January 14, 2015, 13:43 (GMT) |
Depsgraph: Move pose evaluation callbacks from temporary stub to blenkernel Now all the callback functions used by dependency graph are moved to dedicated armature_update.c file in blenkernel. There are also some spline IK functions moved from armature.c to the new file which isn't strictly related on the original change but without any file in blenkernel using functions from the new armature_update there'll be a linking error. It also seems not so crazy idea to concentrate all the pose evaluation in a single file, so overall it's not so much bad. Just annoying stupidnes of compiler forced to do such a move now. |
January 14, 2015, 12:35 (GMT) |
Depsgraph: Use proper debug print for IK solver builder |
January 14, 2015, 12:30 (GMT) |
Depsgraph: Don't re-tag operations for update This way we're saving some time in both tagging (by skipping loop over component operations which are tagged already) and scene evaluation (because now we don't have redundant entries in the directly modified operations). This gives about 30% speedup on loading Koro rig file. |
January 14, 2015, 12:15 (GMT) |
Depsgraph: Move get entry/exit operations to node This way we've got a bit less template magic going on and make it possible to cleanly access entry/exit operation of nodes outside of the builders. |
January 14, 2015, 11:28 (GMT) |
Merge branch 'master' into depsgraph_refactor |
January 13, 2015, 01:47 (GMT) |
Depsgraph: Fixes for drivers with bone targets defined using RNA Paths Now, these drivers don't get blocked waiting for the bone evals to finish. This makes them function better (specifically, allowing them to have some of the same benefits as they did in the old depsgraph) |
January 13, 2015, 01:10 (GMT) |
Depsgraph: Stub in a few more relations to ob transforms to go along with geometry when doing parenting |
January 13, 2015, 00:03 (GMT) |
Depsgraph: Bone constraints depending on the armature transforms were not working |
January 12, 2015, 23:50 (GMT) |
Code cleanup: Remove redundant comment |
January 12, 2015, 23:50 (GMT) |
Depsgraph: Code cleanup for single operation components If a component has just a single operation and no declared entry/exit operation, get_entry/exit_operation() now automatically just uses the operation that does exist. |
January 12, 2015, 23:39 (GMT) |
Depsgraph: Add support for a few more common datatypes to pointer_to_component_node_criteria() This is used by the RNAPathKey lookups, so supporting more types of data here should ensure more node lookups to work successfully. |
January 12, 2015, 12:59 (GMT) |
Depsgraph: Fix for using bones as targets for hook modifier The hook modifier now depends on the bone component of the relevant bone instead of the entire pose result. This makes it possible to interleave relationships where different bones in the same armature depend on the hooked geometry. |
January 12, 2015, 12:39 (GMT) |
Depsgraph: Fill in missing depsgraphUpdate callback for new modifier |
January 12, 2015, 12:15 (GMT) |
Merge branch 'master' into depsgraph_refactor |
January 12, 2015, 10:42 (GMT) |
Merge branch 'master' into depsgraph_refactor |
January 9, 2015, 14:01 (GMT) |
Depsgraph: Attemt to fix compilation error with mingw |
January 9, 2015, 13:48 (GMT) |
Depsgraph: Jotting out notes for how the obdata/geometry evaluation should all fit together |
January 9, 2015, 13:48 (GMT) |
Depsgraph: Remove invalid code + cleanup comments |
January 9, 2015, 13:48 (GMT) |
Depsgraph: Stub back reference to geometry results for lattice deform parenting type Strictly speaking, we really shouldn't still have any actual "virtual modifiers", so this category should really be obsolete. But, adding it back in for now in an attempt to figure out what's going wrong with koro's lattices (some points stick). |
January 9, 2015, 13:48 (GMT) |
Depsgraph: Shrinkwrap constraint needs the target object's transform as well as its geometry |
|