[Insight-developers] MetaImage Test failing on Mac

[Bradley Lowekamp]

Hello,
	
	I remember this test now, it did not work right when I first  
encountered it.

1) The itkMetaImageStreamingWriterIOTest2:

# This test is failing to produce correct output
ADD_TEST(itkMetaImageStreamingWriterIOTest2 ${IO_TESTS}
#  --compare ${ITK_DATA_ROOT}/Input/mri3D.mhd
#            ${ITK_TEST_OUTPUT_DIR}/mri3DWriteStreamed.mha
   itkMetaImageStreamingWriterIOTest
             ${ITK_DATA_ROOT}/Input/mri3D.mhd
             ${ITK_TEST_OUTPUT_DIR}/mri3DWriteStreamed.mha)

The input mha file references a dcm file. And meta IO does not support  
streaming from that format well. Before I made any changes to the  
streaming IO, this test was passing, but not producing the correct  
output file on all systems.

2) The cause of the segfault is that MetaImage::ElementByteOrderSwap  
assumes the size of the buffer is the size of the  image, but in the  
case of streaming this is not true. This problem could have been  
encountered before I made the recent changes to ImageFileReader. I see  
two possible solutions, A) add a MetaImage function to take the buffer  
size as a parameter, or B) have itkMetaImageIO not call  
MetaImage::ElementByteOrderFix, but instead call an itk::ByteSwapper.  
I think option B would be easiest.

Brad



On Jul 6, 2009, at 9:02 AM, Lowekamp, Bradley (NIH/NLM/LHC) [C] wrote:

> I am pretty sure that this is my fault. I am going to look into it  
> this morning.
>
> Brad
>
> On Jul 6, 2009, at 8:22 AM, Luis Ibanez wrote:
>
>> Brad, Julien,
>>
>> The following test
>>
>> itkMetaImageStreamingWriterIOTest2
>>
>> is failing (seg.faulting) on Macs.
>>
>> For example:
>> http://www.cdash.org/CDash/testDetails.php?test=26512580&build=371914
>>
>>
>> May this be related to the recent changes to the
>> IO classes ?
>>
>> or to the recent changes to the MetaImage library ?
>>
>>
>>   Thanks for any hints,
>>
>>
>>         Luis
>>
>>
>
> ========================================================
> Bradley Lowekamp
> Lockheed Martin Contractor for
> Office of High Performance Computing and Communications
> National Library of Medicine
> blowekamp at mail.nih.gov
>
>
> <ATT00001.txt>

========================================================
Bradley Lowekamp
Lockheed Martin Contractor for
Office of High Performance Computing and Communications
National Library of Medicine
blowekamp at mail.nih.gov


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/mailman/private/insight-developers/attachments/20090706/23cdb886/attachment.htm>