Non-ASCII characters in Subject lines

Non-ASCII characters in Subject lines

[Eli Zaretskii]

Is it allowed by the RFCs to have non-ASCII characters in the Subject
line of an email message?

Rmail currently doesn't decode this part, but maybe it should.

Re: Non-ASCII characters in Subject lines

[Sven Joachim]

On 2009-04-25 09:29 +0200, Eli Zaretskii wrote:

> Is it allowed by the RFCs to have non-ASCII characters in the Subject
> line of an email message?

If it is properly encoded as described in RFC 2047, yes.

> Rmail currently doesn't decode this part, but maybe it should.

Gnus does that, see lisp/gnus/rfc2047.el for implementation details.

Sven

Re: Non-ASCII characters in Subject lines

[Eli Zaretskii]

> From: Sven Joachim <svenjoac@gmx.de>
> Cc: emacs-devel@gnu.org
> Date: Sat, 25 Apr 2009 10:01:16 +0200
> 
> On 2009-04-25 09:29 +0200, Eli Zaretskii wrote:
> 
> > Is it allowed by the RFCs to have non-ASCII characters in the Subject
> > line of an email message?
> 
> If it is properly encoded as described in RFC 2047, yes.

No, I'm talking about raw encoded non-ASCII characters, such as the
ones in this Subject line:

  Subject: Forma\347\343o em Hipnose Cl\355nica - LISBOA - \332LTIMAS VAGAS

> > Rmail currently doesn't decode this part, but maybe it should.
> 
> Gnus does that, see lisp/gnus/rfc2047.el for implementation details.

Thanks for the pointer.  But what does Gnus do with Subject lines such
as the one above?

Re: Non-ASCII characters in Subject lines

[Sven Joachim]

On 2009-04-25 12:16 +0200, Eli Zaretskii wrote:

>> From: Sven Joachim <svenjoac@gmx.de>
>> Cc: emacs-devel@gnu.org
>> Date: Sat, 25 Apr 2009 10:01:16 +0200
>> 
>> On 2009-04-25 09:29 +0200, Eli Zaretskii wrote:
>> 
>> > Is it allowed by the RFCs to have non-ASCII characters in the Subject
>> > line of an email message?
>> 
>> If it is properly encoded as described in RFC 2047, yes.
>
> No, I'm talking about raw encoded non-ASCII characters, such as the
> ones in this Subject line:
>
>   Subject: Forma\347\343o em Hipnose Cl\355nica - LISBOA - \332LTIMAS VAGAS
>
>> > Rmail currently doesn't decode this part, but maybe it should.
>> 
>> Gnus does that, see lisp/gnus/rfc2047.el for implementation details.
>
> Thanks for the pointer.  But what does Gnus do with Subject lines such
> as the one above?

It decodes them as well.

Sven

Re: Non-ASCII characters in Subject lines

[Stefan Monnier]

> Is it allowed by the RFCs to have non-ASCII characters in the Subject
> line of an email message?

You mean to have bytes >127 ?  No, it's not allowed in any part of
the header.  Actually, I'm not sure it's "disallowed", but rather its
meaning falls out of the scope of the RFC.

> Rmail currently doesn't decode this part, but maybe it should.

Such headers appear occasionally, and IIUC Gnus decodes them with the
some default coding system, but I don't think it's high priority.


        Stefan

Re: Non-ASCII characters in Subject lines

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: emacs-devel@gnu.org
> Date: Sat, 25 Apr 2009 12:28:42 -0400
> 
> > Is it allowed by the RFCs to have non-ASCII characters in the Subject
> > line of an email message?
> 
> You mean to have bytes >127 ?

Yes.

> > Rmail currently doesn't decode this part, but maybe it should.
> 
> Such headers appear occasionally, and IIUC Gnus decodes them with the
> some default coding system, but I don't think it's high priority.

OK, thanks.

Non-ASCII characters in Subject lines

[Stephen J. Turnbull]

Eli Zaretskii writes:

 > Is it allowed by the RFCs to have non-ASCII characters in the Subject
 > line of an email message?

The wire format of a header must contain no octets outside of the
range 0-127 inclusive.  More recent versions of the standard (ie, RFC
5322) restrict this further, disallowing most ASCII control
characters.  (IIRC, only TAB, CR, and LF are allowed by RFC 5322.)

 > Rmail currently doesn't decode this part, but maybe it should.

That doesn't make sense; there's nothing decodable there, it's raw
text.  Do you mean that Rmail should try to display a character glyph
rather than an octal escape or the like?

Trying to display should be optional (at least in theory, some
terminals may use some of those characters as control characters).
But it should be available, I guess (Postel Principle).

Note that in my experience almost all such headers are sent by
spammers, and the rest by programming wannabes.  Are you sure you
care?

Re: Non-ASCII characters in Subject lines

[Eli Zaretskii]

> From: "Stephen J. Turnbull" <stephen@xemacs.org>
> Cc: emacs-devel@gnu.org
> Date: Sun, 26 Apr 2009 23:32:26 +0900
> 
> Do you mean that Rmail should try to display a character glyph
> rather than an octal escape or the like?

Yes, that's what I meant.

> Are you sure you care?

I wasn't, that's why I asked.