Robot UDP Transmitter Protocol

The protocol sends all SAX events of HTML parser.

Protocol

• UDP
• server port port_s>1024

Initialization and run

1. Client (IP_c:port_c) sends list of URL prefixes it wants to listen to, separated by a single space
2. (if port_c!=port_s) Server responds to (IP_c:port_s) with "HELO" and the list of the recognized prefixes, the tokens are separated by "\n"
3. Server emits all SAX events to (IP_c:port_c) for up to 30 seconds

If a client subscribes several times, the last subscription is valid only.

Shutdown

1. Client sends from its original (IP_c:port_c) an empty packet
2. Server responds with
   • "BYE" if the client is known and was removed from the broadcast list
   • "ERR" if the client is unknown to the server, so the unsubscribe operation was canceled

Events

Event format is:

• ID (4 bytes)
• global counter (4 bytes)
• sequential number of this event (4 bytes)
• command (2 bytes - interpreted as char)
  • 'D' - document start, value then contains URI in ASCII
  • 'd' - document end, value then contains URI in ASCII
  • 'E' - element start
  • 'e' - element end
  • 'G' - general entity start
  • 'g' - general entity end
  • 'X' - empty element
  • 'C' - comment
  • 'T' - text
  • 'a' - abort
• total number of fragments (4 bytes)
• fragment number (4 bytes)
• length of the value (4 bytes)
• value

Last item may be zero length.

One UDP packet may transfer several events.