Patches item #850805, was opened at 2003-11-28 16:23 Message generated for change (Comment added) made by tkikuchi You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=300103&aid=850805&group_id=103 Category: Pipermail

Group: Mailman 2.1 Status: Open Resolution: Postponed Priority: 5 Submitted By: Richard Barrett (ppsys) Assigned to: Nobody/Anonymous (nobody) Summary: Aggressive anti email address harvesting measure

Initial Comment: This patch is prompted by discussion on the mailman- developers list won the following subject: Re: [Mailman-Developers] bugtraq submission warning: email address harvesting exploit For those with deep concerns about email address harvesting this patch offers a more aggressive masking of email addresses in Mailman mail archive files. The patch modifes two files in the standard Mailman distribution: Mailman/Defaults.py and Mailman/Cgi/ private.py and can be applied using the following command from within the Mailman build directory: path -p1 < path-to-patch-file It would be fairly trivial to make enabling this feature per- list configurable rather than it being a site admin decision and I will enhance this patch for that purpose if people show an interest in it being done. The following notes about the patch can be found in Defaults.py. Rather idiosyncratically most of the operational elements of this patch are in that file. My reasoning behind this decision is that if people want to fool with the regexes that are at the heart of this patch they can see what will be affected by the changes more readily if the related bits are in the same place. ##### # Anti-spam email address harvesting prevention measures. # # These measures are to limit the ability of spam generators to acquire # email address from archived material in Mailman's list archives. # Implementation is via a dynamic search and replace for email # addresses, appearing in files of MIME type text/html or text/plain, as # those files are requested. The underlying archive file content as # generated by the archiving software remains unchanged. # # The implementation requires that archive files are all delivered by a # modified private.py CGI script which only requires user authentication # if the list whose archive material is being requested is set up as a private # list. In order to get public archives served by private.py a RewriteRule # like this: # # RewriteRule ^/pipermail/(.*) /mailman/private/$1 [PT] # # needs to be used in the Apache httpd.conf to transparently redirect # public arechive file requests. # # When email addresses are found, the domain part of the addressed is replaced # with a string of 'x' characters. If the local part of the address appears to # have been VERP'ed then the VERP information is similarly obscured. This is # a fairly brutal set of irreversible modifications to any email addresses in # the returned text and will break any mailto: links in the text. # # Th eamil address regex looks for either an '@' character or its HTML escaped # version '%40' as the local-part/domain separator. You should set # ARCHIVER_OBSCURES_EMAILADDRS = 0 and run bin/arch to rebuild existing archives # to prevent that feature interfering with the operation of these harvesting # prevention measures. # # If you decide to change the regexes then copy all of this stuff into # mm_cfg.py and make the changes there. # ##### ----------------------------------------------------------------------

Comment By: Tokio Kikuchi (tkikuchi) Date: 2004-11-14 00:28

Message: Logged In: YES user_id=67709 Sorry, patch itself is for 2.1 (not 3.0). ---------------------------------------------------------------------- Comment By: Tokio Kikuchi (tkikuchi) Date: 2004-11-14 00:27 Message: Logged In: YES user_id=67709 It looks nice but it may be too much for 2.1.x where the archives are generated static. Also, I'm a little bit anxious writing such a function in Defaults.py. It would be better to include this feature in 3.0 or later. ---------------------------------------------------------------------- Comment By: Martin Mokrejs (mmokrejs) Date: 2004-04-20 09:52 Message: Logged In: YES user_id=696559 Ohh, sorry, the patch really wasn't applied. After proper installation, I can confirm it works fine. ---------------------------------------------------------------------- Comment By: Martin Mokrejs (mmokrejs) Date: 2004-04-20 00:00 Message: Logged In: YES user_id=696559 What sould I do with this error caused by the patch? Traceback (most recent call last): File "/usr/local/mailman/scripts/driver", line 87, in run_main main() File "/usr/local/mailman/Mailman/Cgi/private.py", line 165, in main sys.stdout.write(mm_cfg.deny_harvest(f.read())) AttributeError: 'module' object has no attribute 'deny_harvest' ---------------------------------------------------------------------- Comment By: Richard Barrett (ppsys) Date: 2003-11-30 07:02 Message: Logged In: YES user_id=75166 As pointed out on the mailman-developers list, potential users of this patch should be aware that the simple approach to masking email addresses used by this patch will also capture and munge any other strings in the archive data that resemble email addresses. This can include mailto URLs, other URLs and Message ids. This side-effect may make the patch unsuitable for use with your system, although you also need to consider that the patch does not irreversibly change the source pipermail archived material held on the server; the changes are only made in the copy of the archive material sent to the requesting browser by the server. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=300103&aid=850805&group_id=103