Curves heavily scaled down will have render artifacts
Caused by precision issues when computing average normal map
for INDEX3 surface type. Now calculation happens in local object
space instead of world space.
When deleting all keyframes in F-Curves, the corresponding F-Curves are deleted.
If all the F-Curves in an action group were deleted in such a way, the group
wouldn't be removed. This meant that these groups would never be shown (until
F-Curves for these groups were created again), but would still exist, causing
problems when trying to rearrange groups in the animation editors (i.e. groups
would appear to not move). Now these groups get deleted when they get empty.
Removal of this line fixed the bug.
But!
I suspect it opens up for an excaption.
I will try to identify the exception and handle it in a different way.
Notes and image:
http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.66/Usability
- now each editor has own settings for "show panel header" and
"show panel background", and colors+alpha for this.
- this setting used to be global for all editors, but it can conflict
with looks of specific editors.
- Now you can set for editors to show panels with a 100% transparent
tool/properties region.
Note: read XML theme files now might get an error, Campbell will fix.
now blenlib/BLI doesn't depend on any blenkern/BKE functions,
there are still some bad level includes but these are only to access G.background and the blender version define.
With larger header sizes (via DPI setting), splitting an area horizontally (using
left-bottom corner widget, move up) stopped immediate after split, not allowing
to drag it to a position.
Culprit was code to check minimum header size, and area operator using the same
flags. Now ScrVert has two flags - one for internal use, one for the tools to set.
This commit is an attempt to improve collisions between moving Bullet rigid bodies using (concave) triangle mesh bounds.
Instead of using Gimpact, this we create a btCompoundShape with child shape tetrahedra derived from the surface triangles.
For each triangle, we add a fourth vertex using the centroid, shifting inwards using the triangle normal.
If the centroid hits an internal triangle, we stop. The default depth could be exposed as 'advanced' setting in the user interface.
This solution will be a slower than the original/gimpact solution, but a bit more reliable.
In the future, it is better to add HACD, convex decomposition to Blender, for moving concave meshes.
See http://kmamou.blogspot.com and the Bullet SDK's Demos/ConvexDecompositionDemo.
impossible. In this case matrix inversion failed and didn't give a useful result.
Now it falls back to a pseudoinverse in that case, so that moving along the
other axes still works.
There may be other places that can benefit from this, but this is a place where
it has no significant performance impact, doing this in general for e.g. bone
matrices could be quite slow.
On saving a first .blend after startup, the file would load back as if it was not
saved (showing no name in header).
The whole FILEFLAGS and G.relabase_valid and G.file_saved etc is messy.
This commit fixes issues, but only adds more mess :)
Will discuss a nicer implementation of all of this.
- Move to layer (and more popups) didn't scale yet
- User Prefs and render window now open on right location on Mac Retinas
- Brush sizes for painting now scale for Mac Retina
It was strange logic in code from 2010 which forced image aspect to be 1 for
viewer nodes and render results.
Not sure why it's needed, was only used for unwrapping aspect correction,
but render result/viewer images are already handled differently there.
