<...> > As you can see, when the coroutine finishes Copas closes *the socket > that coroutine was awakened with*. So if you do a copas.send, it > yields, and then you exit after coming back, Copas closes the socket > that you sent to. If Javier is here he can tell if that is indeed the > intended behavior or not. :-)
I see. > If you remove the skt:close() line then your code should be fine, > although you will now have to remember to close all sockets, unless > you are ok with letting the GC do it. :-) Well, I do always explicitly connections, so that should be acceptable. Except that I hate to live on patched third party dependencies. What about adding an option in copas? I can produce a patch myself if needed, if you say which fix is the most appropriate in this case. I believe this unexpected behaviour is a bug. If copas handler is tied to a single connection, then it should assert when resumed with wrong connection — this would save users (like me) a lot of debugging... Also: is this limitation documented anywhere? Thank you! Alexander. _______________________________________________ Kepler-Project mailing list Kepler-Project@lists.luaforge.net http://lists.luaforge.net/cgi-bin/mailman/listinfo/kepler-project http://www.keplerproject.org/
