fix: codex app-server crash text reaches the user when stderr lags the exit (#121467)

Same race as the ACP client: _subprocess_died reads stderr_tail() the moment
is_alive() turns False, before the reader thread has the crash lines, so the user
saw 'exited unexpectedly' with no cause (CI on main: test_crash_mid_item...).
stderr_tail() now joins the reader briefly once the process has exited.

agent/transports/codex_app_server.py

@@ -281,6 +281,9 @@ class CodexAppServerClient:
 
     def stderr_tail(self, n: int = 20) -> list[str]:
         """Return last n lines of codex's stderr (for error reports)."""
+        if self._proc.poll() is not None:
+            # The reader can still hold the crash text when the exit is first observed (#121467).
+            self._stderr_reader.join(timeout=1.0)
         with self._stderr_lock:
             return list(self._stderr_lines[-n:])