and today's HORROR stems from one of those Archane Questions that has to do with the TCP RST bit, and a specific protocol, which, upon further analysis, asserts, explicitly that when there is a Channel Error, then the Server will transmit that message to the Client, and send the recommended solution is to send a TCP RST. Which, gosh, golly, would Oblige the Client to shutdown, close() and restart with the CORRECT Channel Information!!!!
That of course is a whole Skank lot easier to code up, since it does not ask that the Application layer coder know DIDDLY about when, where and how, the underlaying TCP did which things at it's Pay Grade!!!! AH!!!!! Why can't Coder's Get In Touch With The Protocol's Inner Child???