Blender Git Commits

More thoughts on AtomicOperationNodes...

Finish off group merging code

First pass at some node copy logic - ID and Group nodes only so far

Added TODO's for group merge - forgot about freeing old nodes!

Experimental idea - "subgraph" node type to deal with duplis needing isolation

Expose "raw" node_create() API method for greater flexibiltiy where required

Skeleton of node copying operation

* Added an API method for copying a node (and its little subgraph)
* TODO: The handling of in/out links needs to be resolved...

Fleshing out algorithm for merging groups

Added default handling for relationships on copied nodes

For now, resolution is to clear lists of links when doing base-node copy
(as we have no/little knowledge of what to do with regards to how this
links in with rest of tree).

However, callers of this node should know a bit more
(i.e. when copying node's subtree, at least level n-1 knows
how to solve the links between nodes at level n, though we still have
problems with links which stray to other nodes outside this one...

Group merging - Adding in code to handle node-hash redirections

Improve comment for nti->match_outer() callback

Code reshuffling for core/general node management routines

Abstract out standard part of IDDepsNode and GroupDepsNode into a separate type

This makes it easier to tell whether they've been defined correctly,
and/or gives us a legitimate type (OuterIdDepsNodeTemplate) to use when
dealing with a node which we know is either one (and the exact difference
doesn't matter).

More code reshuffling - moving around prototypes (and argument definition lists)

DEG_group_cyclic_nodes() -> DEG_group_cyclic_node_pair()

This renaming change makes it clearer what exactly the function does.
That is, instead of being the "global" entrypoint for all cyclic node
grouping operations, this only performs grouping for a single pair of
cyclic nodes...

Add slot in typeinfo for callback to be used when removing a node from graph

The main purpose of this is so that we can replace ID nodes with ID Groups
by removing the old ID nodes and merging their contents into a new group
node.

As a result, shuffled about the order of callbacks a little in typeinfo

Start of logic for node types

Finish logic for finding matching data nodes

Typeinfo defines for node types

DEG_get_typeinfo() -> DEG_get_node_typeinfo()