*  Fix: Work around a crash in Android 10 and 11 that may be triggered when two threads
    concurrently close an SSL socket. This would have appeared in crash logs as
    `NullPointerException: bio == null`.
 *  Fix: Use plus `+` instead of `%20` to encode space characters in `FormBody`. This was a
    longstanding bug in OkHttp. The fix makes OkHttp consistent with major web browsers.
 *  Fix: Don't crash if Conscrypt returns a null version.

    // For efficiency, especially in executors with limited
    // parallelism, check to see if previously submitted tasks are
    // done before submitting more of them. This interleaving
    // plus the exception mechanics account for messiness of main
    // loop.

    try {
      // Record exceptions so that if we fail to obtain any
      // result, we can throw the last exception we got.

   * <p><b>Java 8+ users:</b> use {@link Comparator#naturalOrder} instead.
   */
  @GwtCompatible(serializable = true)
  @SuppressWarnings({"unchecked", "rawtypes"})
  // TODO(kevinb): right way to explain this??
  // plus https://github.com/google/guava/issues/989
  public static <C extends Comparable> Ordering<C> natural() {
    return (Ordering<C>) NaturalOrdering.INSTANCE;
  }

  // Static factories

  /**

    checkNonnegative(initialArraySize, "initialArraySize"); // for GWT.
    return new ArrayList<>(initialArraySize);
  }

  /**
   * Creates an {@code ArrayList} instance to hold {@code estimatedSize} elements, <i>plus</i> an
   * unspecified amount of padding; you almost certainly mean to call {@link
   * #newArrayListWithCapacity} (see that method for further advice on usage).
   *

work results from an entity transaction, each party to that
transaction who receives a copy of the work also receives whatever
licenses to the work the party's predecessor in interest had or could
give under the previous paragraph, plus a right to possession of the
Corresponding Source of the work from the predecessor in interest, if
the predecessor has it or can get it with reasonable efforts.

      byte[] buf = new byte[min(bufSize, MAX_ARRAY_LEN - totalLen)];
      bufs.add(buf);
      int off = 0;
      while (off < buf.length) {
        // always OK to fill buf; its size plus the rest of bufs is never more than MAX_ARRAY_LEN
        int r = in.read(buf, off, buf.length - off);
        if (r == -1) {
          return combineBuffers(bufs, totalLen);
        }
        off += r;

            "Spent " + thread.timeSpentBlocked + " ns blocked; slept for " + toWaitMillis + " ms")
        .that(thread.exception)
        .hasMessageThat()
        .matches(
            "Waited 2 seconds \\(plus "
                + (longWait ? "3" : "1")
                + " seconds, "
                + nanosRegex
                + " nanoseconds delay\\).*");
  }

  public void testToString_completed() throws Exception {

 *
 * <p>When calling {@link ScaleAndIndex#compute} (in {@linkplain ScaleAndIndexes#compute either
 * form}), the memory requirement is 8*N bytes for the copy of the dataset plus an overhead which is
 * independent of N (but depends on the quantiles being computed). When calling {@link
 * ScaleAndIndex#computeInPlace computeInPlace} (in {@linkplain ScaleAndIndexes#computeInPlace

	// create a new label space for this TEXT.
	p.patch()
	p.labels = make(map[string]*obj.Prog)

	// Operand 0 is the symbol name in the form foo(SB).
	// That means symbol plus indirect on SB and no offset.
	nameAddr := p.address(operands[0])
	if !p.validSymbol("TEXT", &nameAddr, false) {
		return
	}
	name := symbolName(&nameAddr)
	next := 1

succeeded except for the part where the new replication controller is renamed back to the old name. You can just do another rolling update using kubectl 1.4 or higher to change the name back: look for a replication controller that has the original name plus a random suffix.

Unfortunately, there is a much rarer second possible failure mode: the replication controller gets renamed to the old name, but there is a duplicated set of pods in the cluster. kubectl will not report an error since it...