[Insight-developers] getting a vtkmetaio error

[Mathieu Malaterre]

I would put a message(${VTK_SOURCE_DIR}) just above the error at least
to know what is the value like... this might help in tracking down
where it is being set.

2cts

On Tue, Dec 7, 2010 at 8:18 PM, Stephen Aylward
<stephen.aylward at kitware.com> wrote:
> I'm stumped.
>
> Any ideas?
>
> s
>
> On Tue, Dec 7, 2010 at 1:23 PM, Mark Roden <mmroden at gmail.com> wrote:
>> Hi all,
>>
>> I just wiped my build trees for gdcm, vtk, and itk, and then did a git
>> pull for all three.  I then built in this order:
>>
>> vtk (java wrapping on, win 64 release)
>> gdcm (java wrapping on, vtk wrapping on, win 64 release)
>> itk (use system gdcm)
>>
>> I still get the same error.
>>
>> Is there something else I should be looking at, some pollution of
>> cmake in general that could cause this?  My dashboard version still
>> had the bug as of last night as well.
>>
>> Thanks,
>> Mark
>>
>> On Tue, Dec 7, 2010 at 6:49 AM, Mark Roden <mmroden at gmail.com> wrote:
>>> Ah, this may be the issue.
>>>
>>> The vtk version on my dashboard is the latest greatest from git, and
>>> those CMakeLists.txt files match.  The vtk version I built last
>>> outside of dashboard is the 5.6.1 release branch, and that's a
>>> different CMakeLists.txt file from the ITK git version.
>>>
>>> This issue had come up before, but sporadically, and I think it was
>>> because I was regularly, and in no particular order, rebuilding vtk
>>> from either git master or from the 5.6.1 release in order to handle
>>> other issues.
>>>
>>> I erased my vtk 5.6 directories and gdcm directories and attempted to
>>> build against the vtk git that I had, but that didn't work.
>>>
>>> I'll try erasing everything and starting over, but it does appear that
>>> the following combination is broken:
>>>
>>> vtk 5.6.1
>>> itk git current with use system gdcm
>>> gdcm git current wrapping vtk
>>>
>>> Now, I've also wrapped gdcm and vtk with java, and not itk.  I don't
>>> think that should have an effect.
>>>
>>> I'll post how it goes.
>>>
>>> Mark
>>>
>>> On Tue, Dec 7, 2010 at 4:27 AM, Stephen Aylward
>>> <stephen.aylward at kitware.com> wrote:
>>>> Hi,
>>>>
>>>> Thanks Mathieu!
>>>>
>>>> If the fix has been in his source all along, I wonder why it has now
>>>> started failing.
>>>>
>>>> Mark - does the CMakeList.txt file in vtk/Utilities/vtkmetaio match
>>>> the one in ITK/Utilities/MetaIO
>>>>
>>>> s
>>>>
>>>> On Tue, Dec 7, 2010 at 3:40 AM, Mathieu Malaterre
>>>> <mathieu.malaterre at gmail.com> wrote:
>>>>> On Tue, Dec 7, 2010 at 5:33 AM, Mark Roden <mmroden at gmail.com> wrote:
>>>>>> I tried rebuilding all, and I also deleted and completely rebuilt my
>>>>>> itk directory.
>>>>>>
>>>>>> I also thought it might be a build ordering thing, since I last built
>>>>>> vtk as 32 bits and I was building itk as 64, but then I rebuilt as 32
>>>>>> and still got the error.
>>>>>>
>>>>>> Should I wipe my entire source tree and start again?  That seems... extreme.
>>>>>
>>>>> I would suggest so. Your dashboard machine 'silverstreak' was already
>>>>> using ITK from the master branch (which included the fix since July)
>>>>> and not the 'release' branch where the recent cherry-pick was done.
>>>>>
>>>>> HTH
>>>>> --
>>>>> Mathieu
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>>
>>>> ==============================
>>>> Stephen R. Aylward, Ph.D.
>>>> Director of Medical Imaging Research
>>>> Kitware, Inc. - North Carolina Office
>>>> http://www.kitware.com
>>>> stephen.aylward (Skype)
>>>> (919) 969-6990 x300
>>>>
>>>
>>
>
>
>
> --
>
> ==============================
> Stephen R. Aylward, Ph.D.
> Director of Medical Imaging Research
> Kitware, Inc. - North Carolina Office
> http://www.kitware.com
> stephen.aylward (Skype)
> (919) 969-6990 x300
>



-- 
Mathieu