- Startup: quick-count current DB vs all corrupt DBs (SELECT COUNT only). Only runs blocking salvage if a corrupt DB has MORE documents than the current DB. When counts match (already salvaged), startup is fast. - Shows 'Repairing' (via kSalvaging phase) in 'yams daemon status' during blocking salvage so users see progress. - RepairService::stop(): add 5s timeout to activeRepairExecutions_ wait to prevent indefinite shutdown hang. - salvageFromAllCorruptDbs/salvageFromCorruptDb: accept SalvageProgressFn callback for progress streaming during copy operations. - countDocumentsInDb: WAL-locked corrupt DBs recovered via ReadWrite open + PRAGMA wal_checkpoint(TRUNCATE) before counting. - cleanOrphanedMetadata: salvage runs FIRST (Phase 1), then orphan scan (Phase 2). Progress wired through to CLI via RepairEvent. |
||
|---|---|---|
| .. | ||
| spdlog | ||
| yams | ||