* {@code Integer.MAX_VALUE} or {@code Integer.MIN_VALUE} is returned, respectively.
   *
   * @since 20.0
   */
  public static int saturatedAdd(int a, int b) {
    return Ints.saturatedCast((long) a + b);
  }

  /**
   * Returns the difference of {@code a} and {@code b} unless it would overflow or underflow in

            b *= b;
          }
      }
    }
  }

  /**
   * Returns the sum of {@code a} and {@code b} unless it would overflow or underflow in which case
   * {@code Long.MAX_VALUE} or {@code Long.MIN_VALUE} is returned, respectively.
   *
   * @since 20.0
   */
  // Whenever both tests are cheap and functional, it's faster to use &, | instead of &&, ||

   * threads take special action by participating in the signal-passing game.
   */

  /*
   * Timeout handling is intricate, especially given our ambitious goals:
   * - Avoid underflow and overflow of timeout values when specified timeouts are close to
   *   Long.MIN_VALUE or Long.MAX_VALUE.
   * - Favor responding to interrupts over timeouts.

> m even if x < m. // // x and m operands must have the same announced length. // //go:norace func (x *Nat) maybeSubtractModulus(always choice, m *Modulus) { t := NewNat().set(x) underflow := t.sub(m.nat) // We keep the result if x - m didn't underflow (meaning x >= m) // or if always was set. keep := not(choice(underflow)) | choice(always) x.assign(keep, t) } // Sub computes x = x - y mod m. // // The length of both operands must be the same as the modulus. Both operands // must already be reduced...

        `TransformedDistribution`.
    *   Fix a bug in `import_meta_graph`'s handling of partitioned variables
        when
    *   Ensure `tf.distributions.Multinomial` doesn't underflow in `log_prob`.
        Before this change, all partitions of an integer variable were
        initialized with the shape of the unpartitioned variable; after this
        change they are initialized correctly.