This was a regression since 55719 and the reason is:
- image_refresh() will tag combo for re-calc when editing mask.
This is a feature, so you could have immediate feedback on masking
and compositing while editing mask. This is done from image_refresh().
- Color management sends NC_WINDOW notifier, so the whole window is
updated when changing color management settings. Image editor used
to refresh space when this notifier is sent.
The same notifier is sent by compositor job to redraw all possible
viewers.
Simple fix: don't refresh image space for NC_WINDOW, there's nothing
in image_refresh() which we would want to happen on NC_WINDOW event.
Global undo/redo now clears the local undo editmode stack entirely.
Error goes back to the 2.3 days, when undo was added.
Global undo refreshes the entire internal database, so all ID pointers get invalid.
This cases editmode undo storage to fail, if it uses ID pointers as well.
The error was that for any Mesh undo stack, a single global undo would make the
mesh stack corrupt. Back in edit mode, on undo you'd lose assigned texture images,
or get bad crashes.
The downside is that people expect this to work... it's a fun feature to maintain
stacks separately. Having this instable is not acceptable though. Needs quite some
redesign work to solve it (like Dalai's kill-the-tface project :)
* Shift-Rclick and holding right click, x or y will trigger constrained
scaling of overlay. Pressing again will revert to uniform scaling.
* Added operator, visible under the mapping drop menu, to fit stencil
aspect ratio to brush image aspect ratio.
* Made it possible to access stencil attributes from python as vec.x,
vec.y. Thanks to kgeogeo for pointing out!
Issue was caused by some mesa drivers does not support GL_RGBA16F.
Now added check around glTexImage2D to verify whether requested
internal format is actually supported. If not blender will fall
back to non-GLSL image display.
Thins brings up some speed improvements:
SPARSE_SCHUR is approx 1.3-1.5x times faster
ITERATIVE_SCHUR is approx 1.2x times faster
For blender this means camera solution go a bit
faster now. Would not have affect on tracking
speed.
modifiers with text/curve/surface objects.
This is because the code here did not yet support the distinction between
generating the tesselated mesh at viewport/render resolution and generating a
separate displist/derivedmesh for render that is not stored in the object but
created and freed on the fly.
- add rna property 'as_bytes' method so you can get a string property as python bytes (bypass encoding).
- make bpy.path.abspath/relpath compatible with bytes.
- add 'relpath' option to bpy_extras.image_utils.load_image(), so you can load an image relative to a path.
Fix for 34494 Blender 2.65 regression test error - compo_map_uv_cubes.blend - stripe/artifact between cubes
Hue and saturation node has an early break when saturarion is 0. When this happened the input 1 color needed to be used. This behaviour was not merged.
When no FSAA is used in the ZCombine. a mask will be created, this mask will be antialiased and based on this mask the colors between the two images are blended. This was also behaviour that was not merged correctly. Now it is back making much better z-combines.
Hope nobody uses these gabs as a work around.
- At Mind -
Jeroen & Monique
- "Add node" was showing on wrong location when used via pulldown menus.
Now this option will put the nodes in center of the view.
- The Curves widget was making itself smaller/bigger based on width of region.
That messes up the layout engine now - especially the code that checks if
there's a scroller needed or not (it went into an eternal feedback loop).
Now this widget has fixed height (like the other larger widgets).
Better would be to allow such large widgets to be scaled vertically individually.
That's for the todo!
while the docs said it followed the settings in the Output panel, other file
formats work now.
Benderplayer still only saves PNG now as documented, but I cleaned up the code
there to reuse existing imbuf functions rather than using own libpng code.
- When changing clip in clip editor, remove all frames
from it's cache to free memory for new clip.
- When changing proxy render settings, free cache as well.
Made it an operator instead of automatic prefetching.
Filling the whole memory with frames is not always
desired behavior.
Now prefetching is available via P-key, or from Clip
panel in toolbox or from Clip menu.
Also enabled prefetching for non-proxied movies.
rather then getting the longest edge, get the edge which which is most different from the 2 others ends up giving more useful results: for an isosceles triangle it returns the base weather its longer or shorter then the other sides.
