2014-03-11 5:30 GMT+01:00 Cory Cohen <cfco...@verizon.net>: > Hello Thomas and others, > > I recently tried to update my version of OCE to the latest tip at GitHub, > and while OCE built fine, it appears that the latest version of PythonOCC > is not compatible with OCE versions more recent than 0.12. > > I was a bit surprised to find the current tip of PythonOCC nearly a full > year behind OCE. Is anyone able to comment on the issues surrounding the > synchronization? Are there technical problems, or is it just a resource > issue? Or perhaps there's some reason why keeping PythonOCC on a more > stable release is attractive? Hopefully PythonOCC development is still > active. If it's not, that would be important to me... > > I ask in part because I've managed to core dump python several times while > working with PythonOCC, and would like to see whether the latest version of > OCE has corrected the problems. It seems likely that these problems are in > OCE and not in PythonOCC. > > I'm willing to work on the bringing PythonOCC release up-to-date, but not > if I'm wasting or duplicating effort. Is the delay caused by some thorny > problem that will take me hours to get to where you're already at, or is it > just a matter of nobody having slogged through the build errors? Is there a > plan for a new release that I'm unaware of? > > Any guidance that you or anyone else on the list can provide would be > appreciated. > > Cory > > Hi Cory,
Thanks for posting! On my osx machine, I run a pythonocc on top of oce-0.14. Sharing this work is not as simple as it should be (see below). The lack of synchronisation between OCE and pythonOCC can be explained as follows: let v_oce be the development speed of OCE, and v_pythonocc the development speed of pythonOCC. OCE involves many more users/devs than OCE so we can consider that v_pythonocc = 0.5*v_oce. As a consequence, both projects are active and evolve, but the distance between them grows with the time: d = (v_oce - v_pythonocc)*t ;) To improve the synchronization between both projects : * first, it's not necessary IMO that pythonocc follows the short oce release cycle ; * second, we should speed up pythonocc developments. Three issues that slows down pythonocc development: (1) * Porting issues are the cause for a high viscosity in the pythonocc dev pipe. So far, I used to maintain both linux/osx/win32 platforms, but I don't have the time anymore. (2) * Another issue is that the library I chose at the beginning of the project, to generate swig files from c++ headers became deprecated (FYI, pyplusplus and pygccxml). This is especially an issue on Windows platforms. (3) * Too many bug reports/contributions/patches deal with some python scripts. pythonocc development issues are not related to its usage. Solutions: * for (1), pythonocc project needs linux and windows maintainers. The problem is that each linux distribution has its own specificities, the same for windows (32/64, XP, 7, 8 etc). * for (2), the solution is to use a pure python C++ header parser: CppParserHeader. * for (3), pythonocc needs people comfortable with C++, Python and Swig. (1) and (3) depends on the community. (2) I started the port from pygxxml to CppHeaderParser ( https://pypi.python.org/pypi/CppHeaderParser/2.4). The port to oce-0.14 is an going work available in the dev branch tp/oce-0.14 on the github repository. Best Regards, Thomas
_______________________________________________ Pythonocc-users mailing list Pythonocc-users@gna.org https://mail.gna.org/listinfo/pythonocc-users
