*  **Canceling batches of calls is now the application's responsibility.**
    The API to cancel calls by tag has been removed and replaced with a more
    general mechanism. The dispatcher now exposes all in-flight calls via its
    `runningCalls()` and `queuedCalls()` methods. You can write code that
    selects calls by tag, host, or whatever, and invokes `Call.cancel()` on the

import okio.Source
import org.codehaus.mojo.animal_sniffer.IgnoreJRERequirement

/**
 * Factory for [calls][Call], which can be used to send HTTP requests and read their responses.
 *
 * ## OkHttpClients Should Be Shared
 *
 * OkHttp performs best when you create a single `OkHttpClient` instance and reuse it for all of
 * your HTTP calls. This is because each client holds its own connection pool and thread pools.

      return uncheckedCastNullableTToT(peekedElement);
    }
  }

  /**
   * Returns a {@code PeekingIterator} backed by the given iterator.
   *
   * <p>Calls to the {@code peek} method with no intervening calls to {@code next} do not affect the
   * iteration, and hence return the same object each time. A subsequent call to {@code next} is
   * guaranteed to return the same object again. For example:
   *

 * ListenableFuture}</a>.
 *
 * <p>The main purpose of {@code ListenableFuture} is to help you chain together a graph of
 * asynchronous operations. You can chain them together manually with calls to methods like {@link
 * Futures#transform(ListenableFuture, Function, Executor) Futures.transform}, but you will often
 * find it easier to use a framework. Frameworks automate the process, often adding features like

 *
 * <p>If {@link #expireAfterWrite expireAfterWrite} or {@link #expireAfterAccess expireAfterAccess}
 * is requested entries may be evicted on each cache modification, on occasional cache accesses, or
 * on calls to {@link Cache#cleanUp}. Expired entries may be counted by {@link Cache#size}, but will
 * never be visible to read or write operations.
 *
 * <p>If {@link #weakKeys weakKeys}, {@link #weakValues weakValues}, or {@link #softValues

      }
    }

    // subsequent calls should call the loader again, not get the old exception
    try {
      cache.getUnchecked("bar");
      fail();
    } catch (UncheckedExecutionException expected) {
    }
    assertEquals(2, callCount.get());
  }

  /**
   * Test-helper method that performs {@code nThreads} concurrent calls to {@code cache.get(key)} or

    val call2 = client.newCall(Request(server.url("/")))
    call2.execute().use { response ->
      assertThat(
        response.body.string(),
      ).isEqualTo("b")
    }

    // Calls succeed after the degraded pong timeout because the degraded pong was received.
    Thread.sleep(TimeUnit.NANOSECONDS.toMillis(Http2Connection.DEGRADED_PONG_TIMEOUT_NS.toLong()))

  }

  private static final int DISTINCT_CHARS = Character.MAX_VALUE - Character.MIN_VALUE + 1;

  /**
   * This is the actual implementation of {@link #precomputed}, but we bounce calls through a method
   * on {@link Platform} so that we can have different behavior in GWT.
   *
   * <p>This implementation tries to be smart in a number of ways. It recognizes cases where the

     */
    @Nullable V get();

    /** Returns the entry which contains this {@link WeakValueReference}. */
    E getEntry();

    /** Unsets the referenced value. Subsequent calls to {@link #get} will return {@code null}. */
    void clear();

    /**
     * Returns a freshly created {@link WeakValueReference} for the given {@code entry} (and on the

### Other notable changes

* GCE provider: Limit Filter calls to regexps rather than large blobs ([#27741](https://github.com/kubernetes/kubernetes/pull/27741), [@zmerlynn](https://github.com/zmerlynn))