To:
"'ietf-provreg@cafax.se'" <ietf-provreg@cafax.se>
From:
"Hollenbeck, Scott" <shollenbeck@verisign.com>
Date:
Thu, 6 Sep 2001 08:13:58 -0400
Sender:
owner-ietf-provreg@cafax.se
Subject:
FW: EPP Command Question
FYI. I received the private message included below back in August, and I intend to address the comment in the next version of the EPP specifications. I'm sending the suggestion to the list so it gets properly archived. The genesis of the suggested text involved a request to add clarifying text so that implementers understood that they may have to deal with commands and responses that don't get exchanged in a tidy one-for-one fashion. <Scott/> -----Original Message----- From: John Immordino [mailto:johni@nametree.com] Sent: Monday, August 06, 2001 3:04 PM To: Hollenbeck, Scott Subject: RE: EPP Command Question For the EPP spec: "Be sure you consider the mechanics of command-response synchronization if you implement an EPP client that supports asynchronous command-response exchanges. If you implement an EPP server, you must support asynchronous command-response exchanges unless you can guarantee synchronous-only clients." For the TCP spec (append to Section 3): "An EPP client may support asynchronous command-response exchanges. In this case, a server may receive, in the course of a single read operation, data that includes multiple client commands or command fragments. The server must scan the incoming client data, extract and execute properly formed commands as described above, and carry over any remaining data as a prefix to the data received in the next read operation. I'll keep my day job...:-) - John