Blender Git Loki

Alembic: Fix compilation error with alembic disabled

Was a mismatch in stub factory methods, presumably after some recent
changes in hair/strands.

workaround for VBOs on Mesa

Merge branch 'alembic' into gooseberry

Conflicts:
source/blender/physics/intern/BPH_mass_spring.cpp

Fix for broken and messy particle time values to get consistent caches.

Particles use 0..100 range for parent particle "times" (curve parameter)
and 0..1 for children. For caches this is now uniformly calculated in
the 0..1 range to avoid breaking child interpolation.

Used the mass parameter for hair simulation.

Optional force and jacobian return values from angular bending springs,
for occasional debugging.

BGE DeckLink: fix extend mode.

The algorithm copied from VideoTexture's fast scaling was working only for
shrinking but we also need expanding here. Fixed that.

Merge branch 'alembic' into gooseberry

Fix for install_deps script: declare "skip-alembic" as a formal
parameter or it won't be parsed at all.

Fix for missing function argument commented out during cleanup.

Child strand deformation based on parent motion states when reading from
caches.

This allows child strands to follow the motion of parent strands as
calculated by the hair simulation. Unlike the particle system path
caching, this method does not regenerate all the children every frame.
Instead, it relies on an initial child setup that defines the base
shape, then applies the parent motion as a weighted linear deformation.

The result is in fact the same as with regular particle clumping, kink,
roughness etc.. Particles also take care to not change the child shape
when the parent moves, but this is not formalized anywhere (which makes
changing particle code quite difficult). So ignoring the particle code
is both permissible and efficient.

Fix for stupid particle times.

Normalize particle 'time' values to 0..1 range when writing to cache,
anything else is useless.

Fix crashes on undo when workflow has a nodetree.

Still getting memory leaks on exit, they likely occur because main is
owner of nodetree but we should give it to workflow instead.

Merge branch 'asset-experiments' into asset-engine

Merge branch 'master' into asset-experiments

Conflicts:
source/blender/blenlib/intern/BLI_filelist.c
source/blender/imbuf/IMB_thumbs.h
source/blender/imbuf/intern/thumbs.c

Workflow shader fixes:

* ID type for GPUWorkflow shaders gets refined correctly
* GPUWorkflow correctly linked to spaceview3d when loading file
* Correct Icons in Node editor
* Creating a node tree works (but it's empty)

BGE DeckLink: small linux compilation fix.

BGE VideoTexture: 3D support+performance boost on DeckLink.

ImageViewport will now skip the unnecessary RGBA32 filtering if the capture
is already in the RGBA format. This will be the case if alpha=True and no
filter is set. This saves a great deal of CPU.
Add swap attribute to DeckLink to skip pixel format conversion between
VideoTexture internal (RGBA) and DeckLink (BGRA). This is to test the
benefit of passing the image directly to the board without intermediate
copy => great boost of performance.
However, it will be necessary to change the VideoTexture internal pixel
format from RGBA to BGRA.

3D stream is now fully supported on DeckLink, the right eye image is
passed via a new 'right' attribute (name not definitive). To generate this
image one can simply put a second camera to figure the right eye and
capture its view with ImageRende.

Cycles: free mesh BVH after packing the scene BVH

The idea is to avoid having original BVH and BVH packed into scene BVH and hence
reduce memory footprint.

This change doesn't give any measurable differences on it's own, but together
with D1215 and D1217 it gives quite nice results.

Reviewers: campbellbarton, brecht, juicyfruit

Subscribers: eyecandy

Differential Revision: https://developer.blender.org/D1218

Cycles: Synchronize images after building mesh BVH

This way memory overhead caused by the BVH building is not so visible and peak
memory usage will be reduced.

Implementing this idea is not so straightforward actually, because we need to
synchronize images used for true displacement before meshes. Detecting whether
image is used for true displacement is not so striaghtforward, so for now all
all displacement types will synchronize images used for them.

Such change brings memory usage from 4.1G to 4.0G with the 01_01_01_D scene
from gooseberry. With 01_01_01_G scene it's 7.6G vs. 6.8G (before and after
the patch).

Reviewers: campbellbarton, juicyfruit, brecht

Subscribers: eyecandy

Differential Revision: https://developer.blender.org/D1217