I guess I answered myself. Subclassing my base protocol to handle the marshalling between sending and receiving appears more transparent. Regards, David David Pratt wrote:
Hi. I am trying to decide where to put message sizing for a custom protocol. I have set up the connections to be as generic as possible so message handling in as much as possible is being done by the connection handler that is aware of the connection it is handling at the time. There are a few things I need to keep track of at the connection (protocol) level such as original handshake and message id's that I am tracking per connection, but for the most part, everything else can be handled pretty generically by the connection handler.
In any case, where to deal with message sizing? My thought was that the connection handler should deal with it since it should also be responsible for encode and decode of messages. So to process or prepare a message size seems like it fits in the connection handler also.
I realize Banana, for example does sizing in the protocol itself but it does not anticipate handing the message off to a handler for such things. Anyway, it would be great if someone could confirm my reasoning or has some comments on this. Many thanks,
Regards, David
_______________________________________________ Twisted-Python mailing list Twisted-Python@twistedmatrix.com http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python