2009/9/19 Simon Loic <simon1l...@gmail.com>

> Hello,
> First, thank you! You did more for me than the opposite so far.
> About, the PAF package, I don't understand exactly why, but you are right :
> no need to add PAF in __init__.py. I don't know for what reason I had to add
> it in the first place, but it totally works without it.
> Concerning the backtrace, I'm glad you appreciate the feedback. Though, I
> can't really help more, as I don't know enough OCC so far to understand.
> The only thing I can think of, would be to compile OCC in debug so that you
> could have further information in the backtrace. If you're interested, let
> me know. In that case I would appreciate some hints on how to efficiently
> turn debug build on (I guess it is to run ./configure --enable-debug,
> right?).
>

Right. But I have an idea about the problem. According to your
experimentations, the occ_unittest and topoloygy_unittest pass. So the
problem does not come from OCC but, according to me, rather from
salomegeometry. It's possible that salomegeometry is compiled without the
D_OCC64 flag, so the adress handler of OCC is not properly inited. Maybe you
should try to recompile salomegeometry with this flag (for instance, first
set CPPFLAGS='D_OCC64' or something like that). If it solves the problems, I
will fix the Makefile.am file of salomegeometry and commit changes to the
repository. Otherwise you can recompile OCC with --enable-debug to have more
information about this issue.

Under MacOSX 64bits, I don't face this issue.


>
> regards,
> Loïc
>

Thomas


>
>
> On Sat, Sep 19, 2009 at 8:42 AM, Thomas Paviot <tpav...@gmail.com> wrote:
>
>> 2009/9/18 Simon Loic <simon1l...@gmail.com>
>>
>>> Hi again,
>>>
>>
>> Hi Loïc,
>>
>> Manu thanks for your feedback.
>>
>>>
>>>
>>> The env_DRAW.sh script contains a set of environment variables that has
>>>> to be set up before you can use OCC. You can copy/paste the content of this
>>>> file to your .bash_profile. The most important one is the CSF_GraphicShr,
>>>> that points to the OCC OpenGL lib, and enables the 3D display.
>>>>
>>>
>>> Ok I had already something equivalen in my .zshrc (I like this shell
>>> better than bash ;-), and CSF_GraphicShr is set correctly.
>>>
>>
>> Ok.
>>
>>
>>>
>>>
>>> The best way is certainly to launch this python script from the
>>>> ./Tools/InteractiveViewer directory. Do you have any difficulty to get it
>>>> run?
>>>>
>>>
>>> It's ok the basic example works (a part from the Zoom window which is
>>> apparently not implemented or not wrapped, right? I get the following
>>> message:
>>> >> AttributeError: 'Viewer3d' object has no attribute 'Zoom_Window'
>>>
>>
>> I will fix it.
>>
>>
>>>
>>>
>>>> Weird, since you followed the instructions available on the PAF tutorial
>>>> (from the wiki). Before you go and test PAF, I suggest you shoud first 
>>>> check
>>>> that pythonOCC works fine. In the /src/unittest directory, there are 3 sets
>>>> of unittests: occ_unittest (test of the basic features), paf_unittest (for
>>>> PAF) and topology_unittest (topology/geometry handling). Run each of these
>>>> scripts and be sure that they all pass:
>>>> python occ_unittest.py
>>>> python paf_unittest.py
>>>> python topology_unittest.py
>>>>
>>>
>>> Actually the paf unit test raise a seg fault error (the two others run
>>> just fine): see the output file attached.
>>> One more thing I forgot to tell. After installing pythonOcc, the PAF
>>> module was missing in the __all__ list in dist-packages/OCC/__init__.py. I
>>> had to ad it manually.
>>>
>>
>> PAF is not a python module, but a python package, e.g. a directory that
>> conatins a set of python modules. According to me, it has nothing to do in
>> the __init__.py script located in the   site-packages/OCC directory.
>>
>> Thanks for the gdb trace. It seems to be a memory management issue related
>> to OCC. I also need the output of the paf_unittest.py script to have more
>> information.
>>
>>
>>>
>>> cheers,
>>>
>>
>> Cheers,
>>
>>
>>>  Loïc
>>>
>>
>> Thomas
>>
>>
>> _______________________________________________
>> 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
>
>
_______________________________________________
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Reply via email to