* final} constants.
 *
 * <p>The {@link Joiner} class provides the inverse operation to splitting, but note that a
 * round-trip between the two should be assumed to be lossy.
 *
 * <p>See the Guava User Guide article on <a
 * href="https://github.com/google/guava/wiki/StringsExplained#splitter">{@code Splitter}</a>.
 *
 * @author Julien Silland
 * @author Jesse Wilson

 * final} constants.
 *
 * <p>The {@link Joiner} class provides the inverse operation to splitting, but note that a
 * round-trip between the two should be assumed to be lossy.
 *
 * <p>See the Guava User Guide article on <a
 * href="https://github.com/google/guava/wiki/StringsExplained#splitter">{@code Splitter}</a>.
 *
 * @author Julien Silland
 * @author Jesse Wilson

	if !is64Bit {
		return
	}

	runTestCases("64-bit", []testCase{
		0: {"-", maxInt, "out of range"},
	})
}

type RunesTest struct {
	in    string
	out   []rune
	lossy bool
}

var RunesTests = []RunesTest{
	{"", []rune{}, false},
	{" ", []rune{32}, false},
	{"ABC", []rune{65, 66, 67}, false},
	{"abc", []rune{97, 98, 99}, false},

    assertThat(transformation.slope()).isWithin(ALLOWED_ERROR).of(slope);
    // Note that we cannot test the actual mapping of points, as the results will be unreliable due
    // to loss of precision with this value of the slope.
  }

  public void testMappingWithSlope_maximalSlope() {
    double x1 = 1.2;
    double y1 = 3.4;
    double slope = Double.MAX_VALUE;

    assertThat(transformation.slope()).isWithin(ALLOWED_ERROR).of(slope);
    // Note that we cannot test the actual mapping of points, as the results will be unreliable due
    // to loss of precision with this value of the slope.
  }

  public void testMappingWithSlope_maximalSlope() {
    double x1 = 1.2;
    double y1 = 3.4;
    double slope = Double.MAX_VALUE;

Do not suppress NullAway errors without a justification.
Prefer rearranging the code to avoid the error, unless there is a provable performance penalty or a significant loss of code readability.
Use `Objects.requireNonNull` statement or `Preconditions` rather than `assert` or `@SuppressWarning` annotation to suppress if possible.

For gradual adoption, enable checks project-by-project.

of `Policy` optionally took in a `tf.compat.v1.mixed_precision.LossScale` in the constructor, which defaulted to a dynamic loss scale for the `"mixed_float16"` policy and no loss scale for other policies. In `Model.compile`, if the model's policy had a loss scale, the optimizer would be wrapped with a `LossScaleOptimizer`. With the non-experimental `Policy`, there is no loss scale associated with the `Policy`, and `Model.compile` wraps the optimizer with a `LossScaleOptimizer` if and only if the...

Previously, there was no formalization of property naming; only loose conventions were followed.
While these conventions served relatively well in practice, they created gaps, resulting in a loss of uniformity in some cases and leaving friction in the process of naming new properties.

### Terminology

**Public properties** are formally a part of the Public API.

      incidental, or consequential damages of any character arising as a
      result of this License or out of the use or inability to use the
      Work (including but not limited to damages for loss of goodwill,
      work stoppage, computer failure or malfunction, or any and all
      other commercial damages or losses), even if such Contributor
      has been advised of the possibility of such damages.

         */
        return sqrtFloor + lessThanBranchFree(halfSquare, x);
    }
    throw new AssertionError();
  }

  private static int sqrtFloor(int x) {
    // There is no loss of precision in converting an int to a double, according to
    // http://java.sun.com/docs/books/jls/third_edition/html/conversions.html#5.1.2
    return (int) Math.sqrt(x);
  }

  /**