ruote log_2010-02-09

2010-02-09 06:41:17 utc anb_ Hello everyone
2010-02-09 06:41:25 utc jmettraux hello !
2010-02-09 07:08:45 utc anb_ John, you found out if you wanted to drop BlockParticipants ?
2010-02-09 07:16:09 utc jmettraux anb_: not sure yet
2010-02-09 07:16:28 utc jmettraux it's good for tutorials and quickstarts
2010-02-09 07:16:52 utc jmettraux anyway, it's easy to refactor a BlockParticipant into a MyClassNameParticiapnt
2010-02-09 07:16:55 utc jmettraux cipant
2010-02-09 07:17:20 utc anb_ agree. can be nice for debugging too
2010-02-09 07:19:56 utc jmettraux ruby 1.9.x has a nice block.source_location, it returns something like [ 'path/to/src.rb', 129 ]
2010-02-09 07:20:08 utc jmettraux it would be fun to "extract" the source of the block
2010-02-09 07:20:26 utc jmettraux but 1.9.x only
2010-02-09 07:21:55 utc jmettraux I have to fix the bug I mentioned yesterday
2010-02-09 07:30:04 utc anb_ and I guess many people are still on 1.8
2010-02-09 07:30:25 utc jmettraux and their hosting services as well
2010-02-09 07:33:47 utc anb_ true
2010-02-09 07:36:44 utc anb_ I have a question regarding workitems. I have a participant that does a storage.put of the workitem and does not reply to the engine. I want the reply to go through a ruote-kit PUT of the updated workitem. But for my understanding, what does make the workflow proceed ? is there some field in the workitem that rk updates?
2010-02-09 07:37:43 utc jmettraux http://github.com/kennethkalmer/ruote-kit/blob/master/lib/ruote-kit/resources/workitems.rb#L48-68
2010-02-09 07:37:48 utc tosch_le hi! specs are running amoc when using Ruote::StorageParticipant.new in ruote-kit :-(
2010-02-09 07:37:50 utc tosch_le No description supplied for example group declared on /home/tsc/git/ruote-kit/vendor/gems/ruby/1.8/gems/ruote-2.1.6/lib/ruote/part/storage_participant.rb:46:in `initialize'
2010-02-09 07:37:53 utc jmettraux just the "proceed" option
2010-02-09 07:38:03 utc anb_ so it's the call to store_participant.reply
2010-02-09 07:38:08 utc jmettraux yes
2010-02-09 07:38:29 utc jmettraux Hi tosch_le : I guess you'll have to ping kennethkalmer for that one
2010-02-09 07:38:52 utc anb_ thanks for the pointers john
2010-02-09 07:39:01 utc jmettraux anb_: you're welcome
2010-02-09 07:39:29 utc tosch_le i suppose it's rspec's fault. somehow, it doesn't like @context...
2010-02-09 07:39:46 utc jmettraux tosch_le: argh
2010-02-09 07:50:16 utc jmettraux tosch_le: those specs were added by Kenneth ?
2010-02-09 07:51:09 utc tosch_le no. rspec complains about a missing spec description in a normal ruote method.
2010-02-09 07:51:41 utc tosch_le and yes, i just tried to run the original specs by kenneth.
2010-02-09 07:51:55 utc jmettraux so they were working for him
2010-02-09 07:52:02 utc jmettraux supposedly
2010-02-09 07:52:39 utc tosch_le supposedly.
2010-02-09 07:53:40 utc jmettraux cloning
2010-02-09 07:54:08 utc jmettraux 1.8.7p72 for me
2010-02-09 07:55:50 utc jmettraux FFFFFFFFFFF-fest ?
2010-02-09 07:56:30 utc jmettraux same thing here
2010-02-09 07:56:45 utc tosch_le https://rspec.lighthouseapp.com/projects/5645/tickets/927-rspec-breaks-the-context-methods-on-my-objects
2010-02-09 07:57:02 utc tosch_le it seems to be a rspec bug
2010-02-09 08:09:49 utc jmettraux tosch_le: I tried to replace @context = with self.context, but same problem
2010-02-09 08:10:53 utc tosch_le even renaming @context to @my_context doesn't help.
2010-02-09 08:11:06 utc jmettraux :)
2010-02-09 08:11:46 utc tosch_le rspec has a serious problem with this...
2010-02-09 08:23:12 utc tosch_le if you're out there some time, Kenneth: I've pushed some changes to my rk fork regarding the storage participant integration.
2010-02-09 08:23:33 utc tosch_le (http://github.com/tosch/ruote-kit/commit/9e63babe77befc5a77aa7e70e83c5c0bf25d891b)
2010-02-09 08:39:23 utc jmettraux tosch_le: have you tried re-installing bundler ?
2010-02-09 08:40:08 utc tosch_le is your answer related to my question on #rvm? or related to the rspec bug?
2010-02-09 08:40:44 utc jmettraux rvm related
2010-02-09 08:41:00 utc tosch_le i just installed it
2010-02-09 08:41:59 utc tosch_le but thanks for the tip, will uninstall the system's bundler installation and try again
2010-02-09 08:42:07 utc tosch_le s/tip/hint/
2010-02-09 08:42:15 utc jmettraux not sure it will work
2010-02-09 08:46:58 utc jmettraux ok
2010-02-09 08:48:53 utc tosch_le ACTION should rtfm first
2010-02-09 08:49:24 utc jmettraux ruote-kit's README say "gem bundle"
2010-02-09 08:49:26 utc jmettraux and it works
2010-02-09 08:51:09 utc tosch_le depends on the version of bundler installed, i suppose
2010-02-09 08:51:31 utc tosch_le tried the recent 0.9.3 and gem bundle failed
2010-02-09 08:51:37 utc jmettraux 0.8.1
2010-02-09 08:51:41 utc tosch_le but bundle install did the trick
2010-02-09 08:51:49 utc jmettraux ok
2010-02-09 08:52:02 utc jmettraux maybe it's time to patch the README ;)
2010-02-09 08:53:50 utc anb_ I had problems with the gem bundle too :)
2010-02-09 08:53:57 utc anb_ yesterday
2010-02-09 08:54:05 utc tosch_le not only the README, but also the gemfile
2010-02-09 08:54:24 utc tosch_le a source entry is seriously missing...
2010-02-09 08:55:38 utc jmettraux have to escape... ttyl !
2010-02-09 08:55:45 utc jmettraux good job on ruote-kit !
2010-02-09 09:11:14 utc kennethkalmer ACTION wipes sweat from his brows
2010-02-09 09:11:18 utc kennethkalmer hi everyone
2010-02-09 09:11:30 utc kennethkalmer thanks tosch_le for the commit
2010-02-09 09:11:52 utc tosch_le np
2010-02-09 09:12:10 utc tosch_le wait for a second one (Gemfile and Readme update to bundler 0.9.x)
2010-02-09 09:12:40 utc kennethkalmer i know, that is on my todo list
2010-02-09 09:12:48 utc kennethkalmer bumped my head with bundler last night as well
2010-02-09 09:18:46 utc tosch_le pull request sent, but i suppose the readme needs some more updates (bundler won't use vendor/gems anymore)
2010-02-09 09:25:30 utc kennethkalmer i'll work through the bundler updates thanks !
2010-02-09 09:31:25 utc anb_ hi kenneth
2010-02-09 09:31:34 utc kennethkalmer hi anb_
2010-02-09 09:43:32 utc tosch_le wow, ruby 1.9.1 is way faster than ubuntu's default 1.8.7 package...
2010-02-09 09:44:33 utc tosch_le kenneth: specs on ruote-kit are failing due to a rspec bug with classes having a context method (like in Ruote::StorageParticipant)
2010-02-09 09:44:57 utc kennethkalmer argh
2010-02-09 09:46:06 utc tosch_le yeah, that's the right expression. it's (looking up the right word) tremendous
2010-02-09 09:46:37 utc tosch_le https://rspec.lighthouseapp.com/projects/5645/tickets/927-rspec-breaks-the-context-methods-on-my-objects
2010-02-09 09:59:14 utc tosch_le kenneth: is there an important reason for pinning ruote's version to 2.1.4 in the jeweler task of rk?
2010-02-09 10:21:57 utc kennethkalmer not at all, I should make it >= 2.1.6
2010-02-09 10:26:42 utc tosch_le with the new RuoteKit.storage_participant, the catchall participant isn't required for using the workitems resource.
2010-02-09 10:26:54 utc tosch_le i've pushed some changes to my fork
2010-02-09 10:27:55 utc tosch_le (that's why i love the Ruote::StorageParticipant.new(engine) syntax: you're truly independent of the storage participants' names)
2010-02-09 10:29:29 utc kennethkalmer tosch_le: you are on fire :)
2010-02-09 10:34:10 utc tosch_le maybe, but when calling _ruote/workitems from within rails, i get an Internal Server Error
2010-02-09 10:34:20 utc tosch_le does rk some logging anywhere?
2010-02-09 10:35:02 utc tosch_le rails is quite un-noisy...
2010-02-09 10:40:29 utc anb_ tosch_le I just had problems calling GET on _ruote/workitems and it was because I used a trailing slash
2010-02-09 10:40:59 utc tosch_le no trailing slash here
2010-02-09 10:41:52 utc tosch_le there's some other problem i built in i suppose. i'm re-enabling the error raising and try again
2010-02-09 10:42:32 utc anb_ oh I also noticed that workitems isn't available if you don't register the catchall
2010-02-09 10:42:56 utc tosch_le yeah and that should be necessary with the updates i made in my rk fork
2010-02-09 10:43:05 utc tosch_le shouldn't
2010-02-09 10:51:12 utc tosch_le ah, the old ruote version was the culprit.
2010-02-09 10:51:26 utc tosch_le when using ruote 2.1.6, everything works fine.
2010-02-09 10:53:30 utc kennethkalmer apologies for being so distant guys, in the middle of shit storm here :/
2010-02-09 10:54:00 utc kennethkalmer tosch_le: i'm hoping to get the basics of /participants & /errors in place this week as well
2010-02-09 10:54:26 utc tosch_le shit storm? sounds amazing :-) but never mind, i'll get along.
2010-02-09 10:54:38 utc kennethkalmer thanks, appreciate it :)
2010-02-09 11:42:00 utc tosch_le http://github.com/tosch/ruote-on-rails
2010-02-09 11:44:04 utc jmettraux very nice !!
2010-02-09 11:44:51 utc kennethkalmer +1
2010-02-09 11:45:04 utc jmettraux http://twitter.com/jmettraux/status/8851293327
2010-02-09 11:47:50 utc jmettraux tosch_le : rails3 ?
2010-02-09 11:48:19 utc tosch_le no, should be rails 2.3.5
2010-02-09 11:48:44 utc tosch_le not familiar with the new rails yet
2010-02-09 11:52:44 utc tosch_le any hints for the example app are welcome
2010-02-09 11:53:05 utc tosch_le out for launch, ttyl
2010-02-09 12:03:30 utc jmettraux happy rocket trip
2010-02-09 12:56:00 utc jmettraux linked : http://github.com/jmettraux/ruote_website/commit/4085993ccbc944433b9875be432543579982457a
2010-02-09 13:11:53 utc anb_ jmettraux, last question before you go to bed :) -> process variables cannot be found in workitems ?
2010-02-09 13:12:05 utc jmettraux no
2010-02-09 13:12:09 utc anb_ sniff
2010-02-09 13:12:10 utc jmettraux they are engine bound
2010-02-09 13:12:33 utc jmettraux workitem payload != engine variables
2010-02-09 13:12:53 utc jmettraux you have to explicitely copy them in the workitem if you want to make them available
2010-02-09 13:13:30 utc anb_ oki, thx
2010-02-09 13:13:30 utc jmettraux does jbpm have a token payload ? (curious suddenly)
2010-02-09 13:14:07 utc anb_ I don't think so, but you have two variable scope, process and task
2010-02-09 13:14:31 utc jmettraux seems similar then
2010-02-09 13:14:47 utc anb_ yes
2010-02-09 13:15:48 utc jmettraux by using http://ruote.rubyforge.org/dollar.html#deep
2010-02-09 13:16:02 utc jmettraux you could have all the process vars in one "container"
2010-02-09 13:16:22 utc jmettraux it's easy to copy that in one command into the workitem
2010-02-09 13:16:59 utc anb_ that's a possibility
2010-02-09 13:17:16 utc anb_ but when does the dollar substitution happen ?
2010-02-09 13:17:39 utc jmettraux when an expression gets executed and reads its attributes
2010-02-09 13:17:43 utc jmettraux runtime
2010-02-09 13:17:49 utc anb_ ok
2010-02-09 13:18:14 utc jmettraux maybe I should just shut up, I don't know why you asked about workitem not seeing variables
2010-02-09 13:18:28 utc jmettraux I don't want to confuse you
2010-02-09 13:19:22 utc anb_ hehe don't worry :)
2010-02-09 13:20:00 utc anb_ well I just need some variables to be seen by all participants and I was investigating using process variables
2010-02-09 13:21:30 utc jmettraux if your participant implementation has att_accessor :context
2010-02-09 13:21:49 utc jmettraux the worker will set it to the worker/engine context at dispatch time
2010-02-09 13:21:55 utc jmettraux you can place vars in that
2010-02-09 13:22:11 utc jmettraux but beware of nuking vars used by the engine/worker
2010-02-09 13:23:36 utc anb_ I'll test what works best for us
2010-02-09 13:23:39 utc anb_ thanks for the suggestion
2010-02-09 13:24:01 utc jmettraux your suggestions are welcome as well
2010-02-09 13:25:59 utc anb_ well sometimes I feel that I'm trying to reproduce one to one what we were doing with jbpm... and that's not necessarily the best way
2010-02-09 13:26:14 utc jmettraux +1
2010-02-09 13:26:37 utc anb_ but it's hard to get rid of old habits
2010-02-09 13:27:39 utc tosch_le +1 ;-)
2010-02-09 13:27:50 utc tosch_le jmettraux: thanks for the links
2010-02-09 13:28:02 utc jmettraux tosch_le: you're welcome
2010-02-09 13:29:00 utc jmettraux tosch_le: thanks for all your work !
2010-02-09 13:33:54 utc anb_ john, I think I like the deep copy into workitem solution
2010-02-09 13:34:57 utc anb_ but what happens when I want to copy back from the workitem to the process vars ?
2010-02-09 13:35:45 utc jmettraux inverse thing
2010-02-09 13:36:09 utc jmettraux set :var => 'x', :field_value => 'fieldname'
2010-02-09 13:36:59 utc anb_ all right
2010-02-09 13:51:11 utc anb_ when I substitute the entire Hash using $, it gives me the key/values concatenated instead of a proper hash/dict. see what I mean ?
2010-02-09 13:51:23 utc jmettraux yes
2010-02-09 13:51:36 utc jmettraux that's why you have to avoid dollar substitution
2010-02-09 13:52:01 utc jmettraux set :field => 'f', :var => 'x'
2010-02-09 13:52:06 utc jmettraux instead of
2010-02-09 13:52:35 utc jmettraux set :field => 'f', :value => '${v:x}'
2010-02-09 13:52:41 utc anb_ great
2010-02-09 13:52:44 utc anb_ understood
2010-02-09 13:53:07 utc jmettraux :)
2010-02-09 13:55:10 utc jmettraux wayneeseguin: thanks for the RT
2010-02-09 13:56:01 utc wayneeseguin jmettraux: of course!
2010-02-09 13:56:05 utc wayneeseguin ruote is awesome! :)
2010-02-09 13:56:17 utc jmettraux oh thanks !
2010-02-09 13:56:29 utc jmettraux not as awesome and useful as rvm !!
2010-02-09 13:56:43 utc wayneeseguin lol
2010-02-09 13:56:53 utc wayneeseguin People just aren't as exposed is all! :)
2010-02-09 13:57:23 utc jmettraux how is the post-railscast load ?
2010-02-09 14:01:24 utc wayneeseguin insane :/
2010-02-09 14:01:29 utc wayneeseguin I put dayjob at risk yesterday
2010-02-09 14:01:37 utc jmettraux ouch
2010-02-09 14:02:53 utc wayneeseguin well sortof
2010-02-09 14:03:05 utc wayneeseguin the dayjob needs me badly so not really
2010-02-09 14:03:11 utc wayneeseguin but that's how much time I spent on it
2010-02-09 14:06:07 utc jmettraux take care
2010-02-09 14:15:10 utc tosch_le jmettraux: what's the best way to set the remote_definition_allowed option?
2010-02-09 14:15:12 utc tosch_le engine_instance.context['remote_definition_allowed'] = true
2010-02-09 14:15:14 utc tosch_le ?
2010-02-09 14:16:06 utc jmettraux $engine = Ruote::Engine.new(Ruote::Worker.new(Ruote::FsStorage.new('work')))
2010-02-09 14:16:09 utc jmettraux wait
2010-02-09 14:16:41 utc jmettraux $engine = Ruote::Engine.new(Ruote::Worker.new(Ruote::FsStorage.new('work', :remote_definition_allowed => true)))
2010-02-09 14:16:57 utc jmettraux or by setting it directly in the storage
2010-02-09 14:17:17 utc jmettraux conf = storage.get('configurations', 'engine')
2010-02-09 14:17:31 utc jmettraux conf['remote_definition_allowed'] = true
2010-02-09 14:17:38 utc jmettraux storage.put(conf)
2010-02-09 14:18:08 utc tosch_le puh. that's not done in a hurry...
2010-02-09 14:19:32 utc tosch_le i'm writing some lines for the ml on ruote on rails and mention launching workflow definitions served via http. i'd like to add a pointer to the steps needed to get that going.
2010-02-09 14:19:49 utc jmettraux ok
2010-02-09 14:20:24 utc tosch_le setting the config option at engine instantiation time is not that easy as this is done by rk.
2010-02-09 14:20:54 utc jmettraux how can I help ?
2010-02-09 14:20:59 utc tosch_le so either changes to rk or monkey patching is required. not really cute.
2010-02-09 14:21:27 utc tosch_le what about a helper method within the engine to set that option?
2010-02-09 14:21:46 utc jmettraux ...
2010-02-09 14:21:48 utc jmettraux it's local
2010-02-09 14:22:10 utc jmettraux so I guess that engine.context['remote_definition_allowed'] = true is OK
2010-02-09 14:22:20 utc jmettraux though
2010-02-09 14:22:33 utc jmettraux it only affects the engine instance
2010-02-09 14:23:09 utc jmettraux if, inside a process definition, there is subprocess :ref => 'http://xxx'
2010-02-09 14:23:23 utc tosch_le just wanted to mention that
2010-02-09 14:23:32 utc jmettraux and an 'other' worker executes, it won't see the remote_definition_allowed
2010-02-09 14:24:37 utc jmettraux the most elegant way, IMHO, is at engine instantiation time
2010-02-09 14:24:48 utc jmettraux I'd change RK
2010-02-09 14:26:54 utc tosch_le ok. kenneth should consider that.
2010-02-09 14:26:56 utc tosch_le or you could clutter ruote's engine.rb: http://gist.github.com/299235
2010-02-09 14:26:58 utc tosch_le ;-)
2010-02-09 14:27:33 utc jmettraux we could make it geneirc
2010-02-09 14:27:35 utc jmettraux generic
2010-02-09 14:27:54 utc jmettraux def configure (key, value)
2010-02-09 14:28:05 utc jmettraux engine.configure('remote_definition_allowed', true)
2010-02-09 14:28:07 utc tosch_le that's a great idea.
2010-02-09 14:28:20 utc jmettraux engine.configure('ruby_eval_allowed', true)
2010-02-09 14:28:36 utc jmettraux waiting for the pull request
2010-02-09 14:28:48 utc tosch_le ok. give me some minutes
2010-02-09 14:28:55 utc jmettraux take your time
2010-02-09 14:29:04 utc jmettraux tomorrow morning GMT +09
2010-02-09 14:30:01 utc tosch_le oh yeah, it's damn late at your side...
2010-02-09 14:41:02 utc tosch_le do you think this method needs a test? would it be a unit or a functional test?
2010-02-09 14:42:19 utc jmettraux functional since it requires a running engine
2010-02-09 14:49:09 utc tosch_le request sent, though test-less
2010-02-09 14:51:18 utc jmettraux OK, I will add one tomorrow morning
2010-02-09 14:51:39 utc tosch_le thanks.
2010-02-09 14:51:40 utc jmettraux many thanks !
2010-02-09 14:57:26 utc jmettraux bye !