ruote log_2010-10-15

2010-10-15 09:30:29 utc jmettraux rebo: hello, if there is no one answering your questions here, don't hesitate to fire questions at http://groups.google.com/group/openwferu-users
2010-10-15 09:31:40 utc rebo ok cool, ill do that
2010-10-15 13:38:37 utc jmettraux tosch_le_: hello
2010-10-15 13:38:45 utc tosch_le_ hello!
2010-10-15 13:41:18 utc jmettraux ready for the week-end ?
2010-10-15 13:42:00 utc tosch_le_ yes, nearly.
2010-10-15 13:42:25 utc tosch_le_ just doing some clean up of my ubuntu installation after upgrading to maverick
2010-10-15 13:42:34 utc jmettraux what's maverick ?
2010-10-15 13:42:48 utc jmettraux in three words ;-)
2010-10-15 13:42:50 utc tosch_le_ the latest ubuntu release
2010-10-15 13:42:57 utc jmettraux aaah, thanks
2010-10-15 13:42:58 utc tosch_le_ 4, sorry
2010-10-15 13:43:15 utc tosch_le_ "maverick meerkat"
2010-10-15 13:43:37 utc jmettraux ouch, ruby5 thing about ruote : http://ruby5.envylabs.com/episodes/122-episode-120-october-15th-2010?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+Ruby5+%28Ruby5%29&utm_content=Google+Reader
2010-10-15 13:44:27 utc tosch_le_ :-)
2010-10-15 13:45:45 utc jmettraux did you see that one http://groups.google.com/group/openwferu-users/browse_thread/thread/599cec25a2f13787 ?
2010-10-15 13:46:21 utc tosch_le_ yes, but i couldn't make much out of it
2010-10-15 13:46:35 utc jmettraux not much info :-(
2010-10-15 13:46:55 utc tosch_le_ i can't see why yajl should conflict with cucumber
2010-10-15 13:47:06 utc jmettraux me neither
2010-10-15 13:47:34 utc tosch_le_ http://github.com/brianmario/yajl-ruby/issues/issue/24
2010-10-15 13:49:10 utc jmettraux excellent, many thanks !!
2010-10-15 13:49:16 utc jmettraux I will push that to the mailing list
2010-10-15 13:51:15 utc jmettraux pushed
2010-10-15 13:57:42 utc gonzalo hi all, i would like to ask a question about ruote-kit... i would like to launch processes given just a number... instead of {:definition => blabla } something like {:id => 12} ... then, inside pdefs folder, process 12-lend-a-bike.rb should be read an launch
2010-10-15 13:58:07 utc gonzalo does ruote-kit have support for this?
2010-10-15 13:58:13 utc jmettraux hello
2010-10-15 13:58:16 utc gonzalo hi
2010-10-15 13:58:26 utc jmettraux it's easy to implement by yourself
2010-10-15 13:58:51 utc gonzalo ok, i'll add it
2010-10-15 13:59:24 utc gonzalo the logic should be in ruote-kit right
2010-10-15 13:59:33 utc gonzalo I should not touch ruote
2010-10-15 13:59:38 utc gonzalo i think...
2010-10-15 14:00:08 utc jmettraux or are you lauching now ?
2010-10-15 14:00:12 utc jmettraux a POST ?
2010-10-15 14:00:18 utc jmettraux or do you have a specific view ?
2010-10-15 14:01:07 utc gonzalo well it is for AMQP entities that should launch processes just with the id
2010-10-15 14:01:28 utc jmettraux you mean, the receiver does the launch ?
2010-10-15 14:01:39 utc gonzalo yes
2010-10-15 14:01:45 utc jmettraux the order comes via AMQP...
2010-10-15 14:01:56 utc gonzalo yes via AMQP
2010-10-15 14:01:59 utc jmettraux then you can modify the AMQP receiver or ruote itself
2010-10-15 14:02:39 utc gonzalo ok thanks john, I just had some doubts on where to put the code...
2010-10-15 14:03:22 utc gonzalo I think I want this behaviour to happen in everycase
2010-10-15 14:03:39 utc jmettraux if I were you, I'd first try to monkeypatch RuoteAMQP::Receiver
2010-10-15 14:03:39 utc gonzalo so maybe the best place to add this logic is the launch method
2010-10-15 14:03:43 utc jmettraux ok
2010-10-15 14:03:59 utc gonzalo umm ok
2010-10-15 14:04:13 utc gonzalo I'll do what you say
2010-10-15 14:04:35 utc jmettraux if you want it everywhere, then monkeypatching Ruote::Engine#launch is better
2010-10-15 14:04:41 utc jmettraux (your first idea)
2010-10-15 14:04:57 utc gonzalo cool! many thanks
2010-10-15 14:06:21 utc jmettraux http://github.com/jmettraux/ruote-amqp/blob/ruote2.1/lib/ruote-amqp/receiver.rb#L104 seems it wants a "definition" to consider it a launchitem though
2010-10-15 14:06:35 utc jmettraux definition => 12 maybe
2010-10-15 14:06:39 utc gonzalo yes
2010-10-15 14:06:44 utc jmettraux 'definition' => 12
2010-10-15 14:06:54 utc gonzalo i think i will check if it is a number
2010-10-15 14:07:01 utc gonzalo just search pdefs dir
2010-10-15 14:07:11 utc gonzalo and find 12-something_asdads.rb
2010-10-15 14:07:17 utc gonzalo and launch it
2010-10-15 14:07:20 utc jmettraux cool
2010-10-15 14:07:33 utc gonzalo that's fine for us
2010-10-15 14:08:14 utc gonzalo thanks for the tips
2010-10-15 14:08:20 utc gonzalo :P
2010-10-15 14:08:32 utc jmettraux you're welcome !
2010-10-15 14:19:47 utc gonzalo do you know what would be the best way for me to modify ruote for my ruote-kit project but still keep up to date with main development of ruote
2010-10-15 14:19:47 utc gonzalo ?
2010-10-15 14:20:12 utc gonzalo maybe i should just clone ruote in /libs and require 'ruote'
2010-10-15 14:20:24 utc gonzalo and remove ruote from ruote-kit's Gemfile
2010-10-15 14:20:31 utc tosch_le best way imho: fork away on github
2010-10-15 14:20:41 utc gonzalo and periodically do git pull to get updates
2010-10-15 14:20:42 utc gonzalo ok
2010-10-15 14:32:13 utc tosch_le best way imho: fork away on github
2010-10-15 14:32:17 utc tosch_le and merge the main development regularly
2010-10-15 14:33:07 utc jmettraux tosch_le: tomorrow I'll be working at the swiss stand in the local Oktoberfest organised by the German community
2010-10-15 14:34:08 utc tosch_le sounds fun. never was at the Oktoberfest myself, though
2010-10-15 14:34:24 utc tosch_le i just don't like drunken masses ;-)
2010-10-15 14:34:44 utc jmettraux it's something more like Bavarian ?
2010-10-15 14:36:29 utc tosch_le yes. far away from my home…
2010-10-15 14:37:28 utc jmettraux you're in Saxony, right ?
2010-10-15 14:37:31 utc tosch_le yes
2010-10-15 14:38:34 utc tosch_le gonzalo: as my internet connection dropped: did you decide how to do your (local) ruote modifications?
2010-10-15 14:44:11 utc gonzalo tosch_le: sorry I was talking to a co-worker. what i'm going to do is just git clone ruote and modify there whatever i need
2010-10-15 14:44:43 utc gonzalo then i just modify ruote-kit's Gemfile so the gem 'ruote' has a path that points to my local copy of ruote
2010-10-15 14:44:48 utc gonzalo does it make sense?
2010-10-15 14:45:09 utc tosch_le yes, it does.
2010-10-15 14:45:30 utc tosch_le bundler ftw
2010-10-15 14:45:57 utc tosch_le i'd create an own branch in the git repo, though
2010-10-15 14:46:08 utc tosch_le for easy updates
2010-10-15 14:47:26 utc gonzalo oh! yes, i should make a new branch too
2010-10-15 14:47:31 utc gonzalo thanks for the tip
2010-10-15 14:48:34 utc tosch_le and don't forget to study http://gembundler.com/git.html
2010-10-15 14:48:49 utc tosch_le but i suppose you've done that already
2010-10-15 14:51:35 utc gonzalo ok many thanks :P
2010-10-15 14:53:04 utc gonzalo i'm givin' a look at http://gembundler.com/man/gemfile.5.html
2010-10-15 15:10:09 utc lbt hmmm is there a "wait" and a corresponding "resume wfid"
2010-10-15 15:14:38 utc tosch_le lbt: sorry, i don't know. if jmettraux doesn't return, please post on the ml
2010-10-15 15:14:58 utc jmettraux ?
2010-10-15 15:15:41 utc jmettraux lbt: hello, do you have more context ?
2010-10-15 15:16:04 utc lbt sure...I have a process that hands off to a 'blocking' participant
2010-10-15 15:16:41 utc jmettraux reply_to_engine(workitem)
2010-10-15 15:16:52 utc lbt hmm
2010-10-15 15:17:04 utc jmettraux you should know that by now ;)
2010-10-15 15:17:13 utc lbt of course... I don't think of it like that
2010-10-15 15:17:45 utc lbt I want the participant to be able to cope with a crash
2010-10-15 15:18:01 utc lbt and then later trigger a continuation
2010-10-15 15:18:08 utc lbt so I simply have to store the wi
2010-10-15 15:18:09 utc jmettraux continuation ?
2010-10-15 15:18:18 utc lbt carry on running
2010-10-15 15:18:31 utc lbt sorry
2010-10-15 15:18:33 utc jmettraux reply_to_engine(workitem)
2010-10-15 15:18:37 utc lbt nothing to see here... move along :)
2010-10-15 15:18:53 utc lbt ACTION goes for a round of coffees ...
2010-10-15 15:18:57 utc jmettraux ok
2010-10-15 15:19:01 utc jmettraux enjoy
2010-10-15 15:19:35 utc tosch_le ACTION is tired, too, and will leave for the weekend now
2010-10-15 15:19:46 utc tosch_le bye!
2010-10-15 15:20:01 utc lbt o/ tosch_le
2010-10-15 15:20:46 utc lbt jmettraux: anyhow.... http://pastie.org/1213979
2010-10-15 15:21:04 utc lbt still a WIP ... but that's what we're looking at
2010-10-15 15:21:16 utc jmettraux looks complicated
2010-10-15 15:21:35 utc lbt it kinda is...
2010-10-15 15:21:54 utc lbt but essentially it's "check, build, test, promote"
2010-10-15 15:22:55 utc jmettraux keeping the URI, I'll try to come up with a simplification tomorrow
2010-10-15 15:23:32 utc lbt OK ... would be happy to hear feedback
2010-10-15 15:23:50 utc jmettraux crash and continuation, that's the default behaviour of ruote
2010-10-15 15:24:09 utc jmettraux a copy of the emitted workitem is kept in the participant expression
2010-10-15 15:24:34 utc jmettraux if a participant doesn't reply, you can re_apply http://ruote.rubyforge.org/process_administration.html
2010-10-15 15:24:38 utc jmettraux or cancel
2010-10-15 15:24:45 utc jmettraux or reply in its stead
2010-10-15 15:24:58 utc lbt yes... we have the r-k viewer running and intend to use it in those situations
2010-10-15 15:25:15 utc jmettraux you can also set a timeout
2010-10-15 15:25:24 utc lbt one facet of remote participant design should be that we need to understand how to deal with a failure
2010-10-15 15:26:13 utc jmettraux understood
2010-10-15 15:26:16 utc lbt nb... on a different subject: the _if expression has confused *everyone* I've shown it to...
2010-10-15 15:26:41 utc jmettraux so what's your suggestion ?
2010-10-15 15:27:37 utc lbt possibly an else_
2010-10-15 15:28:11 utc jmettraux it's already in there
2010-10-15 15:28:19 utc lbt _if do ..... end feels so much like an implicit sequence
2010-10-15 15:28:39 utc jmettraux http://ruote.rubyforge.org/exp/if.html
2010-10-15 15:29:23 utc jmettraux I tend to use do_that_thing :if => "xxx"
2010-10-15 15:30:33 utc lbt I mean http://pastie.org/1223619
2010-10-15 15:31:14 utc lbt syntactic sugar really
2010-10-15 15:31:42 utc jmettraux you would you express that in ruby ?
2010-10-15 15:33:14 utc lbt _else is almost a no-op/comment
2010-10-15 15:34:26 utc jmettraux http://gist.github.com/628398
2010-10-15 15:34:47 utc jmettraux remember, I don't parse, ruby is used to build an AST
2010-10-15 15:35:22 utc lbt true...
2010-10-15 15:35:49 utc lbt I guess it's not obvious that these are equivalent http://pastie.org/1223619
2010-10-15 15:36:18 utc jmettraux true
2010-10-15 15:36:39 utc lbt maybe add those # comments into the docs? ....
2010-10-15 15:37:35 utc lbt I just noticed it's the only place I've had to explain/excuse the syntax :)
2010-10-15 15:37:43 utc jmettraux great
2010-10-15 15:38:26 utc jmettraux noone has ever done lisp/scheme in your team ?
2010-10-15 15:38:32 utc lbt me :)
2010-10-15 15:38:48 utc lbt damned vi users!
2010-10-15 15:39:04 utc jmettraux :) emacs ?
2010-10-15 15:39:24 utc lbt you know... I didn't realise why I just understood it and no-one else did.... that's probably it
2010-10-15 15:39:34 utc lbt yes, that's my church
2010-10-15 15:40:22 utc jmettraux I'm almost finished with a "registerp" expression
2010-10-15 15:40:56 utc lbt I was reviewing that gist earlier
2010-10-15 15:41:00 utc jmettraux I will think about something, I hate having "_if"
2010-10-15 15:42:26 utc lbt :)
2010-10-15 15:43:34 utc jmettraux http://github.com/jmettraux/ruote/commit/8386dba6964df685843b35215ff84e60d07d579d
2010-10-15 15:46:18 utc jmettraux ok, escaping, I'll go on with that tomorrow (unregisterp)
2010-10-15 15:46:23 utc lbt remind me... is the positional registration controlled by the args
2010-10-15 15:46:33 utc jmettraux :position => x
2010-10-15 15:46:58 utc jmettraux yes, tell me if it's too "blind" for you
2010-10-15 15:47:20 utc lbt I did wonder about some kind of 'absolute' section
2010-10-15 15:47:26 utc jmettraux I don't mind dumping that work (any work)
2010-10-15 15:47:29 utc lbt which goes before regexps
2010-10-15 15:49:03 utc jmettraux it's crazy that we have to discuss that, since with rabbitmq in the middle... rabbitmq is the participant list (somehow)
2010-10-15 15:50:24 utc jmettraux I'd say, if it's an absolute, put it at :position => 0
2010-10-15 15:50:37 utc jmettraux if it's a regexp, put it at position => -2
2010-10-15 15:50:38 utc lbt that's fair
2010-10-15 15:50:45 utc jmettraux -1 it the home of the "catchall"
2010-10-15 15:50:52 utc jmettraux if any
2010-10-15 15:51:01 utc lbt good policy
2010-10-15 15:51:16 utc jmettraux I have to go
2010-10-15 15:51:25 utc lbt g'night o/
2010-10-15 15:51:51 utc jmettraux if you have more reflexions / ideas / etc ==> mailing list
2010-10-15 15:52:00 utc lbt OK
2010-10-15 15:52:03 utc jmettraux I have no access to the IRC logs during the week-end
2010-10-15 15:52:24 utc jmettraux especially if you could tell me if "registerp" is OK for you
2010-10-15 15:52:41 utc lbt it certainly looks like it
2010-10-15 15:52:45 utc jmettraux I'll have a look at your process definition and come up with suggestions
2010-10-15 15:52:47 utc jmettraux ciao !