ruote tmp/log_2011-09-21.html

2011-09-20 22:30:32 utc pbh_ hey, has anyone used ruote-on-rails?
2011-09-21 23:32:14 utc jmettraux pbh_: hello, do you have an issue with ruote-on-rail ? If nobody is replying here, feel free to post on the mailing list (http://groups.google.com/group/openwferu-users) or fill an issue report (https://github.com/tosch/ruote-on-rails/issues)
2011-09-21 23:44:30 utc pbh_ hey there
2011-09-21 23:44:53 utc pbh_ no issue, i was just a little confused as to the goal of the ruote-on-rails app
2011-09-21 23:45:10 utc pbh_ since it seems very similar to the default functionality of ruote-kit
2011-09-21 23:45:18 utc jmettraux OK. Well it's an example
2011-09-21 23:45:49 utc pbh_ but with a few modifications (forms via quaderno etc)
2011-09-21 23:46:51 utc pbh_ my other question is what the state of ruote testing is
2011-09-21 23:47:21 utc pbh_ the documentation just says "TODO", and i can't tell whether that means that there's easy test::unit support existing but not documented, or whether that means the support itself is TODO ;-)
2011-09-21 23:47:51 utc pbh_ (though the existing documentation is quite helpful)
2011-09-21 23:50:06 utc jmettraux OK, I should replace that TODO with something consistent
2011-09-21 23:50:17 utc jmettraux rspec and test/unit examples
2011-09-21 23:51:54 utc pbh_ (i did notice that ruote-cukes seems to exist)
2011-09-21 23:52:12 utc pbh_ are there any asserts, like assert_transitions_to_participant or something?
2011-09-21 23:58:32 utc jmettraux no, nothing like that
2011-09-21 23:59:23 utc jmettraux the only useful thing is Engine#wait_for(x)
2011-09-22 00:00:06 utc jmettraux where x can be a String (wfid (workflow identifier)), a Symbol (name of a participant), or an integer (number of workflow actions/msgs)
2011-09-22 00:00:06 utc pbh_ ah
2011-09-22 00:00:38 utc pbh_ ah, so if you want to test whether a workflow transitions to the right participant
2011-09-22 00:00:42 utc pbh_ you could wait_for(1)
2011-09-22 00:00:53 utc jmettraux not really
2011-09-22 00:01:15 utc jmettraux these are not "steps", they are "workflow messages processed"
2011-09-22 00:01:31 utc pbh_ ah, and it might be the case that there's some nonstandard number of replies or dispatches or the like?
2011-09-22 00:01:49 utc jmettraux Ruote.define { alpha }
2011-09-22 00:01:58 utc jmettraux will take 3 msgs to reach alpha
2011-09-22 00:02:17 utc jmettraux launch/apply workflow, then apply participant expression, then dispatch to alpha
2011-09-22 00:02:30 utc jmettraux engine.wait_for(:alpha) is probably what you want
2011-09-22 00:02:55 utc jmettraux example using engine.wait_for(wfid) (String): https://github.com/jmettraux/ruote/blob/master/test/functional/ft_70_take_and_discard_attributes.rb
2011-09-22 00:03:15 utc pbh_ does wait_for have a timeout or anything?
2011-09-22 00:03:51 utc pbh_ e.g., what happens in ruote-cukes when there is no transition to alpha and you do " Then the process should reach alpha" or when you do a wait_for and alpha is never transitioned to?
2011-09-22 00:04:03 utc pbh_ does it just wait forever?
2011-09-22 00:04:07 utc jmettraux yes
2011-09-22 00:04:29 utc jmettraux for my ci needs I wrap the whole in a timeout
2011-09-22 00:05:36 utc pbh_ ah, do you wrap each test or the full test suite with a timeout?
2011-09-22 00:06:16 utc pbh_ (and do you have a gist and/or github link as to what that looks like?)
2011-09-22 00:06:30 utc jmettraux http://ruote-ci.s3.amazonaws.com/ci.html
2011-09-22 00:06:38 utc jmettraux https://github.com/jmettraux/ci
2011-09-22 00:07:16 utc jmettraux in "ci" I'm simply using the ruby timeout stdlib thinggy
2011-09-22 00:07:41 utc jmettraux having the timeout specified when calling wait_for might be a good idea
2011-09-22 00:08:12 utc pbh_ i think you're probably right not to build it in
2011-09-22 00:08:55 utc pbh_ it seems like the more natural fix is to have a hook for when a participant transitions or a message occurs or what have you
2011-09-22 00:10:22 utc pbh_ (and in the meantime wrapping the wait_for call with a standard ruby timeout seems easy and robust enough, as long as you're not dealing with underlying storage that is prone to super long timeouts or anything)
2011-09-22 00:11:20 utc jmettraux true