Re: [Dovecot] Delivered-To header without +extension ?
On 2011-01-21 4:30 AM, Per Jessen wrote: > Charles Marcus wrote: >> On 2011-01-20 4:06 AM, Per Jessen wrote: >>> I've been reading >>> a bit, and I think the issue is that postfix adds X-Original-To when >>> delivering to a mailbox - which delivery via smtp/lmtp isn't. >>> >>> I'm not sure if postfix should be adding it - postfix applies >>> virtual_aliases_maps, then delivers to dovecot via lmtp (set up via >>> virtual_transport) - without X-Original-To, the information >>> of "original recipient" is lost at this point. >> >> Yikes... I've been planning on switching to LMTP for delivery, but >> this would be a show-stopper... >> >> Please keep us updated on what you find out... > It looks like the issue was discussed here: > > http://marc.info/?l=postfix-users&m=118852762117587 > > Wietse concludes that the virtual aliasing would be better done on the > final station, i.e. dovecot. Personally I don't need the X-Original-To > header, but it does seem like it ought to be written by is > chosen as virtual_transport, rather than only virtual or pipe. Thanks... Thoughts Timo? -- Best regards, Charles
Re: [Dovecot] Delivered-To header without +extension ?
Charles Marcus wrote: > On 2011-01-20 4:06 AM, Per Jessen wrote: >> I've been reading >> a bit, and I think the issue is that postfix adds X-Original-To when >> delivering to a mailbox - which delivery via smtp/lmtp isn't. >> >> I'm not sure if postfix should be adding it - postfix applies >> virtual_aliases_maps, then delivers to dovecot via lmtp (set up via >> virtual_transport) - without X-Original-To, the information >> of "original recipient" is lost at this point. > > Yikes... I've been planning on switching to LMTP for delivery, but > this would be a show-stopper... > > Please keep us updated on what you find out... It looks like the issue was discussed here: http://marc.info/?l=postfix-users&m=118852762117587 Wietse concludes that the virtual aliasing would be better done on the final station, i.e. dovecot. Personally I don't need the X-Original-To header, but it does seem like it ought to be written by is chosen as virtual_transport, rather than only virtual or pipe. /Per Jessen, Zürich
Re: [Dovecot] Delivered-To header without +extension ?
On 2011-01-20 4:06 AM, Per Jessen wrote: > I've been reading > a bit, and I think the issue is that postfix adds X-Original-To when > delivering to a mailbox - which delivery via smtp/lmtp isn't. > > I'm not sure if postfix should be adding it - postfix applies > virtual_aliases_maps, then delivers to dovecot via lmtp (set up via > virtual_transport) - without X-Original-To, the information > of "original recipient" is lost at this point. Yikes... I've been planning on switching to LMTP for delivery, but this would be a show-stopper... Please keep us updated on what you find out... -- Best regards, Charles
Re: [Dovecot] Delivered-To header without +extension ?
Am 20.01.2011 10:06, schrieb Per Jessen: > Robert Schetterer wrote: > >> Am 20.01.2011 09:41, schrieb Per Jessen: >>> Tom Hendrikx wrote: >>> On 20/01/11 08:50, Per Jessen wrote: > Per Jessen wrote: > >> Pascal Volk wrote: >> >>> Hi Per, >>> >>> now the +ext is included in the Delivered-To header again: >>> http://hg.dovecot.org/dovecot-2.0/rev/a3a7cc0172fd >>> >> >> Thanks Pascal, that was fast! >> >> Last night, I reverse applied the patch you mentioned earlier to >> 2.0.9, which worked just fine, I'm building it just now. > > Probably superfluous, but nevertheless - it works fine, I'm getting > the > right Deliver-To header including the +extension. Interestingly, > I'm not seeing X-Original-To - isn't that normally added by > postfix? > > > /Per Jessen, Zürich > X-Original-To: header is added by postfix' pipe(8) command, >>> >>> Hmm, it's can't be only pipe() - if I revert to regular virtual >>> delivery to maildir (instead of lmtp to dovecot), I get the >>> X-Original-To header, and that involves no pipe(). >>> and is only available when delivering to a single recipient (_destination_recipient_limit = 1). >>> >>> I'll try that. >>> >>> >>> /Per Jessen, Zürich >>> >> >> if have no idea if this help , but its easy to try, after all you >> loose performance with lmtp if you set 1 here >> >> lmtp_destination_recipient_limit (default: >> $default_destination_recipient_limit) > > Hi Robert > > yes, I've just tried that, but it made no difference. I've been reading > a bit, and I think the issue is that postfix adds X-Original-To when > delivering to a mailbox - which delivery via smtp/lmtp isn't. > > I'm not sure if postfix should be adding it - postfix applies > virtual_aliases_maps, then delivers to dovecot via lmtp (set up via > virtual_transport) - without X-Original-To, the information > of "original recipient" is lost at this point. > sounds plausible to make sure you may ask Wietse perhaps feature request or some magical setup tip may accour *g greetz from Munich to Zuerich > > /Per Jessen, Zürich > -- Best Regards MfG Robert Schetterer Germany/Munich/Bavaria
Re: [Dovecot] Delivered-To header without +extension ?
Robert Schetterer wrote: > Am 20.01.2011 09:41, schrieb Per Jessen: >> Tom Hendrikx wrote: >> >>> On 20/01/11 08:50, Per Jessen wrote: Per Jessen wrote: > Pascal Volk wrote: > >> Hi Per, >> >> now the +ext is included in the Delivered-To header again: >> http://hg.dovecot.org/dovecot-2.0/rev/a3a7cc0172fd >> > > Thanks Pascal, that was fast! > > Last night, I reverse applied the patch you mentioned earlier to > 2.0.9, which worked just fine, I'm building it just now. Probably superfluous, but nevertheless - it works fine, I'm getting the right Deliver-To header including the +extension. Interestingly, I'm not seeing X-Original-To - isn't that normally added by postfix? /Per Jessen, Zürich >>> >>> X-Original-To: header is added by postfix' pipe(8) command, >> >> Hmm, it's can't be only pipe() - if I revert to regular virtual >> delivery to maildir (instead of lmtp to dovecot), I get the >> X-Original-To header, and that involves no pipe(). >> >>> and is only available when delivering to a single recipient >>> (_destination_recipient_limit = 1). >> >> I'll try that. >> >> >> /Per Jessen, Zürich >> > > if have no idea if this help , but its easy to try, after all you > loose performance with lmtp if you set 1 here > > lmtp_destination_recipient_limit (default: > $default_destination_recipient_limit) Hi Robert yes, I've just tried that, but it made no difference. I've been reading a bit, and I think the issue is that postfix adds X-Original-To when delivering to a mailbox - which delivery via smtp/lmtp isn't. I'm not sure if postfix should be adding it - postfix applies virtual_aliases_maps, then delivers to dovecot via lmtp (set up via virtual_transport) - without X-Original-To, the information of "original recipient" is lost at this point. /Per Jessen, Zürich
Re: [Dovecot] Delivered-To header without +extension ?
Am 20.01.2011 09:41, schrieb Per Jessen: > Tom Hendrikx wrote: > >> On 20/01/11 08:50, Per Jessen wrote: >>> Per Jessen wrote: >>> Pascal Volk wrote: > Hi Per, > > now the +ext is included in the Delivered-To header again: > http://hg.dovecot.org/dovecot-2.0/rev/a3a7cc0172fd > Thanks Pascal, that was fast! Last night, I reverse applied the patch you mentioned earlier to 2.0.9, which worked just fine, I'm building it just now. >>> >>> Probably superfluous, but nevertheless - it works fine, I'm getting >>> the >>> right Deliver-To header including the +extension. Interestingly, I'm >>> not seeing X-Original-To - isn't that normally added by postfix? >>> >>> >>> /Per Jessen, Zürich >>> >> >> X-Original-To: header is added by postfix' pipe(8) command, > > Hmm, it's can't be only pipe() - if I revert to regular virtual delivery > to maildir (instead of lmtp to dovecot), I get the X-Original-To > header, and that involves no pipe(). > >> and is only available when delivering to a single recipient >> (_destination_recipient_limit = 1). > > I'll try that. > > > /Per Jessen, Zürich > if have no idea if this help , but its easy to try, after all you loose performance with lmtp if you set 1 here lmtp_destination_recipient_limit (default: $default_destination_recipient_limit) The maximal number of recipients per message for the lmtp message delivery transport. This limit is enforced by the queue manager. The message delivery transport name is the first field in the entry in the master.cf file. Setting this parameter to a value of 1 changes the meaning of lmtp_destination_concurrency_limit from concurrency per domain into concurrency per recipient. -- Best Regards MfG Robert Schetterer Germany/Munich/Bavaria
Re: [Dovecot] Delivered-To header without +extension ?
Tom Hendrikx wrote: > On 20/01/11 08:50, Per Jessen wrote: >> Per Jessen wrote: >> >>> Pascal Volk wrote: >>> Hi Per, now the +ext is included in the Delivered-To header again: http://hg.dovecot.org/dovecot-2.0/rev/a3a7cc0172fd >>> >>> Thanks Pascal, that was fast! >>> >>> Last night, I reverse applied the patch you mentioned earlier to >>> 2.0.9, which worked just fine, I'm building it just now. >> >> Probably superfluous, but nevertheless - it works fine, I'm getting >> the >> right Deliver-To header including the +extension. Interestingly, I'm >> not seeing X-Original-To - isn't that normally added by postfix? >> >> >> /Per Jessen, Zürich >> > > X-Original-To: header is added by postfix' pipe(8) command, Hmm, it's can't be only pipe() - if I revert to regular virtual delivery to maildir (instead of lmtp to dovecot), I get the X-Original-To header, and that involves no pipe(). > and is only available when delivering to a single recipient > (_destination_recipient_limit = 1). I'll try that. /Per Jessen, Zürich
Re: [Dovecot] Delivered-To header without +extension ?
On 20/01/11 08:50, Per Jessen wrote: > Per Jessen wrote: > >> Pascal Volk wrote: >> >>> Hi Per, >>> >>> now the +ext is included in the Delivered-To header again: >>> http://hg.dovecot.org/dovecot-2.0/rev/a3a7cc0172fd >>> >> >> Thanks Pascal, that was fast! >> >> Last night, I reverse applied the patch you mentioned earlier to >> 2.0.9, which worked just fine, I'm building it just now. > > Probably superfluous, but nevertheless - it works fine, I'm getting the > right Deliver-To header including the +extension. Interestingly, I'm > not seeing X-Original-To - isn't that normally added by postfix? > > > /Per Jessen, Zürich > X-Original-To: header is added by postfix' pipe(8) command, and is only available when delivering to a single recipient (_destination_recipient_limit = 1). I don't see any options in the smtp/lmtp manpage that enable adding a header like this, which is logically since both support multiple recipients by protocol spec, iirc. -- Tom signature.asc Description: OpenPGP digital signature
Re: [Dovecot] Delivered-To header without +extension ?
Per Jessen wrote: > Pascal Volk wrote: > >> Hi Per, >> >> now the +ext is included in the Delivered-To header again: >> http://hg.dovecot.org/dovecot-2.0/rev/a3a7cc0172fd >> > > Thanks Pascal, that was fast! > > Last night, I reverse applied the patch you mentioned earlier to > 2.0.9, which worked just fine, I'm building it just now. Probably superfluous, but nevertheless - it works fine, I'm getting the right Deliver-To header including the +extension. Interestingly, I'm not seeing X-Original-To - isn't that normally added by postfix? /Per Jessen, Zürich
Re: [Dovecot] Delivered-To header without +extension ?
Pascal Volk wrote: > Hi Per, > > now the +ext is included in the Delivered-To header again: > http://hg.dovecot.org/dovecot-2.0/rev/a3a7cc0172fd > Thanks Pascal, that was fast! Last night, I reverse applied the patch you mentioned earlier to 2.0.9, which worked just fine, I'm building it just now. /Per Jessen, Zürich
Re: [Dovecot] Delivered-To header without +extension ?
Hi Per, now the +ext is included in the Delivered-To header again: http://hg.dovecot.org/dovecot-2.0/rev/a3a7cc0172fd Regards, Pascal -- The trapper recommends today: fabaceae.1101...@localdomain.org
Re: [Dovecot] Delivered-To header without +extension ?
Pascal Volk wrote: > On 01/19/2011 07:58 PM Per Jessen wrote: >> According to http://www.postfix.org/local.8.html, the Delivered-To >> header is part of the qmail system. >> >> HISTORY >>The Delivered-To: message header appears in the qmail sys- >>tem by Daniel Bernstein. >> >> "In order to stop mail forwarding loops early, the software adds an >> optional Delivered-To: header with the final envelope recipient >> address." >> >> I need the +extension, so I interpret "the final envelope recipient >> address" to include the +extension :-) > > > Hum … Postfix' local adds in fact the Delivered-To header with the > full address, like: addr+...@doma.in > What about the X-Original-To header? Does it also contain the +ext? Yes it does. /Per Jessen, Zürich
Re: [Dovecot] Delivered-To header without +extension ?
On 01/19/2011 07:58 PM Per Jessen wrote: > According to http://www.postfix.org/local.8.html, the Delivered-To > header is part of the qmail system. > > HISTORY >The Delivered-To: message header appears in the qmail sys- >tem by Daniel Bernstein. > > "In order to stop mail forwarding loops early, the software adds an > optional Delivered-To: header with the final envelope recipient > address." > > I need the +extension, so I interpret "the final envelope recipient > address" to include the +extension :-) Hum … Postfix' local adds in fact the Delivered-To header with the full address, like: addr+...@doma.in What about the X-Original-To header? Does it also contain the +ext? Regards, Pascal -- The trapper recommends today: c01dcafe.1101...@localdomain.org
Re: [Dovecot] Delivered-To header without +extension ?
Pascal Volk wrote: > On 01/19/2011 06:48 PM Per Jessen wrote: >> … Was there any specific reason for having dovecot remove the >> +extension? > > Timo and I talked about the Delivered-To w/ or w/o +ext on irc > (#dovecot). I asked him to remove the extension from the Delivered-To > header. Before that I've checked the headers inserted by Postfix. > > Knows someone what is the correct address format in the Delivered-To > header? Any RFC, …? According to http://www.postfix.org/local.8.html, the Delivered-To header is part of the qmail system. HISTORY The Delivered-To: message header appears in the qmail sys- tem by Daniel Bernstein. "In order to stop mail forwarding loops early, the software adds an optional Delivered-To: header with the final envelope recipient address." I need the +extension, so I interpret "the final envelope recipient address" to include the +extension :-) /Per Jessen, Zürich
Re: [Dovecot] Delivered-To header without +extension ?
On 01/19/2011 06:48 PM Per Jessen wrote: > … Was there any specific reason for having dovecot remove the +extension? Timo and I talked about the Delivered-To w/ or w/o +ext on irc (#dovecot). I asked him to remove the extension from the Delivered-To header. Before that I've checked the headers inserted by Postfix. Knows someone what is the correct address format in the Delivered-To header? Any RFC, …? Regards, Pascal -- The trapper recommends today: c01dcofe.1101...@localdomain.org
Re: [Dovecot] Delivered-To header without +extension ?
Pascal Volk wrote: > On 01/19/2011 06:27 PM Per Jessen wrote: >> I'm running postfix 2.6.2 and dovecot 2.0.9, and I've just switched >> to delivering mails to dovecot via lmtp. It looks like the >> Delivered-To: header is slightly different - it doesn't seem to >> include the +extension? Is this intentional or does the +extension >> somehow get in the way? > > The +ext was removed on July 20th 2010: > http://hg.dovecot.org/dovecot-2.0/rev/19231e3585db > > Maybe you can revert that patch to get the +ext back. Thanks, I'll do that. I've switched to letting postfix deliver directly, which will continue to give me the right Delivered-To: header for the moment, but eventually I want to move to lmtp delivery. Was there any specific reason for having dovecot remove the +extension? /Per Jessen, Zürich
Re: [Dovecot] Delivered-To header without +extension ?
On 01/19/2011 06:27 PM Per Jessen wrote: > I'm running postfix 2.6.2 and dovecot 2.0.9, and I've just switched to > delivering mails to dovecot via lmtp. It looks like the Delivered-To: > header is slightly different - it doesn't seem to include the > +extension? Is this intentional or does the +extension somehow get in > the way? The +ext was removed on July 20th 2010: http://hg.dovecot.org/dovecot-2.0/rev/19231e3585db Maybe you can revert that patch to get the +ext back. Regards, Pascal -- The trapper recommends today: c01dcofe.1101...@localdomain.org