I am running dovecot 0.99-14 on a Fedora Core 2 machine. I had a few questions:
- I wanted to upgrade to the dovecot-1.0 release. However, I am not sure if that's really required. dovecot-0.99-14 has been running very well for me for quite some time. Is there a real advantage to switching to the latest release. The reason I'm asking this is because: [..snip..] iv) I am also planning to support web mail for my users (probably through Squirrelmail). Are there any advantages/disadvantages to upgrading dovecot for that?
Naah.. Well, "not really" is the right answer. Dovecot's a back-end interface to email, but, because of its performance improvements over UW-IMAP, all other applications that access mail via dovecot will benefit from the performance gain.
For example, I noticed right off that "overlook" (a fork of the Squirrelmail code with a Outlook-ish interface) was much faster when I had dovecot in place on our server vs UW-IMAP.
[..snip..] So, are there newer version of dovecot (1.0-stable) available in rpm formats which are guaranteed to be safe and un-trojaned? And is there any really good reason why I should upgrade to 1.0-stable?
Well, when 1.0-stable is made available, it is probably best to upgrade to gain all of the improvements and bug fixes between the version you are currently running and the 1.0-stable version. Of course, as with any major release, a good practice is to let the "early adopters" try it out and, once it has past the 'burn in period', then upgrade... but the choice as to when to upgrade is truly in your hands.
For key applications (PHP, apache, sendmail, dns, etc), I personally "roll my own" RPMS. I personally prefer the file organization of RPMs over straight source installs (or other management tools for other platforms, like pkgadd and such for BSD derivatives), and creating your own (for your use) RPM is not all that difficult when you have a base SRPM to work with that you know and trust.
- My second question is about creating the .subscriptions (or subscriptions in 1.0) file. When I moved people over from using pine to using a mail client like Evolution/Thunderbird, I had to create .subscriptions/.mailboxlist file for them so that they could see the folders that they had created in pine. However, people still sometimes access e-mail using pine. They then create new folders in pine, which they are unable to see in their mail client. Is there a way around it? This occurs especially when people check their mail at the start of the month, and pine asks about saving/deleting old sent-mail folders, Trash folders. Is there a way to add these folders automagically to the .subscriptions file?
Well, the only way around that would be to have a script which comes across at a set time throughout the day to synchronize the two files..
- Would I be able to support web mail using mailboxes in mbox format, as I have them now? Or do I have to go for Maildir format?
mbox vs. maildir format is actually a separate issue. Most webmail applications can handle either format.
Since our university introduced web mail, they have had two sets of folders - one with capitalization (Sent, Deleted Messages, etc.) and one without (sent-mail, etc. - these were from our pine days). Is this due to web mail?
Those differences are due to the different defaults within the webmail application and the other client applications. You find that webmail applications tend to set the defaults to "Drafts" and "sent-mail", and, because of those defaults (people tend to leave the defaults as is), you end up with those folder names in the subscriptions/mailboxlist file.
Also, if I do need to convert from mbox to Maildir, what is the standard tool for doing that? I saw at least three different tools/scripts for doing that (mbox2maildir,mb2d,mb2md-2,mb2md.pl, .. ). Which tool (and from where) should I use to convert mailboxes without problems?
I've been looking at the same thing and finally decided that, for our use of email & the mbox format, migrating to maildir did not provide any benefit. Be sure to fully analyze your needs to insure that migrating is worth your efforts; it isn't necessary for everyone.
-Rich