dovecot - Jun 2007 - CMU Sieve pluigin missing proper charset support

Hello,

After testing the Sieve plugin for the LDA, I have noticed that it does
not handle charset issues properly. IETF draft says:

4.3.  Subject and From Parameters

   The ":subject" parameter specifies a subject line to attach to any
   vacation response that is generated.  UTF-8 characters can be used in
   the string argument; implementations MUST convert the string to
   [RFC2047] encoded words if and only if non-ASCII characters are
   present.  Implementations MUST generate an appropriate default
   subject line as specified below if no :subject parameter is
   specified.

(see http://tools.ietf.org/html/draft-ietf-sieve-vacation-07)

The problem is that the dovecot plugin does not convert the string to
RFC2047 encoded words, it treats it literally and thus sends malformed
(from RFC's point of view) message - with improper Subject field
encoding.

Any chances of getting that fixed?

greetings,
-- 
Jan Srzednicki
w at expro.pl

On Fri, 2007-06-01 at 18:49 +0200, Jan Srzednicki wrote:
> The problem is that the dovecot plugin does not convert the string to
> RFC2047 encoded words, it treats it literally and thus sends malformed
> (from RFC's point of view) message - with improper Subject field
> encoding.
> 
> Any chances of getting that fixed?
Well.. Hmm. I'm not entirely sure how Cyrus handled it, or if it didn't
handle it at all. The Sieve code is from Cyrus v2.2. In Cyrus v2.3
there's some new MIME code. Perhaps upgrading the Sieve plugin to Cyrus
v2.3's code would fix this as well.

Also the APIs that need to be used to decode them are a bit annoying to
use. I think I'll simplify them a bit for v1.1 so it won't be necessary
to copy&paste 20 lines of code to use it.

I added this to my TODO, but if you're interested in fixing it now, I
can give some pointers:

 - dovecot-sieve-1.0/src/sieve-cmu.c : getheader() should decode them
 - dovecot-1.0/src/lib-mail/message-header-decode.h API decodes the
RFC2047 MIME words but doesn't convert to UTF8
 - dovecot-1.0/src/lib-charset/charset-utf8.h API provides charset
conversion.
 - dovecot-1.0/src/lib-mail/message-header-search.c uses the above two
APIs

For v1.1 I'll add a new function to message-header-decode.h that returns
a given header directly as UTF-8 string.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL:
<http://dovecot.org/pipermail/dovecot/attachments/20070604/7fe983e7/attachment-0002.bin>