W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1997

Re: chunked enc. (Was: Re: draft minutes, HTTP-WG meetings April 7)

From: David W. Morris <dwm@xpasc.com>
Date: Fri, 11 Apr 1997 14:50:44 -0700 (PDT)
To: Dave Kristol <dmk@bell-labs.com>
Cc: http-wg@cuckoo.hpl.hp.com
Message-Id: <Pine.SOL.3.95.970411144735.29822E-100000@shell1.aimnet.com>


On Fri, 11 Apr 1997, Dave Kristol wrote:

> > In the discussion, the following issues seemed to have sufficient
> > consensus in the meeting that "last call" will be sent to the
> > mailing list for each of them:
> >    "chunked encoding" clarification
> 
> The clarification was that leading zeroes would be allowed in the hexadecimal
> byte count.
> 
> As I thought about this a little, I realized some additional tweaking would
> be required in the draft, both in the syntax and the description.  In
> particular, the syntax probably becomes
> 	chunk-size	= 1*HEX
> but the description will need a caveat that the value of chunk-size must be
> non-zero, except for the last chunk.  Should the last chunk always be denoted
> by "0" CRLF, or by 1*HEX CRLF, where the value of 1*HEX value is zero?  If
> the former, are we comfortable with zero-length, non-final chunks?  There
> could be the ambiguous case of chunk-size being, say "00".

Well, I think the end flag should simply be a length value of ZERO. I
don't see any percentage in making the specification more complex so 
why not treat a length of 0 or 00 or 000 etc. as the end indicator. 

Dave Morris
Received on Saturday, 12 April 1997 05:18:12 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:32:34 EDT