[Insight-developers] Dashboard Cleanup : Timeout of Join Series in Dash11.kitware : 1, 722 Streaming WARNING messages

[Luis Ibanez]

Hi Brad,

I'm wondering if warning the user is useful (or needed) at all.

After all, this check for zero size regions doesn't change the
behavior of previous filters, does it ?

It simply prevents filters from running unnecessary redundant
updates, but it doesn't produce a different output.

Isn't that the case ?


      Luis


-----------------------------------------------------------------------------------
On Wed, Jan 20, 2010 at 11:40 AM, Bradley Lowekamp
<blowekamp at mail.nih.gov> wrote:
> Thanks for looking into that. It would appear the warning message there is printed out way too many times, when ITK_LEGACY_REMOVE is not defined. This is not at all important to the test. This message was added to let users know that the quagmire of updating a zero sized region had been defined. I suppose that the best behavior would be just to warn the users once per object instance of this improvement. This would require a member variable though.
>
> Can you think of a better way to warn the users of this situation?
>
> Thanks,
> Brad
>
> On Jan 20, 2010, at 11:18 AM, Luis Ibanez wrote:
>
>> Hi Brad
>>
>> I logged in dash11 to run the test manually and to
>> figure out what a reasonable time out would be for it.
>>
>>
>> In doing so, I notice that this test is printing out a
>> very large amount of text, that probably accounts
>> for a good fraction of its long execution time.
>>
>>
>> To be more specific:
>>
>> The test is printing out about 19,000 lines of text,
>> most of it is repetitions of the message:
>>
>> ------------------------------------------------------------
>> 492: WARNING: In /home/ibanez/src/Insight/Code/Common/itkImageBase.txx, line 288
>> 492: Image (0x10e6750): Not executing UpdateOutputData due to zero
>> pixel condition RequestedRegion:ImageRegion (0x10e6890)
>> 492:   Dimension: 2
>> 492:   Index: [0, 0]
>> 492:   Size: [0, 0]
>> 492:  BufferedRegion: ImageRegion (0x10e68b8)
>> 492:   Dimension: 2
>> 492:   Index: [0, 0]
>> 492:   Size: [0, 0]
>> ---------------------------------------------------------------
>>
>> This same message is printed out 1,722 times     :-(
>>
>> -
>>
>> A)   Is that message significant to the test ?
>>
>> B)   Could we run the test in a smaller image perhaps ?
>>
>>
>> In its current form the test needs 160 seconds
>> to run in this machine.
>>
>> I'm increasing the timeout to 240 seconds,
>> just to get the test to pass.
>>
>> but we probably should reconsider the structure
>> and purpose of this tests.
>>
>>
>>    Any suggestions ?
>>
>>
>>         Thanks
>>
>>
>>             Luis
>>
>>
>>
>> ------------------------------------------
>> On Wed, Jan 20, 2010 at 8:55 AM, Dave Partyka <dave.partyka at kitware.com> wrote:
>>> Resending as I wasn't a member of insight-developers.
>>>
>>> On Wed, Jan 20, 2010 at 8:42 AM, Dave Partyka <dave.partyka at kitware.com>
>>> wrote:
>>>>
>>>> No objection here if it really takes that long to run. Granted that
>>>> machine is pretty old/slow.
>>>>
>>>> On Wed, Jan 20, 2010 at 8:40 AM, Luis Ibanez <luis.ibanez at kitware.com>
>>>> wrote:
>>>>>
>>>>> The test:
>>>>>
>>>>>
>>>>>          itkJoinSeriesImageFilterStreamingTest
>>>>>
>>>>> http://www.cdash.org/CDash/viewTest.php?onlyfailed&buildid=519006
>>>>>
>>>>>
>>>>>            Is timing out in dash11.kitware.
>>>>>
>>>>> The current timeout is set to 120 seconds.
>>>>>
>>>>> Does anybody has an objection to raising the
>>>>> TIMEOUT value of dash11  ?
>>>>>
>>>>>
>>>>>   Thanks
>>>>>
>>>>>
>>>>>        Luis
>>>>
>>>
>>>
>
>