- Renamed
bNodeTreeDeclarationtobNodeTreeInterface - Changed panel assignment to a panel/parent pointer in sockets and panels respectively
That's already the case for meshes though, and means we don't constantly have to convert back and forth between the old and new versions when saving files for the foreseeable future, which is…
What do you see as the benefit of this, for the code? In the UI it makes some sense, to simplify aligning sockets, but it seems like keeping a clear separation between inputs and outputs…
- Generally makes sense that the panels know about the contained sockets and not the other way around. In practice we may need a backpointer, but that can just be a cache.
Yes, there would be…
Jacques defines it like so:
the *_for_write suffix on functions is only necessary if the function is doing extra work in this compared to the const version. (e.g. custom data layers may be…
IIRC we use the _for_write suffix to be very clear about write access
I'm fleshing out a new struct for node tree declarations that will include the current bNodeTree::inputs, outputs, and panels. The primary API can be in C++ since we have C++ RNA now for…