Open GoogleCodeExporter opened 9 years ago
depends on issue 40
Original comment by tro...@trouch.com
on 27 Mar 2013 at 11:59
FYI, After looking at some aspects behing COMET, long polling and Bayeux I
would prefer WebSockets as the most future-proof solution. Interestingly there
already exists a OpenSource library family that supports some very important
environments like Android, jScript and Python.
See e.g. http://autobahn.ws/ which would allow to make Android native apps
right away and salso provides a JS library.
Only caveat is that the Python version is relying on Twisted which is not
Python 3.0 ready afaik. Assume that CherryPy is.
You should make shure that the official RFC 6455 release of WebSockets is used.
Looks like this https://github.com/Lawouach/WebSocket-for-Python is providing a
WebSocket implementation for Python that can be used on top CherryPy for Python
2 and 3.
If I understand WebSockets right we still need some additional component that
forms the "wire protocol" that is used inside the WebSockets channel. But no
standards exist here, so it could just be the REST/JSON API that WebIOPi
already implements.
Andreas
Original comment by andreas....@googlemail.com
on 23 Apr 2013 at 9:18
Maybe this http://wamp.ws/ is an interesting approach/example for such a wire
protocol that is highly compatible with what WebIOPi already does in the sense
of JSON usage for "*" requests.
Andreas
Original comment by andreas....@googlemail.com
on 23 Apr 2013 at 9:36
don't worry about that ;)
I think to use websocket with classical long-polling fallback.
I will certainly implement the PubSub model on top of everything.
I'm thinking about cherrypy with ws4py, but not sure yet.
You are right twisted does not support python 3.
I don't know wamp.ws, I will check it.
Original comment by tro...@trouch.com
on 23 Apr 2013 at 4:47
Checked wamp.ws, and its pink color remember me something.
I may have already seen it in my watch.
Well, it implement PubSub as well ;)
Now I have to remember why I don't retained it.
Original comment by tro...@trouch.com
on 23 Apr 2013 at 5:04
Original comment by tro...@trouch.com
on 4 Jan 2014 at 8:29
Original issue reported on code.google.com by
tro...@trouch.com
on 23 Nov 2012 at 3:52