Re: [Gnash-commit] /srv/bzr/gnash/trunk r10414: Move stream creation logic from NetStream to NetConnection (will be useful

[Sandro Santilli]

Oh, hold on a sec.
For remoting, that URI is used to postpone actual connection
to NetConnection.call() time.

The way potlatch works is:

        nc = new NetConnection();
        nc.connect("http://xxxxxyyy";);
        nc.call("xxy");

Gnash triggers nc.onStatus with a failure status code
corrently when NetConnection.connect() is passed an URL.
Potlatch doesn't fail due to the fact that onStatus
is assigned *after* the .connect call, so Gnash doesn't find it
(by luck then).

So:

 1) we shouldn't send a failure message on .connect, unless
    something else happens, maybe malformed url ?
 2) we shouldn't send onStatus immediately, but delegate it
    to someone else (eventual network thread for rtmp or processing
    of queued calls for remoting).

--strk;

On Sat, Dec 13, 2008 at 04:57:21PM +0100, Sandro Santilli wrote:
> On Sat, Dec 13, 2008 at 04:54:46PM +0100, Benjamin Wolsey wrote:
> 
> > Then on the attached NetStream: NetStream.play("stream_name");
> > 
> > So in this case, NetConnection would know it was connected to an rtmp
> > server. NetConnection_as::getStream() would interpret the argument
> > passed from NetStream.play() not as an URL, but as a stream name
> > (whatever that is), and accordingly ask the Gnash object that's managing
> > the rtmp connection for that stream.
> > 
> > Have I understood that correctly?
> 
> Yes, as far as I understood so far.
> 
> --strk;

-- 

 GIS & Flash consultant/developer     ()   ASCII Ribbon Campaign
 http://foo.keybit.net/~strk          /\   Keep it simple!