ruote tmp/log_2011-05-03.html

2011-05-03 11:16:26 utc kennethkalmer jmettraux: hello !
2011-05-03 11:16:51 utc kennethkalmer tosch_le: hello !
2011-05-03 11:17:43 utc jmettraux kennethkalmer: hello ! how are you doing ?
2011-05-03 11:17:54 utc kennethkalmer good thanks, and you ?
2011-05-03 11:18:20 utc jmettraux doing fine
2011-05-03 11:18:23 utc jmettraux all good
2011-05-03 11:18:39 utc kennethkalmer scaling well ? :)
2011-05-03 11:18:53 utc jmettraux slowly but surely :)
2011-05-03 11:18:59 utc kennethkalmer awesome
2011-05-03 11:19:25 utc kennethkalmer have a quick question if that is alright, on composite storage and storage in general
2011-05-03 11:19:36 utc jmettraux please
2011-05-03 11:19:59 utc kennethkalmer so i'm using couch for workitems and history, redis for everything else
2011-05-03 11:20:17 utc kennethkalmer however the worker refuses to run
2011-05-03 11:20:25 utc kennethkalmer i'm prepping a gist quickly
2011-05-03 11:23:24 utc kennethkalmer https://gist.github.com/52d2f0d573e616232827
2011-05-03 11:23:55 utc kennethkalmer woops, back in 10 getting yanked away from desk
2011-05-03 11:26:44 utc jmettraux kennethkalmer: if you could tweak your ruote-redis / storage.rb to show your the "doc without a _rev", maybe you some leftover that can't be digested
2011-05-03 11:39:21 utc jmettraux maybe you have some leftover...
2011-05-03 11:47:31 utc kennethkalmer ooh, thanks, will have a look
2011-05-03 11:47:32 utc kennethkalmer then another thing
2011-05-03 11:48:12 utc kennethkalmer should participants only be registered once, like I have in my initializer (unless $RAKE_TASK...) or can all instances of the engine register workers
2011-05-03 11:48:33 utc jmettraux all the instances can register
2011-05-03 11:49:07 utc jmettraux registering once is probably better if you want to keep properly track of what you registered
2011-05-03 11:49:41 utc jmettraux there is the new Engine#participant_list= and its little brother #participant_list
2011-05-03 11:49:52 utc jmettraux that let you do plist manipulation in one go
2011-05-03 11:50:14 utc kennethkalmer saw it in the docs
2011-05-03 11:50:41 utc jmettraux start with the next ruote, Engine#register(&block) will automatically erase the participant list before registering what's in its block
2011-05-03 11:50:56 utc kennethkalmer that is great
2011-05-03 11:50:57 utc jmettraux starting with...
2011-05-03 11:51:00 utc kennethkalmer no side effects
2011-05-03 11:54:18 utc jmettraux I've been working on placing participant [code] in variables, so that a process can have its own participants, just for its own run
2011-05-03 12:02:55 utc kennethkalmer woah
2011-05-03 12:04:25 utc kennethkalmer hmm, cleared the keys and still it is breaking, inspecting more closely now
2011-05-03 12:06:57 utc kennethkalmer got it, updating the gist quickly
2011-05-03 12:07:18 utc kennethkalmer the msgs/* keys has no _rev fields
2011-05-03 12:08:35 utc kennethkalmer here is an example
2011-05-03 12:08:36 utc kennethkalmer \"regex\":\"notify\",\"put_at\":\"2011-05-03 12:01:04.839041 UTC\",\"action\":\"participant_registered\",\"_id\":\"62843!2148446520!2011-05-03!1304424064.839003!000\",\"type\":\"msgs\"}
2011-05-03 12:09:26 utc kennethkalmer ACTION checking for updated ruote versions
2011-05-03 12:10:01 utc jmettraux I'm wondering how this ruote-redis 2.2.0 went through the CI
2011-05-03 12:10:02 utc kennethkalmer nope, i'm fully up to date, peculiar
2011-05-03 12:10:11 utc kennethkalmer i'm on REE
2011-05-03 12:10:14 utc kennethkalmer let me run local tests
2011-05-03 12:10:45 utc jmettraux you're 2.2.0 right ?
2011-05-03 12:10:49 utc kennethkalmer yep
2011-05-03 12:11:00 utc kennethkalmer bundle show | grep ruote
2011-05-03 12:11:01 utc kennethkalmer * ruote (2.2.0)
2011-05-03 12:11:01 utc kennethkalmer * ruote-couch (2.2.0)
2011-05-03 12:11:02 utc kennethkalmer * ruote-kit (2.2.0.3)
2011-05-03 12:11:04 utc kennethkalmer * ruote-redis (2.2.0)
2011-05-03 12:13:24 utc jmettraux kennethkalmer: could you please gist your ruote/redis/storage.rb ?
2011-05-03 12:13:47 utc kennethkalmer the actual file ? sure thing
2011-05-03 12:14:23 utc jmettraux ah
2011-05-03 12:14:25 utc jmettraux no, sorry
2011-05-03 12:14:36 utc kennethkalmer https://gist.github.com/953222
2011-05-03 12:14:37 utc jmettraux no need
2011-05-03 12:14:43 utc kennethkalmer oh, no harm :)
2011-05-03 12:15:02 utc jmettraux now I remember, it's all my fault
2011-05-03 12:16:04 utc jmettraux it's fixed in https://github.com/jmettraux/ruote/commit/a1f36b435b61f824f8fb5fd83d142600812daddd
2011-05-03 12:16:17 utc jmettraux CompositeStorage issue, sorry
2011-05-03 12:18:02 utc kennethkalmer no worries, thanks for legandary support as always
2011-05-03 12:18:51 utc kennethkalmer ok, so I can change my Gemfile to use that commit for ruote, or is it not safe ?
2011-05-03 12:19:14 utc kennethkalmer or I can monkey patch it
2011-05-03 12:19:14 utc jmettraux it should be OK
2011-05-03 12:19:42 utc jmettraux the safer is to suck this new version of CompositeStorage into something like Ispinabox::CompositeStorage for now
2011-05-03 12:20:54 utc jmettraux the 'master' version of ruote-redis might be interesting, it uses Redis#lpop (less operations)
2011-05-03 12:22:34 utc kennethkalmer saw in the changelog
2011-05-03 12:22:47 utc kennethkalmer first delved around looking for an obvious commit fixing the issue before I reported it
2011-05-03 12:25:19 utc jmettraux you're still busy as ever
2011-05-03 12:27:32 utc jmettraux kennethkalmer: coffee ?
2011-05-03 12:27:38 utc kennethkalmer oh yes please !
2011-05-03 12:27:58 utc jmettraux ACTION emits espresso
2011-05-03 12:33:19 utc kennethkalmer ACTION consumes happily
2011-05-03 12:46:32 utc kennethkalmer great, copied composite_storage seems to work, thanks !
2011-05-03 12:46:52 utc jmettraux w00t !
2011-05-03 12:47:01 utc kennethkalmer :D
2011-05-03 12:51:48 utc kennethkalmer man, the memory leaks in this application in development mode is unreal
2011-05-03 12:52:09 utc jmettraux how can I help ?
2011-05-03 12:52:59 utc kennethkalmer a thing that has been plauging IIAB since day one
2011-05-03 12:53:18 utc jmettraux what does your instinct say ?
2011-05-03 12:53:19 utc kennethkalmer so no worries, unless you have a tip for keeping memory leaks down in development env
2011-05-03 12:53:32 utc kennethkalmer gonna try running under passenger solo thing now
2011-05-03 12:53:53 utc kennethkalmer every request in dev mode leaks like ~60MB
2011-05-03 12:55:25 utc jmettraux weird
2011-05-03 12:55:42 utc jmettraux even the smallest GET / ?
2011-05-03 12:56:07 utc kennethkalmer yep, even a GET into middleware
2011-05-03 12:56:19 utc kennethkalmer i think Rails' class reloading is to blame
2011-05-03 12:56:25 utc jmettraux aaah
2011-05-03 12:56:39 utc kennethkalmer we have tons of AR::Base.send(:include, ActsAsSomethingSpecific)
2011-05-03 12:58:31 utc kennethkalmer can't wait to get this app on Rails 3
2011-05-03 12:58:39 utc jmettraux :)
2011-05-03 13:01:48 utc jmettraux you should probably introduce a start option which disable the reloading, just for your personal use
2011-05-03 13:02:24 utc jmettraux well, if you have sessions when you know you won't need it :-(
2011-05-03 13:05:51 utc kennethkalmer yeah, definitely
2011-05-03 13:06:11 utc kennethkalmer testing the passenger standalone now, hoping the smart spawning will help
2011-05-03 13:07:11 utc jmettraux how far is SG from ZA in hours of plane ?
2011-05-03 13:08:26 utc kennethkalmer Singapore ? 11 iirc
2011-05-03 13:08:59 utc jmettraux then, next year, you should do the http://reddotrubyconf.com/ it was fun
2011-05-03 13:09:17 utc kennethkalmer heard it rocked