throw e.getCause();
    }
  }

  /**
   * Verify that the listener completes in a reasonable amount of time, and Asserts that the future
   * throws an {@code ExecutableException} and that the cause of the {@code ExecutableException} is
   * {@code expectedCause}.
   */
  public void assertException(Throwable expectedCause) throws Exception {
    // Verify that the listener executed in a reasonable amount of time.

    // Run a task on the composed Executor that interrupts its thread (i.e. this thread).
    fakePool.execute(
        new Runnable() {
          @Override
          public void run() {
            Thread.currentThread().interrupt();
          }
        });
    // Run a task that expects that it is not interrupted while it is running.
    e.execute(
        new Runnable() {

   * even to an intermediate buffer) should be considerably more efficient than potentially
   * copying the CharSequence to a String and then calling getBytes(Charset) on that String, in
   * reality there are optimizations that make the getBytes(Charset) approach considerably faster,
   * at least for commonly used charsets like UTF-8.
   */

  @Override

      assertWithMessage("quantile at index " + index)
          .that(percentiles().index(index).computeInPlace(dataset))
          .isWithin(ALLOWED_ERROR)
          .of(expectedLargeDatasetPercentile(index));
    }

    // Assert that the dataset contains the same elements after the in-place computation (although
    // they may be reordered). We only do this for one index rather than for all indexes, as it is

      assertWithMessage("quantile at index " + index)
          .that(percentiles().index(index).computeInPlace(dataset))
          .isWithin(ALLOWED_ERROR)
          .of(expectedLargeDatasetPercentile(index));
    }

    // Assert that the dataset contains the same elements after the in-place computation (although
    // they may be reordered). We only do this for one index rather than for all indexes, as it is

   * is transmitted. If you cannot ensure that your client and server both support HTTP/2, do not
   * use this feature.
   *
   * ### Duplex APIs
   *
   * With regular request bodies it is not legal to write bytes to the sink passed to
   * [RequestBody.writeTo] after that method returns. For duplex requests bodies that condition is

  /**
   * Indicates that a collection disallows certain elements (other than {@code null}, whose validity
   * as an element is indicated by the presence or absence of {@link #ALLOWS_NULL_VALUES}). From the
   * documentation for {@link Collection}:
   *
   * <blockquote>
   *
   * "Some collection implementations have restrictions on the elements that they may contain. For

    }

    /**
     * Configures this {@code Builder} to order entries by value according to the specified
     * comparator.
     *
     * <p>The sort order is stable, that is, if two entries have values that compare as equivalent,
     * the entry that was inserted first will be first in the built map's iteration order.
     *
     * @throws IllegalStateException if this method was already called
     * @since 19.0
     */

   * necessary) that that pivot value appears at the start of the slice i.e. at {@code from}.
   * Expects that {@code from} is strictly less than {@code to}.
   */
  private static void movePivotToStartOfSlice(double[] array, int from, int to) {
    int mid = (from + to) >>> 1;
    // We want to make a swap such that either array[to] <= array[from] <= array[mid], or

If there's no `gzip` in the header, it will not try to decompress the body.

That way, the same route class can handle gzip compressed or uncompressed requests.

{* ../../docs_src/custom_request_and_route/tutorial001.py hl[8:15] *}

### Create a custom `GzipRoute` class { #create-a-custom-gziproute-class }

Next, we create a custom subclass of `fastapi.routing.APIRoute` that will make use of the `GzipRequest`.