Floating point framebuffer test fails on Firefox

[OmarShehata]

This test fails in Firefox on Linux. Does not fail in Firefox on Windows, or in Chrome at all.

http://localhost:8080/Specs/SpecRunner.html?spec=Renderer%2FFramebuffer%20draws%20to%20a%20half%20floating-point%20color%20attachment

Renderer/Framebuffer draws to a half floating-point color attachment

Expected context to render 14336,15872,16384,15360, but rendered: 0,0,0,0

This fails in the same way in 1.67 as well: https://cesium.com/downloads/cesiumjs/releases/1.67/Specs/SpecRunner.html?spec=Renderer%2FFramebuffer%20draws%20to%20a%20half%20floating-point%20color%20attachment

Windows WebGL report:

Platform:   Win32
Browser User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/80.0.3987.132 Safari/537.36
Context Name:   webgl
GL Version: WebGL 1.0 (OpenGL ES 2.0 Chromium)
Shading Language Version:   WebGL GLSL ES 1.0 (OpenGL ES GLSL ES 1.0 Chromium)
Vendor: WebKit
Renderer:   WebKit WebGL
Unmasked Vendor:    Google Inc.
Unmasked Renderer:  ANGLE (NVIDIA GeForce GTX 1080 Ti Direct3D11 vs_5_0 ps_5_0)
Antialiasing:   Available
ANGLE:  Yes, D3D9
Major Performance Caveat:   No

Linux WebGL report:

Platform:   Linux x86_64
Browser User Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/79.0.3945.79 Safari/537.36
Context Name:   webgl
GL Version: WebGL 1.0 (OpenGL ES 2.0 Chromium)
Shading Language Version:   WebGL GLSL ES 1.0 (OpenGL ES GLSL ES 1.0 Chromium)
Vendor: WebKit
Renderer:   WebKit WebGL
Unmasked Vendor:    Intel Open Source Technology Center
Unmasked Renderer:  Mesa DRI Intel(R) HD Graphics 630 (Kaby Lake GT2)
Antialiasing:   Available
ANGLE:  No
Major Performance Caveat:   No

[likangning93]

I'm seeing this during the 1.77 release process too, Firefox on Linux, on both Intel and Nvidia.

[mramato]

With WebGL validation on, you actually get a better error:

RuntimeError: WebGL Error:  INVALID_OPERATION: readPixels(0, 0, 1, 1, 6408, 36193, 0,0,0,0);

So that makes is sound like we're doing something wrong on our side.

[ggetz]

Still seeing this after https://github.com/CesiumGS/cesium/issues/10887.