func splitTrace(parsed *parsedTrace) ([]traceviewer.Range, error) {
	// TODO(mknyszek): Split traces by generation by doing a quick first pass over the
	// trace to identify all the generation boundaries.
	s, c := traceviewer.SplittingTraceConsumer(100 << 20) // 100 MiB
	if err := generateTrace(parsed, defaultGenOpts(), c); err != nil {
		return nil, err
	}
	return s.Ranges, nil
}

const DirectioAlignSize = 4096

// CopyAligned - copies from reader to writer using the aligned input
// buffer, it is expected that input buffer is page aligned to
// 4K page boundaries. Without passing aligned buffer may cause
// this function to return error.
//
// This code is similar in spirit to io.Copy but it is only to be
// used with DIRECT I/O based file descriptor and it is expected that

====

That's all.
If you run your application, by executing the `run` task or through a <<#sec:the_distribution,generated start script>>, it will run as module and respect module boundaries at runtime.
For example, reflective access to an internal package from another module can fail.

The configured _main class_ is also baked into the `module-info.class` file of your application Jar.

However, there are some things to consider depending on whether you need module information to be available, and module boundaries to be enforced, during test execution.
In this context, the terms _whitebox testing_ (module boundaries are deactivated or relaxed) and _blackbox testing_ (module boundaries are in place) are often used.
Whitebox testing is used/needed for unit testing and blackbox testing fits functional or integration test requirements.

		return _EXCEPTION_CONTINUE_SEARCH_SEH
	}
	// The Windows SEH machinery will unwind the stack until it finds
	// a frame with a handler for the exception or until the frame is
	// outside the stack boundaries, in which case it will call the
	// UnhandledExceptionFilter. Unfortunately, it doesn't know about
	// the goroutine stack, so it will stop unwinding when it reaches the

   * is unnecessary; for example, {@code copyOf(copyOf(anArrayList))} will copy the data only once.
   * This reduces the expense of habitually making defensive copies at API boundaries. However, the
   * precise conditions for skipping the copy operation are undefined.
   *
   * @throws NullPointerException if any of {@code elements} is null
   * @since 7.0 (source-compatible since 2.0)
   */

Also consider that the more parts of your build logic rely on the dynamic aspects of Groovy, the harder they will be to use from the Kotlin DSL.
You'll find recipes on how to cross the dynamic boundaries from static Kotlin in the <<kotlin_dsl.adoc#sec:interoperability,interoperability section of the Gradle Kotlin DSL documentation>>, regardless of where the dynamic Groovy build logic resides.

   * limits, this call may be executed well before processing the request is able to begin.
   *
   * This will be invoked only once for a single [Call]. Retries of different routes or redirects
   * will be handled within the boundaries of a single [callStart] and [callEnd]/[callFailed] pair.
   */
  open fun callStart(call: Call) {
  }

  /**
   * Invoked prior to a proxy selection.
   *

		for _, i := range instances {
			// Insert all instances for this IP for services within the same namespace. This ensures we
			// match Kubernetes logic where Services do not cross namespace boundaries and avoids
			// possibility of other namespaces inserting service instances into namespaces they do not
			// control.
			if node.Metadata.Namespace == "" || i.Service.Attributes.Namespace == node.Metadata.Namespace {

	// ref: https://github.com/kubernetes/community/blob/master/sig-scalability/slos/pod_startup_latency.md
	//
	// The histogram bucket boundaries for pod startup latency metrics, measured in seconds. These are hand-picked
	// so as to be roughly exponential but still round numbers in everyday units. This is to minimise the number