feat(core)!: use newtype sequencer block hashes

[SuperFluffy]

Summary

Adds a newtype astria_core::sequencerblock::v1::block::Hash to represent Sequencer block hash (an array of 32 bytes).

The standard formatting of the newtype's std::format:Display implementation uses lower case hex, while its alternative selector uses standard base64. Note that the latter was introduced to provide a way to format the hash using the same conventions as pbjson for the protobuf wire types.

use astria_core::sequencerblock::v1::block;
let block_hash = block::Hash::new([42; 32]);
assert_eq!(
    "2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a2a",
    format!("{block_hash}"),
);
assert_eq!(
    "KioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKio=",
    format!("{block_hash:#}"),
};

Background

During work on the Astria Auctioneer it was discovered that representing Sequencer Block Hashes with a distinct type is desirable for two reasons:

1. to distinguish Sequencer block hashes from other objects represented by an array of 32 bytes ([u8; 32]) that show up in the Astria ecosystem, for example Rollup block hashes.
2. to have a unified std::format::Display implementation for all Astria block hashes using a hexadecimal representation.

Changes

• Introduce the astria_core::sequencerblock::v1::block::Hash newtype for [u8; 32] objects representing Astria block hashes, and replace all domain types with it.

Testing

Added doc tests for the Hash constructor and normal and alternative display impl.

Changelogs

Changelogs updated.

Breaking Changelist

• This is a breaking change in astria-core but not in any of the services (because the return value of public getters was changed).
• This is not network breaking.

[ethanoroshiba]

Couple nits but otherwise LGTM. I don't foresee these changes being breaking since the storage types remain the same, but it may be worthwhile to double check this by starting a sync with main, stopping midway and building with this PR, then continuing with the sync.

[ethanoroshiba]

Suggested change

	/// display formatting faithful to pbjson convetion use the alternative formatting selector,
	/// display formatting faithful to pbjson convention use the alternative formatting selector,

[ethanoroshiba]

Is there a reason this can't be done now?

Suggested change

	// TODO: use the block hash's Display impl for this
	state.serialize_field(FIELDS[1], &base64(&*self.0.block_hash))?;
	state.serialize_field(FIELDS[1], &self.0.block_hash.to_string())?;

[SuperFluffy]

Ha, no, thanks for pointing that out! I updated the lines without thinking about about them.