Fix intersecting l2 commitment blocks

[yaziciahmet]

Description

The issue was caused by the fact that when the sequencer is restarted, the sync of ledger db from postgres was problematic. This PR fixes the compare_commitments_from_db method.

Linked Issues

• Fixes Commitments are created on intersecting L2 ranges #842

Testing

Existing test sequencer_crash_and_replace_full_node works now.

[rakanalh]

Since compare_commitments_from_db isn't really just comparing anymore, i suggest calling it, adjust_commitments_from_db or moving the set logic outside of the compare_ method.

[codecov]

Codecov Report

Attention: Patch coverage is 93.75000% with 1 line in your changes missing coverage. Please review.

Project coverage is 78.3%. Comparing base (f43cae1) to head (41df0da).
Report is 5 commits behind head on nightly.

Additional details and impacted files

Files	Coverage Δ
crates/sequencer/src/sequencer.rs	91.1% <93.7%> (+<0.1%)	⬆️

... and 11 files with indirect coverage changes

[rakanalh]

Question: We re-initialize the sequencer ledger db on startup from postgres... i believe the proper bug fix would be for the sequencer ledger DB to not be out of sync to begin with. Why does postgres have information ledger db doesn't have?

[yaziciahmet]

@rakanalh what would you say to naming it sync_commitments_from_db?

[rakanalh]

@rakanalh what would you say to naming it sync_commitments_from_db?

Yea that works i guess.

[yaziciahmet]

Question: We re-initialize the sequencer ledger db on startup from postgres... i believe the proper bug fix would be for the sequencer ledger DB to not be out of sync to begin with. Why does postgres have information ledger db doesn't have?

If you restart a sequencer, there is no problem. The problem is when full node is upgraded to be a sequencer. Full node does not track commitment related stuff in its LedgerDB. That is why this type of sync is necessary.

[rakanalh]

Makes sense. Approving.

Please wait for @eyusufatik to review as well before merging.