It appears that I have three options: 1. Encourage client vendors to change this practice.
2. Announce new messages to all connections promptly. 3. Take Timo suggestion and notice the message underneath. One seems like an uphill battle. Two is the most compliant solution, but will still will not solve the problem with out changes from at least Evolution. Three does solve the issue for OE, but leaves me with some concerns. Will three cause issues with other clients or cause our IMAP server to violate the RFC? -----Original Message----- From: Timo Sirainen <[EMAIL PROTECTED]> To: Mark Crispin <[EMAIL PROTECTED]> Date: 03 Apr 2003 01:01:14 +0300 Subject: Re: Proper Response to UID STORE command? On Thu, 2003-04-03 at 00:50, Mark Crispin wrote: > > Except for the ones that break if it is done. Evolution at least breaks > > if it sees untagged reply instead of "+" after it has sent APPEND. > > If Evolution does that, then it is broken. Yeah, it is. Too many IMAP clients are. > > OE also breaks if it sees EXISTS while it's expecting FETCH > > replies. > > I don't see how OE can possibly do that. Otherwise it wouldn't work with > many servers, including UW imapd. It happens when it has multiple connections open for same mailbox. First connection sees new mail, with second it does UID FETCH <newuid> BODY[]. If there's EXISTS before or after the untagged FETCH reply, it shows "message no longer in server" as message body.
