Morph geometry more like NifSkope (bug #6416)

Alexei Kotovrequested to merge
Capostrophic/openmw:morphyslaw into master

Fixes #6416 (closed).

NifSkope seems to do things more or less the right way so this makes morph geometry imitate it closer.

Morph target zero is now used in a special way. Rather than being ignored altogether, its vertices are used as a baseline for morphing. The source geometry is still used to determine the original bounding box and the animated bounding box in case there are no morph targets.

This seems to fix the TR creature's head animation. This also corrects a small oversight of mine where negative morph weight would not be regarded as a valid morph weight in the condition for the recalculation of the bounding box.

Merge request reports