On 27 Mar 2013, at 08:11, Holger Hans Peter Freyther <[email protected]> wrote:
> On Tue, Mar 26, 2013 at 09:08:09PM +0100, Sven Van Caekenberghe wrote: > >> Maybe I am telling you things you already know, but TCP and UDP are >> fundamentally different. UDP is really one-shot, try-and-hope, completely >> asynchronous. In networking you generally only get a ConnectionClosed when >> you actually read or write (and even then ;-). > > Yes, but at least for the unix platform on close the aio/select handler is > removing the socket from the registration. So there is no way to read the > zero from the dead socket. :) > >> For your concrete question (2): just start a reading process, wrap it in a >> handler and clean up. For UDP this will never happen: you will keep on >> reading nothing. For TCP reading will eventually fail if the other end >> disappears. > > Do you have an example for "handler and clean up"? I resorted to code > like this for now in my 'blocking' read. > > [self dataAvailable] > whileFalse: [[self waitForDataFor: 10] on: ConnectionTimedOut do: > []]. > > But I have not ported my client code so I don't know if the above is > working or not. This way I will get the ConnectionClosed exception at > least within 10 seconds. > > thank you > > holger It is still not clear to me if this is for UDP or TCP, I would be surprised of #dataAvailable or #waitForData: actually work for UDP. I gave you some pointers to examples that do work. Your code snippet looks reasonable to me, but I would set timeouts on the socket(streams) and use regular read/write operations and then loop. Sven
