Previously, calling Project from View for UVs when viewing through a camera produced bad results. You had to scale and stretch the resulting UVs, never resulting in a true "from view" with a camera. Now, what you get in the UV editor is what you saw through the camera.
Separated preview drawing into own ARegion, this should make using View2D possible
The Sequencer now has three view types: Sequencer, Preview and split Sequencer/Preview.
Changing the preview can be done either by the combobox in the header or toggling through those types with CTRL+TAB.
Notes:
* Icon for split Sequencer/Preview view missing still.
* Naming items in the comboboxes can be improved (just Preview instead of Image Preview?)
Next steps:
* bringing back View2D handling (zoom/pan) for image preview
* experimenting with splitting the Preview ARegion for In/Out editing
* option to roll the delta of the arm rig.
* fix to copy metarig type
* renamed EditBone.align() --> EditBone.align_roll()
* Added EditBone.align_orientation(other)
* Added bone.vector: same as (bone.tail - bone.head)
textbox width. I can't tell if this is working correct, the code is very
cryptic, with gotos even, but seems there is not enough space allocated
to permit splitting each character to it's own line.
Also fixes:
* Weight paint subsurf drawing.
* Missing pointer endian conversion in paint brushes.
* Use of unitialized variable in screen version patch.
* Multires modifier without mdisps layer crash.
again on library linked objects. The former is ok because it is actually
the scene data being edited, the latter will not be saved to file but is
useful..
* Added a User-Pref option for the "XYZ to RGB" colour-mode setting for new F-Curves to compliment the one used for Keying Sets. With this option enabled, the builtin Keying Sets also can obey this option.
* Made all places that were previously manually checking the flags for keyframing to use a standard API function to do this now.
* Fixed bug introduced earlier today in commit 25353 by reverting the changes to keyingsets.c. Forgot that delete_keyframe doesn't handle do the "entire array" hack with array_index = -1
* Fixed bug with the insert-keyframe code for the array_index = -1 case, where too many channels were being keyed (i.e. an imaginary channel was often keyed in addition to the valid ones)
- Offset parameter of the path constraint has no effect on the animation
- Path animation was being repeated multiple times even though it was not supposed to
Autokeying for transform functions now gets context-info, allowing for bone paths to be recalculated.
However, the main purpose of this is to allow KeyingSets to eventually have poll functions.
Also put a bit more logic for guessing player paths based on my system.
If anyone can make this a bit more clever/bulletproof, please feel free to
get involved in it, it's all python!
* added renamed files in revision 25337
* renamed BLI_util.h -> BLI_path_util.h for consistency
* cleanup of #includes: removed BLI_blenlib.h in favour of direct includes of the needed headerfiles in a few places.
* removed debug print in sequencer.c
* added missing include in blenkernel/blender.c -> bad dependency, needs to be fixed still
Removed the Start/End Settings and the effect_fader from the sequencer transform strip.
The transform strip can now only be animated directly by keying the Scale X, Scale Y, Translate X, Translate Y and Rotation values.
Caveat:
The uniform scale re-uses the Scale X value, so when changing the animation on Scale X in the non-uniform scaling case, the uniform scaling will be affected too. This was done to not break files for Durian.
Note:
As much as I would have liked to clean up the TransformVars, it caused crashes of earlier versions of blender when opening the file.
The rna identifiers were also kept as to not break any existing animation on those keys.
