ruote tmp/log_2012-01-04.html

2012-01-04 08:17:03 utc anb_ Hello everyone
2012-01-04 08:17:20 utc anb_ Happy new year
2012-01-04 08:20:23 utc jmettraux anb_: hello, happy new year !
2012-01-04 08:20:34 utc anb_ Thanks
2012-01-04 08:21:30 utc anb_ still got ruote running in production, about 5000 processes
2012-01-04 08:21:44 utc jmettraux old version I guess ?
2012-01-04 08:21:54 utc anb_ 2.2.0.3
2012-01-04 08:21:56 utc anb_ I think
2012-01-04 08:22:18 utc jmettraux ah, not so old
2012-01-04 08:22:27 utc anb_ it's not getting the love it deserves though... have no time to update or optimize
2012-01-04 08:22:36 utc anb_ just look at it when it crashes :)
2012-01-04 08:22:48 utc anb_ well, not exactly crashes
2012-01-04 08:22:56 utc anb_ more like bugs I introduced
2012-01-04 08:23:18 utc anb_ speaking of which
2012-01-04 08:24:03 utc anb_ brb
2012-01-04 08:39:38 utc anb_ I needed a non-blocking participant, so I modified a my consume method
2012-01-04 08:40:00 utc anb_ to reply_to_engine(workitem)
2012-01-04 08:40:06 utc anb_ when I want it non-blocking
2012-01-04 08:40:12 utc anb_ and @context.storage.put(doc)
2012-01-04 08:40:17 utc anb_ when I want it blocking
2012-01-04 08:40:36 utc anb_ that's correct ?
2012-01-04 08:41:17 utc anb_ assuming doc is doc = workitem.to_h
2012-01-04 08:41:42 utc jmettraux well
2012-01-04 08:42:26 utc jmettraux it's true that calling reply_to_engine(workitem) will immediately make the flow resume
2012-01-04 08:43:01 utc jmettraux it's true that calling reply_to_engine(workitem) "blocking"
2012-01-04 08:43:06 utc jmettraux ouch
2012-01-04 08:43:46 utc anb_ I'm confused :)
2012-01-04 08:43:48 utc jmettraux if you want a "blocking participant", you should take inspiration from the Ruote::StorageParticipant (even inherit from it)
2012-01-04 08:44:00 utc anb_ that's what I did actually
2012-01-04 08:44:09 utc jmettraux the second "it's true" is to be erased
2012-01-04 08:44:17 utc anb_ all right
2012-01-04 08:44:42 utc anb_ I did inherit from your StorageParticipant, but depending on the workitem parameters, I have it non blocking
2012-01-04 08:45:21 utc jmettraux you're overriding consume as well ?
2012-01-04 08:46:00 utc anb_ yes
2012-01-04 08:46:29 utc anb_ my reply_to_engine and storage.put are used in consume
2012-01-04 08:46:53 utc anb_ I thought it was working well but I'm having a little issue
2012-01-04 08:47:53 utc jmettraux https://gist.github.com/1559149 maybe
2012-01-04 08:50:51 utc anb_ not far : https://gist.github.com/1559159
2012-01-04 08:51:13 utc anb_ dumbed down a bit
2012-01-04 08:51:20 utc anb_ but relevant part is there
2012-01-04 08:52:48 utc jmettraux looks OK
2012-01-04 08:53:09 utc anb_ nice
2012-01-04 08:53:18 utc anb_ ok so my problem might be unrelated
2012-01-04 08:53:31 utc jmettraux what are the symptoms ?
2012-01-04 08:54:05 utc anb_ get a stack trace when opening a process in ruote-kit
2012-01-04 08:54:09 utc anb_ missing method
2012-01-04 08:54:12 utc anb_ let me show you
2012-01-04 08:55:26 utc phaeron sounds like the problem I have been facing , but triggered in a different way ?
2012-01-04 08:56:00 utc phaeron jmettraux: the process without root expression thing
2012-01-04 08:56:06 utc jmettraux phaeron: hello, happy new year
2012-01-04 08:56:17 utc phaeron jmettraux: hello , happy new year to you too :)
2012-01-04 08:56:25 utc jmettraux :-)
2012-01-04 08:56:47 utc jmettraux let's wait to get the whole picture
2012-01-04 08:56:57 utc anb_ https://gist.github.com/1559173
2012-01-04 08:57:49 utc jmettraux anb_: do you know which commit number your ruote version is at ?
2012-01-04 08:57:56 utc jmettraux commit hash
2012-01-04 08:58:25 utc anb_ I'll try to find out. Not very proud of how I deployed it
2012-01-04 08:58:41 utc jmettraux anb_: dinner here, I'll be back in one hour
2012-01-04 08:58:50 utc anb_ bon appetit
2012-01-04 08:59:02 utc jmettraux merci
2012-01-04 08:59:19 utc phaeron ok I'll send email about that issue
2012-01-04 09:01:38 utc anb_ ruote-7790c6a410f6
2012-01-04 09:01:46 utc anb_ that's what I have in my bundler directory
2012-01-04 09:01:59 utc anb_ but I got ruote-7790c6a410f6 and ruote-a5b395f18f87
2012-01-04 09:02:10 utc anb_ how to know which one is used by ruote-kit ?
2012-01-04 09:03:40 utc anb_ the line in question is @expressions.any? ? @expressions.first.fei.wfid : @errors.first.fei.wfid, so that means "first" is Nil
2012-01-04 10:00:53 utc jmettraux anb_: back
2012-01-04 10:02:34 utc jmettraux it seems that @expressions and @errors are empty
2012-01-04 10:04:35 utc anb_ ha
2012-01-04 10:05:11 utc jmettraux weird case
2012-01-04 10:05:18 utc anb_ it seems that the process is not in ruote-kit process list
2012-01-04 10:05:27 utc anb_ but when I list the workitems
2012-01-04 10:05:36 utc anb_ and open related process, I get this crash
2012-01-04 10:06:17 utc jmettraux the workitem is an orphan then ?
2012-01-04 10:06:25 utc anb_ I think so
2012-01-04 10:07:05 utc anb_ since the participant replied to the engine, the process continued (and died) but the workitem is still in storage
2012-01-04 10:07:08 utc anb_ could it be ?
2012-01-04 10:07:11 utc jmettraux it's my fault, I shouldn't go stop you before you get into this issue
2012-01-04 10:07:16 utc jmettraux exactly
2012-01-04 10:13:57 utc anb_ ok, I see
2012-01-04 10:14:00 utc jmettraux anb_: I will fill an issue and fix it on master, hopefully you can backport that as is
2012-01-04 10:14:28 utc anb_ thanks a lot
2012-01-04 10:16:17 utc jmettraux https://github.com/kennethkalmer/ruote-kit/issues/15
2012-01-04 10:16:33 utc jmettraux I cannot work on it tonight, hopefully I can have a go at it tomorrow
2012-01-04 10:19:04 utc anb_ sure, no problem
2012-01-04 10:19:10 utc anb_ thanks for filling the issue
2012-01-04 10:19:51 utc jmettraux you're welcome
2012-01-04 10:19:55 utc jmettraux lots of snow there ?
2012-01-04 10:22:16 utc anb_ yes we had a lot the last two weeks
2012-01-04 10:22:23 utc anb_ but it rained a bit
2012-01-04 10:22:42 utc anb_ how about down there ?
2012-01-04 10:23:22 utc jmettraux it snowed all morning, but at 2pm it was all gone :-(
2012-01-04 10:23:36 utc anb_ too warm ?
2012-01-04 10:23:51 utc anb_ btw, how would you go about deleting the failed orphans manually ?
2012-01-04 10:24:17 utc jmettraux fs storage ?
2012-01-04 10:24:35 utc phaeron snowed +2cm in a couple of hours and it's icing over since then
2012-01-04 10:25:06 utc anb_ fs, yes
2012-01-04 10:25:20 utc jmettraux phaeron: how many hours of sunlight these days ?
2012-01-04 10:26:00 utc jmettraux anb_: you can locate the workitem's json file in the hierarchy and delete it
2012-01-04 10:26:42 utc jmettraux that's all, no need to restart[
2012-01-04 10:26:52 utc anb_ oki
2012-01-04 10:27:35 utc phaeron 5 or 6
2012-01-04 10:27:43 utc phaeron but it's not real sunlight
2012-01-04 10:29:19 utc jmettraux the transition must have had been extreme
2012-01-04 10:30:10 utc phaeron yep
2012-01-04 10:30:34 utc jmettraux phaeron: about your new ruote-kit patch, how about something like = @process.launched_time ? Time.parse(@process.launched_time).localtime.strftime('%Y/%m/%d %H:%M:%S') : '-'
2012-01-04 10:32:35 utc phaeron ok , one less line :)
2012-01-04 10:34:34 utc phaeron jmettraux: I just saw your comment on the ruote branch as well, it was breaking without that line , so it's probably not equivalent. I'll test again without that line just to be sure
2012-01-04 10:53:07 utc jmettraux phaeron: if you want to rework your ruote-kit pull request I'll wait, else I'll take it as is
2012-01-04 10:53:49 utc phaeron jmettraux: I'll have to test it again on the server that has those broken processes
2012-01-04 10:54:02 utc phaeron so I can't do it immediately , but I can do it later today
2012-01-04 10:54:32 utc phaeron code style guidelines are enforced by upstream so if you say do it , I will :)
2012-01-04 10:56:15 utc jmettraux ok, let's not waste time, I'm taking your pull request as is
2012-01-04 10:56:42 utc phaeron ok , thanks :)
2012-01-04 10:57:07 utc phaeron I'll test the ruote one without the line and then we can discuss and I'll do a pull request
2012-01-04 10:57:33 utc phaeron anyway gtg now , will be back online in a couple of hourse
2012-01-04 10:57:40 utc phaeron *hours
2012-01-04 10:57:55 utc jmettraux ok
2012-01-04 11:08:30 utc lbt jmettraux: happy new year to you too :) I think we'll be OK with an upgrade. Will look at the changelog and check with phaeron.
2012-01-04 11:22:22 utc jmettraux lbt: ah great, thanks !