*/
  @JvmField val minimumConcurrentCalls: Int = 0,
  /** How long to wait to retry pre-emptive connection attempts that fail. */
  @JvmField val backoffDelayMillis: Long = 60 * 1000,
  /** How much jitter to introduce in connection retry backoff delays */
  @JvmField val backoffJitterMillis: Int = 100,

used to test the individual components in the lifecycle with data that has expected characteristics
and can be asserted as desired.

You can change/extend these stubs, and tests should not be breaking too much, since most tests
assert using expected values from the stubs. Normally, when you try to use data from the stubs that
have not been properly populated, you'll get a nullpointer in your test and you then have to

        MavenExecutionPlan plan = LifecycleExecutionPlanCalculatorStub.getProjectAExecutionPlan();

        assertNull(plan.findLastInPhase("pacXkage"));
        // Beer comes straight after package in stub, much like real life.
        assertNotNull(plan.findLastInPhase(LifecycleExecutionPlanCalculatorStub.INITIALIZE.getPhase()));
    }

        token_url, data=data, verify=False,
        allow_redirects=False, auth=(client_id, client_secret))

    print('body: ' + id_token_response.text)

    # we can now use the id_token as much as we want to access protected resources.
    tokens = json.loads(id_token_response.text)
    id_token = tokens['id_token']

    response = sts_client.assume_role_with_web_identity(

import org.eclipse.aether.deployment.DeployRequest;
import org.eclipse.aether.installation.InstallRequest;

/**
 * Maven POM transformer.
 * TODO: rename this interface to "Transformer" as it can do much more than just transform POM.
 * @since TBD
 */
public interface PomArtifactTransformer {
    InstallRequest remapInstallArtifacts(RepositorySystemSession session, InstallRequest request);

		// Unlock the 1st lock; ref=1 after this line
		nsLk.unlock("volume", "path", false)

		// Taking another lockMapMutex here allows queuing up additional lockers. This should
		// not be required but makes reproduction much easier.
		nsLk.lockMapMutex.Lock()

		// lk3 blocks.
		lk3ch := make(chan bool)
		go func() {
			lk3ch <- nsLk.lock(ctx, "volume", "path", "source", "opsID", false, 0)
		}()

		// lk4, blocks.

 * </p>
 * <p>
 * Wise voices have suggested that maybe aggregators shouldn't be bound to the ordinary
 * lifecycle at all, in which case we wouldn't be needing this class at all ( and
 * ProjectBuildList.getByTaskSegments). Or maybe they should be introduced in the calculation
 * of the execution plan instead, which seems much nicer.
 * </p>
 * <p>

### Summary { #summary }

I used to say I didn't like history that much... and look at me now giving "tech history" lessons. 😅

</dependencies>
    ]]></configuration>
    <description>
      <![CDATA[
      This is a set of Artifact instances resolved from the set of dependencies for
      the current project.

      NOTE: This will likely contain much more than the direct dependencies of the
      current POM, since Maven uses transitive, or chained, dependency resolution.
    ]]></description>
  </expression>
  <expression>
    <syntax>project.parentArtifact</syntax>

import java.math.BigInteger;
import java.util.Random;
import org.jspecify.annotations.NullUnmarked;

/**
 * Utilities for benchmarks.
 *
 * <p>In many cases, we wish to vary the order of magnitude of the input as much as we want to vary
 * the input itself, so most methods which generate values use an exponential distribution varying
 * the order of magnitude of the generated values uniformly at random.
 *
 * @author Louis Wasserman
 */