[Insight-developers] 32/64 bits universal builds cause itk test failures.

[Mathieu Coursolle]

Commited.

Thanks.

Mathieu

>Mathieu,
>
>If that fixes the problem, check it in. This class has its own random
>number generator and won't affect any other classes.
>
>Thanks,
>
>Bill
>
>
>On Jan 4, 2008 2:15 PM, Mathieu Coursolle < mcoursolle at rogue-
>research.com> wrote:
>In itkRandomImageSource.txx - GenerateThreadedData(...), the random
seed is an
>unsigned long. The random values can then be different in 32 and 64 bits.
>
>Is there a reason why its unsigned long rather than unsigned int. 
>It seems to fix the problem if it is unsigned int.
>
>Thanks.
>
>Mathieu
>
>>Mathieu,
>>
>>The trouble with the failing test is due to the part of the input image 
>>that is random. The test passesbut the regression comparison fails. If
>>the same regression error appears from day-to-day we can check in a
>>second regression image. Don't bother spending time on this one. I'll 
>>monitor the regression error for a few days. If it remains the same,
>>I'll let you know how to check in a second regression image.
>>
>>Bill
>>On Jan 4, 2008 9:19 AM, Mathieu Coursolle < mcoursolle at rogue-
>>research.com> wrote:
>>Bill,
>>
>>I commited a fix too. The size of long is now set explicitly in the
configure
>>file for APPLE, which is needed for universal builds. 
>>
>>In tif_fax3.c, the SIZEOF_LONG is use to define the FILL macro. The same
>>was used
>>in both 32 and 64 bits.
>>
>>Thanks for your fix.
>>
>>There is now one last failing test on our dashboards: 
>>
>>< http://www.itk.org/Testing/Sites/RogueResearch4/Mac10.5-InsightBS-dbg/
>>20080104-0100-Nightly/Results/__Testi 
>>ng_Code_BasicFilters_itkNarrowBandImageFilterBaseTest.html>
>>
>>The output says "Test passed", but the exit code is failed...
>>I'll take a look at it today.
>>
>>Mathieu 
>>
>>>Mathieu,
>>>
>>>Ooops, I just checked in a fix. Can yo0u try it?
>>>
>>>Bill
>>>
>>>On Jan 3, 2008 3:18 PM, Mathieu Coursolle < mcoursolle at rogue-research.com
>>>> wrote:
>>>I am testing a fix for the TIF right now. Same kind has for the Nrrd one.
>>>In the config file, I am setting the size if APPLE is defined.
>>> 
>>>Mathieu
>>>
>>>>I think the tiff is a long problem in TIFFImageIO. I found similar
>>>>problems in BMP.
>>>>
>>>>
>>>>On Jan 3, 2008 1:53 PM, Sean McBride < sean at rogue-research.com > wrote:
>>>>On 1/3/08 1:21 PM, Bill Lorensen said:
>>>>
>>>>>As you probably noticed, I fixed the nrrd configure problem. It's a 
>>>>>combination of the things we talked about.
>>>>
>>>>Yes, thank you!
>>>>
>>>>>I have not closed the bug:
>>>>> http://public.kitware.com/Bug/view.php?id=5904yet until you guys agree
>>>>>that it's working.
>>>>
>>>>Mathieu is giving it a look...
>>>>
>>>>>Oh, I also found a bug or 2 in itkBMPImageIO.cxx that was also causing a
>>>>>failure on the ppc.
>>>>
>>>>Sweet. Any thoughts on why the TIFF tests are failing? It could
perhaps be:
>>>>< http://public.kitware.com/Bug/view.php?id=5469 >
>>>>
>>>>For the other failing test, I have increased the timeout once again.
>>>>
>>>>--
>>>>____________________________________________________________ 
>>>>Sean McBride, B. Eng sean at rogue-research.com
>>>>Rogue Research www.rogue-research.com
>>>>Mac Software Developer Montréal, Québec, Canada
>>>>
>>>
>>
>