// don't execute 'immediately'.  By checking isDone here we avoid that.
    // A corollary to all that is that we don't need to check isDone inside the loop because if we
    // get into the loop we know that we weren't done when we entered and therefore we aren't under
    // an obligation to execute 'immediately'.
    if (!isDone()) {
      Listener oldHead = listeners;

 *  Breaking: Drop support for Kotlin Multiplatform.

    We planned to support multiplatform in OkHttp 5.0, but after building it, we weren't happy with
    the implementation trade-offs. We can't use our HTTP client engine on Kotlin/JS, and we weren't
    prepared to build a TLS API for Kotlin/Native.

    We'd prefer a multiplatform HTTP client API that's backed by OkHttp on Android and JVM, and

	m, fds := readRequest(t, ztunClient)
	// we got am essage from ztun, so it should have observed us being connected
	mt.Assert(ztunnelConnected.Name(), nil, monitortest.Exactly(1))

	// we should get the fd to dev null. note that we can't assert the fd number
	// as the kernel may have given us a different number that refers to the same file.
	assert.Equal(t, len(fds), 1)
	// in theory we should close the fd, but it's just a test..

Each instance of the `SessionLocal` class will be a database session. The class itself is not a database session yet.

But once we create an instance of the `SessionLocal` class, this instance will be the actual database session.

We name it `SessionLocal` to distinguish it from the `Session` we are importing from SQLAlchemy.

We will use `Session` (the one imported from SQLAlchemy) later.

	}
	return nil
}

func realDependencies() *dep.RealDependencies {
	return &dep.RealDependencies{
		CNIMode:          false, // we are in cni, but as we do the netns ourselves, we should keep this as false.
		NetworkNamespace: "",
	}
}

// Remove pod from mesh: pod is not deleted, we just want to remove it from the mesh.
func (s *NetServer) RemovePodFromMesh(ctx context.Context, pod *corev1.Pod) error {

	// RoleARN parameter processing: If a role ARN is given in the request, we
	// use that and validate the authentication request. If not, we assume this
	// is an STS request for a claim based IDP (if one is present) and set
	// roleArn = openid.DummyRoleARN.
	//
	// Currently, we do not support multiple claim based IDPs, as there is no

We are going to include this `APIRouter` in the main `FastAPI` app, but first, let's check the dependencies and another `APIRouter`.

## Dependencies

We see that we are going to need some dependencies used in several places of the application.

So we put them in their own `dependencies` module (`app/dependencies.py`).

Complex logic ahead...
We have three sets of values, in order of precedence (last wins):
1. The builtin values.yaml defaults
2. The profile the user selects
3. Users input (-f or --set)

Unfortunately, Helm provides us (1) and (3) together (as .Values), making it hard to insert (2).

However, we can workaround this by placing all of (1) under a specific key (.Values.defaults).

//
// Below here we do the gradient algorithm. It works as follows:
//
// First we filter the tape to just the subset of operations we want to
// differentiate. In the process of doing so we count how many times each Tensor
// is used as an input to an op (so we know when we're done computing gradients
// for that Tensor). We also count, for each tape entry, how many of its output

      if (retryRoute != null) {
        // Lock in the route because retryRoute() is racy and we don't want to call it twice.
        nextRouteToTry = retryRoute
        return true
      }
    }

    // If we have a routes left, use 'em.
    if (routeSelection?.hasNext() == true) return true

    // If we haven't initialized the route selector yet, assume it'll have at least one route.