bill.lorensen at gmail.com>
>>>>>> Cc: VTK Developers <vtk-developers at vtk.org>
>>>>>> Sent: Wed, Feb 13, 2013 1:41 pm
>>>>>> Subject: Re: [vtk-developers] volTM2DRotateClip recurring test failures
>>>>>>
>>>>>> This test has been failing on certain platforms for at least
>>>>>> five years or so.  The test does a bit of animation, and the
>>>>>> test harness seems to capture the image at the wrong time,
>>>>>> i.e. the captured image is not actually the final image shown
>>>>>> by the test.
>>>>>>
>>>>>> Is there any possibility that the image is being captured
>>>>>> before the final SwapBuffers has taken effect?  I mean,
>>>>>> if the graphics card has to process all vertexes and fragments
>>>>>> and then wait for the next vblank before swapping
>>>>>> the buffers, is it possible that the screen grab might grab
>>>>>> the screen before the swap has actually taken effect?
>>>>>> Especially for a test like this that has a huge number of
>>>>>> textures?
>>>>>>
>>>>>>  - David
>>>>>>
>>>>>> On Wed, Feb 13, 2013 at 10:29 AM, Bill Lorensen <bill.lorensen at gmail.com>
>>>>>> wrote:
>>>>>>> David,
>>>>>>>
>>>>>>> Do you have any insight into the recurring test failures (since at
>>>>>>> least October) for vtkRenderingVolumePython-volTM2DRotateClip.
>>>>>>> They seem to be on some Macs (3 at this time). See for example:
>>>>>>> http://open.cdash.org/testDetails.php?test=177122517&build=2810336
>>>>>>>
>>>>>>> Do we just need an alternate baseline?
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>> Bill
>>>>>> _______________________________________________
>>>>>> 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
>
>
>
> --
> Unpaid intern in BillsBasement at noware dot com