Mesh: optimize object mode face tessellation

- Multi-thread BKE_mesh_recalc_looptri.

- Add BKE_mesh_recalc_looptri_with_normals,
this skips having to calculate normals for ngons.

Exact performance depends on number of faces, size of ngons and
available CPU cores.

For high poly meshes the isolated improvement to BKE_mesh_recalc_looptri
in my tests was between 6.7x .. 25.0x, with the largest gains seen in
meshes containing ngons with many sides.

The overall speedup for high poly meshes containing quads and triangles
is only ~20% although ngon heavy meshes can be much faster.

Mesh: minor optimization to concave quad check for tessellation

Use the face normal (when available) for a faster concave quad test.

Cleanup: clang-format

Generate unique names for output attributes.

bli: generational_arena: insert_with() and try_insert_with()

bli: generational_arena: use better constructor and better if-else

Physics: Added basic visualisation for collisions, plus some cleanups to the Forces code

Merge branch 'master' into soc-2021-vse-strip-thumbnails

Thumbnail Caching

The caching code is updated to allow for separate thumbnail storage. The size
is taken as 0.25*render size as of now. flag set up to control when to store
thumbs. Image scaling done in preprocess stage using already present code for
crop and transform.

UV : Pack islands to box area

Adds a new operator to the UV editor - Pack islands to area
Allows the users to pack selected UV islands to a specified
area in the UV editor

Refer T78398