* objects come from factory methods that are themselves {@code @Nullable}. In reality neither the
   * array nor its contents can be null, but it is not useful to force the use of {@code
   * requireNonNull} or the like just to assert that.
   *
   * <p>{@code EqualsTester} will always check that every object it is given returns false from

     * initiated by the client (MSIE post requests after successful NTLM SSP
     * authentication). If false and the user has not been authenticated yet
     * the client will be forced to send an authentication (server sends
     * HttpServletResponse.SC_UNAUTHORIZED).
     * @return True if the negotiation is complete, otherwise false
     * @throws IOException if an I/O error occurs

   * objects come from factory methods that are themselves {@code @Nullable}. In reality neither the
   * array nor its contents can be null, but it is not useful to force the use of {@code
   * requireNonNull} or the like just to assert that.
   *
   * <p>{@code EqualsTester} will always check that every object it is given returns false from

 * strategies.
 *
 * <ul>
 *   <li>SafeAtomicHelper: uses Atomic FieldsUpdaters to implement synchronization
 *   <li>SynchronizedHelper: uses {@code synchronized} blocks for synchronization
 * </ul>
 *
 * To force selection of our fallback strategies we load {@link AggregateFutureState} (and all of
 * {@code com.google.common.util.concurrent}) in degenerate class loaders which make certain

         * the case of the Samsung bug, and we do that by breaking AtomicReferenceFieldUpdater.
         * Breaking AtomicReferenceFieldUpdater not only forces AggregateFutureState to fall back to
         * another implementation but also forces AbstractFutureState to be able to do the
         * same—hence the try-catch here.
         *

    <dependency>
      <groupId>org.apache.maven.resolver</groupId>
      <artifactId>maven-resolver-transport-file</artifactId>
    </dependency>
    <!-- HTTP/1.1, lowest priority, Java8+ (still must as some ITs force it) -->
    <dependency>
      <groupId>org.apache.maven.resolver</groupId>
      <artifactId>maven-resolver-transport-wagon</artifactId>
    </dependency>
    <!-- HTTP/1.1, medium priority, Java8+ -->
    <dependency>

	return nil
}

// Remove removes tier if it is empty.
func (config *TierConfigMgr) Remove(ctx context.Context, tier string, force bool) error {
	d, err := config.getDriver(ctx, tier)
	if err != nil {
		if errors.Is(err, errTierNotFound) {
			return nil
		}
		return err
	}
	if !force {
		if inuse, err := d.InUse(ctx); err != nil {
			return err
		} else if inuse {
			return errTierBackendNotEmpty
		}

    }

    @Test
    public void testStatistics() throws Exception {
        long initialAllocated = bufferManager.getTotalAllocated();

        // Allocate regions larger than pool buffer size to force new allocations
        RdmaMemoryRegion region1 = bufferManager.getSendRegion(131072); // 128KB > 64KB pool size
        RdmaMemoryRegion region2 = bufferManager.getSendRegion(131072); // Another large allocation

 * selected in the static initializer of AbstractFuture. This is convenient and performant but
 * introduces some testing difficulties. This test exercises the fallback strategies.
 *
 * <p>To force selection of our fallback strategies, we load {@link AbstractFuture} (and all of
 * {@code com.google.common.util.concurrent}) in degenerate class loaders which make certain

To sign off a single commit:

`git commit --amend --signoff`

To sign off one or multiple commits:

`git rebase --signoff origin/master`

Then force push your branch:

`git push --force origin test-branch`

### Fixing sanity check failures after public API changes

If your PR includes any changes to the Gradle Public API, it will cause the binary compatibility check to fail.