W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2007

Re: Thought on ABNF upgrade

From: Henrik Nordstrom <henrik@henriknordstrom.net>
Date: Tue, 20 Mar 2007 11:57:22 +0100
To: Julian Reschke <julian.reschke@gmx.de>
Cc: HTTP Working Group <ietf-http-wg@w3.org>
Message-Id: <1174388242.12435.34.camel@henriknordstrom.net>
lör 2007-03-17 klockan 09:49 +0100 skrev Julian Reschke:

> 1) I think there is agreement that the ABNF syntax needs to be upgraded 
> to the syntax defined in RFC4234.

Yes.

> 3) Related to that, RFC2616 defines special treatment of Linear White 
> Space (LWS), see 
> <http://greenbytes.de/tech/webdav/rfc2616.html#rfc.section.2.1.p.11>). 
> It's up for discussion whether we want to keep that aspect. See 
> <http://lists.w3.org/Archives/Public/ietf-http-wg/2007JanMar/0116.html> 
> for some comments from Roy Fielding about that subject.

I would propose start with an extended form with "list" syntax and
keeping LWS while migrating. This to mimimize the changes making it
easier to follow.

Then look into first if it makes sense dropping the "list" extension or
not, and finally if it's possible to get rid of the implied LWS rule
while keeping the grammar readable. Should be possible.

> The next step would be to run this through a parser. Here's where we 
> have the first problems:

Not entirely sure the existing BNF is in entirely shape for being parsed
without substantial edits. But it worthwhile trying.

>     qdtext         = <any TEXT except <">>
> 
> which IMHO needs to be fixed not to contain the literals "<" and ">". I 
> would like to make that change with the next draft we submit.

In the RFC2616 BNF <"> is the token used for a literal "  (2.2). Easy to
rename if it makes parsing easier, or should not be too hard to fix the
parser to handle it. Used in a number of rules not only qdtext.

Regards
Henrik

Received on Tuesday, 20 March 2007 10:57:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:00 GMT