Blender Git Commits

Force kernel re-compilation check for viewport render

Switch definitively to int32[4] for uuid's, mixing mixing it with char[16] is confusing at least.

This is only on human-level anyway... And endianess should not be an issue
either, hopefully, that way!

Removed the unused particle writer/reader code from Alembic.

We only use particles to initialize strands export now.

Only process hair simulation in "realtime" mode.

This means no motion state data is generated for render mode, this
requires some larger changes to the way render/realtime data is stored
in the caches.

Merge branch 'asset-experiments' into asset-engine

Merge branch 'master' into asset-experiments

Merge branch 'alembic' into gooseberry

Object and particle system selection in the cache hair sim modifier.

Before this the hair sim would be applied indiscriminately to all the
strands in the cache data. Now an object/psys combination from the dupli
cache must be selected. Note that this is not the actual target of the
hair simulation, which still operates on the cache overrides instead.
The Object/Psys only functions as a selectable key, if no matching data
is found in the cache no hairsim will be applied.

[BCYCLES-212 Fix] Make split kernel work on multiple OpenCL devices

Bug fix : Dont use ASSIGN_RAY_STATE macro for first assignment

[BCYCLES-209] Refactor : rename variables sd_shadow -> sd_DL_shadow

[BCYCLES-209] make direct lighting and shadow kernels share members of
shaderData buffers

[BCYCLES-209] make direct lighting and shadow blocked kernels share shaderData global buffers

Print message to console incase we need to split tiles further

Bug fix : Always make max_render_feasible_tile_size a multiple of local work size

WIP better/fixed support of uuid's in asset engines.

Still not working as expected it seems... Not having a real bytes type in RNA is a real PITA. :/

Blender displays fine again without errors, using the custom matrix
stacks.

Still need to replace glRect calls to completely fix it probably.

This still does not use the soc API properly.

Hello non crashy world! Still not 100% correct but at least now it fires
up.

Issue was getting matrices from GL instead of own matrix stack

Trying to solve text crash issues

More fixes...

Amber engine seems to be running nice again in new system, pfew!