Launching the player from the Blender UI now makes use of the "Debug Properties", "Framerate and Profile", and "Deprecation Warnings" options by setting the appropriate command line flags.
When tabs to spaces is enabled and a paste command is undone, the improper number of characters could get removed. Also fixed issue with shift + left/right only selecting a max of 1 character.
- add missing headers from cmake (own omission)
- quiet rna_test.c unused define warnings.
- minor style edits
- spelling corrections and ignore all uppercase words with spell checking script.
rotation of the brush mask texture.
Unfortunately secondary path does not work here because we do not have a
permanent switch to choose between primary-secondary brush texture. Use
operator property instead.
Adding a new node in Node Editor failed for "High DPI" (Only Mac retina now).
- Py script for adding nodes was doing dpi magic, which it shouldn't. It has
been replaced with a (temporary) API call to set the correct cursor location.
(Thanks to Lukas T for helping here)
- The SpaceNode->cursor[2] property now is *only* storing the coordinate
in "adding new node space". Use of this has been removed from the code where
possible, with as only exception the code to draw noodles while adding them.
Special coder note: Nodes should respect the DPI value, and draw larger with
larger buttons if you increase this size. The hack here is that this can only
work nice if also the node positions are scaled accordingly.
A better fix could be to check on scaling the node view itself for it. That
then would also remove this Python API call that was added in this commit.
However, that again might fight with how buttons layout code works now...
needs some careful checking.
For Blender this release is interesting because of:
- Covariance estimation (not used in Blender yet, but now we
might use it for keyframe selection instead of havingown
implementation).
- Significant performance improvements to loss function and
dense linear solvers and automatic differentiation.
Unfortunately, didn't notice speedup of tracking itself,
but camera reconstruction now happens around 2 times faster
on my laptop,
- Better inner iteration step acceptance and stopping.
The issue with wrapping is that it requires correct interpolation of the border pixels. Since interpolation is done at the far left end of the node tree in buffer/image/etc read operations, the wrapping
setting can not be used directly in those operations (otherwise in-line translate operations would cause conflicts). To make wrapping work correctly we need to add a buffer in front of the translate
operation, which can then be interpolated correctly based on wrapping. The WrapOperation becomes a variant of ReadBufferOperation, which uses its wrapping setting to determine the correct "extend" mode
for interpolation of the buffer.
This merges all mix operations into a COM_MixBaseOperation
(naming could be better, but this way it corresponds to what's
going on with math operations.
Same was done with RenderLayers operations.
Overall this gives 20% of bf_compositor library compilation
time decrease. And it was rather annoying to have tens of
files with just a single-line constructors anyway.
TODO:
- All Convert operations could also be merged into a single file,
but that would require adding some ConvertBaseOperation to
reduce code duplication (ideally). Leaving it unchanged for now.
- Some operations' headers seems to be wrongly including MixOperation
header, they need to include NodeOperation instead it seems.
This is rather harmless, but would be nice to doublecheck on this
eventually.
* Avoid special code, when Subsurface is enabled.
Ideally we should only use the function, and get rid of the extra duplicate, but this is slower on CUDA.
