canonstr -> unicode usernames

Bug #265861 reported by Donnc
2
Affects Status Importance Assigned to Milestone
GNU Mailman
Fix Released
Medium
Unassigned

Bug Description

Cgi.subscribe.process_form gets a unicode value for fullname,
from Utils.canonstr() (return unicode(newstr, charset,
'replace')

This came to my attention because it broke OwnerNotification
when the email Message couldn't "flatten" itself with the
unicode payload. Suspect that's only the beginning of the
trouble that will be caused, so I'm hoping we really don't
want unicode usernames values as a rule. I fixed it in
subscribe.py with fullname = fullname.encode('iso8859-1'), but
note that canonstr is used elsewhere like this.

[http://sourceforge.net/tracker/index.php?func=detail&aid=672649&group_id=103&atid=100103]

Revision history for this message
Pcarr-users (pcarr-users) wrote :

I stumbled over the same bug, but in my case it was bailing
only in the confirm.py script.

Revision history for this message
Tokio Kikuchi (tkikuchi) wrote :

If your users choose to input latin-1 charcters in fullname,
it will be better to set iso-8859-1as english charset.

in mm_cfg.py,

def _(s):
    return s
LC_DESCRIPTIONS['en'] = (_('English (USA)'), 'iso-8859-1')
del _

Revision history for this message
Barry Warsaw (barry) wrote :

I think this bug has been fixed in 2.1.1 or 2.1.2 (cvs).

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.