From: Matthias A. <mat...@gm...> - 2006-04-21 14:47:58
|
Ratan Nalumasu <nal...@ya...> writes: > I read the FAQ, man pages and tried several settings to solve a > simple problem "how to retain Seen flag when fetching mail with > fetchmail" and met with complete failure. > > The problem is that when I read the mail on the server using > elm, mailx, or squrrelmail, they all agree on what message is > previously read and what is not. However, when I fetch the mail > using fetchmail, those status bits are disappearing. I am > attaching my .fetchmailrc, the contents of the mailbox before > the fetchmail, output of "fetchmail -v -v", and the mailbox > after the mail is fetched. Any help is welcome. I see you're using IMAP, and unfortunately, fetchmail relies on the \Seen flags stored on the server to determine which messages to download with IMAP -- in your situation for instance, it will not download messages that you've read with squirrelmail or elm at all. There aren't many options out of this situation. You could theoretically use "--fetchall --nokeep", but then squirrelmail could no longer be used. If the server offers POP3 access, you can try POP3 with --uidl (or the uidl option), fetchmail will use "TOP" to retrieve messages, which works on most servers without setting seen flags. Some servers (such as Dovecot) also don't set seen flags in IMAP when the message has been retrieved with POP3. A future fetchmail version (6.4.0 or something) will be able to track the UID to see which messages are read or unread with IMAP as well (and not just POP3), but that code has yet to be written. Finally, fetchmail 6.2.5.* isn't officially supported any more (though it may still be covered by vendor support), and 6.2.5.2 needs additional patches (errata or something) to fix security bugs CVE-2005-3088 and CVE-2005-4348. While they aren't major issues (one affects fetchmailconf and the other multidrop), your vendor should have fixed them. Details at http://www.fetchmail.info/ Anyways, I hope my answer helped you find a way around your problem. If not, try describing what you're trying to achieve, perhaps somebody knows a solution. -- Matthias Andree |