kafka 源代码分析之Message（v0.10）

这里主要更新一下kafka 0.10.0版本的message消息格式的变化。

message 的格式在0.10.0的版本里发生了一些变化（相对于0.8.2.1的版本）这里把0.10.0的message的代码分析一下。

message的格式定义在源代码目录message目录里的Message.scala的源代码目录里。下面看一下主要代码。

object Message {

  /**
   * The current offset and size for all the fixed-length fields
   */
  val CrcOffset = 0
  val CrcLength = 4
  val MagicOffset = CrcOffset + CrcLength
  val MagicLength = 1
  val AttributesOffset = MagicOffset + MagicLength
  val AttributesLength = 1
  // Only message format version 1 has the timestamp field.
  val TimestampOffset = AttributesOffset + AttributesLength      //可以看见最明显的变化在这里。在以前的message格式的基础上head部分多了一个字段。就是时间戳的字段。
  val TimestampLength = 8    //时间戳的字段长度是8个字节。 唯一的变化在这里。
  val KeySizeOffset_V0 = AttributesOffset + AttributesLength   //同时支持了消息的两个版本。v0和v1，v0跟以前一样。没有时间戳的字段。
  val KeySizeOffset_V1 = TimestampOffset + TimestampLength  //v1多了时间戳的字段。
  val KeySizeLength = 4
  val KeyOffset_V0 = KeySizeOffset_V0 + KeySizeLength
  val KeyOffset_V1 = KeySizeOffset_V1 + KeySizeLength
  val ValueSizeLength = 4

//下面的一些定义都是针对消息的相关定义。跟之前也有些变化但是变化不大。
  private val MessageHeaderSizeMap = Map (
    (0: Byte) -> (CrcLength + MagicLength + AttributesLength + KeySizeLength + ValueSizeLength),
    (1: Byte) -> (CrcLength + MagicLength + AttributesLength + TimestampLength + KeySizeLength + ValueSizeLength))

  /**
   * The amount of overhead bytes in a message
   * This value is only used to check if the message size is valid or not. So the minimum possible message bytes is
   * used here, which comes from a message in message format V0 with empty key and value.
   */
  val MinMessageOverhead = KeyOffset_V0 + ValueSizeLength

  /**
   * The "magic" value
   * When magic value is 0, the message uses absolute offset and does not have a timestamp field.
   * When magic value is 1, the message uses relative offset and has a timestamp field.
   */
  val MagicValue_V0: Byte = 0
  val MagicValue_V1: Byte = 1
  val CurrentMagicValue: Byte = 1

  /**
   * Specifies the mask for the compression code. 3 bits to hold the compression codec.
   * 0 is reserved to indicate no compression
   */
  val CompressionCodeMask: Int = 0x07
  /**
   * Specifies the mask for timestamp type. 1 bit at the 4th least significant bit.
   * 0 for CreateTime, 1 for LogAppendTime
   */
  val TimestampTypeMask: Byte = 0x08
  val TimestampTypeAttributeBitOffset: Int = 3

  /**
   * Compression code for uncompressed messages
   */
  val NoCompression: Int = 0

  /**
   * To indicate timestamp is not defined so "magic" value 0 will be used.
   */
  val NoTimestamp: Long = -1

  /**
   * Give the header size difference between different message versions.
   */
  def headerSizeDiff(fromMagicValue: Byte, toMagicValue: Byte) : Int =
    MessageHeaderSizeMap(toMagicValue) - MessageHeaderSizeMap(fromMagicValue)

}

　　这里就简单记录一下v0.10.0的message格式上的变化。对于Message.scala源代码里的方法，变量等就不再做详细说明了。