Philippe, I've never had much use for UDP, but the common things should work the same, at least to a point. UDP is the protocol that does not guarantee packet order or even delivery, right? So anything remotely stream related is out, and there might be other things that should legitimately be different, if I'm thinking of the right thing. While I'm displaying ignorance, there is a negotiated rate serial protocol that ends up being implemented using sockets. I can't think of the name of it but know where to look for a reminder. It is something I would like to see supported "some day." Bill -----Original Message----- From: pharo-project-bounces@lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of Philippe Marschall Sent: Wednesday, August 26, 2009 11:46 AM To: pharo-project@lists.gforge.inria.fr Subject: Re: [Pharo-project] new socket API Btw, while we're at it: TCP sockets and UDP sockets have a different interface. Some messages can only be sent to TCP sockets others only to UDP sockets. Yet they are both represented by the same Socket class. Cheers Philippe Schwab,Wilhelm K wrote:
Stef,
I am not sure exactly what should be next, but I do not think we can simply kick it back to squeak-dev any more than we can the GUI/IDE, streams, etc., as sockets are very important. I am slowly chipping away at things that are standing in the way of my first Pharo success story; the impact of sockets has been minimal, but that will change. When I get going, I will be looking carefully at how things behave when peers do not respond (only the calling thread should suffer). Dolphin's socket streams are nice; Pharo seems to lean toward a single read-write stream, which should be fine as long as it manages not to hobble the image when the wrong wire gets kicked. SSL via OpenSSL should be easy to use; that will be an interesting adventure, but it needs to happen.
With all due respect to those who have disagreed, I'm still highly suspicious that the networking problems have far more to do with IPv6 vs. 4 than the OS.
Bill
-----Original Message----- From: pharo-project-bounces@lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of Stéphane Ducasse Sent: Wednesday, August 26, 2009 7:21 AM To: Pharo-project@lists.gforge.inria.fr Subject: Re: [Pharo-project] new socket API
so what is the next step?
stef
On Aug 26, 2009, at 8:19 AM, Michael Rueger wrote:
Philippe Marschall wrote:
There's gotta be a better way to do this.
The whole API with arrays where you have to know the index, ByteArray objects, integer flags and so forth seems like a blast from the past.
We could put some default behavior wrapper methods around this. But this is really a discussion for squeak-dev, as the IPv6 implementation originates from etoys/olpc.
Michael
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project