Mailman 3 July 2000 - Mailman-Developers

Mailman Database support
by Andy Lowe 27 Jul '00

27 Jul '00

I would like to take the lead on this project if you don't mind. I have to have a solution by the 9th of August so I think I will be leading the grunt code writing while looking to those of you more knowledgable about the Mailman code base for ideas :) As I look at it there are two sections of data. 1: the list of recipients. 2. the List configuration (the list of mail lists may be a third, but for now I am assuming it is tied into the list configurations). The list configurations data is basically all in the form of variable=value while the list of recipients is a list of recipients with a few values attached (digest, hide from list etc. . .) I believe one easy way to do this is to replace the Maillist.Save, Maillist.Load, and Maillist.Create methods. This should be pretty easy, and will allow development of other areas to continue with problems, but has some problems. I am running lists with 15 thousand plus users, and it seems like a great waste to have to do a query against a DB and load all that info into memory for every web page etc. . . thing which mailman does. And it offers problems with concurrent access. I think it would be better / harder to divide mailman into a front end and two back ends, one for the list configurations and one for the list of recipients with a standard api for each back end. Then when the front end starts, it looks in mm_cfg.py for the type of back end used for the list configurations and imports the correct back end for that type. If the front end then needs to access the list of recipients, it imports the module for the back end of the particular database. The problem with this solution is that it appears to require a near total rewrite of parts of mailman. Another aspect which needs to be addressed is fixed DB support vs. dynamic DB support. For example, I have a table in an Oracle database which I have to use. So in effect I need to be able to give mailman insert, remove and select statements so that it will know how to query my database and tables. I don't need to create tables or any such stuff. This is what I call dynamic support. Most people I think will need fixed support. They will want mailman to generate /access it's own tables etc. . . They could care less about how the tables are set up. I have said alot of things in this message I am not certain about, so please correct me where I went astray, in the mean time I will try to sit down tonight and write down a potential rough draft of the requirements / API's assuming we take the difficult path. Any input would be greatly appreciated. Would anyone be interested in writing back ends for other things then Oracle? I can see RDBs, LDAP, and maybe RADIUS, but I do not understand ZOPE well enough to visualize the connection. Enough rambling. .. A. L.

2 1

broken latest CVS "Email my Pasword To Me
by Mentor Cana 27 Jul '00

27 Jul '00

The latest CVS has the link broken: Here is the URL I get when I click to "Email my Password To Me": http://www.alb-net.com/handle_opts/test-l/mentor@alb-net.com where is the path to /mailman ? later, mentor

2 4

Re: [Mailman-Developers] Content-Type and Content-Transfer-Encoding headers
by J C Lawrence 27 Jul '00

27 Jul '00

On Thu, 27 Jul 2000 13:48:06 +0400 Michael Sobolev <mss(a)transas.com> wrote: > Content-Type: text/plain; charset=koi8-r > And, as our users use different MUA, Content-Transfer-Encoding may > well be different from 8bit, and it's usually equal to base64 or > quoted-printable. Mailman will nto help you here. You need to use an external archiver that handles character set definitions properly. MHonArc is such an animal: http://www.mhonarc.org/ > The problem is two folded. The mailing list engine just appends > footer maillist administrator creates and this usually results in > garbage... This is a known problem as the list signature code is not MIME aware. The only useful solution is to turn off list signatures. -- J C Lawrence Home: claw(a)kanga.nu ---------(*) Other: coder(a)kanga.nu http://www.kanga.nu/~claw/ Keys etc: finger claw(a)kanga.nu --=| A man is as sane as he is dangerous to his environment |=--

2 1

Content-Type and Content-Transfer-Encoding headers
by Michael Sobolev 27 Jul '00

27 Jul '00

Gentlemen, As far as I see, this question arised several time, but no answer was given. Here is the problem. I tried to create few mailing lists for my company. Our native language is Russian, which means that Content-Type header usually has something like like as its value: Content-Type: text/plain; charset=koi8-r And, as our users use different MUA, Content-Transfer-Encoding may well be different from 8bit, and it's usually equal to base64 or quoted-printable. The problem is two folded. The mailing list engine just appends footer maillist administrator creates and this usually results in garbage, as --- mailing list blah... usually is meaningless if interpreted according to Content-Transfer-Encoding header if the latter is equal to base64 or quoted-printable. The second important issue is maillist archives. Neither `charset' value from Content-Type header nor Conten-Transfer-Encoding header are used while creating html pages for messages. Which makes the archives completely unusable. I found a hack for the first problem (I just make no footer for the mailing lists) and may avoid the second problem by using an external archiver. But this is not a solution, as it looks like mailman (2.0 beta4 at the moment) cannot be used in environments where non iso-8859-1 or us-ascii charset are used... I'd appreciate any comments. Thanks, -- Misha

1 0

Re: [Mailman-Developers] config_list bug with subscribe_policy
by Hrvoje Niksic 27 Jul '00

27 Jul '00

bwarsaw(a)beopen.com (Barry A. Warsaw) writes: > | # legal values are: > | # 0 = "confirm" > | # 1 = "require approval" > | # 2 = "confirm+approval" > | subscribe_policy = 1 > > Hrvoje> The number is correct, but the comment is not; the list in > Hrvoje> question does not require approval -- it merely asks for > Hrvoje> confirmation. > > Hrvoje> config_list should probably have a special-case for > Hrvoje> subscribe_policy, as the graphical tools do. Otherwise > Hrvoje> the comment is very misleading for people who hope to > Hrvoje> use config_list to import lists to Mailman. > > Try this patch. Yes, it seems to work fine now. Thanks! > - i = 0 > + # TBD: This is disgusting, but it's special cased > + # everywhere else anyway... > + if varname == 'subscribe_policy' and \ > + not mm_cfg.ALLOW_OPEN_SUBSCRIBE: > + i = 1 > + else: > + i = 0 > for choice in data[2]: > print '# ', i, '= "%s"' % choice > i = i + 1 I propose that you remove this kludge by setting the values to *always* be in the [0, 3) range, except that setting it to zero is disallowed if the site disallows it. In that case, the graphical tools could simply gray out the 0th option.

1 0

config_list bug with subscribe_policy
by Hrvoje Niksic 27 Jul '00

27 Jul '00

The subscribe_policy parameter is a bit specific in that it can be 0, 1, 2, or 3, except that 0 can be turned off on a per-site basis (which is done by default, at least under Debian). The web GUI makes sure that the RadioButtonArray gets set up with the current setting of subscribe_policy correctly laid out, which depends on whether the site-wide option is turned on or off. However, the config_list utility does not seem to understand the convention, and prints the following: # legal values are: # 0 = "confirm" # 1 = "require approval" # 2 = "confirm+approval" subscribe_policy = 1 The number is correct, but the comment is not; the list in question does not require approval -- it merely asks for confirmation. config_list should probably have a special-case for subscribe_policy, as the graphical tools do. Otherwise the comment is very misleading for people who hope to use config_list to import lists to Mailman.

2 1

Mailman and Oracle
by Andy Lowe 26 Jul '00

26 Jul '00

I have a project at work where I have to get some sort of mail list manager working with information stored in our Oracle database. I would like to use Mailman for this purpose since I am comfortable programming in Python and I like Mailman. I would like to know if anyone is working on database support for Mailman and or if you are interested in integrating my work back into the official release of Mailman. If anyone is interested in using my efforts, then I will be happy to take direction from the development team as to how to implement the features I need. If you are not interested, then I will just hack it to work with my system. I would like to develop a system where you can choose from several back ends (Marshal, Oracle MySql etc. . .) on a per list or per site basis. Including list preferences, user preferences etc. . . I don't know how much desire there would be for this since the oracle implementation would seem to require installing the oracle client on the mail list machine and installing DCOracle. I am interested in any ideas, suggestions etc. . . anyone has on the subject. Sincerely, Andy Lowe

3 2

Re: [Mailman-Developers] Attachment blocking?
by J C Lawrence 26 Jul '00

26 Jul '00

On Wed, 26 Jul 2000 11:16:30 -0700 Risto Kotalampi <risto(a)kotalampi.com> wrote: > One of my list just recently dsitributed a vbs virus to > everyone. Any plans to have an option on the list that would block > all attachments going to the list? You can do this now by inserting a filter in your aliases file before passing to Mailman's wrapper. -- J C Lawrence Home: claw(a)kanga.nu ---------(*) Other: coder(a)kanga.nu http://www.kanga.nu/~claw/ Keys etc: finger claw(a)kanga.nu --=| A man is as sane as he is dangerous to his environment |=--

1 0

Attachment blocking?
by Risto Kotalampi 26 Jul '00

26 Jul '00

One of my list just recently dsitributed a vbs virus to everyone. Any plans to have an option on the list that would block all attachments going to the list? Risto

1 0

lowercase listnames?
by Dan Mick 26 Jul '00

26 Jul '00

newlist lowercases the listname before creating it. Is there any good reason for this? Unix isn't case-insensitive; do we really want Mailman to be?

5 6