variables from your model are also going to be untrusted. That means that if
your code interacts with the filesystem, network, etc. and uses checkpointed
variables as part of those interactions (ex: using a string variable to build a
filesystem path), a maliciously created checkpoint might be able to change the
targets of those operations, which could result in arbitrary
read/write/executions.

### Running a TensorFlow server

// Help template for caching feature.
var (
	defaultHelpPostfix = func(key string) string {
		return config.DefaultHelpPostfix(DefaultKVS, key)
	}

	// Help provides help for config values
	Help = config.HelpKVS{
		config.HelpKV{
			Key:         Bitrot,
			Description: `perform bitrot scan on drives when checking objects during scanner` + defaultHelpPostfix(Bitrot),
			Optional:    true,
			Type:        "on|off",
		},

The main logic with ADRs is to describe (architectural) decisions made:

* To provide best practices and solutions we (as the *build tool* team) want to promote.
* To avoid asking the same thing multiple times during code review.
* To explain *rejected solutions*, for now, and future development, in case they are proposed again.

ADRs can be written by any team.
Like code, they should be reviewed by any other relevant teams.

  })
  private int size;

  // "" means no fixed seed
  @Param("1234")
  private SpecialRandom random;

  @Param({"ImmutableSetImpl", "HashSetImpl"})
  private SetImpl impl;

  // the following must be set during setUp
  private CollectionBenchmarkSampleData sampleData;

  @BeforeExperiment
  void setUp() {
    sampleData = new CollectionBenchmarkSampleData(true, random, 0.8, size);
  }

  @Benchmark

@Experimental
public interface ModelTransformer {

    /**
     * Apply a transformation on the file model.
     *
     * @param model the input model
     * @param problems the problem collector to report any issues encountered during transformation
     * @return the transformed model, or the input model if no transformation is needed
     */
    @Nonnull
    Model transform(

     * @throws InvokerException if an error occurs during the invocation process
     */
    int invoke(@Nonnull R invokerRequest) throws InvokerException;

    /**
     * Closes and disposes of this {@link Invoker} instance, releasing any resources it may hold.
     * This method is called automatically when using try-with-resources statements.
     *

     */
    public ParserException(String message) {
        super(message);
    }

    /**
     * Constructs a new ParserException with the specified detail message and cause.
     *
     * @param message the detail message explaining the cause of the exception
     * @param cause the underlying cause of the exception
     */
    public ParserException(String message, Throwable cause) {
        super(message, cause);

        } else {
            trackingDir = event.getFile().getParentFile().toPath().resolve(".tracking");
        }

        String baseName;
        String ext = missing ? ".miss" : ".dep";
        Path trackingFile = null;

        String indent = "";
        ArrayList<String> trackingData = new ArrayList<>();

        if (collectStepTrace == null && plugin != null) {
            ext = ".plugin";

   */
  SipHashFunction(int c, int d, long k0, long k1) {
    checkArgument(
        c > 0, "The number of SipRound iterations (c=%s) during Compression must be positive.", c);
    checkArgument(
        d > 0, "The number of SipRound iterations (d=%s) during Finalization must be positive.", d);
    this.c = c;
    this.d = d;
    this.k0 = k0;
    this.k1 = k1;
  }

  @Override
  public int bits() {

 */
@Deprecated(since = "4.0.0")
public interface ModelCache {

    <T> T computeIfAbsent(String groupId, String artifactId, String version, String tag, Supplier<T> data);

    <T> T computeIfAbsent(Source path, String tag, Supplier<T> data);