[Insight-developers] Uncovered Code

Hans Johnson hans-johnson at uiowa.edu
Wed Sep 7 10:41:51 EDT 2005


Luis,

I have reworked the building of FFTW and all the CMakeList and
FindFFTW.cmake code to allow users to select either the single, the double,
or both versions of FFTW.

I have broken backwards compatibility a little (there was no way around it)
that does not affect part of the ABI that users are supposed to be aware of.

Instead of specifying USE_FFTW, you must either specify USE_FFTWF and/or
USE_FFTWD to select the version of the library that you want.

The advantage of the new system is that it will force compiler errors at
compile time of the library.  Previously the errors only showed up at link
time, and the errors were difficult to decipher.

These changes have been working on our Iowa builds for 3 weeks now, so they
have been tested pretty well.  I hope to have them committed by the end of
today.

Regards,
Hans



On 9/7/05 8:20 AM, "Luis Ibanez" <luis.ibanez at kitware.com> wrote:

> 
> 
> Hi John, Torsten,
> 
> Thanks for pointing this out.
> 
> Zion ITK Gcc 3.4, the machine that is running the coverage in
> the Dashboard, didn't have FFTW activated in its CMakeCache.
> 
> This has been added now, so the next submission of coverage
> will include those methods that require FFTW calls.
> 
> 
> 
>     Thanks
> 
> 
>       Luis
> 
> 
> 
> ----------------------
> Torsten Rohlfing wrote:
>> 
>> Hans:
>> 
>> I don't see how that is the case. The itkCurvatureRegistrationFilter
>> uses double precision data internally and makes all calls to "fftw_XXX"
>> functions, which are in the double-precision library.
>> 
>> TR
>> 
>> Hans Johnson wrote:
>> 
>>> Torsten,
>>> 
>>> This test requires the floating point version of fftw to be installed
>>> on the
>>> system that does coverage.
>>> 
>>> Hans
>>> 
>>> 
>>> 
>>> On 9/2/05 12:16 PM, "Torsten Rohlfing" <torsten at synapse.sri.com> wrote:
>>> 
>>>  
>>> 
>>>> Bill:
>>>> 
>>>> The first two entries on your list,
>>>> 
>>>> ./Algorithms/itkCurvatureRegistrationFilter.txx date: 2005-05-03 author:
>>>> ibanez
>>>> ./Algorithms/itkFastSymmetricForcesDemonsRegistrationFunction.txx date:
>>>> 2005-05-03 author: ibanez
>>>> 
>>>> were written by me. I also sent Luis a test back then, and it seems its
>>>> in the CVS repository
>>>> (Testing/Code/Algorithms/itkCurvatureRegistrationFilterTest.cxx). This
>>>> tests both above classes. Am I missing something?
>>>> 
>>>> Torsten
>>>>   
>>> 
>>> 
>>>  
>>> 
>> 
>> 
> 




More information about the Insight-developers mailing list