ruote tmp/log_2011-06-03.html

2011-06-03 22:58:26 utc lucas-howcast Hey guys, I'm having some issues with setting the timeout for a specifc task and using the :on_timeout => "some_previous_process". It goes back to the process but then skips steps after..
2011-06-03 22:58:31 utc lucas-howcast does this ring a bell to anyone?
2011-06-03 22:59:33 utc jmettraux lucas-howcast: hello, welcome to #ruote
2011-06-03 22:59:42 utc lucas-howcast Hey, thx
2011-06-03 22:59:56 utc jmettraux do you have a gist of the process ?
2011-06-03 23:00:02 utc lucas-howcast sure, one sec
2011-06-03 23:06:36 utc jmettraux lucas-howcast: another question, what version of ruote ?
2011-06-03 23:07:32 utc lucas-howcast jmettraux: here's the gist:
2011-06-03 23:07:34 utc lucas-howcast and the version is...
2011-06-03 23:07:54 utc lucas-howcast 2.1.11
2011-06-03 23:11:48 utc lucas-howcast let me know if the gist is confusing, I can give you more info about it
2011-06-03 23:13:48 utc jmettraux which timeout is "not working as expected" ?
2011-06-03 23:13:57 utc jmettraux which on_timeout
2011-06-03 23:14:33 utc lucas-howcast this one:
2011-06-03 23:14:34 utc lucas-howcast filmmaker :task => 'accept', :timeout => '${filmmaker_accept_timeout}', :on_timeout => 'assigning'
2011-06-03 23:14:46 utc lucas-howcast it seems it works though, beause it does go back to "assigning"
2011-06-03 23:14:52 utc lucas-howcast but if a timeout ocurred,
2011-06-03 23:15:09 utc lucas-howcast then instead of going to "make video" which would be the following task
2011-06-03 23:15:27 utc lucas-howcast it skipps it and goes to the third task which would be "coordinate review"
2011-06-03 23:15:37 utc lucas-howcast that does not happen if no timeout ocurred
2011-06-03 23:16:28 utc lucas-howcast so the order should be: assign - accept - make video - review
2011-06-03 23:16:39 utc lucas-howcast if there's a timeout on accept, it goes back to assign, which is OK
2011-06-03 23:16:45 utc lucas-howcast but then skips make video and goes directly to review
2011-06-03 23:16:52 utc lucas-howcast makes sense?
2011-06-03 23:17:52 utc jmettraux I'm in a middle of a meeting, It should be over in a few minutes, I'll get back to you right after it
2011-06-03 23:18:06 utc lucas-howcast sure, thank you very much!
2011-06-03 23:22:59 utc jmettraux maybe the process is confused, there are 3 cursors
2011-06-03 23:23:35 utc jmettraux you seem to only want to pass orders to the main cursor
2011-06-03 23:24:12 utc jmettraux wait, you're using :ref, it should be ok
2011-06-03 23:24:22 utc lucas-howcast k
2011-06-03 23:27:15 utc jmettraux your 'assigning' subprocess, you want it to rewind if the filmmaker timed out ?
2011-06-03 23:28:12 utc lucas-howcast if the filmmaker didn't 'accept' then someone needs to do the 'assign' process again. Does that answer your question?
2011-06-03 23:28:38 utc lucas-howcast so, yeah I guess it needs to reweind
2011-06-03 23:28:42 utc lucas-howcast rewind*
2011-06-03 23:28:53 utc lucas-howcast if it timed out or if the filmmaker didn't accept it
2011-06-03 23:29:14 utc jmettraux I'd drop the on_timeout and simply let the rewind trigger (hopefully ${accepted} is not set)
2011-06-03 23:30:43 utc lucas-howcast would that work with the timeout?
2011-06-03 23:31:11 utc jmettraux when there is no timeout specified, it just sets the __timed_out__ field and goes on
2011-06-03 23:31:22 utc jmettraux it would thus reach the rewind
2011-06-03 23:32:03 utc lucas-howcast I see, I thought maybe the timeout without an :on_timeout would cancel it or something
2011-06-03 23:32:18 utc lucas-howcast great, then I'll try that
2011-06-03 23:32:29 utc lucas-howcast do you know why is this happening though?
2011-06-03 23:32:38 utc lucas-howcast I ask in case I see this again, since I have other timeouts in there
2011-06-03 23:33:08 utc jmettraux I'm not sure what's going on, but I prefer help you simplify your process first
2011-06-03 23:33:36 utc lucas-howcast Gotcha, I'll try that right now then
2011-06-03 23:33:42 utc lucas-howcast thx a lot, I'll let you know how it goes
2011-06-03 23:33:48 utc jmettraux on_timeout "bubbles" to a participant or a subprocess
2011-06-03 23:34:13 utc jmettraux it replaces the expression that timed out with the given participant or subprocess
2011-06-03 23:34:41 utc jmettraux once the sub or the participiant is done, it [should] resume(s) after the timed out expression
2011-06-03 23:35:25 utc lucas-howcast right, I thought that too. Hopefully just triggering the rewind will suffice
2011-06-03 23:35:35 utc jmettraux in your 'reviewing' subprocess, I'd remove the cursor (and the :ref => 'main' altogether), like you did for 'filming'
2011-06-03 23:36:02 utc lucas-howcast you think that may cause conflicts at some point?
2011-06-03 23:36:19 utc jmettraux no, it's just a matter of having less "code"
2011-06-03 23:36:25 utc jmettraux making it easier to read
2011-06-03 23:36:30 utc lucas-howcast ah, gotcha
2011-06-03 23:36:50 utc lucas-howcast yeah sorry if it's not too clean, I'm a Ruote newbie still :)
2011-06-03 23:36:55 utc lucas-howcast thx for the tips too
2011-06-03 23:36:58 utc jmettraux in 'filming' as well, I'd remove the on_timeout
2011-06-03 23:37:25 utc jmettraux you have this jump philosophy in your process, I'd go on with it
2011-06-03 23:37:47 utc lucas-howcast yeah that's possible too
2011-06-03 23:38:07 utc jmettraux you could do something like "jump :to => 'assigning', :if => '${__timed_out__}'"
2011-06-03 23:38:35 utc lucas-howcast oh,sweet
2011-06-03 23:39:49 utc jmettraux but __timed_out__ is a bit ugly when appearing in a process, maybe simply jump if there is no ${review}
2011-06-03 23:40:07 utc jmettraux ok, looks good
2011-06-03 23:40:11 utc lucas-howcast still good to know as a last measure
2011-06-03 23:40:17 utc jmettraux exactly