refactor(lstar): fold triplicated staleness filter into the extractor#972
Merged
tcoratger merged 1 commit intoJun 12, 2026
Merged
Conversation
The head.slot > latest_finalized.slot staleness predicate was applied at three call sites that each pre-filtered a payload pool before handing it to the latest-vote extractor. Move the predicate inside the extractor so the cutoff lives in exactly one place, and pass each pool directly. Iteration order and results are unchanged: the surviving keys appear in the same insertion order whether filtered before the loop or skipped inside it, so the first-seen-wins LMD tie-break is preserved. The fork-choice test-store checker inspects raw pool content before pruning and must stay unfiltered, so it gets an inlined unfiltered latest-vote loop instead of routing through the now-filtered extractor. Verified byte-identical: regenerated the full tests/consensus/lstar/fork_choice suite before and after the change and diffed the fixtures with no differences, with the two-seed determinism check passing in both runs. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
The fork-choice staleness predicate
head.slot > latest_finalized.slotwas applied at three separate call sites, each pre-filtering a payload pool with an identical dict comprehension before handing it to the latest-vote extractor:This fold moves the predicate inside the extractor, so the cutoff exists in exactly one place. Each call site now passes its pool directly plus the finalized slot.
Why it is behavior-preserving
The extractor walks
aggregated_payloads.items()in dict insertion order. Pre-filtering with a comprehension preserves the insertion order of the surviving keys; skipping inside the loop visits the same keys in the same order andcontinues on the filtered-out ones. The set of(attestation_data, proofs)pairs processed, and their order, is identical. The "first vote seen wins on equal slots" LMD tie-break therefore resolves identically.Test-framework checker
The fork-choice test-store checker (
store_checks.py) deliberately inspects raw pool content before pruning, so it must remain unfiltered. Routing it through the now-filtered extractor changed its semantics (it dropped a validator whose head sits at the finalized slot, breakingtest_finalization_prunes_stale_aggregated_payloads). It now uses an inlined unfiltered latest-vote loop, preserving its original behavior with no dependency on the filtered extractor.Equivalence verification
tests/consensus/lstar/fork_choicesuite (110 cases) before and after the change into separate output directories.diff -rqof the two fixture trees reports no differences (byte-identical).--check-determinism) passes in both runs, confirming LMD ordering is unaffected.just checkpasses (ruff, format, ty, codespell, mdformat).🤖 Generated with Claude Code