Using these ideas, JWT can be used for way more sophisticated scenarios.

In those cases, several of those entities could have the same ID, let's say `foo` (a user `foo`, a car `foo`, and a blog post `foo`).

				time.Unix(0, 3).UTC(),
				time.Unix(0, 3).UTC(),
				time.Unix(0, 3).UTC(),
			},
			time.Unix(0, 3).UTC(),
			4,
		},
		{
			// 3. Tests common time obtained when elements have a mixture of
			// sentinel values and don't have read quorum on any of the values.
			[]time.Time{
				time.Unix(0, 3).UTC(),
				time.Unix(0, 3).UTC(),
				time.Unix(0, 2).UTC(),
				time.Unix(0, 1).UTC(),
				time.Unix(0, 3).UTC(),

            continue@eachQueue
          }

          // If we already have more than one task, that's enough work for now. Stop searching.
          readyTask != null -> {
            multipleReadyTasks = true
            break@eachQueue
          }

          // We have a task to execute when we complete the loop.
          else -> {
            readyTask = candidate
          }

And `threading.local` is not compatible with the new async features of modern Python.

!!! note "Technical Details"
    `threading.local` is used to have a "magic" variable that has a different value for each thread.

  * In practice, this largely means that references are fully qualified in the API. IP Addresses (generally) have a network associated with them, node names have a cluster associated with them, etc.

See the [XDS Evolution](https://docs.google.com/document/d/1V5wkeBHbLSLMzAMbwFlFZNHdZPyUEspG4lHbnB0UaCg/edit) document for more history and details.

// scanners.
func (s *Scanner) Scan() bool {
	if s.done {
		return false
	}
	s.scanCalled = true
	// Loop until we have a token.
	for {
		// See if we can get a token with what we already have.
		// If we've run out of data but have an error, give the split function
		// a chance to recover any remaining, possibly empty token.
		if s.end > s.start || s.err != nil {

  private final transient int size;

  /*
   * We have some considerable complexity in these create methods because of
   * Builder.buildKeepingLast(). The same Builder might be called with buildKeepingLast() and then
   * buildOrThrow(), or vice versa. So in particular, if we modify alternatingKeysAndValues to
   * eliminate duplicate keys (for buildKeepingLast()) then we have to ensure that a later call to

!!! danger
    For simplicity, here we are just adding the scopes received directly to the token.

    But in your application, for security, you should make sure you only add the scopes that the user is actually able to have, or the ones you have predefined.

=== "Python 3.10+"

    ```Python hl_lines="155"
    {!> ../../../docs_src/security/tutorial005_an_py310.py!}
    ```

=== "Python 3.9+"

    ```Python hl_lines="155"

    ```

This will make FastAPI skip the response model generation and that way you can have any return type annotations you need without it affecting your FastAPI application. 🤓

## Response Model encoding parameters

Your response model could have default values, like:

=== "Python 3.10+"

    ```Python hl_lines="9  11-12"

    # then two meshes must have different Mesh IDs:
    # - Meshes will have their telemetry aggregated in one place
    # - Meshes will be federated together
    # - Policy will be written referencing one mesh from the other
    #
    # If an administrator expects that any of these conditions may become true in
    # the future, they should ensure their meshes have different Mesh IDs
    # assigned.
    #