W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1996

Re: ISSUES LIST: Chunked Encoding

From: <hallam@w3.org>
Date: Fri, 15 Mar 96 14:46:30 -0500
Message-Id: <9603151946.AA32239@zorch.w3.org>
To: hardie@nasa.gov, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Cc: hallam@w3.org
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/9

>Specifically, we have an open question on whether we will support both
>chunked encoding and multi-part MIME for persistent connections.  We
>also have an open question about whether and how to fix the current
>specification of chunked encoding so that it allows the kinds of
>footer information the Phillip Hallam-Baker needs.

No, this is exactly the opposite of what I am saying. Chunked as specified
is OK, it includes footers and there is a minor tweak to the grammar
which I beleive should be made which would hold open interesting developments in 

The problem is not how to "fix" chunked. The problem is how to keep people from 
taking the footers out. The message I put out 8 weeks ago was a result of 
misreading the draft. Roy cleared this up but it does not appear that everyone 
noted this.

>If we agree we will support both chunked encoding (as a MUST) and
>multi-part MIME (perhaps as a SHOULD), can we leave messages which
>require footers to multi-part MIME?  If this is possible, it seems
>that we could retain some of the ease of implementation that the Area
>Directors indicated chunking would buy us, and help us get persistent
>connections deployed more quickly

No, multiparts are intollerable over an 8-bit clean connection. They are not
acceptable as the sole means of incorporating footers.

I suggest we:

1) Leave chunked in as a MUST *including* *footers* (as is currently the case).

2) Make the technical change to the chunk a grammar to allow chunk attributes to 
be incorporated at a later date.

3) Relegate MIME multipart support to SHOULD.
	[If indeed coding time is an issue].

I am not proposing an onerous quantity of code. It is simply requiring 1.1 
parsers to tolerate possible 1.2 extensions. It has no impact on the normal flow 
of control, it is merely a question of what is considered an error.

I really do not think there is any major problem here. I think that it is simply 
a case of miscommunication which is easy enough via e-mail. 

It appears that the louder I tell people that removing footers from chunked 
would be a very stupid mistake the more people suggest this course of action as 
a means of obtaining consensus!

If people want consensus and want chunked in 1.1 then can we please have a 
discussion that addresses the *technical* issues. 

Received on Friday, 15 March 1996 18:07:25 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:40:16 UTC