Re: Warn header (P6) vs RFC2047 encoding

tor 2009-07-23 klockan 12:22 +0200 skrev Julian Reschke:

> I think this should be dropped the same way we already did it for the 
> Reason Phrase (see <http://trac.tools.ietf.org/wg/httpbis/trac/ticket/94>).

Reason Phrase is different as this generally is not intended to be
displayed in user agents other than for debugging/tracing/status
indications, while Warning is really intended to be displayed to the
user with information the user is supposed to read and understand.

So Warning SHOULD support localized text outside English, or it will
have an even harder time to get accepted than it already has..

What I remember from the prior discussion was to remove iso-8859-1 and
2047 from the general header definition to open up for new headers to be
specified using UTF-8 if they like, not that we should drop 2047 support
in existing headers defined as having human oriented text (Warning,
Auth*, maybe more). But it was long ago...

However I think we maybe want to revisit the general header syntax
description, making sure we do provide the intended guidance for those
who want to write extension headers requiring natural language content..
current wording kind of suggest such headers is not allowed.

Regards
Henrik

Received on Thursday, 23 July 2009 19:30:47 UTC