* most once unless the underlying {@code get()} throws an exception. The supplier's serialized
   * form does not contain the cached value, which will be recalculated when {@code get()} is called
   * on the deserialized instance.
   *
   * <p>When the underlying delegate throws an exception then this memoizing supplier will keep
   * delegating calls until it returns valid data.

   * most once unless the underlying {@code get()} throws an exception. The supplier's serialized
   * form does not contain the cached value, which will be recalculated when {@code get()} is called
   * on the deserialized instance.
   *
   * <p>When the underlying delegate throws an exception then this memoizing supplier will keep
   * delegating calls until it returns valid data.

## Decision

### Types of properties that should not use lazy types

The guidelines below do not apply to all properties. There are cases where the lazy types should not be used:

#### Non-calculated values

These are inappropriate uses of lazy types:

```groovy
class Example {
    Property<String> getSomeProperty()

    Example() {
        getSomeProperty().set("value")

    /**
     * Specifies a single quantile index to be calculated, i.e. the k in the kth q-quantile.
     *
     * @param index the quantile index, which must be in the inclusive range [0, q] for q-quantiles
     */
    public ScaleAndIndex index(int index) {
      return new ScaleAndIndex(scale, index);
    }

    /**
     * Specifies multiple quantile indexes to be calculated, each index being the k in the kth
     * q-quantile.

     * Supports Integer, Long, Float, Double, and String representations of numbers.
     *
     * @param map the document data as a map of field names to values
     * @return the calculated boost value as a float, or 0.0f if evaluation fails
     */
    public float getValue(final Map<String, Object> map) {
        if (map == null || map.isEmpty()) {
            return 0.0f;
        }

Especially including efforts that are normally less visible, and in many cases more arduous, like helping others with questions and reviewing Pull Requests with translations.

The data is calculated each month, you can read the [source code here](https://github.com/fastapi/fastapi/blob/master/scripts/).

Here I'm also highlighting contributions from sponsors.

	if opts.MTime.IsZero() {
		fi.ModTime = UTCNow()
	}

	// Save successfully calculated md5sum.
	// for replica, newMultipartUpload would have already sent the replication ETag
	if fi.Metadata["etag"] == "" {
		if opts.UserDefined["etag"] != "" {
			fi.Metadata["etag"] = opts.UserDefined["etag"]
		} else { // fallback if not already calculated in handler.
			fi.Metadata["etag"] = getCompleteMultipartMD5(parts)
		}
	}

        assertEquals(0, qrList.getStart());
        assertEquals(10, qrList.getPageSize());
        assertEquals(0, qrList.getOffset());

        // Page info should be calculated (allRecordCount=100, pageSize=10 => 10 pages)
        // But since size() < pageSize, collapsing logic applies and allPageCount becomes currentPageNumber

   * unless the user is actually calling a parse method.
   */
  private static final class ParseOverflowDetection {
    private ParseOverflowDetection() {}

    // calculated as 0xffffffffffffffff / radix
    static final long[] maxValueDivs = new long[Character.MAX_RADIX + 1];
    static final int[] maxValueMods = new int[Character.MAX_RADIX + 1];

  }

  public void testScale_indexes_varargs_compute_doubleCollection_snapshotsIndexes() {
    // This test is the same as testScale_indexes_varargs_compute_doubleCollection except that the
    // array of indexes to be calculated is modified between the calls to indexes and compute: since
    // the contract is that it is snapshotted, this shouldn't make any difference to the result.
    int[] indexes = {0, 10, 5, 1, 8, 10};