The qpopper list archive ending on 24 Feb 2000
Topics covered in this issue include:
1. Re: What does this mean in log file???
"Justin M. Wilson" <justin at justin dot net>
Thu, 17 Feb 2000 14:38:49 -0500 (EST)
2. RE: What does this mean in log file???
"Karim Naguib" <knaguib at clickhere dot com>
Thu, 17 Feb 2000 13:42:58 -0600
3. RE: What does this mean in log file???
John Stevenson <JSteven1 at swst dot com>
Thu, 17 Feb 2000 13:48:10 -0600
4. Qpopper process keeps the lock!
"Luis Fernando Barrera" <luba at assist.com dot gt>
Thu, 17 Feb 2000 14:09:40 -0600
5. Re: Qpopper process keeps the lock!
Gerhard Gonter <gonter at wendy.wu-wien.ac dot at>
Thu, 17 Feb 2000 22:02:01 +0100 (MEZ)
6. Re: Qpopper process keeps the lock!
Markus Dobel
Thu, 17 Feb 2000 22:42:06 +0100
7. Re: Qpopper process keeps the lock!
Alan Brown <alan at manawatu.gen dot nz>
Fri, 18 Feb 2000 13:40:59 +1300 (NZDT)
8. Re: Qpopper process keeps the lock!
Randall Gellens <randy at pensive dot org>
Thu, 17 Feb 2000 19:12:06 -0800
9. Re: Qpopper process keeps the lock!
Randall Gellens <randy at pensive dot org>
Thu, 17 Feb 2000 19:16:40 -0800
10. Re: Qpopper process keeps the lock!
Catalin BOIE <catab at deuroconsult dot ro>
Fri, 18 Feb 2000 10:34:33 +0200 (EET)
11. a new addition to the code... for having POP only accounts.
"david's sysop user." <D.Bussenschutt at mailbox.gu.edu dot au>
Fri, 18 Feb 2000 19:41:54 +1000
12. Re: Qpopper process keeps the lock!
Gerhard Gonter <gonter at wendy.wu-wien.ac dot at>
Fri, 18 Feb 2000 10:24:21 +0100 (MEZ)
13. patches/modifications for qpopper 2.53?
Georg Schwarz <schwarz at physik.tu-berlin dot de>
Fri, 18 Feb 2000 14:49:08 +0100 (MET)
14.
"Peter Palombi" <pal at dnet dot it>
Fri, 18 Feb 2000 13:53:09 +0100
15. qpopper and solaris
Patricia Max <pam at lampinc dot com>
Fri, 18 Feb 2000 10:20:12 -0700 (MST)
16. Re: Qpopper process keeps the lock!
Randall Gellens <randy at pensive dot org>
Fri, 18 Feb 2000 11:08:10 -0800
17. Re: patches/modifications for qpopper 2.53?
Randall Gellens <randy at pensive dot org>
Fri, 18 Feb 2000 11:14:38 -0800
18.
Randall Gellens <randy at pensive dot org>
Fri, 18 Feb 2000 11:16:08 -0800
19. Re: qpopper and solaris
Randall Gellens <randy at pensive dot org>
Fri, 18 Feb 2000 11:23:40 -0800
20. Qpopper process keeps the lock! v2
"Luis Fernando Barrera" <luba at assist.com dot gt>
Fri, 18 Feb 2000 14:54:24 -0600
21. Re: Qpopper process keeps the lock! v2
Randall Gellens <randy at pensive dot org>
Fri, 18 Feb 2000 16:34:30 -0800
22. bulletins override compiled value
"Mike McKeage" <mckeage at ebtech dot net>
Sat, 19 Feb 2000 08:40:21 -0500
23. qpopper 2.53's popauth doesn't work?
Georg Schwarz <schwarz at physik.tu-berlin dot de>
Sat, 19 Feb 2000 16:13:59 +0100 (MET)
24. Re: bulletins override compiled value
Fergal Daly <fergal at esatclear dot ie>
Sat, 19 Feb 2000 15:23:47 +0000
25. how do I unsubscribe?
Andy <war1 at jps dot net>
Sat, 19 Feb 2000 10:34:25 -0800
26. Qpopper + MySQL
Paulo Santos <psantos at gol.com dot br>
Sat, 19 Feb 2000 17:31:26 -0300
27. Re: bulletins override compiled value
Randall Gellens <randy at pensive dot org>
Sat, 19 Feb 2000 13:38:46 -0800
28. Re: qpopper 2.53's popauth doesn't work?
Randall Gellens <randy at pensive dot org>
Sat, 19 Feb 2000 13:41:28 -0800
29. Re: how do I unsubscribe?
Randall Gellens <randy at pensive dot org>
Sat, 19 Feb 2000 13:42:49 -0800
30. Re: Qpopper + MySQL
Randall Gellens <randy at pensive dot org>
Sat, 19 Feb 2000 13:47:59 -0800
31. comments/patches for 3.0b34
schwarz at physik.tu-berlin dot de (Georg Schwarz)
Sun, 20 Feb 2000 11:08:57 +0100
32. popper/dbm_jacket.h not used?
Georg Schwarz <schwarz at physik.tu-berlin dot de>
Sun, 20 Feb 2000 22:50:42 +0100 (MET)
33. Virtual E-Mail Domain
"Peter Palombi" <pal at dnet dot it>
Mon, 21 Feb 2000 08:46:24 +0100
34. Re: Qpopper process keeps the lock!
Catalin BOIE <catab at deuroconsult dot ro>
Mon, 21 Feb 2000 09:53:19 +0200 (EET)
35. Re: popper/dbm_jacket.h not used?
Randall Gellens <randy at qualcomm dot com>
Mon, 21 Feb 2000 10:47:12 -0800
36. Re: Qpopper process keeps the lock!
Randall Gellens <randy at qualcomm dot com>
Mon, 21 Feb 2000 11:03:48 -0800
37. Appropriate list?
"Neil R. Smith" <neils at ariel.met.tamu dot edu>
Mon, 21 Feb 2000 18:17:15 -0600
38. No pop serving on an NIS client?
"Neil R. Smith" <neils at ariel.met.tamu dot edu>
Mon, 21 Feb 2000 19:03:03 -0600
39. Re: Appropriate list?
"Jeremy C. Reed" <reed at wcug.wwu dot edu>
Mon, 21 Feb 2000 17:50:55 -0800 (PST)
40. Re: Appropriate list?
Fergal Daly <fergal at esatclear dot ie>
Tue, 22 Feb 2000 12:33:58 +0000
41. qpopper 3.0b34 man page
Georg Schwarz <schwarz at physik.tu-berlin dot de>
Tue, 22 Feb 2000 14:08:43 +0100 (MET)
42. CHECK_SHELL?
Georg Schwarz <schwarz at physik.tu-berlin dot de>
Tue, 22 Feb 2000 15:12:41 +0100 (MET)
43. Re: Appropriate list?
"Nordh, Valter" <valter at che.chalmers dot se>
Tue, 22 Feb 2000 16:00:40 +0100
44. Re: No pop serving on an NIS client?
Randall Gellens <randy at qualcomm dot com>
Tue, 22 Feb 2000 09:34:32 -0800
45. Patch for Qualcomm popper 3.0b34
Georg Schwarz <schwarz at physik.tu-berlin dot de>
Wed, 23 Feb 2000 00:24:50 +0100 (MET)
46. Re: CHECK_SHELL?
David Bussenschutt <D.Bussenschutt at mailbox.gu.edu dot au>
Wed, 23 Feb 2000 10:03:16 +1000
47. Re: qpopper 2.53's popauth doesn't work?
"Kenneth Porter" <shiva at well dot com>
Tue, 22 Feb 2000 18:15:58 -0800
48. qpopper3.0b34
Mats Luspa <mats.luspa at irf dot se>
Thu, 24 Feb 2000 13:06:41 +0100 (MET)
49. RFP: Request for patch
Joaquin Vidal Balanza <joaquin.vidal at f-integra dot org>
Thu, 24 Feb 2000 13:46:36 +0100
50. Re: RFP: Request for patch
PSantos <psantos at gol.com dot br>
Thu, 24 Feb 2000 11:27:04 -0300
Date: Thu, 17 Feb 2000 14:38:49 -0500 (EST)
From: "Justin M. Wilson" <justin at justin dot net>
Subject: Re: What does this mean in log file???
http://www.eudora.com/qpopper/faq.html#stats
----
Justin M. Wilson
justin at justin dot net
justin at cavtel dot net
justin.wilson at vcu dot edu
On Thu, 17 Feb 2000, John Stevenson wrote:
> Hello. Can someone please inform me as to what each of the 0's mean in this
> log entry?
>
> Feb 17 00:00:49 mail popper[29402]: Stats: jsteven1 0 0 0 0
>
> Thank you,
> John Stevenson
>
From: "Karim Naguib" <knaguib at clickhere dot com>
Subject: RE: What does this mean in log file???
Date: Thu, 17 Feb 2000 13:42:58 -0600
Unsubscribe.
From: John Stevenson <JSteven1 at swst dot com>
Subject: RE: What does this mean in log file???
Date: Thu, 17 Feb 2000 13:48:10 -0600
Thank you!!!
John.
-----Original Message-----
From: Justin M. Wilson [mailto:justin at justin dot net]
Sent: Thursday, February 17, 2000 1:39 PM
To: John Stevenson
Cc: Subscribers of Qpopper
Subject: Re: What does this mean in log file???
http://www.eudora.com/qpopper/faq.html#stats
----
Justin M. Wilson
justin at justin dot net
justin at cavtel dot net
justin.wilson at vcu dot edu
On Thu, 17 Feb 2000, John Stevenson wrote:
> Hello. Can someone please inform me as to what each of the 0's mean in
this
> log entry?
>
> Feb 17 00:00:49 mail popper[29402]: Stats: jsteven1 0 0 0 0
>
> Thank you,
> John Stevenson
>
From: "Luis Fernando Barrera" <luba at assist.com dot gt>
Subject: Qpopper process keeps the lock!
Date: Thu, 17 Feb 2000 14:09:40 -0600
Hi,
I have a strange problem... When a user for a distant location (i.e. not =
from the LAN), tries
to read a big message (grater than 500KB), from the POP server, it just =
got stalled. Sometimes
the client reads the mail, but in the disconnect phase, also got =
stalled. The outlook client freezes!
I read the debug/trace file from the QPopper process, but it said =
nothing. Everything seems
right but the process would'nt die until 15 minutes o more; even I have =
a 5min timeout on the
client and 2min timeout on the Server. The .user.pop file remains on the =
spool directory...
If I read the same mailbox from a machine in the LAN, it works fine...
Any Ideas?
I tried with Qpopper 2.53 and Qpopper 3.0b34 with Linux Red Hat 6.0...
Thanks in advance!
Luis Fernando Barrera
luba at assist.com dot gt
From: Gerhard Gonter <gonter at wendy.wu-wien.ac dot at>
Subject: Re: Qpopper process keeps the lock!
Date: Thu, 17 Feb 2000 22:02:01 +0100 (MEZ)
According to Luis Fernando Barrera:
> I read the debug/trace file from the QPopper process, but it said nothing. Everything seems
> right but the process would'nt die until 15 minutes o more; even I have a 5min timeout on the
> client and 2min timeout on the Server. The .user.pop file remains on the spool directory...
This matches my observations. Outlook (some versions thereof?) seems
to give up the transfer after exactly 15 minutes, WHILE TRANSFERING
DATA, claiming, that transfer has stopped. I watched such sessions on
a simple protocol analysator. Of course, my observation can be flawed,
but I don't think so.
> If I read the same mailbox from a machine in the LAN, it works fine...
Of course, as long as the transfer of a single message takes less than
15 minutes.
> Any Ideas?
>
> I tried with Qpopper 2.53 and Qpopper 3.0b34 with Linux Red Hat 6.0...
I fear you can try whatever you want, the problem is IMO outlook.
+gg
--
Gerhard.Gonter at wu-wien.ac.at Fax: +43/1/31336/702 g dot gonter at ieee dot org
Zentrum fuer Informatikdienste, Wirtschaftsuniversitaet Wien, Austria
Date: Thu, 17 Feb 2000 22:42:06 +0100
From: Markus Dobel
Subject: Re: Qpopper process keeps the lock!
Gerhard Gonter wrote:
>
> According to Luis Fernando Barrera:
> > I read the debug/trace file from the QPopper process, but it said nothing. Everything seems
> > right but the process would'nt die until 15 minutes o more; even I have a 5min timeout on the
> > client and 2min timeout on the Server. The .user.pop file remains on the spool directory...
>
> This matches my observations. Outlook (some versions thereof?) seems
> to give up the transfer after exactly 15 minutes, WHILE TRANSFERING
> DATA, claiming, that transfer has stopped. I watched such sessions on
> a simple protocol analysator. Of course, my observation can be flawed,
> but I don't think so.
maybe the problem is somewhere in between. i had some problems like
these years ago with one of my dialup providers and netscape. like
outlook and many other clients, netscape has a feature to check for new
mail every x seconds. and it checks for mail exactly every x seconds, no
matter, if there's a mail download running or not.
the pop server of my provider dropped the first connection and accepted
the second one, the first download stalled.
without having tested it here and without knowing, which pop server my
provider had (it's >4 years ago), maybe qpopper does the same? if so, it
should be patched not to open any further sessions for same account
while another session is open.
regards, markus
Date: Fri, 18 Feb 2000 13:40:59 +1300 (NZDT)
From: Alan Brown <alan at manawatu.gen dot nz>
Subject: Re: Qpopper process keeps the lock!
On Thu, 17 Feb 2000, Gerhard Gonter wrote:
> This matches my observations. Outlook (some versions thereof?) seems
> to give up the transfer after exactly 15 minutes, WHILE TRANSFERING
> DATA, claiming, that transfer has stopped.
Have you checked the message collection timeouts settings?
AB
Date: Thu, 17 Feb 2000 19:12:06 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: Qpopper process keeps the lock!
At 10:42 PM +0100 2/17/00, Markus Dobel wrote:
> the pop server of my provider dropped the first connection and accepted
> the second one, the first download stalled.
>
> without having tested it here and without knowing, which pop server my
> provider had (it's >4 years ago), maybe qpopper does the same?
Qpopper rejects a second log-in by the same user, giving an [IN-USE] error.
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
Friends help you move. Real friends help you move bodies.
Date: Thu, 17 Feb 2000 19:16:40 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: Qpopper process keeps the lock!
At 2:09 PM -0600 2/17/00, Luis Fernando Barrera wrote:
> I read the debug/trace file from the QPopper process, but it said
> nothing. Everything seems
> right but the process would'nt die until 15 minutes o more; even I
> have a 5min timeout on the
> client and 2min timeout on the Server. The .user.pop file remains
> on the spool directory...
First, you probably want to increase the timeouts at both ends.
Second, are you saying that the client disconnects but the Qpopper
process stays active, with the .user.pop file around, for fifteen
minutes?
If you are not using server mode, and if the user has a very large
spool file, it may take several minutes after a session ends for the
Qpopper process to finish cleaning up and to then remove the
.user.pop file. See <http://www.eudora.com/qpopper/faq.html#cancel>.
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
If I knew what was going on, I'd be indignant!
Date: Fri, 18 Feb 2000 10:34:33 +0200 (EET)
From: Catalin BOIE <catab at deuroconsult dot ro>
Subject: Re: Qpopper process keeps the lock!
On Thu, 17 Feb 2000, Randall Gellens wrote:
> Qpopper rejects a second log-in by the same user, giving an [IN-USE] error.
Just a tought: why we not give the error when a user issue a command AFTER
he/she was authenticated?
Right now is:
On console1
telnet host pop3
USER x
+OK
PASS y
+OK
LIST
+OK
OK, first user authenticated.
On the console2
USER x
+OK
PASS y
-ERR [IN-USE] /usr/poptemp/.x.pop lock busy! Is another session active? (11)
So, I think we can let the user authenticate, but when he/she try to do
something (LIST, TOP, DELE etc.) to give this error!
I think will be better.
Is this thing catched in a RFC (locking, I mean)?
Thanks!
---
Catalin(ux) BOIE
catab at deuroconsult dot ro
http://www2.deuroconsult.ro/~catab
Fax: (561) 760-6212
ICQ: 45443500
Date: Fri, 18 Feb 2000 19:41:54 +1000
From: "david's sysop user." <D.Bussenschutt at mailbox.gu.edu dot au>
Subject: a new addition to the code... for having POP only accounts.
This is a multi-part message in MIME format.
--------------E7612CD6503CACC34B032542
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Hi Qpopper masters,
When a users account is disabled on our Solaris Systems they can still
POP mail (using qpopper lastest stable version) with standard install.
We disable the users accounts by changing the users shell to an invalid
shell (a valid shell is defined in the /etc/shells file). (man shells)
By disabling this way, we can provide users with a meaningful message
telling them why their account is disabled. eg change the shell to
"/DISABLED", and create a /DISABLED file:
---------------8<---cut-here----8<---------------------------------
#!/usr/ucb/tail +5
# file: /DISABLED
#
# Special Script, run as a users shell when their account is disabled
#
Your account has been disabled!
You are required to see Computing Support, room XXXX in XXXX
Building, as soon as physically possible to have it re-enabled.
---------------8<---cut-here----8<---------------------------------
We deliberately DO NOT disable the accounts by using the account expiry
parameters in the /etc/shadow password file because then the users (who
aren't necessarily intelligent) dont get a meaningful error (they just
get a 'access denied' or 'login denied').
qpopper does not recognise this as being a valid method of disabling
access.
I noticed the hook in the source code for the CHECK_SHELL option, and
this was almost what we needed but not quite.
In case you want to add this into the distribution attached are the
context diffs. (to version 2.53).
They are diffs to: popper.h pop_pass.c and INSTALL
let me know what you think.
David.
--------------E7612CD6503CACC34B032542
Content-Type: text/plain; charset=us-ascii;
name="diffs2"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="diffs2"
*** INSTALL.orig Fri Feb 18 18:17:50 2000
--- INSTALL Fri Feb 18 18:28:59 2000
***************
*** 493,507 ****
desired.
u) CHECK_SHELL - Enable this compile time feature
! to lock out users via the shell value. A user
! shell of /QPOPPER/ANY/SHELL allows the user
! access but blocks other programs that check
! shells.
!
! v) GDBM - This value uses the GNU's GDBM library
! instead of NDBM.
NOTES:
--- 493,519 ----
desired.
u) CHECK_SHELL - Enable this compile time feature
! to lock out users who have an invalid shell defined.
! (ie their shell in the /etc/passwd is NOT listed in
! the list of valid shells /etc/shells). This can be
! overwritten/deactivated by adding the wildcard
! /POPPER/ANY/SHELL/ to the list of valid shells
! (ie /etc/shells) exactly as written.
!
! v) MAILONLY_SHELL - Enable this compile time feature
! to allow users (on a user-by-user basis) to POP with
! an invalid shell(ie create POP-only accounts).
! Setting a users shell(in /etc/passwd) to
! /POPPER/MAILONLY allows a user to have mail access
! but blocks other programs that check shells (such as
! ftp and telnet) because it is an invalid shell.
! This option only has effect if CHECK_SHELL is also
! defined, otherwise it is meaningless to allow those
! already allowed.
+ w) GDBM - This value uses the GNU's GDBM library
+
NOTES:
*** pop_pass.c.orig Fri Feb 18 18:18:08 2000
--- pop_pass.c Fri Feb 18 18:20:50 2000
***************
*** 601,607 ****
/* You can default the shell, but I don't think it's a good idea */
/* shell = "/usr/bin/sh"; */
return(pop_msg(p, POP_FAILURE, "No user shell defined"));
!
for (shellvalid = 0; !shellvalid && ((cp = getusershell()) != NULL);)
if (!strcmp(cp, WILDCARD_SHELL) || !strcmp(cp, shell))
shellvalid = 1;
--- 601,611 ----
/* You can default the shell, but I don't think it's a good idea */
/* shell = "/usr/bin/sh"; */
return(pop_msg(p, POP_FAILURE, "No user shell defined"));
!
! #ifdef MAILONLY_SHELL
! if (!strcmp(shell,_MAILONLY_SHELL)) shellvalid = 2;
! else
! #endif
for (shellvalid = 0; !shellvalid && ((cp = getusershell()) != NULL);)
if (!strcmp(cp, WILDCARD_SHELL) || !strcmp(cp, shell))
shellvalid = 1;
***************
*** 608,614 ****
endusershell();
if (!shellvalid)
! return(pop_msg(p, POP_FAILURE, "\"%s\": shell not found.", p->user));
#endif
if ((p->kerberos ? auth_user_kerberos(p, pw) : auth_user(p, pwp))
--- 612,618 ----
endusershell();
if (!shellvalid)
! return(pop_msg(p, POP_FAILURE, "\"%s\": shell not found (\"%s\").",p->user,shell));
#endif
if ((p->kerberos ? auth_user_kerberos(p, pw) : auth_user(p, pwp))
*** popper.h.orig Fri Feb 18 18:18:21 2000
--- popper.h Fri Feb 18 18:21:33 2000
***************
*** 147,152 ****
--- 147,156 ----
#define WILDCARD_SHELL "/POPPER/ANY/SHELL/"
#endif
+ #ifdef MAILONLY_SHELL
+ #define _MAILONLY_SHELL "/POPPER/MAILONLY"
+ #endif
+
extern int errno;
#if !(defined(BSD) && (BSD >= 199306)) && !defined(__USE_BSD)
--------------E7612CD6503CACC34B032542--
From: Gerhard Gonter <gonter at wendy.wu-wien.ac dot at>
Subject: Re: Qpopper process keeps the lock!
Date: Fri, 18 Feb 2000 10:24:21 +0100 (MEZ)
According to Alan Brown:
> > This matches my observations. Outlook (some versions thereof?) seems
> > to give up the transfer after exactly 15 minutes, WHILE TRANSFERING
> > DATA, claiming, that transfer has stopped.
>
> Have you checked the message collection timeouts settings?
How do you ask a user to check what s/he has condifugred in outlook?
The best answer they can usually provide a series of BMP's embedded in
a Word file...
Outlook problems are usually solved very easily by installing Eudora
or Netscape. That's the prefered solution here ;)
+gg
--
Gerhard.Gonter at wu-wien.ac.at Fax: +43/1/31336/702 g dot gonter at ieee dot org
Zentrum fuer Informatikdienste, Wirtschaftsuniversitaet Wien, Austria
Subject: patches/modifications for qpopper 2.53?
Date: Fri, 18 Feb 2000 14:49:08 +0100 (MET)
From: Georg Schwarz <schwarz at physik.tu-berlin dot de>
Are there already any patches for Qualcomm's popper 2.53 pop3 daemon which
could address some of the following issues?
- use a private apop database for each user (such as ~/.poppassword)?
- run the password checking and everything after that under the user's
UID?
- allow apop from everywhere, but cleartext authentification only from
127.0.0.0,255.0.0.0 (the latter to enable ssl or ssh tunnelling)?
- use the user's home directory for any temporary files?
Any hints are welcome. Thanks.
--
Georg Schwarz schwarz at physik.tu-berlin dot de
Institut für Theoretische Physik Tel. +49 30 314-24254, FAX -21130
Technische Universität Berlin http://home.pages.de/~schwarz/
Sekr. PN 7-1, Hardenbergstr. 36, D-10623 Berlin IRC: kuroi
From: "Peter Palombi" <pal at dnet dot it>
Date: Fri, 18 Feb 2000 13:53:09 +0100
Hi all,
I would use qpopper and sendmail8.9.3
but i have now a problem
I must use two domain and i must insert a new account(for exemple peter with
with e-mail address peter@domain1
) and other account (for exemple jim with e-mail address jim@domain2)
how can i make this
peter
thanks
pal at dnet dot it
Date: Fri, 18 Feb 2000 10:20:12 -0700 (MST)
From: Patricia Max <pam at lampinc dot com>
Subject: qpopper and solaris
Are there any reported problems with qpopper3.0b30, solaris2.7 at
patch level 106541-08, sendmail-8.9.3, Eudora, NT 4, Outlook
Express, etc.? I have had several problems reported to me in the last 6
weeks from people who access email from both unix and non-unix machines!
1. I have a user who reads his email using Openwindows mailtool
from solaris 2.6 via nfs mount from the mail server running 2.7.
When he also reads his email from his nt machine with Outlook
Express, he winds up with duplicate copies of his email in his system
mailbox on the 2.7 server. This did not happen before I upgraded the
mail server to 2.7, the popper to 3.0b30, and his machine to 2.6, all in
about 2 days right after Christmas!
2. I have another user who claims that when he uses Eudora to read
email from home, the only eudora option that works to ensure that
he only downloads new messages is Status: headers. It ignores
POP3 LAST command and First message not read by this machine. If
he selects either of the last 2, then all mail is downloaded each time.
3. I have a third user who claims that the download of his email
to his nt machine is much slower now then it was before I upgraded the
server and popper. I was not able to duplicate his problem using
netscape, but he also uses Outlook Express.
I have popper running with a -s option, but I can't find any information
in my current syslogs about what popper is doing. There does not seem to
be any information in my syslog file, but the man page for popper says
that at the end of each session information will be logged.
Does anyone have any suggestions?
Thanks in advance.
Patricia Max
LAMP Inc
Date: Fri, 18 Feb 2000 11:08:10 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: Qpopper process keeps the lock!
At 10:34 AM +0200 2/18/00, Catalin BOIE wrote:
> On Thu, 17 Feb 2000, Randall Gellens wrote:
>
>> Qpopper rejects a second log-in by the same user, giving an [IN-USE] error.
> Just a tought: why we not give the error when a user issue a command AFTER
> he/she was authenticated?
>
> Right now is:
>
> On console1
> telnet host pop3
> USER x
> +OK
> PASS y
> +OK
> LIST
> +OK
>
> OK, first user authenticated.
> On the console2
> USER x
> +OK
> PASS y
> -ERR [IN-USE] /usr/poptemp/.x.pop lock busy! Is another session active? (11)
>
> So, I think we can let the user authenticate, but when he/she try to do
> something (LIST, TOP, DELE etc.) to give this error!
> I think will be better.
> Is this thing catched in a RFC (locking, I mean)?
Why would it be better to give the error later?
At any rate, the current behavior is specified in RFC 1939.
Note that it is possible to have a server which does not use
exclusive locking, but instead permits multiple simultaneous POP
sessions, giving each one a view of the mailbox as if it was fixed at
the time of authentication. Messages marked deleted in one session
remain visible to other sessions previously active; when all sessions
which could see the deleted message exit the message is physically
removed.
However, such implementations usually are "black box" servers which
do not permit any access to the internals, which is quite the
opposite of Qpopper and standard Unix mail software.
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
The challenge of computer science is: How NOT to make a mess of it.
-Edsger W. Dijkstra
Date: Fri, 18 Feb 2000 11:14:38 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: patches/modifications for qpopper 2.53?
At 2:49 PM +0100 2/18/00, Georg Schwarz wrote:
> - use the user's home directory for any temporary files?
In Qpopper 3.0 you can change the directories used for temporary
files. See the INSTALL file, or just run ./configure --help.
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
There is no human problem which could not be solved if people would
simply do as I advise. -- Gore Vidal
Date: Fri, 18 Feb 2000 11:16:08 -0800
From: Randall Gellens <randy at pensive dot org>
At 1:53 PM +0100 2/18/00, Peter Palombi wrote:
> Hi all,
>
> I would use qpopper and sendmail8.9.3
> but i have now a problem
> I must use two domain and i must insert a new account(for exemple peter with
> with e-mail address peter@domain1
> ) and other account (for exemple jim with e-mail address jim@domain2)
>
> how can i make this
>
> peter
>
> thanks
> pal at dnet dot it
See <http://www.eudora.com/qpopper/faq.html#virt.dom--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
Do not drink coffee in early A.M. It will keep you awake until noon.
Date: Fri, 18 Feb 2000 11:23:40 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: qpopper and solaris
At 10:20 AM -0700 2/18/00, Patricia Max wrote:
> Are there any reported problems with qpopper3.0b30, solaris2.7 at
> patch level 106541-08, sendmail-8.9.3, Eudora, NT 4, Outlook
> Express, etc.?
I'd suggest upgrading to Qpopper 3.0b34.
> I have had several problems reported to me in the last 6
> weeks from people who access email from both unix and non-unix machines!
>
> 1. I have a user who reads his email using Openwindows mailtool
> from solaris 2.6 via nfs mount from the mail server running 2.7.
> When he also reads his email from his nt machine with Outlook
> Express, he winds up with duplicate copies of his email in his system
> mailbox on the 2.7 server. This did not happen before I upgraded the
> mail server to 2.7, the popper to 3.0b30, and his machine to 2.6, all in
> about 2 days right after Christmas!
You could enable debug tracing to see what is going on. You could do
this on a different port, just for this user, if you want. Do a
'make clean', then re-run ./configure, adding '--enable-debugging',
and add '-t tracefile' to the inetd.conf line. The trace info will
be written to the file specified as 'tracefile'.
> 2. I have another user who claims that when he uses Eudora to read
> email from home, the only eudora option that works to ensure that
> he only downloads new messages is Status: headers. It ignores
> POP3 LAST command and First message not read by this machine. If
> he selects either of the last 2, then all mail is downloaded each time.
There was a bug in 3.0b30 that could cause this if server_mode is not
used. Upgrade to b34.
> 3. I have a third user who claims that the download of his email
> to his nt machine is much slower now then it was before I upgraded the
> server and popper. I was not able to duplicate his problem using
> netscape, but he also uses Outlook Express.
>
> I have popper running with a -s option, but I can't find any information
> in my current syslogs about what popper is doing. There does not seem to
> be any information in my syslog file, but the man page for popper says
> that at the end of each session information will be logged.
You can use the --with-log-facility=name parameter to ./configure if
you like. Also, see the information logging in the INSTALL file.
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
An algorithm must be seen to be believed. -- Donald Knuth
From: "Luis Fernando Barrera" <luba at assist.com dot gt>
Subject: Qpopper process keeps the lock! v2
Date: Fri, 18 Feb 2000 14:54:24 -0600
Hi,
>
> > I read the debug/trace file from the QPopper process, but it said
> > nothing. Everything seems
> > right but the process would'nt die until 15 minutes o more; even I
> > have a 5min timeout on the
> > client and 2min timeout on the Server. The .user.pop file remains
> > on the spool directory...
>
> First, you probably want to increase the timeouts at both ends.
One of the problems is that Outlook only accepts at most 5min timeout...
>
> Second, are you saying that the client disconnects but the Qpopper
> process stays active, with the .user.pop file around, for fifteen
> minutes?
That's right. I noticed that the outlook client gets a timeout and tries
for a second time, but off course the client gets the "POP Busy, Account =
file locked..." error.
It seems that the Qpopper did not pay attention to the timeout setting I =
gave it...
>
> If you are not using server mode, and if the user has a very large
> spool file, it may take several minutes after a session ends for the
> Qpopper process to finish cleaning up and to then remove the
> .user.pop file. See <http://www.eudora.com/qpopper/faq.html#cancel>.
I'm using server mode... I guess... How can you tell the diference? I =
saw that
even in the server mode, Qpopper creates a .user.pop file, with size =
0...
Luis Fernando Barrera
luba at assist.com dot gt
Date: Fri, 18 Feb 2000 16:34:30 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: Qpopper process keeps the lock! v2
At 2:54 PM -0600 2/18/00, Luis Fernando Barrera wrote:
> One of the problems is that Outlook only accepts at most 5min timeout...
As someone else mentioned, most Outlook problems can be solved by
upgrading to Eudora, which is now available totally free. See
<http://www.eudora.com>.
> That's right. I noticed that the outlook client gets a timeout and tries
> for a second time, but off course the client gets the "POP Busy,
> Account file locked..." error.
> It seems that the Qpopper did not pay attention to the timeout
> setting I gave it...
What does a trace of the session show? If either the client or
Qpopper times out, the session should be terminated. In normal (not
server) mode, or if some but not all messages were deleted, and if
the spool is quite large, it may take a few minutes for Qpopper to
finish recreating the spool. In server mode where all messages are
deleted, or none are deleted, or if NO_UPDATE_ON_ABORT is set, this
should be very quick.
>
> > If you are not using server mode, and if the user has a very large
>> spool file, it may take several minutes after a session ends for the
>> Qpopper process to finish cleaning up and to then remove the
>> .user.pop file. See <http://www.eudora.com/qpopper/faq.html#cancel>.
>
> I'm using server mode... I guess... How can you tell the diference?
> I saw that
> even in the server mode, Qpopper creates a .user.pop file, with size 0...
Yes, even in server mode there will be a .user.pop file, but if no
messages are deleted (or if all messages are deleted) Qpopper does
not copy the spool to it and then back, which saves a lot of time.
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
Language is a virus from outer space. --William S. Burroughs
From: "Mike McKeage" <mckeage at ebtech dot net>
Date: Sat, 19 Feb 2000 08:40:21 -0500
Subject: bulletins override compiled value
Hello
The qpopper FAQ says this about invoking bulletins:
1.By adding --enable-bulletins=path when running
./configure,
I did this, and it works, except most of my users do not have a
home directory.
2.Using the command line option -b bulldir. This overrides the
compiled value if it is defined.
Does this command line option override #1 ? If I add this command
option, and create the 2 empty database files, will it quit using the
home dir method, and use the database method?
Thank You
P.S. If anyone has handy a simple bash script to read a passwd
file, and create home directories, I would be eternally grateful.
The passwd file (using shadow as well) does specify home dir's,
just that most of them do not really exist.
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
I am Billgatus of Borg.
You will be assimiILLEGAL OPERATION
Subject: qpopper 2.53's popauth doesn't work?
Date: Sat, 19 Feb 2000 16:13:59 +0100 (MET)
From: Georg Schwarz <schwarz at physik.tu-berlin dot de>
popauth from qpopper 2.53 does not seem to work. Trying to add a
user's password I always get:
[root@zeus qpopper2 dot 53]# ./popauth -user schwarz
Changing POP password for schwarz.
New password:
Retype new password:
./popauth: /tmp/popauth: unable to open POP authorization DB
With the latest beta version it works. Is this a known bug? If so, what cau=
ses
it, and how could it be fixed?
BTW, this is RH Linux 5.2 on Intel.
--
Georg Schwarz schwarz at physik.tu-berlin dot de
Institut für Theoretische Physik Tel. +49 30 314-24254, FAX -21130
Technische Universität Berlin http://home.pages.de/~schwarz/
Sekr. PN 7-1, Hardenbergstr. 36, D-10623 Berlin IRC: kuroi
Date: Sat, 19 Feb 2000 15:23:47 +0000
From: Fergal Daly <fergal at esatclear dot ie>
Subject: Re: bulletins override compiled value
At 08:40 19/02/00 -0500, Mike McKeage wrote:
>Hello
>
>The qpopper FAQ says this about invoking bulletins:
>
> 1.By adding --enable-bulletins=path when running
>./configure,
>
>I did this, and it works, except most of my users do not have a
>home directory.
>
> 2.Using the command line option -b bulldir. This overrides the
>compiled value if it is defined.
The first option specifies where the bulletins are stores, the second
specifies where the records of who has seen the bulletins are stored, they
are 2 completely different things.
>Does this command line option override #1 ? If I add this command
>option, and create the 2 empty database files, will it quit using the
>home dir method, and use the database method?
>
>Thank You
>
>P.S. If anyone has handy a simple bash script to read a passwd
>file, and create home directories, I would be eternally grateful.
>The passwd file (using shadow as well) does specify home dir's,
>just that most of them do not really exist.
haven't tried running it but this should do the trick
#! /usr/bin/perl
foreach my $user (`cat /etc/passwd`)
{
chomp($user);
my ($uid, $gid, $homedir) = (split(":", $user))[2, 3, 6];
if (not -e $homedir)
{
mkdir($homedir, 0700);
chown($uid, $gid, $homedir);
}
}
Date: Sat, 19 Feb 2000 10:34:25 -0800
From: Andy <war1 at jps dot net>
Subject: how do I unsubscribe?
How do I unsubscribe?
Date: Sat, 19 Feb 2000 17:31:26 -0300
From: Paulo Santos <psantos at gol.com dot br>
Subject: Qpopper + MySQL
Olá All,
I'm running Qpopper + Mysql.
All run OK.. all users are getting your mails into a Mysql DB.
The Authentication is from MySQL too (there are not users into PASSWD).
./configure --enable-specialauth --enable-bulletins=/var/spool/mail/bulletins
--enable-shy --enable-log-login --enable-servermode --enable-bulldb=/var/spool/mail/bulletins
but, the users are in a table, then (I think) the qpopper.log report:
.....
Sat Feb 19 17:22:33 2000 [7406] Unable to create .popbull file (2)
.....
ps: I runned without mysql (another machine and with passwd) and run OK.
How can I clean this error?
Thanks!
Saudacoes,
Paulo Santos
mailto:psantos at gol.com dot br
Date: Sat, 19 Feb 2000 13:38:46 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: bulletins override compiled value
At 8:40 AM -0500 2/19/00, Mike McKeage wrote:
> Hello
>
> The qpopper FAQ says this about invoking bulletins:
>
> 1.By adding --enable-bulletins=path when running
> ./configure,
>
> I did this, and it works, except most of my users do not have a
> home directory.
Then you may want to use a bulletins database. Use the
--enable-bulldb=path ./configure option instead of --enable-bulletins.
>
> 2.Using the command line option -b bulldir. This overrides the
> compiled value if it is defined.
>
> Does this command line option override #1 ? If I add this command
> option, and create the 2 empty database files, will it quit using the
> home dir method, and use the database method?
The command line option overrides the path. It does not turn on the
bulldb option. To do that, use the new ./configure option as above.
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
Violence is the last refuge of the incompetent.
Salvador Hardin
Date: Sat, 19 Feb 2000 13:41:28 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: qpopper 2.53's popauth doesn't work?
At 4:13 PM +0100 2/19/00, Georg Schwarz wrote:
> popauth from qpopper 2.53 does not seem to work. Trying to add a
> user's password I always get:
>
> [root@zeus qpopper2 dot 53]# ./popauth -user schwarz
> Changing POP password for schwarz.
> New password:
> Retype new password:
> ./popauth: /tmp/popauth: unable to open POP authorization DB
>
> With the latest beta version it works. Is this a known bug? If so,
> what causes
> it, and how could it be fixed?
> BTW, this is RH Linux 5.2 on Intel.
Yes, it is a known bug in 2.53, and as you noted, it is fixed in 3.0.
See the doc/Changes file to find out which specific beta, if you are
interested. It can be fixed by using 3.0. which is much more stable,
especially on Linux, than 2.53.
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
To define recursion, we must first define recursion.
Date: Sat, 19 Feb 2000 13:42:49 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: how do I unsubscribe?
At 10:34 AM -0800 2/19/00, Andy wrote:
> How do I unsubscribe?
Every message sent by the list has this header:
List-Unsubscribe:
<mailto:qpopper-request at lists.pensive dot org?body=unsubscribe>
Simply click the link.
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
A successful tool is used to do something undreamed of by its author.
--Johnson
Date: Sat, 19 Feb 2000 13:47:59 -0800
From: Randall Gellens <randy at pensive dot org>
Subject: Re: Qpopper + MySQL
At 5:31 PM -0300 2/19/00, Paulo Santos wrote:
> Olá All,
>
> I'm running Qpopper + Mysql.
> All run OK.. all users are getting your mails into a Mysql DB.
> The Authentication is from MySQL too (there are not users into PASSWD).
>
> ./configure --enable-specialauth --enable-bulletins=/var/spool/mail/bul=
letins
> --enable-shy --enable-log-login --enable-servermode
> --enable-bulldb=/var/spool/mail/bulletins
>
> but, the users are in a table, then (I think) the qpopper.log report:
>
> .....
> Sat Feb 19 17:22:33 2000 [7406] Unable to create .popbull file (2)
> .....
>
> ps: I runned without mysql (another machine and with passwd) and run OK.
>
> How can I clean this error?
> Thanks!
>
As you say, the error is fixed in current 3.0 betas. But the MySQL
patch hasn't been updated. The MySQL patch is a third-party patch,
meaning it is not supported by Qualcomm. So you much choose:
1. Stop using MySQL and upgrade to current versions of Qpopper.
2. Fit the MySQL patch to current version yourself.
3. Try to back-fit the bug fix to the old version of Qpopper (very dangerou=
s!)
4. Wait until someone else updates the MySQL patch.
>
--
--
Randall Gellens Randy at Pensive dot Org
---------------------- (randomly-selected tag) ---------------------
Certe, Toto, sentio nos in Kansate non iam adesse. --Dorothea
Subject: comments/patches for 3.0b34
From: schwarz at physik.tu-berlin dot de (Georg Schwarz)
Date: Sun, 20 Feb 2000 11:08:57 +0100
Here are a few issues re. qpopper 3.0b34:
* configure.in says:
dnl 03/04/00 [rg]
dnl - Restore check for gperf, but warn (not error) if not found.
apparently the programmers did get some input from the future :-)
Should be 02/04/00, shouldn't it?
* popper/pop_msg.c includes:
pop_log ( p, POP_NOTICE, fn, ln,
( isdigit ( p->client[0] ) ? "%s@[%s]: %s" : "%s@%s: %s" =
),
( p->user ? p->user : "(null)" ), p->client, message);
it appears that the programmers wanted to print IP numbers in [] (whether t=
his
makes sense is another question; it would make sense to me if we were deali=
ng
with email addresses; however, here there are no email addresses involved a=
t
all).
Now thiscodes checks whether the first character of p->client is a digit. T=
hat
way, however, valid FQDN that start on a digit (such as 4com.com, for examp=
le)
are mistaken as IPs.
If one wants to print IPs in [] I suggest replacing
isdigit( p->client[0] )
by
( strcmp( p->client, p->ipaddr ) == 0)
or something like that.
the required string.h is already included anyway in pop_msg.c. Maybe, with
isdigit() no longer used one could get rid of some include file here?
The following patch introduces two new flags (one should change the config
facilities and add configure options for them as well):
NOCLEARTEXT: does not allow clear text passwords for anybody
ONLYLOCALCLEARTEXT: clear text passwords are only allowed if the clients
connects via the local loopback interface 127.0.0.0/8 (to enable, for examp=
le
ssh or ssl tunnekling of pop3).
Maybe one might come up with better names for those flags. Also, in the lat=
ter
case one might argue to allow cleartext logins via loopback even if an apop
entry exists. Right now, in the unpatched 3.0b34, it is possible to confirm
guessed user names from outside via the pop3d of the guessed user happens =
to
have an entry in the apop database.
zeus 55% /tmp/qpopper3.0b34-p>diff -u ../qpopper3.0b34/popper/pop_user.c
popper/pop_user.c
--- ../qpopper3.0b34/popper/pop_user.c Fri Feb 11 01:21:22 2000
+++ popper/pop_user.c Sat Feb 19 19:13:00 2000
@@ -153,6 +153,7 @@
* If this call fails then the database is not accessable (doesn't
* exist?) in which case we can ignore an APOP user trying to
* access Qpopper with a cleartext password.
+ * (unless NOCLEARTEXT is specified)
*/
if (((pw = getpwnam(p->user)) != NULL) &&
@@ -240,7 +241,25 @@
#endif /* AUTHON */
if ( p->AuthType == noauth ) /* If authentication method is unkno=
wn (i.e.
*/
- p->AuthType = plain; /* not Kerberos) then assume plain te=
xt
*/
+ { /* not Kerberos) then assume plain text =
*/
+
+#ifdef ONLYLOCALCLEARTEXT /* clean text passwords only allowed whe=
n */
+ /* connecting from loopback interface */
+ if(strncmp("127.",p->ipaddr,4) != 0) /* not local */
+#endif /* ONLYLOCALCLEARTEXT */
+#if defined(NOCLEARTEXT) || defined (ONLYLOCALCLEARTEXT)
+ return ( pop_auth_fail ( p, POP_FAILURE, HERE,
+ "[AUTH] You must use stronger
authentication "
+ "such as AUTH SCRAM-MD5 or APOP to co=
nnect
"
+"to this server" ) );
+#endif /* NOCLEARTEXT || ONLYLOCALCLEARTEXT */
+#ifdef ONLYLOCALCLEARTEXT
+ else /* client access from loopback interface 127.X.X.X */
+#endif
+#ifndef NOCLEARTEXT
+ p->AuthType = plain;
+#endif
+ }
p->AuthState = user; /* User command "succesful"
*/
/* Tell the user that the password is required */
if I should send that diff separately please let me know.
Two more issues I'd love to see implemented:
- temporary files being created in the user's home directory (even if mail =
is
not spooled to the home directory), resolving the home directory from the p=
asswd
entry and taking into account that the home directory might be NFS mounted =
(i.e.
requiring a change of the effective uid).
- option to have a separate apop db for each user in his home directory (so=
me
potential NFS issues as above) owned by him (would also require a few chang=
es in
popauth).
--
Georg Schwarz schwarz at physik.tu-berlin.de, kuroi at cs dot tu-berlin dot de
Institut für Theoretische Physik +49 30 314-24254, FAX -21130
Technische Universität Berlin http://home.pages.de/~schwarz/
Subject: popper/dbm_jacket.h not used?
Date: Sun, 20 Feb 2000 22:50:42 +0100 (MET)
From: Georg Schwarz <schwarz at physik.tu-berlin dot de>
Is it correct that popper/dbm_jacket.h in 3.0b34 is not used at all and
should be removed?
--
Georg Schwarz schwarz at physik.tu-berlin dot de
Institut für Theoretische Physik Tel. +49 30 314-24254, FAX -21130
Technische Universität Berlin http://home.pages.de/~schwarz/
Sekr. PN 7-1, Hardenbergstr. 36, D-10623 Berlin IRC: kuroi
From: "Peter Palombi" <pal at dnet dot it>
Subject: Virtual E-Mail Domain
Date: Mon, 21 Feb 2000 08:46:24 +0100
Hi,
Qpopper can support the virtual E-Mail Domain and how?
I can save the mail in two cdirectory(sendmail) /var/spool/mail
and /var/spool/vmail but how can i catch the mail in all two directory
Thanks
Peter
Date: Mon, 21 Feb 2000 09:53:19 +0200 (EET)
From: Catalin BOIE <catab at deuroconsult dot ro>
Subject: Re: Qpopper process keeps the lock!
On Fri, 18 Feb 2000, Randall Gellens wrote:
> Date: Fri, 18 Feb 2000 11:08:10 -0800
> From: Randall Gellens <randy at pensive dot org>
> To: Catalin BOIE <catab at deuroconsult dot ro>,
> Subscribers of Qpopper <qpopper at lists.pensive dot org>
> Subject: Re: Qpopper process keeps the lock!
>
> At 10:34 AM +0200 2/18/00, Catalin BOIE wrote:
>
> > On Thu, 17 Feb 2000, Randall Gellens wrote:
> >
> >> Qpopper rejects a second log-in by the same user, giving an [IN-USE] error.
> > Just a tought: why we not give the error when a user issue a command AFTER
> > he/she was authenticated?
> >
> > Right now is:
> >
> > On console1
> > telnet host pop3
> > USER x
> > +OK
> > PASS y
> > +OK
> > LIST
> > +OK
> >
> > OK, first user authenticated.
> > On the console2
> > USER x
> > +OK
> > PASS y
> > -ERR [IN-USE] /usr/poptemp/.x.pop lock busy! Is another session active? (11)
> >
> > So, I think we can let the user authenticate, but when he/she try to do
> > something (LIST, TOP, DELE etc.) to give this error!
> > I think will be better.
> > Is this thing catched in a RFC (locking, I mean)?
>
> Why would it be better to give the error later?
Because Outlook (probably Netscape too) don't know if were a server error
(INUSE) or bad password.
>
> At any rate, the current behavior is specified in RFC 1939.
Because POP3 doesn't have error codes (like sendmail or ftp etc.) a mail
agent can't tell if was an error or only a bad password.
>
> Note that it is possible to have a server which does not use
> exclusive locking, but instead permits multiple simultaneous POP
> sessions, giving each one a view of the mailbox as if it was fixed at
> the time of authentication. Messages marked deleted in one session
> remain visible to other sessions previously active; when all sessions
> which could see the deleted message exit the message is physically
> removed.
Yes. This is what I want: multiple logins, BUT when the secont login (or
third etc.) trys to delete something (or any command) to give this error.
I think will be better.
Can you tell me on what page the PASS command must return -ERR when the
mail file is locked, please?
Thanks!
---
Catalin(ux) BOIE
catab at deuroconsult dot ro
http://www2.deuroconsult.ro/~catab
Fax: (561) 760-6212
ICQ: 45443500
Date: Mon, 21 Feb 2000 10:47:12 -0800
From: Randall Gellens <randy at qualcomm dot com>
Subject: Re: popper/dbm_jacket.h not used?
At 10:50 PM +0100 2/20/00, Georg Schwarz wrote:
>Is it correct that popper/dbm_jacket.h in 3.0b34 is not used at all and
>should be removed?
It is not yet used, this is true.
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly-selected tag: ---------------
Justify my text? I'm sorry but it has no excuse.
Date: Mon, 21 Feb 2000 11:03:48 -0800
From: Randall Gellens <randy at qualcomm dot com>
Subject: Re: Qpopper process keeps the lock!
At 9:53 AM +0200 2/21/00, Catalin BOIE wrote:
> > >> Qpopper rejects a second log-in by the same user, giving an
>[IN-USE] error.
>> > Just a tought: why we not give the error when a user issue a
>>command AFTER
> > > he/she was authenticated?
> > Why would it be better to give the error later?
>
>Because Outlook (probably Netscape too) don't know if were a server error
>(INUSE) or bad password.
>
>>
>> At any rate, the current behavior is specified in RFC 1939.
>
>Because POP3 doesn't have error codes (like sendmail or ftp etc.) a mail
>agent can't tell if was an error or only a bad password.
The "[IN-USE]" is an extended error response code, as specified in
RFC 2449 (see <ftp://ftp.isi.edu/in-notes/rfc2449.txt>). The whole
point is that the "[IN-USE]" tells the client that this error means
the authentication was OK, but another session is in progress.
> Note that it is possible to have a server which does not use
> > exclusive locking, but instead permits multiple simultaneous POP
>> sessions, giving each one a view of the mailbox as if it was fixed at
>> the time of authentication. Messages marked deleted in one session
>> remain visible to other sessions previously active; when all sessions
>> which could see the deleted message exit the message is physically
>> removed.
>
>Yes. This is what I want: multiple logins, BUT when the secont login (or
>third etc.) trys to delete something (or any command) to give this error.
>I think will be better.
No, that would be much worse. RFC 1939 states that when a POP client
authenticates, an exclusive lock is obtained on the maildrop. If
this is unsuccessful, and error is returned. What I said above is
that some POP servers are structured so that there is no single
exclusive lock, instead, multiple simultaneous POP sessions can be in
progress, and each one has the illusion that it has an exclusive
lock. The state of the maildrop at the time of authentication is
preserved for that session. For example, if new messages arrive,
they are invisible to sessions already in progress, but can be seen
by subsequent sessions. Any client in any session can delete a
message; the message remains visible to other simultaneous sessions,
and is physically removed when all sessions when can see it have
terminated.
This is very different from delaying the error. If you give a
delayed error, you confuse the clients. They have no idea why the
error happened. In the non-exclusive model, the clients work just
fine. The error is not delayed; it is avoided.
As I said, this non-exclisive model is generally used with black-box
servers, very different from the open model used with Qpopper.
>Can you tell me on what page the PASS command must return -ERR when the
>mail file is locked, please?
See the bottom of page 4 of RFC 1939:
>Once the POP3 server has determined through the use of any
>authentication command that the client should be given access to the
>appropriate maildrop, the POP3 server then acquires an exclusive-
>access lock on the maildrop, as necessary to prevent messages from
>being modified or removed before the session enters the UPDATE
>state. If the lock is successfully acquired, the POP3 server
>responds with a positive status indicator. The POP3 session now
>enters the TRANSACTION state, with no messages marked as deleted.
>If the maildrop cannot be opened for some reason (for example, a
>lock can not be acquired, the client is denied access to the
>appropriate maildrop, or the maildrop cannot be parsed), the POP3
>server responds with a negative status indicator.
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly-selected tag: ---------------
He had occasional flashes of silence that made his conversation
perfectly delightful. --Sydney Smith
Date: Mon, 21 Feb 2000 18:17:15 -0600
From: "Neil R. Smith" <neils at ariel.met.tamu dot edu>
Subject: Appropriate list?
I've sent questions about making popper 2.53 on an SGI to this
list twice, but have received no response. Does this mean that
no one has anything to offer, or that this is not the
appropriate list for the questions? If I am asking in the wrong
place, please accept my appologies and advise as to a more
appropriate list.
Thanks, -Neil
--
Neil R. Smith, Comp. Sys. Mngr. neils at ariel.met.tamu dot edu
Dept. Atmospheric Sci., Texas A&M Univ. 979/845-6272 FAX:979/862-4466
Date: Mon, 21 Feb 2000 19:03:03 -0600
From: "Neil R. Smith" <neils at ariel.met.tamu dot edu>
Subject: No pop serving on an NIS client?
Well, I think I've just deduced part of my "password incorrect"
errors - the server that I wish to run popper on is an NIS client,
and the /etc/passwd has no users in it; just the std. + entry
as the last line. When I create a valid user entry in the
/etc/passwd, with valid password, as an entry prior to the + entry,
then my built popper works for that user.
I must have missed this point in the FAQ or INSTALL doc. Is this
true for popper? It can't be run on an NIS client?
(Note: the /etc/passwd file is not shadowed)
Thanks, -Neil
--
Neil R. Smith, Comp. Sys. Mngr. neils at ariel.met.tamu dot edu
Dept. Atmospheric Sci., Texas A&M Univ. 979/845-6272 FAX:979/862-4466
Date: Mon, 21 Feb 2000 17:50:55 -0800 (PST)
From: "Jeremy C. Reed" <reed at wcug.wwu dot edu>
Subject: Re: Appropriate list?
Neil,
On Mon, 21 Feb 2000, Neil R. Smith wrote:
> I've sent questions about making popper 2.53 on an SGI to this
> list twice, but have received no response. Does this mean that
You may want to try the comp.sys.sgi.admin newsgroup because I saw some of
the same "Giving up after printing 50 warnings" questions there at
http://www.deja.com. You may also want to try with the Qpopper 3.0 Beta.
Jeremy C. Reed
http://www.reedmedia.net
http://bsd.reedmedia.net
Date: Tue, 22 Feb 2000 12:33:58 +0000
From: Fergal Daly <fergal at esatclear dot ie>
Subject: Re: Appropriate list?
"Neil R. Smith" wrote:
>
> I've sent questions about making popper 2.53 on an SGI to this
> list twice, but have received no response. Does this mean that
> no one has anything to offer, or that this is not the
> appropriate list for the questions? If I am asking in the wrong
> place, please accept my appologies and advise as to a more
> appropriate list.
> Thanks, -Neil
There was a similar problem a while ago, I don't know if my suggestion
solved it but one thing you could do is adding -w to the CFLAGS (and
maybe LDFLAGS if it's there) in the makefiles and removing any other
options that begins with -w. This should suppress all warnings from gcc,
so it won't reach the limit of 50.
I think the problem is that you're linking to the SGI c library rather
than the gnu c library and gcc doesn't like some of the stuff that the
SGI header files contain, maybe...
Fergal
Subject: qpopper 3.0b34 man page
Date: Tue, 22 Feb 2000 14:08:43 +0100 (MET)
From: Georg Schwarz <schwarz at physik.tu-berlin dot de>
a brief remark re. qpopper 3.0b34's popauth man page: it mentions a -help
option which is not implemented. Should be removed IMHO.
PS: I've just added compioler switches for the following features:
- short options in popauth
- private db for apop (also if home is mounted via NFS)
- fallback to "normal" authentification for APOP users
- blocking of cleartext passwords
- blocking of cleartext passwords except for clients connecting from loopba=
ck
I'm currently testing it, but I'll soon send in patches.
Now that if've dug into the code (at least a bit), does anybody have
some other detail in mind that might need improvement?
--
Georg Schwarz schwarz at physik.tu-berlin dot de
Institut für Theoretische Physik Tel. +49 30 314-24254, FAX -21130
Technische Universität Berlin http://home.pages.de/~schwarz/
Sekr. PN 7-1, Hardenbergstr. 36, D-10623 Berlin IRC: kuroi
Subject: CHECK_SHELL?
Date: Tue, 22 Feb 2000 15:12:41 +0100 (MET)
From: Georg Schwarz <schwarz at physik.tu-berlin dot de>
Has anybody managed to compile qpopper 3.0beta with CHECK_SHELL set on
anything but Linux (and maybe *BSD)? It seems to use the getusershell()
system function is not seem to be available on other systems, is it?
Are there any other functions to use instead?
--
Georg Schwarz schwarz at physik.tu-berlin dot de
Institut für Theoretische Physik Tel. +49 30 314-24254, FAX -21130
Technische Universität Berlin http://home.pages.de/~schwarz/
Sekr. PN 7-1, Hardenbergstr. 36, D-10623 Berlin IRC: kuroi
Date: Tue, 22 Feb 2000 16:00:40 +0100
From: "Nordh, Valter" <valter at che.chalmers dot se>
Subject: Re: Appropriate list?
--=====================_79278843==_.ALT
Content-Type: text/plain; charset="us-ascii"; format=flowed
I have compiled qpopper 2.53 on a SGI system, running IRIX 6.5.5
I used:
./configure --prefix=/machine --enable-apop=/etc/apop.auth
--with-popuid=your_user
The option --with-popuid=cheuser didn't work, so I had to change in config.h
# diff config.h config.h-generated
< #define POPUID "your_user"
---
> #define POPUID "pop"
Lots of warnings, but no errors....
If you are planning to use apop I have some useful scripts and ideas as well.
/Valter
At 01:17 2000-02-22, Neil R. Smith wrote:
>I've sent questions about making popper 2.53 on an SGI to this
>list twice, but have received no response. Does this mean that
>no one has anything to offer, or that this is not the
>appropriate list for the questions? If I am asking in the wrong
>place, please accept my appologies and advise as to a more
>appropriate list.
> Thanks, -Neil
>--
>Neil R. Smith, Comp. Sys. Mngr. neils at ariel.met.tamu dot edu
>Dept. Atmospheric Sci., Texas A&M Univ. 979/845-6272 FAX:979/862-4466
--=====================_79278843==_.ALT
Content-Type: text/html; charset="us-ascii"
<html>
<font size=3>I have compiled qpopper 2.53 on a SGI system, running IRIX
6.5.5<br>
I used:<br>
./configure --prefix=/machine --enable-apop=/etc/apop.auth
--with-popuid=your_user<br>
The option --with-popuid=cheuser didn't work, so I had to change in
config.h<br>
<br>
# diff config.h config.h-generated <br>
< #define POPUID "your_user"<br>
---<br>
> #define POPUID "pop"<br>
<br>
Lots of warnings, but no errors....<br>
<br>
If you are planning to use apop I have some useful scripts and ideas as
well.<br>
<br>
/Valter<br>
<br>
<br>
At 01:17 2000-02-22, Neil R. Smith wrote:<br>
<blockquote type=cite cite>I've sent questions about making popper 2.53
on an SGI to this<br>
list twice, but have received no response. Does this mean
that<br>
no one has anything to offer, or that this is not the <br>
appropriate list for the questions? If I am asking in the
wrong<br>
place, please accept my appologies and advise as to a more <br>
appropriate list.<br>
Thanks, -Neil <br>
-- <br>
Neil R. Smith, Comp. Sys.
Mngr.<x-tab> </x-tab><x-tab> </x-tab>neils at ariel.met.tamu dot edu<br>
Dept. Atmospheric Sci., Texas A&M
Univ.<x-tab> </x-tab>979/845-6272
FAX:979/862-4466</font></blockquote><br>
</html>
--=====================_79278843==_.ALT--
Date: Tue, 22 Feb 2000 09:34:32 -0800
From: Randall Gellens <randy at qualcomm dot com>
Subject: Re: No pop serving on an NIS client?
At 7:03 PM -0600 2/21/00, Neil R. Smith wrote:
>Well, I think I've just deduced part of my "password incorrect"
>errors - the server that I wish to run popper on is an NIS client,
>and the /etc/passwd has no users in it; just the std. + entry
>as the last line. When I create a valid user entry in the
>/etc/passwd, with valid password, as an entry prior to the + entry,
>then my built popper works for that user.
>
>I must have missed this point in the FAQ or INSTALL doc. Is this
>true for popper? It can't be run on an NIS client?
>
>(Note: the /etc/passwd file is not shadowed)
>
>Thanks, -Neil
>--
>Neil R. Smith, Comp. Sys. Mngr. neils at ariel.met.tamu dot edu
>Dept. Atmospheric Sci., Texas A&M Univ. 979/845-6272 FAX:979/862-4466
I've built and run Qpopper on several NIS client systems. It seems
NIS replaces the libraries, so when Qpopper asks for password
entries, it actually gets them from NIS. It gets what appears to be
a classic passwd entry, with encrypted password and everything in the
right fields. I didn't do anything special to Qpopper.
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly-selected tag: ---------------
Kids Make Nutritious Snacks
--Newspaper headline
Subject: Patch for Qualcomm popper 3.0b34
Date: Wed, 23 Feb 2000 00:24:50 +0100 (MET)
From: Georg Schwarz <schwarz at physik.tu-berlin dot de>
--ELM951261890-19978-0_
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Attached please find a patch for Qualcomm's qpopper 3.0b34 which adresses
the following issues:
- bug fix: pop_msg.c would mistake client FQDNs starting on a digit
(such as 1host.example.com) as IPs.
- new features, enabled by compile time defines (appropriate configure
switches should be added!):
FALLBACKCLEARTEXT: even if a user has set an APOP (or similar) password
he can still as well log in via clear text passwords supplied by the
system's passwd data base (if that is allowed; see below)
NOCLEARTEXT: do not allow clear text passwords for anybody
ONLYLOCALCLEARTEXT: do not allow clear text passwords except for clients
conneting via the loopback interface (127.0.0.0/8); can be used to
enable tunnneling via ssh or ssl.
PRIVATEDB: have an individual APOP (or similar) database for each user in
their home directories. popauth no longer needs to be suid. popuid
is not used here.
CHEUID: the APOP database is accessed with the respective user's euid
(probably only makes sense if PRIVATEDB is used as well). Useful when
haing nfs mounted home directories, for example.
POPAUTHSHORTOPT: adds short (one letter) options to popauth.
I hope those changes will make it into the regular release.
feedback is appreciated!
--
Georg Schwarz schwarz at physik.tu-berlin dot de
Institut für Theoretische Physik Tel. +49 30 314-24254, FAX -21130
Technische Universität Berlin http://home.pages.de/~schwarz/
Sekr. PN 7-1, Hardenbergstr. 36, D-10623 Berlin IRC: kuroi
--ELM951261890-19978-0_
Content-Type: *unknown*/
Content-Disposition: attachment; filename=qpopper3.0b34-patch.gz
Content-Description: qpopper3.0b34-patch.gz
Content-Transfer-Encoding: base64
--ELM951261890-19978-0_--
Date: Wed, 23 Feb 2000 10:03:16 +1000
From: David Bussenschutt <D.Bussenschutt at mailbox.gu.edu dot au>
Subject: Re: CHECK_SHELL?
I can't say for sure for 3.0x..., but I compiled 2.53 fine with CHECK_SHELL
(and also with the patches I posted) on solaris 2.6 last week, so it must
be implememted on solaris systems. (unless something else broke it in
qpopper from 2.53-> 3.0x). Any other information anyone?
David.
At 03:12 PM 2/22/00 +0100, you wrote:
>Has anybody managed to compile qpopper 3.0beta with CHECK_SHELL set on
>anything but Linux (and maybe *BSD)? It seems to use the getusershell()
>system function is not seem to be available on other systems, is it?
>Are there any other functions to use instead?
>--
>Georg Schwarz schwarz at physik.tu-berlin dot de
>Institut für Theoretische Physik Tel. +49 30 314-24254, FAX -21130
>Technische Universität Berlin http://home.pages.de/~schwarz/
>Sekr. PN 7-1, Hardenbergstr. 36, D-10623 Berlin IRC: kuroi
>
>
--------------------------------------------------------------------
David Bussenschutt Email: D.Bussenschutt at mailbox.gu.edu dot au
Senior Computing Support Officer & Systems Administrator/Programmer
Location: Griffith University. Information Technology Services
Brisbane Qld. Aust. (TEN bldg. rm 1.33) Ph:(07)38757079
--------------------------------------------------------------------
From: "Kenneth Porter" <shiva at well dot com>
Date: Tue, 22 Feb 2000 18:15:58 -0800
Subject: Re: qpopper 2.53's popauth doesn't work?
On Sat, 19 Feb 2000 16:13:59 +0100 (MET), Georg Schwarz wrote:
>popauth from qpopper 2.53 does not seem to work. Trying to add a
>user's password I always get:
>
>With the latest beta version it works. Is this a known bug? If so, what causes
>it, and how could it be fixed?
>BTW, this is RH Linux 5.2 on Intel.
It's a file-locking problem. popauth opens the file first with the gdb
API and then with fopen, but gdb locks the file exclusively so the
fopen fails. The fopen is superfluous as the resulting handle is never
used. The 3.0 code eliminates the fopen.
I tried using 2.53 briefly on RH5.2, but never got it working. 3.0 has
been pretty solid except for a small number of releases. I've been
waiting a day or two after each release to let the more adventurous
shake them out, so I've managed to avoid the bad ones. Check out my
source RPM package at http://rpmfind.net/linux/RPM/qpopper.html.
Ken
mailto:shiva at well dot com
http://www.sewingwitch.com/ken/
From: Mats Luspa <mats.luspa at irf dot se>
Subject: qpopper3.0b34
Date: Thu, 24 Feb 2000 13:06:41 +0100 (MET)
Hello!
I noticed a small error in the qpopper3.0b34.tar.Z package that I got from
ftp://sunsite.doc.ic.ac.uk/Mirrors/ftp.qualcomm.com/eudora/servers/unix/popper/
When you compile to use KERBEROS (-DKERBEROS), then in file pop_init.c in
function authenticate row 133 there was written return( P_FAILURE ). It
should be return( POP_FAILURE ).
Correct me if I'm wrong.
/Regards Mats
--
--------------------------------------------------------------------
Mats Luspa Phone: +46 (0)980 79 022
Institutet foer Rymdfysik Fax: +46 (0)980 79 050
Swedish Institute of Space Physics email: mats.luspa at irf dot se
Box 812
SE-981 28 Kiruna, Sweden
--------------------------------------------------------------------
Date: Thu, 24 Feb 2000 13:46:36 +0100
From: Joaquin Vidal Balanza <joaquin.vidal at f-integra dot org>
Subject: RFP: Request for patch
Hello,
I need a patch (for qpopper of course) that allow pop user access
without having an account on my host. I have a hundred mail accounts to
manage.
Do you know any patch to do this?
Thanks in advance,
Joaquin
Date: Thu, 24 Feb 2000 11:27:04 -0300
From: PSantos <psantos at gol.com dot br>
Subject: Re: RFP: Request for patch
Ola Joaquin,
Thursday, February 24, 2000, 9:46:36 AM, voce escreveu:
JVB> I need a patch (for qpopper of course) that allow pop user access
JVB> without having an account on my host. I have a hundred mail accounts to
JVB> manage.
JVB> Do you know any patch to do this?
how about to use a MySQL+ Qpopper?
the only problem: the last patch (MySQL+Qpopper) was at 3.0b23
release. after that, avery week there are a new beta release and th
guys who make this kind of patch stoped to made it!
http://www.riverstyx.net/qpopmysql/ had the patch and instructions.
BUT, remember: 3.0b23 release... the last is 3.0b34 !!!!!!!!!!!!!!!
Obrigado e Saudações,
---------------------
Paulo Santos
um GOL na Internet Gratuita
mailto:psantos at gol.com dot br