*  Fix: Don't miss whole operation timeouts when calls redirect.
 *  Fix: Don't leak connections if web sockets have malformed responses or if `onOpen()` throws.
 *  Fix: Don't retry when request bodies fail due to `FileNotFoundException`.
 *  Fix: Don't crash when URLs have IPv4-mapped IPv6 addresses.
 *  Fix: Don't crash when building `HandshakeCertificates` on Android API 28.

 * {@link #callAndAssertBlocks} method is ever called in a test, the lock-like object must have a
 * method equivalent to {@link java.util.concurrent.locks.ReentrantLock#hasQueuedThread(Thread)}. If
 * the {@link #callAndAssertWaits} method is ever called in a test, the lock-like object must have a
 * method equivalent to {@link
 * java.util.concurrent.locks.ReentrantLock#hasWaiters(java.util.concurrent.locks.Condition)},

in a future release! If you’re skipping releases, it’ll be much easier if you upgrade to OkHttp 4.0
as an intermediate step.

#### Vars and Vals

Java doesn’t have language support for properties so developers make do with getters and setters.
Kotlin does have properties and we take advantage of them in OkHttp.

 * **Address**: certificatePinner, connectionSpecs, dns, hostnameVerifier, protocols, proxy,

 * {@link #callAndAssertBlocks} method is ever called in a test, the lock-like object must have a
 * method equivalent to {@link java.util.concurrent.locks.ReentrantLock#hasQueuedThread(Thread)}. If
 * the {@link #callAndAssertWaits} method is ever called in a test, the lock-like object must have a
 * method equivalent to {@link
 * java.util.concurrent.locks.ReentrantLock#hasWaiters(java.util.concurrent.locks.Condition)},

   * someone still manages to pass null, let's continue to have the method work.
   *
   * (PatternFilenameFilter is of course one of those classes that shouldn't be a publicly visible
   * class to begin with but rather something returned from a static factory method whose declared
   * return type is plain FilenameFilter. If we made such a change, then the annotation we choose

        } catch (ResultOffsetExceededException e) {
            assertEquals(expectedMessage, e.getMessage());
            assertNull(e.getCause());
        } catch (Exception e) {
            fail("Should have caught ResultOffsetExceededException");
        }
    }

    public void test_throwAndCatchAsFessSystemException() {
        // Test catching as parent class FessSystemException

        } catch (SearchQueryException e) {
            assertEquals(expectedMessage, e.getMessage());
            assertNull(e.getCause());
        } catch (Exception e) {
            fail("Should have caught SearchQueryException");
        }
    }

    public void test_throwWithCauseAndCatch() {
        // Test throwing with cause and catching
        String expectedMessage = "Test with cause";

@GwtCompatible
public final class Futures extends GwtFuturesCatchingSpecialization {

  // A note on memory visibility.
  // Many of the utilities in this class (transform, withFallback, withTimeout, asList, combine)
  // have two requirements that significantly complicate their design.
  // 1. Cancellation should propagate from the returned future to the input future(s).
  // 2. The returned futures shouldn't unnecessarily 'pin' their inputs after completion.

    tests, either. This causes it to fail to find AtomicLongMapTest.

    Our workaround is to tell GWT that util.concurrent and all other packages
    have prod supersource, even if they have none. GWT is happy to ignore us
    when we specify a nonexistent path.

    (I hope that this workaround does not cause its own problems in the future.)
-->
<super-source path="super"/>

   * tasks. Second, the returned list will always be empty, as any submitted task is considered to
   * have started execution. This applies also to tasks given to {@code invokeAll} or {@code
   * invokeAny} which are pending serial execution, even the subset of the tasks that have not yet
   * started execution. It is unclear from the {@code ExecutorService} specification if these should