@simon-says-build/harbormaster 0.12.0 → 0.14.0

package/README.md

@@ -154,3 +154,48 @@ extensions) in other directories. `acquire`: on top of the daemon's scrubbed
 env, under the caller's explicit `--fn-env`. `exec`: on top of the caller's
 ambient env, so a key exported in some shell can't leak into the suite.
 `python3 test/project_env.py`.
+
+## An idle simulator nobody leases is shut down (0.13.0)
+
+The world sweep (every `FBPORTS_SWEEP_INTERVAL`, 5 min) now shuts down a
+booted simulator that meets all three conditions:
+
+- **no lease holds it**;
+- **booted longer than `FBPORTS_IDLE_SIM_HOURS`** (default 6; `0` turns this
+  off), measured by its `launchd_sim`;
+- **no app is running in it**.
+
+It is shut down only, never deleted, so whoever booted it can boot it again.
+
+On 2026-10-20 "Impulse-agent" had sat on the Mac Studio for 32 hours,
+booted outside any lease by an app's pinned simulator name, holding two of
+its ten cores with no app installed. Leases stopped only what they created.
+
+`status` reports what the sweep last saw and did: `idleSims {booted, unleased,
+shutDown[]}`. The CLI prints a `simulators:` line. Tested by
+`python3 test/idle_sims.py`.
+
+## An idb companion whose simulator is gone is ended (0.14.0)
+
+idb starts one `idb_companion --udid <UDID>` per simulator it drives, with a
+socket at `/tmp/idb/<UDID>_companion.sock`. A lease deletes its simulator, but
+nothing ended that simulator's companion. On 2026-10-20 the Mac Studio had 19
+companions running for devices that no longer existed, some from Oct 19, and
+109 stale sockets. (A wedged companion also cost a walk2gether run a phone the
+same day; the driver now restarts one that refuses, simon-says #1782.)
+
+The world sweep now ends a companion that meets both conditions:
+
+- **its UDID is not in `simctl list`**, read again just before ending it;
+- **it is older than 10 minutes**, so a device created after the inventory
+  read keeps its companion.
+
+It also removes a socket for a gone device when no companion holds it and the
+file is older than 10 minutes. Nothing else in the directory is touched
+(`FBPORTS_IDB_DIR`, default `/tmp/idb`). When `simctl list` fails or comes back
+empty, the sweep skips this step: an unreadable inventory never counts as
+"every device is gone".
+
+`status` reports `companions {running, ended, socketsRemoved}`, and the CLI
+prints an `idb companions:` line when the last sweep ended or removed anything.
+Tested by `python3 test/companions.py`.

package/harbormaster

@@ -84,7 +84,7 @@ import uuid
 from datetime import datetime, timezone
 from pathlib import Path
 
-VERSION = "0.12.0"
+VERSION = "0.14.0"
 
 # Overridable for tests / parallel installs; default is the shared user daemon.
 # The dir keeps the old "firebase-ports" name: pre-rename binaries still installed
@@ -125,6 +125,19 @@ BOOT_BACKSTOP = 600          # only for a genuinely wedged boot; readiness is li
 # How often the daemon sweeps the WORLD for resources that escaped the ledger:
 # listeners on pool ports with no lease, orphaned firebase runtimes, stray sims.
 SWEEP_INTERVAL = int(os.environ.get("FBPORTS_SWEEP_INTERVAL", "300"))
+# A booted simulator NO lease holds, up this long with no app running in it, is
+# shut down (never deleted: its owner can boot it again). 0 turns it off.
+# 2026-10-20: "Impulse-agent", booted outside any lease by an app's pinned
+# simulator name, held two of the Mac Studio's ten cores for 32 hours.
+IDLE_SIM_HOURS = float(os.environ.get("FBPORTS_IDLE_SIM_HOURS", "6"))
+# idb's per-simulator companions (0.14.0): one `idb_companion --udid <UDID>`
+# and its /tmp/idb/<UDID>_companion.sock per device it ever drove. A lease
+# deletes its simulator; nothing ended the companion. One older than the grace
+# whose device no longer exists is ended, and a socket for a gone device that no
+# companion holds is removed. The grace keeps a just-created device (newer than
+# the inventory read) out of it.
+IDB_SOCK_DIR = os.environ.get("FBPORTS_IDB_DIR", "/tmp/idb")
+COMPANION_GRACE_S = 600
 # Process-name signatures of firebase-tools' own children. firebase-tools spawns
 # these `detached: true` — their OWN process groups — so killpg on the suite can
 # structurally never reach them (the v1 "killpg fixes it" assumption was wrong:
@@ -459,6 +472,98 @@ def _sim_inventory():
         return None
 
 
+def _etime_s(etime):
+    """ps etime ([[dd-]hh:]mm:ss) in seconds, or None."""
+    try:
+        days, _, rest = etime.strip().rpartition("-")
+        parts = [int(x) for x in rest.split(":")]
+        while len(parts) < 3:
+            parts.insert(0, 0)
+        return (int(days) if days else 0) * 86400 + parts[0] * 3600 + parts[1] * 60 + parts[2]
+    except (ValueError, AttributeError):
+        return None
+
+
+_DEVICE_RE = re.compile(r"/CoreSimulator/Devices/([0-9A-Fa-f-]{36})/")
+
+
+def _booted_sims(ps_text):
+    """{udid: {"upS": seconds its launchd_sim has run, "app": an app is running}}
+    for every booted simulator, read off one `ps -axo etime=,command=` listing.
+    A device's launchd_sim is its boot; a process under its
+    data/Containers/Bundle/Application is an app running in it."""
+    out = {}
+    for line in ps_text.splitlines():
+        etime, _, cmd = line.strip().partition(" ")
+        m = _DEVICE_RE.search(cmd)
+        if not m:
+            continue
+        udid = m.group(1).upper()
+        d = out.setdefault(udid, {"upS": None, "app": False})
+        if "launchd_sim" in cmd.split(" ", 1)[0]:
+            up = _etime_s(etime)
+            if up is not None:
+                d["upS"] = max(d["upS"] or 0, up)
+        if "/data/Containers/Bundle/Application/" in cmd:
+            d["app"] = True
+    return out
+
+
+def idle_unleased_sims(booted, leased, names, hours=IDLE_SIM_HOURS):
+    """The booted simulators to shut down: no lease holds them, up longer than
+    `hours`, and no app running. Pure. `names` is {udid: name} (for the report)."""
+    if not hours or hours <= 0:
+        return []
+    out = []
+    for udid, d in booted.items():
+        if udid in leased or d.get("app") or d.get("upS") is None:
+            continue
+        if d["upS"] >= hours * 3600:
+            out.append({"udid": udid, "name": names.get(udid, ""), "upH": round(d["upS"] / 3600, 1)})
+    return out
+
+
+_COMPANION_UDID_RE = re.compile(r"--udid[ =]([0-9A-Fa-f-]{36})")
+_SOCK_RE = re.compile(r"^([0-9A-Fa-f-]{36})_companion\.sock$")
+
+
+def _companions(ps_text):
+    """[{pid, udid, upS}] for every idb_companion bound to one simulator, read off
+    one `ps -xo pid=,etime=,command=` listing."""
+    out = []
+    for line in ps_text.splitlines():
+        parts = line.split(None, 2)
+        if len(parts) < 3 or "idb_companion" not in parts[2].split(" ", 1)[0]:
+            continue
+        m = _COMPANION_UDID_RE.search(parts[2])
+        if not m:
+            continue
+        try:
+            pid = int(parts[0])
+        except ValueError:
+            continue
+        out.append({"pid": pid, "udid": m.group(1).upper(), "upS": _etime_s(parts[1])})
+    return out
+
+
+def dead_companions(companions, known, grace=COMPANION_GRACE_S):
+    """The companions to end: their device is not in `known` (every simulator's
+    UDID, upper-case) and they are older than `grace`. Pure."""
+    return [c for c in companions
+            if c["udid"] not in known and c.get("upS") is not None and c["upS"] >= grace]
+
+
+def stale_companion_sockets(names, known, live):
+    """The socket file names to remove: a device not in `known` that no live
+    companion (`live`, UDIDs) holds. Pure."""
+    out = []
+    for n in names:
+        m = _SOCK_RE.match(n)
+        if m and m.group(1).upper() not in known and m.group(1).upper() not in live:
+            out.append(n)
+    return out
+
+
 def _group_alive(pgid):
     """Any member of the process group left? EPERM counts as alive."""
     try:
@@ -997,7 +1102,11 @@ class Broker:
                                 "ports": l["ports"], "projectId": l.get("project_id"),
                                 "sim": l.get("sim_udid"), "expires": l.get("expires"),
                                 "label": l.get("label"), "log": l.get("log")}
-                               for l in self.leases.values()]}
+                               for l in self.leases.values()],
+                    # What the idle-simulator sweep last saw and did (SSB's Hosts
+                    # card reads it: a leak is a number, not a chart to diagnose).
+                    "idleSims": getattr(self, "_idle_sims", None),
+                    "companions": getattr(self, "_companions", None)}
 
     def reap(self):
         with self.lock:
@@ -1143,6 +1252,9 @@ class Broker:
                         and udid not in leased_udids):
                     findings.append({"kind": "sim", "udid": udid, "name": name,
                                      "evict": udid in mine})
+        if apply:
+            self._sweep_idle_sims(inv)
+            self._sweep_companions(inv)
         if apply and findings:
             pids = {f["pid"] for f in findings
                     if f["kind"] in ("listener", "orphan") and f.get("evict")}
@@ -1178,6 +1290,75 @@ class Broker:
                           + f" — {f['cmd']}", flush=True)
         return findings
 
+    def _sweep_idle_sims(self, inv):
+        """Shut down booted simulators no lease holds that have sat idle past
+        IDLE_SIM_HOURS. Shutdown only — the device and its data stay, and its
+        owner boots it again. Recorded for status whatever it finds."""
+        if not IDLE_SIM_HOURS or IDLE_SIM_HOURS <= 0:
+            return
+        try:
+            r = subprocess.run(["ps", "-axo", "etime=,command="], capture_output=True, text=True, timeout=10)
+        except (subprocess.SubprocessError, OSError):
+            return
+        booted = _booted_sims(r.stdout)
+        leased = {str(l.get("sim_udid")).upper() for l in self.leases.values() if l.get("sim_udid")}
+        names = {u.upper(): n for u, n in (inv or {}).items()}
+        idle = idle_unleased_sims(booted, leased, names)
+        done = []
+        for s in idle:
+            try:
+                subprocess.run(["xcrun", "simctl", "shutdown", s["udid"]], capture_output=True, text=True, timeout=60)
+                done.append(s)
+                print(f"idle sim {s['name'] or '?'} ({s['udid'][:8]}) shut down: no lease, "
+                      f"booted {s['upH']}h, no app running", flush=True)
+            except (subprocess.SubprocessError, OSError) as e:
+                print(f"idle sim {s['udid'][:8]}: shutdown failed: {e}", flush=True)
+        unleased = [u for u in booted if u not in leased]
+        prev = getattr(self, "_idle_sims", None) or {}
+        self._idle_sims = {"at": time.time(), "booted": len(booted), "unleased": len(unleased),
+                           "afterHours": IDLE_SIM_HOURS,
+                           "shutDown": (done + (prev.get("shutDown") or []))[:20]}
+
+    def _sweep_companions(self, inv):
+        """End idb companions whose simulator is gone and remove their sockets.
+        Skipped when the inventory is indeterminate: an empty or failed
+        `simctl list` must never read as "every device is gone"."""
+        if not inv:
+            return
+        try:
+            r = subprocess.run(["ps", "-xo", "pid=,etime=,command="], capture_output=True, text=True, timeout=10)
+        except (subprocess.SubprocessError, OSError):
+            return
+        comps = _companions(r.stdout)
+        dead = dead_companions(comps, {u.upper() for u in inv})
+        if dead:
+            # Re-read: a device created since the first read keeps its companion.
+            again = _sim_inventory()
+            dead = dead_companions(dead, {u.upper() for u in again}) if again else []
+        killed = sorted(_kill_wave({c["pid"] for c in dead})) if dead else []
+        if killed:
+            print(f"idb companions ended (device gone): {len(killed)} — "
+                  + ", ".join(c["udid"][:8] for c in dead[:8]), flush=True)
+        live = {c["udid"] for c in comps if c["pid"] not in set(killed)}
+        removed = 0
+        try:
+            names = os.listdir(IDB_SOCK_DIR)
+        except OSError:
+            names = []
+        now = time.time()
+        for n in stale_companion_sockets(names, {u.upper() for u in inv}, live):
+            path = os.path.join(IDB_SOCK_DIR, n)
+            try:
+                if now - os.lstat(path).st_mtime >= COMPANION_GRACE_S:
+                    os.unlink(path)
+                    removed += 1
+            except OSError:
+                pass
+        if removed:
+            print(f"idb sockets removed (device gone): {removed}", flush=True)
+        self._companions = {"at": now, "running": len(comps), "ended": len(killed),
+                            "socketsRemoved": removed}
+
     def doctor(self, fix=False):
         """Report (and with fix=True, evict) everything the sweep can see —
         the command that would have made 51 orphaned JVMs visible on day one."""
@@ -1728,7 +1909,17 @@ def main():
         if not _up():
             print("daemon not running.")
             return
-        leases = _call("GET", "/status")[1].get("leases", [])
+        st = _call("GET", "/status")[1]
+        leases = st.get("leases", [])
+        idle = st.get("idleSims")
+        if idle:
+            print(f"simulators: {idle['booted']} booted, {idle['unleased']} with no lease"
+                  + (f"; shut down idle: " + ", ".join(f"{x['name'] or x['udid'][:8]} ({x['upH']}h)" for x in idle["shutDown"][:5])
+                     if idle.get("shutDown") else ""))
+        comp = st.get("companions")
+        if comp and (comp.get("ended") or comp.get("socketsRemoved")):
+            print(f"idb companions: {comp['running']} running; last sweep ended {comp['ended']}"
+                  f" and removed {comp['socketsRemoved']} socket(s) for gone devices")
         if not leases:
             print("nothing out.")
             return

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@simon-says-build/harbormaster",
-  "version": "0.12.0",
+  "version": "0.14.0",
   "description": "A localhost daemon that leases each caller its own port block — isolated Firebase emulator sandboxes, Metro/Expo ports, private iOS simulators — reaped when the caller exits. No port collisions, no shared instances, no pkill.",
   "bin": {
     "harbormaster": "harbormaster",
@@ -27,11 +27,11 @@
   "license": "MIT",
   "repository": {
     "type": "git",
-    "url": "git+https://github.com/simon-says-build/harbormaster.git"
+    "url": "git+https://github.com/simon-says-studio/harbormaster.git"
   },
-  "homepage": "https://github.com/simon-says-build/harbormaster#readme",
+  "homepage": "https://github.com/simon-says-studio/harbormaster#readme",
   "bugs": {
-    "url": "https://github.com/simon-says-build/harbormaster/issues"
+    "url": "https://github.com/simon-says-studio/harbormaster/issues"
   },
   "os": [
     "darwin",