generic type for shaderMaterial, exposing uniform props

[xinaesthete]

Why

I and at least one other person were having trouble with the type of shaderMaterial, as discussed in this StackOverflow question.

Actually, as of this writing I'm not really entirely sure what the relationship to useRef is, but this seems like the right approach for the returned type of shaderMaterial.

What

I've added a type parameter to shaderMaterial, inferred from the type of the uniforms argument, and intersected that with the returned material type. As such, the type of the returned value should reflect the fact that each of the uniforms is available as a prop of appropriate type.

Checklist

• Documentation updated (example)
• Storybook entry added (example)
• Ready to be merged

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
drei	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Aug 9, 2023 11:30am

[codesandbox-ci]

This pull request is automatically built and testable in CodeSandbox.

To see build info of the built libraries, click here or the icon next to each commit SHA.

Latest deployment of this branch, based on commit af564ee:

Sandbox	Source
nervous-waterfall-wjhryw	Configuration
Ground reflections and video textures	Configuration
arc-x-pmndrs-colors	Configuration

[xinaesthete]

So when I saw failing CI stuff before I couldn't see the actual errors and I wasn't sure if it was to do with actual problems in the actual code vs not running properly because of who I was or something.

I realised the thing that was failing was an existing use of shaderMaterial in MeshReflectionMaterial, where the actual original call to shaderMaterial had envMap: null, ..., resolution: THREE.Vector2 which then were incompatible with their use...

So I still think the change I made in the PR is basically a good one - I'd certainly prefer to have this type inferred, and for that to - but there could be user-code that would get broken in a similar way if it got through to a release.

I'm also not entirely sure that the way I fixed issues with existing code in my last commit is necessarily the best - still had some niggles with getting it .