IRC Logs for #circuits-dev Monday, 2013-02-11

mehereRight, the outer condition prevents adding handlers if the channel is a manager, unless the component happens to be that manager. Or (maybe) simpler: if the channel happens to be a manager, a component's handlers are currently only taken into consideration as handlers if that (manager-)channel happens to be the component.02:18
mehereWe simply have to make up our minds about the semantics. I don't see the reason for the current implementation. IMHO if you send an event to a component (i.e. use the component as channel) this should be like sending it to id(component). Except that I'd have to use something like ("object" + str(id(component))" to make this really unique and that it wouldn't look as good in debug output.02:23
mehereBut currently there's a "funny" restriction. Only handlers of the component itself will receive events sent using the component as channel. Of course, this can also be defined as a feature, Though the use case is not that obvious to me.02:24
mehereMy use case is: I have a component as a "view" of some particular component (the "model"). So I'm interested in events sent to that component. This allows me to use the component as a unique channel. I don't have to create specially named channels for that purpose with id() or uuid4(). But, of course, I can do this. So the whole point is not urgent. I've just been surprised about the current behaviour.02:28
prologicmehere, so does remoeving that outter condition resolvethis for your use-case?09:53
prologicallowing us to use the component as a unique channel?09:53
prologicbecause I do happen to agree with you :)09:54
prologicthe current semantics were probably not all that thought through very well to be honest09:54
prologicmehere, ping?10:47
prologichmm10:48
prologicyour client doesn't respond to CTCP TIME10:48
prologic:/10:48
prologicoh well answer when you're here:)10:48
prologichttp://codepad.org/aHvAhEZG10:48
prologicThis patch against tip works here10:48
prologicall 72 core tests pass10:48
prologicAnd it seems to have the behavior your after10:49
*** anunnaki has quit IRC13:30
*** anunnaki has joined #circuits-dev14:05
mehereLooks ok.14:07
*** anunnaki has quit IRC16:10
*** anunnaki has joined #circuits-dev16:11
prologicmehere, this patch passes all the unit tests16:28
prologicmehere, what were the ones you thought would fail?16:29
prologicI only have ipv6 issues here where some ipv6 tests are failing because of "network Unreachable" errors16:29

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!