SV: SV: SV: Charset error in subject when using Danish letters (æøå / ÆØÅ)

To summarize: It is only the subject in the outgoing mail from RT that has wrong characters - the mail body looks fine (including Nordic national characters).
Everything inside RT looks fine and the subject is shown correctly in the webui. Even checked the database tables and they’re fine too.
So my guess is, that it has to do with the mail-sender (whereever that is :))

Apache is configured with default charset as UTF-8
AddDefaultCharset UTF-8
A bunch of additional charset is added in the httpd.conf
including latin1.

In RT_SiteConfig.pm following is set:
@EmailInputEncodings = qw(iso-8859-1 us-ascii utf-8)
unless (@EmailInputEncodings);
Set($EmailOutputEncoding , ‘iso-8859-1’);

Does this answer your questions?

Med venlig hilsen / Best regards
Brian Kjelin Olsen
Schilling A/S

-----Oprindelig meddelelse-----Fra: Jesse Vincent [mailto:jesse@bestpractical.com]
Sendt: 18. maj 2007 18:47
Til: Brian Kjelin Olsen
Cc: RT Users
Emne: Re: SV: SV: [rt-users] Charset error in subject when using Danish letters (æøå / ÆØÅ)

No, I haven’t heard of any having the same problem with Danish
letters in the subject only.

If there are other message headers with Danish characters, do they
get similarly mangled? Are the characters typed in as unicode or
latin-1? Does it happen only with one mail client or with many? Are
there clients it doesn’t happen iwth?

Med venlig hilsen / Best regards
Brian Kjelin Olsen
Schilling A/S

-----Oprindelig meddelelse-----
Fra: Jesse Vincent [mailto:jesse@bestpractical.com]
Sendt: 16. maj 2007 18:35
Til: Brian Kjelin Olsen
Cc: RT Users
Emne: Re: SV: [rt-users] Charset error in subject when using Danish
letters (æøå / ÆØÅ)

Hi Jesse

I’m sorry that this reply was delayed, but several other tasks were
impediment for continuing this issue. Nevertheless, it is still an
issue we haven’t bin able to solve even though several people have
suggested changes of various settings.

You asked for the Perl and Mime version and about the database and
webui setup.

  • MimeTools version is: 5.420. The encode version was: 2.01 but is
    upgraded to 2.20
  • Perl version is: v5.8.5 built for i386-linux-thread-multi
  • DB (Oracle) and webui looks alright - no strange conversions…

Everything looks right - except for the subject in e-mails returned
from RT…
There are no problems with Danish letters in the mail body.

Has anyone else seen this issue? (Finding commonalities would help
track it down)

Med venlig hilsen / Best regards
Brian Kjelin Olsen
Schilling A/S

-----Oprindelig meddelelse-----
Fra: Jesse Vincent [mailto:jesse@bestpractical.com]
Sendt: 25. april 2007 15:40
Til: Brian Kjelin Olsen
Cc: RT Users
Emne: Re: [rt-users] Charset error in subject when using Danish
letters (æøå / ÆØÅ)

During our correspondence with our customers by mail through the RT
system, the Danish letters (æøå / ÆØÅ) is getting unreadable when
they are placed in the subject.
Any ideas would be most welcome.

Example: “Subject: Fwd: [support.schilling.dk #120841]
Sÿgefunktion pÃ¥ kunde. ÿBLEGRÿDÿ - ÿblegrÿdÃ¥”
Correct subject: “Subject: Fwd: [support.schilling.dk #120841]
Søgefunktion på kunde. ÆBLEGRØDÅ - æblegrødå”

What Perl version are you running? Are you current on MIME::Tools and
Encode? Is it right in the database and/or webui?

To summarize: It is only the subject in the outgoing mail from RT
that has wrong characters - the mail body looks fine (including
Nordic national characters).
Everything inside RT looks fine and the subject is shown correctly
in the webui. Even checked the database tables and they’re fine too.
So my guess is, that it has to do with the mail-sender (whereever
that is :))

Apache is configured with default charset as UTF-8
AddDefaultCharset UTF-8
A bunch of additional charset is added in the httpd.conf
including latin1.

In RT_SiteConfig.pm following is set:
@EmailInputEncodings = qw(iso-8859-1 us-ascii utf-8)
unless (@EmailInputEncodings);
Set($EmailOutputEncoding , ‘iso-8859-1’);

If you set the output encoding to utf-8 as a test, does it work
correctly?

PGP.sig (186 Bytes)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

To summarize: It is only the subject in the outgoing mail from RT that has wrong characters - the mail body looks fine (including Nordic national characters).
Everything inside RT looks fine and the subject is shown correctly in the webui. Even checked the database tables and they’re fine too.
So my guess is, that it has to do with the mail-sender (whereever that is :))

Apache is configured with default charset as UTF-8
AddDefaultCharset UTF-8
A bunch of additional charset is added in the httpd.conf
including latin1.

In RT_SiteConfig.pm following is set:
@EmailInputEncodings = qw(iso-8859-1 us-ascii utf-8)
unless (@EmailInputEncodings);
Set($EmailOutputEncoding , ‘iso-8859-1’);

Does this answer your questions?

Med venlig hilsen / Best regards
Brian Kjelin Olsen
Schilling A/S

I had problems with subject encoding and 2-byte (japanese) characters.
especially with a mix of single-byte (ASCII) and 2-byte (japanese)
characters in the subject. This problem was specifically noticable with
long subject lines that would need to be split in the headers.

it’s been awhile since i’ve used RT (I changed jobs) so I don’t remember
specifically the situation, but if I remember correctly, RT didn’t
properly MIME encode the subject. either it didn’t properly terminate
and re-initiate the encoding per-line during a split, or if it did,
there were problems with it.

eventually I had to install Jcode.pl and use its encoding routines to
fix the problem for japanese. This is probably related to the problems
you’re seeing, although not likely Jcode.pl will have much effect on non
2-byte characters.

Either way, I suggest looking at the mail source to see the encoded

HTH

Alan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGV6j/E2gsBSKjZHQRAsxTAKDNG1AHcDOiZkMP6i7BV/GJL5fK9gCeNPNv
+PVYB7h0vcOPi7VXr6W+Jlk=
=0qC7
-----END PGP SIGNATURE-----