Re: [hybi] Websockets sub-protocol registry

[Ted Hardie <[e-mail]>]

On Mon, Mar 14, 2016 at 1:31 AM, Anne van Kesteren <[e-mail]> wrote:

> On Fri, Mar 11, 2016 at 5:42 PM, Ted Hardie <[e-mail]> wrote:
> > RTCWEB re-uses the websockets sub-protocol registry for identifying
> > sub-protocols using its data channels.  During discussion on the list, we
> > realized that the method by which these are compared is not specified and
> > that it is therefore not clear whether IANA could or should register two
> > values which differed only in case.
> >
> > Would anyone object to specifying that IANA should not register two
> values
> > that differed only in case?  Is there a specification which determines
> > whether the match is case-sensitive or case-insensitive that we missed?
>
> Why would you assume it to be case-insensitive? Typically any kind of
> matching is case-sensitive unless otherwise specified.
>

We simply don't want to assume it; we'd like it to be specified, at least
in the RTCWEB case.  We're happy to specify that it be case-insensitive
match.  Based on a cursory look at libwebsockets and a few python libs, it
looks like that is the commoning match for the original usage as well.

Back to the IANA side of the question: Any objection to specifying that
IANA should not register two values that differed only in  case?  The
current instructions aren't clear on that.

regards,

Ted Hardie