Sounds good. A hook to override the add_test calls would do the trick.

Adopting MIDAS in MITK itself for managing large test data files may be
worth considering as an alternative to the current monolithic MITK-Data
tarball. It would help solve the problem of maintaining a fork that
diverged from some arbitrary MITK commit and for which the
MITK-Data_*.tar.gz file no longer exists. But then I guess that's not an
officially supported use case of MITK and not a major problem to begin
with. Maybe the benefits are not worth the added complexity - I'll let the
MITK team decide.

Regards,
Taylor


On Wed, May 28, 2014 at 12:07 PM, Sascha Zelzer <[email protected]
> wrote:

> Hi Taylor,
>
> Thanks for considering to contribute your MIDAS support to MITK. As you
> know, we are currently not using MIDAS in MITK itself so while I think it
> is a nice feature in general, I am hesitating a little to include it in
> MITK directly because of our unfamiliarity with it and the additional
> maintenance.
>
> I will bring the topic up in one of our next meetings anyway. We could
> also add a CMake hook in the mitkMacroCreateModuleTests.cmake file which
> would allow you to override the add_test() call and hook in your own MIDAS
> version externally (midas_add_test).
>
> Best,
> Sascha
>
>
> On 05/05/2014 08:55 PM, Taylor Braun-Jones wrote:
>
>> Hi all,
>>
>> I have added MIDAS support to my MITK fork, but since I prefer my fork to
>> diverge as little as possible from the mainline master, I thought I'd run
>> by the idea of incorporating the change into MITK itself. The required
>> change looks something like this:
>>
>> https://github.com/nocnokneo/MITK/commit/31b760768fb8021e59df37fda44c0a
>> 90ef70b508
>>
>> Though it would need to be rebased against the latest MITK cmake
>> scripts... My fork is a bit behind.
>>
>> The change allows test arguments like MIDAS{file1.nrrd.md5} to be used.
>> When a test target is run then file1.nrrd gets downloaded locally and the
>> local path to file1.nrrd is substituted for MIDAS{file1.nrrd.md5} when the
>> test is actually executed. If there are no test argument strings matching
>> MIDAS{...} then there is no change to the current behavior.
>>
>> Thoughts?
>>
>> Regards,
>> Taylor
>>
>>
>
------------------------------------------------------------------------------
Time is money. Stop wasting it! Get your web API in 5 minutes.
www.restlet.com/download
http://p.sf.net/sfu/restlet
_______________________________________________
mitk-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mitk-users

Reply via email to