* Tiny tweak of GL commands used when drawing F-Curves (single GL_LINES instead of multiple GL_LINE_STRIPS) to hopefully improve the performance with heaps of handles drawn a bit
* Spelling fix for initTransDataCurveHandes -> initTransDataCurveHandles
The code for transforming a mixture of keyframes with bezier and non-bezier interpolation was crashing.
The old code only took all the handles when a keyframe was bezier, and one when it was not; but sometimes this underestimated the situation (the first handle is only really used if the previous keyframe was bezier, as per the standard evaluation rules for these, but it didn't really check for this). Now, it just adds them whenever, since there is the possibility that keyframes may be moved before other unselected ones, in which case the handles may become invalid.
Thanks to Lee (from Durian, who found the crash), and Jess Balint (who had submitted a patch with some steps towards fixing this)
The point where it zooms to can now be specified as operator properties and will be used if zoom to pointer is on. (when not specified, it will use the x and y of the event)
* Some code cleanups to match layout file code guidelines.
* Ported Operator Labels to op file itself.
* Added some Tooltips, theeth: Feel free to change/improve them as you like. :)
The dark bar for sliders will now be drawn in a way that takes into account the state of the slider (this includes info about keyframing and/or drivers). This means no more ambiguarity about whether these sliders are animated or not!
I've had to make a little hack here to make sure that the sliders bar will still be visible after the state has been blended, but this should probably be solved nicer by adding a separate setting for blending-factor in this case.
Added a base-template for editing/creating RNA-Paths. This is now used for KeyingSets and Driver UI's, so that when the actual magic gets put in, it will work.
- Small changes to make ghost_cocoa compatible with 10.4 + gcc4.0 (the initial goal was to be 10.5+ compatible, but 10.4 is finally also possible)
- Main window title is now in Apple document window title style (proxy icon + filename)
- fix for top menu "Blender" sub-menu not anchored correctly in 10.5
New current_orientation property that returns the current transform orientation data (if any)
New UI elements to rename and remove custom orientation (visible only when needed)
* Enums with an _itemf callback now never get context NULL passed in,
rather a fixed list of enum items are defined which should contain
all items (if possible), from which the _itemf callback can then use
a subset.
This commit implements more of the CopyPose capabilities in
Blender. It is now possible to select which axis will be
constrained in position and orientation to obtain
interesting effects. Another option selects if the axis
are relative to the end effector or to the target.
Unlocking a position axis means that the coordinate along
this axis is not constrained and can take any value.
Unlocking the Y axis of the End Effector produces an
'aiming' effect: the end effector is oriented towards
the target but without stretching.
Unlocking a rotation axis means that the end effector can
freely rotation along that axis. Unlocking the Y axis
produces a 'tangent' effect: the end effector aligns with
the Y axis of the target but can rotate along that axis.
A 'floor' effect is possible if the position Z axis of the
target is unlocked. More effects are possible an can be
combined.
it would show you the last selected shape key until doing
another operation. This is confusing, and the Pin button allows
you to do the same kind of shape browsing.
* Converted Drivers UI to mostly use the Layout Engine
* All the buttons that perform actions are not operators yet (the code for that would be quite icky still)
* I've added some (commented out) calls for the property definitions of Driver properties to perform Depsgraph updates. I've left these commented out until we have the option to turn off auto-updates, since with driver editing, that could be very dangerous.
* Drivers can now (in theory) use any ID-block, using the Any-ID template added earlier. However, be warned that the stupid depsgraph won't be able to cope with most of these cases.
TODO:
- more fancy widgets for RNA-Path and Index will come later
