
* Optimized RNA property lookups and path resolving, still can be much better, but now the 1000 IPO example on bf-taskforce25 runs at reasonable speed. * Also an optimization in the depsgraph when dealing with many objects, this was actually also a bottleneck here.