}
                            }
                        } finally {
                            deleteDir() // clean up after ourselves to reduce disk space
                        }
                    }
                }
            }
        }
    }
}

// run the parallel ITs
parallel(runITsTasks)

//     Some of the Veeam products use Multi Delete operations. This setting can reduce how many objects are included in one
//     multi-delete operation. The same registry key setting overwrites the storage-defined setting.
//     Optional value, default 1000, range: 1-unlimited (S3 standard maximum is 1000 and should not be set higher)
//
//   - <StorageConcurrentTasksLimit>

   * This test creates a long random sequence of inputs, then a lot of differently configured sinks
   * process it; all should produce the same answer, the only difference should be the number of
   * process()/processRemaining() invocations, due to alignment.
   */
  @AndroidIncompatible // slow. TODO(cpovirk): Maybe just reduce iterations under Android.
  public void testExhaustive() throws Exception {

                .filter(q -> StringUtil.isNotBlank(q) && q.length() <= maxQueryLength)
                .forEach(oq -> split(oq, " ")
                        .get(s -> s.filter(StringUtil::isNotBlank).reduce((q1, q2) -> escape(q1, "(", ")") + OR + escape(q2, "(", ")")))
                        .ifPresent(q -> {
                            appendQuery(queryBuf, q);
                        })));

   * storedPermits becomes 10.0 (assuming maxStoredPermits >= 10.0). At that point, a request of
   * acquire(3) arrives. We serve this request out of storedPermits, and reduce that to 7.0 (how
   * this is translated to throttling time is discussed later). Immediately after, assume that an
   * acquire(10) request arriving. We serve the request partly from storedPermits, using all the

  }

  /**
   * Returns a larger power of 2 hashtable size given the current mask.
   *
   * <p>For hashtable sizes less than or equal to 32, the returned power of 2 is 4x the current
   * hashtable size to reduce expensive rehashing. Otherwise the returned power of 2 is 2x the
   * current hashtable size.
   */
  static int newCapacity(int mask) {
    return ((mask < 32) ? 4 : 2) * (mask + 1);
  }

        if (sources == null) {
            return null;
        }
        return sources.stream()
                .reduce((a, b) -> {
                    throw new IllegalStateException(String.format(
                            "No unique Source for %s:%s: %s and %s",

That way I could find the best ways to reduce code duplication as much as possible, to have completion everywhere, type and error checks, etc.

All in a way that provided the best development experience for all the developers.

## Requirements

/**
 * One or more variables that together maintain an initially zero {@code long} sum. When updates
 * (method {@link #add}) are contended across threads, the set of variables may grow dynamically to
 * reduce contention. Method {@link #sum} (or, equivalently, {@link #longValue}) returns the current
 * total combined across the variables maintaining the sum.
 *

/**
 * One or more variables that together maintain an initially zero {@code long} sum. When updates
 * (method {@link #add}) are contended across threads, the set of variables may grow dynamically to
 * reduce contention. Method {@link #sum} (or, equivalently, {@link #longValue}) returns the current
 * total combined across the variables maintaining the sum.
 *