[rt-users] Character encoding problems with RT 3.0.7_1

[Jari Lehtonen]

Hello,

we've recently implemented Request Tracker using the following components:

- FreeBSD 5.1-RELEASE
- Perl 5.8.2 (from ports)
- Apache 1.3.29 (from port apache+mod_ssl-1.3.29+2.8.16=
- mod_perl (from port mod_perl-1.28)

and everything works quite well, even the superuser is relatively happy, 
with the exception of one problem: scandinavian characters are all garbled 
up in SOME tickets sent via email. I've done some tests and came to a 
conclusion that @EmailInputEncodings is case sensitive; that is, it thinks 
that "iso-8859-1" and "ISO-8859-1" are different character sets. Is this 
correct?

Anyway, after inserting ISO-8859-1 into that array emails sent from pine 
(configured to report its character set as ISO-8859-1) come through without 
problems.

Another issue is in the notification emails RT sends back to requestors; 
namely that subject lines containing 8-bit characters are sent base64 
encoded. Is this intentional, or a bug?

Example:

  Subject:
 =?ISO-8859-1?B?W3Rlc3QudXR1LmZpICMyNF0gTHn2ZOTkbiB2aWVs5CBsaXPk5CBw9mtr9uQ
gcGVzIOTkbiAodGVzdGkp?=

When it actually should read something like:

  Subject: =?ISO-8859-1?Q?Ly=F6d=E4=E4n_viel=E4_lis=E4=E4_p=F6kk=F6=E4_pes?=
 =?ISO-8859-1?Q?=E4=E4n_=28testi=29?=

Note: these subjects are not exactly the same even when decoded. They are 
here just to point out the difference in coding used.

-- 
        Jari Lehtonen
        Unix & Network services
        Computing Center
        University of Turku, Finland