feat: expose Signature<T>

[leruaa]

Motivation

Needed for alloy-rs/alloy#1077 and paradigmxyz/reth#9593

Solution

PR Checklist

• Added Tests
• Added Documentation
• Breaking changes

[leruaa]

While attempting to apply this refactoring to alloy itself, I start to understand why the k256 feature is not additive. That to avoid to impl traits like SignableTransaction for multiple Signature types leading to the "multiple impls ... found" error.

I'm trying to find a workaround.

[mattsse]

@leruaa perhaps we can introduce a helper trait in alloy for the required rsv function and just that for the alloy encode signature fns instead?

[mattsse]

this would introduce a ton of friction

I agree the memoized type is a bit horribly, but we can probably solve our problems alloy internally by introducing a signature encodeable trait there so we can implement it for reth's signature type as well.
can deal with the generic signature snafu later

[mattsse]

but perhaps we should consider removing the generic entirely, I find this very useless because this is effectively just duplicating the rsv again

core/crates/primitives/src/signature/sig.rs

Lines 301 to 305 in adcf7ad

	/// Returns the inner ECDSA signature.
	#[inline]
	pub const fn inner(&self) -> &S {
	&self.inner
	}

[leruaa]

@mattsse fyi and for posterity, the branch in alloy applying this refactoring in alloy-core: alloy-rs/alloy@main...leruaa:alloy:signature_refactoring