return cfg, fmt.Errorf("unknown value: '%s'", idleSpeed)
	}

	// Stick to loading deprecated config/env if they are already set, and the Speed value
	// has not been changed from its "default" value, if it has been changed honor new settings.
	if kvs.GetWithDefault(Speed, DefaultKVS) == "default" {
		if kvs.Get(Delay) != "" && kvs.Get(MaxWait) != "" {
			if err = lookupDeprecatedScannerConfig(kvs, &cfg); err != nil {
				return cfg, err
			}

		return 0
	}
	return limit
}

func availableMemory() (available uint64) {
	available = 2048 * blockSizeV2 * 2 // Default to 4 GiB when we can't find the limits.

	if runtime.GOOS == "linux" {
		// Honor cgroup limits if set.
		limit := cgroupMemLimit()
		if limit > 0 {
			// A valid value is found, return its 90%
			available = (limit * 9) / 10
			return available
		}

**In our tests we also found ext4 does not honor POSIX O_DIRECT/Fdatasync semantics, ext4 trades performance for consistency guarantees. Please avoid ext4 in your setup.**

 *  New: `OkHttpClient.webSocketCloseTimeout` configures how long a web socket connection will wait
    for a graceful shutdown before it performs an abrupt shutdown.

 *  Fix: Honor `RequestBody.isOneShot()` in `MultipartBody`

 *  Fix in `okhttp-coroutines`: Don't leak response bodies in `executeAsync()`. We had a bug where

	disks := er.getDisks()
	// Assume (N/2 + 1) quorum for Delete()
	// this is a theoretical assumption such that
	// for delete's we do not need to honor storage
	// class for objects that have reduced quorum
	// due to storage class - this only needs to be honored
	// for Read() requests alone that we already do.
	writeQuorum := len(disks)/2 + 1

        assertEquals(0, response1.getNum());
        assertEquals(0, response1.getTotal());

        SuggestResponse response2 = suggester.suggest().setQuery("honor").setSuggestDetail(true).execute().getResponse();
        assertEquals(1, response2.getNum());
        assertEquals(1, response2.getTotal());
        assertEquals("honorificabilitudinitatibus", response2.getWords().get(0));

          }
        }
      }
    }

    override fun priority(
      streamId: Int,
      streamDependency: Int,
      weight: Int,
      exclusive: Boolean,
    ) {
      // TODO: honor priority.
    }

    override fun pushPromise(
      streamId: Int,
      promisedStreamId: Int,
      requestHeaders: List<Header>,
    ) {
      pushRequestLater(promisedStreamId, requestHeaders)

    between Kotlin's `assert()` and Java's `assert()`. (Kotlin always evaluates the argument; Java
    only does when assertions are enabled.)

 *  Fix: Honor `RequestBody.isOneShot()` in `HttpLoggingInterceptor`.


## Version 4.2.2

_2019-10-06_

 *  Fix: When closing a canceled HTTP/2 stream, don't send the `END_STREAM` flag. This could cause

instances will stomp on each other, corrupt the response cache, and possibly crash your program.

Response caching uses HTTP headers for all configuration. You can add request headers like `Cache-Control: max-stale=3600` and OkHttp's cache will honor them. Your webserver configures how long responses are cached with its own response headers, like `Cache-Control: max-age=9600`. There are cache headers to force a cached response, force a network response, or force the network response to be validated...

- Fixes a regression in the pod binding subresource to honor the `metadata.uid` precondition.