WebGLCapabilities: add textureTypeReadable() and textureFormatReadable()

[sguimmara]

Fixed #28086

Description

WebGL only supports one texture type for reading: UnsignedByteType, and only one pixel format: RGBAFormat. Any other combination is optional and implementation dependent.

This PR exposes 2 conditional checks that previously ocurred in WebGLRenderer.readRenderTargetPixels():

• the texture type check
• the pixel format check

as two publicly accessible methods on WebGLCapabilities:

• textureTypeReadable()
• textureFormatReadable()

To be able to determine if the render target can be read, we can now use those two methods. If at least one of them returns false, we might have the opportunity to convert the texture to another format before calling readRenderTargetPixels().

This contribution is funded by Oslandia

[github-actions]

📦 Bundle size

Full ESM build, minified and gzipped.

Filesize dev	Filesize PR	Diff
672.7 kB (166.7 kB)	672.8 kB (166.7 kB)	+137 B

🌳 Bundle size after tree-shaking

Minimal build including a renderer, camera, empty scene, and dependencies.

Filesize dev	Filesize PR	Diff
452.4 kB (109.2 kB)	452.5 kB (109.3 kB)	+128 B

[sguimmara]

@Mugen87 is there any documentation to update somewhere ?

[Mugen87]

No, there isn't a doc page for WebGLCapabilities.

Apart from that, I'm curious to see if other devs prefer an API on WebGLRenderer level. We eventually need a similar API in context of WebGPURenderer (since it has an optional WebGL 2 backend). A method like getMaxAnisotropy() is placed in the Renderer class and calls the implementation from the respective backend implementation (meaning this.backend.getMaxAnisotropy()).

The PR in its current state is fine in context of WebGLRenderer. For WebGPURenderer the solution might be slightly different. Especially since I believe WebGPU hasn't the same readback restrictions like WebGL.