Henry Gomersall wrote: > I was thinking about thistoday (triggered by the original email!). > > How realistic is it to create an API-extractor definition -> Cython > tool, combined with the (possibly more developed) pxd generation tool: > http://www.google-melange.com/gsoc/project/google/gsoc2012/redbrain1123/28002 > > (which seemingly is in an alpha state) > > Is there more to creating QT bindings than that? How much effort would > that be? > > It strikes me that Cython solves many of the difficult problems already, > and it supports C++ pretty well now.
The above is a SoC project which means it's just been developed by a student and probably never used, so it would undoubtedly require some work. Otherwise it may not be a bad approach. I don't know if the custom code and hints from PySide could be used with that. If not, it's really just a complete replacement of PySide, which is probably more work than is needed. My general sense is the problems in PySide may be things that are not solved by this anyway. Many of the bugs seem to bein the manually written code and hints and not the basic C++ wrapper machinery. You still need to find the API to wrap, and then know what to do with objects that get passed in and out (wrapping them correctly, managing their life cycle, and in some cases tweaking or translating parameters). At least some of this is defined by Qt semantics and not basic C++ behavior so no generic binding/calling solution is automatically going to solve those. I could be wrong about some of this. :-) - Stephan _______________________________________________ PySide mailing list PySide@qt-project.org http://lists.qt-project.org/mailman/listinfo/pyside
