* Missing null checks may cause NullPointerException
* Bridging with Kotlin code is polluted with platform types
* Some public APIs have incorrect nullability annotations (both overly restrictive and overly permissive)

While there are competing ways to represent absence (`Optional`, "Null Object" pattern, method overloads),
it is unlikely that we'll be able to remove `null` entirely.

# Print a stacktrace when a test is killed
test --test_env="GTEST_INSTALL_FAILURE_SIGNAL_HANDLER=1"


# Default options should come above this line.

# Android configs. Bazel needs to have --cpu and --fat_apk_cpu both set to the
# target CPU to build transient dependencies correctly. See
# https://docs.bazel.build/versions/master/user-manual.html#flag--fat_apk_cpu
common:android --config=clang_local

            //
            // So that is why we cancel only in the case of CAS success.
            taskFuture.cancel(false);
          }
        };
    // Adding the listener to both futures guarantees that newFuture will always be set. Adding to
    // taskFuture guarantees completion if the callable is invoked, and adding to outputFuture
    // propagates cancellation if the callable has not yet been invoked.

    // aProject is whitelisted, it should not be returned as a downstream project for itself. bProject and cProject
    // are not whitelisted, so they should return their downstream projects, both have dProject and eProject as
    // downstream projects. Which would result in dProject and eProject being returned twice, but now the results are
    // made unique.
    @Test

// ARMMRCOffset implements the peculiar encoding of the MRC and MCR instructions.
// The difference between MRC and MCR is represented by a bit high in the word, not
// in the usual way by the opcode itself. Asm must use AMRC for both instructions, so
// we return the opcode for MRC so that asm doesn't need to import obj/arm.
func ARMMRCOffset(op obj.As, cond string, x0, x1, x2, x3, x4, x5 int64) (offset int64, op0 obj.As, ok bool) {
	op1 := int64(0)

[OkHttp 3.11][OkHttp311]
------------------------

_2018-07-12_

Added a new extra strict RESTRICTED_TLS configuration inspired by [Google Cloud’s similar policy][googlecloud_ssl_policy]. It is appropriate when both the host platform
(JVM/Conscrypt/Android) and target webserver are current.

##### RESTRICTED_TLS cipher suites

 * TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256
 * TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256

   * <p>If this graph is directed, {@code endpoints} must be ordered.
   *
   * <p><b>{@code edge} must be unique to this graph</b>, just as a {@code Map} key must be. It must
   * also be non-null.
   *
   * <p>If either or both endpoints are not already present in this graph, this method will silently
   * {@link #addNode(Object) add} each missing endpoint to the graph.
   *

import com.google.common.collect.ImmutableList;
import com.google.errorprone.annotations.CanIgnoreReturnValue;

/**
 * Standard implementation of {@link MutableNetwork} that supports both directed and undirected
 * graphs. Instances of this class should be constructed with {@link NetworkBuilder}.
 *
 * <p>Time complexities for mutation methods are all O(1) except for {@code removeNode(N node)},

   * @return the value that was previously associated with the key, or {@code null} if there was no
   *     previous entry. (If the bimap contains null values, then {@code forcePut}, like {@code
   *     put}, returns {@code null} both if the key is absent and if it is present with a null
   *     value.)
   */
  @CanIgnoreReturnValue
  @Nullable V forcePut(@ParametricNullness K key, @ParametricNullness V value);

  // Bulk Operations

  /**

  }

  /**
   * This is an extraordinary test case. Here's what it's trying to simulate.
   * - 2 requests happen concurrently to a host that can be coalesced onto a single connection.
   * - Both request discover no existing connection. They both make a connection.
   * - The first request "wins the race".
   * - The second request discovers it "lost the race" and closes the connection it just opened.