[horde] problem with accents

[Leena Heino]

> I do have the same problem with umlauts. (See my post "[horde] Umlaut
> Problem") It seems to me, that somehow the old horde version was able
> to switch charsets encoding of the webspages according to the
> (default or user) language chosen. Horde 3.0.5 seems to lack that
> capability and just uses UTF-8 for all the pages.

It seems that in Imp 3.x the email message was by default shown with 
current locales' charset eg. in finnish with charset iso-8859-1. This 
meant that most of the email messages from most of the western countries 
with broken or missing encoding was shown mostly correctly. This also 
meant that email with utf-8 characters would by default be shown 
containing strange characters.

The deeper problem lies in the fact that some email programs and 
organizations sents email without any information of the message's 
character encoding and have unencoded and unmarked 8-bit characters in 
headers. These messages violate the RFC 2822 standard.

If messages with these unencoded and unmarked 8-bit characters end up in 
your mail store they will be shown as is in Imp 4.0.x. Imp 4.0.x sents 
html-pages that are marked as being encoded in utf-8, but if your mail 
store have these unencoded and unmarked 8-bit characters then the 
resulting html-pages will have these characters as is and in most of the 
cases these characters are invalid utf-8 character sequences. These 
invalid utf-8 characater sequences might cause some very unexpected 
results in different web browsers. IE6 shows a blank page or a page that 
contains unexpected characters, but you could still read the message with 
IE's view page source. Firefox just shows either ?-character or diamond 
shape character in place of these invalid utf-8 character sequences.

To correct this situation you could refuse to accept email with unecoded 
and unmarked 8-bit characters. You could modify the messages by dropping 
the 8th bit from the unencoded and unmarked 8-bit characters or replace 
them with safe 7-bit characters. Or you could modify these messages by 
adding the bare minimum necessary MIME-headers and if needed encode 
message headers.

-- 
   Leena Heino              University of Tampere / Computer Centre
   ( liinu at uta.fi )      ( http://www.uta.fi/laitokset/tkk )