Re: [courier-users] Looped DSNs when empty sender and broken user's forward
Sam Varshavchik wrote: Alessandro Vesely writes: sensible information, coordinating mail filters, et cetera. Besides properly setting the Return-Path, a policy might need to mandate a specific outgoing IP, a smart relay, some kind of authorization token, or other features. Would it be worth to plan such a general solution for Courier? That's what maildrop is for. Once a decision is made to forward mail, there's no substantial difference between what maildrop does and what Courier does. Courier invokes submit, while maildrop invokes sendmail which invokes submit. I guess you're right. I shall get rid of naive .courier forwarding and reserve .courier files for very specific tasks, such as, e.g., mailing lists. (A mailing list _is a_ type of forwarding.) BTW, sendmail is invoked through a maildrop variable, which provides an additional way for setting up whatever policy, orthogonally across all forwarding recipes, except .courier driven ones. Thank you for the tip - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08 ___ courier-users mailing list courier-users@lists.sourceforge.net Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
Re: [courier-users] Looped DSNs when empty sender and broken user's forward
On Cz, 2008-07-03 at 07:57 -0700, Gordon Messmer wrote: Double-bounces aren't usually a problem. Courier already detects the situation that a message can't be delivered, and the DSN can't be delivered either. Those bounces go to the postmaster. If your postmaster account is the one bouncing mail, then I suggest you set up a postmaster address that doesn't use mailfilters that are prone to breaking. Just create a dot-courier file in your home directory and specify a delivery rule that indicates a Maildir path. alias: postmaster: ptecza-postmaster /home/ptecza/.courier-postmaster: ./Maildir/.Postmaster/. Hi Gordon, Thank you very much for your reply! Unfortunately, my postmaster account is used not only by me. It's simply an alias for all admins of our mail system. I asked Sam, because I couldn't see any good reasons for setting SENDER=$FROM. My best regards, Pawel signature.asc Description: This is a digitally signed message part - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08___ courier-users mailing list courier-users@lists.sourceforge.net Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
Re: [courier-users] Looped DSNs when empty sender and broken user's forward
On Cz, 2008-07-03 at 18:26 -0400, Sam Varshavchik wrote: Paweł Tęcza writes: So, my question is: why don't you set SENDER='[EMAIL PROTECTED]' or SENDER='[EMAIL PROTECTED]' or something similar if SENDER variable is empty? In that case, if user's forward is broken, then Courier is trying to deliver a DSN message to a postmaster, not to a user with not working forward. I think it's reasonable solution, because information about broken user's forward is valuable for mail system administrator. What's your opinion? Yes, I think that's reasonable. The original reason for resetting the return address would be to be notified if forwarded mail bounces. I expected people to set up forwarding rules carefully, and take care not to indiscriminantly forward all mail, rather only ones that match narrow filtering criteria. But I suppose that's too much to expect, from people, so I guess this will have to change. I'll change the return address on forwarded mail to a null address, which will bounce to the postmaster's mailbox automatically. Hello Sam, Thanks a lot for your feedback! I'm glad to see you will change it :) Have a nice coding day, Pawel signature.asc Description: This is a digitally signed message part - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08___ courier-users mailing list courier-users@lists.sourceforge.net Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
Re: [courier-users] Looped DSNs when empty sender and broken user's forward
Sam Varshavchik writes: Paweł Tęcza writes: So, my question is: why don't you set SENDER='[EMAIL PROTECTED]' or SENDER='[EMAIL PROTECTED]' or something similar if SENDER variable is empty? In that case, if user's forward is broken, then Courier is trying to deliver a DSN message to a postmaster, not to a user with not working forward. I think it's reasonable solution, because information about broken user's forward is valuable for mail system administrator. Yes, I think that's reasonable. The original reason for resetting the return address would be to be notified if forwarded mail bounces. I expected people to set up forwarding rules carefully, and take care not to indiscriminantly forward all mail, rather only ones that match narrow filtering criteria. But I suppose that's too much to expect, from people, so I guess this will have to change. I'll change the return address on forwarded mail to a null address, which will bounce to the postmaster's mailbox automatically. That seems a somewhat simplistic solution to me... What about plain courier deliveries, i.e. forwarding without maildrop? IMHO, there should be a generic framework for setting forwarding policies and establish criteria for determining the type of any forwarding rule. For example, newsletter exploders, anonymous mailboxes, changed email address, secondary mail servers, all result in different types of forwarding. The reasons why different policies are needed, originate from concerns such as determining the responsibility for forwarding spam, avoiding to disclose sensible information, coordinating mail filters, et cetera. Besides properly setting the Return-Path, a policy might need to mandate a specific outgoing IP, a smart relay, some kind of authorization token, or other features. Would it be worth to plan such a general solution for Courier? I mean, we all have various forwarding rules scattered here and there, and don't want to have to upgrade them one by one whenever a new antispam forwarding mode comes about. Would it be convenient/feasible/cool to plan something in advance? - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08 ___ courier-users mailing list courier-users@lists.sourceforge.net Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
Re: [courier-users] Looped DSNs when empty sender and broken user's forward
Alessandro Vesely writes: What about plain courier deliveries, i.e. forwarding without maildrop? IMHO, there should be a generic framework for setting forwarding policies and establish criteria for determining the type of any forwarding rule. For example, newsletter exploders, anonymous mailboxes, changed email address, secondary mail servers, all result in different types of forwarding. The How do you define them, specifically. You'll end up with maildrop. reasons why different policies are needed, originate from concerns such as determining the responsibility for forwarding spam, avoiding to disclose sensible information, coordinating mail filters, et cetera. Besides properly setting the Return-Path, a policy might need to mandate a specific outgoing IP, a smart relay, some kind of authorization token, or other features. Would it be worth to plan such a general solution for Courier? I mean, we all That's what maildrop is for. Once a decision is made to forward mail, there's no substantial difference between what maildrop does and what Courier does. Courier invokes submit, while maildrop invokes sendmail which invokes submit. pgpw6krk2cZAy.pgp Description: PGP signature - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08___ courier-users mailing list courier-users@lists.sourceforge.net Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
[courier-users] Looped DSNs when empty sender and broken user's forward
Hello Sam,
I would like to ask you about setting SENDER=$FROM in user's mailfilter
file when SENDER variable is empty (please see
maildir_filter_saverules() function in maildir/maildirfilter.c file):
import SENDER
if ($SENDER eq )
{
SENDER=$FROM
}
Unfortunately, that maildrop filter code can be dangerous for a mail
server if a user has broken forward and he received a message from
empty address.
Recently we have had that situation in our mail system. Courier server
couldn't deliver a message to user's external account, so it was trying
to deliver DSN to our user, because the maildrop filter code has set
his internal address as a return address. The DSN message wasn't
delivered too, because user's forward didn't work, etc, etc. In result
we had looped DSNs from MAILDER-DAEMON...
I noticed it watching mail queue peak on our Munin graph. Fortunately
we have strong servers, so it wasn't noticeable for the rest our users.
So, my question is: why don't you set SENDER='[EMAIL PROTECTED]'
or SENDER='[EMAIL PROTECTED]' or something similar if SENDER
variable is empty?
In that case, if user's forward is broken, then Courier is trying
to deliver a DSN message to a postmaster, not to a user with not
working forward. I think it's reasonable solution, because information
about broken user's forward is valuable for mail system administrator.
What's your opinion?
My best regards,
Pawel
signature.asc
Description: This is a digitally signed message part
-
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08___
courier-users mailing list
courier-users@lists.sourceforge.net
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
Re: [courier-users] Looped DSNs when empty sender and broken user's forward
Double-bounces aren't usually a problem. Courier already detects the situation that a message can't be delivered, and the DSN can't be delivered either. Those bounces go to the postmaster. If your postmaster account is the one bouncing mail, then I suggest you set up a postmaster address that doesn't use mailfilters that are prone to breaking. Just create a dot-courier file in your home directory and specify a delivery rule that indicates a Maildir path. alias: postmaster: ptecza-postmaster /home/ptecza/.courier-postmaster: ./Maildir/.Postmaster/. - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08 ___ courier-users mailing list courier-users@lists.sourceforge.net Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
Re: [courier-users] Looped DSNs when empty sender and broken user's forward
Paweł Tęcza writes: So, my question is: why don't you set SENDER='[EMAIL PROTECTED]' or SENDER='[EMAIL PROTECTED]' or something similar if SENDER variable is empty? In that case, if user's forward is broken, then Courier is trying to deliver a DSN message to a postmaster, not to a user with not working forward. I think it's reasonable solution, because information about broken user's forward is valuable for mail system administrator. What's your opinion? Yes, I think that's reasonable. The original reason for resetting the return address would be to be notified if forwarded mail bounces. I expected people to set up forwarding rules carefully, and take care not to indiscriminantly forward all mail, rather only ones that match narrow filtering criteria. But I suppose that's too much to expect, from people, so I guess this will have to change. I'll change the return address on forwarded mail to a null address, which will bounce to the postmaster's mailbox automatically. pgp8TF13IRkyd.pgp Description: PGP signature - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08___ courier-users mailing list courier-users@lists.sourceforge.net Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
