ruote log_2010-05-26

2010-05-26 06:43:00 utc lbt jmettraux: thanks :) I'm looking at rabbitmq as the core for an integration automation project. Possibly looking at ruote and ruote-amqp as workflow
2010-05-26 07:02:53 utc jmettraux lbt: ah, nice. Not sure about the status of ruote-amqp these days though
2010-05-26 07:03:44 utc lbt ah :
2010-05-26 07:04:07 utc lbt well, I'll look at it (and the amount of code in r-a) this morning
2010-05-26 07:04:33 utc jmettraux oh cool
2010-05-26 07:05:10 utc lbt essentially I'm looking to automate a CI build process with build iteration cycles, full image builds, automated tests and publication
2010-05-26 07:05:20 utc lbt most functionality is in pre-existing systems
2010-05-26 07:05:38 utc jmettraux very nice
2010-05-26 07:05:42 utc lbt so it is about saying "when the build finishes make an image"
2010-05-26 07:05:54 utc lbt when the image is ready; send to test system
2010-05-26 07:06:38 utc lbt the idea is to use rabitmq to accept the events and trigger wf in ruote
2010-05-26 07:06:42 utc lbt hence ra
2010-05-26 07:06:54 utc jmettraux ok
2010-05-26 07:09:19 utc tosch_le hi!
2010-05-26 07:09:30 utc tosch_le ruote-amqp seems to be ok for kenneth: http://ruote.rubyforge.org/irclogs/ruote_2010-05-04.txt
2010-05-26 07:10:20 utc lbt tosch_le: excellent - that's reassuring :)
2010-05-26 07:14:13 utc lbt not many people here but the nice welcome more than makes up for it ;)
2010-05-26 07:15:05 utc jmettraux hey, it's a just a small project
2010-05-26 07:15:36 utc lbt :D
2010-05-26 07:16:23 utc lbt for now...
2010-05-26 07:16:34 utc tosch_le unfortunately kenneth isn't around, he could have shared his experiences with ruote and amqp
2010-05-26 07:17:12 utc jmettraux three_tee as well ( http://github.com/threetee/ruote-rails-example )
2010-05-26 07:17:24 utc lbt meh... digging in is the best way to learn... I got rabbitmq/python talking yesterday....
2010-05-26 07:17:33 utc lbt if only I knew ruby better...
2010-05-26 07:18:43 utc jmettraux no worries, it isn't that hard :)
2010-05-26 07:19:34 utc lbt it's used in the suse OBS which is the build system in question
2010-05-26 07:20:38 utc jmettraux are you working for a big software company ?
2010-05-26 07:20:58 utc lbt :)
2010-05-26 07:21:02 utc lbt meego
2010-05-26 07:21:11 utc lbt intel/nokia
2010-05-26 07:21:28 utc jmettraux very nice !
2010-05-26 07:22:12 utc lbt *nod* .... it's a great project... have you heard of it?
2010-05-26 07:22:41 utc jmettraux sorry m(_ _)m ... no
2010-05-26 07:22:50 utc jmettraux just browsed it
2010-05-26 07:23:47 utc lbt *nod* ... it's nice to have linux on your phone :)
2010-05-26 07:23:56 utc jmettraux +1
2010-05-26 07:24:30 utc jmettraux tosch_le: good morning btw, seen my email answer ?
2010-05-26 07:24:44 utc lbt biab
2010-05-26 07:24:53 utc tosch_le yeah, seen it, thanks.
2010-05-26 07:25:14 utc tosch_le just preparing an answer
2010-05-26 08:06:55 utc jmettraux sorry for bothering about the list, I think it's very important to keep all the conversations open
2010-05-26 08:07:59 utc tosch_le it's ok. wasn't sure about the list because it's a very specific question, but i think you're right.
2010-05-26 08:10:52 utc jmettraux :)
2010-05-26 08:11:58 utc tosch_le filtering workitems by field values works now in rk
2010-05-26 08:12:07 utc jmettraux very cool !!
2010-05-26 08:12:09 utc tosch_le have to update rk-client, though
2010-05-26 08:12:09 utc jmettraux woot !
2010-05-26 08:12:56 utc tosch_le thanks for the JSON hint
2010-05-26 08:13:31 utc tosch_le that does the trick even without "type" component
2010-05-26 08:13:35 utc jmettraux just sent an answer, you're welcome
2010-05-26 08:16:20 utc tosch_le thanks for the answer
2010-05-26 08:17:24 utc tosch_le switching from json to rufus-json will be the next entry on my rk todo list
2010-05-26 08:18:05 utc jmettraux not much work I hope
2010-05-26 08:20:02 utc tosch_le just search & replace, i expect
2010-05-26 08:20:13 utc jmettraux +1
2010-05-26 08:20:29 utc jmettraux o.to_json ~~> Rufus::Json.encode(o)
2010-05-26 08:21:00 utc tosch_le my only worry is the Gemfile
2010-05-26 08:21:08 utc jmettraux ?
2010-05-26 08:21:40 utc tosch_le if rufus-json expects an already required backend, i'll have to put a "gem 'json'" line in there
2010-05-26 08:21:55 utc jmettraux gem 'yajl-ruby'
2010-05-26 08:22:14 utc tosch_le yeah, but yajl-ruby won't run on all rubies, won't it?
2010-05-26 08:22:27 utc jmettraux indeed
2010-05-26 08:22:41 utc tosch_le i wouldn't use such a dependency as default
2010-05-26 08:24:08 utc jmettraux what about a Gemfile that'd look (fragment) like http://gist.github.com/414220
2010-05-26 08:25:16 utc jmettraux maybe with something that says "make sure one of them is uncommented"
2010-05-26 08:26:29 utc tosch_le hmm, here we are with the 1001 ways to use rk: when using as gem, the gemfile isn't to be edited
2010-05-26 08:27:26 utc jmettraux sorry, only using it from rails3's Gemfile
2010-05-26 08:27:32 utc jmettraux gem 'ruote-kit', :git => 'http://github.com/tosch/ruote-kit.git'
2010-05-26 08:27:44 utc jmettraux ;)
2010-05-26 08:28:22 utc tosch_le couldn't rufus-json try to load some backends?
2010-05-26 08:28:31 utc jmettraux it does
2010-05-26 08:28:45 utc jmettraux it tries yajl, json then activesupport
2010-05-26 08:29:05 utc tosch_le yeah, but it doesn't the require itself, does it?
2010-05-26 08:29:17 utc jmettraux no, it doesn't
2010-05-26 08:31:03 utc tosch_le i'd like to have rk running out of the box, without configuration hassle
2010-05-26 08:31:38 utc jmettraux so the blocker is 'yajl-ruby' not working on jruby ?
2010-05-26 08:32:08 utc tosch_le yes. that seems strange to me, i'm not even using jruby...
2010-05-26 08:32:21 utc tosch_le i'd like to ask kenneth about that.
2010-05-26 08:32:34 utc jmettraux is the Gemfile really used in "gem mode" ?
2010-05-26 08:32:52 utc tosch_le yes, bundler should respect the gemfile
2010-05-26 08:33:04 utc tosch_le and since gem loading is done by bundler...
2010-05-26 08:33:04 utc jmettraux true
2010-05-26 08:33:27 utc jmettraux I'd be for having gem 'yajl-ruby' around
2010-05-26 08:33:32 utc tosch_le what about that: i don't include a gem 'json' whatever line in the gemfile
2010-05-26 08:33:33 utc jmettraux jruby people are seldom
2010-05-26 08:33:40 utc tosch_le but some commented lines in there
2010-05-26 08:33:49 utc jmettraux +1
2010-05-26 08:34:03 utc tosch_le i do include a require 'json' in the config.ru, though
2010-05-26 08:34:05 utc jmettraux (that's what I was recommending before, no ? ;-) 9
2010-05-26 08:34:16 utc tosch_le yeah :-)
2010-05-26 08:34:16 utc jmettraux +1 +1
2010-05-26 08:34:23 utc jmettraux config.ru ftw
2010-05-26 08:34:47 utc tosch_le that way, 'rackup' runs out of the box (if json is installed)
2010-05-26 08:35:04 utc jmettraux great
2010-05-26 08:35:12 utc jmettraux (and if json is not borked)
2010-05-26 08:35:16 utc tosch_le ruote-on-rails works, too, only when using an own gemfile project, some work is needed
2010-05-26 08:35:44 utc tosch_le yeah, stepped about that yesterday. there should be big red warning signs around json 1.4.3
2010-05-26 08:36:15 utc jmettraux I'd love to send a bug report, but I couldn't isolate the thing
2010-05-26 08:36:17 utc jmettraux :(
2010-05-26 08:36:22 utc tosch_le thankfully i remembered the posts about that on the ruote ml so that i did find the culprit soon
2010-05-26 08:36:41 utc jmettraux how are your work projects doing ?
2010-05-26 08:36:48 utc tosch_le there's already an issue in json's github tracker
2010-05-26 08:36:55 utc jmettraux great !
2010-05-26 08:37:49 utc tosch_le doing some progress with my only ruote related project, but still far away from using ruote to its limits ;-)
2010-05-26 08:38:09 utc tosch_le and i'm still having to do far too much php these days...
2010-05-26 08:38:16 utc tosch_le and yours?
2010-05-26 08:38:28 utc jmettraux doing well
2010-05-26 08:38:40 utc jmettraux jammed with requirement changes
2010-05-26 08:38:43 utc jmettraux but surfacing
2010-05-26 08:58:46 utc jmettraux I want to release ruote 2.1.10 very soon
2010-05-26 08:58:50 utc jmettraux have to find some time
2010-05-26 09:03:14 utc tosch_le wishlist_for_rufus-json << 'there should be one error class for parser errors'
2010-05-26 09:03:48 utc jmettraux OK
2010-05-26 09:05:02 utc jmettraux http://github.com/jmettraux/rufus-json/issues/issue/1 let me meditate
2010-05-26 09:06:52 utc jmettraux biab
2010-05-26 10:08:48 utc jmettraux tosch_le: I'm sorry, I've been thinking as I was walking, but why do you need a unique exception for parse (or encoding errors) ?
2010-05-26 10:12:23 utc tosch_le for easy recovery from parser errors. i'd like to use the json encoding only as soft option. sure, checking for '{' at the beginning and '}' at the ending will be sufficient in the most cases, but i know there may be stupid users
2010-05-26 10:12:49 utc jmettraux you can catch any exception
2010-05-26 10:12:54 utc jmettraux (IMHO)
2010-05-26 10:12:56 utc tosch_le i know.
2010-05-26 10:13:06 utc tosch_le but that has an ugly code smell
2010-05-26 10:13:25 utc jmettraux I don't think so, sorry
2010-05-26 10:13:33 utc jmettraux try to convince me ;)
2010-05-26 10:14:09 utc tosch_le i've seen a nice article about that somewhere... wait, i'll see if i find it again
2010-05-26 10:14:45 utc jmettraux I'm afraid of diluting the exception message
2010-05-26 10:15:01 utc jmettraux and average users are already extremely poor at trace reading
2010-05-26 10:15:27 utc tosch_le you could pass the exception message on
2010-05-26 10:16:29 utc tosch_le i'll fork rufus-json and show you... later ;-)
2010-05-26 10:20:28 utc tosch_le ok, i don't find that article.
2010-05-26 10:21:26 utc tosch_le in short: i like to recover only from errors i know that it's ok to recover from.
2010-05-26 10:22:03 utc tosch_le there may be some error while json decoding which is not desirable to be rescued from
2010-05-26 10:22:49 utc tosch_le in that case, parser errors are ok, i just pass the strings on, then.
2010-05-26 10:23:20 utc jmettraux this makes sense then
2010-05-26 10:24:04 utc tosch_le other errors, which i don't know yet, may be not ok. the user should get an error message so that he may repair things by himself. if i rescue from every exception, the user will be left helpless
2010-05-26 10:24:38 utc jmettraux you're only wrapping a parse operation
2010-05-26 10:24:51 utc jmettraux parse(File.read(x)) is of course trickier
2010-05-26 10:25:01 utc jmettraux parse("...") is another thing
2010-05-26 10:25:22 utc jmettraux you know you'll only incur a parser error
2010-05-26 10:25:33 utc tosch_le yeah, i know. but i don't know what the f****** json backend is doing in the end. better be on the safe side. and be good example for fellow devs.
2010-05-26 10:26:02 utc jmettraux the problem I have to deal now is how to wrap the root error
2010-05-26 10:26:25 utc tosch_le let me have a try on that
2010-05-26 10:26:27 utc jmettraux well I guess I could just say "x" is not a valid JSON and amen
2010-05-26 10:29:42 utc jmettraux I can do it, there's not much magic, I'll just wrap the cause in a 'cause' attribute
2010-05-26 10:31:09 utc tosch_le which 'cause'?
2010-05-26 10:31:25 utc jmettraux the original parse error, intercepted
2010-05-26 10:31:44 utc jmettraux be it yajl / json / json_pure or active_support
2010-05-26 10:32:23 utc tosch_le i would copy the message to the new exception
2010-05-26 10:32:40 utc tosch_le that way, rufus-json is still a wrapper
2010-05-26 10:32:56 utc tosch_le : it just wraps the error classes, too.
2010-05-26 10:33:04 utc jmettraux ok
2010-05-26 10:33:07 utc jmettraux let me do it
2010-05-26 10:33:59 utc jmettraux or you want to go on ? Seen your fork
2010-05-26 10:34:22 utc tosch_le let me try and see if you like my solution
2010-05-26 10:34:37 utc jmettraux ok
2010-05-26 10:53:32 utc tosch_le hey, active_support backend really doesn't care about json syntax...
2010-05-26 10:53:48 utc tosch_le it parses {123:456} and even '12345sfddfrt#
2010-05-26 10:53:50 utc tosch_le '
2010-05-26 10:54:27 utc tosch_le line breaks don't mess the results, too...
2010-05-26 10:55:07 utc jmettraux :(
2010-05-26 10:55:57 utc tosch_le and hey, even " within the string is no problem
2010-05-26 10:56:16 utc tosch_le seems that it eats everything...
2010-05-26 10:56:33 utc tosch_le so no test case for as...
2010-05-26 10:56:49 utc jmettraux :(
2010-05-26 11:02:50 utc tosch_le http://github.com/tosch/rufus-json/commit/ec903d40bd68e96ec3c1d56a368e70b111967cd4
2010-05-26 11:03:50 utc jmettraux excellent
2010-05-26 11:04:02 utc jmettraux very clean
2010-05-26 11:04:47 utc tosch_le :)
2010-05-26 11:05:09 utc jmettraux I would have never thought of using lambdas for wrapping the exceptions as well
2010-05-26 11:05:18 utc jmettraux that's great
2010-05-26 11:05:45 utc jmettraux OK, if you tell me you're OK with a release now, I pull, package and release
2010-05-26 11:05:46 utc jmettraux wdyt ?
2010-05-26 11:05:51 utc tosch_le i would not have thought of that, too, if just putting the exception class there would have worked...
2010-05-26 11:06:08 utc tosch_le it's perfectly fine for me.
2010-05-26 11:07:01 utc tosch_le i'll update my rk fork to use the new release afterwards
2010-05-26 11:07:08 utc jmettraux :)
2010-05-26 11:07:44 utc tosch_le still some way to go till a beta release of rk, though
2010-05-26 11:08:04 utc jmettraux I'll be using it in prod very soon ;)
2010-05-26 11:08:16 utc tosch_le :-)
2010-05-26 11:08:46 utc tosch_le i suppose documentation is still a weak point
2010-05-26 11:10:49 utc tosch_le but step after step...
2010-05-26 11:16:11 utc jmettraux tosch_le : http://rubygems.org/gems/rufus-json
2010-05-26 11:16:46 utc tosch_le great, thanks!
2010-05-26 11:17:23 utc jmettraux thanks to you !
2010-05-26 11:33:43 utc tosch_le http://github.com/tosch/ruote-kit/commit/0562fc382e4a243cef0c806151c7f1ac002261eb
2010-05-26 11:33:59 utc tosch_le will wait for a new release, though
2010-05-26 11:34:20 utc tosch_le i'd like to get some feedback on the new filter params
2010-05-26 11:34:37 utc tosch_le and some more features (like /errors ;-) ) would be fine
2010-05-26 11:35:04 utc tosch_le will continue with updating rk-client tomorrow
2010-05-26 11:36:34 utc jmettraux OK
2010-05-26 11:36:41 utc jmettraux I will try to do /errors
2010-05-26 14:27:30 utc kschiess I am unsure about all the logging/tracing stuff in ruote. My goal would be to somehow produce output at every step of a process.
2010-05-26 14:27:51 utc kschiess I got something like 's_logger' => [ 'ruote/log/test_logger', 'Ruote::TestLogger' ], but am not sure if I understand that...
2010-05-26 14:28:12 utc tosch_le hello kschiess!
2010-05-26 14:28:21 utc kschiess hi ;)
2010-05-26 14:28:26 utc tosch_le unfortunately, jmettraux is not around
2010-05-26 14:28:36 utc kschiess I can see that ;)
2010-05-26 14:28:55 utc tosch_le hi could answer your question better than me.
2010-05-26 14:28:56 utc kschiess Thought I'd try you guys while reading the code... looks like old/new mix to me ;)
2010-05-26 14:29:43 utc tosch_le i'd advise you to post a question on the mailing list
2010-05-26 14:30:39 utc kschiess This is the kind of thing I am hesitant about asking on a list.. looks like I didn't do my homework, when in fact I tried very hard ;)
2010-05-26 14:31:28 utc tosch_le john knows the documentation is missing some points, so his answer will help him to complete the docs ;-)
2010-05-26 14:32:17 utc kschiess I will do so if really I can't find my way around. Might post the answer anyway then ;)
2010-05-26 14:33:08 utc kschiess tnx
2010-05-26 14:34:01 utc tosch_le np
2010-05-26 14:38:57 utc tosch_le ok, now i've got some time to have a look into the code.
2010-05-26 14:39:07 utc tosch_le what did you already try to do?
2010-05-26 14:46:20 utc tosch_le kschiess: 's_logger' => [ 'ruote/log/test_logger', 'Ruote::TestLogger' ] is a configuration option set in the storage
2010-05-26 14:47:11 utc kschiess That should log something, right? (it doesn't).
2010-05-26 14:47:16 utc tosch_le http://github.com/jmettraux/ruote/blob/ruote2.1/test/functional/base.rb#L27 includes an example for setting that option
2010-05-26 14:47:48 utc kschiess I guess that's where I got it from.
2010-05-26 14:48:09 utc kschiess But that must be a bad alley to walk. I can't find the canonical small logging example.
2010-05-26 14:48:32 utc tosch_le wait, have to update my sources...
2010-05-26 14:50:38 utc tosch_le you have to call @engine.context.logger.noisy = true
2010-05-26 14:50:55 utc tosch_le see http://github.com/jmettraux/ruote/blob/ruote2.1/test/functional/base.rb#L35
2010-05-26 14:51:15 utc tosch_le and #L92
2010-05-26 14:51:56 utc kschiess I'll try that in a second, thank you!
2010-05-26 14:52:06 utc tosch_le hope it helps
2010-05-26 14:57:36 utc kschiess Ok, stole hacked something together, http://pastie.org/978090 - but I would still like to hear john about this. Will post to ML.
2010-05-26 14:58:13 utc kschiess This kinda works, but nothing like what I was looking for.
2010-05-26 14:59:08 utc kschiess Anyho, tnx for the help and tata, tosch_le