Blender Git Commits

Fix silly mistake in nearestVert

Merge branch 'blender2.8' into transform-manipulators

Conflicts:
source/blender/editors/space_view3d/view3d_edit.c
source/blender/editors/space_view3d/view3d_ops.c
source/blender/editors/transform/transform_manipulator.c

Merge branch 'blender2.8' into custom-manipulators

Conflicts:
source/blender/editors/space_view3d/view3d_edit.c
source/blender/editors/space_view3d/view3d_ops.c
source/blender/editors/transform/transform_manipulator.c

Cycles: Fix building of CUDA kernels after last commit

Cycles: Deduplicate sample range setting and getting

Cycles: Add generic work item functions for OpenCL

Cycles: Add generic work item functions for CUDA

Change angle function calls in sdef

Change angle function call in 3d to 2d mapping function

Make caching a bit more sane (fixes plasticity issue)

This cleans up the cloth caching logic a bit.

Now, a simulation step only occurs if the current time is exactly one
frame after the last simulated step, thus preventing time gaps and
runtime data corruption. Also, now the cloth is only re-initialized
if the cache is actually outdated, instead of always being reset when
on the first frame.

This solves a bug where the cache was freed when moving the time to
a point within the cached period.

These changes also fix the issue of non-cached runtime simulation state
properties (e.g. plasticity) being reset when going to the first frame
or retaining a future state when simulating over an existing cache with
time gaps.

Outliner: initial work for render-layers

Remove Base references, still need to do some adjustments

FOREACH_VISIBLE_OBJECT macro

Also adding a new flag value for ObjectBase to store visibility.

I still need this to be synced, but the idea is to centralize the
logic of tree evluation, and keep the visibility cached.

Initial syncing of selection and visibility flag

When we make a collection invisible or unselectable, selected object have to be re-evaluated
Same goes for when we add a new object, its base visibility has to be
refreshed.

TODO also add a call for this when we remove a scene collection, or
unlink a scene layer.

Added object mode wires to new drawing pipeline.

Cycles: Implement new NLM reconstruction kernels

Use ObjectBase only to delete objects

Cycles: Rendering multiple tiles at once for OpenCL mega kernel

Basically the same as for CUDA. No major performance difference, but in
some cases this is a little faster than single tile rendering.

Cycles: Implement the get_ideal_tile_size function

Since devices will be rendering many tiles at once the tile size chosen
is on the smaller side.

A limited selection of tiles sizes is provided as an option. This may be
removed later, for now its mostly useful for debugging.

Add get_ideal_tile_size function to RenderEngine

This allows render engines to specify an ideal tile size to use while
rendering. The function is optional, if provided the returned tile size
will override the scene settings, if not provided the original behavior
will be used.

Cycles: Rendering multiple tiles at once for CUDA

There's no major performance difference from rendering a single large
tile. However if there are too many tiles (such as if the tiles are
16x16) the overhead of updating tile display can become a bit
noticeable. This should be easy to fix by restructuring the tile update
code a little.