Split shader compilation logic into new wgpu-shaders crate

[inner-daemons]

Connections
Closes #8842
Waiting for #9064

Description
Splits shader compilation into a wgpu-shaders crate so that naga can more easily be removed as a dependency, and the compilation logic can be reused for e.g. precompiling shaders.

Testing
Too early

Squash or Rebase?
Squash

Checklist

• Run cargo fmt.
• Run taplo format.
• Run cargo clippy --tests. If applicable, add:
  • --target wasm32-unknown-unknown
• Run cargo xtask test to run tests.
• If this contains user-facing changes, add a CHANGELOG.md entry.

[inner-daemons]

I'm gonna shout with excitement into the void here:

THE REFACTOR NOW WORKS FOR ALL BUT DX12 IN WGPU HAL!!!! ONLY TOOK LIKE 3K LINES

[inner-daemons]

Ok nvm, webassembly 1.0 is unhappy, but I'm counting it

[teoxoy]

To achieve the goal outlined in #8842 I don't see why there need to be any changes to the naga crate. Why is the new wgpu-shader-types crate needed? IIUC the new wgpu-shaders crate should contain the wgpu-hal-naga-glue code that's currently in wgpu-hal but that should be it.

[inner-daemons]

@teoxoy I think this has been answered but just to document that, wgpu-hal may need to use some types that naga uses without depending on naga. ShaderStage is the best example, they both use it, but if naga can be removed as a dependency then ShaderStage should live in a crate that can be depended on separately by both (wgpu-types can't be depended on by naga). It should probalby be renamed to naga-types however.