What steps will reproduce the problem? 1. Have a contact who has specified non-latin nickname on the ICQ server 2. request "User info" for this contact 3. you will be kicked off, with error trace in the log (specify "-d" option)

What is the expected output? What do you see instead?

-

What version of the product are you using? On what operating system?

Version: 0.8a-4 .deb package on Debian etch

Please provide any additional information below.

===== File "/usr/share/jabber-pyicq/src/legacy/icqt.py", line 80, in gotUserInfo self.oscarcon.gotvCard(self.oscarcon.userinfoCollection[id]) File "/usr/share/jabber-pyicq/src/legacy/glue.py", line 497, in gotvCard self.legacyList.updateNickname(usercol.userinfo, usercol.nick) File "/usr/share/jabber-pyicq/src/legacy/buddies.py", line 170, in updateNickname c.updateNickname(nick, push=True) File "/usr/share/jabber-pyicq/src/contact.py", line 145, in updateNickname if unicode(self.nickname, errors='replace') != "nickname":

exceptions.TypeError: decoding Unicode is not supported

Apparently when Contact.updateNickname is called, the argument has been already converted to UTF-8. When a unicode object is passed to "unicode" function, Python raises exception even if you specify "errors='replace'". This can be verified by trying to execute "unicode(u'',errors='replace')"

The following diff helps, though I don't know how "correct" it is"

--- src/contact.py.orig 2008-01-23 16:56:23.000000000 +0300 +++ src/contact.py 2008-01-23 17:03:02.000000000 +0300 @@ -142,7 +142,7 @@ self.contactList.legacyList.deauthContact(self.jid)

    def updateNickname(self, nickname, push=True):

- if unicode(self.nickname, errors='replace') != "nickname": + if self.nickname != "nickname": self.nickname = nickname # will re-remove this if it's removed from JEP-0172. #self.sendNickname()

It seems fixed in nearly released 0.8b

Yes - this is fixed in 0.8b (SVN changeset 230).