You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Originally reported at pmndrs/gltfjsx#277, running optimize on the attached models causes the window panels to be rotated 90º and out of position. Disabling instancing with --no-instance prevents the issue, which suggests that instance() may have a bug.
Likely unrelated to the transform issue, but also want to note that in the transformed glb, at least one node (Window.021.0) has no mesh but has a EXT_mesh_gpu_instancing extension with some number of instances. I flagged this in KhronosGroup/glTF#2404 but that PR did not reach a resolution so I still don't know if this is supposed to be valid or not. At the very least it seems non-optimal wrt output file size.
Thanks @zeux! It was indeed related. The original file already used EXT_mesh_gpu_instancing, and the implementation of the instance() function was not prepared for that, detaching the existing meshes but leaving EXT_mesh_gpu_instancing extensions behind. Will be fixed by #1507.
Describe the bug
Originally reported at pmndrs/gltfjsx#277, running
optimize
on the attached models causes the window panels to be rotated 90º and out of position. Disabling instancing with--no-instance
prevents the issue, which suggests that instance() may have a bug.https://github.com/user-attachments/files/17029993/venice-transformed.zip
To Reproduce
Expected behavior
Visual consistency in the instanced model.
Versions:
The text was updated successfully, but these errors were encountered: