panda3d python side is a mess!
I am trying to generate better documination by simply importing all the files and reading the doc attrib but it turns out that most of them execute alot of stuff… making this impossible. I dont think its right for a python library to do this … make lots of automated python tools not work properly.
I suppose this has to do with the all the globals constructed during the import (DirectStart spring to mind).
I remember that david wrote once that it was because one dev wanted things to be “convenient”. If I remember correctly, david is not a fan of this approach.
( Personnally, I would like a more explicit approach. )
Isn’t it possible the dynamically detect and import a submodule of a known (already imported) module. A kind of module walker (as in os.path.walk).
With a bit of magic (DirectStart again), that should work, no ?
Later,
Maybe forking the panda3d’s python side and reworking how the global stuff is managed … but then thats alot of work!