refactor: avoid overwriting blocks when clear-on-start is false

[whankinsiv]

Description

Refactors chain_store to stop overwriting blocks when clear-on-start is false. After the first full sync, subsequent syncs run 2-3x faster because persistence now resumes from last_stored_block + 1.

To prevent corrupt state when switching networks, this adds handle_first_block, which compares the first block received from cardano.block.available with the stored block and returns an error if the block bodies differ.

Related Issue(s)

Speeds up development and testing for anything that depends on chain_store. Implemented primarily to accelerate testing of the /addresses/{address}/utxos endpoint in #335.

How was this tested?

• Verified persistence resumes at last_stored_block + 1
• Observed the 2-3x speedup on repeated runs.

Checklist

• My code builds and passes local tests
• I added/updated tests for my changes, where applicable
• I updated documentation (if applicable)
• CI is green for this PR

Impact / Side effects

Major sync performance improvement (2-3x) after initial sync. Behavior unchanged when clear-on-start = true.

Reviewer notes / Areas to focus

All logic changes are fairly straight forward.