[sylpheed:32881] Re: unicode-1-1-utf-7 question

Hiroyuki Yamamoto hiro-y at kcn.ne.jp
Wed Mar 4 11:30:15 JST 2009


On Wed, 4 Mar 2009 00:13:36 +0100
Marc Finet <m.dreadlock at gmail.com> wrote:

> Hello,
> I'm a sylpheed user for a little while and just discovered a mail
> that was treated strangely by sylpheed (see attached 29 file), i.e.
> not rendering utf7 (i didn't check it was correct utf-7) like '+AOk-'
> as i expected 'é'. Well i'm not used to encodings so it might not
> appear good, it's just the accented 'e'.
> The problem seems to come from unicode-1-1-utf7 encoding format, as
> seen in file/mail. It appears "inline" as =?unicode-1-1-utf-7?... or
> in charset=unicode-1-1-utf-7. I'm definitely not familiar with mail
> RFCs and norms and what i quickly saw in sylpheed's code (2.6.0 on
> gentoo) told me that UTF_7 seems to be defined (CS_UTF_7) so i guessed
> supported.
> Now my questions are:
>  - is this mail valid ? (RFC regarding such encoding as google told
> me is rfc1642) ? otherwise i will 'complain' about sender (I didn't
> check yet sender user-agent) and none of my following question have
> sense...
>  - is sylpheed compliant with this RFC ?
>  - is sylpheed compliant with utf-7 ?
>  - what could be done (i could give some help; i'm just a
> C-developer but interested in lending a hand, given a few hints on
> work to be done, i.e. utf-7 support first, then unicode-1-1-utf7 as
> an alias or something like that...) ?

UNICODE-1-1-UTF-7 is defined in RFC 1642, but the RFC is obsoleted by
RFC 2152. In RFC 2152, it is told to use the charset tag "UTF-7".
In conclusion, the name "unicode-1-1-utf7" should not be used.

Sylpheed can handle UTF-7 encoding, but it depends on the iconv() API
provided by glibc or libiconv. Libiconv supports the name
"UNICODE-1-1-UTF-7", but glibc doesn't.

If UNICODE-1-1-UTF-7 is already widely used, I can make a workaround in

Hiroyuki Yamamoto <hiro-y at kcn.ne.jp>

More information about the Sylpheed mailing list