}
	return c, kyberSharedSecret(ss, c), nil
}

func kyberSharedSecret(K, c []byte) []byte {
	// Package mlkem768 implements ML-KEM, which compared to Kyber removed a
	// final hashing step. Compute SHAKE-256(K || SHA3-256(c), 32) to match Kyber.
	// See https://words.filippo.io/mlkem768/#bonus-track-using-a-ml-kem-implementation-as-kyber-v3.
	h := sha3.NewShake256()
	h.Write(K)
	ch := sha3.Sum256(c)
	h.Write(ch[:])

	HasSHA1   bool // K{I,L}MD-SHA-1 functions
	HasSHA256 bool // K{I,L}MD-SHA-256 functions
	HasSHA512 bool // K{I,L}MD-SHA-512 functions
	HasSHA3   bool // K{I,L}MD-SHA3-{224,256,384,512} and K{I,L}MD-SHAKE-{128,256} functions
	HasVX     bool // vector facility. Note: the runtime sets this when it processes auxv records.
	HasVXE    bool // vector-enhancements facility 1
	HasKDSA   bool // elliptic curve functions

	if r, ok := w.(*response); ok {
		return r.conn.rwc, true
	}
	return nil, false
}

func init() {
	// Set the default rstAvoidanceDelay to the minimum possible value to shake
	// out tests that unexpectedly depend on it. Such tests should use
	// runTimeSensitiveTest and SetRSTAvoidanceDelay to explicitly raise the delay
	// if needed.
	rstAvoidanceDelay = 1 * time.Nanosecond
}

	HasSHA1   bool // K{I,L}MD-SHA-1 functions
	HasSHA256 bool // K{I,L}MD-SHA-256 functions
	HasSHA512 bool // K{I,L}MD-SHA-512 functions
	HasSHA3   bool // K{I,L}MD-SHA3-{224,256,384,512} and K{I,L}MD-SHAKE-{128,256} functions
	HasVX     bool // vector facility
	HasVXE    bool // vector-enhancements facility 1
	_         CacheLinePad
}

func init() {
	archInit()
	initOptions()
	processOptions()
}

[[finding_problems]]
== Finding problems with task output caching

Below we describe a step-by-step process that should help shake out any problems with caching in your build.

=== Ensure incremental build works

First, make sure your build does the right thing without the cache.
Run a build twice without enabling the Gradle build cache.

        verifyHasChildren(whenReadyEvaluated, initScriptAppId, 'init', expectedGradleOps)
    }

    def 'no extra executions for composite builds'() {
        // This test does two things:
        // - shake out internal listener registration that isn't using InternalListener.
        //   There are a lost of listeners registered through the methods that we've decorated in the composite build code

				}
			}
		}()

		// Short-lived goroutines exercise different code paths (goroutines with
		// status _Gdead, for instance). This churn doesn't have behavior that
		// we can test directly, but does help to shake out data races.
		ready.Add(1)
		var churn func(i int)
		churn = func(i int) {
			SetGoroutineLabels(WithLabels(ctx, Labels(t.Name()+"-churn-i", fmt.Sprint(i))))
			if i == 0 {
				ready.Done()

// Verifies that constants generated post shape inference are propagated.
// get_shape result in this test.
module attributes {tf.versions = {producer = 179 : i32}} {

  // CHECK-LABEL: func @main
  func.func @main(%arg0: tensor<10x19xf32>, %arg1: tensor<19x10xf32> {mhlo.is_same_data_across_replicas = true}) -> tensor<?xi64> {
    %0 = "tf.Shape"(%arg0) : (tensor<10x19xf32>) -> tensor<2xi64>

// RUN: tf-opt -split-input-file -verify-diagnostics --tf-shape-inference %s | FileCheck %s

module attributes {tf.versions = {producer = 888 : i32}} {
// CHECK-LABEL: testConv2dShapeValidPadding
func.func @testConv2dShapeValidPadding(%arg0: tensor<1x112x80x128xf32>, %arg1: tensor<128x3x3x128xf32>, %arg2: tensor<128xf32>) -> tensor<1x?x?x128xf32> {

// RUN: tf-opt -split-input-file -verify-diagnostics -tf-tpu-annotate-dynamic-shape-inputs %s | FileCheck %s

// Test that annotate the inputs of the cluster func to be dynamic shaped.

module attributes {tf.devices = ["/job:worker/replica:0/task:0/device:CPU:0", "/job:worker/replica:0/task:0/device:TPU_SYSTEM:0", "/job:worker/replica:0/task:0/device:TPU:0"]} {
   func.func @main(
      %arg0: tensor<2048xi64> {tf.device = "/job:localhost/replica:0/task:0/device:CPU:0"},