public Object[] toArray() {
    /*
     * If we could, we'd declare the no-arg `Collection.toArray()` to return "Object[] but elements
     * have the same nullness as E." Since we can't, we declare it to return nullable elements, and
     * we can override it in our non-null-guaranteeing subtypes to present a better signature to
     * their users.
     *

# ADR-0008 - Use NullAway for null checking

## Date

2025-08-06

## Context

Our codebase uses `null` extensively to represent an absence of value or optionality of the argument.
We utilize nullness annotations in Java code and somewhat rely on IDE warnings to guide us.
However, the Java code we have is not fully annotated, which causes several consequences:
* IDE warnings can be misleading

          ? NullnessAnnotationReader.FROM_DECLARATION_AND_TYPE_USE_ANNOTATIONS
          : NullnessAnnotationReader.FROM_DECLARATION_ANNOTATIONS_ONLY;

  /**
   * Looks for declaration nullness annotations and, if supported, type-use nullness annotations.
   *
   * <p>Under Android VMs, the methods for retrieving type-use annotations don't exist. This means

      return (delegate != null)
          ? delegate.keySet().toArray()
          : ObjectArrays.copyAsObjectArray(requireKeys(), 0, size);
    }

    @Override
    @SuppressWarnings("nullness") // b/192354773 in our checker affects toArray declarations
    public <T extends @Nullable Object> T[] toArray(T[] a) {
      if (needsAllocArrays()) {
        if (a.length > 0) {

   * but we know that our values will never be `null`.
   */

  @Override
  public Object[] toArray() {
    return snapshotElementsToList().toArray();
  }

  @Override
  @SuppressWarnings("nullness") // b/192354773 in our checker affects toArray declarations
  public <T extends @Nullable Object> T[] toArray(T[] array) {
    return snapshotElementsToList().toArray(array);
  }

  /*

   * require its own suppression, since it is also unsound. toArray(T[]) is just a mess from a
   * nullness perspective. The signature below at least has the virtue of being relatively simple.
   */
  @SuppressWarnings("nullness")
  public final <T extends @Nullable Object> T[] toArray(T[] other) {
    checkNotNull(other);
    int size = size();

    if (other.length < size) {

   * Verifies that the specified expected result is always produced by collecting the specified
   * inputs, regardless of how the elements are divided.
   */
  @SafeVarargs
  @CanIgnoreReturnValue
  @SuppressWarnings("nullness") // TODO(cpovirk): Remove after we fix whatever the bug is.
  public final CollectorTester<T, A, R> expectCollects(R expectedResult, T... inputs) {
    List<T> list = Arrays.asList(inputs);

    }

    @Override
    public E intern(E sample) {
      while (true) {
        // trying to read the canonical...
        @SuppressWarnings("rawtypes") // using raw types to avoid a bug in our nullness checker :(
        InternalEntry entry = map.getEntry(sample);
        if (entry != null) {
          Object canonical = entry.getKey();
          if (canonical != null) { // only matters if weak/soft keys are used

    GraphConnections<N, V> connectionsU = nodeConnections.get(nodeU);
    V value = (connectionsU == null) ? null : connectionsU.value(nodeV);
    // TODO(b/192579700): Use a ternary once it no longer confuses our nullness checker.
    if (value == null) {
      return defaultValue;
    } else {
      return value;
    }
  }

 * in com.google.guava:guava differs from the "frozen" copy in the listenablefuture artifact. This
 * could in principle cause problems for some users. Still, we expect that the benefits of the
 * nullness annotations in particular will outweigh the costs. (And it's worth noting that we have
 * released multiple ListenableFuture.class files that are not byte-for-byte compatible even from