Noeri Huisman

That's unfortunate. I just tried both the `dmabuf` and `glReadPixel` patches on the Firefox tip (`r700545`) and both did still work for me (although the `dmabuf` one did crash after...

> If that helps @mrxz I can setup a cloud instance with an NVIDIA GPU then share access. Thanks for the offer @Utopiah, but the biggest challenge will be finding...

WIP branch available (non-functional): https://github.com/mrxz/webxr-linux/tree/chromium

@dmarcos This issue can be closed, see https://github.com/aframevr/aframe/issues/5311

Thanks to the awesome work of @rcelyte there's now a 'mostly functional' patch to get WebXR working with Chromium (#6). This has been merged into master, hence closing this.

> So, if I understand correctly, this behavior is specific to GLFW, right? If you would be using the SDL application instead (if it's possible for you to test that),...

Useful reference for how ImGui handles this issue: https://github.com/ocornut/imgui/issues/6034#issuecomment-1369620644 They've been inferring the modifiers from the key state since 2015 onwards. Basically ignoring the modifiers provided by GLFW altogether and...

Well, only suggesting -1 since that's already how it works ([docs](https://aframe.io/docs/1.4.0/components/renderer.html#properties)): https://github.com/aframevr/aframe/blob/17f06f6ec71ea914aa889a38b2a0935c0b7d4571/src/core/scene/a-scene.js#L783 Personally I think the limit can be removed entirely. Most people probably don't expect this cap and only...

> Instead of removing the behavior right away we flag it as needsRemove and then clean up at the end of the render loop That was actually the first approach...

Moving this responsibility to the entity is going to make things harder, I'm afraid. Currently the "bookkeeping" burden only happens when adding or removing behaviours, meaning that the scene's `tick`...