Persist wallet sync progress at checkpoints

[3esmit]

🎯 Purpose

WalletCore::sync_to_block currently persists the complete wallet storage after
every processed block. A wallet catching up over a large range therefore
performs one full local storage write per block, creating unnecessary disk I/O
and serialization work for every consumer of the wallet core.

Fixes #631

⚙️ Approach

• Add a local checkpoint counter to block synchronization.
• Persist after every 100 successfully processed blocks.
• Persist any remaining tail before a successful sync returns.
• Keep account state and the last_synced_block watermark in the existing
  single persistence operation. If sync is interrupted between checkpoints,
  reopening resumes from the last durable checkpoint and replays only the
  unpersisted tail.
• Leave explicit lifecycle saves unchanged.

No wallet storage format, protocol behavior, or FFI/API surface changes.

🧪 How to Test

cargo +nightly fmt --all -- --check
cargo +1.94.0 test -p wallet --lib
RISC0_SKIP_BUILD=1 cargo +1.94.0 clippy -p wallet --lib --all-features -- -D warnings

All passed. The wallet unit suite includes coverage for full checkpoint chunks
and a final partial tail.

🔗 Dependencies

None.

🔜 Future Work

None.

📋 PR Completion Checklist

• Complete PR description
• Implement the core functionality
• Add/update tests
• Add/update documentation and inline comments

[Copilot]

Pull request overview

Reduces wallet sync disk I/O by persisting wallet storage at fixed “checkpoint” intervals during WalletCore::sync_to_block, while still ensuring any remaining tail progress is durably saved before returning successfully.

Changes:

• Introduces a small checkpoint counter to trigger persistence every 100 processed blocks.
• Updates sync_to_block to persist only at checkpoints and once at the end if there’s a remaining tail.
• Adds a unit test covering full checkpoint chunks and a final partial tail.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.