[teamleaderleo]

Leo, 2026-11-04: "i feel like the restarted after nightly update does too many layout shifts/flicker renderings and takes too long to reconnect acpmux or whatever". Data: manaflow-ai/cmuxterm-hq#6338 (comments 6597127823 and 6626332314), bench in #34697.

After an update relaunch on macOS the old acpmux daemon sometimes logs stopping: SIGTERM (a restart handoff) about 1.1 s after the app quit and about 5 s before the new app starts. Its agent host is lost with it, so the next daemon resumes the streaming turn on a fresh agent. Nothing in the log says who sent the signal. The app, the bench and acpmux itself have no code that sends SIGTERM to the daemon on that path. The daemon and its hosts each lead their own session, so a process-group kill can't explain it.

This PR makes the daemon say who sent it. A SA_SIGINFO action next to tokio's own records si_pid; both go through signal-hook-registry, which runs every action registered for the signal. The action does only an atomic store, so it is async-signal-safe. The stop line then reads stopping: SIGTERM from pid N (command), parent M (command) (a restart handoff). A fleet bench run at this head (relaunch-reconnect:warm+update:4) is collecting the sender for the fix that follows.

Test

• Red/green: cargo test -p acpmux --test daemon_lifecycle a_sigterm_stop_names_its_sender fails at the red commit (no sender on the stop line) and passes at the head.
• Unit: daemon::term_sender::tests (ps row parsing; this process is described with its command).

Observed / Not run

• See the run results below.

Performance

metric	before	after	delta	harness
work per SIGTERM	0	1 atomic store + 2 ps reads on the stop path	stop path only	code read; no hot path

Risk

Low: only the daemon's SIGTERM stop path changes, plus one extra signal action. If the action can't be registered, the daemon logs a warning and stops as before.

Changelog

none

Agent session: agent session

🤖 Generated with Claude Code

---

^{Open in [code]smith · Need help on this PR? Tag @codesmith-bot with what you need.}

---

Summary by cubic

Names the SIGTERM sender on the acpmux daemon's stop line so an unexplained SIGTERM after a NIGHTLY update relaunch on macOS can be traced.

The daemon now logs stopping: SIGTERM from pid N (command), parent M (command) instead of the plain stopping: SIGTERM (a restart handoff). A SA_SIGINFO action records the sender's pid via signal-hook-registry; the action only does an atomic store, so it is async-signal-safe. It is registered only after tokio's own stream exists, so SIGTERM's default action is never replaced. The sender's command and parent are read from ps at stop time. The lookup is a reviewed process-scan exception: it reads one kernel-named pid once on the stop path and enumerates nothing. If the action can't be registered, the daemon warns and stops as before.

Low risk: only the SIGTERM stop path changes, plus one extra signal action.

^{Written for commit 9a2e1da. Summary will update on new commits.}

[cursor]

Bugbot needs on-demand usage enabled

Bugbot uses usage-based billing for this team and requires on-demand usage to be enabled.

A team admin can enable on-demand usage in the Cursor dashboard.

[github-actions]

cmux status _{(one comment, edited in place; each line is updated by the job that owns it)}

Review 18057fe: HQ feed card · gallery before/after runs · checks

App of this head (app.zip, kept 1 day)

gh workflow run reload-build.yml -R manaflow-ai/cmux --ref feat-cmux-next -f tag=pr35147 -f ref=lane/acpmux-update-sigterm -f platform=macos -f backend_mode=local

web and size checks: green

Shipped in:

• cmux NEXT nightly 1.0.0-nightly.4182499138501 (e0aa712b3b)
• cmux NEXT nightly 1.0.0-nightly.4184087390301 (08052d331e)
• cmux NEXT nightly 1.0.0-nightly.4184224384601 (a3abfc4160)
• cmux NEXT nightly 1.0.0-nightly.4184409870901 (78ee0e108e)

[blacksmith-sh]

Found 2 test failures on Blacksmith runners:

Failures

Test	View Logs
nightly cmux-tui resolve/nightly cmux-tui resolve	View Logs
the short nightly decide gate must use the pooled Linux route with a Blacksmith fallbac k/ the short nightly decide gate must use the pooled Linux route with a Blacksmith fallbac k	View Logs

^{Need help on this PR? Tag @codesmith-bot with what you need.}

[teamleaderleo]

Review (codex, router-review): SHIP at 18057fe

Reviewer: codex gpt-6-astra high, fast tier.

Follow-ups: 1. They land as checklist items in this PR's one follow-up tracking issue in this repo (Follow-ups: #N ..., label follow-up; add to an open one before filing, never one issue per finding), linking this receipt; the lane that opened the PR fixes them forward.

SHIP at 18057fe

Visual verdict: no images; Leo tries the build.

• [P1] [follow-up] [category:reliability] [finding:F1] [outcome:pending] If term_sender::record() succeeds but Tokio’s subsequent SIGTERM registration fails at daemon.rs:592-596, the signal-hook handler replaces the default disposition while next_signal(None) waits forever. SIGTERM then no longer stops the daemon. Retain the registration handle and restore/unregister on Tokio setup failure, or explicitly terminate in that case.

Focused tests were not runnable because this box has no cargo.

Follow-ups:

• F1: Restore SIGTERM’s termination behavior when Tokio signal registration fails.

[teamleaderleo]

Review (codex, router-review): SHIP at 4985441

Reviewer: codex gpt-6-astra high, fast tier.

Follow-ups: 1. They land as checklist items in this PR's one follow-up tracking issue in this repo (Follow-ups: #N ..., label follow-up; add to an open one before filing, never one issue per finding), linking this receipt; the lane that opened the PR fixes them forward.

SHIP at 4985441
Visual verdict: no images; Leo tries the build

1. [P2] [follow-up] [category:correctness] [finding:F1] [outcome:pending] cmux-tui/crates/acpmux/src/daemon/term_sender.rs:24: Tokio’s earlier callback can wake the shutdown task on another thread before this callback stores SENDER. The multithreaded daemon can consequently read zero and omit the sender. Running both callbacks within one signal delivery does not guarantee ordering against other threads. Ensure sender capture precedes notification, and cover cross-thread delivery in a test.

Rust tests could not run because cargo is unavailable. The process-scan guard passed.

Follow-ups:

• F1: Synchronize sender capture with shutdown notification.

[cmux-bot]

cmux-bot approves 4985441: an independent review passed this exact head (router-review receipt by teamleaderleo). A push needs a new review; a later BLOCK at this head dismisses this approval.

A newer verdict blocks 9a2e1da (no independent SHIP at the head).

[teamleaderleo]

Review (codex gpt-6-astra high, router-review range e830594..9a2e1da): SHIP at 9a2e1da

Visual verdict: unchanged by this diff (no UI).

No correctness, regression, security or test findings. The head differs from the SHIP at 4985441 only by the lander's base merge.

Follow-ups: none

[teamleaderleo]

Finding outcome: comment-6627533347/F1 fixed c9f0649