[Insight-developers] Latest WrapITK merged in ITK

[Bradley Lowekamp]

Gaetan,

Over night on my mac laptop I was able to build Wrap ITK with swig 1.3.31, OK. It was submitted as an experiential under malawi.local. There is only one failing test due to a time out, but a whole bunch of warning. I am going to configure a new nightly build on a mac with this system default version of swig. I feel that this version of swig needs to be supported given in comes with the macs, unless there is a very compelling benefit to requiring a newer version.

The build with swig 1.3.29 did not submit as expected.

Brad

On Jan 25, 2011, at 4:41 PM, Gaëtan Lehmann wrote:

> 
> Le 25 janv. 11 à 22:31, Bradley Lowekamp a écrit :
> 
>> 
>> On Jan 25, 2011, at 3:16 PM, Gaëtan Lehmann wrote:
>> 
>>> 
>>> Brad,
>>> 
>>> Sorry, I'm not sure how but I managed to miss your mail.
>> 
>> No worries.
>> 
>>> 
>>> We are recommending the version 1.3.38 or newer, but haven't forced a minimum version as the required version depends on many things — for example, the python version used, the integration of the doxygen documentation in the python one, etc.
>> 
>> Hrmm.. This explains why I have not been able to build WrapITK today. Well I have 1.3.31 which is the version on my mac, and on the cluster, I get 1.3.29. Looking for release date of these I find
>> 
>> 1.3.38 1/31/2009
>> 1.3.31 11/20/2006
>> 1.3.29 3/21/2006
>> 
>> I am lazy, how needed are the newer feature of swig? I don't like to have to build, and keep around more non-system tools then really needed. Too easy to get too many different versions of things laying around. Any ways it could be more compatible with older versions of swig?
> 
> We can probably restrict the flag used to fit the swig version, but there might still be some bugs in older swig versions.
> 
> Can you submit an experimental build, so I can try to figure out what needs to be done?
> 
> Gaëtan
> 
>> 
>> Brad
>> 
>>> 
>>> At some point, we may have to force a minimum version though, to make the debugging possible from the report of our users.
>>> 
>>> 
>>> About the performance, swig provides a -O options which makes the generated code a lot faster. From the swig doc and for python:
>>> 
>>>     -O              - Enable the following optimization options:
>>>                        -modern -fastdispatch -nosafecstrings -fvirtual -noproxydel
>>>                        -fastproxy -fastinit -fastunpack -fastquery -modernargs -nobuildnone
>>> 
>>> I'm not sure which version adds those optimizations.
>>> 
>>> Gaëtan
>>> 
>>> Le 23 janv. 11 à 14:59, Bradley Lowekamp a écrit :
>>> 
>>>> Gaetan,
>>>> 
>>>> You said that you  got a lot of performance gain from a newer version of SWIG. I don't see which version number you recommend. I have SWIG 1.3.31 that came with my mac. Which version do you recommend?
>>>> 
>>>> Brad
>>>> 
>>>> On Jan 23, 2011, at 5:24 AM, Gaëtan Lehmann wrote:
>>>> 
>>>>> 
>>>>> Dear developers,
>>>>> 
>>>>> I've juste merged the WrapITK branch from
>>>>> 
>>>>> https://github.com/ComplexSystemsModeling/ITK/tree/wrapitk
>>>>> 
>>>>> in the ITK main repository. I would have prefer to merge the branch
>>>>> without changes, to be able to continue the development in the same
>>>>> place, but unfortunately I had to rebase the branch to add the
>>>>> required prefix in the commit name. When we began to work on WrapITK
>>>>> branch, this was not enforced by a hook.
>>>>> 
>>>>> The CSwig wrapper has been removed. The most important changes are
>>>>> listed at
>>>>> 
>>>>> http://code.google.com/p/wrapitk/wiki/Release030
>>>>> 
>>>>> A few other thing are not listed there - mostly improved wrapping
>>>>> coverage and an explicit instantiation prototype.
>>>>> 
>>>>> So far everything seems to be OK on the dashboard.
>>>>> 
>>>>> Please let me know any problem you may find.
>>>>> 
>>>>> Regards,
>>>>> 
>>>>> Gaëtan
>>>>> 
>>>>> 
>>>>> -- 
>>>>> Gaëtan Lehmann
>>>>> Biologie du Développement et de la Reproduction
>>>>> INRA de Jouy-en-Josas (France)
>>>>> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
>>>>> http://voxel.jouy.inra.fr  http://www.itk.org
>>>>> http://www.mandriva.org  http://www.bepo.fr
>>>>> 
>>>>> <PGP.sig><ATT00001..txt>
>>>> 
>>> 
>>> -- 
>>> Gaëtan Lehmann
>>> Biologie du Développement et de la Reproduction
>>> INRA de Jouy-en-Josas (France)
>>> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
>>> http://voxel.jouy.inra.fr  http://www.itk.org
>>> http://www.mandriva.org  http://www.bepo.fr
>>> 
>> 
>> ========================================================
>> Bradley Lowekamp
>> Lockheed Martin Contractor for
>> Office of High Performance Computing and Communications
>> National Library of Medicine
>> blowekamp at mail.nih.gov
>> 
>> 
> 
> -- 
> Gaëtan Lehmann
> Biologie du Développement et de la Reproduction
> INRA de Jouy-en-Josas (France)
> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
> http://voxel.jouy.inra.fr  http://www.itk.org
> http://www.mandriva.org  http://www.bepo.fr
> 

========================================================
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/20110126/7cff5b09/attachment.htm>