-
-
Notifications
You must be signed in to change notification settings - Fork 35.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Object3D: Respect matrixWorldAutoUpdate in matrix update methods #28533
Object3D: Respect matrixWorldAutoUpdate in matrix update methods #28533
Conversation
📦 Bundle sizeFull ESM build, minified and gzipped.
🌳 Bundle size after tree-shakingMinimal build including a renderer, camera, empty scene, and dependencies.
|
Do you mind rebasing the PR and #28534? Unfortunately, they are based on a state with broken CI. If you grab the latest commits from dev, we should easier see if something breaks. |
@CodyJasonBennett Thanks for filing this PR! |
Sorry to interfer in a closed PR but I have a hard time understanding the new behaviour. Now with the new condition on line 592 the matrixWorld is never updated whereas I use This is done because I need the camera matrix to be updated before rendering, I cannot wait for the WebGLRenderer to kick in. And I wanted to avoid updating the matrix twice on each frame. Alternative I considered : do it entirely manually this.camera.updateMatrix();
this.camera.matrixWorld.copy(this.camera.matrix); Thanks |
Yes, there is a difference between TBH, I'm not super happy with how So until there is a new world matrix update routine, you indeed have to manage the world matrix manually if you set |
Thanks for those informations. Still I think the "force" flag should be removed entirely has it does nothing but clear the "needsMatrixWorldUpdate" property without actually upadting the matrix. |
I also had a hard time figuring out why all my objects are messed up after a three.js upgrade. I narrowed it down to this PR, which came with r166. I also figured out that the world matrix cannot be updated manually anymore if the "matrixWorldAutoUpdate" is set to false. However, for performance reasons, that's how I did it in the past 2 years, I updated the local-, and worldmatrices manually whenever it was necessary. Now, with r166, I'm using this hack as a workaround:
|
Related issue: #21387
Description
Since #24028, whenever a parent object had its world matrices updated by three, it would update child matrices unconditionally by the local override of
force
(#27245, #27261). Additionally, if three reaches a child withmatrixWorldAutoUpdate = false
, its descendant children aren't processed despite their configuration (#27261 (comment)).I've since fixed both issues by moving the check for
matrixWorldAutoUpdate
when managing the parent world matrix and preserving the original functionality prior to #24028. Unit tests aren't properly isolated, which hid this bug prior, so I also reset hidden states in-between cases, which now catch these regressions.cc @DolphinIQ, @krispya