[Mailman-Users] [Fwd: TypeError: us-ascii with python2.4 and mailman 2.1.8-1 (debian)]

[Tokio Kikuchi]

Hi developers,

This particular problem is caused by a bug in email 4.0.1 package which 
was fixed in the most recent subversion repository.
http://svn.python.org/view/python/trunk/Lib/email/message.py?rev=54333&r1=50840&r2=54333

Maybe it's time to think of next bug fix release of mailman 2.1.10 as 
soon as the email 4.0.2(?) is out.



Justin Warren wrote:
> Hi folks,
> 
> I hit the same sort of problem as a couple of other people apparently
> have, eg:
> 
> http://mail.python.org/pipermail/mailman-users/2006-November/054246.html
> 
> I hit this when Mailman attempts to send digests.
> 
> I thought I'd tracked the issue to the way Charset is used within
> Mailman/Scrubber.py, on line 193 (some modifications by me to print
> debug output):
> 
>     charset = None                                                                                                     
>     lcset = Utils.GetCharSet(mlist.preferred_language)                                                                 
>     print "getting charset for preferred language:", repr(lcset)                                                       
>     lcset_out = Charset(lcset).output_charset or lcset                                                                 
>     print "lcset_out is:", repr(lcset_out) 
> 
> Which yields:
> 
> getting charset for preferred language: 'us-ascii'
> lcset_out is: u'us-ascii'
> 
> It appears that Charset returns a unicode string for
> its .output_charset, which python2.4 email.Message.set_charset() does
> not recognise as a string, and it isn't a Charset either, so a TypeError
> is raised.
> 
> I don't know enough about why .output_charset is being used, or why it
> is unicode, but it seems that changing line 193 to read:
> 
>     lcset_out = Charset(lcset) or lcset
> 
> would prevent this error.
> 
> However, something else is also returning a unicode string, since after
> changing Scrubber, you end up with this error:
> 
> Traceback (most recent call last):
>   File "/usr/lib/mailman/cron/senddigests", line 94, in ?
>     main()
>   File "/usr/lib/mailman/cron/senddigests", line 86, in main
>     mlist.send_digest_now()
>   File "/var/lib/mailman/Mailman/Digester.py", line 60, in send_digest_now
>     ToDigest.send_digests(self, mboxfp)
>   File "/var/lib/mailman/Mailman/Handlers/ToDigest.py", line 142, in send_digests
>     send_i18n_digests(mlist, mboxfp)
>   File "/var/lib/mailman/Mailman/Handlers/ToDigest.py", line 339, in send_i18n_digests
>     mcset = msg.get_content_charset('')
>   File "/usr/lib/python2.4/email/Message.py", line 805, in get_content_charset
>     charset = unicode(charset, 'us-ascii').encode('us-ascii')
> TypeError: decoding Unicode is not supported
> 
> I haven't worked out what this other thing is as yet.
> 
> Alternately, patching email/Message.py (line 803) to ignore a charset
> that is a unicode string seems to work, but is probably fragile or
> broken in other ways:
> 
>         # charset character must be in us-ascii range                                                                  
>         try:                                                                                                           
>             if not isinstance(charset, unicode):                                                                       
>                 charset = unicode(charset, 'us-ascii').encode('us-ascii')                                              
>         except UnicodeError:                                                                                           
>             return failobj 
> 
> I hope this helps someone more knowledgable about Mailman to work out
> what the underlying problem is.
> 


-- 
Tokio Kikuchi, tkikuchi at is.kochi-u.ac.jp
http://weather.is.kochi-u.ac.jp/