using [spotless], backed by the Eclipse formatter. **We strongly recommend
installing using the [Eclipse Code Formatter] plugin** so that you can
apply the correct formatting directly in IntelliJ.  The configuration for
the plugin is held in `.idea/eclipseCodeFormatter.xml` and should be
automatically applied, but manual instructions are below in case you you
need them.

   1. Open **Preferences > Other Settings > Eclipse Code Formatter**

 *  Fix: Don't interpret trailers as headers after an HTTP/100 response. This was a bug only when
    the HTTP response body itself is empty.
 *  Fix: Don't crash when a fast fallback call has both a deferred connection and a held connection.
 *  Fix: `OkHttpClient` no longer implements `Cloneable`. It never should have; the class is
    immutable. This is left over from OkHttp 2.x (!) when that class was mutable. We're using the

 *     .readTimeout(500, TimeUnit.MILLISECONDS)
 *     .build();
 * Response response = eagerClient.newCall(request).execute();
 * ```
 *
 * ## Shutdown Isn't Necessary
 *
 * The threads and connections that are held will be released automatically if they remain idle. But
 * if you are writing a application that needs to aggressively release unused resources you may do
 * so.
 *

enough to get through the door, she ran out of the house, and
found quite a crowd of little animals and birds waiting outside.
The poor little Lizard, Bill, was in the middle, being held up by
two guinea-pigs, who were giving it something out of a bottle.
They all made a rush at Alice the moment she appeared; but she
ran off as hard as she could, and soon found herself safe in a
thick wood.

	plkctx, err := partIDLock.GetLock(ctx, globalOperationTimeout)
	if err != nil {
		return PartInfo{}, err
	}

	ctx = plkctx.Context()
	defer partIDLock.Unlock(plkctx)

	// Read lock for upload id, only held while reading the upload metadata.
	uploadIDRLock := er.NewNSLock(bucket, pathJoin(object, uploadID))
	rlkctx, err := uploadIDRLock.GetRLock(ctx, globalOperationTimeout)
	if err != nil {
		return PartInfo{}, err
	}

    assertThat(server.takeRequest().exchangeIndex).isEqualTo(1)
    assertThat(server.takeRequest().exchangeIndex).isEqualTo(0)

    // Let the first request proceed. It should discard the the held HTTP/2 connection and get a new
    // one.
    latch.countDown()
    val response3 = responses.take()
    assertThat(response3).isEqualTo("ABC")
    assertThat(server.takeRequest().exchangeIndex).isEqualTo(1)

    closing it doesn't also close the underlying stream. We had a bug where
    OkHttp would attempt to reuse a connection but couldn't because it was still
    held by a prior request.
 *  Fix: Ensure network interceptors always have access to the underlying
    connection.
 *  Fix: Use `X509TrustManagerExtensions` on Android 17+.

  /** Marker interface for {@link InternalEntry} implementations for weak values. */
  interface WeakValueEntry<K, V, E extends InternalEntry<K, V, E>> extends InternalEntry<K, V, E> {
    /** Gets the weak value reference held by entry. */
    WeakValueReference<K, V, E> getValueReference();
  }

  @SuppressWarnings("unchecked") // impl never uses a parameter or returns any non-null value
  static <K, V, E extends InternalEntry<K, V, E>>

    memcpy(buf, new_handle.c_str(), new_handle.size() + 1);
    *handle = buf;
  }
}

void TF_DeletePRunHandle(const char* handle) {
  delete[] handle;
  // TODO(suharshs): Free up any resources held by the partial run state.
}

void TF_SessionPRun(TF_Session* session, const char* handle,
                    const TF_Output* inputs, TF_Tensor* const* input_values,

  /** Marker interface for {@link InternalEntry} implementations for weak values. */
  interface WeakValueEntry<K, V, E extends InternalEntry<K, V, E>> extends InternalEntry<K, V, E> {
    /** Gets the weak value reference held by entry. */
    WeakValueReference<K, V, E> getValueReference();
  }

  @SuppressWarnings("unchecked") // impl never uses a parameter or returns any non-null value
  static <K, V, E extends InternalEntry<K, V, E>>