The steering actuator was filling its m_obstacle member when it was created (i.e., conversion time), which meant it had the wrong pointer after the actuator was replicated. Now m_obstacle is reassigned when the actuator is replicated.
- updating normals in py/api's mesh.transform() wasn't working and gave annoying print, disable this, script authors can call calc_normals explicitly if they need.
Note: Supporting obstacles which can be enabled/disabled as animated propoerty is not likely to happen. So I marked this as "Won't fix"/TODO.
I also reverted last commit on this bug because it didn't work and disabled the property from UI to avoid confusion.
fix [#30535] Shortest Path Select not working well in vertex mode.
regression from after bmesh merge, Ctrl+Right mouse for selecting shortest path is meant for edge mode only.
This affected legacy solver. Note that even if I don't expect any, this fix *may* break some other IK setups, so animators, please check your IK rigs using legacy solver. ;) Easy to revert if needed, anyway.
Note that iTaSC solver is also buggy with some scaled parent, but here IK'ed bones just can’t get scaled in any way (either from parent or from there own transformation)... Will try to find what's happenning, though the code is even harder than legacy solver to follow :/
from user freedback it seems there are 2 use cases, both valid.
* Select geometry and cut the selection in half (as 2.4x worked)
* Point-to-point define the faces to cut, dont cut through everything (only cut what you see).
With the second, since you are already selecting the edges to cut and snapping to them. only cutting the selecting is limiting/annoying.
Modifying these options while the knife tool runs doesn't work well, so expose under 2 keys, K, Shift-K.
- When the rip extends into a fan, pick the opposite edge in the fan (rather then 2 along)
- When stepping over the fan to find the rip edge, walk in the direction closest to the mouse (generally works nicer)
That whole "element re-arrange" area could use some love, though (e.g. make a single vert operator with more options, as faces; and make faces work in edit mode, probably no more reasons to switch to object mode?). Post-release TODO, anyway.
Also spotted a glitch in BLI_rand code, imho, which makes first element of an array not guarantied to be shuffled by BLI_array_randomize()... No correction though, as this might/would affect other parts of the code!