Mailman 3 March 2000 - Mailman-Developers

mailman 2.0beta1 problem: Mail delivery failed: returning message to sender (fwd)
by Gergely Madarasz March 22, 2000

March 22, 2000

Hello, I've just upgraded my test machine to 2.0beta1 ... I sent a test message to the list, it was delivered correctly to its members and then I got the error message attached below. Anyway it seems quite strange to me... first, why does scripts/post write anything to stdout, second, why does it write "starting" twice ? The logs/post file has the following: Mar 22 19:02:22 2000 (6870) post to test from gorgo(a)sztaki.hu, size=444, success Additionally in logs/error I have the following: Mar 22 19:02:22 2000 post(6870): starting Mar 22 19:02:22 2000 post(6870): ending -- Madarasz Gergely gorgo(a)sztaki.hu gorgo(a)linux.rulez.org It's practically impossible to look at a penguin and feel angry. Egy pingvinre gyakorlatilag lehetetlen haragosan nezni. HuLUG: http://mlf.linux.rulez.org/ ---------- Forwarded message ---------- Date: Wed, 22 Mar 2000 19:02:23 +0100 From: Mail Delivery System <Mailer-Daemon(a)thunderchild.aszi.sztaki.hu> To: gorgo(a)sztaki.hu Subject: Mail delivery failed: returning message to sender This message was created automatically by mail delivery software. A message that you sent could not be delivered to one or more of its recipients. The following address(es) failed: test(a)thunderchild.aszi.sztaki.hu: generated |/var/lib/mailman/mail/wrapper post test The following text was generated during the delivery attempt: ------ |/var/lib/mailman/mail/wrapper post test ------ starting ending starting ------ This is a copy of the message, including all the headers. ------ Return-path: <gorgo(a)sztaki.hu> Received: from lutra.sztaki.hu [193.225.86.1] by thunderchild.aszi.sztaki.hu with esmtp (Exim 3.12 #1 (Debian)) id 12XpSX-0001ml-00; Wed, 22 Mar 2000 19:02:21 +0100 Received: from localhost by sztaki.hu (PMDF V5.2-32 #42635) with ESMTP id <0FRU004014RXGU(a)sztaki.hu> for test(a)thunderchild.aszi.sztaki.hu; Wed, 22 Mar 2000 19:02:21 +0100 (MET) Date: Wed, 22 Mar 2000 19:02:21 +0100 (MET) From: Gergely Madarasz <gorgo(a)sztaki.hu> Subject: proba To: test(a)thunderchild.aszi.sztaki.hu Message-id: <Pine.GSO.4.10.10003221902010.19257-100000(a)lutra.sztaki.hu> MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII humm -- Madarasz Gergely gorgo(a)sztaki.hu gorgo(a)linux.rulez.org It's practically impossible to look at a penguin and feel angry. Egy pingvinre gyakorlatilag lehetetlen haragosan nezni. HuLUG: http://mlf.linux.rulez.org/

2 1

beta & smtp
by Ricardo Kustner March 22, 2000

March 22, 2000

I just upgraded the cvs version i had installed (which is a snapshot from almost a month ago i think) and when i tried to approve 11 posts, I got a mailman error... something about a connection refused (see below)... maybe mailman is trying more connections than postfix allows... but the weird thing is that i never had this problem before, so this has to do with some recent stmp changes. I'll investigate postfix & mailman further to see if i can figure out what exactly is going wrong... Ricardo. admin(32515): mlist.HandleRequest(request_id, v, comment) admin(32515): File "/usr/local/mailman/Mailman/ListAdmin.py", line 124, in HandleRequest admin(32515): self.__handlepost(data, value, comment) admin(32515): File "/usr/local/mailman/Mailman/ListAdmin.py", line 174, in __handlepost admin(32515): self.Post(msg) admin(32515): File "/usr/local/mailman/Mailman/MailList.py", line 1293, in Post admin(32515): Mailman.Handlers.HandlerAPI.DeliverToList(self, msg) admin(32515): File "/usr/local/mailman/Mailman/Handlers/HandlerAPI.py", line 55, in DeliverToList admin(32515): func(mlist, msg) admin(32515): File "/usr/local/mailman/Mailman/Handlers/SMTPDirect.py", line 41, in process admin(32515): conn = smtplib.SMTP(mm_cfg.SMTPHOST, mm_cfg.SMTPPORT) admin(32515): File "/usr/local/mailman/Mailman/pythonlib/smtplib.py", line 182, in __init__ admin(32515): (code, msg) = self.connect(host, port) admin(32515): File "/usr/local/mailman/Mailman/pythonlib/smtplib.py", line 216, in connect admin(32515): self.sock.connect(host, port) admin(32515): error: (111, 'Connection refused') -- Ricardo. --

4 6

Re: [Mailman-Developers] Mailman keying on Sender: Field
by J C Lawrence March 22, 2000

March 22, 2000

On Wed, 22 Mar 2000 08:11:11 -0600 Roberto Ullfig <rullfig(a)uchicago.edu> wrote: > I noticed that Mailman keys on the Sender: field when you email in > a subscription request. Is this really necessary? Is it possible > to reconfigure Mailman to key on the From: field instead? Is this > more or less desirable? Since confirmations are always in place it > seems to me that there's enough protection as it is without using > the Sender: field. There's a documented value you can edit in mm_cfg.py to change this. -- J C Lawrence Home: claw(a)kanga.nu ----------(*) Other: coder(a)kanga.nu --=| A man is as sane as he is dangerous to his environment |=--

1 0

Re: [Mailman-Users] confidentiality features
by Barry A. Warsaw March 22, 2000

March 22, 2000

Please trim followups to mailman-developers(a)python.org... >>>>> "mjinks" == <mjinks(a)bsd.uchicago.edu> writes: mjinks> First of all, in testing it appears that although the mjinks> "From:" header is stripped, the "X-Sender:" still comes mjinks> through, as does the sender's IP address. We're in a mjinks> university, so back-tracing an IP address to a nearby mjinks> individual is not that tough, and could certainly be seen mjinks> as a gap for confidentiality. Could X-Sender and IP also mjinks> be stripped? Have I missed some detail? Mailman doesn't know about X-Sender, so it doesn't by default strip those. You'd probably also have to grok through Received: headers and others to truly anonymize the message. mjinks> On the other hand, this user wants to be able to peek mjinks> behind the curtain in those cases where confidentiality mjinks> needs to be violated -- for example if a poster to the mjinks> list claims to be suicidal or otherwise seems dangerous. mjinks> I showed the user that he'd see the real return address if mjinks> we set the list to be moderated, but he didn't like that mjinks> idea. Any clues? Can identifying information be made mjinks> available, say, to the list owner only? Here's the approach I would explore, using the 2.0 code base. Write a message handler module (see Mailman/Handlers/* for examples), calling it Anonymize.py. Put this in the pipeline such that every message passes through it before it's delivered to the archiver, or usenet, or the list. In this module, you'd strip or munge all the headers you want. To support lifting the curtain, you can store (securely, on the file system) whatever information gleaned out of the message is necessary to link the poster with the email message. You might even key both off of a randomly generated ID, which you'd poke into the outgoing message. Now that I think about it, it wouldn't be too hard to extend this idea into a general framework for email anonymizing. Given an MTA that can handle highly dynamic updates to it's alias database, you're "randomly generated ID" above would be used to create a Reply-To: address pointing back to your anonymizing service. Your add-on stuff would handle delivering back to the original user based on the random address. Oooh, what a neat idea! Let me know if you decide to play around with this. It'd be really cool to add some more anonymizing features in a future version. -Barry

1 0

Mailman keying on Sender: Field
by Roberto Ullfig March 22, 2000

March 22, 2000

I noticed that Mailman keys on the Sender: field when you email in a subscription request. Is this really necessary? Is it possible to reconfigure Mailman to key on the From: field instead? Is this more or less desirable? Since confirmations are always in place it seems to me that there's enough protection as it is without using the Sender: field. -- Roberto Ullfig : rullfig(a)uchicago.edu Systems Administrator Networking Services and Information Technologies University of Chicago

3 3

Re: [Mailman-Developers] Betas coming soon
by Edmund Lau March 22, 2000

March 22, 2000

Well, I just read the Emails on the new beta version. I hate to ask it, but was there a timetable on when we might expect it to be fully released? Just wondering cuz I just sent an Email out to my users that I'm installing 1.1 tomorrow and I'm not sure if I should just wait for 2.0. Will the upgrade path be difficult? No pressure whatsoever! ;) -Ed

4 5

htmlformat.py missing </body>, quotes on attributes
by Jim Tittsler March 22, 2000

March 22, 2000

CVS of 2000-03-21 19:00 EST My /mailman/listinfo page was missing a </BODY> tag. It looks like it is missing from all pages generated by mailman/htmlformat.py As long as I was looking at the HTML source, I noticed that it quoted the attribute values for some of the tags (like for the BODY tag), but not those for table, table rows, and cells. That led to things like 100% and #99CCFF not being quoted literals. add </body> tag quote all attribute values -- Jim Tittsler, Tokyo Python Starship http://starship.python.net/crew/jwt/

1 0

PATCH: NEWS typo
by Benjamin B. Thomas March 21, 2000

March 21, 2000

I noticed a typo in the NEWS file during the lastest round of cvs commits: Cheers, benjy Index: NEWS =================================================================== RCS file: /projects/cvsroot/mailman/NEWS,v retrieving revision 1.14 diff -C2 -r1.14 NEWS *** NEWS 2000/03/21 06:04:14 1.14 --- NEWS 2000/03/21 21:54:55 *************** *** 30,34 **** this is not guaranteed (or handled) by Mailman currently. Be careful also if your smtpd is on a different host than the ! Mailman host. In practice, mail lossage has not be observed. For this reason cron/run_queue is no longer needed (see the --- 30,34 ---- this is not guaranteed (or handled) by Mailman currently. Be careful also if your smtpd is on a different host than the ! Mailman host. In practice, mail lossage has not been observed. For this reason cron/run_queue is no longer needed (see the

1 0

ToUsenet should use the new nntplib, modereader
by Jim Tittsler March 21, 2000

March 21, 2000

Barry, Posting also requires the server to be in 'mode reader' (at least our INN does). I think you only updated the cron/gatenews... Jim -- Jim Tittsler, Tokyo Python Starship http://starship.python.net/crew/jwt/

1 0

Re: [Mailman-Developers] Failed authentication
by J C Lawrence March 17, 2000

March 17, 2000

On Mon, 13 Mar 2000 23:08:38 +0100 (MET) Gergely Madarasz <gorgo(a)caesar.elte.hu> wrote: > On Mon, 13 Mar 2000 coder(a)kanga.nu wrote: >> On Mon, 13 Mar 2000 14:49:53 +0100 (MET) Gergely Madarasz >> <gorgo(a)caesar.elte.hu> wrote: >>> Please try the 3:3.0.15-2 packages. The 1.2 package set was an >>> NMU, and had some other problems too, probably due to an >>> incomplete or broken build environment. >> >> Got a set of those for Potato as versus Woody? Potato doesn't >> seem to have anything more recent than -1.2 (and I'm not about to >> roll a production box up to Woody just yet). > Potato should have -2, the packages were installed on master at > 8:41 GMT, some mirrors might be still out of date though... Okay, I just upgraded to the -2 packages, restarted Apache, and everything was heppy. The -1.1 and -1.2 packages are broken. -- J C Lawrence Internet: claw(a)kanga.nu ----------(*) Internet: coder(a)kanga.nu ...Honorary Member of Clan McFud -- Teamer's Avenging Monolith...

1 0

Re: [Mailman-Developers] French Templates.
by J C Lawrence March 16, 2000

March 16, 2000

I'm now getting lock failures upon approving messages held for moderation: Traceback (innermost last): File "/var/lib/mailman/scripts/driver", line 112, in run_main main() File "/usr/lib/mailman/Mailman/Cgi/admindb.py", line 112, in main HandleRequests(doc) File "/usr/lib/mailman/Mailman/Cgi/admindb.py", line 200, in HandleRequests list.HandleRequest(request, v, form[comment_key].value) File "/usr/lib/mailman/Mailman/ListAdmin.py", line 122, in HandleRequest self.HandlePostRequest(request_data[2:], value, comment) File "/usr/lib/mailman/Mailman/ListAdmin.py", line 131, in HandlePostRequest self.Post(msg, 1) File "/usr/lib/mailman/Mailman/MailList.py", line 1329, in Post self.ArchiveMail(msg) File "/usr/lib/mailman/Mailman/Archiver/Archiver.py", line 206, in ArchiveMail lock.lock() File "/usr/lib/mailman/Mailman/LockFile.py", line 245, in lock os.unlink(self.__tmpfname) OSError: [Errno 2] No such file or directory: '/var/lib/mailman/locks/mud-dev.archiver.lock.bush.16685' check_perms reports clean, Apache (v1.1-4) runs as user www-data.www-data (Debain defaults), but the Mailman user is list.list. $ pwd /var/lib/mailman/locks $ ls -lasp total 36 4 drwxrwsr-x 2 root list 4096 Mar 16 14:35 ./ 4 drwxrwsr-x 6 root list 4096 Mar 12 21:46 ../ 4 -rw-rw-r-- 1 root list 69 Mar 16 09:48 <site>.lock 4 -rw-rw-r-- 1 list list 74 Mar 16 14:12 mmqueue_run.lock 4 -rw-rw-r-- 2 www-data list 79 Mar 16 14:31 mud-dev.archiver.lock 4 -rw-rw-r-- 1 www-data list 79 Mar 16 14:31 mud-dev.archiver.lock.bush.24449 4 -rw-rw-r-- 2 www-data list 79 Mar 16 14:31 mud-dev.archiver.lock.bush.24499 4 -rw-rw-r-- 1 list list 70 Mar 16 14:35 mud-dev.lock This looks fishy to me. Blowing away the contents of the directory doesn't fix the problem. -- J C Lawrence Internet: claw(a)kanga.nu ----------(*) Internet: coder(a)kanga.nu ...Honorary Member of Clan McFud -- Teamer's Avenging Monolith...

1 0

PC / Internet fax solution
by port22＠rnhnet.com March 16, 2000

March 16, 2000

I stopped by your web site and thought you may be interested in our PC/internet based fax service. You can send 1 to 1,000,000 faxes in just minutes to any destination in the US, Canada, and Europe for as low as 7.4 cents per minute. Our user-friendly submission software lets you queue your job on our server within minutes. Why not try the system for FREE. We will be happy to send 100 Faxes on your behalf to your fax list immediately. Simply go to http://www.efaxcast.com/submit.htm

1 0

Pb with URL sent to modify your options
by Boris Granveaud March 16, 2000

March 16, 2000

Hi, I noticed a problem with the URL that is sent with the introduction mail (when you have just subscribed to a list). The URL is like this: http://www.python.org/mailman/options/mailman-developers/granveau@worldnet.… If you receive with Hotmail and you are using Internet Explorer, the URL doesn't work. The browser tries to open http://www.python.org/mailman/options/mailman-developers/ I think that if you replace granveau(a)worldnet.fr with granveau__at__worldnet.fr, it would work. Can anybody confirm this and provide a patch? Thank you, Boris.

1 0

cookies
by Ricardo Kustner March 15, 2000

March 15, 2000

Hi, this was just suggested to me by somebody from mailman-users: > Please write a patch which puts the string "Cookie could not be set" on the > web page so that I can see that pressing submit will not work :-) i think thats a good point... it would safe some user questions if MM tells exactly why the authorisation failed. I'm not sure if this subject has been touched already, but what allowing admins to choose between cookies and Basic-Authorization... it should be possible to use Basic-auth, right? Ricardo. --

1 0

File handle shortages
by J C Lawrence March 15, 2000

March 15, 2000

I'm now getting this, with no other changes in the system, when approving messages held for moderation. Same kernel, same file handle limits as before...: IOError writing outgoing queue exceptions.IOError/[Errno 23] Too many open files in system: '/var/lib/mailman/data/mm_q.10890.13' Traceback (innermost last): File "/var/lib/mailman/scripts/contact_transport", line 58, in ? File "/usr/lib/mailman/Mailman/OutgoingQueue.py", line 155, in enqueueMessage File "/usr/lib/mailman/Mailman/Utils.py", line 829, in open_ex File "/usr/lib/mailman/Mailman/Utils.py", line 824, in open_ex IOError: [Errno 23] Too many open files in system: '/var/lib/mailman/data/mm_q.10890.13' IOError writing outgoing queue exceptions.IOError/[Errno 23] Too many open files in system: '/var/lib/mailman/data/mm_q.10918.9' Traceback (innermost last): File "/var/lib/mailman/scripts/contact_transport", line 58, in ? File "/usr/lib/mailman/Mailman/OutgoingQueue.py", line 155, in enqueueMessage File "/usr/lib/mailman/Mailman/Utils.py", line 829, in open_ex File "/usr/lib/mailman/Mailman/Utils.py", line 824, in open_ex IOError: [Errno 23] Too many open files in system: '/var/lib/mailman/data/mm_q.10918.9' Content-type: text/html ... Traceback (innermost last): File "/var/lib/mailman/scripts/driver", line 112, in run_main File "/usr/lib/mailman/Mailman/Cgi/admindb.py", line 112, in main File "/usr/lib/mailman/Mailman/Cgi/admindb.py", line 200, in HandleRequests File "/usr/lib/mailman/Mailman/ListAdmin.py", line 122, in HandleRequest File "/usr/lib/mailman/Mailman/ListAdmin.py", line 131, in HandlePostRequest File "/usr/lib/mailman/Mailman/MailList.py", line 1329, in Post File "/usr/lib/mailman/Mailman/Archiver/Archiver.py", line 206, in ArchiveMail File "/usr/lib/mailman/Mailman/LockFile.py", line 200, in lock File "/usr/lib/mailman/Mailman/LockFile.py", line 156, in __read IOError: [Errno 23] Too many open files in system: '/var/lib/mailman/locks/mud-dev.archiver.lock.bush.10877' The mail (at least) does seem to get thru. -- J C Lawrence Home: claw(a)kanga.nu ----------(*) Other: coder(a)kanga.nu --=| A man is as sane as he is dangerous to his environment |=--

1 0

French Templates.
by Hausermann Laurent March 15, 2000

March 15, 2000

Hi, I currently finalizing a communuty web server (for a former student association) in French, and I cannot find any ressources about French Templates for mailman 1.1. Do you now about any translation project ? Thanks a lot Laurent Hausermann

1 0

Re: [Mailman-Developers] Failed authentication
by J C Lawrence March 14, 2000

March 14, 2000

On Sun, 12 Mar 2000 20:30:17 -0800 J C Lawrence <claw(a)kanga.nu> wrote: > I have no bloody idea. POSTS to PHP work just fine. POSTs to > Perl CGIs work fine. Once I add the PHP3 module to Apache's > config, POSTs to Mailman stop. > Colour me confused. > As soon as I get few other things resolved, I'm going to see what > happens if I backlevel PHP, or even try PHP4... I've identified the exact setup that causes the behaviour under Debian/potato: If I install the following .debs for PHP everything works fine: php3-gd_3%3a3.0.15-1_i386.deb php3_3%3a3.0.15-1_i386.deb php3-mysql_3%3a3.0.15-1_i386.deb If however I install the current Debian/potato set: php3_3%3a3.0.15-1.2_i386.deb php3-gd_3%3a3.0.15-1.2_i386.deb php3-mysql_3%3a3.0.15-1.2_i386.deb with the same Apache configs in both cases (a LoadModule line and two AddTypes as per the default http.conf and srm.conf), or I install the prior the prior 2.0.15-1.1 package set, HTTP POSTs to MailMan CGIs no longer happen due to the fact that nothing reaches the stdin of the CGIs. It doesn't seem to matter whether I'm running: apache_1.3.9-12_i386.deb apache-common_1.3.9-12_i386.deb or: apache_1.3.9-11_i386.deb apache-common_1.3.9-11_i386.deb or the various libc6 packages in that date range. What I havem't tried, and what I'm going to leave up to Gergely Madarasz (CC'ed on this message) the Debian package maintainer for both the PHP3 and MailMan packages, is finding out if rebuilding everything with a current (or even the same) glibc version makes the problem go away (it hopefully being reproducable at his end). I'm willing to supply copies of my config files if there's interest. I'm going to bed. -- J C Lawrence Home: claw(a)kanga.nu ----------(*) Other: coder(a)kanga.nu --=| A man is as sane as he is dangerous to his environment |=--

3 4

Re: Bogus directory
by Jim Tittsler March 13, 2000

March 13, 2000

On Fri, Mar 03, 2000 at 10:49:45AM -0300, Rodolfo Pilas wrote in mailman-users: [...] > However I am in receipt of a message about bogus directory: [...] > List todos.tecnico has a bogus archive_directory: > /var/spool/mailman/archives/private/todos.tecnico > List todos.sistemas has a bogus archive_directory: > /var/spool/mailman/archives/private/todos.sistemas The "bogus archive directory" issue seems to come up fairly often. Perhaps it would be worthwhile to add a check to cron/nightly_gzip to find out if it is simply a matter of there not yet being any articles to archive. Mailman could check post_id or last_post_time. I've attached one possibility. -- Jim Tittsler, Tokyo Python Starship http://starship.python.net/crew/jwt/

2 1

Re: [Mailman-Developers] Failed authentication
by J C Lawrence March 13, 2000

March 13, 2000

On Sun, 12 Mar 2000 09:50:13 -0800 J C Lawrence <claw(a)kanga.nu> wrote: > The ~mailman/cgi-bin/admin is not getting anything on stdin. Not > good. Current findings: Installation of the PHP3 apache modules under Debian/potato kills POST methods for MailMan. I don't know why. Problem is, I need both. -- J C Lawrence Home: claw(a)kanga.nu ----------(*) Other: coder(a)kanga.nu --=| A man is as sane as he is dangerous to his environment |=--

2 2

a more uniform interface to private and public archives
by Todd Pfaff March 13, 2000

March 13, 2000

for mailman-1.1, the installation documentation suggests to create, for example, an 'Alias' path for the web server to the mailman public archives and a 'ScriptAlias' cgi-bin path to the private archives. what i'd like to suggest is that we make the interface more uniform by eliminating the 'Alias' path and access both private and public archives via a single cgi-bin interface. if the archive is private we require authentication, if not we simply bypass the authentication. i've done this with my mailman installation by doing the following: - created a new version of the MailMan/Cgi/private.py program - in mm_cfg.py, set PUBLIC_ARCHIVE_URL = '/mailman/private' PRIVATE_ARCHIVE_URL = '/mailman/private' - these could then be collapsed into one ARCHIVE_URL - we could also replace PUBLIC_ARCHIVE_FILE_DIR = os.path.join(PREFIX, 'archives/public') PRIVATE_ARCHIVE_FILE_DIR = os.path.join(PREFIX, 'archives/private') with one ARCHIVE_FILE_DIR, and we could also get rid of the public and private subdirectories altogether. in the new private.py i check listobj.archive_private and if it's set to 1 i do the usual private authentication as before. if it's not 1, i set is_auth to 1 and fall through. that's it. very clean and simple. does anyone see any problems with this? i think it certainly makes things more clear and straightforward. -- Todd Pfaff \ Email: pfaff(a)mcmaster.ca Computing and Information Services \ Voice: (905) 525-9140 x22920 ABB 132 \ FAX: (905) 528-3773 McMaster University \ Hamilton, Ontario, Canada L8S 4M1 \

2 1

Re: [Mailman-Developers] Failed authentication
by J C Lawrence March 13, 2000

March 13, 2000

On Sun, 12 Mar 2000 00:19:18 -0800 J C Lawrence <claw(a)kanga.nu> wrote: > Findings: read_urlencoded never reads anything from stdin. Zero. > Zilch. Nada. There's nothing there. The key edits: ... > According to that Apache is sending all the right data, it just > ain't getting there. I've also upgraded libc6 recently (Debian > system FWIW with packages libc6_2.1.3-[567]_i386.deb (notice regex > range)). I downgraded libc6 to the package versions (5 and 6) > that were installed when it did work to no effect. I also rebuilt > the Mailman C based CGI programs under the installed version of > libc6, and copied them over (in case something odd was going on) > to no observed effect. I changed cgi-wrapper.c to read: #include <stdio.h> #include "common.h" /* passed in by configure */ #define SCRIPTNAME SCRIPT #define LOG_IDENT "Mailman cgi-wrapper (" ## SCRIPT ## ")" /* GID that CGI scripts run as. See your Web server's documentation. */ #define LEGAL_PARENT_GID CGI_GID const char* logident = LOG_IDENT; char* script = SCRIPTNAME; const int parentgid = LEGAL_PARENT_GID; int main(int argc, char** argv, char** env) { int status; char* fake_argv[3]; char ac[31]; /* !!!!!!!!!!!edited!!!!!!!!!!!!! */ memset (ac, 0, 31); /* !!!!!!!!!!!edited!!!!!!!!!!!!! */ running_as_cgi = 1; if (getgid()>=100 && getgid()!=65534) check_caller(logident, parentgid)$ /* for these CGI programs, we can ignore argc and argv since they * don't contain anything useful. `script' will always be the driver * program and argv will always just contain the name of the real * script for the driver to import and execute (padded with two dummy * values in argv[0] and argv[1] that are ignored by run_script(). */ fake_argv[0] = NULL; fake_argv[1] = NULL; fake_argv[2] = script; fread (ac, 1, 30, stdin); /* !!!!!!!!!!!!edited!!!!!!!!!!!! */ fprintf (stderr, "---[%s]---",ac); /* !!!!!!!!!!!!edited!!!!!!!!!!!! */ status = run_script("driver", 3, fake_argv, env); fatal(logident, status, "%s", strerror(errno)); return status; } The ~mailman/cgi-bin/admin is not getting anything on stdin. Not good. -- J C Lawrence Home: claw(a)kanga.nu ----------(*) Other: coder(a)kanga.nu --=| A man is as sane as he is dangerous to his environment |=--

2 1

Re: [Mailman-Developers] Failed authentication
by J C Lawrence March 12, 2000

March 12, 2000

On Sun, 12 Mar 2000 15:24:06 +0100 (MET) Gergely Madarasz <gorgo(a)caesar.elte.hu> wrote: > On Sun, 12 Mar 2000, J C Lawrence wrote: >> Background data as this is also going to the Debian Mailman >> package maintainer: >> >> Mailman's web interface suddenly stopped working for me. The >> forms would display but I am unable authenticate not matter what >> browser (Netscape 4.05 and 4.7, Lynx, W3M or Mozilla M14) I use. > Hmm... unable to authenticate ? That means you get to the password > screen each time you try to submit a form ? This usually happens > if you don't access your admin interface on the Base url set for > your list... Nope, that's not the problem (and I'm accessing on the Base Url to boot). The problem is that the Mailman CGIs are not getting the form submission data from Apache. Apache appears to be supplying it (ScriptLog says it is), but by the time the Python code runs its no longer there. -- J C Lawrence Home: claw(a)kanga.nu ----------(*) Other: coder(a)kanga.nu --=| A man is as sane as he is dangerous to his environment |=--

1 0

Re: [Mailman-Developers] Failed authentication
by J C Lawrence March 12, 2000

March 12, 2000

Background data as this is also going to the Debian Mailman package maintainer: Mailman's web interface suddenly stopped working for me. The forms would display but I am unable authenticate not matter what browser (Netscape 4.05 and 4.7, Lynx, W3M or Mozilla M14) I use. On Thu, 09 Mar 2000 13:11:06 -0800 J C Lawrence <claw(a)kanga.nu> wrote: > On Thu, 9 Mar 2000 15:55:44 -0500 (EST) bwarsaw > <bwarsaw(a)cnri.reston.va.us> wrote: >>>>>>> "JCL" == J C Lawrence <claw(a)kanga.nu> writes: JCL> Aye, what's especially odd is that forms are process properly JCL> elsewhere (PHP for instance). So the Apache side and the JCL> modules sides are Okay its just something with the Mailman CGI JCL> wrapper. I guess. >> I still want to know at what point the cgi data is getting >> corrupted. Is it between Apache and the C wrapper? Is it >> between the C wrapper and invoking the Python driver script? Is >> it between the driver script and the specific Mailman/Cgi module? > I'm really not sure, and I'm not too sure how to find out either. > Could you give me some hints where to poke this? I've done some more digging, implemented ScriptLog in Apache, inserted sys.exit(1) in the admin script to force a failure right after the FieldStorage call, and then instrumented pythonlib/cgi.py. Findings: read_urlencoded never reads anything from stdin. Zero. Zilch. Nada. There's nothing there. The key edits: --<cut>-- def read_urlencoded(self): """Internal: read data in query string format.""" qs = self.fp.read(self.length) qs=str(qs) sys.stderr.write (qs) sys.stderr.write ('=======') sys.stderr.write (str(self.length)) ...etc --<cut>-- --<cut>-- %% [Sat Mar 11 23:58:57 2000] POST /lists/admin/library/ HTTP/1.0 %% 500 /var/lib/mailman/cgi-bin/admin %request Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, image/png, */* Accept-Charset: iso-8859-1,*,utf-8 Accept-Encoding: gzip Accept-Language: en Connection: Keep-Alive Content-length: 40 Content-type: application/x-www-form-urlencoded Cookie: mud-dev:admin="(lp1\012S'216.148.243.2'\012p2\012aI952846157\012aI952856957\012aI-96722254\012a." Host: www.kanga.nu Referer: http://www.kanga.nu/lists/admin/library/ User-Agent: Mozilla/4.7 [en] (X11; I; Linux 2.2.13 i686; Nav) adminpw=xxxxx&request_login=Let+me+in... %response %stderr =======40 --<cut>-- According to that Apache is sending all the right data, it just ain't getting there. I've also upgraded libc6 recently (Debian system FWIW with packages libc6_2.1.3-[567]_i386.deb (notice regex range)). I downgraded libc6 to the package versions (5 and 6) that were installed when it did work to no effect. I also rebuilt the Mailman C based CGI programs under the installed version of libc6, and copied them over (in case something odd was going on) to no observed effect. Note: Plopping the cgi.py in as a standalone cgi tests out as working using GET methods (ie encode vars on the URL). I haven't built a framework yet to test POST for that module. I guess the next step is to either build that framework, or test-drive the C CGIs to make sure stdin really has the right bumph sitting there waiting. -- J C Lawrence Home: claw(a)kanga.nu ----------(*) Other: coder(a)kanga.nu --=| A man is as sane as he is dangerous to his environment |=--

2 1

Re: [Mailman-Developers] Posting to 2 lists breaks newsgating for the second one
by Harald Meland March 10, 2000

March 10, 2000

[Jim Hebert] > I filed a jitterbug report about this a couple weeks back... > > Applix is fairly motivated about using mailman for their internal lists > but needs the ability to crosspost on the mail side and have it show up in > all the right newsgroups too. (If this doesn't make sense, visit the > jitterbug report below.) I suspect this has something to do with the fact that two separate Usenet messages can't have the same Message-Id. If you could have a look in your logs/post Mailman log, you might actually find some foundation for my suspicion (due to the except: clause around line 125 of Mailman/Handlers/ToUsenet.py). If this indeed is the problem at hand, I'm not sure how best to fix it. The optimal solution would be if mail "crossposts" actually turned into proper Usenet crossposts, and not multiple (nearly) identical posts to different newsgroups. However, I don't know enough about NNTP to say if implementing something like that is at all possible. A quick band-aid solution might be to change the Message-Id a wee bit (in a way that makes the change differ from list to list) before passing the message to the NNTP server. > So, I got them to put some money behind the bug report, if someone will > fix it: > > http://www.python.org/mailman-bugs/incoming?id=202;user=guest > http://www.cosource.com/cgi-bin/cos.pl/wish/info/295 > > (Sorry if this is better posted to the -users list, I just joined > both lists and don't have a clear sense of which one would have been > better...) I haven't had time to read -users for a looooong time, but I think -developers is the proper place to post specific feature requests like this. > All in all, we're loving mailman! =) Great :) -- Harald

1 0

Re: [Mailman-Developers] Mailman password and administrative reminders
by Harald Meland March 10, 2000

March 10, 2000

[David Birnbaum] > Here's the headers for the reminder: > > Return-Path: <mailman-owner(a)chelsea.net> > Received: from killian.chelsea.net (localhost [127.0.0.1]) > by killian.chelsea.net (8.9.3/8.9.3) with ESMTP id FAA15246 > for <donad(a)gaigo.sundai.ac.jp>; Wed, 1 Mar 2000 05:52:23 -0500 (EST) > Date: Wed, 1 Mar 2000 05:52:23 -0500 (EST) > Message-Id: <200003011052.FAA15246(a)killian.chelsea.net> > From: mailman-owner(a)chelsea.net > Subject: thebody.com mailing list memberships reminder > To: ###A USER### > X-No-Archive: yes > Precedence: bulk > X-Mailman-Version: 1.1 > Precedence: bulk > List-Id: The Test List <test.chelsea.net> > > It should be from (I think), thebody-admin(a)thebody.com, with an Errors-To: > header. > > Not sure why List-Id: is The Test List...that's odd. That's because the script, while iterating through all lists, looks for the first "public" list. When found, that list is remembered, and that list's (fasttrack) delivery pipeline is used for sending out all the messages. As the "test" list is usually created pretty early on in any particular Mailman installation, it is not entirely uncommon that this list's List-Id: gets stuck on all password reminders... Normally it is nice that Mailman indicates which list's fasttrack pipeline you received a particular message through, but I can see that it might be confusing in this situation. However, a proper remedy is complicated by the lack of a list-independent Mailman mail address -- and thus, as Barry mentioned, also by the lack of a user database. -- Harald

2 1