release has many critical bug fixes and is safe to run in production. We're eager to stabilize our
new APIs so we can get out of alpha.

 *  New: Support Java 21's virtual threads (‘OpenJDK Project Loom’). We changed OkHttp's internals
    to use `Lock` and `Condition` instead of `synchronized` for best resource utilization.

   * might not (and the answer might change while we're building the array).
   *
   * TODO: cpovirk - Is this an issue anywhere anymore? It looks to have been fixed for Java 8
   * (https://bugs.openjdk.org/browse/JDK-7121314) and before Lollipop
   * (https://r.android.com/47508). We *would* need to worry for J2KT, whose own concurrency support
   * is evolving (b/381065164, b/458160722), but this class is @J2ktIncompatible.

  }

  /*
   * TODO: b/381289911 - Make the Iterator overload use StringJoiner (including Android or not)—or
   * some other optimization, given that StringJoiner can over-allocate:
   * https://bugs.openjdk.org/browse/JDK-8305774
   */

  // TODO: b/381289911 - Optimize MapJoiner similarly to Joiner (including Android or not).

  /**

         */
        helper = new SynchronizedHelper();
      }
    }
    ATOMIC_HELPER = helper;

    // Prevent rare disastrous classloading in first call to LockSupport.park.
    // See: https://bugs.openjdk.org/browse/JDK-8074773
    @SuppressWarnings("unused")
    Class<?> ensureLoaded = LockSupport.class;

    // Log after all static init is finished; if an installed logger uses any Futures methods, it

          helper = new SynchronizedHelper();
        }
      }
    }
    ATOMIC_HELPER = helper;

    // Prevent rare disastrous classloading in first call to LockSupport.park.
    // See: https://bugs.openjdk.org/browse/JDK-8074773
    @SuppressWarnings("unused")
    Class<?> ensureLoaded = LockSupport.class;

    // Log after all static init is finished; if an installed logger uses any Futures methods, it

   * and released to the public domain, as explained at
   * http://creativecommons.org/licenses/publicdomain
   *
   * As of 2010/06/11, this method is identical to the (package private) hash method in OpenJDK 7's
   * java.util.HashMap class.
   */
  // Copied from java/com/google/common/collect/Hashing.java
  private static int smear(int hashCode) {
    hashCode ^= (hashCode >>> 20) ^ (hashCode >>> 12);

   * noted, but the initial table is a zero-length array.
   *
   * <p>This test may fail miserably on non-OpenJDK environments...
   */
  @J2ktIncompatible
  @GwtIncompatible // reflection
  @AndroidIncompatible // relies on assumptions about OpenJDK
  public void testNewHashMapWithExpectedSize_wontGrow() throws Exception {
    // before jdk7u40: creates one-bucket table

    assertThat(verifier.verify("tel.com", session)).isFalse()
    assertThat(verifier.verify("k.com", session)).isFalse()
  }

  @Test fun specialKInExternalCert() {
    // OpenJDK related test.
    platform.assumeNotConscrypt()

    // Expecting actual:
    //  ["℡.com", "K.com"]
    // to contain exactly (and in same order):
    //  ["���.com", "���.com"]

   * is what most users want and expect it to do.
   *
   * <p>This behavior can't be broadly guaranteed, but has been tested with OpenJDK 1.7 and 1.8.
   *
   * @param expectedSize the number of elements you expect to add to the returned set
   * @return a new, empty hash set with enough capacity to hold {@code expectedSize} elements
   *     without resizing

 * Added support for compressed data frames.
 * Process push promise callbacks in order.
 * Update to http/2 draft 12.
 * Update to HPACK draft 07.
 * Add ALPN support. Maven will use ALPN on OpenJDK 8.
 * Update NPN dependency to target `jdk7u60-b13` and `Oracle jdk7u55-b13`.
 * Ensure SPDY variants support zero-length DELETE and POST.
 * Prevent leaking a cache item's InputStreams when metadata read fails.