Re: [hybi] #1: HTTP Compliance

[Roberto Peon <[e-mail]>]

On Tue, May 18, 2010 at 6:35 PM, Jamie Lokier <[e-mail]> wrote:

> Roberto Peon wrote:
> >
> >    On Tue, May 18, 2010 at 5:31 AM, Greg Wilkins <[1][e-mail]>
> >    wrote:
> >
> >    On 18 May 2010 14:12, Jamie Lokier <[2][e-mail]> wrote:
> >    > I'm arguing that the second object should be able to reuse the
> >    > connection resulting from the first one's failure to negotiate on
> >    the
> >    > same port
> >
> >      Exactly!
> >      If the handshake is HTTP compliant, then the connection for a
> >      websocket handshake could be taken from the existing pool of
> >      idle connections to a host.  That would save the time needed to
> >      establish the connection.
> >      If the handshake is HTTP compliant, then a rejected websocket
> >      handshake could put the connection back into the idle pool to be
> >      available for subsequent handshakes or XHRs.
> >      If the handshake is not HTTP compliant, then the connection
> >      needs to be created on every handshake and disposed as
> >      untrusted after every rejection.
> >
> >    Which, to use the technical phrase, sucks.
>
> What do you mean?  I don't understand your point at all.
>
> Do you mean that the current design preventing reuse sucks, or that
> doing reuse would suck (even though you wouldn't have to if you didn't
> want to)?
>

Heh.

And in case I'm misunderstanding the joke when it was a serious question,
preventing reuse sucks...

-=R


>
> -- Jamie
>