for (int i = 0; i < threadCount; i++) {
            final int index = i;
            threads[i] = new Thread(() -> {
                try {
                    // Multiple threads accessing context simultaneously
                    for (int j = 0; j < 50; j++) {
                        context.getConfig();
                        context.getBufferCache();
                        context.getDfs();

        final AtomicInteger failureCount = new AtomicInteger(0);
        final List<Exception> exceptions = new ArrayList<>();

        // Create multiple threads that all try to connect simultaneously
        for (int i = 0; i < threadCount; i++) {
            new Thread(() -> {
                try {
                    startLatch.await(); // Wait for signal to start
                    conn.connect();

  var idleAtNs = Long.MAX_VALUE

  /**
   * Returns true if this is an HTTP/2 connection. Such connections can be used in multiple HTTP
   * requests simultaneously.
   */
  internal val isMultiplexed: Boolean
    get() = http2Connection != null

  /** Prevent further exchanges from being created on this connection. */
  override fun noNewExchanges() {
    withLock {

        byte[] buffer = new byte[1000];
        int numThreads = 10;
        Thread[] threads = new Thread[numThreads];

        // When - multiple threads calling methods simultaneously
        for (int i = 0; i < numThreads; i++) {
            final int threadIndex = i;
            threads[i] = new Thread(() -> {
                for (int j = 0; j < 100; j++) {

  @GwtIncompatible // NullPointerTester
  public void testNullPointers() {
    /*
     * Don't bother testing: Preconditions defines a bunch of methods that accept a template (or
     * even entire message) that simultaneously:
     *
     * - _shouldn't_ be null, so we don't annotate it with @Nullable
     *
     * - _can_ be null without causing a runtime failure (because we don't want the interesting

                } finally {
                    completeLatch.countDown();
                }
            }).start();
        }

        // Start all threads simultaneously
        startLatch.countDown();

        // Wait for completion
        assertTrue(completeLatch.await(5, TimeUnit.SECONDS));

        // Authenticator should be closed
        assertTrue(auth.isClosed());

state.

This is a slower operation that does not use replication queues and is designed to walk the namespace and replicate objects one at a time so as not to impede server load. Ideally, resync should not be initiated for multiple buckets simultaneously - progress of the syncing can be monitored by looking at `mc replicate resync status alias/bucket --remote-bucket <arn>`. In the event that resync operation failed to replicate some versions, they would be picked up by the healing mechanism in-built...

 * Since connections may be reused, the proxy selection, DNS, and connect events may not be present
 * for a call. In future releases of OkHttp these events may also occur concurrently to permit
 * multiple routes to be attempted simultaneously.
 *
 * Events and sequences of events may be repeated for retries and follow-ups.
 *
 * All event methods must execute fast, without external locking, cannot throw exceptions, attempt

    assertEvents(
      "plan 3 TCP connect canceled",
    )

    taskFaker.assertNoMoreTasks()
  }

  /**
   * This test causes two connections to become available simultaneously, one from a TCP connect and
   * one from the pool. We must take the pooled connection because by taking it from the pool, we've
   * fully acquired it.
   *

Use an HTTP POST to send a request body to a service. This example posts a markdown document to a web service that renders markdown as HTML. Because the entire request body is in memory simultaneously, avoid posting large (greater than 1 MiB) documents using this API.

=== ":material-language-kotlin: Kotlin"
    ```kotlin
      private val client = OkHttpClient()

      fun run() {