feat: add AsWebGpuErrorType and ErrorType APIs

[ErichDonGubler]

Connections

• Resolves Firefox bug 1840926.

Description
Describe what problem this is solving, and how it's solved.

TODO:

• Make Deno consume this.
• Check that Firefox can consume this.
• Inform Servo when this merges.

Testing
Explain how this change is tested.

Checklist

• Run cargo fmt.
• Run taplo format.
• Run cargo clippy. If applicable, add:
  • --target wasm32-unknown-unknown
  • --target wasm32-unknown-emscripten
• Run cargo xtask test to run tests.
• Add change to CHANGELOG.md. See simple instructions inside file.

[sagudev]

This will simplify error handling in servo too, currently we hack it: https://github.com/servo/servo/blob/ee63174d6ff0b3b7d9b255fc47c72a82ae63bc09/components/webgpu/gpu_error.rs#L76

That reminds me this should also be used for classification in wgpu:

wgpu/wgpu/src/backend/wgpu_core.rs

Lines 267 to 300 in c110bf2

	fn handle_error_inner(
	&self,
	sink_mutex: &Mutex<ErrorSinkRaw>,
	source: ContextErrorSource,
	label: Label<'_>,
	fn_ident: &'static str,
	) {
	let source_error: ErrorSource = Box::new(wgc::error::ContextError {
	fn_ident,
	source,
	label: label.unwrap_or_default().to_string(),
	});
	let mut sink = sink_mutex.lock();
	let mut source_opt: Option<&(dyn Error + 'static)> = Some(&*source_error);
	let error = loop {
	if let Some(source) = source_opt {
	if let Some(wgc::device::DeviceError::OutOfMemory) =
	source.downcast_ref::<wgc::device::DeviceError>()
	{
	break crate::Error::OutOfMemory {
	source: source_error,
	};
	}
	source_opt = source.source();
	} else {
	// Otherwise, it is a validation error
	break crate::Error::Validation {
	description: self.format_error(&*source_error),
	source: source_error,
	};
	}
	};
	sink.handle_error(error);
	}

[ErichDonGubler]

thought: This should probably go into wgpu_types? Not sure.

[ErichDonGubler]

thought: I'm minorly concerned here about the Lost variant of this type. This seems to have meaningful handling in Firefox. Perhaps this is a variant we need to add to ErrorType?

[sagudev]

In servo we have this: https://github.com/servo/servo/blob/ba061ec2b0ef7124a5e64ec11a406cbc45cac02f/components/webgpu/wgpu_thread.rs#L389, I think instead of adding new variant, as_webgpu_error_type could return Option when wgpu error is not considered as error.

EDIT: Hm, this sounds like ErrorType::Internal.

[ErichDonGubler]

@sagudev: Internal errors are in the same error hierarchy as OOM and validation errors. Device loss, however, is usually reported through. I need to investigate whether reporting device loss as something distinct from an internal error is still necessary, if that's properly hooked up.

CC @teoxoy and @jimblandy for Firefox thinking.

[teoxoy]

Since we now handle driver induced device loss (#6229) properly we shouldn't need to do anything special with the lost variant anymore. It can be ignored.

[teoxoy]

We still need an error type for it I guess but that can also be ignored in downstream code.