| 2013-02-14 20:37:40 utc | Skippy1 | Do participant instances persist between different launches in the engine? Or is a new one instantiated whenever it's called in a process? |
| 2013-02-14 21:04:32 utc | jmettraux | Skippy1: hello, each time a participant gets instantiated to service a workitem |
| 2013-02-14 21:05:04 utc | jmettraux | or "it's always a fresh participant instance that services a workitem" |
| 2013-02-14 21:05:08 utc | Skippy1 | cool |
| 2013-02-14 21:05:27 utc | Skippy1 | just worried about setting instance variables in my participants... |
| 2013-02-14 21:21:12 utc | jmettraux | they will get lost |
| 2013-02-14 21:22:07 utc | jmettraux | you can pass options when registering the participants, they can use those options to get their "context" |
| 2013-02-14 21:22:18 utc | jmettraux | connect to a db, configure themselves, etc... |
| 2013-02-14 21:22:28 utc | Skippy1 | yeah I want transient instance variables |
| 2013-02-14 21:37:08 utc | bigfleet | Anybody around who can walk me through attaching a ProcessObserver subclass to a ruote dashboard properly? |
| 2013-02-14 21:38:40 utc | jmettraux | bigfleet: hello and welcome to #ruote |
| 2013-02-14 21:39:11 utc | bigfleet | jmettraux: Thanks! This may not be my last question-- I'm getting my feet wet, doing my learnings. :-) |
| 2013-02-14 21:39:33 utc | jmettraux | I'm in a meeting right now, but I'll try to help |
| 2013-02-14 21:39:50 utc | bigfleet | Please wait, I'll be hanging around for an answer. |
| 2013-02-14 21:41:21 utc | jmettraux | what doc are you using to try to hook this observer? |
| 2013-02-14 21:42:08 utc | bigfleet | https://github.com/jmettraux/ruote/blob/master/lib/ruote/observer.rb |
| 2013-02-14 21:42:20 utc | bigfleet | http://ruote.rubyforge.org/process_observation.html |
| 2013-02-14 21:48:20 utc | jmettraux | please look at |
| 2013-02-14 21:48:21 utc | jmettraux | https://github.com/jmettraux/ruote/blob/master/test/functional/ft_76_observer.rb |
| 2013-02-14 21:48:58 utc | jmettraux | and https://github.com/jmettraux/ruote/blob/master/test/functional/ft_77_process_observer.rb |
| 2013-02-14 21:51:07 utc | bigfleet | OK, I see |
| 2013-02-14 21:51:39 utc | jmettraux | ACTION adds todo for self, explain how to hook observers in their rdoc and in their documentation |
| 2013-02-14 21:51:46 utc | bigfleet | I suspect I have attached it correctly then, just didn't match my expectations of behavior (that is OK, I'm learning) |
| 2013-02-14 21:51:56 utc | bigfleet | If an error is raised in on_launch, will that just disappear? |
| 2013-02-14 21:52:09 utc | bigfleet | I do have ruote_kit and expected an error |
| 2013-02-14 21:55:07 utc | jmettraux | yes, the observer should keep its errors to itself |
| 2013-02-14 21:55:27 utc | bigfleet | That ought to be easy to verify |
| 2013-02-14 21:58:30 utc | bigfleet | I did not see the log message I expected, although I do see I am initializing the Observer |
| 2013-02-14 21:59:04 utc | bigfleet | If I attach it to a dashboard, and define on_launch, should I execute for launch of that dashboard's participants? |
| 2013-02-14 21:59:14 utc | bigfleet | I suppose I am asking "on_launch of what?" |
| 2013-02-14 21:59:55 utc | jmettraux | just checked, if there is an error in your observer the worker should choke on $stdout |
| 2013-02-14 22:00:24 utc | jmettraux | so you launch a new process and your on_launch observer method is not called? |
| 2013-02-14 22:00:43 utc | bigfleet | That is how it appears, but it could be user error |
| 2013-02-14 22:00:59 utc | bigfleet | I do not need to interrupt your meeting, I can continue experimentation |
| 2013-02-14 22:01:27 utc | jmettraux | ok |