Blender Git Loki

Merge branch 'undo-experiments' into undo-experiments-swap-reread-datablocks

Conflicts:
source/blender/blenloader/intern/readfile.c

Merge branch 'master' into new-object-types

Cleanup: Interpolate operator

Replace loops and remove old unused code

Merge branch 'readfile_factorize_common_id_processing' into undo-experiments

Merge branch 'master' into readfile_factorize_common_id_processing

store default value per CPPType

Readfile: refactor/factorize more handling of common ID data.

- Move handling of private ID data (nodetree and master_collection)
under generic ID code. This shortens code a bit, but mostly avoids having
to modify all type-specific callback functions if/when we have to add
generic processing to IDs there.

- Seriously factorize `expand_xxx` area, in the same way we were already
doing in `direct_link_xxx` and `lib_link_xxx` areas.

Differential Revision: https://developer.blender.org/D6764

NodeTree: Add access to the address of an ID's nodetree pointer.

Zero initialize before default construction to make it more deterministic

Merge branch 'greasepencil-object' into greasepencil-refactor

Merge branch 'master' into greasepencil-object

Merge branch 'master' into undo-experiments

Merge branch 'greasepencil-object' into greasepencil-refactor

GPencil: Add missing flag after merge

The vertex paint mode was not in master.

Merge branch 'master' into greasepencil-object

GPencil: Initial implementation of the layer mask relationship

Now, there is a new parameter to define the mask used by layer.

Still pending to remove old `use_mask` property in C and python (not removed yet to keep running the masking UI).

Volume: initial support for frame sequences

There is frame duration/start/offset just like images. A difference is
that there are more playback modes: clip, extend, repeat, ping-pong.

Still missing is auto detect of frame sequences, the duration has to be
manually entered now.

Volumes: debug logging for grid file load/unload

Volumes: show active grid in workbench

Better handling of usages of usercount increment in readfile liblink.

Some places could/should still need proper ID refcounting in liblink:
* UI data (spaces) - although this should be strictly 'user one' only...
* Deprectaed data/pointers which are not handled by libquery code.

Not sure whether IPO pointers really need usercount actually...