[Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender  trunk/blender/source/blender/ python/intern/bpy_rna.c: disallow setting RNA attributes while drawing, this is bad practice so enforcing here has the benifit of making sure
bartius.crouch at gmail.com
Wed Dec 8 09:49:50 CET 2010
What might be helpful is to have an example of recommended practice.
Speaking for myself: right now it's often a matter of experimentation to get
things to work and this might result in solutions that depend on bad
For example, if you wish your add-on to display a panel with a custom value
depending on what scene you are in, it would seem logical to add a custom
property to the Scene ID class.
Initialising this property from within the panel draw code is bad practice,
but what to do? Initialising upon registration fails when the add-on is
enabled and saved to the default .blend (context doesn't exist yet when
add-on is run at startup). And not initialising of course results in an
error when the panel (displaying the custom property) is drawn.
> This raises the question: are most extension authors reading bf-committees
> mailing list?, should we discuss python api design on bf-python?
Bf-python is probably the most logical place to discuss python api design,
but you raise a good point: how many script writers are actually subscribed
to it? For the record, I'm reading both bf-committers and bf-python.
More information about the Bf-committers