added object_tfm_protected_backup, object_tfm_protected_restore so its easier to transform the object and respect protected channels (otherwise you need checks everywhere for each channel which is verbose).
* Code cleanup, changed some splits to row's, much nicer when you only use 2 properties in a row.
* if > elif
* Removed some duplicated code for an if/elif/else check
Did some minor separator() changes too.
* Fix excessive fireflies in Velvet BSDF (patch by David).
* Disable some unused SSE code
* Remove RTTI disabling flags for now, this is giving some compile issues and
was only needed of OSL which we're not using yet.
* Added BKE_dynamicpaint.h to CMake list.
* Changed some unsigned integers back to signed, as OpenMP can't iterate unsigned values anyway.
* Fixed more warnings.
* Add back option to bundle CUDA kernel binaries with builds.
* Disable runtime CUDA kernel compilation on Windows, couldn't get this working,
since it seems to depend on visual studio being installed, even though for
this particular case it shouldn't be needed. CMake only at the moment.
* Runtime compilation on linux/mac should now work if nvcc is not installed in
the default location, but available in PATH.
This bug is caused by exactly the same reason as #26316: differences in how new vertices/edges
are getting calculated first and how they're adding later. In some cases extra vertices are
creating which weren't counted before.
This patch prevents crash in such situations, but result mesh can be a bit wrong.
This should work fine in bmesh, so think it's acceptable to have such workaround
before actual fix coming with bmesh.
ended up having to add a new pointer into the uiBlock (which I'd rather have avoided), but setting the uiLayoutSetContextPointer(..) was complicated to properly use for submenus and popus.
Beforehand, a few problems were in view: some of the indexing was changed towards the end to use the more efficient stack structure, but needed to use the correct def group indices.
Also, the designated selected group would use its own value to acquire the standard to base change distributed to the others.
Lastly, the total_change was used rather than -left_overs in the formula to compute the new designated weight within the means of the locks' allowed changes.
Now, while maintaining the ratios of the selection, it correctly returns left over change that could not be redistributed to the unlocked groups.
* OpenEXR doesn't need debug suffix
* Fix libmv template issue when linking by removing duplicate libmv inclusion. I wonder how this never turned up in release builds as well.
Important: Since OpenImageIO went into trunk, OpenEXR, possibly along with other image libraries will need to be turned on too because OIIO depends on them.