Blender Git Loki

parsing of voro++ data should work now, but a crappy mesh appears, hmm

attempt to parse voro++ data from string, not working yet

Fix for point cloud construction: Use a point struct to avoid difficult
float arrays.
Also added Frac prefixes to prevent potential name collisions.

A few small fixes for the voro++ C API, which seem to have fixed a
freezing issue when building with clang.

Also min/max initialization for point cloud construction.

cmake cleanup: removed the ++ suffix from VORO cmake variables and
defines.

Stub code for voro++ integration.

Not functional yet, but needs testing for potential voro++ bug.

Removed the ShardList return value of the fracture function.

This is not used atm and it's a bit ugly to return such allocated
values. If necessary this can be added back later.

Updated the pseudo-fracture method to new Mesh types.

New method for constructing the DerivedMesh for the fracture modifier.
Uses a simple union of all the shard mesh data.

Create the initial fracmesh geometry by copying the input DerivedMesh
as a single shard.

Removed the BMesh reference from FracMesh.
With the new Mesh data types this will no longer be needed.

Use plain Mesh types (MVert, MPoly, MLoop) in shards instead of copying
BMesh element arrays.

This will make Shard work more like a mini-Mesh struct and allow storing
of mesh data in blend files.

Larger construction methods currently disabled, TODO.

Partial revert of previous commit: DerivedMesh should be released by
setting needsFree and then calling dm->release.

fixed majority of memory leaks (of fracture)

Moved the render_mesh DerivedMesh from FracMesh into the
FractureModifierData.

This keeps the derived mesh inside the modifier system and is more
in line with common principles in other modifiers.

Also did lots of style cleanup.

basic foundation for fracture modifier / operator laid; todo fix memoryleaks...

modifier initial setup

initial commit for new fracture system

Let the smoke cache work on a smoke domain rather than the smoke
modifier data. This is a bit cleaner, it avoids referring to modifiers
which are a separate concern and will probably be replaced as the
container for smoke data at some point anyway.

Remaining RNA type mappings to pointcache readers/writers.