Re: [apps-discuss] Fun with URLs and regex

[Nico Williams <[e-mail]>]

On Wed, Jan 28, 2015 at 04:51:05PM -0500, Sam Ruby wrote:
> 
> 
> On 01/28/2015 04:40 PM, Roy T. Fielding wrote:
> >On Jan 28, 2015, at 1:14 PM, Julian Reschke wrote:
> >>On 2015-01-28 06:36, Mark Nottingham wrote:
> >>>... which brings about another interesting observation -- only http and https define fragments in their syntax; the other schemes do not.
> >>>...
> >>
> >>It's because you asked for that in <https://lists.w3.org/Archives/Public/ietf-http-wg/2013AprJun/0187.html>, and apparently were successful in convincing Roy.
> >
> >It is a very very very old debate regarding whether the fragment is part
> >of a URI or something attached to the end of a URI, but that was resolved
> >in RFC3986 (since the only thing that really matters here is that a fragment
> >is going to be parsed as such regardless of the scheme).
> >
> >HTTP was merely updated to reflect what STD66 calls a URI.
> 
> Based on this discussion, I am gathering that the correct way to
> validate a URI with a known scheme is as follows:
> 
>   return new RegExp("^" + known[scheme] + "($|#" + fragment +
> ")").test(string)
> 
> Anybody care to confirm or deny?

For dereferenceable URI schemes that looks right to me.

It's not really right for mailto: URIs (which aren't dereferenceable).

I suppose data: URIs could have fragments.

Nico
--