ruote log_2010-09-30

2010-09-30 06:44:57 utc jmettraux tosch_le: hello
2010-09-30 06:45:09 utc tosch_le hello!
2010-09-30 06:45:19 utc tosch_le still having fun with lsof?
2010-09-30 06:45:31 utc jmettraux done and fixed
2010-09-30 06:45:41 utc jmettraux how are you doing ?
2010-09-30 06:46:11 utc tosch_le fine, thanks. will voice over some screencasts today :) how are _you_ doing?
2010-09-30 06:46:39 utc tosch_le and what was the fix?
2010-09-30 06:48:34 utc jmettraux the fix http://github.com/jmettraux/rufus-jig/blob/master/lib/rufus/jig/adapters/net_persistent.rb#L32-75 is about closing the http connections referenced in all the threads
2010-09-30 06:49:45 utc jmettraux do you have some time to test a thing or two ?
2010-09-30 06:49:55 utc jmettraux (don't be afraid to say no)
2010-09-30 06:50:14 utc tosch_le i won't, it's okay, i've got some time
2010-09-30 06:51:09 utc jmettraux feel inclined to test ruote-dm vs mysql ?
2010-09-30 06:51:40 utc tosch_le sounds like great job for me :-)
2010-09-30 06:52:20 utc jmettraux let me push a few things
2010-09-30 06:54:35 utc jmettraux just pushed the latest ruote and ruote-dm
2010-09-30 06:55:10 utc jmettraux ruote-dm's README contains the instructions on testing
2010-09-30 06:55:24 utc jmettraux ruote-dm/test/functional_connection.rb contains the connection details
2010-09-30 06:56:01 utc jmettraux and you can uncomment the auto_upgrade! code in there to let ruote-dm setup the db for you (if properly connected)
2010-09-30 06:56:25 utc tosch_le ok, give me some minutes and i will start testing
2010-09-30 06:56:25 utc jmettraux coffee ?
2010-09-30 06:56:31 utc jmettraux take your time
2010-09-30 06:56:44 utc tosch_le tea would be better today
2010-09-30 06:57:05 utc tosch_le http://gist.github.com/604147 <-- i'd implemented the shutdown_all_threads thing that way
2010-09-30 06:57:16 utc tosch_le seems a bit more dry to me
2010-09-30 06:58:01 utc jmettraux :) That's the version that should go to drbrain
2010-09-30 06:58:13 utc tosch_le yes, i suppose your right
2010-09-30 06:58:21 utc tosch_le s/your/you're/
2010-09-30 06:58:28 utc jmettraux def shutdown(t=Thread.current)
2010-09-30 06:58:37 utc tosch_le even better, yes
2010-09-30 06:58:37 utc jmettraux def shutdown(thread=Thread.current)
2010-09-30 06:59:17 utc jmettraux I looked at the test suite of net-http-persistent yesterday evening and tried to set up a few tests about that, but it broke my head
2010-09-30 06:59:55 utc jmettraux this morning I just went straight to rufus-jig and did my monkey patching (once the lsof taught me that connections were kept around)
2010-09-30 07:00:15 utc tosch_le monkey patching ftw
2010-09-30 07:00:17 utc tosch_le :-)
2010-09-30 07:00:25 utc jmettraux so, you are right
2010-09-30 07:00:49 utc jmettraux sorry for the lazy shortcut
2010-09-30 07:01:17 utc jmettraux OK, I'm out for a coffee, I'll be back
2010-09-30 07:01:33 utc tosch_le no need to be sorry, i'm just checking my ruby foo
2010-09-30 07:15:29 utc jmettraux back
2010-09-30 07:20:31 utc tosch_le "Ruote::Dm::Document.auto_upgrade!" isn't available by default anymore, the dm-migrations gem needs to be installed and required before
2010-09-30 07:21:01 utc jmettraux ouch, what about putting
2010-09-30 07:21:15 utc jmettraux require "dm-migrations" # gem install dm-migrations
2010-09-30 07:21:22 utc jmettraux right before the auto_upgrade! ?
2010-09-30 07:21:45 utc tosch_le that's what i've done to get the tests running
2010-09-30 07:22:10 utc jmettraux can I add it on my side ?
2010-09-30 07:22:36 utc tosch_le also, i had to call "ruby test/functional/test.rb -- --dm" instead of "ruby test/functional/test.rb --dm" to make the tests run
2010-09-30 07:23:03 utc jmettraux minitest
2010-09-30 07:23:11 utc jmettraux OK, I'm changing that on my side
2010-09-30 07:25:13 utc jmettraux pushed
2010-09-30 07:26:37 utc tosch_le unit tests were fine with ruby enterprise edition
2010-09-30 07:26:53 utc tosch_le ruby 1.8.7 (2010-04-19 patchlevel 253) [i686-linux], MBARI 0x8770, Ruby Enterprise Edition 2010.02
2010-09-30 07:27:02 utc tosch_le running functional tests now
2010-09-30 07:30:38 utc jmettraux works with me for 1.9.1p378 and psql
2010-09-30 07:31:40 utc tosch_le tests still running; they really take their time on my machine
2010-09-30 07:32:14 utc tosch_le but it's fun to see the db table changing
2010-09-30 07:32:29 utc jmettraux mysql admin or some GUI ?
2010-09-30 07:32:42 utc tosch_le mysql query browser
2010-09-30 07:39:36 utc kennethkalmer http://adamdeane.wordpress.com/2010/09/29/open-source-bpm/
2010-09-30 07:39:40 utc kennethkalmer afternoon !
2010-09-30 07:40:07 utc jmettraux kennethkalmer: hello Kenneth
2010-09-30 07:40:35 utc jmettraux I'm following this dude, looking at ruote will make him cringe ;-)
2010-09-30 07:41:04 utc jmettraux his previous post, http://adamdeane.wordpress.com/2010/09/27/open-source-bpm-safari-trip/ is fun as well
2010-09-30 07:41:12 utc jmettraux thanks for the referral !
2010-09-30 07:42:53 utc kennethkalmer np, reading the safari trip
2010-09-30 07:43:10 utc jmettraux tosch_le: oh oh, you've just forked net-http-persistent
2010-09-30 07:43:46 utc tosch_le is this dangerous?
2010-09-30 07:43:50 utc jmettraux no no
2010-09-30 07:44:01 utc tosch_le thought i'd have a look at those tests
2010-09-30 07:44:10 utc jmettraux I will be very happy if your patch makes it
2010-09-30 07:44:23 utc tosch_le and perhaps add one function and change another ;-)
2010-09-30 07:44:41 utc tosch_le let's see if i get some tests written
2010-09-30 07:45:03 utc jmettraux looking forward to it !
2010-09-30 07:45:48 utc tosch_le functional tests for ruote-dm/mysql still running, but there's at least one 'F'
2010-09-30 07:46:19 utc jmettraux OK
2010-09-30 08:30:53 utc tosch_le jmettraux: http://github.com/tosch/net-http-persistent/commit/ab9a9244d81f274142261290fada190bfbaaa328 wdyt?
2010-09-30 08:31:26 utc tosch_le functional tests for ruote-dm/mysql seems to be hanging
2010-09-30 08:31:34 utc jmettraux :(
2010-09-30 08:31:43 utc jmettraux your patch looks excellent
2010-09-30 08:32:46 utc tosch_le http://gist.github.com/604228
2010-09-30 08:33:37 utc tosch_le are there timing issues likely in the tests?
2010-09-30 08:33:48 utc jmettraux in that one yes
2010-09-30 08:34:48 utc jmettraux could you try to run ruby test/functional/rt_1_listen.rb -- --dm ?
2010-09-30 08:34:51 utc jmettraux please
2010-09-30 08:35:14 utc tosch_le yes, of course
2010-09-30 08:35:16 utc tosch_le running now
2010-09-30 08:35:42 utc tosch_le the configurations entry is in the db
2010-09-30 08:35:58 utc tosch_le but nothing else visible happens
2010-09-30 08:36:25 utc jmettraux stuck ?
2010-09-30 08:36:41 utc tosch_le yes
2010-09-30 08:37:20 utc jmettraux consistent behaviour ftw
2010-09-30 08:37:47 utc jmettraux what happens if you uncomment this line http://github.com/jmettraux/ruote/blob/ruote2.1/test/functional/rt_1_listen.rb#L46 ?
2010-09-30 08:37:50 utc tosch_le still hanging around, that one
2010-09-30 08:38:09 utc jmettraux I'd try a pull request with your net-http-persistent
2010-09-30 08:38:36 utc jmettraux I hope it makes it into the original repo, in a form or the other
2010-09-30 08:42:48 utc jmettraux the uncomment trick reveals nothing ?
2010-09-30 08:43:11 utc tosch_le writing the pull request, wait a moment
2010-09-30 08:43:18 utc jmettraux :-)
2010-09-30 08:45:22 utc tosch_le http://github.com/drbrain/net-http-persistent/pull/1
2010-09-30 08:46:23 utc tosch_le http://gist.github.com/604228
2010-09-30 08:46:34 utc tosch_le rt-1-listen w/ noisy
2010-09-30 08:46:56 utc jmettraux strange
2010-09-30 08:47:15 utc jmettraux what happens if you
2010-09-30 08:47:27 utc jmettraux ruby test/functional/rt_1_listen.rb -- --dm -d
2010-09-30 08:48:23 utc tosch_le no change
2010-09-30 08:49:43 utc jmettraux weird
2010-09-30 08:50:18 utc jmettraux what happens if you uncomment the noisy up a few lines ?
2010-09-30 08:51:26 utc tosch_le http://gist.github.com/604228
2010-09-30 08:52:24 utc jmettraux I've just installed mysql 5.1 on my linux office box
2010-09-30 08:52:34 utc jmettraux but that'll have to wait for tomorrow
2010-09-30 08:52:57 utc jmettraux pull request : you don't need to mention my name ;-)
2010-09-30 08:53:03 utc jmettraux he'll hate us both ;-)
2010-09-30 08:53:13 utc tosch_le too late ;-p
2010-09-30 08:53:28 utc jmettraux lol
2010-09-30 08:53:34 utc jmettraux weird issue with this rt_
2010-09-30 08:53:43 utc jmettraux seems like the tracker is dead
2010-09-30 08:54:16 utc tosch_le yes, wait_for never returns
2010-09-30 08:54:45 utc jmettraux maybe putting some puts/p into http://github.com/jmettraux/ruote/blob/ruote2.1/lib/ruote/svc/tracker.rb could help see clearer
2010-09-30 08:55:14 utc jmettraux running with --fs to see the expected behaviour and then with --dm to see the fail behaviour
2010-09-30 08:55:41 utc jmettraux I have to escape for a while
2010-09-30 08:56:21 utc tosch_le see you later!
2010-09-30 08:56:28 utc jmettraux ciao !
2010-09-30 09:49:31 utc jmettraux tosch_le: sorry, which version of mysql on which platform are you running ?
2010-09-30 09:50:57 utc tosch_le MySQL 5.1.41-3ubuntu12.6
2010-09-30 09:53:32 utc tosch_le updated http://gist.github.com/604228
2010-09-30 09:53:53 utc tosch_le one log with --fs and one with --dm and some puts in the Tracker class
2010-09-30 09:55:02 utc jmettraux the receive seems to happen before the tracker is registered
2010-09-30 09:58:14 utc jmettraux well not really
2010-09-30 09:59:49 utc jmettraux puzzling
2010-09-30 10:00:24 utc jmettraux but at least consistent
2010-09-30 10:00:59 utc jmettraux cheese fondue time
2010-09-30 10:01:08 utc tosch_le you're lucky
2010-09-30 10:01:13 utc tosch_le :-)
2010-09-30 10:05:29 utc tosch_le leaving for lunch: chinese cookshop
2010-09-30 10:05:37 utc jmettraux bon appetit !
2010-09-30 10:05:49 utc tosch_le merci
2010-09-30 10:06:52 utc lbt jmettraux: quick update ... the chinese team can't cc the openweferu... great firewall :(
2010-09-30 10:07:07 utc jmettraux lbt: ah too bad
2010-09-30 10:07:24 utc jmettraux thanks anyway
2010-09-30 10:07:40 utc lbt I've asked them to join the ml I mentioned yesterday and 'announce' the perf results... so you can ack that with some questions if you are on already :)
2010-09-30 10:09:03 utc jmettraux ah yes, I saw one of your guys registered on the ml, that's great
2010-09-30 10:10:15 utc lbt Victor is writing something now :)
2010-09-30 10:10:30 utc lbt sadly they're on hols next week :/
2010-09-30 10:14:04 utc lbt http://wiki.meego.com/Release_Infrastructure/BOSS/Performance/Results
2010-09-30 10:14:43 utc lbt they put some trend results up showing how ruote behaves over simulated long term running
2010-09-30 10:15:37 utc lbt ie after 12months of running 1k wflows/month the engine is a bit slower than it was at the start but not badly
2010-09-30 10:15:54 utc jmettraux irrealist
2010-09-30 10:16:05 utc jmettraux they started testing 1 month ago
2010-09-30 10:16:44 utc lbt ?
2010-09-30 10:17:28 utc jmettraux how can they say they've testing it for 12 months ?
2010-09-30 10:17:34 utc jmettraux time machine ?
2010-09-30 10:17:36 utc lbt no.. sorry
2010-09-30 10:17:51 utc lbt if we assume a normal usage is 1k/month ...
2010-09-30 10:18:07 utc lbt then load the system with 12k flows in a few hours
2010-09-30 10:18:16 utc lbt that simulates 'cycles'
2010-09-30 10:18:39 utc lbt and is intended to show memory leaks or issues due to growth of the DB or stuff
2010-09-30 10:19:54 utc lbt cf time to add each of the first 50 entries to a SQL index vs time to add the 1millionth entry :)
2010-09-30 10:20:22 utc jmettraux ah understood, thanks
2010-09-30 10:24:00 utc jmettraux when Victor shows up, I'll ask for clarifications
2010-09-30 10:24:24 utc jmettraux then when I'll have time, I'll reproduce the tests, but without any amqp stuff
2010-09-30 10:24:54 utc lbt *nod*
2010-09-30 10:25:19 utc lbt I actually asked them to do that but I don't think they know the ruby test framework
2010-09-30 10:28:08 utc jmettraux no need for that, I'll just write a one file script
2010-09-30 10:28:29 utc jmettraux 20_000.times { engine.launch(pdef) }
2010-09-30 10:30:38 utc lbt :D
2010-09-30 10:32:35 utc lbt would it be useful to have a load test area in the spec/ to allow people to run these kinds of tests
2010-09-30 10:32:53 utc lbt again, the focus isn't on speed, more reliability and endurance
2010-09-30 10:33:49 utc jmettraux there is already, but not documented
2010-09-30 10:33:56 utc jmettraux and very dusty
2010-09-30 10:40:07 utc lbt heh
2010-09-30 11:26:59 utc tosch_le back from lunch
2010-09-30 11:36:29 utc jmettraux tosch_le: I have the same issue with mysql 5.1 on osx
2010-09-30 11:36:48 utc tosch_le :-( but at least its reproducable
2010-09-30 11:39:25 utc tosch_le how do you run the tests on ruby 1.9.2? it complains "no such file to load -- test/functional/base (LoadError)"
2010-09-30 11:42:45 utc jmettraux ruby -I. test/functional/test.rb
2010-09-30 11:43:46 utc tosch_le ah, that's better, thanks
2010-09-30 11:44:34 utc tosch_le same issue on ruby-1.9.2
2010-09-30 11:44:52 utc tosch_le ruby 1.9.2p0 (2010-08-18 revision 29036) [i686-linux]
2010-09-30 11:49:52 utc jmettraux great
2010-09-30 11:52:41 utc tosch_le doc = @context.storage.get_trackers
2010-09-30 11:52:56 utc tosch_le in lib/ruote/svc/tracker.rb returns an empty hash
2010-09-30 11:53:03 utc jmettraux aah
2010-09-30 11:55:12 utc jmettraux nothing about trackers in the db ?
2010-09-30 11:55:49 utc jmettraux typ="variables"
2010-09-30 11:56:11 utc tosch_le no, just the engine configuration
2010-09-30 11:56:28 utc tosch_le typ="configurations"
2010-09-30 11:56:33 utc tosch_le one single entry
2010-09-30 12:00:16 utc tosch_le add_tracker never gets called (at least not before this wait_for)
2010-09-30 12:00:23 utc jmettraux aaah
2010-09-30 12:04:17 utc tosch_le i've got the feeling the table gets purged when @engine.shutdown or start_new_engine is called
2010-09-30 12:04:55 utc jmettraux so add_tracker is called before the shutdown ?
2010-09-30 12:05:12 utc tosch_le yes
2010-09-30 12:05:12 utc jmettraux (back for real, was cleaning dishes)
2010-09-30 12:05:34 utc jmettraux ok, so there is a purge going on, but why not with pgsql ?
2010-09-30 12:05:52 utc tosch_le got it running now: disabled the purge and auto_upgrade! calls in functional_connection.rb of ruote-dm
2010-09-30 12:06:28 utc jmettraux aaaaaaaaaah
2010-09-30 12:06:33 utc jmettraux lol
2010-09-30 12:06:39 utc tosch_le s/purge/all.destroy/
2010-09-30 12:07:05 utc jmettraux confirmed here
2010-09-30 12:07:06 utc tosch_le and that's another mystery solved :-)
2010-09-30 12:07:11 utc jmettraux excellent
2010-09-30 12:07:14 utc jmettraux well done !
2010-09-30 12:07:25 utc tosch_le puts ftw
2010-09-30 12:07:58 utc jmettraux \o/
2010-09-30 12:15:36 utc jmettraux may I ask you for another batch of tests with mysql ?
2010-09-30 12:16:46 utc tosch_le just running the complete functional test suite w/ 1.9.2 now
2010-09-30 12:17:03 utc tosch_le this will take its time, I suppose
2010-09-30 12:17:14 utc tosch_le what shall i test after that?
2010-09-30 12:17:34 utc jmettraux there are three multi workers tests
2010-09-30 12:17:58 utc tosch_le how do i run them?
2010-09-30 12:18:13 utc jmettraux . test/functional/crunner.sh 0 -- --dm
2010-09-30 12:18:17 utc jmettraux . test/functional/crunner.sh 1 -- --dm
2010-09-30 12:18:20 utc jmettraux . test/functional/crunner.sh 2 -- --dm
2010-09-30 12:18:32 utc jmettraux each of those scripts runs the test continuously
2010-09-30 12:18:51 utc tosch_le do i launch them in parallel or in order?
2010-09-30 12:18:53 utc jmettraux as soon as there is a error, it breaks
2010-09-30 12:18:55 utc jmettraux in order
2010-09-30 12:19:07 utc jmettraux this runner runs the test continuously
2010-09-30 12:19:22 utc jmettraux after 100 successful runs, I consider the test green
2010-09-30 12:19:35 utc jmettraux so that's at least 300 tests
2010-09-30 12:19:39 utc tosch_le will start as soon as the functional tests are through
2010-09-30 12:19:44 utc jmettraux I interrupt with ctrl-c
2010-09-30 12:20:12 utc jmettraux they make sure multi-worker don't conflict
2010-09-30 12:20:21 utc tosch_le hopefully i get them done in the next 40 minutes, have to leave then to the sound studio
2010-09-30 12:20:40 utc jmettraux are you playing an instrument ?
2010-09-30 12:21:40 utc tosch_le no, i will voice over some screencasts
2010-09-30 12:21:59 utc tosch_le it's not really a studio but an office of radio journalists
2010-09-30 12:22:07 utc jmettraux aaaaaaaaaaaaaah
2010-09-30 12:22:16 utc jmettraux english -> Deutsch ?
2010-09-30 12:22:19 utc tosch_le better mic, better sound
2010-09-30 12:22:24 utc tosch_le :-)
2010-09-30 12:22:26 utc tosch_le german
2010-09-30 12:24:03 utc jmettraux rails screencasts ;-) ?
2010-09-30 12:24:17 utc tosch_le unfortunately not
2010-09-30 12:24:29 utc tosch_le those ones are for scooterplan.net
2010-09-30 12:25:34 utc jmettraux looks interesting
2010-09-30 12:26:09 utc jmettraux rent-a-bike
2010-09-30 12:26:42 utc jmettraux I have a ct_2 weakness with dm and pgsql, now testing with mysql
2010-09-30 12:27:05 utc tosch_le functional tests still running here
2010-09-30 12:30:41 utc jmettraux ct_2 failing in the same way for mysql and pgsql
2010-09-30 12:30:46 utc jmettraux diving into the issue
2010-09-30 12:48:07 utc tosch_le my computer is … slow. functional tests still running
2010-09-30 12:49:22 utc jmettraux happy voice over !
2010-09-30 12:50:25 utc tosch_le functional tests result with ruby 1.9.2p0: http://gist.github.com/604522
2010-09-30 12:52:37 utc jmettraux ft_42 should work after http://github.com/jmettraux/ruote/commit/5293c2fe83d9f4770d78200a4224aca6d933d94d
2010-09-30 12:54:02 utc tosch_le forgot to update ruote this morning, just updated ruote-dm, sorry
2010-09-30 12:55:02 utc tosch_le will re-test tomorrow if you like
2010-09-30 12:55:15 utc tosch_le ct_0 is at #42
2010-09-30 12:55:26 utc jmettraux ct_2 should be the only one failing
2010-09-30 12:55:51 utc tosch_le will see, ct_1 and ct_2 are for tomorrow, too
2010-09-30 12:56:05 utc jmettraux OK, many thanks !!
2010-09-30 12:56:23 utc jmettraux if I solve it meanwhile, I will release (before your morning)
2010-09-30 12:56:24 utc tosch_le never mind, it's fun!
2010-09-30 12:56:34 utc tosch_le :-) sounds great
2010-09-30 12:57:01 utc tosch_le will you release rk, too or shall i do /that/ tomorrow?
2010-09-30 12:57:12 utc jmettraux you're the judge ;)
2010-09-30 12:57:43 utc tosch_le i'll do it, that way i may update the gemfile and readme as well
2010-09-30 12:57:49 utc jmettraux +1
2010-09-30 12:58:26 utc tosch_le #73 and no errors
2010-09-30 13:02:17 utc tosch_le ct_0: 105 runs, no errors on ruby 1.8.7 (2010-04-19 patchlevel 253) [i686-linux], MBARI 0x8770, Ruby Enterprise Edition 2010.02
2010-09-30 13:02:26 utc jmettraux :)
2010-09-30 13:02:33 utc tosch_le well done!
2010-09-30 13:03:44 utc tosch_le good night and happy bug hunting!
2010-09-30 13:03:57 utc jmettraux ciao !