Blender Git Loki

Minor cleanup.

Merge branch 'master' into temp_custom_loop_normals

Rework threaded code...

Now small meshes do not use threading at all (allows to be 6 times quicker on a 200 verts cube e.g.),
and threading code use ThreadQueue as it should have been doing since the beginning.
We do not gain much here, though, two or three percents at most on a 1M verts monkey
(using a mempool might help here, though, have to check this yet).

Review update: some small performance enhancements, and cleanup.

Bump OpenEXR to 2.2 for windows MSVC 2013

Review updates: Replace 'bbox center' option by an offset one.

Warp modifier: remove redundant NULL check

Warp modifier: avoid sqrt checking radius

Fix T43524: Warp modifier curve clamped to 0-1

Cycles: OSL kernel now also ignores maybe-uninitialized

This is a bit of a mess because of the header dependency hell, but
since the tweak is small enough it's gonna be fine.

Cycles: Further tweaks to T43511 to solve compilation error on 32bit platforms

Cycles: Solve dependency between camera and object synchronization

IN theory object might depend on camera location (spatial adaptive subdivisions
for example) which became not possible to achieve after camera in volume support.

Should be no functional changes for artists.

Cycles: Ignore -Wmaybe-uninitialized from the kernel in release builds

This warning provided too much false-positive issues in release version of the
kernel, making it really easy to miss actual warnings.

First round of changes from campbell's review.

Mostly renaming, not everything addressed yet.

Selection: avoid uninitialized memory read with occlusion queries

Fix for hair child roughness.

The fixed-function roughness was not getting used once the curve
roughness has been enabled at least once.

Adjust GLSL drawing of window coordinates from camera view (use camera
bounds instead of window bounds). Fixes remaining part of T43346

Fix for unfreed effector weights in hair dynamics.

The internal cloth modifier effector weights get replaced temporarily
to make the cloth sim use the particle weight settings instead. But
the particle sim was not putting back the original weights, which can be
non-NULL in case the cloth sim allocated these already. Messy design ...

Fix for unfreed memory due to unnecessary early exit in the free
function.

Fix for unfreed memory of internal hair cloth modifier pointcache.