Expand description
State recovery and stale-state detection.
devflow recover reads the existing state file, determines if the
agent process is still running, and either reports status or
offers to clean up / restart.
Structs§
- Recovery
Status - Result of inspecting an existing workflow state.
Enums§
- Recover
Error - Errors produced by recover operations.
Constants§
- STALE_
THRESHOLD - Maximum age before a state is considered stale (>24h).
Functions§
- clean
- Clean up stale or abandoned workflow state.
- clean_
phase - Explicitly clean ONE phase, regardless of staleness — the operator’s escape hatch for a wedged-but-fresh run. Clears its state and cron record; warns (but proceeds) when the recorded agent still looks alive.
- format_
age - Format a unix-seconds timestamp’s age as a human-readable string
(“5m ago”). Public since 14c:
devflow statusreuses it for elapsed time and event recency. - inspect_
all - Inspect every active phase state, producing one recovery status per phase
(sorted by phase number). Errors with
RecoverError::NothingToRecoverwhen no phase has persisted state. - is_
stale_ state - Check whether a state is stale: >24h old with no running agent.