have to make sure that it's a single process running those previous steps *even* if afterwards, you start **multiple processes** (multiple workers) for the application itself. If those steps were run by **multiple processes**, they would **duplicate** the work by running it in **parallel**, and if the steps were something delicate like a database migration, they could cause conflicts with each other.

Of course, there are some cases where there's no problem in running the previous steps multiple...

      assertThat(queueSize).isBetween(0L, messageCount * message.size)
      // Expect to fail before enqueueing 32 MiB.
      assertThat(messageCount).isLessThan(32L)
    }

    // Confirm all sent messages were received, followed by a client-initiated close.
    val server = serverListener.assertOpen()
    for (i in 0 until messageCount) {
      serverListener.assertBinaryMessage(message)
    }

  }

  /**
   * Returns an immutable multimap containing the specified entries. The returned multimap iterates
   * over keys in the order they were first encountered in the input, and the values for each key
   * are iterated in the order they were encountered. If two values for the same key are {@linkplain
   * Object#equals equal}, the first value encountered is used.
   *

    * Old behavior:
      * environment variables explicitly whitelisted via --docker-env-metadata-whitelist were exported as `container_env_*=*`. Default is zero so by default non were exported
      * all docker labels were exported as `container_label_*=*`
    * New behavior:
      * Only `container_name`, `pod_name`, `namespace`, `id`, `image`, and `name` labels are exposed

     * - Temporarily change this test:
     * --- Comment out assertEquals.
     * --- System.out.println(input + " " + simplifyPath(input));
     * --- fail(). (If the test were to pass, its output would be hidden.)
     * - Run the test.
     * - Pull the relevant lines of output from the test into a testcases file.
     * - Test the output:
     * --- cat testcases | while read L; do

 *       to close resources will be thrown from the finally block. The throwable from the try block
 *       will be thrown.
 *   <li>If no exceptions or errors were thrown in the try block, the <i>first</i> exception thrown
 *       by an attempt to close a resource will be thrown.
 *   <li>Any exception caught when attempting to close a resource that is <i>not</i> thrown (because

     * - Temporarily change this test:
     * --- Comment out assertEquals.
     * --- System.out.println(input + " " + simplifyPath(input));
     * --- fail(). (If the test were to pass, its output would be hidden.)
     * - Run the test.
     * - Pull the relevant lines of output from the test into a testcases file.
     * - Test the output:
     * --- cat testcases | while read L; do

# Classes as Dependencies

Before diving deeper into the **Dependency Injection** system, let's upgrade the previous example.

## A `dict` from the previous example

In the previous example, we were returning a `dict` from our dependency ("dependable"):

//// tab | Python 3.10+

```Python hl_lines="9"
{!> ../../docs_src/dependencies/tutorial001_an_py310.py!}
```

////

//// tab | Python 3.9+

```Python hl_lines="11"

    fun status(status: String) =
      apply {
        this.status = status
      }

    /**
     * Removes all HTTP headers including any "Content-Length" and "Transfer-encoding" headers that
     * were added by default.
     */
    fun clearHeaders() =
      apply {
        headers = Headers.Builder()
      }

    /**

	// the regex used to parse out the pod UID and container ID from a
	// cgroup name. It assumes that any ".scope" suffix has been trimmed off
	// beforehand.  CAUTION: we used to verify that the pod and container id were
	// descendants of a kubepods directory, however, as of Kubernetes 1.21, cgroups
	// namespaces are in use and therefore we can no longer discern if that is the