Comparing changes

A spawn system bolting modules onto a prop at runtime rests on a chain of matrix identities that all fail quietly: a skipped matrix_parent_inverse, a basis built with the cross product the wrong way round, a transform frozen on the parent after the sockets were placed. Nothing in the gallery pinned that chain down — prop-origin-transform covers pivots and parent-inverse-orrery covers parent-inverse under animation, but neither asserts that a module dropped on a named socket with an identity local transform lands where the artist put it. The socket orientation check is deliberately non-circular: the mount pad mesh is swung onto its declared normal by a quaternion, while the socket basis is built by explicit Gram-Schmidt, and the check compares the socket's +Z against a Newell normal recomputed from raw vertex coordinates. Two independent derivations of the same declared value, so a basis-construction bug cannot hide behind the spec that produced it. Two hazards surfaced while authoring and are asserted rather than merely described, so they cannot change under us silently: applying a transform on an Empty root has no object data to bake into, so Blender pushes it down into every child (all 7 parent-inverses cleared, root scale handed to each child's local matrix) while preserving world matrices; and a child left selected during that apply gets the transform applied twice, drifting sockets 2.335 m. Signed-off-by: fOuttaMyPaint <tmhospitalitystrategies@gmail.com> Signed-off-by: fOuttaMyPaint <TMhospitalitystrategies@gmail.com>

Navigation Menu

Comparing changes

Open a pull request

Commits on Jul 24, 2026

This comparison is taking too long to generate.

Uh oh!