*
   * @since 20.0
   */
  public static final MediaType WAX_AUDIO = createConstant(AUDIO_TYPE, "x-ms-wax");

  /**
   * Real Audio. For more information, see <a
   * href="http://service.real.com/help/faq/rp8/configrp8win.html">this link</a>.
   *
   * @since 20.0
   */
  public static final MediaType VND_REAL_AUDIO = createConstant(AUDIO_TYPE, "vnd.rn-realaudio");

  /**

## Dependencies

### Added
_Nothing has changed._

### Changed
- github.com/google/cel-go: [v0.12.5 → v0.12.6](https://github.com/google/cel-go/compare/v0.12.5...v0.12.6)
- k8s.io/system-validators: v1.7.0 → v1.8.0
- sigs.k8s.io/apiserver-network-proxy/konnectivity-client: v0.0.33 → v0.0.35

### Removed
_Nothing has changed._

- github.com/google/cadvisor: [v0.51.0 → v0.52.1](https://github.com/google/cadvisor/compare/v0.51.0...v0.52.1)
- github.com/google/cel-go: [v0.22.0 → v0.23.2](https://github.com/google/cel-go/compare/v0.22.0...v0.23.2)
- github.com/google/gnostic-models: [v0.6.8 → v0.6.9](https://github.com/google/gnostic-models/compare/v0.6.8...v0.6.9)

    *   tf.optimizers => tf.keras.optimizers
    *   Add tf.keras.layers.AbstractRNNCell as the preferred implementation of
        RNN cell for TF v2. User can use it to implement RNN cell with custom
        behavior.
    *   Adding `clear_losses` API to be able to clear losses at the end of
        forward pass in a custom training loop in eager.

And it generates OpenAPI schemas.

That's how it works in Flask, Starlette, Responder, etc.

But then, we have again the problem of having a micro-syntax, inside of a Python string (a big YAML).

The editor can't help much with that. And if we modify parameters or Marshmallow schemas and forget to also modify that YAML docstring, the generated schema would be obsolete.

/// info

APISpec was created by the same Marshmallow developers.

          // TODO(lukes): consider building everything in terms of ListenableScheduledFuture then
          // the AbstractService could monitor the future directly. Rescheduling is still hard...
          // but it would help with some of these lock ordering issues.
          scheduleFailure = e;
          toReturn = new FutureAsCancellable(immediateCancelledFuture());
        } finally {
          lock.unlock();
        }

    public UnmodifiableIterator<Entry<K, V>> iterator() {
      return multimap.entryIterator();
    }

    @Override
    boolean isPartialView() {
      return false;
    }

    // redeclare to help optimizers with b/310253115
    @SuppressWarnings("RedundantOverride")
    @Override
    @J2ktIncompatible
    @GwtIncompatible
        Object writeReplace() {
      return super.writeReplace();
    }
  }

    return from(Arrays.asList(elements));
  }

  /**
   * Construct a fluent iterable from another fluent iterable. This is obviously never necessary,
   * but is intended to help call out cases where one migration from {@code Iterable} to {@code
   * FluentIterable} has obviated the need to explicitly convert to a {@code FluentIterable}.
   *

    return from(Arrays.asList(elements));
  }

  /**
   * Construct a fluent iterable from another fluent iterable. This is obviously never necessary,
   * but is intended to help call out cases where one migration from {@code Iterable} to {@code
   * FluentIterable} has obviated the need to explicitly convert to a {@code FluentIterable}.
   *

Ví dụ:

* <a href="https://code.visualstudio.com/docs/python/environments#_select-and-activate-an-environment" class="external-link" target="_blank">VS Code</a>
* <a href="https://www.jetbrains.com/help/pycharm/creating-virtual-environment.html" class="external-link" target="_blank">PyCharm</a>

/// tip

Bạn thường chỉ cần làm điều này **một lần**, khi bạn tạo môi trường ảo.

///