I've had to add a special operator for this, since the generic region-flip operator doesn't seem to be getting the right region from RNA. This operator explicitly searches for an appropriate header region before beginning. I suspect by default, operators get the main region set as being context target...
Renaming "Sync Audio" to "Realtime Playback" to make it clearer about its purpose. Internally, it's still called "sync_audio" since that's strictly what it represents, but that could be changed later if there is a need.
This contains two entries - one which calls the region flip operator, and one which calls the maxmize area operator.
Unfortunately, there seem to be some context issues which are causing the wrong region to get activated for use by the region flipping, meaning that nothing happens.
Also, fixed own typo in 3d-view header/menu code...
Playback of animation will now be restored to the frame that playback started on when stopping if the SCREEN_OT_animation_cancel operator is used (i.e. ESCKEY).
All other ways of stopping playback (pause button, alt-a) should act more like toggles, and do not reset the playhead.
http://wiki.blender.org/index.php/User:Theeth/Snapping
The new widget isn't there yet, but the rest works, including:
Increment (gears) is now a snap mode
Ctrl click to toggle snap on or off (in transform and 3d view)
Shift+Ctrl click to select snap mode (3d view only)
Snap status (on/off) is persisted in tool settings, no need to always turn it back on when entering transform.
It's still possible to have something resembling the old system by editing the transform modal keymap to snap on on ctrl press and snap off on key release.
- Started cleaning up the sequencer drawing code by firstly decoupling the different draw modes from each other (i.e. timeline view doesn't call image view).
- Also separated out a few distinct few phases in sequencer-timeline drawing into different functions instead of being lumped in the single one. Recoded part of this to make it less ugly too...
- Made markers get drawn again in the sequencer timeline view
This patch, by Elia Sarti (vekoon), simply adds the possibility to specify the final array component of the RNA path in the path itself,
e.g. using location[0] or location["x"] or even location.x,
instead of specifying this using an "array_index"
This should be easier for users to understand the driver system. The array-indices have been kept (but hidden from the UI under standard situations) since they are theoretically a tad faster than the in-path lookups still, and are easier for internal-tools to set for now...
the RNA constraint api was checking the current context when getting the constraint driver path and renaming constraints.
this made scripts not work properly so changed this to search for the constraint pose channel user within the object (if the object its self is not the user).
When metaballs were added to the render, if they weren't the basis ball, they'd be skipped, leaving a render object with no geometry. Now it doesn't add an object in the first place.
- Drivers on added to the 'armature' datablock (i.e. keyframing some settings for a "Bone" as opposed to "PoseBone") now evaluate correctly. Added proper recalcs for this case too.
- Also fixed some memory leaks and loading problems I encountered with the test file provided. After having problems loading the test file, I ended up reproducing and finding the error.
Symptom: first button drag gets stuck on values.
Cause: Buttons didn't set grab bounds, to use the whole window. But Ghost didn't do anything in that case, it left the bounds value as is. It only affected the first cursor grab, because cursor ungrab sets it to -1, which forces it to use the whole window size (checking every mouse move) for subsequent grabs.
Solution: When NULL, init bounds to window size (and don't query every mouse move).
This is working around an Apple OpenGL driver bug, chatted to Martin about this in IRC, seems ok. Hopefully can revert this if/when Apple comes up with a driver fix.
