[Standards-JIG] Re: Closing idle streams (server comparison chart)

Carlo v. Loesch CvL at mail.symlynX.com
Wed May 31 19:53:29 UTC 2006


Carlo v. Loesch typeth:
| Matthias Wimmer typeth:
| | At least you can already close the sending half of a TCP/IP connection 
| | after you sent the </stream:stream>. It is enough to keep the receiving 
| | half of the connection open.
| 
| Oh you are right, thank you.
| You mean the outgoing TCP line though.

No wait, it's not that simple.

If you properly close the socket sending </stream:stream>, then the
other side will want to reply with its own </stream:stream>. This
will cause a TCP error and the whole transaction will look like
there was an error even though there wasn't, and poof we're back
in fake-unreliability-land where we are just trying to escape from.

Either we do the proper handshake, or we declare the kill-the-socket
method on outgoing links legal, or we use the good-old connection-timeout
which by definition implies a socket shutdown and thus won't trigger any
replies.




More information about the Standards mailing list