On Mon, 2007-06-25 at 18:54 +0000, Aaron Stone wrote: > On Mon, Jun 25, 2007, [EMAIL PROTECTED] said: > > > > Jorge Bastos: > > > I'd like to have another thing in all daemons on dbmail. > > > I was testing this, i did: > > > telnet server 110 > > > and waited some time to see if the connection go away with a timeout or > > > so, but it stayed alive. > > > How about to have a config value to have this timeout and kill the > > > connection when this value reach's? > > > My idea, 0 disable's timeout, other value defines the timeout and kill's > > > the connection when it reach's that value. > > > > Aaron Stone: > > Please discuss new ideas on the mailing list first! The timeout that > > applies in this instance is the same as the timeout once you're already > > connected. It might make sense to have a separate, shorter timeout prior > > to login. This needs mailing list discussion to work out the details. > > So, yeah, I think 30 - 60 seconds for the login timeout makes sense. Does > anybody feel strongly that it should be configurable? Is it typical for a > client UI to hold open the connection while waiting for the user to supply > a new password? (In which case a longer timeout makes sense, to give the > user time to respond).
I have in the past done manual APOP stuff in testing, which takes cut/pasting from the login banner and back. 60 seconds ought to be plenty to do things like that .. probably a little nicer than 30 seconds. I've not done the same under imap except for a plaintext LOGIN .. dunno if it would potentially take longer there for different auth types or not. In any case, make sure the manual/telnet "UI" isn't affected too much by the value selected. -- Jesse Norell Kentec Communications, Inc. [EMAIL PROTECTED] _______________________________________________ Dbmail-dev mailing list Dbmail-dev@dbmail.org http://twister.fastxs.net/mailman/listinfo/dbmail-dev
