Implement Clone for ShaderModuleDescriptor

[daxpedda]

Checklist

• Run cargo clippy.
• Run RUSTFLAGS=--cfg=web_sys_unstable_apis cargo clippy --target wasm32-unknown-unknown if applicable.
• Add change to CHANGELOG.md. See simple instructions inside file.

Connections
Depends on gfx-rs/naga#2013.

Description
I am playing around with pre-compiled shaders and encountered an issue with storing them in statics, ShaderSource::Naga needs to own it, so cloning solves the problem.

Testing
None.

[daxpedda]

Now that gfx-rs/naga#2013 was merged, I updated the relevant parts to depend on it.

To explain this PR a bit further, personally it doesn't make sense to me without #2903. This PR can further refine the process by not having the user build ShaderModuleDescriptor from scratch every time, a very small quality-of-life improvement.

With this PR:

const SHADER: &'static ShaderModuleDescriptor = ...;

// The module itself is not cloned here, because it uses `Cow::Borrowed(..)`.
device.create_shader_module(SHADER.clone());

Without this PR:

const MODULE: &'static Module = ...;

// Now you have to write it out, even though it could be part of the const.
device.create_shader_module(ShaderModuleDescriptor {
    label: None,
    source: ShaderSource::Naga(Cow::Borrowed(MODULE)),
});

I must say after writing out this example and already locally using #2903 successfully, I don't think the quality-of-life improvement this PR offers is good. Calling clone requires the user to actually understand that nothing is cloned in a case like this, which to me adds unnecessary mental load when reading through code.

Maybe somebody else can see the value in this, otherwise feel free to close.

[cwfitzgerald]

Closing in favor of #2903

[daxpedda]

Just for clarification, as I said, I don't have a stake in this anymore: #2903 doesn't replace this. This PR (#2902) adds more quality-of-life to #2903.

[daxpedda]

@cwfitzgerald just wanted to make sure you take a second look at this after #2890 was merged. I can rebase this PR if it is still desired.

[cwfitzgerald]

Yes please rebase, then we're g2g

[daxpedda]

@cwfitzgerald done, you probably have to reopen for CI to run.

[cwfitzgerald]

Did you push? Force pushes should show up in the event log (and trigger CI)

[daxpedda]

Yeah I pushed, I think nothing shows up because the PR is closed. I can open another one?

[cwfitzgerald]

Huh, yeah go ahead, I can't re-open it either