Solve struct ellipsoid caused the model to dark

[verybigzhouhai]

Solved the problem that struct ellipsoid caused the model to darken on Qualcomm platform

[cesium-concierge]

Thanks for the pull request @verybigzhouhai!

• ✔️ 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.
• ❔ Unit tests were not updated.
  • Make sure you've updated tests to reflect your changes, added tests for any new code, and ran the code coverage tool.

Reviewers, don't forget to make sure that:

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

[verybigzhouhai]

Fixed #7651

[verybigzhouhai]

@OmarShehata Hi，Now I am modifying the relevant code of the ellipsoid struct. Do you think it is acceptable?

[verybigzhouhai]

This is ready!

[OmarShehata]

Just to close the loop, this is the same fix as implemented in #7911 right?

[verybigzhouhai]

@OmarShehata yes, it's exactly the same.

[verybigzhouhai]

@OmarShehata Will you review this?

[OmarShehata]

I can confirm this fixes the issue at #6622. I believe this also fixes #7651 and #7403.

Testing on the physically based Sandcastle, in the hosted Sandcastle on Android:

In this branch:

@verybigzhouhai my main notes here are:

• I would avoid copying the value manually everywhere it's used in the code. Consider adding something in the shader constants that can be re-used like https://github.com/AnalyticalGraphicsInc/cesium/blob/master/Source/Shaders/Builtin/Constants/degreesPerRadian.glsl
• I think this can be removed if it's not being used anywhere https://github.com/AnalyticalGraphicsInc/cesium/blob/master/Source/Shaders/Builtin/Structs/ellipsoid.glsl

[verybigzhouhai]

@OmarShehata I am very happy that the function is normal in your testing.
I did use some constants to replace the structure at the beginning, but it caused a lot of unit tests to fail. I don't know how to avoid or need to modify a lot of unit test code, so I used the current modification method.
Even if we use repeatable constants, I don't think this is the best way to modify them, because the structure itself is hoping to make the program simpler, so I think the best way to do this is to dynamically before compiling the shader. Replace the structure, but this approach may not be that simple, I have raised a similar issue here, https://stackoverflow.com/questions/56479172/about-remove-replace-struct , if we can do this, could make the code Keep the original elegance.
What do you think about the above, if nothing, I will modify it again according to what you mentioned.

[OmarShehata]

@verybigzhouhai I'm curious about the unit tests that fail using that. Can you list a few of them and the errors you get?

[verybigzhouhai]

@OmarShehata I found some errors of my own, which will not lead to unit test failure now. Now I add two constants and delete files that do not need to be used. Updates have been submitted，you can check whether they meet the requirements.

[OmarShehata]

This looks great @verybigzhouhai ! Just a minor doc comment and I think this will be ready to merge.

[OmarShehata]

@verybigzhouhai one more thing: Can you update CHANGES.md to add a bullet point under "fixes"? It should say something like:

• Fixed a bug that caused 3D models to appear darker on Android devices. #7944

[verybigzhouhai]

Thank you very much for your guidance, @OmarShehata now it should be ready to be merged.

[OmarShehata]

Great - I'm confident merging this since it's largely just rewriting structs to use builtin constants. It works in Chrome, Firefox, Edge, and IE11. It works correctly in Chrome on Android now.

@lilleyse if you want to skim the changes later when you get a chance in case I missed any shader styling thing I wasn't aware of, that would be helpful!

[slchow]

If this is a first-time contribution, we'd like to share it. @verybigzhouhai do you have a Twitter handle?

[verybigzhouhai]

@slchow twitter handle @bigzhouhai https://mobile.twitter.com/bigzhouhai