The addons panel draw function calls addon_utils.modules() which in turn retrieves a list of fake modules from the script paths every time. This can become costly when network paths are included for
addons. Solution is to put the scanning process into a dedicated "refresh" function and disable it in frequently called draw and filter functions, i.e. in these cases the cached addons_fake_modules list
will be used instead.
Note that this may lead to invalid addon lists if script paths are changed (which is not working 100% without restart anyway according to Campbell). For this there is now a "Refresh" operator button in
the addons preferences. If necessary and feasible such forced refreshes can be added later too.
* Remove difflib's quick_ratio() call, from quick tests looks it is four times slower than ratio(), and yet it gives the same results...
* Moved code to find all po's in a dir into its own func (will be used by i18n addon tools too).
Also workaround a nasty bug, where unregistered py classes remain listed in relevant __subclasses__() calls, which would lead to crash with python addons i18n tools (main translation was not affected, as messages extracting tools are executed in a brand new "factory startup" Blender ;) ).
We have a glicth with colormanagement's spaces descriptions, though, looks like they are clamped at 64 chars (see raw space), will see that later, if it’s solvable.
Fix for keyingsets tips, and make them (and a few others) findable by i18n messages extracting code (for some reasons, their bl_rna.description are void???).
Most notable difference from now on will be that all py is handled from current blender's resource dirs, no more from source dir. Better for consistency, and avoid e.g. cycles' addon to be checked twice...
Notes:
* Everything is still a bit raw and sometimes hackish.
* Not every feature implemented yet.
* A bunch of cleanup is still needed.
* Doc needs to be updated too!
Also enhanced the ugly py ast parsing code, so that it now can handle (up to some extent) "name" function nodes, and add bpy.app.translations.pgettext func familly to extracted ones...
This ast py parsing becomes more and more ugly, should probably try to refactor it a bit. :/
Also generate rigid body constraint types in py bullet code from RNA enum values (simpler than having to sync the code when something is changed here!).
Side note: RNA API about icons still needs to expose icons for enum values, and conversion funcs between icon_name and icon_value!
