"Jason R. Mastaler" <[EMAIL PROTECTED]> writes: > Drew Raines <[EMAIL PROTECTED]> writes: > >> * new/100000.PID.host Quarantined message whose >> sender was sent a confirmation >> >> * new/100001.PID.host:2,F Quarantined message whose sender >> has sent other unconfirmed, >> pending messages >> >> * cur/100002.PID.host:2,T Confirmed message >> >> * cur/100003.PID.host:2,S Confirmed message whose sender >> has sent other unconfirmed, >> pending messages > > The questions is: is it safe to do this when these "info" fields > mean something completely different to an MUA than they do to > TMDA? My gut tells me it isn't.
I'm not clear on what exactly you want the MUA to do with them. I don't see functionality with any stock MUA beyond viewing and deleting messages in the pending queue. If that indeed is all we can do, then the naming conventions don't need to be terribly complex. However, I have another brainstorm. We can always mix and match identifiers, and DJB requests the letters be in ASCII order. If we need to, we could make ridiculous conjunctions such as: 100000.PID.host:2,DFST for a confirmed message with other sender messages in the queue. This seems unnecessary since you apparently have a more elaborate directory/filename scheme cooking up in another message. I'm not sure we can be too conservative, though. Since there's not a lot of applications mangling maildirs in this manner, we don't have much to go on. Best bet is to forsee future conflicts and write code expecting to modify it (like you don't do that already). If TMDA sets some strong precedents, other software may adapt to it. _________________________________________________ tmda-workers mailing list ([EMAIL PROTECTED]) http://tmda.net/lists/listinfo/tmda-workers
