fix: let go of an adopted tab whose tmux session ended (release review)

When a foreign session ended, remain-on-exit kept Codeman's own tmux client
on the wrapper's dead pane, so the tab kept its pid and busy, input went
nowhere, both attach routes refused it, and a new session of that name only
led back to the dead tab. paneExit stays out of it by design.

The reconnect watcher's tick now detaches a client from a dead wrapper. That
is an ordinary PTY exit with status 0, so the pid goes null and selecting the
tab runs the attach again: it joins the target if it is back and otherwise
prints the existing guard text. Checked on an isolated instance and in a
headless browser: boot recovery, target gone, target back, view detached, tab
closed, and no restart loop while the tab sits unselected.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Codeman maintainer committed 2026-11-06 07:44:22 +01:00
1 parent 82593a4dd6
commit e7c68f1340
3 files changed
+125 -1

No files matched your search

+1 -1
View File
@@ -111,7 +111,7 @@ Input arriving during a wake is buffered: a chunk over 4 KB is dropped whole, ne
**No workspace.** ⚠️ An adopted session keeps its connection on `adopt.remote`/`adopt.docker`, so `session.remote` and `session.docker` are undefined, and its `workingDir` is the neutral `/tmp`. Code that branches on those two reads an in-container or over-ssh adoption as a local session working in `/tmp`. So anything that treats `workingDir` as the session's workspace must also test `session.isAdopted` (`adopt` on a pure view): uploads, the upload GC, git status, git fetch and its scheduler, the image watcher, Read My Mind's git signals, the Discord folder, `last-response`, guarded input and Ralph's `@fix_plan.md` and `@ralph_prompt.md` routes all do. The file surfaces (Files panel, File Viewer and its edit mode) still treat `/tmp` as the workspace, which is harmless: it is world-writable anyway, and adoption is admin-only in multi-user mode. A wrapper reports `statusLineExporter: 'remote'`, which is also what keeps keep-warm off it.
**Lifecycle.** `paneExitApplies` is false for a wrapper, so an attach that ended is never `paneExit` and the sweep never closes the tab. Boot recovery re-attaches a dead wrapper, which prints that the session is gone and exits. Reboot restore refuses one, since all Codeman ever had was an attach command. ⚠️ A wrapper that cannot start fails the start, and selecting the tab tries it again (`_failAdoptedStart()`). It never takes the direct-PTY fallback, which would run a new local `claude --dangerously-skip-permissions` or a server-account shell in `/tmp` under the foreign session's name, and `runPrompt()` refuses it for the same reason.
**Lifecycle.** `paneExitApplies` is false for a wrapper, so an attach that ended is never `paneExit` and the sweep never closes the tab. ⚠️ `remain-on-exit` still keeps Codeman's own tmux client on the dead pane, where a pid and `busy` hid it and both attach routes refused it, so `detachEndedAdoptedAttaches()` detaches that client on the reconnect watcher's tick: the tab closes nothing, its pid goes null, and selecting it (or "Go to tab" once a session of that name is back) runs the guard again. Boot recovery re-attaches a dead wrapper, which prints that the session is gone and exits. Reboot restore refuses one, since all Codeman ever had was an attach command. ⚠️ A wrapper that cannot start fails the start, and selecting the tab tries it again (`_failAdoptedStart()`). It never takes the direct-PTY fallback, which would run a new local `claude --dangerously-skip-permissions` or a server-account shell in `/tmp` under the foreign session's name, and `runPrompt()` refuses it for the same reason.
### Docker cases
+55
View File
@@ -499,6 +499,19 @@ export function hasObservablePaneSession(sessions: Iterable<MuxSession>): boolea
return false;
}
/**
* Which adopted wrappers have a dead pane with a tmux client still attached, from rows of
* `#{pane_dead} #{session_attached} #{session_name}`. Exported for unit testing.
*/
export function endedAdoptedAttaches(rows: string, adopted: ReadonlySet<string>): string[] {
const ended: string[] = [];
for (const row of rows.split('\n')) {
const [dead, attached, muxName] = row.trim().split(' ');
if (dead === '1' && Number(attached) > 0 && adopted.has(muxName)) ended.push(muxName);
}
return ended;
}
/**
* Resolve a target pane id from `tmux list-panes -F '#{pane_id}:#{pane_active}'`.
* Prefers the active pane and falls back to the first valid pane.
@@ -3989,9 +4002,51 @@ export class TmuxManager extends EventEmitter implements TerminalMultiplexer {
} catch (err) {
console.error('[TmuxManager] Remote reconnect watcher error:', err);
}
void this.detachEndedAdoptedAttaches().catch((err) => {
console.error('[TmuxManager] Adopted attach watcher error:', err);
});
}, intervalMs);
}
/**
* Let go of a wrapper whose pane died (#510). That pane holds the attach to someone
* else's tmux session, so it dies when that session ends, the transport drops or the
* view gets detached, and `paneExit` rightly reports none of them. But `remain-on-exit`
* keeps Codeman's own tmux client on the dead pane, so the tab kept its pid and `busy`,
* input went nowhere and both attach routes refused it.
*
* Detaching that client is an ordinary PTY exit with status 0: the pid goes null, the tab
* stays open, and selecting it starts the wrapper again, which joins the target if it is
* back and otherwise prints that it is gone. Only Codeman's own socket is touched.
*/
async detachEndedAdoptedAttaches(): Promise<void> {
const adopted = new Set<string>();
for (const session of this.sessions.values()) {
if (session.adopt) adopted.add(session.muxName);
}
if (adopted.size === 0) return;
for (const muxName of endedAdoptedAttaches(await this.readAttachRows(), adopted)) {
await this.detachClients(muxName);
}
}
/** Rows for {@link endedAdoptedAttaches}. Empty when tmux did not answer. */
protected async readAttachRows(): Promise<string> {
if (IS_TEST_MODE) return '';
const format = '#{pane_dead} #{session_attached} #{session_name}';
const { stdout } = await execAsync(`${this.tmux()} list-panes -a -F '${format}' 2>/dev/null || true`, {
encoding: 'utf-8',
timeout: EXEC_TIMEOUT_MS,
});
return stdout;
}
/** Detach every tmux client from one of Codeman's own sessions. */
protected async detachClients(muxName: string): Promise<void> {
if (IS_TEST_MODE || !isValidMuxName(muxName)) return;
await execAsync(`${this.tmux()} detach-client -s "${muxName}"`, { timeout: EXEC_TIMEOUT_MS });
}
stopRemoteReconnectWatcher(): void {
if (this.remoteReconnectInterval) {
clearInterval(this.remoteReconnectInterval);
+69
View File
@@ -18,6 +18,7 @@ import {
parsePaneRows,
derivePaneExits,
hasObservablePaneSession,
endedAdoptedAttaches,
type PaneRow,
resolveActivePaneTarget,
} from '../src/tmux-manager.js';
@@ -1721,3 +1722,71 @@ describe('hasObservablePaneSession', () => {
// the rule is enforced; that pairing is pinned in session-pane-exit.test.ts,
// where a real Session can answer for itself.
});
describe('an adopted wrapper whose pane died', () => {
const adopted = new Set(['codeman-aaaa1111']);
class WatchedManager extends TmuxManager {
rows = '';
detached: string[] = [];
protected override async readAttachRows(): Promise<string> {
return this.rows;
}
protected override async detachClients(muxName: string): Promise<void> {
this.detached.push(muxName);
}
adopt(sessionId: string, adopt?: MuxSession['adopt']): void {
const session = {
sessionId,
muxName: `codeman-${sessionId}`,
pid: 1,
createdAt: 0,
workingDir: '/tmp',
mode: 'shell',
attached: false,
adopt,
};
(this as unknown as { sessions: Map<string, MuxSession> }).sessions.set(sessionId, session as MuxSession);
}
}
const adoption = {
location: 'local',
socketPath: '/tmp/s',
targetSession: 'work',
viewSession: 'codeman-view-aaaa1111',
} as MuxSession['adopt'];
it('reads as ended only with a dead pane and a client still on it', () => {
expect(endedAdoptedAttaches('1 1 codeman-aaaa1111\n', adopted)).toEqual(['codeman-aaaa1111']);
expect(endedAdoptedAttaches('1 2 codeman-aaaa1111', adopted)).toEqual(['codeman-aaaa1111']);
// Nothing to detach: a second try would only print `no current client` every tick.
expect(endedAdoptedAttaches('1 0 codeman-aaaa1111', adopted)).toEqual([]);
expect(endedAdoptedAttaches('0 1 codeman-aaaa1111', adopted)).toEqual([]);
expect(endedAdoptedAttaches('', adopted)).toEqual([]);
});
it("leaves every other session's dead pane alone, since that one is an agent's exit", () => {
expect(endedAdoptedAttaches('1 1 codeman-bbbb2222', adopted)).toEqual([]);
});
it('detaches the client from a dead wrapper on the tick', async () => {
const manager = new WatchedManager();
manager.adopt('aaaa1111', adoption);
manager.adopt('bbbb2222');
manager.rows = '1 1 codeman-aaaa1111\n1 1 codeman-bbbb2222\n';
await manager.detachEndedAdoptedAttaches();
expect(manager.detached).toEqual(['codeman-aaaa1111']);
});
it('does not read tmux while nothing is adopted', async () => {
const manager = new WatchedManager();
manager.adopt('bbbb2222');
const read = vi.spyOn(manager as unknown as { readAttachRows: () => Promise<string> }, 'readAttachRows');
await manager.detachEndedAdoptedAttaches();
expect(read).not.toHaveBeenCalled();
});
});