[04:53] *** Lawouach_ has joined #kamaelia |
[05:31] *** salmon_ has joined #kamaelia |
[06:47] *** MS-- has joined #kamaelia |
[06:55] *** Uraeus has joined #kamaelia |
[07:40] *** Uraeus_ has joined #kamaelia |
[07:42] < Lawouach_> grumble, KeyboardInterrupt and Axon aren't friends |
[08:06] *** Lawouach__ has joined #kamaelia |
[08:43] < MS--> Lawouach: what behaviour would you expect OOI? |
[08:44] < MS--> I'm guessing "DIE DIE DIE!" :-) |
[08:44] < Lawouach> MS--: It seems that the scheduler doesn't trap it an escape abruptely when it happens |
[08:45] < Lawouach> No I'd like "Let me clean it up and then DIE" |
[08:46] < MS--> :) |
[08:46] < MS--> ie it should call .stop() on all components. |
[08:46] < MS--> and then shutdown |
[08:46] < MS--> That is a bit nasty really isn't it |
[08:47] < Lawouach> yeah but currently it makes things rather harsh as it is |
[08:47] < Lawouach> I'm running a script that will create hundreds of XMPP account which I'd like to be deleted whenever the script stops or is stopped |
[08:48] < Lawouach> I've banged my head at it for a couple of days now and it's almost impossible to predict what can happen on KeayboardInterrupt |
[08:49] < MS--> Just trying to think where to put KeyboardInterrupt handling |
[08:50] < MS--> If it happens inside a component, it's fairly obvious where to catch it |
[08:50] < MS--> if it happens outside, it's less obvious |
[08:50] < MS--> (ie in the scheduler) |
[08:51] < Lawouach> indeed |
[08:51] < Lawouach> it'd be nice if you could think about it in your refactoring :) |
[08:52] < MS--> Will do |
[08:52] < MS--> currently thinking about fwiw |
[08:52] < Lawouach> ta :) |
[08:54] < MS--> Early versions of the scheduler were designed such that instead of simply catching StopIteration and removing the component, that they would catch everything and remove the component. ie not crash |
[08:54] < MS--> the system (that is) |
[08:54] < MS--> That was at a much earlier stage though, when it was useful for debugging the system |
[08:55] *** Uraeus has joined #kamaelia |
[08:55] < MS--> Naturally that was a problem with KeyboardInterrupt as well |
[08:55] < MS--> so that led to the decision to simply catch StopIteration and then left I guess |
[08:56] < MS--> I think we're probably getting to the stage where resilience in servers/etc is preferable over debugging - especially given it's clearer how you would insert that. |
[09:44] < Lawouach__> Seems fair |
[13:17] *** salmon_ has joined #kamaelia |
[15:01] *** Lawouach has joined #kamaelia |
[15:47] *** vmlemon_ has joined #kamaelia |
[15:49] < vmlemon_> Hi |
[17:03] *** eikenberry_ has joined #kamaelia |
[17:13] *** eikenberry has joined #kamaelia |
[17:13] *** salmon_ has joined #kamaelia |
[17:28] *** Davbo has joined #kamaelia |
[17:34] *** eikenberry_ has joined #kamaelia |
[18:05] *** eikenberry has joined #kamaelia |
[18:10] < vmlemon_> Hi eikenberry |
[18:10] < eikenberry> hey |
[18:12] *** vmlemon_ is surprised by how buggy the Symbian Foundation Developer site is, even for an "early-look" closed beta |
[18:13] *** vmlemon_ wonders what's happening with Kamaelia, these days |
[18:27] *** Davbo is strugling to pick a dissertaion project |
[18:29] < MS--> Davbo: "The importance of cheese in the developmental history of 3D gaming systems" |
[18:29] < MS--> (bonus points if anyone gets the reference) |
[18:30] *** vmlemon_1 has joined #kamaelia |
[18:37] < Davbo> Did John Carmack do a bit of curdling on the side? |
[18:37] *** Davbo doesn't get the reference |
[18:37] < Davbo> :) |
[18:52] < Davbo> looking at the list of possible projects and tempted by semantic web stuff, but it's not very "computer sciency" |
[18:54] < MS--> Davbo: http://www.angusm.demon.co.uk/AGDB/DBM1/Merc.html |
[18:56] < Davbo> haha, Kraft, awesome. |
[20:12] *** Davbo has joined #kamaelia |