Direction for raycaster may broken on A-Frame 1.0

[dala00]

I have been specified direction of raycaster on a game. After upgrade to A-Frame1.0, that direction is no longer effective.

• A-Frame Version: 1.0.0
• Platform / Device: Oculus Quest
• Reproducible Code Snippet or URL: https://glitch.com/~aframe-raycaster-test
  (Working version by A-Frame 0.9.2 https://glitch.com/~aframe-raycaster-test-092)

[felipenmoura]

Here too.
This is critical for me as well!

First, A-frame simply stoped working on the latest version of chrome due to a navigator.xr.requestDevice problem, what forced me to update to 1.0.
Now it works on Chrome, but the raycaster is all completely out of sync with the coordinations! Specially for AR! (using aframe-ar 2.0.8).

Any fix for that?!

[dala00]

Oh, same flow at all

[felipenmoura]

What I have ends up like this:

<a-scene
        id="scene-el"
        inspector="https://cdn.jsdelivr.net/gh/aframevr/aframe-inspector@master/dist/aframe-inspector.min.js"
        arjs="debugUIEnabled: false; detectionMode: mono_and_matrix; matrixCodeType: 3x3;"
        vr-mode-ui="enabled: false"
        cursor="fuse: false; rayOrigin: mouse"
        raycaster="objects: [data-click-on] [data-clickable]"
        loading-screen="dotsColor: blue; backgroundColor: white;"
      >
  
    <a-entity data-click-on="true">
        <a-entity data-clickable="true" ... ></a-entity>
    </a-entity>
    <a-entity camera id="main-camera" cursor="fuse: false; rayOrigin: mouse"></a-entity>
</a-scene>

But these clickable elements are actually created dynamically and appended to the main entity (in case that might matter).

I had a project running (with clients even), and now it is suddenly useless with no interactions at all :/
Any way I can help fixing this quicker?
Thanks.

[felipenmoura]

I created a pseudo-polyfill that works, @dala00 .
The "official" polyfill didn't work either, then I realised AFrame only uses the requestDevice method to validate two properties, immersive and exclusive.

Try loading A-frame 0.9.2 right after these lines:

navigator.xr.requestDevice = navigator.xr.requestDevice || function () {
  return new Promise((resolve, reject) => {
    resolve({
      supportsSession: new Promise((resolve, reject) => {
        resolve({
          immersive: true,
          exclusive: true
        });
      })
    });
  });
};

Pros
It works!! And quickly solves both problems with Chrome (which was not loading due to the navigator.xr.requestDevice is not a function error) and raycaster (which was triggering clicks and touches in misplaced areas of the scene).

Cons
It might trigger some false positives for devices that do not have support to that.
I'm still looking on ways to validate only those two fields.

[klausw]

Are both of you using Oculus Quest? As far as I know its current WebXR implementation is missing WebXR Gamepads Module support, so you don't get button or axis states, that'll be added in the next Quest browser update. However, poses and WebXR "select" events should still work.

I tried https://aframe-raycaster-test.glitch.me on Daydream, and I do see a ray (pointed about 45 degrees down from the controller forward axis), and the controller model shows button presses, though I don't see any interactions happen as a result.

[dala00]

Oh, thank you for information! Yes, Oculus Quest.

Then I wait browser update and close this.

[klausw]

I think it may be too early to close this, I don't know if the known Quest issue is what's causing this problem. The missing gamepads module means that you can't read button values or axis values, but the controller pose and ray position should still be working. I'm not sure if A-Frame internally uses WebXR "select" events (which work), or if it reads gamepad button values (which doesn't currently work on Quest).

[dala00]

OK, I continue to check and wait other information if exists.

[klausw]

FYI, @Artyom17 just posted this half an hour ago: https://twitter.com/abolgar/status/1207469356461916161

Oculus Browser 7.1 has been released with XR Gamepad Module support and a fix of the bug affected three.js examples.

[dala00]

Oh!? Thanks, I try at night.