* description of CL 265462958.
   */
  // TODO(user): Write more tests for memory retention.
  @ForOverride
  @OverridingMethodsMustInvokeSuper
  void releaseResources(ReleaseResourcesReason reason) {
    checkNotNull(reason);
    /*
     * All elements of `futures` are completed, or this future has already completed and read
     * `futures` into a local variable (in preparation for propagating cancellation to them). In

    }
  }

  // If you have a FinalizableReferenceQueue that is a static field of one of the classes of your
  // app (like the FrqUser class above), then the app's ClassLoader will never be gc'd. The reason
  // is that we attempt to run a thread in a separate ClassLoader that will detect when the FRQ
  // is no longer referenced, meaning that the app's ClassLoader has been gc'd, and when that

            Thread.currentThread().setContextClassLoader(projectRealm);
        }
    }

    // TODO I'm really wondering where this method belongs; smells like it should be on MavenProject, but for some
    // reason it isn't ? This localization is kind-of a code smell.

    public static String getKey(MavenProject project) {
        return project.getGroupId() + ':' + project.getArtifactId() + ':' + project.getVersion();
    }

/* jcifs smb client library in Java
 * Copyright (C) 2002  "Michael B. Allen" <jcifs at samba dot org>
 *                   "Jason Pugsley" <jcifs at samba dot org>
 *                   "skeetz" <jcifs at samba dot org>
 *                   "Eric Glass" <jcifs at samba dot org>
 *                   and Marcel, Thomas, ...
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public

       * fastest (more than twice as fast as the Java implementation, vs ~1.5x with non-final static
       * fields, on x86_32) under the Hotspot server compiler. The reason is obviously that the
       * non-final fields need to be reloaded inside the loop.
       *
       * And, no, defining (final or not) local variables out of the loop still isn't as good

 * Generally speaking, this class reduces object allocation and memory consumption at the price of
 * moderately increased constant factors of CPU. Only use this class when there is a specific reason
 * to prioritize memory over CPU.
 *
 * @author Dimitris Andreou
 * @author Jon Noack
 */
@GwtIncompatible // not worth using in GWT for now
@ElementTypesAreNonnullByDefault

   * 0}.
   *
   * @throws IllegalArgumentException if {@code a < 0} or {@code b < 0}
   */
  public static int gcd(int a, int b) {
    /*
     * The reason we require both arguments to be >= 0 is because otherwise, what do you return on
     * gcd(0, Integer.MIN_VALUE)? BigInteger.gcd would return positive 2^31, but positive 2^31 isn't
     * an int.
     */

    int newCapacity =
        (oldCapacity < 64) ? (oldCapacity + 1) * 2 : IntMath.checkedMultiply(oldCapacity / 2, 3);
    return capAtMaximumSize(newCapacity, maximumSize);
  }

  /** There's no reason for the queueSize to ever be more than maxSize + 1 */
  private static int capAtMaximumSize(int queueSize, int maximumSize) {
    return Math.min(queueSize - 1, maximumSize) + 1; // don't overflow
  }

/**
 * Static utility methods pertaining to {@link InetAddress} instances.
 *
 * <p><b>Important note:</b> Unlike {@code InetAddress.getByName()}, the methods of this class never
 * cause DNS services to be accessed. For this reason, you should prefer these methods as much as
 * possible over their JDK equivalents whenever you are expecting to handle only IP address string
 * literals -- there is no blocking DNS penalty for a malformed string.
 *

   * Preconditions.checkNotNull.
   *
   * (verifyNotNull has many fewer "problem" callers, so we could try to be stricter. On the other
   * hand, verifyNotNull arguably has more reason to accept nullable arguments in the first
   * place....)
   */

  /**
   * Ensures that {@code reference} is non-null, throwing a {@code VerifyException} with a default
   * message otherwise.
   *