gradleInvocation.implicitLauncherJvmArgs.clear();
    }

    /**
     * Returns the JVM opts that should be used to start a forked JVM.
     */
    private void calculateLauncherJvmArgs(GradleInvocation gradleInvocation) {
        // Add JVM args that were explicitly requested
        gradleInvocation.launcherJvmArgs.addAll(commandLineJvmOpts);

            setActiveProfiles((Collections.unmodifiableList(project.getActiveProfiles())));
        }

        if (project.getAttachedArtifacts() != null) {
            // clone properties modifiable by plugins in a forked lifecycle
            setAttachedArtifacts(new ArrayList<>(project.getAttachedArtifacts()));
        }

        if (project.getCompileSourceRoots() != null) {
            // clone source roots

If your tests do share resources, they could interfere with each other in random and unpredictable ways.

==== Fork tests into multiple processes

By default, Gradle runs all tests in a single forked VM.
If there are a lot of tests, or some tests that consume lots of memory,
your tests may take longer than you expect to run. You can increase the
heap size, but garbage collection may slow down your tests.

==== Improved classpath separation for worker processes

The system classpath for worker daemons started by the <<worker_api.adoc#worker_api, Worker API>> when using `PROCESS` isolation has been reduced to a minimum set of Gradle infrastructure.
User code is still segregated into a separate classloader to isolate it from the Gradle runtime.

[[sec:test_execution]]
== Test execution

Gradle executes tests in a separate ('forked') JVM, isolated from the main build process. This prevents classpath pollution and excessive memory consumption for the build process. It also allows you to run the tests with different JVM arguments than the build is using.

[[rel5.0:worker_api]]
=== [5.0] Worker API: working directory of a worker can no longer be set

Since JDK 11 no longer supports changing the working directory of a running process, setting the working directory of a worker via its fork options is now prohibited.

All workers now use the same working directory to enable reuse.

	workerSharedMemSize = 100 << 20 // 100 MB
)

// worker manages a worker process running a test binary. The worker object
// exists only in the coordinator (the process started by 'go test -fuzz').
// workerClient is used by the coordinator to send RPCs to the worker process,
// which handles them with workerServer.
type worker struct {
	dir     string   // working directory, same as package directory

	// At this point, the pod worker is aware of which pods are not desired (SyncKnownPods).
	// We now look through the set of active pods for those that the pod worker is not aware of
	// and deliver an update. The most common reason a pod is not known is because the pod was
	// deleted and recreated with the same UID while the pod worker was driving its lifecycle (very

	"fmt"
	"io"
	"os"
	"sync"

	"github.com/fsnotify/fsnotify"
)

type worker struct {
	mu sync.RWMutex

	// watcher is an fsnotify watcher that watches the parent
	// dir of watchedFiles.
	dirWatcher *fsnotify.Watcher

	// The worker maintains a map of channels keyed by watched file path.
	// The worker watches parent path of given path,
	// and filters out events of given path, then redirect

implementation-class=com.example.worker.WorkerPlugin...