- From: Mark Nottingham <mnot@mnot.net>
- Date: Fri, 22 Jan 2010 11:53:37 +1100
- To: Julian Reschke <julian.reschke@gmx.de>
- Cc: mjb@asplake.co.uk, Apps Discuss <discuss@apps.ietf.org>, HTTP Working Group <ietf-http-wg@w3.org>
I agree in principle, but practically, it's almost a foregone conclusion that people will produce these unquoted, and implementations will consume them happily. I'd rather align with what is done in practice, even if it does preclude using a generic parser... On 22/01/2010, at 2:08 AM, Julian Reschke wrote: > Mike Burrows wrote: >> Hi Julian (and all), >> Which made have a closer look at existing parameters (sorry for not >> bringing this up earlier): >> | ( "type" "=" type-name "/" subtype-name ) >> I believe this *definitively* needs quoting, as "/" is a separator >> character in HTTP and thus can not appear in a token. >> But is 'type-name "/" subtype-name' a token? I didn't interpret as such - in fact I assumed (perhaps erroneously) that it was in fact two tokens separated by "/". >> Regards, >> Mike >> ... > > You are right, it's not. But, in general, parameter values either take tokens or quoted-strings, so a generic parser of HTTP parameters would fail to parse this correctly. > > Maybe that's not a big issue in practice, because people use custom parser per header field, but it would be good if we could avoid that. > > Best regards, Julian > _______________________________________________ > Apps-Discuss mailing list > Apps-Discuss@ietf.org > https://www.ietf.org/mailman/listinfo/apps-discuss -- Mark Nottingham http://www.mnot.net/
Received on Friday, 22 January 2010 00:54:11 UTC