so if i understand this correctly, pyopengl consist of an auto-generated "raw" opengl layer and another layer on top of it which makes it more "pythonic"? and the auto-generated ctypes "raw" layer could in theory be replaced with an auto-generated pyrex "raw" layer? and this could make pyopengl a bit faster? :) sounds interesting...
i also find the pypy project very fascinating. i am looking forward to see the first parts of it getting production ready and also to see how fast they can get python with their jit work.
> would it be possible to write a tool which automatically converts
> ctype wrappers to pyrex? wouldn't that be nice?
> i am no computer scientist though and maybe this is a stupid idea
> which isn't feasible. :)
In fact, PyPy is able to translate a subset of ctypes into a number of
code types including C code with type inferencing allowing for
significant code simplification... i.e. the code should get close to the
speed of directly C-coded source code.
The wrappers are also built from gcc-xml output (for the "raw" API), I
*believe* pyrex already has a mechanism that allows for the same
transformation (i.e . build a raw API from raw gcc-xml). Which is to
say, we could, in theory, provide a way to register a "raw" API plugin
for PyOpenGL that would allow for entirely replacing the core-code
operations (or something like that). That way a compiled system could
provide a run-time pluggable replacement for the pure-python code... in
Have to get back to work now.
Mike C. Fletcher
Designer, VR Plumber, Coder
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
PyOpenGL-Users mailing list