Hi,
Is there a way with dovecot to warn users who are close to reach their quota?
Although I never used it, I think cyrus supports something like that by droping a warning message into the INBOX when delivering a message that brings the total size over a certain percentage.
Would it be feasible to implement something like that with an extra plugin or by changing the quota plugin?
Cheers,
Nicolas
Hi List,
I recently installed dovecot 1.0rc15 on centos, installation went fine, it installed under /usr/local/ but I cannot start the dovecot service automatically. I tried with chkconfig but it told me the dovecot service is not supported by chkconfig. Can anybody give me a tip about how to get done with this?
I will appreciate any help.
Regards,
--Ivan.
"Look both ways before crossing the Net"
On Wed, Feb 14, 2007 at 08:02:15AM -0500, Ivan Arteaga wrote:
Hi List,
I recently installed dovecot 1.0rc15 on centos, installation went fine, it installed under /usr/local/ but I cannot start the dovecot service automatically. I tried with chkconfig but it told me the dovecot service is not supported by chkconfig. Can anybody give me a tip about how to get done with this?
I will appreciate any help.
Install http://dl.atrpms.net/all/dovecot-1.0-2_45.rc22.el4.at.i386.rpm and use
chconfig dovecot on
Regards,
--Ivan.
"Look both ways before crossing the Net"
-- Axel.Thimm at ATrpms.net
Ivan Arteaga wrote:
I recently installed dovecot 1.0rc15 on centos, installation went fine, it installed under /usr/local/ but I cannot start the dovecot service automatically.
Take a look to the start/stop-Scripts in /etc/init.d/. Take one of these (maybe "atd" or whatever is installed) and copy it to "dovecot". Such a script is simple to understand normally, so you can edit it to start/stop your manually installed dovecot easily. (Maybe there comes such a script within the dovecot-tarball? I am using the deb-package that is included in debian/sid that does set up everything automatically.)
Then you need to place symlinks to the runlevels to start/stop the daemon. Here on a debian-system there are:
/etc/rc0.d/K21dovecot -> ../init.d/dovecot /etc/rc1.d/K21dovecot -> ../init.d/dovecot /etc/rc2.d/S21dovecot -> ../init.d/dovecot /etc/rc3.d/S21dovecot -> ../init.d/dovecot /etc/rc4.d/S21dovecot -> ../init.d/dovecot /etc/rc5.d/S21dovecot -> ../init.d/dovecot /etc/rc6.d/K21dovecot -> ../init.d/dovecot
Maybe this helps?
Gert
Helo, here i my debian startup script:
and to runlevel copy script (in /etc/init.d directory): update-rc.d dovecot defaults 98 02
#!/bin/sh -e
PATH=/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/dovecot/bin/dovecot DAEMON=/usr/local/dovecot/sbin/dovecot NAME=dovecot DESC="IMAP server"
test -x $DAEMON || exit 0
case "$1" in start) echo -n "Starting $DESC: $NAME" start-stop-daemon --start --quiet --oknodo --exec $DAEMON echo "." ;; stop) echo -n "Stopping $DESC: $NAME" start-stop-daemon --stop --quiet --oknodo --exec $DAEMON echo "." ;; restart|force-reload) echo -n "Restarting $DESC: $NAME" start-stop-daemon --stop --quiet --exec $DAEMON sleep 1 start-stop-daemon --start --quiet --exec $DAEMON echo "." ;; *) echo "Usage: $0 {start|stop|restart|force-reload}" >&2 exit 1 ;; esac
exit 0
Dne středa 14 únor 2007 14:28 Gert Brinkmann napsal(a):
Ivan Arteaga wrote:
I recently installed dovecot 1.0rc15 on centos, installation went fine, it installed under /usr/local/ but I cannot start the dovecot service automatically.
Take a look to the start/stop-Scripts in /etc/init.d/. Take one of these (maybe "atd" or whatever is installed) and copy it to "dovecot". Such a script is simple to understand normally, so you can edit it to start/stop your manually installed dovecot easily. (Maybe there comes such a script within the dovecot-tarball? I am using the deb-package that is included in debian/sid that does set up everything automatically.)
Then you need to place symlinks to the runlevels to start/stop the daemon. Here on a debian-system there are:
/etc/rc0.d/K21dovecot -> ../init.d/dovecot /etc/rc1.d/K21dovecot -> ../init.d/dovecot /etc/rc2.d/S21dovecot -> ../init.d/dovecot /etc/rc3.d/S21dovecot -> ../init.d/dovecot /etc/rc4.d/S21dovecot -> ../init.d/dovecot /etc/rc5.d/S21dovecot -> ../init.d/dovecot /etc/rc6.d/K21dovecot -> ../init.d/dovecot
Maybe this helps?
Gert
Hi.,
At /etc/rc.d/init.d put a file named dovecot
############ ############ #!/bin/bash # # /etc/rc.d/init.d/dovecot # # Starts the dovecot daemon # # chkconfig: - 65 35 # description: Dovecot Imap Server # processname: dovecot # Source function library. . /etc/init.d/functions
test -x /usr/local/sbin/dovecot || exit 0
RETVAL=0 prog="Dovecot Imap"
start() { echo -n $"Starting $prog: " daemon /usr/local/sbin/dovecot RETVAL=$? [ $RETVAL -eq 0 ] && touch /var/lock/subsys/dovecot echo }
stop() { echo -n $"Stopping $prog: " killproc /usr/local/sbin/dovecot RETVAL=$? [ $RETVAL -eq 0 ] && rm -f /var/lock/subsys/dovecot echo }
# # See how we were called. # case "$1" in start) start ;; stop) stop ;; reload|restart) stop start RETVAL=$? ;; condrestart) if [ -f /var/lock/subsys/dovecot ]; then stop start fi ;; status) status /usr/local/sbin/dovecot RETVAL=$? ;; *) echo $"Usage: $0 {condrestart|start|stop|restart|reload|status}" exit 1 esac
exit $RETVAL
############ ############
And then "chkconfig dovecot on"
Luis.
-----Mensaje original----- De: dovecot-bounces@dovecot.org [mailto:dovecot-bounces@dovecot.org] En nombre de Ivan Arteaga Enviado el: Miércoles, 14 de Febrero de 2007 10:02 Para: 'Dovecot Mailing List' Asunto: [Dovecot] Start dovecot at boot.
Hi List,
I recently installed dovecot 1.0rc15 on centos, installation went fine, it installed under /usr/local/ but I cannot start the dovecot service automatically. I tried with chkconfig but it told me the dovecot service is not supported by chkconfig. Can anybody give me a tip about how to get done with this?
I will appreciate any help.
Regards,
--Ivan.
On Wed, 2007-02-14 at 11:48 +0100, Nicolas Boullis wrote:
Is there a way with dovecot to warn users who are close to reach their quota?
No.
Although I never used it, I think cyrus supports something like that by droping a warning message into the INBOX when delivering a message that brings the total size over a certain percentage.
Would it be feasible to implement something like that with an extra plugin or by changing the quota plugin?
Yes, I think it should be somewhat easy to implement into the quota plugin. Go ahead ;)
Although you'd have to somehow also check if the mailbox already has such a warning. But I guess it should be pretty easy too by searching for a message with the same subject.
Timo Sirainen wrote:
Yes, I think it should be somewhat easy to implement into the quota plugin. Go ahead ;)
I can try, but last time I tried to understand how plugins work, I somehow got lost... Is there any documentation I should read?
Although you'd have to somehow also check if the mailbox already has such a warning. But I guess it should be pretty easy too by searching for a message with the same subject.
Wouldn't it be easier to simply drop the warning if the usage was below x% before appending the message ans above after appending?
Cheers,
Nicolas
PS: no need to CC me, I do read the list
On Thu, 2007-02-15 at 14:21 +0100, Nicolas Boullis wrote:
Yes, I think it should be somewhat easy to implement into the quota plugin. Go ahead ;)
I can try, but last time I tried to understand how plugins work, I somehow got lost... Is there any documentation I should read?
Not really. In general the plugins will use the exact same APIs as Dovecot internally uses. lib-storage/mail-storage.h is the most important API.
Although you'd have to somehow also check if the mailbox already has such a warning. But I guess it should be pretty easy too by searching for a message with the same subject.
Wouldn't it be easier to simply drop the warning if the usage was below x% before appending the message ans above after appending?
I guess that would work too.
PS: no need to CC me, I do read the list
You can set Reply-To to the list for your messages in http://dovecot.org/cgi-bin/mailman/options/dovecot
I can't remember who wants Ccs and who don't, so I always just hit reply-to-all. And since I sometimes reply to messages weeks after the post, I don't think most people would notice the reply if I only send it to the list.
PS: no need to CC me, I do read the list
You can set Reply-To to the list for your messages in http://dovecot.org/cgi-bin/mailman/options/dovecot
I can't remember who wants Ccs and who don't, so I always just hit reply-to-all. And since I sometimes reply to messages weeks after the post, I don't think most people would notice the reply if I only send it to the list.
How long till a change takes effect? I changed mine over an hour ago, and still my replies default to the sender, not the list...
--
Best regards,
Charles
On Thu, 2007-02-15 at 13:07 -0500, Charles Marcus wrote:
PS: no need to CC me, I do read the list
You can set Reply-To to the list for your messages in http://dovecot.org/cgi-bin/mailman/options/dovecot
I can't remember who wants Ccs and who don't, so I always just hit reply-to-all. And since I sometimes reply to messages weeks after the post, I don't think most people would notice the reply if I only send it to the list.
How long till a change takes effect? I changed mine over an hour ago, and still my replies default to the sender, not the list...
Your last 3 messages did have the Reply-To: list header. So the change happens immediately.
I can't remember who wants Ccs and who don't, so I always just hit reply-to-all. And since I sometimes reply to messages weeks after the post, I don't think most people would notice the reply if I only send it to the list.
How long till a change takes effect? I changed mine over an hour ago, and still my replies default to the sender, not the list...
Your last 3 messages did have the Reply-To: list header. So the change happens immediately.
<smacks head>
I was hoping this would work for messages *I* *replied* *to*, not people who replied to *mine*...
Sorry... ;)
On Fri, 2007-02-16 at 10:49 -0500, Charles Marcus wrote:
I can't remember who wants Ccs and who don't, so I always just hit reply-to-all. And since I sometimes reply to messages weeks after the post, I don't think most people would notice the reply if I only send it to the list.
How long till a change takes effect? I changed mine over an hour ago, and still my replies default to the sender, not the list...
Your last 3 messages did have the Reply-To: list header. So the change happens immediately.
<smacks head>
I was hoping this would work for messages *I* *replied* *to*, not people who replied to *mine*...
But that's kind of the point of the header, to let everyone decide for themselves if they want to be Ccd or not. I don't think you should assume that everyone in the list want replies to go only to the list. I know in other mailing lists I do want people to Cc me because I don't read all the mails in the list and I could easily forget I had replied to some thread.
I was hoping this would work for messages *I* *replied* *to*, not people who replied to *mine*...
But that's kind of the point of the header, to let everyone decide for themselves if they want to be Ccd or not. I don't think you should assume that everyone in the list want replies to go only to the list. I know in other mailing lists I do want people to Cc me because I don't read all the mails in the list and I could easily forget I had replied to some thread.
I know - but this goes back to the old argument, to munge or not to munge.
I'm one of those that believes that - at least for discussion lists like this - the reply to should default to the list - and make the sender work if they want it to go direct to the individual.
As it is now, almost every message I send, I have to manually delete the 'To:', and then change the 'CC:' to 'To:'... just as I did on this one.
Of course, it would be nice if Thunderbird implemented a 'Reply-to-list' button/feature, but it isn't the only one that doesn't have one...
I hate getting duplicates, which is what happens when someone just hits 'Reply all', and is why I take the time to the above every time.
Anyway, no worries - dovecot is well worth this little aggravation... :)
--
Best regards,
Charles
On Fri, 2007-02-16 at 13:26 -0500, Charles Marcus wrote:
As it is now, almost every message I send, I have to manually delete the 'To:', and then change the 'CC:' to 'To:'... just as I did on this one.
But I have Reply-To set to this list, so this shouldn't be necessary? Or do you mean that Thunderbird's reply-to-all actually replies to the From address as well as to the Reply-To address? Or have you lost the Reply-To headers somewhere on the way?
Of course, it would be nice if Thunderbird implemented a 'Reply-to-list' button/feature, but it isn't the only one that doesn't have one...
I hate getting duplicates, which is what happens when someone just hits 'Reply all', and is why I take the time to the above every time.
Yes, but even though you want them it doesn't mean that everyone else wants them. That's why I always hit Reply-to-all and if anyone is bothered by the duplicates they can enable the Reply-to-list flag.
Timo Sirainen wrote:
On Fri, 2007-02-16 at 13:26 -0500, Charles Marcus wrote:
As it is now, almost every message I send, I have to manually delete the 'To:', and then change the 'CC:' to 'To:'... just as I did on this one.
But I have Reply-To set to this list,
Not for this one... There is no 'Reply-To' header. I don't think this is a Tbird problem either, because some of them *do* have it...
Weird... I just checked, and apparently your pref got reset somehow. The last message from you that has the Reply-to header set to the list was:
2/2/07, at 6:27am Subject: Re: [Dovecot] Quota crashing w/ gdb backtrace MessageID: 1170415621.3315.25.camel@hurina
The very next message, a mere three minutes later, there is no Reply-to header - and no message after that has one either.
That's why I always hit Reply-to-all and if anyone is bothered by the duplicates they can enable the Reply-to-list flag.
Yes - but I didn't even know about the ability to set these prefs until your last email about it... ;)
--
Best regards,
Charles
On Fri, 2007-02-16 at 14:00 -0500, Charles Marcus wrote:
Timo Sirainen wrote:
On Fri, 2007-02-16 at 13:26 -0500, Charles Marcus wrote:
As it is now, almost every message I send, I have to manually delete the 'To:', and then change the 'CC:' to 'To:'... just as I did on this one.
But I have Reply-To set to this list,
Not for this one... There is no 'Reply-To' header. I don't think this is a Tbird problem either, because some of them *do* have it...
Okay, I suppose my Mailman patching didn't go that great. Now it's set for me again, lets see if it stays on for you.. :)
On Fri, 2007-02-16 at 20:38 +0200, Timo Sirainen wrote:
On Fri, 2007-02-16 at 13:26 -0500, Charles Marcus wrote:
As it is now, almost every message I send, I have to manually delete the 'To:', and then change the 'CC:' to 'To:'... just as I did on this one.
But I have Reply-To set to this list, so this shouldn't be necessary?
I see you've added a Reply-To header later. The canonical response in this case is for someone to reference: http://www.unicom.com/pw/reply-to-harmful.html
Yes, but even though you want them it doesn't mean that everyone else wants them. That's why I always hit Reply-to-all and if anyone is bothered by the duplicates they can enable the Reply-to-list flag.
Or get a mail client that eliminates duplicates based on message IDs, or use a simple procmail recipe to eliminate duplicates based on message IDs, or any number of other solutions...
The fact that you're now doing Reply-To munging and yet arguing that people do the Right Thing for duplicates seems inconsistent.
Every time this discussion comes up anywhere, the answer is still the same... There's no reason for Reply-To munging except that people have bad software and refuse to fix it and would rather have a workaround implemented for them, at the cost of those of us who use decent software. I realize this is your list and you can do what you want, but my vote is strongly against Reply-To munging.
Richard
On Fri, 2007-02-16 at 15:16 -0600, Richard Laager wrote:
I see you've added a Reply-To header later. The canonical response in this case is for someone to reference: http://www.unicom.com/pw/reply-to-harmful.html
Yes, I've read that page. It says basically that it makes replying privately to the sender difficult. I'm sure it does, but in this list I don't think pretty much anyone ever wants to reply to someone only privately.
Except possibly me once in a while when I want some larger logs or whatever. Perhaps I should drop the Reply-To header from my messages. Although the current way works well for me, and I'm a bit lazy in configuring any duplicate-prevention systems.
Yes, but even though you want them it doesn't mean that everyone else wants them. That's why I always hit Reply-to-all and if anyone is bothered by the duplicates they can enable the Reply-to-list flag.
Or get a mail client that eliminates duplicates based on message IDs, or use a simple procmail recipe to eliminate duplicates based on message IDs, or any number of other solutions...
I don't really like this solution. First of all they all seem to decide that whatever message is received first is the one that stays. And usually the private reply is received first and the list reply later. And because I filter my mails with the List-ID header (the only right way to do it), this duplication avoidance system clearly breaks.
Another possible way would be to check the message's To and Cc headers and delete the mail if it contains dovecot@dovecot.org but didn't contain List-ID header. This way might lose mails, especially if the person who sent it wasn't joined to the list. So I don't really want to do that either.
Third and the only really working way would be not to do anything before both messages have been received, and only then decide which one to drop. But how should this be done? Should there be a queue where the messages are waiting before they're written to the actual mailbox? How long should they stay in the queue? It can't be too short, but too long is also annoying. Are there any existing solutions that does it this way? I somehow doubt it.
This all could be solved nicely if there was a header that means "don't include me in the reply-to-all list". I guess the Mail-Followup-To header would do this. Hmm. Do enough clients nowadays support that? IIRC mutt supported it, Thunderbird seems to be supporting it now that I'm googleing, what about others? If support for it is good enough I could change Mailman to add that header instead of Reply-To.
On Sat, Feb 17, 2007 at 12:33:40AM +0200, Timo Sirainen wrote:
On Fri, 2007-02-16 at 15:16 -0600, Richard Laager wrote:
I see you've added a Reply-To header later. The canonical response in this case is for someone to reference: http://www.unicom.com/pw/reply-to-harmful.html
A canonical response to that is
http://www.metasystema.net/essays/reply-to.mhtml
This all could be solved nicely if there was a header that means "don't include me in the reply-to-all list".
Active threads tend to pick up contributors: how would you try to honor all of the preferences of all of the people who might end up in the cc list? You can't really trust the existing CC list to have been properly vetted (for some version of "proper").
With regard to reply-to munging and cc preferences you just get into an unmanageable matrix of conflicting opinions and wishes. It's just better to:
be deliberate about where you're sending, so you aren't hapharzardly at the whim of your tools and environment. If you make a mistake, at least do it on purpose :)
manage your own inbox, don't ask others to remember your preferences.
You really can't please everybody.
Incidentally, my approach, for any list where I think discussions ought to stay on list, is to guarantee a Mail-Followup-To header in message as it comes into my inbox, and to use a mail reader that looks at it when I ask it to reply to the list and which conversely ignores it when I ask it to reply to the sender. (I've gotten flack for this from people who for some reason think I'm not allowed to add this header on my own mail sitting in my own inbox, but naturally I don't agree with them.) This gives me the choice of behaviors that I want. And I still review the outgoing header before I let it fly.
Hmm, I guess this must be my day for giving bicycle shed opinions.
http://en.wikipedia.org/wiki/Wikipedia:Avoid_Parkinson's_Bicycle_Shed_Effect
-mm-
On Saturday 17 February 2007 00:13, Mark E. Mallett wrote:
On Sat, Feb 17, 2007 at 12:33:40AM +0200, Timo Sirainen wrote:
On Fri, 2007-02-16 at 15:16 -0600, Richard Laager wrote:
I see you've added a Reply-To header later. The canonical response in this case is for someone to reference: http://www.unicom.com/pw/reply-to-harmful.html
A canonical response to that is
My mind was not changed, sorry. First, RFC 822 is obsolete, and RFC 2822 does not contain the quoted paragraph. Instead it says that '[w]hen the "Reply-To:" field is present, it indicates the mailbox(es) to which the author of the message suggests that replies be sent.' - the author, not the mailing list software. There are other fields for that.
The rest of the arguments assume that lack of a "reply to list" or "follow up" command is the natural order of things. Sadly, almost no mailers do provide such a command, but munging Reply-To doesn't exactly increase pressure on vendors to implement it.
I guess this is a fight between ideal and reality, i.e. between agreeing on a direction and going there and staying where we are because no single individual has an incentive to take the leap. This is where the market fails.
Interestingly, Kmail seems to have some "anti-munging" feature - when the Reply-To address equals the List-Post address, I guess, the reply to author command ignores the Reply-To field. This doesn't work on some other mailing lists I'm on, which don't have List-* header fields.
-- Magnus Holmgren holmgren@lysator.liu.se (No Cc of list mail needed, thanks)
"Exim is better at being younger, whereas sendmail is better for Scrabble (50 point bonus for clearing your rack)" -- Dave Evans
Hi,
Timo Sirainen wrote:
Although I never used it, I think cyrus supports something like that by droping a warning message into the INBOX when delivering a message that brings the total size over a certain percentage.
Would it be feasible to implement something like that with an extra plugin or by changing the quota plugin?
Yes, I think it should be somewhat easy to implement into the quota plugin. Go ahead ;)
This is work in progress, but I just figured out the the quota implementation in 1.0(rc) is quite different from the quota implementation in CVS HEAD. Where should I better try to implement those warnings?
Cheers,
Nicolas
participants (11)
-
Axel Thimm
-
Charles Marcus
-
Gert Brinkmann
-
Ivan Arteaga
-
lampacz@gmail.com
-
Luis Venegas
-
Magnus Holmgren
-
Mark E. Mallett
-
Nicolas Boullis
-
Richard Laager
-
Timo Sirainen