[08:33:00] <mindactuponmind>	 hey, I read online that the wikimedia app was based on phonegap, but when I got the code, it seemed to be native ios with no phonegap.
[08:33:03] <mindactuponmind>	 Is this correct?
[08:42:28] <multichill>	 Good whatever it is :-)
[08:42:43] * multichill  is in timezone limbo

[08:46:43] <mindactuponmind>	 OK, I've solved the mystery....  I downloaded the wrong code (wikipedia-iphone instead of WikipediaMobile).
[08:46:55] <mindactuponmind>	 I got the correct code, and it does in fact use phonegap.
[16:48:47] <tfinc>	 greetings allj
[16:48:50] <tfinc>	 its good to be back
[16:49:03] <awjr>	 hi tfinc
[16:49:09] <awjr>	 good vacation?
[16:49:48] <tfinc>	 it was a damn good time
[16:49:52] <awjr>	 :D
[16:52:48] <flyingclimber>	 hangout is up https://plus.google.com/hangouts/_/69105685903c1aa48b3524b0875f564c85bb08d1?authuser=1&hl=en
[16:53:53] <flyingclimber>	 MaxSem:  awjr --^
[16:54:02] <awjr>	 thanks be there in a sec
[16:55:45] <flyingclimber>	 awjr: excellent
[16:56:19] <flyingclimber>	 jdlrobson: were hanging out in the standup area. come by
[17:08:55] <tfinc>	 MaxSem:  awjr. thanks for the standup updates about the db move. looking at http://www.mediawiki.org/wiki/Wiki_Loves_Monuments_mobile_application/ToDo i'm vague on where we actually are. is there a mail in my inbox that summarizes it ?
[17:13:38] <awjr>	 tfinc: there are a couple of components that are still open 1) Max has gotten the API stuff running on labs, which we need to perf test today to get metrics to ops 2) in the same vein, ops is pushing us to host this in labs - they have some decommissioned boxes we might be able to use to host it, but they are waiting for us to get them more specific #s about our needs 3) building out the admin_tree portion of the API is nearly done, pending som
[17:14:14] <MaxSem>	 tfinc, updated. I hope to finish the first bullet point today
[17:14:44] <MaxSem>	 do we have some Labs availability metrics?
[17:15:12] <MaxSem>	 for example mobile-testing was inaccessible for a few weeks, then suddenly returned from life
[17:15:25] <MaxSem>	 was it something within labs itself,
[17:15:26] <MaxSem>	 ,
[17:15:28] <MaxSem>	 ?
[17:15:39] <awjr>	 those are good questions for Ryan_Lane
[17:16:39] <MaxSem>	 firs of all, did anyone on this channel actually fix the instance, or it became accessible again "by itself"?
[17:20:14] <multichill>	 awjr: Hi
[17:20:20] <multichill>	 Probably more on topic here.
[17:20:22] <awjr>	 multichill: hello
[17:20:24] <awjr>	 :)
[17:20:36] <MaxSem>	 okay, not me, not you. not Patrick?
[17:20:45] <awjr>	 MaxSem: not me
[17:21:00] <multichill>	 Say for example us -> us-ca -> Orange County
[17:21:09] <multichill>	 We have several of those, see https://en.wikipedia.org/wiki/Orange_county
[17:21:14] <MaxSem>	 preilly, you didn't fix mobile-testing either?
[17:22:04] <multichill>	 awjr: And is the id persistent?
[17:22:25] <awjr>	 multichill: i think that depends on what you mean by persistent
[17:22:43] <MaxSem>	 these ids get reassigned on next rebuild
[17:22:44] <awjr>	 the id will persist until data gets updated
[17:22:49] <awjr>	 eha
[17:22:51] <awjr>	 er
[17:22:51] <awjr>	 yaeh
[17:24:41] <multichill>	 hmmmm
[17:24:45] <awjr>	 multichill: i still don't totally follow what you're getting at. to get to Orange County in the API, you need to know the parent admval (eg us-ca)
[17:25:29] <multichill>	 And if you go one level deeper?
[17:25:37] <multichill>	 So everything under Orange County?
[17:25:51] <MaxSem>	 refer to it by id, and it will work
[17:26:32] <awjr>	 well, there is no admval = 'Orange County' - at least for CA the value looks like '[[Orange County, California]]'
[17:30:57] <awjr>	 multichill i hear your point though - you are concerned about locations that share the same level and the same exact name
[17:34:30] <awjr>	 multichill: can you explain the use case(s) of this API?
[17:35:05] <gerrit-wm>	 New patchset: Jdlrobson; "reduce height of navigation bar so users eyes don't bleed" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/16428
[17:38:28] <multichill>	 awjr: The primary usecase is browsing in the mobile app
[17:38:44] <multichill>	 To find the items in a region
[17:40:36] <awjr>	 multichill: right, so you need to know the parent adm level and the parent adm name to find the subdivisions within a given region
[17:41:05] <multichill>	 Would it be bad to leave the breadcrumbs in the url? http://toolserver.org/~erfgoed/api/api.php?action=adminlevels&adm0=us&adm1=us-mo&adm2=Orange_County ?
[17:41:58] <multichill>	 MaxSem: Not sure if it expands like that, could also be [[Orange County, California|Orange County]] -> Orange County
[17:42:25] <awjr>	 multichill: i dont understand what you mean by 'leave breadcrumbs in the url'
[17:42:59] <multichill>	 Like https://en.wikipedia.org/wiki/Breadcrumb_%28navigation%29
[17:43:19] <multichill>	 us -> us-ca -> orange county -> ...
[17:44:08] <awjr>	 multichill i understand what a breadcrumb is, but are you suggesting that the requestor will need to list out all parents/grandparents/etc of what they are looking for?
[17:47:45] <MaxSem>	 who'll be implementing browsing on the app side?
[17:48:09] <MaxSem>	 jdlrobson, will it be you? ^^
[17:48:13] <tfinc>	 awjr:  http://www.mediawiki.org/wiki/Mobile_default_for_sister_projects#Sibling_project_automatic_redirect_progress_matrix needs an update
[17:50:13] <tfinc>	 philinje: we need a date for the commons mobile default move.
[17:50:21] <multichill>	 Yes. Max would be 4 (adm0-4)
[17:50:25] <tfinc>	 philinje: how much time do you need in order to prep for it ?
[17:50:38] <awjr>	 tfinc just updated
[17:50:56] <multichill>	 awjr: And based on the admX you know the level
[17:52:20] <awjr>	 multichill: right. so the problem is that you can have multiple locations with the same name at the same level
[17:52:26] <multichill>	 yep
[17:52:41] <awjr>	 multichill but i think before we can figure out the smart way to deal with that, we need to understand how exactly the api is going to be uesd
[17:52:45] <multichill>	 And exposing an internal non-persistent id might not be the best solution
[17:53:19] <multichill>	 AFAIK (jdlrobson can correct me) we have two options when you start:
[17:53:29] <multichill>	 * By current location (we have that right now)
[17:53:37] <multichill>	 * Navigate (that's now the campaigns)
[17:53:46] <awjr>	 for instance if the app will be making consecutive requests, drilling down from adm0 to adm4 might require a different approach than if the app just makes a blind request for children of adm3
[17:53:51] <multichill>	 Navigate will show all the countries (adm0)
[17:54:10] <multichill>	 A user clicks a country (us) and gets presented the next level, etc
[17:54:21] <awjr>	 yeah, so navigate involves drilling down
[17:54:42] <awjr>	 which could have a solution like what i think you wre getting at with the 'breadcrumb' in the URL?
[17:54:44] <multichill>	 Yes, ddddddrrrrrrilling
[17:54:59] * MaxSem  unplugs multichill

[17:55:05] <tfinc>	 awjr: mw.org works but not for the main page. do we know why? i bet its a redirect issue
[17:55:12] <multichill>	 yup. So no parameter is the country, adm0 set gives a list of adm1 etc etc
[17:55:42] <multichill>	 So having to keep track of the upper levels shouldn't be a problem
[17:56:11] <preilly>	 tfinc: the redirect only happens for /wiki/ requests
[17:57:43] <gerrit-wm>	 Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/16428
[17:59:07] <tfinc>	 preilly: i still don't get it even if i go to http://www.mediawiki.org/wiki/MediaWiki
[17:59:16] <awjr>	 tfinc mediawiki.org hasn't been enabled
[17:59:23] <awjr>	 for automattic redirection i maen
[17:59:32] <awjr>	 oh but i said it was on the matrix
[17:59:35] <awjr>	 let me fix that
[18:00:01] <awjr>	 gah i screwed up the matrix
[18:00:03] <awjr>	 i hate wiki tables
[18:02:04] <awjr>	 ok should be fixed
[18:05:02] <tfinc>	 awjr: thanks
[18:05:25] <MaxSem>	 awjr, you have much better ping to Labs than me (I hope) - can you run the load test from your machine>
[18:05:26] <MaxSem>	 ?
[18:07:13] <awjr>	 MaxSem im averaging 142ms ping to labs
[18:07:27] <MaxSem>	 I'm 200+
[18:07:37] <MaxSem>	 we could spam it concurrently:)
[18:07:38] <awjr>	 how were you performing the load test?
[18:08:16] <MaxSem>	 tools/stress-test.php in the WLM repo
[18:09:02] <MaxSem>	 ehm, it ceased to return anything for me
[18:10:43] <awjr>	 MaxSem have you ever used ab before?
[18:11:17] <jdlrobson>	 sorry only just catching up… are we talking about wiki loves monuments and drilling down from a country to a specific city multichill / awjr ?
[18:11:21] <awjr>	 you can do a better stress test with it than with what's in stress-test.php, as you can make concurrent connections
[18:11:24] <jdlrobson>	 If so I haven't thought much about this yet
[18:11:28] <awjr>	 jdlrobson yes
[18:11:33] <multichill>	 exactly
[18:11:43] <multichill>	 That's going to be a nice UI challenge
[18:12:46] <MaxSem>	 awjr, my test also analyses what the server returns
[18:13:29] <MaxSem>	 and makes random, plausible requests, not just hammering the same URL
[18:13:46] <jdlrobson>	 UI challenge - in what sense?
[18:15:15] <multichill>	 jdlrobson: The levels differ per country
[18:15:22] <MaxSem>	 awjr, but yeah - you'll have to launch several instances of it to test:)
[18:15:34] <multichill>	 Not sure if it's a good experience everywhere to completely drill down
[18:15:55] <jdlrobson>	 I see what you mean
[18:16:19] <jdlrobson>	 In the app clicking back should always take you up to the level you were before
[18:16:34] <tfinc>	 preilly: is this still the most active ticket on this https://rt.wikimedia.org/Ticket/Display.html?id=3221 ?
[18:16:40] <preilly>	 tfinc: yes
[18:19:10] * tfinc  hates rt with a passion

[18:19:34] <multichill>	 jdlrobson: And maybe an option to just show the items when you're not completely at the end of the tree?
[18:19:37] <jdlrobson>	 tfinc: I don't seem to have an account on RT… I'm sure I did but I can't login. Who should I speak to?
[18:19:39] * preilly  agrees wholeheartedly 

[18:20:08] <jdlrobson>	 multichill: - I guess it depends on how big the data set is
[18:20:23] <jdlrobson>	 if you have 1000 monuments, more drilling down makes sense
[18:20:32] <multichill>	 Let's worry about that later, first the simple drilling down.
[18:20:37] <jdlrobson>	 but if you only have 10 monuments we should show multiple levels
[18:20:39] <multichill>	 We can ask all countries for feedback
[18:20:41] <jdlrobson>	 but multichill completely agree :D
[18:21:14] <tfinc>	 jdlrobson: i can open a ticket about it. do you need one now ?
[18:21:50] <jdlrobson>	 tfinc: I just feel left out when people talk about RT and I can't see what on earth they are talking about
[18:23:51] <tfinc>	 well if our rt wasn't a closed system then we wouldn't have this problem. /me looks at Ryan_Lane and rest of ops
[18:24:59] <tfinc>	 Ryan_Lane: will we ever get an instance of rt that is world readable for most queues ?
[18:25:00] <jdlrobson>	 hisses at Ryan_Lane: ;-)
[18:25:24] <Ryan_Lane>	 probably not
[18:25:32] * preilly  boo hises at Ryan_Lane and jdlrobson for hissing 

[18:25:33] <Ryan_Lane>	 request an account for him
[18:25:39] <awjr>	 MaxSem: ab supports concurrency and will give you much more detailed results
[18:25:53] <Ryan_Lane>	 ab kind of sucks :)
[18:26:11] <awjr>	 Ryan_Lane can you suggest a better benchmarking tool?
[18:26:12] <Ryan_Lane>	 it's good for small tests, though
[18:26:13] <jdlrobson>	 is there any reason we couldn't have a bot that pushes RT tickets to bugzilla ?
[18:26:30] <MaxSem>	 awjr, can it also issue random requests that follow the API logics?
[18:26:34] <Ryan_Lane>	 or you guys could just get accounts
[18:26:43] <Ryan_Lane>	 awjr: ab is fine for small tests
[18:27:02] <tfinc>	 awjr: MaxSem: are you guys doing these perf tests againt's a labs instance ?
[18:27:08] <awjr>	 MaxSem you can script around it to do that
[18:27:10] <awjr>	 tfinc yes
[18:27:12] <Ryan_Lane>	 heh
[18:27:16] <Ryan_Lane>	 labs is worthless for performance tests
[18:27:24] <preilly>	 awjr, MaxSem: not a great idea
[18:27:48] <MaxSem>	 preilly, that's what Asher told us to do XD
[18:27:59] <tfinc>	 awjr: virts aren't going to give you a good perf read . it'll way under report what you can actually do
[18:28:19] <MaxSem>	 my test was written specifically to perform different requests
[18:28:50] <jdlrobson>	 MaxSem: did you see Nasir Khan's mail about enlisting Bangladesh ? I'm not quite sure what is process is...
[18:29:07] <Ryan_Lane>	 the biggest problem with testing in labs is that you won't get consistent results
[18:29:35] <awjr>	 yeah that makes sense. i've learned this before about virts but had forgotten
[18:29:56] <Ryan_Lane>	 based on scheduling, IO load, and CPU load you'll get either great results or shitty results
[18:30:17] <Ryan_Lane>	 and performance tests depend on the fact that your results will be similar every run
[18:31:28] <awjr>	 MaxSem where did the 10 req/s number come from (for WLM app traffic)? and was that number for API requests or for the app in general?
[18:32:36] <MaxSem>	 awjr, 160k photos last year * 10 times possible growth * 10 search requests per upload / contest duration = 7 requests/sec
[18:32:51] <MaxSem>	 I've roundeed it up to 10 just to make sure
[18:33:27] <tfinc>	 jdlrobson: i cut an rt ticket to get you an account
[18:34:02] <awjr>	 MaxSem do you have any idea if the traffic was consistent last year, or what spikes were like?
[18:34:27] <MaxSem>	 awjr, there was no app last time
[18:35:02] <MaxSem>	 two 10x coefficients should take care of uneven distribution
[18:36:35] <awjr>	 MaxSem ok and how close are we to having complete monument data?
[18:36:50] <MaxSem>	 this is a question to multichill
[18:37:24] <MaxSem>	 jdlrobson, http://lists.wikimedia.org/pipermail/mobile-l/2012-July/005634.html ?
[18:38:10] <multichill>	 awjr: https://commons.wikimedia.org/wiki/Commons:Monuments_database/Statistics is what we have and https://bugzilla.wikimedia.org/show_bug.cgi?id=38314 is for what needs to happen
[18:38:14] <jdlrobson>	 yup MaxSem
[18:38:28] <multichill>	 Worse/best case double size would be my guess
[18:40:18] <tfinc>	 https://rt.wikimedia.org/Ticket/Display.html?id=3221 updated
[18:40:53] <MaxSem>	 jdlrobson, replied
[18:41:03] <awjr>	 ok MaxSem what other numbers can we give ops than aside from disk usage/growth and traffic requirements?
[18:42:15] <MaxSem>	 awjr, I think a "usual" DB server should be enough. but that's pretty vague
[18:43:18] <jdlrobson>	 Thanks MaxSem
[18:43:32] <awjr>	 MaxSem: let asher know the expected traffic and how much disk storage we'll need; i think we're going to get whatever box they have available
[18:44:48] <MaxSem>	 the 10 requests/sec numver has already been mentioned, disk space requirements are negligible
[18:45:28] <MaxSem>	 so we should just request best of whatever we have? :P
[18:46:14] <awjr>	 MaxSem just reiterate the traffic and come up with the storage requirements for the rt ticket so we can get it closed and move on.
[18:46:51] <awjr>	 in the meantime we should keep moving forward with getting the labs instance set up the way we will want to get whatever box we get to bet set up
[18:47:05] <multichill>	 Guys, I'm going to bed. Haven't slept last night :P
[18:47:48] <awjr>	 goodnight multichill
[19:02:26] <tfinc>	 awjr: i taked with phil about the commons sibling move and he doesn't want us to block on product for it. he wont have any time to look at it so i say we just schedule it
[19:02:48] <awjr>	 sounds good to me
[19:02:49] <tfinc>	 philinje: awjr : how are we doing at getting metrics for these moves ?
[19:03:17] <tfinc>	 i'm eager to see how much traffic their actually getting
[19:07:28] <awjr>	 tfinc good question - are those stats automatically generated somewhere like on stats.w.o?
[19:08:13] <tfinc>	 awjr: http://dumps.wikimedia.org/other/pagecounts-ez/projectcounts/readme.txt
[19:08:41] <tfinc>	 thats how we used to count them
[19:08:45] <tfinc>	 drdee would know best
[19:10:07] <tfinc>	 preilly: how are we doing on the j2me app ?
[19:11:15] <preilly>	 tfinc: the code is looking clean
[19:11:31] <preilly>	 tfinc: they also seem to be making good progress as well
[19:12:00] <preilly>	 tfinc: the last code I saw was from 3 days ago
[19:13:07] <preilly>	 tfinc: the last thing I saw was the added ability to select a new language to view an article in and the Language Dialog
[19:13:22] <caxthelm>	 preilly, tfinc: I'm happy at our progress.  The Aplha is due this week but I feel we have already surpassed it.
[19:13:31] <caxthelm>	 alpha*
[19:13:53] <preilly>	 caxthelm: yeah I'd agree
[19:15:20] <caxthelm>	 preilly: barring any issues on other things that need my assistance, I should continue to stay well ahead of project estimates.
[19:17:50] <caxthelm>	 tfinc: do you all have a set of test pages you like to use?  I put the build into our test with a request to look for HTML oddities and problems, but that might be expedited if you all have a list of articles you all like to test with.
[19:18:12] <tfinc>	 caxthelm: hmmm
[19:19:56] <tfinc>	 caxthelm: http://en.wikipedia.org/wiki/Barack_Obama works nicely as a really long article
[19:23:09] <MaxSem>	 as the matter of fact, it's our trademark "pushed to the limits" page
[19:23:27] <caxthelm>	 ok, any ones with a lot of odd formatting?
[19:23:50] <tfinc>	 caxthelm: hmm
[19:24:06] <tfinc>	 jdlrobson: whats your favorite odd formatting page on the wiki ?
[19:24:22] <jdlrobson>	 odd in which sense?
[19:24:40] <caxthelm>	 barack Obama has a couple of tables right at the front...that is going to be fun.
[19:25:47] <jdlrobson>	 https://pt.wikipedia.org/wiki/GRCSES_Acad%C3%AAmicos_do_Tucuruvi  < this page makes my eyes bleed
[19:26:43] <tfinc>	 caxthelm: --^
[19:26:51] <preilly>	 jdlrobson: probably subway map pages
[19:27:30] <jdlrobson>	 ohh that sort of thing
[19:28:01] <preilly>	 jdlrobson: like http://en.wikipedia.org/wiki/Baker_Street_and_Waterloo_Railway for example
[19:28:07] <jdlrobson>	 http://en.wikipedia.org/wiki/New_Haven_Line
[19:28:29] <jdlrobson>	 mm that did use to have a map but not anymore..
[19:28:34] <preilly>	 jdlrobson: like this section http://en.wikipedia.org/wiki/Baker_Street_and_Waterloo_Railway#Opening
[19:28:45] <jdlrobson>	 ohh yeh it's under the route map
[19:28:48] <jdlrobson>	 in the info box
[19:28:56] <jdlrobson>	 yep preilly ^
[19:29:14] <jdlrobson>	 also anything with maps on it
[19:29:17] <caxthelm>	 jdlrobson: lovely, am I seeing a bright red backing for the tables on that academicos page?
[19:29:21] <jdlrobson>	 where a dot is placed where the location is
[19:29:31] <jdlrobson>	 yep caxthelm - that's what makes your eyes bleed :)
[19:29:47] <preilly>	 or directly at http://en.wikipedia.org/wiki/Template:New_Haven_Line_map
[19:31:39] <caxthelm>	 I'm not terribly worried about images, they get collapsed into a button & dialog to save memory.  I'll double check that maps are treated the same.
[19:32:55] <MaxSem>	 Ryan_Lane, what would you say about running something production-ish on Labs?
[19:33:19] <preilly>	 MaxSem: he would say no
[19:33:47] <Ryan_Lane>	 indeed. I'd say no
[19:33:51] <Ryan_Lane>	 it's not meant for production
[19:34:44] <MaxSem>	 so was my opinion, too
[19:35:45] <MaxSem>	 what does "get your deploy stack configuration initially codified in labs and don't worry about hardware"
[19:35:50] <MaxSem>	 mean?
[19:36:15] <Ryan_Lane>	 do the initial installation in labs
[19:36:19] <Ryan_Lane>	 then puppetize it
[19:36:29] <Ryan_Lane>	 we'll handle the hardware
[19:37:17] <caxthelm>	 Thank you all, I've added the pages to our testing.
[19:37:19] <MaxSem>	 what needs to be puppetized, exactly? box config (mysql, apache, etc.) or the web app itself?
[19:37:56] * tfinc  ponders food

[19:39:45] <tfinc>	 MaxSem: just ask the ops team. they will know best what currently has a puppet config vs us needing to write new ones
[19:40:28] <MaxSem>	 I'm trying to collect as much information as possible to ask a right question:P
[19:41:03] <tfinc>	 fair enough
[19:45:24] <awjr>	 MaxSem i've looked over some of the puppet configs for misc boxes - it looks like we'll need to specify packages that we'll need installed, provide config files (like for apache), and define any specific permissioning - but i am not particularly familiar with puppet
[19:45:51] <MaxSem>	 me neither
[19:46:45] <MaxSem>	 though default Apache config  will be sufficient - just read from /var/www/ which will contain a single directory with a single symlink in it
[19:57:43] <awjr>	 MaxSem we can probably use already existing puppet definitions for apache and mysql on top of whatever other custom stuff we'll need
[19:58:55] <caxthelm>	 MaxSem: can you recommend which API we should be using to get a list of languages supported? (i.e. http://meta.wikimedia.org/wiki/List_of_Wikipedias)
[19:59:29] <caxthelm>	 Preferably something a low memory device can paginate through.
[20:00:29] <MaxSem>	 action=sitematrix
[20:00:51] <MaxSem>	 awwwwwwww
[20:00:59] <MaxSem>	 it doesn't support pagination
[20:01:19] <MaxSem>	 nor even selection by group
[20:02:03] <MaxSem>	 caxthelm, what is the memory footprint you're aiming at? (not just this information, whole app)
[20:03:53] <caxthelm>	 MaxSem, the smallest device has max phone memory of 507kb.  I'm not aiming THAT low right now, but we have to keep it in mind to make porting go smoothly
[20:04:46] <MaxSem>	 heh
[20:05:03] <MaxSem>	 I have experience of programming for 2kb chips :]
[20:05:31] <caxthelm>	 well doing a 10kb html page has crashed the c3-00 (the first targeted device)
[20:05:44] <caxthelm>	 brb, meeting
[20:05:57] <preilly>	 hmm
[20:06:05] <MaxSem>	 s4 sucks:(
[20:06:10] <MaxSem>	 *S40
[20:11:06] <tfinc>	 MaxSem: is may suck feature wise compared to todays phones but the majority of the world still uses them
[20:14:36] <MaxSem>	 caxthelm, we could try adding pagination to action=siteinfo
[20:14:49] <MaxSem>	 ping me when you're back
[20:26:22] <caxthelm>	 MaxSem, I am back.
[20:26:45] <MaxSem>	 caxthelm, what information from http://meta.wikimedia.org/wiki/List_of_Wikipedias do you need, exactly?
[20:26:52] <caxthelm>	 MaxSem, action=siteinfo?
[20:27:19] <MaxSem>	 action=siteinfo will eat all your RAM:)
[20:27:38] <tfinc>	 caxthelm: i just gave the latest checked in build a spin in the J2ME simulator. nice work
[20:28:22] <tfinc>	 philinje: if your still having trouble running it in on device i suggest giving it a go in the simulator to get an idea of look and functionality
[20:28:57] <MaxSem>	 caxthelm, I'm trying to figure out how to feed you only inforrmation you want. for this, I need to know what you need:)
[20:29:18] <MaxSem>	 do ypu need lust a list of Wikipedia languages?
[20:29:25] <MaxSem>	 owww
[20:29:30] <MaxSem>	 s/lust/just/
[20:30:03] <caxthelm>	 tfinc: a tester is looking into what could be happening on the n95.  MaxSem: yeah I just need to know what languages (name and code) the user is allowed to select from to display the entire wiki app in.
[20:32:51] <MaxSem>	 caxthelm, so from https://en.wikipedia.org/wiki/Special:ApiSandbox#action=sitematrix&format=json&smtype=language you need something like [{code:.., name:...}] ?
[20:33:36] <MaxSem>	 caxthelm, also - would pagination help you in this case?
[20:33:37] <caxthelm>	 yeah
[20:33:49] <caxthelm>	 pagination would be extremely helpful, yes
[20:35:47] <caxthelm>	 Best-case scenario would be to have the various languages sorted by region, hemisphere, or something to assist the user in finding the one they want.  Otherwise alphabetical by country code would work.
[20:36:39] <caxthelm>	 the application should already pull which language the phone is set to and default to that.
[20:38:01] <MaxSem>	 caxthelm, okay - I'll do it. region information is currenty absent, unfortunately
[20:38:38] <caxthelm>	 MaxSem, wonderful thank you sir.
[20:39:45] <awjr>	 MaxSem have you given any thought to logging/monitoring for the WLM API host?
[20:42:49] <MaxSem>	 awjr, adding just request logging is easy, the question is whether we need to classify them
[20:49:54] <awjr>	 MaxSem: yeah, and we should also figure out getting monitoring set up as well
[20:50:11] <awjr>	 MaxSem have you already started working on puppet config, or should i get the ball rolling on that?
[20:50:50] <MaxSem>	 I'm writing a reply to RT right now
[20:51:06] <MaxSem>	 so we're both have no prior experience with puppet?
[20:51:36] <awjr>	 correct
[20:51:51] <awjr>	 although i think preilly has some and we should be able to get some guidance from ops as needed
[20:53:02] <MaxSem>	 "dear ops, what should we start to RTFM?" :)
[20:53:26] <awjr>	 heh
[20:53:45] <awjr>	 yeah i've been trying to go over what dcoumentation i can find on labswiki and wikitech and it's pretty thin
[20:54:22] <wikipedia-github>	 [WikipediaMobile] jdlrobson pushed 2 new commits to master: http://git.io/-8OZ2g
[20:54:22] <wikipedia-github>	 [WikipediaMobile/master] (bug 38535) Share article in proper language after interwiki - YuviPanda
[20:54:22] <wikipedia-github>	 [WikipediaMobile/master] Merge pull request #269 from yuvipanda/fix-interlang-sharing - Jon Robson
[20:54:45] <mw-jenkinsbot>	 Project WikipediaMobile - Nightly builds build #362: SUCCESS in 13 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/362/
[20:54:46] <mw-jenkinsbot>	 yuvipanda: (bug 38535) Share article in proper language after interwiki
[21:03:04] <MaxSem>	 preilly, I've prepared some notes regarding GeoData deployment: http://etherpad.wikimedia.org/GeoData
[21:20:03] <preilly>	 MaxSem: okay cool
[21:20:58] * awjr  slaps rt

[21:21:17] <awjr>	 that's better
[21:25:28] <MaxSem>	 hmm, if we start from http://dpaste.org/8IXKW/ ...
[21:32:50] <awjr>	 MaxSem: s/role/misc/g
[21:39:41] <preilly>	 awjr: why not just add a m record for donate?
[21:41:06] <awjr>	 preilly donate is a funky, special instance used in fundraising. it's not like your average MW instance.
[21:41:36] <awjr>	 we should coordinate with the fundraising team before enabling mobile anything for fundraising-related stuff
[21:42:28] <preilly>	 awjr: okay did you ping Asher to create a new binary and push it live for the redirector in squid?
[21:42:56] <awjr>	 preilly i did
[21:45:46] <preilly>	 awjr: okay great
[22:07:25] <jdlrobson>	 team are we deploying today?
[22:10:35] <awjr>	 not that i am aware of
[22:10:48] <awjr>	 jdlrobson are there changes that need to go out?
[22:11:03] <jdlrobson>	 There was some changes to the beta
[22:11:12] <jdlrobson>	 linSmith and pchang__ were keen to test
[22:11:53] <awjr>	 oh i see
[22:17:28] <awjr>	 jdlrobson: im not really able to get changes pushed out inside of our window right now since i wasn't prepared for this - if preilly or MaxSem can't pick this up right now, we should schedule something later in the week or wait til next Monday
[22:17:52] <jdlrobson>	 ok lets do it later in the week then - unless pchang__ needs more urgency
[22:17:56] <MaxSem>	 do we still have a window on Tuesdays?
[22:18:17] <MaxSem>	 hmm, we do
[22:20:39] <awjr>	 jdlrobson MaxSem I should be able to do it tomorrow
[22:20:42] <preilly>	 jdlrobson: we can just do it tomorrow
[22:46:30] <MaxSem>	 lol http://what-if.xkcd.com/1/
[22:47:21] <Platonides>	 I guess you saw Yoda?
[23:22:26] <tfinc>	 awjr: MaxSem how is the wlm db work wrapping up today ?
[23:22:39] <tfinc>	 do you guys need anything else from me for the puppet and profiling work ?
[23:22:55] <MaxSem>	 working on puppetization
[23:24:12] <awjr>	 tfinc max has been working on puppetization and i've been checking in/helping where i can while working on some changes for the admin_tree api
[23:25:01] <tfinc>	 how many componetns do we have to puppetize?
[23:27:35] <awjr>	 aside from getting the base system set up with apache, mysql, monitoring, logging, svn, users, there's the API code itself, configuration for it, and periodic execution of update scripts
[23:28:06] <MaxSem>	 we're not going to puppetize the API itself, it's too fluid
[23:29:32] <tfinc>	 can we re-use puppet configs from our own infrastructure for the basic building blocks (apache,mysql, etc ) ?
[23:30:22] <awjr>	 for the most part yes, but what all that is requires some untangling
[23:37:18] <tfinc>	 awjr: MaxSem what do you guys project as the lead time to get all the puppetization done?
[23:37:43] * MaxSem  have never done it before

[23:38:17] <tfinc>	 and im guessing ops wont let us do this without using puppet
[23:38:19] <MaxSem>	 today is over, tomorrow I will start testing what I've wrote so far
[23:39:24] <MaxSem>	 then it should be run past ops to figure out what else needs to be done
[23:39:26] <awjr>	 my guess is we'll probably need the rest of the week to get it done - since neither of us are familiar with puppet, it's going to require a fair amount of testing/review/back and forth with ops people
[23:40:39] <awjr>	 i dont think that's a whole week of working full time on it, mind you, just a full week for all the back-and-forth
[23:41:45] <awjr>	 once the puppetazation is done though, the whole thing will be pretty portable and should be trivial to make go on wahtever hardware we get
[23:42:03] <tfinc>	 thats a lot of time to use up on this
[23:42:35] <tfinc>	 for something that will drastically change when we find that people use these tools
[23:42:57] <tfinc>	 using a day to scope this is fine but a while week is too costly
[23:42:59] <tfinc>	 whole*
[23:43:12] <MaxSem>	 we could take a risk and leave it be on Toolserver
[23:43:35] <awjr>	 tfinc not a whole week's worth of work, just a week's time to get it wrapped up, but i hear what you're saying
[23:43:37] <tfinc>	 well. why can't we just use the stock puppet packages that our cluster uses for apache, mysql, etc and just do the rest by hand
[23:44:29] <preilly>	 or even the standard ubuntu packages for precise
[23:44:32] <MaxSem>	 yeah, we will definitely puppetize as little as possible
[23:45:56] <awjr>	 it would go a lot faster if either of us had puppet experience or if it all was better documented, but figuring out which bits of all of the different apache puppet declarations to use, etc requires untangling from reading through the manifests
[23:46:10] <tfinc>	 preilly: agreed. even simpler then. awjr MaxSem: is there any reason why we can't keep it that simple for just a two month run of this backend infrastructure ?
[23:46:25] <tfinc>	 awjr: thats why i like preilly's idea of just using stock precise with its own packages
[23:46:27] <MaxSem>	 fun stuff is that deploying WLM manually on an empty LAMP server with root access is just 30 minutes
[23:46:51] <preilly>	 exactly
[23:46:59] <tfinc>	 see i like that much better for a system were just going to re-architect
[23:47:00] <awjr>	 yeah, i have no problem with that but we'll have to circle back with ops; asher made it sound crucial that we puppetize the whole thing
[23:47:12] <awjr>	 i agree, it's a helluva lot easier
[23:47:34] <tfinc>	 this is no way in hell that were going to keep it in its current state
[23:48:43] <MaxSem>	 how soon after the end of the contest are we going to shut it down?
[23:49:24] <awjr>	 tfinc: im happy to push back on the rt ticket and make a case for not puppetizing
[23:49:57] <tfinc>	 preilly and i just talked to asher and he agrees that we should re-use whatever is available
[23:50:10] <tfinc>	 doing all this by hand will just take too much time
[23:50:32] <awjr>	 you mean to not puppetize?
[23:50:54] <tfinc>	 by hand from scratch*
[23:52:04] <awjr>	 yeah gotcha
[23:53:41] <MaxSem>	 so, just minimum minimorum of a puppet?
[23:56:50] <tfinc>	 yes, anything more is just a waste of time for a project that will only last for 2 months
[23:57:49] <tfinc>	 as this only has to live in production sept-oct