ruote tmp/log_2012-07-30.html

2012-07-30 09:46:31 utc hartog hi all!
2012-07-30 09:47:25 utc hartog jmettraux: did you start work on a 'raw' observer? I'd be happy to take your code sample and put it in my branch
2012-07-30 10:20:08 utc jmettraux hartog: hello, not yet, ok if I do it within 3 hours?
2012-07-30 10:22:06 utc hartog np; i was thinking of Ruote::ProcessObserver and Ruote::ProcessObserver::Simple (the latter being fully candy waxed) - would that fit your expectations
2012-07-30 10:22:40 utc jmettraux sounds good
2012-07-30 10:23:05 utc hartog ok!
2012-07-30 10:24:23 utc jmettraux wait, there is something that is a bit odd: it's not really process observing, it's more like engine observing
2012-07-30 10:24:38 utc jmettraux instances of our observers observe all the processes
2012-07-30 10:25:18 utc jmettraux well, that could be OK since you tend to drop the "s" in english in that case "process observer" and not "processes observers"
2012-07-30 10:25:21 utc jmettraux wdyt?
2012-07-30 10:26:18 utc hartog hmmmmm
2012-07-30 10:27:34 utc jmettraux Observer
2012-07-30 10:27:46 utc jmettraux Ruote::Observer, Ruote::Observer::Simple
2012-07-30 10:28:10 utc jmettraux ok, ProcessObserver is not bad
2012-07-30 10:30:51 utc hartog the essence of my sugarcoating is to take away the abstract - so ProcessObserver would fit me better there; otoh Observer (or MessageObserver) would be just right for your case
2012-07-30 10:32:47 utc jmettraux that'd be intersting, I could do the Observer and you could do the ProcessObserver
2012-07-30 10:33:44 utc hartog agreed
2012-07-30 10:34:24 utc jmettraux I'm also thinking of doing the pre_notify anyway, it's evil, but since it wouldn't make it into the ProcessObserver it can be kept only for special needs
2012-07-30 10:34:31 utc jmettraux ok with you?
2012-07-30 10:40:00 utc jmettraux btw, when you add files, put your name instead of mine for the copyright and made in germany is cool
2012-07-30 10:43:48 utc hartog the pre_notify sounds good
2012-07-30 10:43:59 utc hartog (i would be 'abusing' it anyway ;))
2012-07-30 10:47:47 utc jmettraux ok, the next thing we have to do: think about your "launched_at" request
2012-07-30 10:48:16 utc jmettraux I think it'd make a worthy addition to the workitem thinggy
2012-07-30 10:50:21 utc hartog i would love to see such basic info in ruote, and I would like to have workitems that optionally do not disappear when the flow ends
2012-07-30 10:50:58 utc hartog but the latter I could arrange myself using the pre_notify ... *evil laughter*
2012-07-30 10:54:17 utc jmettraux sorry, I don't want to archive workitems, somehow people could use a variant of the StorageHistory to do that
2012-07-30 11:00:57 utc hartog there's a plan!
2012-07-30 11:02:33 utc hartog but then I would love to see 'flunked_at', 'canceled_at', 'terminated_at', etc since I could then query the storage history for such infos
2012-07-30 11:03:16 utc jmettraux well a process is launched once
2012-07-30 11:03:25 utc jmettraux it can be cancelled more than twice
2012-07-30 11:03:29 utc jmettraux more than once
2012-07-30 11:04:02 utc hartog you know, Matz invented the array for that ;)
2012-07-30 11:04:13 utc hartog ACTION chuckles
2012-07-30 11:04:23 utc jmettraux the workitem would grow heavier and heavier
2012-07-30 11:05:00 utc hartog from where I am standing, the ruote workitem is ultra light weight as it is
2012-07-30 11:05:10 utc jmettraux great
2012-07-30 11:05:30 utc jmettraux I don't want a growing cost passing around the workitems
2012-07-30 11:05:42 utc jmettraux I'd like the history cost to be paid only when needed
2012-07-30 11:06:08 utc hartog hence; write your own observer and add that info
2012-07-30 11:06:42 utc hartog or rather; track that info
2012-07-30 11:08:19 utc jmettraux granted, embedding the history in the workitem is nice, but if the flow lasts two year with many many events, it's going to be fat
2012-07-30 11:08:56 utc jmettraux ok, pushed: https://github.com/jmettraux/ruote/commit/86db03ea06c6b6a0b33cb8af765c6bfa315181ab
2012-07-30 11:09:22 utc jmettraux taking a break now
2012-07-30 11:09:29 utc jmettraux I'll be back a bit later
2012-07-30 12:31:00 utc jmettraux hartog: back, but going to bed actually
2012-07-30 12:31:13 utc hartog good night then!
2012-07-30 12:31:31 utc jmettraux good day! :-)
2012-07-30 12:31:37 utc hartog i had one question earlier on; but I forgot
2012-07-30 12:31:41 utc hartog :D
2012-07-30 15:04:22 utc hartog g'night all!
2012-07-30 20:51:03 utc hartog jmettraux: now i'm off to bed ;)
2012-07-30 20:51:16 utc jmettraux have a good night!
2012-07-30 20:51:19 utc hartog looking forward to your response
2012-07-30 20:51:30 utc hartog cu in < 12hrs
2012-07-30 20:51:35 utc hartog bbye!
2012-07-30 20:51:38 utc jmettraux ciao!