Class StompDecoder

java.lang.Object
org.springframework.messaging.simp.stomp.StompDecoder

public class StompDecoder extends Object
Decodes one or more STOMP frames contained in a ByteBuffer.

An attempt is made to read all complete STOMP frames from the buffer, which could be zero, one, or more. If there is any left-over content, i.e. an incomplete STOMP frame, at the end the buffer is reset to point to the beginning of the partial content. The caller is then responsible for dealing with that incomplete content by buffering until there is more input available.

Since:
4.0
Author:
Andy Wilkinson, Rossen Stoyanchev
  • Constructor Details

    • StompDecoder

      public StompDecoder()
  • Method Details

    • setHeaderInitializer

      public void setHeaderInitializer(@Nullable MessageHeaderInitializer headerInitializer)
      Configure a MessageHeaderInitializer to apply to the headers of Messages from decoded STOMP frames.
    • getHeaderInitializer

      @Nullable public MessageHeaderInitializer getHeaderInitializer()
      Return the configured MessageHeaderInitializer, if any.
    • decode

      public List<Message<byte[]>> decode(ByteBuffer byteBuffer)
      Decodes one or more STOMP frames from the given ByteBuffer into a list of Messages. If the input buffer contains partial STOMP frame content, or additional content with a partial STOMP frame, the buffer is reset and null is returned.
      Parameters:
      byteBuffer - the buffer to decode the STOMP frame from
      Returns:
      the decoded messages, or an empty list if none
      Throws:
      StompConversionException - raised in case of decoding issues
    • decode

      public List<Message<byte[]>> decode(ByteBuffer byteBuffer, @Nullable MultiValueMap<String,String> partialMessageHeaders)
      Decodes one or more STOMP frames from the given buffer and returns a list of Messages.

      If the given ByteBuffer contains only partial STOMP frame content and no complete STOMP frames, an empty list is returned, and the buffer is reset to where it was.

      If the buffer contains one or more STOMP frames, those are returned and the buffer reset to point to the beginning of the unused partial content.

      The output partialMessageHeaders map is used to store successfully parsed headers in case of partial content. The caller can then check if a "content-length" header was read, which helps to determine how much more content is needed before the next attempt to decode.

      Parameters:
      byteBuffer - the buffer to decode the STOMP frame from
      partialMessageHeaders - an empty output map that will store the last successfully parsed partialMessageHeaders in case of partial message content in cases where the partial buffer ended with a partial STOMP frame
      Returns:
      the decoded messages, or an empty list if none
      Throws:
      StompConversionException - raised in case of decoding issues
    • skipEol

      protected void skipEol(ByteBuffer byteBuffer)
      Skip one ore more EOL characters at the start of the given ByteBuffer. STOMP, section 2.1 says: "The NULL octet can be optionally followed by multiple EOLs."