?, ? extends OneSizeTestContainerGenerator<BiMap<K, V>, Entry<K, V>>>
          parentBuilder) {
    List<TestSuite> derived = super.createDerivedSuites(parentBuilder);
    // TODO(cpovirk): consider using this approach (derived suites instead of extension) in
    // ListTestSuiteBuilder, etc.?
    derived.add(
        MapTestSuiteBuilder.using(new MapGenerator<K, V>(parentBuilder.getSubjectGenerator()))

     * down. twiceSignifFloor will contain the top SIGNIFICAND_BITS + 2 bits, and signifFloor the
     * top SIGNIFICAND_BITS + 1.
     *
     * It helps to consider the real number signif = absX * 2^(SIGNIFICAND_BITS - exponent).
     */
    int shift = exponent - SIGNIFICAND_BITS - 1;
    long twiceSignifFloor = absX.shiftRight(shift).longValue();

and there are no missing null checks.

## Decision

Use "NullAway" in JSpecify mode to check for null-related errors in Java code.

Consider improper uses of `null` detected by NullAway a compilation error (not a warning, and not a separate CI job to test).

Do not suppress NullAway errors without a justification.

   *     unimplement<i>able</i> (for example, consider a {@code Converter<Chicken, ChickenNugget>}),
   *     then this is not logically a {@code Converter} at all, and should just implement {@link
   *     Function}.
   */
  @ForOverride
  protected abstract A doBackward(B b);

  // API (consumer-side) methods

  /**

If there are compelling use cases for early closing in dependencies with `yield`, I would consider adding a new way to opt in to early closing.

### Dependencies with `yield` and `except`, Technical Details { #dependencies-with-yield-and-except-technical-details }

This is out of scope for this ADR.  Migrating an existing property to lazy types is being handled in a different way.

### Conventions

Once a new property has been introduced, you need to consider what its conventions will be.

   * but not the other, the shorter iterable is considered to be less than the longer one. For
   * example, a lexicographical natural ordering over integers considers {@code [] < [1] < [1, 1] <
   * [1, 2] < [2]}.
   *
   * <p>Note that {@code ordering.lexicographical().reverse()} is not equivalent to {@code
   * ordering.reverse().lexicographical()} (consider how each would order {@code [1]} and {@code [1,
   * 1]}).
   *

 * <code>S-1-5-21-1496946806-2192648263-3843101252-1029</code> but they may
 * also be resolved to yield the name of the associated Windows account
 * such as <code>Administrators</code> or <code>MYDOM\alice</code>.
 * <p>
 * Consider the following output of <code>examples/SidLookup.java</code>:
 *
 * <pre>
 *        toString: S-1-5-21-4133388617-793952518-2001621813-512
 * toDisplayString: WNET\Domain Admins
 *         getType: 2

            log.warn("NTLMv1 LM response is deprecated and insecure. Consider using NTLMv2.");
            return NtlmUtil.getPreNTLMResponse(tc, getPasswordAsCharArray(), chlng);
        case 2:
            // NTLMv1 with NTLM response - still insecure
            log.warn("NTLMv1 NTLM response is deprecated and insecure. Consider using NTLMv2.");
            return NtlmUtil.getNTLMResponse(getPasswordAsCharArray(), chlng);

	// Bind to the lookup user account
	if err = l.LDAP.LookupBind(conn); err != nil {
		return nil, err
	}

	// Check if the passed in username is a valid DN.
	if !l.ParsesAsDN(username) {
		// We consider it as a login username and attempt to check it exists in
		// the directory.
		bindDN, err := l.LDAP.LookupUsername(conn, username)
		if err != nil {
			if strings.Contains(err.Error(), "User DN not found for") {