val tags = EmptyTags
    val atomicTags = AtomicReference<Tags>(tags)
    val result =
      atomicTags.computeIfAbsent(String::class) {
        // 'Race' by making another computeIfAbsent call. In practice this would be another thread.
        assertThat(atomicTags.computeIfAbsent(Integer::class) { 5 as Integer }).isEqualTo(5)
        "a"
      }
    assertThat(result).isEqualTo("a")

  Note that commits might be squashed by a maintainer on merge.
- [ ] Write unit tests that match behavioral changes, where the tests fail if the changes to the runtime are not applied.
  This may not always be possible but is a best-practice.
- [ ] Run `mvn verify` to make sure basic checks pass.
  A more thorough check will be performed on your pull request automatically.
- [ ] You have run the [Core IT][core-its] successfully.

   * very flexible regarding wildcard rules, but this flexibility is not something currently used
   * in practice. To simplify the implementation, we've avoided implementing the flexible rules in
   * favor of supporting what's actually used in practice. That means if these assertions ever fail,
   * the implementation will need to be revisited to support a more flexible rule.
   */

                VersionRange.createFromVersion("1.0"),
                "type",
                null,
                "system",
                "compile");

        // this one should never happen in practice...
        Artifact artifact5 = factory.createDependencyArtifact(
                "test-grp", "test-artifact-5", VersionRange.createFromVersion("1.0"), "type", null, "system", "system");

 * hint, so this technique may fail at the whim of the JDK implementation, for example if a user
 * specified the JVM flag {@code -XX:+DisableExplicitGC}. But in practice, it works very well for
 * ordinary tests.
 *
 * <p>Failure of the expected event to occur within an implementation-defined "reasonable" time
 * period or an interrupt while waiting for the expected event will result in a {@link

      Thread.interrupted();
    }
  }

  /**
   * awaitFullGc() is not quite as reliable a way to ensure calling of a specific finalize method as
   * the more direct await* methods, but should be reliable enough in practice to avoid flakiness of
   * this test. (And if it isn't, we'd like to know about it first!)
   */
  public void testAwaitFullGc() {
    CountDownLatch finalizerRan = new CountDownLatch(1);
    WeakReference<Object> ref =

# Calls

The HTTP client’s job is to accept your request and produce its response. This is simple in theory but it gets tricky in practice.

## [Requests](https://square.github.io/okhttp/4.x/okhttp/okhttp3/-request/)

Each HTTP request contains a URL, a method (like `GET` or `POST`), and a list of headers. Requests may also contain a body: a data stream of a specific content type.

## [Responses](https://square.github.io/okhttp/4.x/okhttp/okhttp3/-response/)

 * hint, so this technique may fail at the whim of the JDK implementation, for example if a user
 * specified the JVM flag {@code -XX:+DisableExplicitGC}. But in practice, it works very well for
 * ordinary tests.
 *
 * <p>Failure of the expected event to occur within an implementation-defined "reasonable" time
 * period or an interrupt while waiting for the expected event will result in a {@link

    // This approach does mean that we break out of the fast path code in cases
    // where we don't strictly need to, but this situation will almost never
    // occur in practice.
    if (safeMin >= Character.MIN_HIGH_SURROGATE) {
      // The safe range is empty or the all safe code points lie in or above the
      // surrogate range. Either way the character range is empty.

   * processed during inflation. If this happens, we need to either reject the message or discard
   * the unreachable data. We choose to discard it!
   *
   * In practice this could happen if the encoder doesn't strip the [0x00, 0x00, 0xff, 0xff] suffix
   * and that ends up repeated.
   *
   * https://github.com/square/okhttp/issues/8551
   */
  @Test