---

Common examples of CPU bound operations are things that require complex math processing.

For example:

* **Audio** or **image processing**.
* **Computer vision**: an image is composed of millions of pixels, each pixel has 3 values / colors, processing that normally requires computing something on those pixels, all at the same time.

        when {
          it.sourceCodePoint1 < codePoint -> -1
          it.sourceCodePoint0 > codePoint -> 1
          else -> 0
        }
      }

    // Code points must be in 0..0x10ffff.
    require(index in mappings.indices) { "unexpected code point: $codePoint" }

    val mapping = mappings[index]
    var result = true

    when (mapping.type) {
      TYPE_IGNORED -> Unit

    /**
     * Constructs a new DataSerializer.
     * <p>
     * Initializes the ThreadLocal Kryo instances with appropriate configuration.
     * The Kryo instances are configured to not require class registration for
     * flexibility, and debug mode warnings are enabled when debug logging is active.
     * </p>
     */
    public DataSerializer() {
        kryoThreadLocal = ThreadLocal.withInitial(() -> {

			err = msgp.WrapError(err, "Tiers", za0001)
			return
		}
	}
	return
}

// MarshalMsg implements msgp.Marshaler
func (z *TierConfigMgr) MarshalMsg(b []byte) (o []byte, err error) {
	o = msgp.Require(b, z.Msgsize())
	// map header, size 1
	// string "Tiers"
	o = append(o, 0x81, 0xa5, 0x54, 0x69, 0x65, 0x72, 0x73)
	o = msgp.AppendMapHeader(o, uint32(len(z.Tiers)))
	for za0001, za0002 := range z.Tiers {

some connectivity failures. In this case, the `connectFailed()` event is not terminal and not followed by `callFailed()`. Event listeners will receive multiple events of the same type when retries are attempted.

A single HTTP call may require follow-up requests to be made to handle authentication challenges, redirects, and HTTP-layer timeouts. In such cases multiple connections, requests, and responses may be attempted. Follow-ups are another reason a single call may trigger multiple events...

It has a `GET` operation that could return an error.

It has a `POST` operation that could return several errors.

Both *path operations* require an `X-Token` header.

//// tab | Python 3.10+

```Python
{!> ../../docs_src/app_testing/app_b_an_py310/main.py!}
```

////

//// tab | Python 3.9+

```Python

		err = msgp.WrapError(err, "TargetReplicationResyncStatus")
		return
	}
	return
}

// MarshalMsg implements msgp.Marshaler
func (z *SiteResyncStatus) MarshalMsg(b []byte) (o []byte, err error) {
	o = msgp.Require(b, z.Msgsize())
	// map header, size 6
	// string "v"
	o = append(o, 0x86, 0xa1, 0x76)
	o = msgp.AppendInt(o, z.Version)
	// string "ss"
	o = append(o, 0xa2, 0x73, 0x73)
	o, err = z.Status.MarshalMsg(o)

      "Code $code is reserved and may not be used."
    } else {
      null
    }

  fun validateCloseCode(code: Int) {
    val message = closeCodeExceptionMessage(code)
    require(message == null) { message!! }
  }

  fun acceptHeader(key: String): String = (key + ACCEPT_MAGIC).encodeUtf8().sha1().base64()

  @Throws(IOException::class)
  private fun writeControlFrame(
    opcode: Int,
    payload: ByteString,
  ) {
    if (writerClosed) throw IOException("closed")

    val length = payload.size
    require(length <= PAYLOAD_BYTE_MAX) {
      "Payload size must be less than or equal to $PAYLOAD_BYTE_MAX"
    }

    val b0 = B0_FLAG_FIN or opcode
    sinkBuffer.writeByte(b0)

    var b1 = length

and can heal one object at a time. For RAID, healing can be done only at the volume level which translates into high downtime. As MinIO encodes each object individually, it can heal objects incrementally. Storage servers once deployed should not require drive replacement or healing for the lifetime of the server. MinIO's erasure coded backend is designed for operational efficiency and takes full advantage of hardware acceleration whenever available.

![Erasure](https://github.com/minio/m...