Re: [Insight-users] ITK 3.16 Packages for Ubuntu Karmic Koala (9.10)
BTW, I cannot build ITK+CSwig+Tcl8.5 anymore. I'll check if switching
to wrapitk work around the compilation issue with Tk_PhotoSetSize
within cswig code.
cheers
On Mon, Nov 2, 2009 at 9:50 AM, Mathieu Malaterre
<mathieu.malaterre@gmail.com> wrote:
> Steve,
>
> If you have been reading this thread, what is your point of view ?
> Do you think you can move to ITK/WrapITK (this would solve the numpy
> vs numarray issue as a side effect) ?
>
> thanks,
>
> 2009/11/2 Gaëtan Lehmann <gaetan.lehmann@jouy.inra.fr>:
>>
>> Le 2 nov. 09 à 09:16, Mathieu Malaterre a écrit :
>>
>>> Luis,
>>>
>>> That is terrific ! At least Kitware's roadmap is clear ! Thanks so
>>> much. Gaetan, could you remind me (us?) why you choose to developg
>>> google/WrapITK instead of directly working on ITK/WrapITK ?
>>
>> Sure: I made drastic changes in the wrapping process, and it was absolutely
>> impossible to make it without breaking things.
>> Also, I made most of the work on my spare time, and so I had no idea about
>> when things may be usable again (if usable at all).
>> Last point: cableswig has been replaced by swig (the unmodified one) and a
>> custom program - igenerator.py - to parse the output of gccxml and produce a
>> swig interface. Unfortunately, igenerator.py is written in python, so it
>> adds a dependency on python to build java or tcl wrappers.
>>
>>> Is it
>>> possible to backport google/WrapITK into ITK ?
>>
>> There are still a few problems:
>> * dependency on python due to igenerator.py. It should be possible to recode
>> it in C++, but it's a work with a very low priority on my todo list, so it's
>> very unlikely to be done soon. But is it really a problem?
>> * there is a bug - I think in swig - which makes some methods not be wrapped
>> in tcl. See this test failure for example:
>> http://www.cdash.org/CDash/viewTest.php?onlyfailed&buildid=461742
>>
>>>
>>> Thanks again !
>>>
>>> On Sun, Nov 1, 2009 at 10:19 PM, Luis Ibanez <luis.ibanez@kitware.com>
>>> wrote:
>>>>
>>>> Mathieu,
>>>>
>>>> As we move forward to ITK 4.0, it is likely that WrapITK will
>>>> become the official way of wrapping ITK, and at that point
>>>> Paul's way of of packaging should probably become the
>>>> standard.
>>>>
>>>> At this point, however, the fault is really on us, as developers of
>>>> the toolkit, for not having made our minds about fully supporting
>>>> one wrapping system or the other.
>>>>
>>>> Hopefully, ITK 4.0 will give us the opportunity to clean up
>>>> many corners of the toolkit.
>>>>
>>>> Luis
>>>>
>>>>
>>>> -----------------------------------------------------------
>>>> On Wed, Oct 28, 2009 at 4:34 PM, Paul Novotny <paul.novotny@gmail.com>
>>>> wrote:
>>>>>
>>>>>> Hi Paul,
>>>>>>
>>>>>> I still do not understand why you insist on building an outside
>>>>>> ubuntu package.
>>>>>
>>>>> Well, I build them for my own internal use, and just decided to release
>>>>> them to the public. Since there are people who use and appreciate it, I
>>>>> continue to do it.
>>>>>
>>>>>>
>>>>>> Ref:
>>>>>> https://launchpad.net/ubuntu/+source/insighttoolkit
>>>>>>
>>>>>> What is wrong with the one in ubuntu ? Why not work with the
>>>>>> ubuntu-team to get issues fixed ? Why not work with the debian-med
>>>>>> team to get it fixed (earlier in the pipeline) ?
>>>>>
>>>>> There is nothing wrong with the ubuntu/debian version. They appear to be
>>>>> using the CSWIG wrapping version. I have used and continue to use
>>>>> WrapITK in my own work, and therefor create the packages. So there is
>>>>> nothing to 'fix'.
>>>>>
>>>>>> BTW, I did post a couple of days ago about usage of CSWIG vs WrapITK
>>>>>> vs WrapITK. It seems you have chosen the solution #2 (the unmaintained
>>>>>> WrapITK shipped with ITK). Could you comment on your choice please ?
>>>>>
>>>>> As far as why I chose #2, well, I think the post from Gaëtan Lehmann you
>>>>> referenced perfectly sums up why (if you don't mind me paraphrasing).
>>>>> CSWIG is older and doesn't have as much itk coverage. Wrapitk, great
>>>>> python integration, better itk coverage, and is tested and stable. The
>>>>> new WrapITK, while faster, cleaner, and with better code coverage is
>>>>> unstable, and isn't as well tested.
>>>>>
>>>>> So, in the end, I don't understand why you don't like it that I release
>>>>> these packages?
>>>>>
>>>>> -Paul
>>>>>
>>>>> _____________________________________
>>>>> Powered by www.kitware.com
>>>>>
>>>>> Visit other Kitware open-source projects at
>>>>> http://www.kitware.com/opensource/opensource.html
>>>>>
>>>>> Kitware offers ITK Training Courses, for more information visit:
>>>>> http://www.kitware.com/products/protraining.html
>>>>>
>>>>> Please keep messages on-topic and check the ITK FAQ at:
>>>>> http://www.itk.org/Wiki/ITK_FAQ
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://www.itk.org/mailman/listinfo/insight-users
>>>>>
>>>>
>>>
>>>
>>>
>>> --
>>> Mathieu
>>
>> --
>> 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
>>
>>
>
>
>
> --
> Mathieu
>
--
Mathieu
Reply to: