W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2008

Re: i94, was: Does Reason-Phrase allow LWS?

From: Frank Ellermann <nobody@xyzzy.claranet.de>
Date: Thu, 3 Apr 2008 21:29:06 +0200
To: ietf-http-wg@w3.org
Message-ID: <ft3b1n$p7i$1@ger.gmane.org>

Julian Reschke wrote:

>>> There's an overlap with issue 74

>> I don't understand section 5 in RFC 3987.  Are HTTP
>> implementors forced to grok IRI comparison ?  What
>> has this to do with I18N for <Reason-Phrase> ?
 
> That's what I'm asking you :-)

LOL.  Okay, then I'd say <Reason-Phrase> used to be a
string of Latin-1 octets carefully avoiding CR and LF.
There could be a Latin-1 IRI in <Reason-Phrase>, but
Latin-1 IRIs are not sexy, unlike KOI8-R IRIs. ;-)

In other words, as long as HTTP headers can't do "raw"
UTF-8 we can ignore IRIs.  Folks wanting them anyway 
find the recipe to transform any IRI into an URI in
RFC 3987.

As soon as it's an URI it's US-ASCII and we can forget
Latin-1 wrt IRIs.  In a <Reason-Phrase> or elsewhere.

Using IRIs or URIs within an in essence unstructured
part of a header field has similar problems as using
IRIs or URIs in text/plain mail or news bodies:  Put
them in angle brackets, maybe use <URL: .. >, or not.

In the case of a header field they could end up in an
2047/2231 <encoded-word>, a seriously bad idea.  But
hopefully not bad enough to talk about it in 2616bis.
I've never seen a 2047/2231 encoded IRI.  

>> Sanity check, we don't want folding there, right ?
> I don't think so. That's what i94 is about.

Good.
Received on Thursday, 3 April 2008 19:27:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:46 GMT