[Insight-developers] ITK SCIFIO

[Matt McCormick]

Awesome, thanks Mark.

I have pushed the fixes up as Patch Set 3 if you want to test them, Kedar:

  http://review.source.kitware.com/#/c/9450/

Thanks,
Matt

On Wed, Apr 17, 2013 at 3:58 PM, Mark Hiner <hiner at wisc.edu> wrote:
> Hi Kedar,
>
>   I was wrong about the dimensions being off. The itkSCIFIOImageIOTest uses
> a streaming reader with a default of 4 divisions, thus the size 512 y is
> correct.
>
>   So it's just that the pixel size was being computed incorrectly. I was
> mapping SCIFIO and ITK pixel/component types incorrectly. Fixed that, and
> now this conversion test works for me locally.
>
>   Matt: Curtis said you handled merging work to the SCIFIO branch before. I
> sent a request for push access to ITK.git but you could also just update
> http://review.source.kitware.com/#/c/9450/ with the code from my
> scifio-fixes branch.
>
>   Anyway, just let me know if you'd like me to do anything to merge my code
> back to the ITK repo.
>
>   I'm considering the SCIFIOImageIO fixed for now. I still want to make both
> the c++ and java code prettier more manageable, but I don't think I have
> time to work on that right now, since they're non-functional improvements.
>
>   Please let me know if you come across any more bugs though!
>
> Thanks again,
> Mark
>
>
> On Wed, Apr 17, 2013 at 2:32 AM, Kedar Grama <gbkedar at gmail.com> wrote:
>>
>> Hi Mark,
>>
>> I turned on the following flags
>> ITK_USE_64BITS_IDS
>> ITK_USE_REVIEW
>> and compiled in Release mode on both windows and linux.
>>
>> Thanks,
>>
>> -Kedar
>>
>>
>> On Tue, Apr 16, 2013 at 9:28 PM, Mark Hiner <hinerm at gmail.com> wrote:
>>>
>>> Hi Kedar,
>>>
>>>>
>>>> I just wanted to send a quick note saying that I know Mark is working on
>>>> this problem. He told me earlier today that he had managed to duplicate the
>>>> bug (after initial unsuccessful efforts). He will reply back once he has
>>>> more information.
>>>
>>>
>>> Just wanted to confirm that I did reproduce the bug. Thanks for verifying
>>> it works with the normal BF command line tools. I'm hoping it's just a
>>> problem with the ITKBridgePipes java code.
>>>
>>> Looking at the error handling at line 746:
>>>
>>>  errorMessage += std::string( pipedata, pipedatalength );
>>>
>>> since the keepReading flag isn't changed here, I was thinking that it
>>> could be eating Java exceptions and then hangs waiting for input that never
>>> comes, because the Java side has crashed.
>>>
>>> Anyway I can't finish testing this tonight because I ruined my ITK
>>> install and don't have time to rebuild w/ tests. So I will investigate
>>> tomorrow morning.
>>>
>>> Side note - do you adjust any of the cmake flags with your build?
>>>
>>> Thanks,
>>> Mark
>>
>>
>