// Postponed routes are always tried last. For example, if we have 2 proxies and all the
      // routes for proxy1 should be postponed, we'll move to proxy2. Only after we've exhausted
      // all the good routes will we attempt the postponed routes.
      val proxy = nextProxy()
      for (inetSocketAddress in inetSocketAddresses) {
        val route = Route(address, proxy, inetSocketAddress)

* First, make sure you **understand the problem** that the pull request is trying to solve. It might have a longer discussion in a GitHub Discussion or issue.

* There's also a good chance that the pull request is not actually needed because the problem can be solved in a **different way**. Then you can suggest or ask about that.

### Don't worry about style { #dont-worry-about-style }

     * both say in their comments that they're not really sure what the right approach is. We go
     * with Chrome's behavior (which also experimentally seems to match what IE does), but if you
     * actually want to have a good chance of things working, please avoid double-quotes, newlines,
     * percent signs, and the like in your field names.
     */
    internal fun StringBuilder.appendQuotedString(key: String) {
      append('"')

    }

    /**
     * Careful here: if threads are mutating the atomicLongArray while this method is executing, the
     * final long[] will be a "rolling snapshot" of the state of the bit array. This is usually good
     * enough, but should be kept in mind.
     */
    public static long[] toPlainArray(AtomicLongArray atomicLongArray) {
      long[] array = new long[atomicLongArray.length()];

github.com/ncw/directio
https://github.com/ncw/directio
----------------------------------------------------------------
Copyright (C) 2012 by Nick Craig-Wood http://www.craig-wood.com/nick/

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal

   *
   * <p><b>Note:</b> the overhead of measurement can be more than a microsecond, so it is generally
   * not useful to specify {@link TimeUnit#NANOSECONDS} precision here.
   *
   * <p>It is generally not a good idea to use an ambiguous, unitless {@code long} to represent
   * elapsed time. Therefore, we recommend using {@link #elapsed()} instead, which returns a
   * strongly-typed {@link Duration} instance.
   *

 *          <br>
 *          <i>NOTE: This will only be used in very special cases, using a highly controlled vocabulary of possible
 *          values. (Elements like this are why it's a good idea to use the descriptor tools.)</i>
 *      </td>
 *   </tr>
 *   <tr>
 *      <td>phase</td>
 *      <td>defaultPhase = LifecyclePhase.&lt;phase&gt;</td>
 *      <td>No</td>

      response.close()
    }
  }

  /**
   * Test that if we set a bad implementation on the OkHttpClient directly, that we can avoid the failure
   * by setting a good override.
   */
  @Test
  fun testOverrideBadImplementation(
    override: OverrideParam =
      burstValues(
        OverrideParam.Authenticator,
        OverrideParam.Cache,

   * Returns a new builder with a hint for how many distinct keys are expected to be added. The
   * generated builder is equivalent to that returned by {@link #builder}, but may perform better if
   * {@code expectedKeys} is a good estimate.
   *
   * @throws IllegalArgumentException if {@code expectedKeys} is negative
   * @since 33.3.0
   */
  public static <K, V> Builder<K, V> builderWithExpectedKeys(int expectedKeys) {

   *
   * <p><b>Note:</b> the overhead of measurement can be more than a microsecond, so it is generally
   * not useful to specify {@link TimeUnit#NANOSECONDS} precision here.
   *
   * <p>It is generally not a good idea to use an ambiguous, unitless {@code long} to represent
   * elapsed time. Therefore, we recommend using {@link #elapsed()} instead, which returns a
   * strongly-typed {@link Duration} instance.
   *