Re: Call for Consensus: Remove "reference set" from HPACK (to address #552)

On Wed, Jul 16, 2014 at 06:45:26AM +0000, Poul-Henning Kamp wrote:
> In message <20140716054520.GC14304@1wt.eu>, Willy Tarreau writes:
> >On Wed, Jul 16, 2014 at 01:52:35PM +1000, Greg Wilkins wrote:
> 
> >The problem I'm seeing with this change is that you need one extra-byte to
> >send litteral headers fields that are not present in the static table. Maybe
> >on the other hand it rarely happens that they're not in the static table ?
> 
> Sounds like a good way to persuade people to not create new headers all
> the time :-)

Good point :-)

> Which reminds me:
> 
> Should we fold "X-Forwarded-For" to "Forwarded-For" as part of the
> compression ?

No we must not do that, because X-whatever as its name implies is not
standardized and can be used any other way. Also there are a very large
number of systems around which make use of it and which will not yet be
able to parse Forwarded, which has a different representation.

Willy

Received on Wednesday, 16 July 2014 09:26:23 UTC