ruote log_2010-09-09

2010-09-09 08:35:48 utc dhf aloha. ok I give. I know you can get an array of all storage participants by their complete participant name like 'link_JUSTDESTINY' is it possible to get a list of storage participants that start with 'link_'?
2010-09-09 08:36:09 utc jmettraux hello
2010-09-09 08:37:14 utc jmettraux you'd probabaly have a unique storageparticipant
2010-09-09 08:37:34 utc dhf there could be hundreds of participants starting with 'link_' some could have the same name but most are unique. Imy storage partcipant is catchall
2010-09-09 08:37:45 utc jmettraux and then filter its content for all the workitems whose participant name starts with "link_"
2010-09-09 08:38:27 utc jmettraux engine.storage_participant.all.filter { |wi| wi.participant_name.match(/^link_/) }
2010-09-09 08:38:31 utc jmettraux or something like that
2010-09-09 08:39:45 utc dhf so get a list of all and filter? that was what I was thinking I had to go to but was hoping I did not need to read the all. but not a problem
2010-09-09 08:40:03 utc jmettraux let me look at the methods
2010-09-09 08:40:18 utc dhf nothing i saw is storage participant
2010-09-09 08:41:05 utc jmettraux http://github.com/jmettraux/ruote/blob/ruote2.1/lib/ruote/part/storage_participant.rb#L170-266
2010-09-09 08:47:29 utc jmettraux if those method do not suffice, please tell me and I'll find a better solution
2010-09-09 08:56:28 utc dhf the first works. the second is exact match only. not regex. It would return 'link_DESTINY' if I use that as the parameter. Get to find all that start with 'link_' I can use the all and filter method it is basically what the query does anyway.
2010-09-09 09:02:37 utc dhf the reason I am using the participant name like that is I need to find all taht start with link and do the link on the oldest participant then bring up each one with the same name and rehydrate each of them to finish. They are all pending for someone to establish the account. that must be a manual process. Once done, all pending need to be finished. Each one processes a unique document. It might work to name it 'link' only and use the query to find a
2010-09-09 09:02:38 utc dhf with teh same name
2010-09-09 09:05:42 utc jmettraux ok, I usually hand that to a specific user or group of users
2010-09-09 09:05:53 utc jmettraux pile-wise
2010-09-09 09:07:06 utc dhf I agree with group. How do you specify a group in ruote?
2010-09-09 09:07:15 utc jmettraux you don't
2010-09-09 09:07:27 utc jmettraux participant "group_alpha"
2010-09-09 09:07:40 utc jmettraux engine.storage_participant.by_participant('group_alpha')
2010-09-09 09:13:08 utc dhf then use the field and name query to rehydrate after the work is done on the first one? The rest just ned to go on to next step. the manual work has been done on the first one. I think that will work. I can actually a query based upon the name and the launch time.
2010-09-09 09:13:41 utc jmettraux what is "rehydrate" ? Refill the task list ?
2010-09-09 09:18:44 utc dhf sorry. yes reactivate the pending process. They are al waiting for some one to manually create the account. once done. I will go through each pending one for the same person and reply with the same account information in the workitem, so the rest of the process can complete. Prior to ruote it was stored in a DB work table. the person name as the primary key and the information necessary to finish processing after the account was created. nots of co
2010-09-09 09:18:44 utc dhf like the ruote approach better. it is just a pending stem in the process. That way it does not matter what the next steps are.
2010-09-09 09:19:39 utc jmettraux :-)
2010-09-09 14:01:32 utc lbt http://github.com/jmettraux/ruote/wiki seems to have gone away
2010-09-09 14:04:15 utc jmettraux lbt: hello, sorry
2010-09-09 14:04:23 utc jmettraux I have a copy here locally
2010-09-09 14:04:34 utc lbt :)
2010-09-09 14:04:57 utc jmettraux I triggered the upgrade to the new "gollum" wiki engine github haz
2010-09-09 14:04:58 utc jmettraux has
2010-09-09 14:05:08 utc lbt *poof*
2010-09-09 14:05:29 utc jmettraux yeah, it's reacting weirdly
2010-09-09 14:05:49 utc jmettraux I want to put it back in order soon
2010-09-09 14:06:03 utc lbt well, no rush here. I was just pointing to it
2010-09-09 14:06:10 utc jmettraux thanks !
2010-09-09 14:14:29 utc jmettraux lbt: how's the team progressing ?
2010-09-09 14:15:26 utc lbt Mainly getting up to speed I think
2010-09-09 14:15:42 utc lbt I'm expecting a review early next week
2010-09-09 14:16:10 utc lbt (that's on the perf testing side)
2010-09-09 14:16:55 utc lbt we've launched BOSS (ruote+rabbit+... really) internally and are starting to use it with triggers from the build system
2010-09-09 14:17:17 utc jmettraux :-)
2010-09-09 14:17:18 utc lbt So our test team are writing a participant as we speak
2010-09-09 14:17:43 utc lbt We've agreed to use it for the meego.com community build system QA processes
2010-09-09 14:18:10 utc lbt so when apps are built etc, they'll go through various reviews and auto-tests
2010-09-09 14:18:30 utc lbt including manual breakout if needed and auto-promotion
2010-09-09 14:18:34 utc lbt that's all BOSS
2010-09-09 14:18:49 utc lbt then the core distro will also be using it for their work.
2010-09-09 14:19:16 utc jmettraux so a small initial process to let the team get a feel for it ?
2010-09-09 14:19:17 utc lbt I need to setup some infra for them now... I need to do some security design work for that
2010-09-09 14:19:25 utc lbt yes... absolutely
2010-09-09 14:19:42 utc lbt we're also looking at a git-store for the processes
2010-09-09 14:20:13 utc lbt so we'll have a launcher... and it'll pull a process from git based on certain conditions and launch via AMQP
2010-09-09 14:20:24 utc lbt event driven
2010-09-09 14:20:34 utc lbt so that process can grow
2010-09-09 14:20:54 utc jmettraux cool
2010-09-09 14:21:00 utc lbt yeah... :)
2010-09-09 14:21:15 utc lbt We really aren't pushing ruote yet though
2010-09-09 14:21:27 utc lbt tiny processes, very linear
2010-09-09 14:22:03 utc jmettraux versioning the process, that's nice, I'm always suggesting it to people, but they "want to put it in the database"
2010-09-09 14:22:30 utc jmettraux there's a URI for anything these days....
2010-09-09 14:22:37 utc lbt he
2010-09-09 14:22:41 utc lbt h
2010-09-09 14:23:06 utc lbt https://gist.github.com/1d2371319225d9eec0b4
2010-09-09 14:23:53 utc jmettraux very nice !
2010-09-09 14:24:41 utc jmettraux tight
2010-09-09 14:25:05 utc lbt thankyou
2010-09-09 14:25:37 utc lbt it's the approach... clearly some work to do yet :)
2010-09-09 14:26:07 utc lbt I'm getting interested in a mechanism to do notification and human interaction
2010-09-09 14:26:39 utc lbt often a "yes/no" for a particular question (with enough context to allow them to make the decision)
2010-09-09 14:26:52 utc lbt bbias
2010-09-09 14:29:03 utc lbt the other area I'm starting to ponder is signing :)
2010-09-09 14:29:56 utc jmettraux aaah
2010-09-09 14:29:57 utc lbt having parts of the wi signed in some way to allow trust to be maintained in a distributed workflow.
2010-09-09 14:30:17 utc lbt but that's very much at the participant level I think
2010-09-09 14:30:40 utc lbt it's an interesting and fun problem
2010-09-09 14:30:46 utc jmettraux :-)
2010-09-09 14:31:59 utc lbt anyhow... back to some DB design ... :)
2010-09-09 14:32:12 utc lbt and some coffee.... mmmm ..... nice wife!
2010-09-09 14:32:20 utc jmettraux enjoy !
2010-09-09 14:32:27 utc lbt o/