users@websocket-spec.java.net

[jsr356-users] [jsr356-experts] Re: Valid ServerEndpoint.value() path specs?

From: Joakim Erdfelt <joakim_at_intalio.com>
Date: Fri, 19 Apr 2013 15:13:22 -0700

On Fri, Apr 19, 2013 at 3:10 PM, Mark Thomas <mark_at_homeinbox.net> wrote:

> On 11/04/2013 02:27, Mark Thomas wrote:
> > On 10/04/2013 00:48, Joakim Erdfelt wrote:
> >
> >> These forms bring up a few questions.
> > My $0.02
> >
> >> Should we prevent path navigation segments? ("/../", "/./", and "//")
> > Yes
>
> Based on my previous message I think I have changed my mind for //.
> However, that creates various problems for implementations. A simpler
> option might be to say that variables are always required and must have
> length > 0. That would remove a number of edge cases. Is that reasonable?
>
>
I like that.
A PathParam can only be considered a PathParam if it has a value of length
> 0.

+1 from me.