- To add test cases for the new code. If you have questions about how to do it, please ask on our [Slack](https://slack.min.io) channel.
- To run `make verifiers`
- To squash your commits into a single commit. `git rebase -i`. It's okay to force update your pull request.
- To run `make test` and `make build` completes.

### Commit changes

) {
    QUICK_FEEDBACK_LINUX_ONLY(
        "Quick Feedback - Linux Only",
        "Run checks and functional tests (embedded executer, Linux)",
        "QuickFeedbackLinuxOnly",
    ),
    QUICK_FEEDBACK("Quick Feedback", "Run checks and functional tests (embedded executer, Windows)", "QuickFeedback"),
    PULL_REQUEST_FEEDBACK("Pull Request Feedback", "Run various functional tests", "PullRequestFeedback"),
    READY_FOR_NIGHTLY(

  private var nextQueueName = 10000
  private var coordinatorWaiting = false
  private var coordinatorWakeUpAt = 0L

  /**
   * When we need a new thread to run tasks, we call [Backend.execute]. A few microseconds later we
   * expect a newly-started thread to call [Runnable.run]. We shouldn't request new threads until
   * the already-requested ones are in service, otherwise we might create more threads than we need.
   *

# "-devYYYYMMDD" to the end of the version string for nightly builds.
if [[ "$TFCI_NIGHTLY_UPDATE_VERSION_ENABLE" == 1 ]]; then
  export TF_VER_FULL="$(tfrun bazel run //tensorflow/tools/ci_build:calculate_full_version -- --wheel-type nightly)"
else
  export TF_VER_FULL="$(tfrun bazel run //tensorflow/tools/ci_build:calculate_full_version -- --wheel-type release)"
fi

# Note on gsutil commands:

@NullMarked
public interface ListenableFuture<V extends @Nullable Object> extends Future<V> {
  /**
   * Registers a listener to be {@linkplain Executor#execute(Runnable) run} on the given executor.
   * The listener will run when the {@code Future}'s computation is {@linkplain Future#isDone()
   * complete} or, if the computation is already complete, immediately.
   *

  }

  private static class SingleCallListener implements Runnable {

    private boolean expectCall = false;
    private final CountDownLatch calledCountDown = new CountDownLatch(1);

    @Override
    public void run() {
      assertTrue("Listener called before it was expected", expectCall);
      assertFalse("Listener called more than once", wasCalled());
      calledCountDown.countDown();
    }

    void expectCall() {

## Quick Start

import okhttp3.internal.threadFactory
import okhttp3.internal.unmodifiable

/**
 * Policy on when async requests are executed.
 *
 * Each dispatcher uses an [ExecutorService] to run calls internally. If you supply your own
 * executor, it should be able to run [the configured maximum][maxRequests] number of calls
 * concurrently.
 */
class Dispatcher() {
  /**

public final class PostStreamingWithPipe {
  public static final MediaType MEDIA_TYPE_MARKDOWN
      = MediaType.get("text/x-markdown; charset=utf-8");

  private final OkHttpClient client = new OkHttpClient();

  public void run() throws Exception {
    final PipeBody pipeBody = new PipeBody();

    Request request = new Request.Builder()
        .url("https://api.github.com/markdown/raw")
        .post(pipeBody)
        .build();

デフォルトでは、**auto-reload** が有効です。コードを変更するとサーバーが自動で再読み込みされます。これはリソースを多く消費し、無効時より安定性が低くなる可能性があります。開発時のみに使用してください。また、IP アドレス `127.0.0.1`（マシン自身のみと通信するための IP、`localhost`）で待ち受けます。

## `fastapi run` { #fastapi-run }

`fastapi run` を実行すると、デフォルトで本番モードで起動します。

デフォルトでは、**auto-reload** は無効です。また、IP アドレス `0.0.0.0`（利用可能なすべての IP アドレスを意味します）で待ち受けるため、そのマシンと通信できる任意のクライアントから公開アクセスが可能になります。これは、たとえばコンテナ内など、本番環境で一般的な実行方法です。