Closed Canleskis closed 2 years ago
Hi, Thank you very much for the PR! :pray:
I'll try to review it this week.
I have not taken the time to work out what problems this could cause, but I assume there needs to be some checks done in some places before assigning to the GlobalTransform.
As long as all existing tests pass, we can merge. If it does cause problems, we can address them separately (incl. adding relevant test/benchmark)
Of course, I am nevertheless interested to understand the potential problems and what needs to be done to avoid them. But I won't block this PR just because we don't know everything.
:tada: This issue has been resolved in version 4.0.0-alpha.4 :tada:
:tada: This issue has been resolved in version 4.0.0 :tada:
Bevy 0.8's GlobalTransform is now backed by an Affine3A, in this PR I just naively convert it back to a transform, rotation and scale when needed using the to_scale_rotation_translation() method. This could cause problems (?), as discussed here: https://github.com/dimforge/bevy_rapier/pull/219
I have not taken the time to work out what problems this could cause, but I assume there needs to be some checks done in some places before assigning to the GlobalTransform.