M. Nawijn a écrit :
> On Thu, Mar 12, 2009 at 6:08 PM, Arthur Magill <arthur.mag...@epfl.ch> wrote:
>   
>> hey Marco,
>>
>> That sounds like a smart idea. I think you've probably got further than I
>> have. I have OCC6.3.0 installed and running, but I haven't got very far with
>> pyOCC yet. I'll try and have a go tonight and see where I get to. I'm
>> building on i386 at home - I have an x86_64 at work, but it needs a bit of
>> work before I can install there.
>>
>> Do you have any tips before I start?
>>     
> Only one tip actually. If you try to build with SCons you have to configure 
> your
> environment in 2 places, the SConstruct file and the the
> environment,py file. The
> reason is that the SConstruct files runs the SWIG_generator.py file
> which in turn
> imports environment.py. I already reported this and it will be
> resolved in the near
> future.
>   
The SWIG_generator is called only if the related SWIG files are not 
found. Latest svn rev provides all required SWIG files. It's then not 
necessary to tweak the environment.py script.
> When I run SCons, it runs OK quit far, however it fails at somewhere
> at the end where
> it wants to build the wrapper for some X.. header (don't remember the
> exact name).
>   
Do you mean XSControl? This module also fails for me. It's however a 
*very* important module since most of the IGESControl(Reader/Writer) and 
STEPControl(Reader/Writer) objects inherit from the XSControl_Controler 
class. I made a small mistake in my SWIG_generator script; the fix is 
quite easy. I have to regenerate all the .i files and upload the result 
to the svn repository. Hope I'll have a few time tomorrow to do that.
> Tonight I will try to solve the SCons/SWIG annoyance that every
> wrapper is regenerated
> when only a single change is made.
>   
This is really an issue. I guess scons allows to customize such a behaviour.

Thomas
> Good luck,
>
> Marco
>
>   
>>        Arthur
>>
>>
>> M. Nawijn wrote:
>>     
>>> Hello Arthur,
>>>
>>> I think the two of us should hook-up a little (digitally speaking ofcourse
>>> ;) ).
>>> I am using Fedora 9 also on a x86_64 architecture. I am having a little
>>> bit of
>>> a rough time compiling pythonOCC, although I did not have to much of a
>>> trouble
>>> to compile OCC itself. I am very interested in creating yum
>>> installation packages for
>>> both OCC and PythonOCC, but I would first like to see a clean build.
>>>
>>> Regards,
>>>
>>> Marco
>>>
>>> On Thu, Mar 12, 2009 at 4:12 PM, Arthur Magill <arthur.mag...@epfl.ch>
>>> wrote:
>>>       
>>>> Hi Jelle,
>>>>
>>>> This is where the Linux world is really confusing if you're not used to
>>>> it, but I'll try and give you an overview ;-)
>>>>
>>>> Different distributions use different methods. The main once are .deb
>>>> (Debian and Ubuntu) and RPM (Fedora, SuSE).
>>>>
>>>> Actual distribution is then usually done through repositories - each
>>>> distro has an official one and often several unofficial ones. The
>>>> official one takes a while to get into, and involves things like legal
>>>> checks on security and the free-ness of the package (this is partly why
>>>> I'd like to see OCC go LGPL or similar). Unofficial repos vary from very
>>>> strict to 'hey, this is cool'. Once a package is in the official repo,
>>>> you get to type something like
>>>>
>>>>  > yum install OpenCASCADE
>>>>
>>>> and your computer checks all dependencies on other librarys, etc, and
>>>> installs everything you need. Cool when it works.
>>>>
>>>> OpenCASCADE now has an official build in Debian (thanks to the work of
>>>> Adam Powell and friends), and I think that has already trickled through
>>>> to Ubuntu (Ubuntu is heavily based on Debian). There is an RPM available
>>>> for SuSE but I don't know anything about it. I'm considering trying to
>>>> put together a Fedora RPM, but I don't have much experience here.
>>>>
>>>> There is also a package for Gentoo, largely put together by Daniel
>>>> Tourde. Gentoo is slightly odd because it doesn't really distribute
>>>> binaries - everyone compiles there own, but there is a very
>>>> sophisticated build system to make sure everything goes right.
>>>>
>>>> The neat thing about official builds is that you then know exactly where
>>>> to find things. Given the current state, I don't think OpenCASCADE will
>>>> be all that predictable on Linux for a little while yet, but this does
>>>> have the advantage that anyone who has managed to install it probably
>>>> knows what they're doing! It should be possible to include a small
>>>> script with pyOCC defining where things are, and telling users that they
>>>> should edit this to fit their system - this isn't such an odd request in
>>>> the Linux world. How does that sound?
>>>>
>>>> ...now, I must actually have another go at installing pyOCC rather than
>>>> just talking about it...
>>>>
>>>> Arthur
>>>>
>>>> jelle feringa wrote:
>>>>         
>>>>> great, this helps for the wiki.
>>>>> how are binary packages distributed for linux?
>>>>> it would be good to have packages for generic architectures / distros.
>>>>> pythonocc compiles in about 1/24th of the time of OCC, so it would be
>>>>> helpful surely.
>>>>> ( its not as if catia users compile catia ( they wish ) before
>>>>> installing it, I'm seeing it from that perspective )
>>>>>
>>>>> i'm not a linux fella, any thoughts? is this a feasible idea?
>>>>>
>>>>> thanks for the input Arthur.
>>>>>
>>>>> -jelle
>>>>>
>>>>> On Thu, Mar 12, 2009 at 3:26 PM, Arthur Magill <arthur.mag...@epfl.ch>
>>>>> wrote:
>>>>>
>>>>>    Hi Jelle,
>>>>>
>>>>>    Yes, I've used tips from Dan Heeks (and several others on the
>>>>>    OpenCASCADE list). In case it is helpful to anyone else here on the
>>>>>    list, I'm using Fedora 10, and the following worked for me:
>>>>>
>>>>>    <---snip--->
>>>>>
>>>>>    export CASROOT=/home/magill/occ/OpenCASCADE6.3.0/ros
>>>>>
>>>>>    cd ${CASROOT}
>>>>>    aclocal
>>>>>    automake
>>>>>    autoconf
>>>>>    cp /usr/share/libtool/ltmain.sh make/
>>>>>
>>>>>    flags="--with-gl-include=/usr/include --with-gl-library=/usr/lib"
>>>>>    flags="$flags --with-xmu-include=/usr/include/X11"
>>>>>    flags="$flags --with-xmu-library=/usr/lib"
>>>>>    flags="$flags --with-tcl=/usr/lib"
>>>>>    flags="$flags --with-tk=/usr/lib "
>>>>>    flags="$flags --with-java-include=/usr/lib/jvm/java/include "
>>>>>    flags="$flags --disable-debug --enable-production"
>>>>>    flags="$flags --enable-wrappers=yes"
>>>>>    flags="$flags --enable-wok=no"
>>>>>    flags="$flags --enable-draw=yes"
>>>>>
>>>>>    ./configure $flags --prefix=/usr/local/occ63
>>>>>    make
>>>>>    su
>>>>>    make install
>>>>>
>>>>>    <---snip--->
>>>>>
>>>>>    I don't think I need WOK, so I disabled that. I needed to pull the
>>>>> devel
>>>>>    package for Xmu (I guess also tcl-devel and tk-devel), but everything
>>>>>    else I needed was already present.
>>>>>
>>>>>    I'll be installing onto an eight-core x86_64 machine when I get some
>>>>>    time to upgrade it - I'll let you all know how it goes.
>>>>>
>>>>>    Arthur
>>>>>
>>>>>
>>>>>    Jelle Feringa wrote:
>>>>>     > Guys,
>>>>>     >
>>>>>     > I recall seeing some comments by Dan Heeks on building OCC,
>>>>>    perhaps useful?
>>>>>     >
>>>>>     > If you're lucky to have a mulit core machine, don't forget to use
>>>>>     > make -j 8
>>>>>     >
>>>>>     > where 8 is the number of cores.
>>>>>     >
>>>>>     > -jelle
>>>>>     >
>>>>>     >
>>>>>
>>>>>  
>>>>> https://www.blogger.com/comment.g?blogID=4746036612237842643&postID=7242152714259315360
>>>>>
>>>>>  
>>>>> <https://www.blogger.com/comment.g?blogID=4746036612237842643&postID=7242152714259315360>
>>>>>     >
>>>>>
>>>>>  
>>>>> <https://www.blogger.com/comment.g?blogID=4746036612237842643&postID=7242152714259315360
>>>>>
>>>>>  
>>>>> <https://www.blogger.com/comment.g?blogID=4746036612237842643&postID=7242152714259315360>>
>>>>>     >
>>>>>     > To build Open CASCADE, I downloaded the source code.
>>>>>    www.opencascade.org <http://www.opencascade.org>
>>>>>     > <http://www.opencascade.org> ( it was called OpenCASCADE_src.tgz
>>>>>    ) and
>>>>>     > extracted it. I got packages called tk8.4 and tk8.4-dev from
>>>>> Synaptic
>>>>>     > Package Manager
>>>>>     >
>>>>>     > I typed in a terminal:
>>>>>     > cd Desktop/OpenCASCADE6.3.0/ros
>>>>>     > mkdir build
>>>>>     > cd build
>>>>>     > flags="--with-gl-include=/usr/include --with-gl-library=/usr/lib"
>>>>>     > flags="$flags --with-xmu-include=/usr/X11R6/include/X11"
>>>>>     > flags="$flags --with-xmu-library=/usr/X11R6/lib"
>>>>>     > flags="$flags --with-tcl=/usr/share/tcltk/tcl8.4"
>>>>>     > flags="$flags --with-tk=/usr/share/tcltk/tk8.4"
>>>>>     > flags="$flags --disable-debug --enable-production"
>>>>>     > ../configure $flags
>>>>>     > sudo make
>>>>>     > sudo make install
>>>>>     > sudo ldconfig
>>>>>     >
>>>>>
>>>>>
>>>>>    --
>>>>>
>>>>>    Arthur Magill, PhD
>>>>>    RF Engineer
>>>>>
>>>>>    Centre d'Imagerie BioMédicale (CIBM)
>>>>>    Laboratory for Functional and Metabolic Imaging (LIFMET)
>>>>>    Ecole Polytechnique Fédérale de Lausanne (EPFL)
>>>>>    Station 6, CH F1 532 (Bâtiment CH),
>>>>>    CH-1015 Lausanne,
>>>>>    Switzerland
>>>>>
>>>>>      tel: +41 21 693 0569
>>>>>      fax: +41 21 693 7960
>>>>>    email: arthur.mag...@epfl.ch
>>>>>
>>>>>    _______________________________________________
>>>>>    Pythonocc-users mailing list
>>>>>    Pythonocc-users@gna.org <mailto:Pythonocc-users@gna.org>
>>>>>    https://mail.gna.org/listinfo/pythonocc-users
>>>>>
>>>>>
>>>>>           
>>>> --
>>>>
>>>> Arthur Magill, PhD
>>>> RF Engineer
>>>>
>>>> Centre d'Imagerie BioMédicale (CIBM)
>>>> Laboratory for Functional and Metabolic Imaging (LIFMET)
>>>> Ecole Polytechnique Fédérale de Lausanne (EPFL)
>>>> Station 6, CH F1 532 (Bâtiment CH),
>>>> CH-1015 Lausanne,
>>>> Switzerland
>>>>
>>>>  tel: +41 21 693 0569
>>>>  fax: +41 21 693 7960
>>>> email: arthur.mag...@epfl.ch
>>>>
>>>> _______________________________________________
>>>> Pythonocc-users mailing list
>>>> Pythonocc-users@gna.org
>>>> https://mail.gna.org/listinfo/pythonocc-users
>>>>
>>>>         
>> --
>>
>> Arthur Magill, PhD
>> RF Engineer
>>
>> Centre d'Imagerie BioMédicale (CIBM)
>> Laboratory for Functional and Metabolic Imaging (LIFMET)
>> Ecole Polytechnique Fédérale de Lausanne (EPFL)
>> Station 6, CH F1 532 (Bâtiment CH),
>> CH-1015 Lausanne,
>> Switzerland
>>
>>  tel: +41 21 693 0569
>>  fax: +41 21 693 7960
>> email: arthur.mag...@epfl.ch
>>
>>     
>
> _______________________________________________
> Pythonocc-users mailing list
> Pythonocc-users@gna.org
> https://mail.gna.org/listinfo/pythonocc-users
>
>   

_______________________________________________
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Reply via email to