From c4647dda6b482b419d8b120f0a14b59b326acece Mon Sep 17 00:00:00 2001
From: ocpway <ocpway@local>
Date: Thu, 30 Jul 2026 04:43:45 +0300
Subject: [PATCH] Fix hang on click and reconnect after daemon restart

Two root causes shared one symptom: the waybar module froze after a
click and needed a waybar restart to recover.

1. oc-waybar-ctl used socat's UNIX-SENDTO, a bidirectional relay that
   can block on a datagram socket. waybar runs the click command on
   its per-module event thread and stops updating the module until the
   command returns, so a hung socat froze the module. Replace the
   socat/nc chain with a bounded python3 sendto (2s timeout). python3
   is guaranteed present since the daemon needs it.

2. The daemon ran oc raw json with a blocking subprocess.run on the
   main loop, so a slow network fetch (up to 60s) stalled the 5s
   rotation and queued clicks. Move fetch_ocd to a worker thread; the
   main loop adopts the result when it lands and never blocks. Rotation
   and click draining stay live during a fetch.

Reconnect after a daemon restart now works because the click helper no
longer wedges waybar's event thread; the module's interval-1 cat
re-reads state.json as soon as the new daemon rewrites it (verified:
state.json returns within 1s).
---
 scripts/ocd-waybar     | 25 ++++++++++++++++++++++---
 scripts/ocd-waybar-ctl | 33 ++++++++++++++++++---------------
 2 files changed, 40 insertions(+), 18 deletions(-)

diff --git a/scripts/ocd-waybar b/scripts/ocd-waybar
index 26e33f3..73c8d94 100755
--- a/scripts/ocd-waybar
+++ b/scripts/ocd-waybar
@@ -20,6 +20,7 @@ import signal
 import socket
 import subprocess
 import sys
+import threading
 import time
 from datetime import datetime, timezone
 from html import escape
@@ -303,6 +304,11 @@ class Daemon:
         self.last_key = None
         self.sock = None
         self.running = True
+        # The ocd poll runs in a worker thread so the main loop never
+        # blocks on a slow network fetch. Rotation and clicks stay live.
+        self._fetch_thread = None
+        self._fresh_report = None
+        self._fetch_lock = threading.Lock()
 
     def event_deadline(self):
         now = time.monotonic()
@@ -327,14 +333,27 @@ class Daemon:
     def lock(self):
         self.locked_until = time.monotonic() + LOCK_SECS
 
+    def _fetch_worker(self):
+        rep = fetch_ocd()
+        if rep:
+            with self._fetch_lock:
+                self._fresh_report = rep
+
     def maybe_tick(self):
         now = time.monotonic()
-        if now >= self.next_poll:
+        # Kick off a poll on a worker thread; never block the main loop.
+        if now >= self.next_poll and (self._fetch_thread is None or not self._fetch_thread.is_alive()):
             self.next_poll = now + POLL_SECS
-            rep = fetch_ocd()
+            self._fetch_thread = threading.Thread(target=self._fetch_worker, daemon=True)
+            self._fetch_thread.start()
+        # Adopt a finished poll's result if one is ready.
+        if self._fetch_thread is not None and not self._fetch_thread.is_alive():
+            self._fetch_thread = None
+            with self._fetch_lock:
+                rep = self._fresh_report
+                self._fresh_report = None
             if rep:
                 self.report = rep
-                # keep index in range
                 n = len(rep.get("accounts") or [])
                 if n and self.index >= n:
                     self.index %= n
diff --git a/scripts/ocd-waybar-ctl b/scripts/ocd-waybar-ctl
index f347f46..2e7f5bb 100755
--- a/scripts/ocd-waybar-ctl
+++ b/scripts/ocd-waybar-ctl
@@ -1,30 +1,33 @@
 #!/bin/sh
 # ocd-waybar-ctl: send a command to the ocd-waybar daemon.
 # Usage: ocd-waybar-ctl {advance|lock|poll|quit}
-# wired into the waybar custom/ocd module as on-click / on-click-right.
+# Wired into the waybar custom/ocd module as on-click / on-click-right.
+#
+# This must never hang: waybar runs it on a click and stops updating the
+# module until it returns. socat's UNIX-SENDTO relay can block waiting on
+# the datagram socket, so we use a small python sendto with a hard timeout
+# instead. python3 is guaranteed present because the daemon itself needs it.
 
 set -eu
 cmd="${1:-advance}"
-sock="${XDG_RUNTIME_DIR:-/run/user/$(id -u)}/ocd-waybar.sock"
-
-if [ "$#" -gt 0 ]; then
-  cmd="$1"
-fi
-
 case "$cmd" in
   advance|lock|poll|quit) ;;
   *) echo "ocd-waybar-ctl: unknown command '$cmd'" >&2; exit 2 ;;
 esac
 
-if [ ! -S "$sock" ]; then
-  echo "ocd-waybar-ctl: daemon socket not found at $sock" >&2
-  exit 1
-fi
+sock="${XDG_RUNTIME_DIR:-/run/user/$(id -u)}/ocd-waybar.sock"
+[ -S "$sock" ] || { echo "ocd-waybar-ctl: daemon socket not found at $sock" >&2; exit 1; }
 
-printf '%s\n' "$cmd" | socat - UNIX-SENDTO:"$sock" 2>/dev/null \
-  || printf '%s\n' "$cmd" | nc -U -u -w1 "$sock" 2>/dev/null \
-  || python3 - "$sock" "$cmd" <<'PY'
+python3 - "$sock" "$cmd" <<'PY'
 import socket, sys
+sock, cmd = sys.argv[1], sys.argv[2]
 s = socket.socket(socket.AF_UNIX, socket.SOCK_DGRAM)
-s.sendto((sys.argv[2] + "\n").encode(), sys.argv[1])
+s.settimeout(2.0)
+try:
+    s.sendto((cmd + "\n").encode(), sock)
+except (OSError, socket.timeout) as e:
+    print("ocd-waybar-ctl: send failed: %s" % e, file=sys.stderr)
+    sys.exit(1)
+finally:
+    s.close()
 PY
\ No newline at end of file