On 2.6.2012, at 2.19, Adrien W. de Croy wrote: >> That's possible of course, but also pretty kludgy. IMAP in general doesn't >> require clients to do such things. > The flip-side to that is that implementation design decisions shouldn't > necessarily force the protocol. > Other clients don't have this problem, so can you blame the protocol.
I am talking about the protocol, not the client implementation. The protocol should try to behave predictably and without special exceptions as much as possible. Jan's client behaves the way an "ideal IMAP client" would behave. Saying "other IMAP clients don't have this problem" doesn't matter, because pretty much all IMAP clients behave badly or horribly badly. The protocol shouldn't be designed for those. >> And anyway it shouldn't ignore them entirely, because there might be other >> EXPUNGEs reported that didn't come from a MOVE. >> > Not sure that's possible/legal. I don't think a server is allowed to send > unsolicited EXPUNGE responses whilst in another command. It depends on the command. The way UID MOVE is currently defined it is possible for other EXPUNGEs to be sent. (And I think it should stay that way, otherwise it would be an exception to how the protocol normally behaves. Exceptions are bad.) _______________________________________________ imap5 mailing list [email protected] https://www.ietf.org/mailman/listinfo/imap5
