Blender Git Loki

Sculpt: Fix gcc compile error

Sculpt: brush UI and debug changes

* You can now edit brush input mappings
inside the main workspace brush panels.
* PBVH_CHECK_NAN now limits how many reports
it prints from a given source file.
* Fixed various problems with shift-smooth strength
setting.
* Fixed a NaN in the smoothing code.

Sculpt: fix memory corruption

* Fixed a nasty bit of memory corruption
* Along the way, added ASAN support to
bmesh customdata blocks.
* Each CD layer is padded by 32 bytes
inside the bmesh data block.
* Also fixed a few minor errors in
mempool's asan support.
* Tried and failed to fix numerical
stability issues with lasso/box
trim brushes.

Sculpt: fix bug with building paths for
mapping curve inside the ui

Sculpt: fix versioning error and add cutoff option
for square and cutoff mapping functions.

Sculpt: forgot yet again to increment the subversion

Sculpt: brush input mapping improvements

* Input mappings now take a premultiply factor
to scale the input data prior to evaluation;
* Mapping data can also now be fed through a
(wave) function prior to evaluation.
* The UI now has seperate inputs and outputs
sections for input mapping to avoid confusion.
* Added a distance mapping and implemented the speed
mapping.
* Also fixed original data bug in color filter.

Merge branch 'master' into asset-greasepencil

Sculpt: fix missing NULL check for wire edges

Sculpt: fix crash in PBVH_FACES

Sculpt: forget to increment subversion in
last commit

Sculpt: brush input mappings improvements

Cleaned up brush channel input mappings:

* BrushMapping now stores .min/.max
* BrushMappingDef .min/max now sets BrushMapping min/max
instead of changing the curve preset bounds.
* Fixed how BKE_brush_channel_eval_mappings evaluates
the mappings stack. Mappings now blend directly
with channel value instead of accumulating a multiplier
that's applied at the end.
* Consequently, BrushMapping->blendmode now defaults to MA_BLEND_MULT.
* Exposed BrushMapping->blendmode in RNA and UI.
Note that it doesn't support every MA_BLEND_ type,
it provides its own EnumPropertyItem list of
supported blendmodes.
* Added a random input method, BRUSH_MAPPING_RANDOM.
* Fixed BRUSH_MAPPING_ANGLE being given data in the wrong
range (all channels should be 0..1, not -pi..pi).

Other changes:
* Improved the uv smooth brush. It's still hidden behind an
experimental pref.
* Added a SCULPT_temp_customlayer_has function to check if a temporary
customdata attribute layer exists.
* Fixed a bunch of broken sliders in the paint_toolsystem_common.py.

Sculpt: smothing ops now slide UVs

* Wrote a new function, SCULPT_reproject_cdata,
to reproject loop customdata after smoothing.
* SCULPT_reproject_cdata is only called if UV
layers exist.
* All of the smoothing tools (hopefully all)
use it.
* This change is necassary to properly support vector
displacement maps in the future; otherwise DynTopo
will introduce lots of noise into the uv tangent
space.

Prepare ground for bool and int attributes.

Separate attribute and VBO types for the conversion.

Also do not promote readily all attributes to float4.

Merge branch 'master' into temp-abc-features

Fix precision issues and a bug in vec_roll_to_mat3_normalized.

When the input vector gets close to -Y, y and theta becomes totally
unreliable. It is thus necessary to compute the result in a different
way based on x and z. The code already had a special case, but:

- The threshold for using the special case was way too low.
- The special case was not precise enough to extend the threshold.
- The special case math had a sign error, resulting in a jump.

This adds tests for the computation precision and fixes the issues
by adjusting the threshold, and replacing the special case with one
based on a quadratic Taylor expansion of sqrt instead of linear.

Replacing the special case fixes the bug and results in a compatibility
break, requiring versioning for the roll of affected bones.

Differential Revision: https://developer.blender.org/D9551

Split and extend unit tests for vec_roll_to_mat3_normalized.

Separate the huge test into huge logical parts and add more cases
to check. Also add a utility to check that the matrix is orthogonal,
with arbitrary epsilon values and calculations in double.

A couple of tests deliberately fail, to be fixed in following commits.

Ref D9551

Fix T82455: vec_roll_to_mat3_normalized returns NaN when nor close to -Y.

In this case theta is completely unsafe to use, so a different
threshold based on x and z has to be used to avoid division by zero.

Ref D9551

Sculpt: fix adding shapekeys in sculpt mode
zeroing verts.