Epic — v0.4 production-hardening, phase P0: Query, Transaction & Catalog Correctness
The bugs a user hits in the first hour: wrong results on the most common read/write shapes, data loss on restart, and DDL that bricks the data directory. These are already filed individually (below); this epic groups them as the P0 workstream. Goal: the engine returns correct results for common read/write shapes and survives routine DDL.
Wrong results / stale reads
Data loss / boot failure (critical)
Multi-database / graph / DDL scoping
Success: read-your-writes/deletes holds on WHERE pk= and scans; GROUP BY and qualified predicates return correct rows; non-default databases are reachable; DROP is complete and never bricks boot; document_strict survives restart.
Tracked in .planning/v1-production-readiness/GITHUB-ISSUES.md.
Epic — v0.4 production-hardening, phase P0: Query, Transaction & Catalog Correctness
The bugs a user hits in the first hour: wrong results on the most common read/write shapes, data loss on restart, and DDL that bricks the data directory. These are already filed individually (below); this epic groups them as the P0 workstream. Goal: the engine returns correct results for common read/write shapes and survives routine DDL.
Wrong results / stale reads
Data loss / boot failure (critical)
Multi-database / graph / DDL scoping
Success: read-your-writes/deletes holds on
WHERE pk=and scans; GROUP BY and qualified predicates return correct rows; non-default databases are reachable; DROP is complete and never bricks boot;document_strictsurvives restart.Tracked in
.planning/v1-production-readiness/GITHUB-ISSUES.md.