Mailman 3 May 2014 - Mailman-Developers

Re: [Mailman-Developers] GSOC 2014: CI tool for the Mailman suite and postorius improvements
by varun sharma 23 May '14

23 May '14

Sorry for being out of touch for so long, i was having exams during community bonding period,so i couldn't spare time for discussions. But after doing 3 irc meetings with my mentors,i have now pretty clear understanding on how to proceed with my project but in order to get further suggestions from community, i am posting all the details of my project: 1. CI tool for Mailman Suite (name yet to decide): The first part that i will be doing for next 4 weeks is CI tool for Mailman Suite. I have already started work on that and created a launchpad repository at https://code.launchpad.net/~varun/mailman/ci_tool. I will try to adhere to my timeline, so by the end of this week i'll be finishing with all the unit tests of ci-tool. Right now i've written the rough implementation of pull and test commands using python's optparse module http://bazaar.launchpad.net/~varun/mailman/ci_tool/view/head:/ci-tool: ci-tool -p all (will pull all the projects which includes mailman, mailman.client, postorius, postorius_standalone, hyperkitty and kittystore) ci-tool -p hyperkitty (for pulling a particular project) ci-tool -t all (will run all the unit tests from all the projects) ci-tool -t postorius (run tests specifically) The tool is running fine but buildbot is unable to detect the failed tests in case of running them from ci-tool, i guess i need to write plugin for nose2 to fix this problem. Link to one of such build is http://buildbot.sharmalabs.com/builders/runtests/builds/116 The basic commands that i think it should have are 'pull', 'test' and 'try' but suggestions are welcomed for any new addition. I request all the creative (and non-creative) members of the community to suggest some cool names for the ci-tool :) 2. The UI Testing Framework for Postorius: I have also started working on UI framework and after discussing the possible testing frameworks that we can use in that, we decided to go with selenium. I wrote 5 basis tests for testing the ui of postorius with selenium webdriver ( http://bazaar.launchpad.net/~varun/mailman/ci_tool/view/head:/test_postoriu…) and the i am yet to expand it to hyperkitty. The idea is to include this framework into ci-tool as an optional feature. I will be very happy to hear the feedback from the community, so that i'll be able to make changes to my project accordingly and at right time. Many thanks to my mentors Florian, Steve and Sneha for helping me in making this project feasible. @abhilsh: Sorry, i was really busy during that time,so couldn't replay accordingly and i'll keep that in mind for future discussions. @barry: Thanks barry, I'll make sure the test suite works fine with all the major MTAs, DBMSs and *nix platforms. Thanks Varun

1 0

Re: [Mailman-Developers] Fixing DMARC problems with .invalid munge
by Franck Martin 21 May '14

21 May '14

The trouble with .invalid is that it is a domain that do not accept emails. Therefore why should you accept emails from a domain that does not allow you to reply to it? It is bound in the future to create issues when people move to more serious/ubiquitous domain reputation schemes.

10 20

Re: [Mailman-Developers] Fixing DMARC problems with .invalid munge
by Franck Martin 17 May '14

17 May '14

You can also apply this patch: http://bazaar.launchpad.net/~mlm-author/mailman/2.1-author/revision/1341?re… Rather than injecting an invalid domain in the From: and weakening more the security of email...

2 1

Anyone tried the DMARC mail address translucent forwarder hack?
by John Levine 17 May '14

17 May '14

> and a really evil one where you append a name >that points to a server that rewrites the address and remails it, e.g. >mmeyer(a)yahoo.com.remail.lists.org -> mmeyer(a)yahoo.com. This is apparently what LISTSERV does, give or take details of the syntax of the forwarding address. Has anyone tried this with Mailman? R's, John PS: You don't have to tell me all the reasons it's not a good idea. I'm just wondering how bad an idea it turns out to be in practice.

3 3

Re: [Mailman-Developers] Fixing DMARC problems with .invalid munge
by Franck Martin 17 May '14

17 May '14

Upgrade to 2.1.18 Toute connaissance est une réponse à une question. > On May 16, 2014, at 20:43, "Bob Puff" <bob(a)nleaudio.com> wrote: > > So guys... Is there a simple little hack we can do within MM 2.1 to try to > mitigate this issue, by adding .invalid or some other extension? I've got a > few lists that are getting to the point where MM sends the probe email, and > then figures it is not a bouncing address, but a lot of emails are not being > delivered. > > Bob > > _______________________________________________ > Mailman-Developers mailing list > Mailman-Developers(a)python.org > https://mail.python.org/mailman/listinfo/mailman-developers > Mailman FAQ: http://wiki.list.org/x/AgA3 > Searchable Archives: http://www.mail-archive.com/mailman-developers%40python.org/ > Unsubscribe: https://mail.python.org/mailman/options/mailman-developers/franck%40peachym… > > Security Policy: http://wiki.list.org/x/QIA9

2 1

Re: [Mailman-Developers] Fixing DMARC problems with .invalid munge
by Franck Martin 17 May '14

17 May '14

You are really mixing everything... Toute connaissance est une réponse à une question. > On May 16, 2014, at 20:09, "John Levine" <johnl(a)taugh.com> wrote: > > In article <1856298671.144791.1400292991012.JavaMail.zimbra(a)peachymango.org> you write: >> The trouble with .invalid is that it is a domain that do not accept emails. Therefore why should you accept emails from a domain >> that does not allow you to reply to it? >> >> It is bound in the future to create issues when people move to more serious/ubiquitous domain reputation schemes. > > Everyone I know who's tried to do spam filtering by SMTP callbacks to > verify sender addresses has stopped, for the dual reasons that it > doesn't work, and it's abusive. I can't imagine why you think people > are likely to resume using it. Considering the amount of mail I get > with non-replyable addresses like donotreply(a)bigbank.com, I'm not the > only person who doesn't think this is a problem. > > Or if they do, it's easy enough to add domain suffixes that satisfy > whatever even more broken reputation scheme we need to defeat. > > R's, > John > _______________________________________________ > Mailman-Developers mailing list > Mailman-Developers(a)python.org > https://mail.python.org/mailman/listinfo/mailman-developers > Mailman FAQ: http://wiki.list.org/x/AgA3 > Searchable Archives: http://www.mail-archive.com/mailman-developers%40python.org/ > Unsubscribe: https://mail.python.org/mailman/options/mailman-developers/franck%40peachym… > > Security Policy: http://wiki.list.org/x/QIA9

1 0

syslog(debug) disable logging in production?
by Sylvain Viart 16 May '14

16 May '14

Hi, I think that syslog() call don't pass through system syslog. I use Mailman/Logging/Syslog.py from Mailman.Logging.Syslog import syslog Is there a way to keep the code without disabling it? In the mailman/Mailman folder (2.1.15) debug tagged syslog calls are commented. Of course I can write some code: DEBUG = False def debug(msg, *args, **kws): if DEBUG == 1: syslog.write_ex('debug', msg, args, kws) def process(mlist, msg, msgdata=None): global DEBUG if hasattr(mlist, 'debug'): DEBUG = mlist.debug […] enabling with: config_list -i <(echo mlist.debug=1) listname I did, in fact… :-) Did I miss some nice feature? Regards, Sylvain.

2 1

Dev version of mailman-bundler
by Abhilash 15 May '14

15 May '14

Hi All, Right now I mailman-bundler uses the PyPI packages for Mailman3, Postorius, HK, KittyStore and MailmanClient to fetch and deploy the entire mailman suite. I think it would be nice to have a version of it where latest code is pulled from the respective sources. I will try to work on it it but, I just wanted to know if it will be used at all by anyone else? -- thanks, Abhilash

2 1

modifying payload using same encoding and Content-Transfer-Encoding
by Sylvain Viart 15 May '14

15 May '14

Hi, I don't manage to figure out how to keep the same Content-Transfer-Encoding, while rewriting the content? msg.set_payload() doesn't seems to trigger any encoding, base64 for example was set. not using decode=True, return a base64 bloc… I just want to write it back. I should be able to know that the payload was encoded… related = MIMEMultipart('related') html_footer = HTML_ATTACHMENT_HOLDER % {'HTML_HERE': html_attach} html_footer += '</body>' old_content = msg.get_payload(decode=True) new_content = re.sub(r'</body>', html_footer, old_content) if old_content != new_content: syslog('debug', 'add html footer') else: syslog('debug', 'no html footer added') charset = msg.get_content_charset() syslog('debug', 'get_content_charset: %s, Content-Transfer-Encoding: %s', charset, msg['Content-Transfer-Encoding']) msg.set_payload(new_content, charset) related.attach(msg) related.attach(clip_payload) The output produced must be wrong: Content-Transfer-Encoding: base64 MIME-Version: 1.0 Content-Type: text/html; charset="utf-8" <html> <head> <meta content="text/html; charset=UTF-8" http-equiv="Content-Type"> </head> <body bgcolor="#FFFFFF" text="#000000"> sysloged output: May 12 17:45:32 2014 (1795) get_content_charset: utf-8, Content-Transfer-Encoding: base64

2 4

Re: [Mailman-Developers] Mailman-Suite demo server
by Abhilash Raj 08 May '14

08 May '14

On Thu, May 8, 2014 at 2:25 PM, Sylvain Viart <sylvain(a)opensource-expert.com > wrote: > Hi Abhilash, > > > On 08/05/2014 10:33, Abhilash Raj wrote: > >> the mm3 web-ui) is a django app which uses mozzila persona for user >> authentication purposes. ( Google and >> other links were not working the last time I checked ). >> > > Ok. Interesting. > > > >> Well theoretically it is possible to add a separate authentication module >> to >> postorius and hyperkitty as they both use the same django project while >> deploying >> using mailman-bundler. But AFAIK it is not among one of the todos. (You >> will find >> a long discussion on user authentication if you dig though mm-dev >> archives) >> > > A bit surprising to me. If I understand what you said correctly, you mean > that mailman standalone (local) auth for mailing list subscription is, for > now, delegated to some foreign generic auth module ? Here mozilla stuff. > > So I can't subscribe to you test list, without sharing that with mozilla. > > I'm right? > > Yes, authentication is delegated to mozzila persona as of now. Funny… And scary. :-) > > Thanks for your reply and setting up a demo. > The UI is a Sooo great improvement of the old good, but boring, mailman > cgi. :-P > > Sylvain. > Yes it is a great improvement I agree. If you are so worried about sharing your email with mozzila I can create an account for you and you can see the after-login part in postorius and HK? thanks, Abhilash

1 0