ruote tmp/log_2012-09-07.html

2012-09-07 20:51:48 utc Skippy1 Anyone getting stack level too deep errors when registering a participant in 2.3.0?
2012-09-07 20:54:24 utc danryan Skippy1: hi skippy, I'm using 2.3.0 and haven't noticed that. Mind gisting your code?
2012-09-07 20:55:21 utc Skippy1 well it never get's past this line:
2012-09-07 20:55:23 utc Skippy1 RUOTE_DASHBOARD.register( :publish_to_queue, CAPEngineWorker::Participants::AmqpPublish )
2012-09-07 20:56:53 utc danryan Gist the stacktrace as well please if you don't mind.
2012-09-07 20:57:01 utc Skippy1 doesn't seem anything untowards about it
2012-09-07 20:57:14 utc Skippy1 how do I get the stacktrace?
2012-09-07 20:59:15 utc danryan Skippy1: whatever spits out from "stack level too deep" is it
2012-09-07 20:59:27 utc Skippy1 I just noticed that a gem i'm using uses polyglot
2012-09-07 20:59:36 utc Skippy1 well there ain't much
2012-09-07 20:59:52 utc Skippy1 SystemStackError: stack level too deep
2012-09-07 20:59:54 utc Skippy1 from /home/farrel/.rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/irb/workspace.rb:80
2012-09-07 20:59:54 utc Skippy1 Maybe IRB bug!!
2012-09-07 20:59:55 utc Skippy1 that's it
2012-09-07 21:00:07 utc Skippy1 anyway about polyglot
2012-09-07 21:00:14 utc Skippy1 it has a Polyglot.register method
2012-09-07 21:00:32 utc Skippy1 could that be interfering or overwriting with the register in Ruote?
2012-09-07 21:05:05 utc Skippy1 removed the gems that use polyglot from my gemfile but that didn't help
2012-09-07 21:38:03 utc jmettraux Skippy1: hello and welcome to #ruote
2012-09-07 21:38:28 utc jmettraux do you have a stracktrace for the issue?
2012-09-07 21:38:51 utc Skippy1 not really
2012-09-07 21:38:54 utc Skippy1 just those three lines
2012-09-07 21:38:59 utc jmettraux ok
2012-09-07 21:39:30 utc Skippy1 this is on 1.9.2-p290 as well
2012-09-07 21:41:37 utc jmettraux could you add some print output before line 80 in your /home/farrel/.rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/irb/workspace.rb to see what's handed to the eval in there?
2012-09-07 21:42:21 utc Skippy1 that I think is just the eval loop in irb
2012-09-07 21:42:30 utc Skippy1 when I run from command line it complains about a line in daemin-kit
2012-09-07 21:43:12 utc Skippy1 i assume from where it's require'ing the file that registers my participants
2012-09-07 21:43:50 utc jmettraux well, when run from the command line there should be a different error message / stacktrace to look at
2012-09-07 21:44:46 utc jmettraux let's forget the irb case then and focus on the daemon-kit thing. Could you please a gist of the error?
2012-09-07 21:45:08 utc jmettraux ... please create a gist of the error?
2012-09-07 21:45:10 utc Skippy1 http://pastebin.com/30axjDKQ
2012-09-07 21:46:25 utc Skippy1 here's where the participants are registered
2012-09-07 21:46:26 utc Skippy1 http://pastebin.com/nunJSXb3
2012-09-07 21:46:35 utc Skippy1 you can see a "b" in the first pastebin
2012-09-07 21:46:47 utc Skippy1 but none of the "a" after it are output
2012-09-07 21:48:21 utc jmettraux may I see the code of the AmqpPublish?
2012-09-07 21:48:49 utc jmettraux other question: you didn't say, but you seem to imply that it worked before 2.3.0, but it worked with what version/commit of ruote?
2012-09-07 21:50:48 utc Skippy1 I pulled in 2.3.0 from git
2012-09-07 21:51:09 utc Skippy1 but I've changed to using 2.3.0 from gems and still nothing
2012-09-07 21:51:19 utc jmettraux and before that?
2012-09-07 21:52:15 utc Skippy1 seemed to work
2012-09-07 21:52:28 utc jmettraux and what version/commit seemed to work?
2012-09-07 21:52:45 utc Skippy1 it's been a month or two since I did a bundle update
2012-09-07 21:52:53 utc Skippy1 but I did one when ruote went to 2.3.0 officially
2012-09-07 21:52:56 utc Skippy1 http://pastebin.com/CtvBZfub
2012-09-07 21:52:59 utc Skippy1 that's the participant
2012-09-07 21:53:16 utc jmettraux thanks
2012-09-07 21:56:22 utc jmettraux the only change you made was moving to ruote 2.3.0? No other changes? If you switch back to the previous ruote things work as before?
2012-09-07 21:57:14 utc Skippy1 let me use 2.2
2012-09-07 21:57:33 utc Skippy1 I'm just trying on ruby 1.9.3
2012-09-07 21:57:38 utc Skippy1 quick
2012-09-07 21:57:47 utc jmettraux no, please use the same ruote as before, look at what was in your Gemfile.lock before you updated to ruote 2.3.0
2012-09-07 21:58:47 utc Skippy1 I had gem( 'ruote', '2.3.0', :git => 'ruote-github-here' )
2012-09-07 21:58:53 utc Skippy1 in my Gemfile beforehand
2012-09-07 21:59:04 utc Skippy1 so I was kinda following edge
2012-09-07 21:59:15 utc jmettraux ok, and your Gemfile.lock was pointing to a given SHA...
2012-09-07 21:59:16 utc Skippy1 but hadn't done a bundle update in a while
2012-09-07 21:59:20 utc Skippy1 yeah
2012-09-07 21:59:50 utc jmettraux so really, let me ask again, you only upped ruote to 2.3.0 official and it breaks? No side story?
2012-09-07 22:01:06 utc Skippy1 not that I know of it was working about 2 days ago.. I did a bundle update this evening and I get the errors
2012-09-07 22:01:17 utc Skippy1 a whole bunch of other gems got upgraded as well
2012-09-07 22:01:28 utc Skippy1 I initially thought it was polyglot
2012-09-07 22:01:28 utc jmettraux ok
2012-09-07 22:01:34 utc Skippy1 it has a 'register' method
2012-09-07 22:01:43 utc Skippy1 and it does some funky stuff with require
2012-09-07 22:01:51 utc jmettraux ok
2012-09-07 22:02:11 utc Skippy1 but I removed all the gems that depend on it from my gemfile and same error happens
2012-09-07 22:02:24 utc jmettraux I suggest you continue the "puts a" vs "puts b" investigation inside of Ruote.register (and then further down) to isolate the problematic point
2012-09-07 22:02:35 utc jmettraux use "bundle open ruote" to help you do that
2012-09-07 22:03:22 utc jmettraux ping me here when you've pinpointed the root
2012-09-07 22:06:40 utc Skippy1 will try
2012-09-07 22:08:41 utc jmettraux :-)
2012-09-07 22:36:29 utc Skippy1 jmettraux: it's getting last on this line in Dashboard#register_participant
2012-09-07 22:36:31 utc Skippy1 pa = @context.plist.register(regex, participant, opts, block)
2012-09-07 22:36:45 utc Skippy1 not sure where the context or plist is...
2012-09-07 22:42:07 utc Skippy1 okay found it
2012-09-07 22:45:27 utc Skippy1 jmettraux: think I found it
2012-09-07 22:45:42 utc Skippy1 in participant_list.rb
2012-09-07 22:45:58 utc Skippy1 in register
2012-09-07 22:46:00 utc Skippy1 line 125
2012-09-07 22:46:22 utc Skippy1 it never gets past that and just loops to the beginning
2012-09-07 22:51:39 utc Skippy1 is it because my dashboard works directly on a storage and not on a woker?
2012-09-07 22:52:45 utc jmettraux looking at it, sorry, to soon to draw conclusions. If it worked before, it should work again
2012-09-07 22:53:17 utc jmettraux I need one more piece of info: the git diff of your Gemfile.lock from "working" to "now"
2012-09-07 22:53:18 utc jmettraux please
2012-09-07 22:55:38 utc jmettraux ok, if you could please add a "p r" to that participant_list.rb file, right before line 125, and try again and tell me what's happening
2012-09-07 22:55:42 utc Skippy1 http://pastebin.com/L1q1z6aq
2012-09-07 22:56:08 utc jmettraux thanks :-)
2012-09-07 22:56:38 utc Skippy1 just getting this over and over
2012-09-07 22:58:01 utc Skippy1 http://pastebin.com/Dt1rWLzg
2012-09-07 22:58:17 utc Skippy1 as far as I can tell it's not changing
2012-09-07 22:58:44 utc jmettraux always _rev=>55?
2012-09-07 22:58:54 utc Skippy1 yup
2012-09-07 22:59:44 utc jmettraux ok, there something wrong in ruote-mon
2012-09-07 23:00:09 utc Skippy1 ah
2012-09-07 23:00:31 utc jmettraux the potential quickfix: remove that document (_id: participant_list) from MongoDB and then retry
2012-09-07 23:00:44 utc Skippy1 ok lemme wipe my data dirs
2012-09-07 23:00:48 utc Skippy1 they're local
2012-09-07 23:00:52 utc jmettraux ok
2012-09-07 23:01:29 utc Skippy1 ok it's preallocating new files
2012-09-07 23:03:11 utc Skippy1 ok now it works
2012-09-07 23:03:21 utc Skippy1 so it was definitely something in mongo...
2012-09-07 23:03:41 utc jmettraux it's certainly my fault (ruote-mon)
2012-09-07 23:05:04 utc jmettraux please tell me if it happens again, I'll try to come up with a repro case on my side
2012-09-07 23:05:13 utc jmettraux thanks for your time!
2012-09-07 23:05:18 utc Skippy1 no probs
2012-09-07 23:20:50 utc jmettraux danryan: thanks for your help :-)
2012-09-07 23:43:07 utc briahn10 jmettraux: can you share a little bit of your swf backend?
2012-09-07 23:43:26 utc jmettraux hello, sorry, I can't, it belongs to my employer
2012-09-07 23:43:28 utc briahn10 *talk a little bit
2012-09-07 23:43:29 utc briahn10 i mean
2012-09-07 23:44:26 utc jmettraux well, it uses SWF as a backend
2012-09-07 23:44:41 utc briahn10 ah, perhaps a general question about swf (unrelated to route)?
2012-09-07 23:44:47 utc jmettraux ok
2012-09-08 00:00:46 utc briahn10 actually, i don't think it's necessarily related to swf :|
2012-09-08 00:00:50 utc briahn10 sorry about that
2012-09-08 00:01:04 utc briahn10 just thinking about different ways to structure activity worker code tied to one workflow
2012-09-08 00:02:32 utc briahn10 how to deploy it, how to scale the # of workers, etc
2012-09-08 00:08:13 utc jmettraux "activity worker" in the SWF sense?
2012-09-08 00:18:03 utc briahn10 yeah
2012-09-08 00:18:41 utc jmettraux I don't think there is a need to tie an activity worker to one workflow, activity workers just picks activities
2012-09-08 00:18:59 utc jmettraux whatever the workflow
2012-09-08 00:20:15 utc jmettraux if you want specific activity workers to do specific activities you can probably use different tasklists
2012-09-08 00:20:55 utc briahn10 right, i'm imagining trying to scale anything but a few activity workers is going to require chef or some other management tool
2012-09-08 00:21:30 utc jmettraux right