On Mon, 23 Jan 2006, Jesse Chan-Norris wrote:
On Jan 22, 2006, at 6:44 AM, Timo Sirainen wrote:
On Mon, 2006-01-16 at 18:29 -0500, jcn-dovecot@pith.org wrote: I'm having a strange problem that appeared when I upgraded from�.99something to 1.0alpha3 and then disappeared for a while and then� re-appeared when I upgraded to 1.0alpha5.
I am using mbox and IMAP in Apple's Mail.app to read my mail. When I� delete a message in Mail.app, the message gets marked as deleted in the� spool and it disappears from the mail listing in Mail. However, the� message number does not change until I either manually flush the deleted� messages through mail, or I go offline which deletes the messages from the� spool.
I have mbox_lazy_writes set to 'no' but this didn't seem to have any� effect on the behavior. I never saw this type of behavior with UW-IMAP� either, which is what I migrated from.
What do you mean by "message number doesn't change"?
To me this sounds like it all works as it should. Deleted messages are really deleted only until expunge command is given, which is done when you manually select "purge expunged messages" command or whatever it was called in Mail.
Maybe before with UW-IMAP you just had some "automatically expunge messages" option set?
I don't think that UW-IMAP was expunging messages because I could open up the mbox with pine and see all of the messages flagged as "Deleted" but not yet expunged. And when I try a similar test with Fastmail.fm (delete the message in Mail, then check using the web interface) the mail is marked as deleted in the web interface and does not show up in Mail.app but the messages are re-numbered appropriately.
Granted, I have no idea how Mail.app does its numbering (especially since Thunderbird doesn't have message numbering), but is it possible that there's more information that the IMAP server should be sending back that Dovecot isn't?
I've attached a screenshot from my Mail.app that shows the numbering issue. The gaps in the message sequence are where a deleted (but not expunged) message was.
jcn