How about replacing all ','s with '\0's? '\0' is a disallowed character everywhere, unlike ',' which requires special parsing because of quoted strings. I'd rather that the lower layers of the protocol stack not have to do quoted string parsing. That belongs to the part of the stack that is parsing the individual fields, not the code that has to determine the field boundaries. -=R On Fri, Nov 22, 2013 at 10:17 PM, Tatsuhiro Tsujikawa <tatsuhiro.t@gmail.com > wrote: > > > > On Sat, Nov 23, 2013 at 1:57 AM, RUELLAN Herve <Herve.Ruellan@crf.canon.fr > > wrote: > >> +1 >> >> I like both #1 and #3 proposals, with a preference for #3 as it doesn't >> introduce any new mechanism. >> Not using #2 means that we will be able to drop some emission ordering >> stuff from HPACK. >> >> > I also prefer #3 to #1 with the same reason. > > Best, > Tatsuhiro Tsujikawa > > > >> Hervé. >> >> > -----Original Message----- >> > From: Amos Jeffries [mailto:squid3@treenet.co.nz] >> > Sent: vendredi 22 novembre 2013 15:20 >> > To: ietf-http-wg@w3.org >> > Subject: Re: #305 Header ordering >> > >> > On 22/11/2013 4:37 p.m., Mark Nottingham wrote: >> > > To be clear - what makes me somewhat comfortable with this approach is >> > that the default is that order is preserved; only if you know that >> ordering *is* >> > insignificant are you allowed to break it up. >> > > >> > >> > +1. >> > >> > Amos >> > >> > >> > >> >> >> >
Received on Saturday, 23 November 2013 06:38:39 UTC