Problems with national characters

[mla]

There seems to be some kind of problem with national characters, e.g. the Swedish 'Å', 'Ä', 'Ö'.

When a compose a mail in Zimbra containing these characters, the get mangled at the receiving end (outside of Zimbra).

An example:

> Kolla på http://www.zimbra.com
>
> Verkar ganska lovande, fast ÅÄÖ kanske blir fel?
>
> mla
> ---
> Mats Löfström

It looks like ÅÄÖ is exchanged for two bytes characters...

Digging further in, this is what I find (when viewing a Zimbra mail in vi):

Å => \303\205
Ä => \303\204
Ö => \303\226
å => \303\245
ä => \303\244
ö => \303\266

This is what they _should_ be:

Å => \305
Ä => \304
Ö => \326
å => \345
ä => \344
ö => \366

[KevinH]

I just tried sending from gmail to zimbra and then Zimbra to gmail with ÅÄÖ and it worked fine. Can you explain your exact use case?

[dkarp]

Originally Posted by mla

There seems to be some kind of problem with national characters, e.g. the Swedish 'Å', 'Ä', 'Ö'.
It looks like ÅÄÖ is exchanged for two bytes characters...

If you look at the headers for the Zimbra-generated message, you should see a line that looks like this:

Content-Type: text/plain; charset=utf-8

The characters in the message body are in the UTF-8 character set, not in ISO-8859-1.

You can't change the default character set at present. We'll add that to our feature wishlist...

[mla]

Originally Posted by KevinH

I just tried sending from gmail to zimbra and then Zimbra to gmail with ÅÄÖ and it worked fine. Can you explain your exact use case?

From my Zimbra installation to a Solaris based host with sendmail, mail read and replied in mutt, back to Zimbra host.

Reading and replying with Horde gives the same result.

[Noise]

You can't change the default character set at present. We'll add that to our feature wishlist...

Do that. It should be one of your priorities, as there is not just american/english language in the world.

[dkarp]

Originally Posted by Noise

Do that. It should be one of your priorities, as there is not just american/english language in the world.

Actually, we send messages in UTF-8, a charset that encodes Unicode and hence can represent pretty much every character in every language. Switching to a smaller charset (e.g. iso-8859-1) will actually make the system less portable, not more. Modern MUAs don't have any problem with UTF-8, but according to the up-thread posters there are enough old mailers that sysadmins may need to use a small, local charset to make certain users happy.