Re: Header order [was: An analysis of mustUnderstand and related issues]

* Jean-Jacques Moreau <moreau@crf.canon.fr> [2001-05-17 09:37+0200]
> Since then, I have been thinking about the following scenario, and I am
> wondering how it fits with your proposal. Let's suppose a given intermediary
> processes one header block (A), reinserts a copy of that block (A') targeted
> at the Next actor (which is not the ultimate destination), and adds two new
> header blocks (B and C) targeted at the ultimate destination.
> 
> Where should the different blocks be inserted within the message? Should A',
> B and C be inserted where A used to be (this would mean a simpler and more
> efficient XMLP processor)? Or should B and C be inserted *instead* at the
> end of the header section (which would mean editing the message in two
> different places, plus adding the necessary intelligence to the processor to
> do this smart editing)?

I believe that intermediaries should be able to insert blocks anywhere
they want. If order is important, then B and/or C might need to be
some place other than after A' or at the end, e.g.:

  Received block order:			A E F G
  Message passed to the next hop:	B A' E F C G

-- 
Hugo Haas - W3C
mailto:hugo@w3.org - http://www.w3.org/People/Hugo/ - tel:+1-617-452-2092

Received on Thursday, 17 May 2001 11:26:02 UTC