[SM-USERS] Problems with encoding of e-mail addresses

[Rafael M. G. <r.m...@us...>]

Hello

We are having some problems with how squirrelmail process some e-mail
addresses when they are encoded in this way:

=?US-ASCII?Q?Martinez Guerrero=2c_Rafael?= <r.m...@us...>

This line is the samme as this one:
"martinez guerrero, rafael" <r.m...@us...>

When squirrelmail 'replay to' one of these encoded addresses it
translates them to:

"martinez" guerrero, "rafael" <r.m...@us...>

As you can see, two addresses. The first one with the e-mail part like
'guerrero'. In our system, users can send emails to other users at the
university only using the username (without the @domain part). 

If you are unlucky, a user is replaying to the address in the example
and 'guerrero' exist in the system as a valid username, the user will be
sending two e-mails, one to the rigth address (r.m...@us...)
and the second to a user with a username like 'guerrero' (that in most
cases they do not know)

Can we do anything to change this behavior?. Suggestions?. Other e-mail
clients translate this encoded addresses in the rigth way.

PD.- We can ask users to verify the TO header before they send an e-
mail. But this is real problem, users do not verify the TO header when
they are replaying to an e-mail.

Thanks in advance
Regards
-- 
Rafael Martinez, <r.m...@us...>
Center for Information Technology Services
University of Oslo, Norway

PGP Public Key: http://folk.uio.no/rafael/