### Containers with Multiple Processes and Special Cases { #containers-with-multiple-processes-and-special-cases }

Of course, there are **special cases** where you could want to have **a container** with several **Uvicorn worker processes** inside.

In those cases, you can use the `--workers` command line option to set the number of workers that you want to run:

        // These test cases ensure that the new java.util.Base64 implementation
        // produces the same output as the old custom implementation

        // Test case 1: Standard padding
        final byte[] data1 = "abc".getBytes();
        assertEquals("YWJj", Base64Util.encode(data1));

        // Test case 2: Single padding
        final byte[] data2 = "abcd".getBytes();

    /**
     * {@return the dependency scope for the given identifier, or {@code null} if none}.
     * The identifiers are usually in lower cases with {@code '-'} instead of {@code '_'}
     * as word separator.
     *
     * @param id the identifier of the scope (case-sensitive)
     */
    public static DependencyScope forId(String id) {
        return IDS.get(id);
    }

    private final String id;

        } catch (Exception // sneaky checked exception
            | Error atomicReferenceFieldUpdaterFailure) {
          // Some Android 5.0.x Samsung devices have bugs in JDK reflection APIs that cause
          // getDeclaredField to throw a NoSuchFieldException when the field is definitely there.
          // For these users fallback to a suboptimal implementation, based on synchronized. This

{* ../../docs_src/response_model/tutorial001_01_py310.py hl[16,21] *}

Comme cela se passe côté Rust, la performance sera bien meilleure que si cela était fait avec le Python classique et la classe `JSONResponse`.

Lorsque vous utilisez un `response_model` ou un type de retour, FastAPI n'utilise ni le `jsonable_encoder` pour convertir les données (ce qui serait plus lent) ni la classe `JSONResponse`.

   * (closed) or exclusive (open).
   *
   * @since 14.0
   */
  public static <C extends Comparable<?>> Range<C> upTo(C endpoint, BoundType boundType) {
    switch (boundType) {
      case OPEN:
        return lessThan(endpoint);
      case CLOSED:
        return atMost(endpoint);
    }
    throw new AssertionError();
  }

  /**
   * Returns a range that contains all values strictly greater than {@code endpoint}.

     * it'll either be a CACHE_MISS, in which case we
     * simply give up, or it'll be a Method, in which
     * case, we return it.</p>
     * <p>If nothing is found, then we must actually go
     * and introspect the method from the MethodMap.</p>
     * @param name Method name.
     * @param params Method parameters.
     * @return The found method.
     * @throws MethodMap.AmbiguousException in case of duplicate methods.
     */

{* ../../docs_src/query_params_str_validations/tutorial014_an_py310.py hl[10] *}

## Custom Validation { #custom-validation }

There could be cases where you need to do some **custom validation** that can't be done with the parameters shown above.

In those cases, you can use a **custom validator function** that is applied after the normal validation (e.g. after validating that the value is a `str`).

				// In this case we will just collapse this into "eq" for all use cases.
				parsedPolicy.Conditions.Policies = append(parsedPolicy.Conditions.Policies, struct {
					Operator string
					Key      string
					Value    string
				}{
					policyCondEqual, "$" + strings.ToLower(k), toString(v),
				})
			}
		case []any: // Handle array types.

     * Covers all defined error classes and codes, including default cases.
     */
    private static Stream<Arguments> provideErrorClassAndCodeForGetErrorString() {
        return Stream.of(
                // SUCCESS
                Arguments.of(NbtException.SUCCESS, 0, "SUCCESS"),

                // ERR_NAM_SRVC
                // Note: FMT_ERR case has a bug in NbtException.getErrorString() - missing break statement causes fallthrough