The way it works now is:
- Border select is using TWEAK, for which you need to hold LMB down
and move mouse. This prevents operators using RELEASE/CLICK events
from being fired.
- LMB select is using RELEASE event, this is no selection happens
before border if you do border select. And this prevents any
operator uses CLICK event from being fired.
- Delect all happens by CLICK, which would be fired only in case no
operators were handled with PRESS/RELEASE.
This is a bit cards-house, but this is how events currently works and
wouldn't want inventing something bigger now.
The do-version handling for Userdef is outside file reading, which makes
it needed to store the file version in UserDef, so it gets the correct
version to handle.
Thanks Antonis R. for pointing at the omission!
Also removed the mindboggling define. If you do such, then make it
like "MAIN_VERSION_OLDER_THAN() or so.
In general version hacking could be limited much better... ask me
before even thinking to add one, most optimal is to do it in a way
it's not depending on a version ever - forward/backward compatible.
This also revealed another bug, as you could not explicitely set default context to text_ctxt UI func parameter (None is not accpeted by RNA string props), so I had to change default context from py POV to "*" instead of None.
Anyway, that physics UI translation remains weak, as the trick used here (helper func) prevents message extractor script to directly find them. Currently it works because specified labels are also defined elsewhere, but it would be nice to have some kind of "translation markers" in py code too (similar to our N_/CTX_N_ C macros, unfortunately python does not have preprocessing ;) )...
automatically flip the unpinned faces to point down too, instead of trying to
unwrap them in the other direction and giving bad results.
If there's a mix of faces pinned up and down it will pick the direction with the
biggest area.
Adding non-RGBA float buffers as texture crashes MipMap code.
Images with less than 4 channels not supported (yet). Not sure
if we ever should even... normals, Z, vectors, UV, Index, all
don't like downsampling.
In some cases objects might have rigid bodies but aren't in the rigid
body group, just add objects to the group then.
Also allow changing the rigid body type by using the add acive/passive
buttons. This avoids having to remove objects just to change type.
Remore group notifiers, they're already sent by add/remove group operators.
Add pointcache notifiers.
Fixes UI not being updated properly in some cases.
Tweak for new option for Relative Bone parenting
(which transforms child object based on rest pose, so you can change bones
in editmode to define pivot)
In the original commit it was made default, but that was too invisble for
users. Now it's an option in the Make Parent menu to choose. Communicates
a new feature better.
