You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 68866c7
Browse filesBrowse the repository at this point in the historyBrowse files
Add stop-wm command to shut down the shared workspace server
The CLI could start a shared workspace server (start-wm-server
--detach --keep-alive) and recover it (restart-wm and friends), but had
no counterpart to stop it. A keep-alive server outlives its clients by
design, so the only way to end one was to kill the process by hand.
- Add `stop-wm` alongside the other recovery commands. Like them it
requires --shared-server: stopping a private per-command server would
report success while leaving the workspace an editor or agent uses
untouched.
- Refuse with RecoveryFailed when no server is running, naming how to
start one, rather than reporting a stop that did nothing.
- Send server/shutdown through the new ApiClient.shutdown and tolerate
the server closing the socket before the response is read, as
replace_running_server already does; the wait that follows verifies
the stop.
- Add wm_lifecycle.wait_until_stopped, which polls running_port in a
thread (its synchronous probe blocks for the full timeout on a
filtered port) and fails with a timeout naming the port.
- Document the command in docs/cli.md and update "all four" to "all
five".
0 commit comments