Files
routstrd/src/daemon/http
redshift 00c823519d wallet: shared recovery tracking, bounded waits, safe shutdown
- Share one wallet-lifetime tracker (family:id keys) across startup
  settlement, targeted recovery, explicit mint-quote recovery, and the
  pending-mint sweep, so a timed-out observation is never retried
  underneath.
- Bound targeted recovery per operation (15s) and per pass (60s); a
  timeout reports timedOut and keeps the underlying work tracked.
- Executing-send recovery holds coco's per-operation lock until the
  underlying drive actually settles, not until the caller times out.
- Dispose now rejects new/queued recovery, drains queued and in-flight
  work, and only then closes coco/database/ownership; a bounded caller
  wait reports incomplete shutdown instead of closing early.
- Fail-closed checks for coco private recovery internals before any
  local cleanup writes.
- Remove the periodic recheck; recovery stays startup- or
  operator-triggered.
2026-10-02 16:36:30 +08:00
..
2026-06-11 09:17:13 +08:00