Fix the transformation of plane's normal

[baothientran]

Fixes #9109

This PR will convert the transformation matrix to a normal matrix before multiplying with the plane's normal. Most of the fix happens in glsl czm_transformPlane() function and Cesium.Plane.transform()

@lilleyse Can you please take a look at it?

[cesium-concierge]

Thanks for the pull request @baothientran!

• ✔️ Signed CLA found.
• ❔ CHANGES.md was not updated.
  • If this change updates the public API in any way, please add a bullet point to CHANGES.md.

Reviewers, don't forget to make sure that:

• Cesium Viewer works.
• Works in 2D/CV.
• Works (or fails gracefully) in IE11.

[baothientran]

It is weird that prettier-check on CI is failing some of the files that I don't modify

[lilleyse]

The normal matrix should probably be a Matrix3 everywhere. This keeps it consistent with czm_normal which is also a Matrix3.

[lilleyse]

@baothientran only one suggestion here (which is small but affects a lot of files). I tested non-uniform models and point clouds and it works well.

[IanLilleyT]

Since we are doing inverse transpose in a lot of different areas now, it could be useful to move that into the Matrix3 and Matrix4 files and add a couple unit tests, similar to what you did in PlaneSpec.js.

Also, I know @lilleyse suggested using Matrix3 but I feel like Matrix4 makes more sense. For clipping planes, we would no longer need to pass an additional u_normalClippingPlanesMatrix. Instead, we just pass u_clippingPlanesMatrix which has the inverse transpose built in. Then czm_transformPlane becomes more simple and we can do return transform * clippingPlane; (if my understanding of https://stackoverflow.com/questions/7685495/transforming-a-3d-plane-using-a-4x4-matrix is correct). This would also remove the need for calcNormalClippingPlanesMatrix functions, since the inverse transpose will happen as part of calcClippingPlanesMatrix. Unless I'm overlooking something I think this will make the code change in this PR a lot smaller.

Thoughts @baothientran @lilleyse ?

[baothientran]

@IanLilleyT it is indeed a lot simpler to do transform * clippingPlane with the transform = inverse(transpose(originalMatrix)). I will update it right away

[baothientran]

@IanLilleyT there is a little complication. It seems like gltf_clippingPlanesMatrix is also used to calculate IBL in this code. I'm not really sure about how IBL works. The PBR material behavior may be changed if I transpose and inverse the clipping plane matrix.

[baothientran]

@IanLilleyT I close this and open another PR #9135 instead. The approaches between two PR are quite difference, so it may be better to start a new one to prevent accidental changes