test(qwp): join the I/O worker before asserting it dead in blocked-send close test

close() returns when the worker counts shutdownLatch down inside ioLoop's
finally -- before the worker's exit tail and actual thread termination --
so Thread.isAlive() can be observably true for a scheduling beat after
close() returns (linux-x64 failure in build 252736). Quiescence is the
latch plus the cleanup-before-countdown ordering (already asserted), not
thread death; every sibling test already uses join-then-assert. Reproduced
by sleeping 200ms after the countdown: the old assert fails exactly like
CI; the joined assert passes under the same window.

Author: bluestreak01

core/src/test/java/io/questdb/client/test/cutlass/qwp/client/sf/cursor/CursorWebSocketSendLoopBlockedSendCloseTest.java

@@ -95,7 +95,14 @@ public void testCloseBreaksBlockedSendBeforeJoiningWorker() throws Exception {
 
                 Thread ioThread = client.sendThread.get();
                 Assert.assertNotNull(ioThread);
-                Assert.assertFalse("I/O worker must be dead when close returns", ioThread.isAlive());
+                // close() returns when the worker counts shutdownLatch down --
+                // the worker's last action before its exit tail -- so the
+                // thread can be observably alive for a scheduling beat after
+                // close() returns. Quiescence is the latch plus the cleanup
+                // asserts below (cleanup runs before the countdown), not
+                // thread death: join briefly instead of asserting the race.
+                ioThread.join(TimeUnit.SECONDS.toMillis(5));
+                Assert.assertFalse("I/O worker did not exit after close returned", ioThread.isAlive());
                 Assert.assertEquals("full client cleanup must run exactly once", 1, client.cleanupCount.get());
                 Assert.assertEquals("the I/O worker must own cleanup before publishing exit",
                         ioThread, client.cleanupThread.get());