Excerpts from Simon Schampijer's message of Mon May 23 10:01:01 +0200 2011:
> > I guess we don't need to subclass Transport, but we probably need to > > call socket.setdefaulttimeout() during Sugar start-up: [...] > Hmm, yes sounds about right. I don't think we have to do it during sugar > startup (not sure about other side-effects), just after the 'else' is > fine. The side effects are exactly the reason I want it to run on start-up. That way things will always behave the same, instead of potentially breaking in odd ways only after the user has used Register during the current session (which is rather hard to diagnose - to the user it's a one-time glitch, not reproducible). Another, less risky idea would be to call socket.setdefaulttimeout(None) afterwards (using try-final) to reset the value to the default (no timeout). > In any case I guess the real solution is to make the registration > asynchronous: http://bugs.sugarlabs.org/ticket/2289#comment:4 but I did > not want to block on it. +1 on both. If somebody provides a patch, we can push it to master. Sascha -- http://sascha.silbe.org/ http://www.infra-silbe.de/
signature.asc
Description: PGP signature
_______________________________________________ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
