[Dovecot] UTF-8 in dovecot-sieve ?

[Arthur Pétry]

Hi,

Reading http://tools.ietf.org/html/rfc3028#section-2.7.2 I thought  
that out-of-the-box dovecot can filter messages by matching header in  
UTF-8.

But when I tested it, it failed :
(I'm using 1.0.5 with sieve plugin)

Here is my .dovecot.sieve file :

----------
require ["fileinto", "imapflags"];

if allof(header :contains "subject" "filtre") {
        addflag "test_1";
}

if allof(header :contains "subject" "là") {
        addflag "test_2";
}

if allof(header :comparator "i;ascii-casemap" :contains "subject"  
"hé") {
        addflag "test_3";
}

----------

Only the first of the three rules do work.

When I try to match a subject string with accent encoded in ISO-8859  
like "hélà" : "=?ISO-8859-1?Q?h=E9l=E0?=" there is no match

Am I doing something wrong ?