align filestorage with constant-size proof verification#373
align filestorage with constant-size proof verification#373dime-squanderer wants to merge 4 commits into
Conversation
# Conflicts: # core/Cargo.lock # core/indexer/Cargo.toml # core/indexer/src/bin/generate_por_fixtures.rs # core/indexer/src/database/queries.rs # core/indexer/src/database/sql/schema.sql # core/indexer/tests/contracts/file_storage_tests/proof_verification_e2e.rs # core/indexer/tests/database.rs # core/indexer/tests/fixtures/por_proof_fixtures.json # native-contracts/binaries/filestorage.wasm.br # native-contracts/binaries/registry.wasm.br # native-contracts/binaries/staking.wasm.br # native-contracts/binaries/token.wasm.br
|
bugbot run |
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit f0084f0. Configure here.
| shard_challenges.update_next_seq(|seq| seq.saturating_add(1)); | ||
| shard_challenges | ||
| .by_seq() | ||
| .set(&shard_seq, challenge_id.clone()); |
There was a problem hiding this comment.
Duplicated shard initialization logic across two functions
Low Severity
The ~30-line block that initializes ProverChallenges / ShardChallenges, assigns the next shard_seq, increments next_seq, and inserts into by_seq is copy-pasted between generate_challenges_for_block and create_challenge_for_agreement. A shared helper (e.g., taking the model and returning (shard_id, shard_seq)) would reduce maintenance burden and eliminate the risk of one copy diverging from the other during future edits.
Additional Locations (1)
Reviewed by Cursor Bugbot for commit f0084f0. Configure here.
2 participants
Note
High Risk
High risk because it changes on-chain filestorage contract storage/layout and
verify_proof/join/leave APIs, plus modifies the indexer DB schema and ledger reconstruction ordering; mistakes could break proof verification or require state migrations.Overview
Aligns proof verification with constant-size proofs by removing proof-derived challenge lists and instead verifying against a caller-specified
(shard_id, start_seq, end_seq)range of challenges deterministically indexed per prover/shard, with new view APIs to query shard sequences and challenge IDs.Updates filestorage participation semantics so
join_agreement/leave_agreementderivenode_idfromctx.signer().to_string(), and extendschallenge-datawithshard_id/shard_seqwhile indexing challenges during generation/creation.Makes indexer file-ledger reconstruction stable by adding a
ledger_indexcolumn (unique, ordered) tofile_metadata, persisting it on insert, and loading rows ordered byledger_index; updates WIT host bindings/tests accordingly and switcheskontor-cryptoto a git dependency/branch. Removes static POR proof fixtures and generates proofs inline in E2E tests due to prover-id changes.Reviewed by Cursor Bugbot for commit f0084f0. Bugbot is set up for automated code reviews on this repo. Configure here.