corebreaker / stratodb

98%

main: 98%

Pull #5

circleci

Address PR #5 review: clarify test intents for the coverage additions

Follow the reviewer's five minor points, all in test code — no production change. In table_key's ord_matches_the_encoded_byte_order, document that the final Some-vs-None index pair is separated by its tag (INDEX_DUP < INDEX_UNIQUE), not by the then_with entity arm, which never sees a mixed pair and returns Equal for two unique keys. In backend's corrupt_entries_are_rejected, use a distinct unpopulated key for the non-node data value so it is unambiguous the rejection is the "data entry is not a node" arm rather than the object-marker idempotent no-op. In inner's cached_indexes_reports_a_poisoned_mutex_on_either_lock, note that the second-lock case depends on cached_indexes releasing the get lock before load runs. In derive's rename_all block, add field-casings and variant-casings sub-section separators. In writer's exercise, re-insert the just-removed key to also drive the forwarded put path after a delete.

905 of 905 new or added lines in 26 files covered. (100.0%)

6976 of 7099 relevant lines covered (98.27%)

439380.88 hits per line