mirror of
https://github.com/zhayujie/chatgpt-on-wechat.git
synced 2026-10-25 17:36:29 +08:00
Rudy CelekliandClaude Opus 5
9755a67014
_update_is_stale coerced the stored worker_pid with int() and no guard, so a
status file carrying a non-numeric value raised out of the "an update is
already in progress" check in start_update:
ValueError: invalid literal for int() with base 10: 'abc'
The function exists so that a running status left behind by a dead worker
does not block every later update, which is exactly the situation where the
status file is least trustworthy. read_update_status above it already
tolerates a missing file, an OSError, invalid JSON and a non-dict payload, so
the module validated the container but not the fields it then coerced.
Fall through to the timestamp check when the pid will not parse. Reporting
the run stale instead would unblock an update that may still be running,
which is the opposite of what the guard is for: a corrupt pid is no evidence
the worker died. A fresh run therefore still blocks, and an old one is still
reported stale.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>