qmail-spp-bugs Mailing List for qmail-spp - qmail SMTP plugin patch
Brought to you by:
pavcio
You can subscribe to this list here.
2004 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(2) |
Aug
|
Sep
(3) |
Oct
|
Nov
|
Dec
|
---|---|---|---|---|---|---|---|---|---|---|---|---|
2006 |
Jan
|
Feb
|
Mar
(2) |
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
(1) |
Sep
|
Oct
|
Nov
|
Dec
|
2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(2) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
2010 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(1) |
Sep
|
Oct
|
Nov
|
Dec
|
From: vislavie <vis...@gm...> - 2010-08-10 16:50:43
|
Hi All I setup netqmail-1.06 with vpopmail-5.4.27 qmail-authentication-071 qmail-spp-042 plugin rblspp It worked :-) The smtp-daemon is running from recordio with daemontools so I can see that the authentication works fine. Then I tried the authlogger plugin and got the message "authlogger: pid 14000 - smtp user authenticated as xxx@xxx.xx" that's ok and "503 you're already authenticated (#5.5.0)" in log. Thunderbird don't send the message and popup a window with the "503" Server said message, Outlook shows me the loginwindow in a loop, stupid outlook. I wrote a testplugin, the same. If fprinf put output to stderr the error occurs. Any ideas? Chears Joc |
From: Jim W. <Ji...@hy...> - 2007-12-03 13:00:48
|
=20 After I installed the greylisting plugin I was looking at the database entries it had created and noticed that the recipient address was the same as the sender. After some testing I found that moving: =20 if (!(spp_val =3D spp_rcpt(allowed))) return; =20 from within the smtp_mail section of qmail-smtp.c to the smtp_rcpt section fixed the problem. At this point I have it at the bottom of that function, which has been patched for LDAP lookups. I'm not to sure if it should be higher up, but for the moment, it works. =20 =20 =20 |
From: Pawel F. <pj...@as...> - 2007-05-27 12:11:21
|
Hi Chris, On Wednesday 23 May 2007 17:30, Christoph Probst wrote: > I haven't looked into the SPP code but I guess that after the RCPT command > the plugins are executed prior to the rcpthosts/morercpthosts check. I > strongly recommend to switch these checks. Well that's not a bug, that's a feature ;-) Actually, the rcpthosts is checked *prior* to executing the plugins, but the result is used *after*. This is to make it possible for a plugin to be aware of this result and optionally skip this rcpthosts check (using the "O" command). I don't remember why I've added it, probably someone requested this, but please give me some more time to meditate on it and come up with an ultimate solution :-). As a temporary workaround, you may write a tiny plugin checking if SMTPRCPTHOSTSOK=1 and exiting with well-known "E553 sorry, that domain isn't in my list (...)". By adding it as first to the chain of plugins, you should minimize the impact of this bug. Any ideas welcome, as always. Bye, -- Pawel Foremski pj...@as... |
From: Christoph P. <chr...@ne...> - 2007-05-23 15:52:13
|
Hi, I'm using netqmail 1.05-r4 as shipped with Gentoo Linux and that includes=20 support for SPP. Today there was a spam wave hitting my server on one domai= n=20 name and suddenly the load went crazy. Trying to remove the effected domain from 'morercpthosts' didn't help and I= =20 found that all plugins were still executed although my server wasn't=20 responsible for these domains anymore. I haven't looked into the SPP code but I guess that after the RCPT command = the=20 plugins are executed prior to the rcpthosts/morercpthosts check. I strongly= =20 recommend to switch these checks. Quick workaround was to move the Plugins to be executed after the DATA comm= and=20 but that is definitly the wrong place for a 'RCPT not found' plugin ... Please copy all replies to me as I'm not subscribed to this mailinglist. Cheers, Chris =2D-=20 Kontakt-Details: http://www.christoph-probst.com/kontakt/ PGP-FP: B171 7EA4 988C DD90 1601 D21C 5279 2FAF 9978 AF86 |
From: Biro D. <da...@da...> - 2006-08-08 06:42:22
|
Hi! I've noticed that when using the vpopmail_check_recipient plugin, the catchall feature of qmailadmin doesn't work. The catchall setting in qmailadmin writes the following code in the .qmail-default: | /var/vpopmail/bin/vdelivermail '' /var/vpopmail/domains/something.com/catchallaccount Is there a workaround for this? Thanks in advance, Daniel |
From: Pawel F. <pj...@as...> - 2006-04-10 12:48:48
|
On Sunday 26 March 2006 11:09, Pawel Foremski wrote: > If it's not urgent, please be patient as I'm in constant lack of time. If > it is, yell at me ;-) > > I'll try to fix this till end of this month. Case closed. -- Pawel Foremski pj...@as... |
From: Pawel F. <pj...@as...> - 2006-03-26 09:09:36
|
On Thursday 23 March 2006 23:21, Roberto Alsina wrote: > The problem is when you try to implement a rcpt plugin that checks for > invalid local addresses. > > When you are authenticated, or you have relay, you have relayclient > enabled, so allowed is never set, and remains (at least here ;-) set to 0. Acked. If it's not urgent, please be patient as I'm in constant lack of time. If it is, yell at me ;-) I'll try to fix this till end of this month. Thanks. Bye, -- Pawel Foremski pj...@as... |
From: Roberto A. <ra...@kd...> - 2006-03-23 22:21:42
|
Hello, I wrote a plugin that reimplements the functionality of chkuser2 using SPP= =20 (http://www.interazioni.it/opensource/chkuser/). If anyone wants a beta=20 version let me know :-) But, I have run into something that looks like a SPP bug. In qmail-smtpd.c, the patch puts this in smtp_rcpt: if (!relayclient) allowed =3D addrallowed(); And later if (!(spp_val =3D spp_rcpt(allowed) return; The problem is when you try to implement a rcpt plugin that checks for inva= lid=20 local addresses. When you are authenticated, or you have relay, you have relayclient enabled= ,=20 so allowed is never set, and remains (at least here ;-) set to 0. So, all addreses seem to be non-rcpthosts, and no validation occurs. I changed it to=20 allowed =3D addrallowed(); and have not had any adverse effects, authentication and relaying still see= m=20 to work pretty much in the usual manner. I am using a quite heavily patched qmail, but the auth patch is from=20 http://www.fehcom.de/qmail.html =2D-=20 =A0("\''/").__..-''"`-. . =A0 =A0 =A0 =A0 Roberto Alsina =A0`9_ 9 =A0) =A0 `-. ( =A0 =A0).`-._.`) =A0r...@kd... =A0(_Y_.)' ._ =A0 ) `._`. =A0" -.-' =A0 KDE Developer (MFCH) =A0 _..`-'_..-_/ /-'_.' (l)-'' ((i).' ((!.' =A0 Buenos Aires - Argentina Imminentizing the eschaton since 1971. |
From: Pawel F. <pj...@gn...> - 2004-09-23 14:14:35
|
On Tuesday 21 of September 2004 17:29, Pawel Foremski wrote: > Nobody can - it's a bug - thanks for reporting it :P. I'm just about to > release 0.33 version, but please test this first: Well, I couldn't wait more - qmail-spp has been just released :). -- Pawel Foremski pj...@gn... http://pjf.dotgeek.org/ |
From: Pawel F. <pj...@gn...> - 2004-09-21 15:29:55
|
On Tuesday 21 of September 2004 10:24, Derek Sims wrote: > if [ `env | grep ^RELAYCLIENT` ]; then Why don't you simply use if [ -n "$RELAYCLIENT" ]; then > export LOCAL='Y' Another note: you don't need to do this if you run > echo "SLOCAL=Y" just after :) > fi > I get the following in the LOGFILE: > 20040921080755 HELO+xxx.xxx.xxx.xxx, LOCAL set to Y by RELAYCLIENT > 20040921080755 FROM+xxx.xxx.xxx.xxx, LOCAL:Y > Z > > Can anyone explain why I get an apparent value in $LOCAL of "Y\nZ" ? Nobody can - it's a bug - thanks for reporting it :P. I'm just about to release 0.33 version, but please test this first: Before compilation add line data.s[data.len - 1] = 0; just after case 'S': /* set */ and just after case 'U': /* unset */ (just after these two lines - twice), finally you should have: case 'S': /* set */ data.s[data.len - 1] = 0; if (!env_put(data.s + 1)) die_nomem(); break; case 'U': /* unset */ data.s[data.len - 1] = 0; if (!env_unset(data.s + 1)) die_nomem(); there. Recompile qmail-smtpd and try it - you'll be the first running qmail-spp 0.33 :P. Thanks once more, cheers! -- Pawel Foremski pj...@gn... http://pjf.dotgeek.org/ |
From: Derek S. <de...@in...> - 2004-09-21 08:24:51
|
The filters are all running as bash scripts I have the following code in the "helo" filter: if [ `env | grep ^RELAYCLIENT` ]; then echo "$NOW HELO+$TCPREMOTEIP, LOCAL set to Y by RELAYCLIENT" >> $LOGFILE export LOCAL='Y' echo "SLOCAL=Y" fi Then in the "mail" filter I have: # set flag for internal mail - kludge because qmail-spp does not seem to set env vars properly if [ -n "$LOCAL" ]; then internal='Y' fi #internal=$LOCAL echo "$NOW FROM+$TCPREMOTEIP, LOCAL:$LOCAL" >>$LOGFILE I get the following in the LOGFILE: 20040921080755 HELO+xxx.xxx.xxx.xxx, LOCAL set to Y by RELAYCLIENT 20040921080755 FROM+xxx.xxx.xxx.xxx, LOCAL:Y Z Can anyone explain why I get an apparent value in $LOCAL of "Y\nZ" ? TIA, Derek |
From: Pawel F. <pj...@gn...> - 2004-07-08 12:07:21
|
Hi, this is message to test whether this mailing list works good. -- Pawel Foremski pj...@gn... http://pjf.dotgeek.org/ |
From: Pawel F. <pj...@gn...> - 2004-07-08 11:50:20
|
Hi, this is message to test whether this mailing list works good. -- Pawel Foremski pj...@gn... http://pjf.dotgeek.org/ |