Blender Git Commits

Fixed the particle attribute state data access functions. These have to
use an index as key rather than a particle ID, since the latter is not
available in a single attribute state without access to the parent
particle state.

Added accessor functions in particle state which allow extending the
particle state. These can be used later in python wrappers to support
more immediate collection-style access.

Unused RNA code for exposing particles directly in the state as a
collection. This doesn't work atm because for lookup it would need to
allocate a particle iterator, which can not be freed afterward. This
can probably be done nicer as a pure python wrapper instead of RNA.

Ensure the attribute states are synced to the attributes of the particle
system when adding/removing attribute descriptors. This is not a hard
requirement really, if an attribute has no state data that would still
work, but in most cases it should stay in sync.

Collection property for accessing the attribute states in a particle
state.

Replaced the attribute state array with a ListBase. Eventually this
should use a hash table for O(1) lookup, but for now a linked list is
the easiest implementation.

Copy the attribute descriptor to attribute states as well. The storage
overhead is negligible and this allows sanity type check assert as well
as correct RNA refining without access to the original attribute
descriptor.

New function for adding a particle based on particle id.

Removed leftover #undefs.

Added 2 functions for looking up particles by their ID in the psys.

Copy the particle state when copying an attribute.

Added a comment in DNA about the concept of state/descriptor separation
and possible future optimization.

Use a fixed "id" attribute, this will be required for all sorts of
particle functions and identify particles unambiguously.

Iterator access functions for particles.

New function for getting a particle iterator at a given index. This will
be used for binary search in the particle id attribute layer.

Implemented a simple iterator API for particle systems using a plain
index and the ID attribute for determining the state size. Have to see
how this plays out in terms of performance ...

Collection properties for actual data in particle buffers.

Removed unused data type mapping function in RNA, this is now directly
refined through the type of the attribute/state.

Removed deprecated DNA data inside #if 0 block, C preprocessor doesn't
work with makesdna.

Some renaming and DNA reorganization. Particle attributes now have a
descriptor as well as state data.