[00:29:07] <Nemo_bis>	 Hm, wait, maybe it was a fi.wiki gadget?
[00:29:24] <Nemo_bis>	 (Today fi.wiki confused me quite a bit)
[08:26:15] <Al__>	 I can't clear my watch list :(  The error I get is: "Our servers are currently experiencing a technical problem. This is probably temporary and should be fixed soon. Please try again in a few minutes."
[10:39:05] <Cladis>	 hi
[10:39:45] <Cladis>	 what ratelimit is for creating pages from the sime ip? too lazy to try to find it in config
[11:51:55] <whym>	 ja.wikipedia has been keep overriding Typography Refresh. is there any problem with it going forward?
[11:57:18] <Glaisher>	 'edit' => array(
[11:57:19] <Glaisher>	 			// 8 ed./min per each non-autoconfirmed, or group thereof from same IP
[11:57:19] <Glaisher>	 			'ip' => array( 8, 60 ),
[11:57:21] <Glaisher>	 Cladis:  ^
[11:58:38] <Nemo_bis>	 whym: you'd better ask on the bug report you had filed
[11:59:30] <Nemo_bis>	 whym: also, it's about time to file a bug report on how to handle the thing in the 1.24 tarball (font changes were removed from 1.23 branch but IIRC they're still in master; they didn't get worked on or stabilised though, AFAIK)
[12:04:57] <Thehelpfulone>	 hmm, on https://wikimania2014.wikimedia.org/wiki/Main_Page#Democratic_Media - (scroll down) to the democratic media section, is there any way to get the images to squeeze onto one row?
[12:07:01] <Glaisher>	 Thehelpfulone:  smaller images?
[12:07:43] <Thehelpfulone>	 Glaisher: I mean is there a gallery to automatically resize them?
[12:08:23] <Thehelpfulone>	 oh sorry found it, that heightspx works
[12:08:25] <Glaisher>	 height parameter
[12:12:14] <Glaisher>	 er, Thehelpfulone it seems that it depends on screen resolution or size.. with your last change it's still on two different rows on my screen
[12:12:40] <Thehelpfulone>	 hmm, to be safe should I bring it down to something like 100px?
[12:14:50] <whym>	 Nemo_bis: thanks for the suggestion, I have added some update on it.
[12:15:10] <Glaisher>	 Thehelpfulone:  Too small. The caption on Heather's image is protruding upwards with 100px
[12:16:13] <Thehelpfulone>	 Glaisher: can you play around with it on your screen to see what works?
[12:21:28] <Glaisher>	 Thehelpfulone: Same result when it's one row. I guess there's no easy way to achieve a consistency across various screen sizes.
[12:21:50] <Glaisher>	 Also, the first and thw last sections are also appearing as two rows here. Better to keep it as it is, I think.
[12:22:16] <Thehelpfulone>	 ok, thanks
[15:16:53] <Josve05a>	 "4Nyfilmcritique thanked you for your edit on [No page]. 1 minute ago"   [No page]?
[15:18:16] <Reedy>	 Josve05a: What wiki?
[15:18:46] <Josve05a>	 enwp
[15:20:06] <Josve05a>	 Reedy - http://i.snag.gy/X0Pq5.jpg
[15:21:37] <Bsadowski1>	     10:14, July 24, 2014 Nyfilmcritique (talk | contribs | block) thanked Josve05a (talk | contribs | block)
[15:21:52] <Bsadowski1>	 Log doesn't say which page
[15:22:36] <Josve05a>	 What if the age has  been deleted? Would it still say which page it was?
[15:23:17] <Krenair>	 It's used when the Title is null
[15:23:25] <Krenair>	 Not sure when you might get that...
[15:24:04] <Josve05a>	 Well, I did get that.
[15:24:12] <Bsadowski1>	 Might be this?: https://en.wikipedia.org/wiki/Special:Undelete/Zest.md
[15:24:29] <Bsadowski1>	 It's when you interacted with him
[15:24:40] <Bsadowski1>	 So it looks like the page was deleted
[15:25:15] <Bsadowski1>	 (del/undel) 10:14, July 24, 2014 NawlinWiki (talk | contribs | block) deleted page Zest.md 
[15:25:18] <Josve05a>	 Is the  page named [[Zest.md]]? (I do not have "permission" to check that link... :P
[15:25:25] <Bsadowski1>	 But I do.
[15:27:12] <Josve05a>	 hmmm, I saw a bz-report about something about this...
[15:27:54] <Josve05a>	 50829
[15:29:05] <Josve05a>	 Thank you, Ijust wanted to know which page it was, so that seems to be the most likely.
[16:13:03] <jorn>	 short question about API:Search and generators...
[16:13:10] <jorn>	 do generators lose the order of a search?
[16:13:17] <jorn>	 search: http://en.wikipedia.org/w/api.php?format=json&action=query&redirects&list=search&srlimit=10&srsearch=cain%20abel
[16:13:41] <jorn>	 generator for those pages: http://en.wikipedia.org/w/api.php?format=json&action=query&redirects=&indexpageids=&prop=info|categories&clcategories=Category:Disambiguation%20pages&generator=search&gsrlimit=10&gsrsearch=cain%20abel
[16:16:22] <jorn>	 (if you do the same with gsrlimit=1 you'll see the first search result on top)
[16:39:51] <chrismcmahon>	 Reedy: do you remember why we added db slaves in beta labs? https://github.com/wikimedia/operations-mediawiki-config/commit/38990c671fd3b8d15f31a7c819e7bdd52ecef3ef I think it's causing https://bugzilla.wikimedia.org/show_bug.cgi?id=65486
[16:40:11] <Reedy>	 chrismcmahon: yup
[16:40:17] <Reedy>	 the simple answer is to match production
[16:40:37] <Reedy>	 There was a discussion to have a permenantly lagged slave to test race conditions
[16:40:44] <chrismcmahon>	 Reedy: I think the problem is that we don't run update.php in prod. 
[16:40:48] <Reedy>	 let me find the bug for it
[16:41:12] <chrismcmahon>	 Reedy:  having the db go read-only in beta several times per day is a pita.
[16:41:36] <Reedy>	 is that still happening after those schema changes?
[16:41:54] <chrismcmahon>	 Reedy: it happens pretty much daily, I can find you an example from overnight if you want
[16:42:20] <Reedy>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=38945
[16:42:36] <Reedy>	 That's not the bug I was actually thinking of though
[16:42:46] <chrismcmahon>	 Reedy: I happen to have it up, this was overnight yesterday with basically zero load on beta https://integration.wikimedia.org/ci/view/BrowserTests/job/browsertests-VisualEditor-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce/120/console
[16:44:10] <chrismcmahon>	 Reedy: yeah, that one from Niklas shouldn't happen where we do other work.
[16:44:50] <Reedy>	 There's a similar one logged by someone else
[16:46:45] <Reedy>	 chrismcmahon: found it
[16:46:45] <Reedy>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=60058
[16:46:58] <Reedy>	 Which then is needed for https://bugzilla.wikimedia.org/show_bug.cgi?id=57583
[16:48:06] <chrismcmahon>	 oh hell
[16:48:15] <Nemo_bis>	 which is what was filed for https://bugzilla.wikimedia.org/show_bug.cgi?id=38945
[16:48:32] <chrismcmahon>	 I know why 60058 was filed
[16:48:50] <Nemo_bis>	 which is needed to debug https://bugzilla.wikimedia.org/show_bug.cgi?id=46716 and its dozens siblings
[16:49:32] <chrismcmahon>	 Reedy Nemo_bis I think the point was to have replag but not have the beta db become readonly several times per day.
[16:49:42] <Reedy>	 haha, right :)
[16:49:48] <Reedy>	 But that's the reason for adding the slaves
[16:50:03] <Reedy>	 well, slave, singular
[16:50:30] <chrismcmahon>	 Reedy: but having slave db and running update.php are incompatible a working db it seems
[16:50:48] <chrismcmahon>	 incompatible with having a working db all the time
[16:51:08] <Reedy>	 those schema changes before were an issue as it was trying to run 30 or whatever sets simultaneously
[16:51:26] <Reedy>	 Most of the time there aren't any schema changes made
[16:51:31] <Reedy>	 I'm not sure why it's getting so lagged out apparently so frequently
[16:54:13] <chrismcmahon>	 Reedy: several times per day usually. it is the source of many flaky test failures
[16:54:21] <Reedy>	 :/
[16:54:36] <Reedy>	 I can't believe we're generating that much db traffic
[16:54:44] <Reedy>	 and update.php should mostly be a noop
[16:55:05] <Reedy>	 if it's becoming an issue, we can comment it out of the MW config
[16:55:19] <Reedy>	 it'll still replicate, but mw won't use it
[16:56:48] <chrismcmahon>	 Reedy: I'm not entirely sure I understood that last, but go for it
[16:57:10] <Reedy>	 heh
[16:57:17] <Reedy>	 So, mysql replication is a seperate thing entirely to mediawiki
[16:57:24] <chrismcmahon>	 sure
[16:58:01] <Reedy>	 if we comment out -db2 from the labs db config, mediawiki will stop using it (so then wikis won't go readonly due to replag), but we'll still have the "working" slave
[16:58:40] <Reedy>	 I wonder what the threshold value is for a wiki to become readonly
[16:59:12] <chrismcmahon>	 wfm
[16:59:55] <chrismcmahon>	 Reedy: yeah, the readonly is the actual problem
[17:00:20] <Reedy>	 "If all slaves are lagged by more than 30 seconds, MediaWiki will stop writing to the database"
[17:00:38] <Reedy>	 I can't remember if this is adjustable
[17:00:44] <Reedy>	 if it is, I'd like to try increasing that first
[19:08:30] <mutante>	 how to resolve "Enable irc feed for wikitech.wikimedia.org site" ?:)
[19:10:25] <Reedy>	 mutante: UDP packets from the host need to make it to argon
[19:11:10] <Reedy>	 Then something like
[19:11:10] <Reedy>	 $wgRCFeeds['default'] = array(
[19:11:10] <Reedy>	 	'formatter' => 'IRCColourfulRCFeedFormatter',
[19:11:10] <Reedy>	 	'uri' => "udp://$208.80.154.160:9390/#wikitec",
[19:11:11] <Reedy>	 	'add_interwiki_prefix' => false,
[19:11:11] <Reedy>	 	'omit_bots' => false,
[19:11:11] <Reedy>	 );
[19:11:56] <Krenair>	 without the $ in the IP
[19:12:10] <Reedy>	 and #wikitech spelt properly
[19:13:20] <mutante>	 Reedy: thanks! 
[19:15:19] <Krenair>	 should probably be #wikitech.wikimedia
[19:15:43] <Reedy>	 heh
[19:23:55] <mutante>	 Reedy: confirmed the network part works, i can netcat -u from virt1000 (wikitech) to argon
[19:24:04] <mutante>	 9391 or whatnot
[19:24:18] <mutante>	 9390 
[19:26:54] <mutante>	 Reedy: that was using the new RCStream module, right
[19:27:24] <Reedy>	 "new"
[19:27:31] <mutante>	  15          'uri'       => "redis://rcstream.eqiad.wmnet:6379/rc.$wgDBname",
[19:27:34] <mutante>	 that
[19:27:42] <Reedy>	 Uh..
[19:27:43] <Reedy>	 That's not IRC
[19:27:43] <mutante>	 puppet/modules/rcstream
[19:28:01] <mutante>	 but but "contains a simple and self-contained software stack for streaming 5 recent changes from a MediaWiki instance.
[19:29:01] <Nemo_bis>	 "simple and self-contained software stack" smells like it's not our good old hacky IRC feed
[19:30:12] <mutante>	 ok, trying a change in mw-config as you originally said, then -> gerrit
[19:32:57] <mutante>	 2709     // RCStream / stream.wikimedia.org
[19:33:36] <Bsadowski1>	 "An error occurred during a connection to stream.wikimedia.org. SSL received a record that exceeded the maximum permissible length. (Error code: ssl_error_rx_record_too_long) "
[19:34:05] <Bsadowski1>	 No SSL suppport?
[19:34:07] <mutante>	 404 on http
[19:34:14] <mutante>	 not sure yet what we'd expect there
[19:34:16] <Krenair>	 Bsadowski1, did you try to connect via a browser or something?
[19:34:22] <Bsadowski1>	 lol yeah
[19:34:28] <Krenair>	 ...
[19:35:03] <mutante>	 we might still want to disable that the httpd answers
[19:35:19] <Krenair>	 we could redirect it to the docs :p
[19:35:23] <Krenair>	 https://wikitech.wikimedia.org/wiki/RCStream
[19:35:29] <mutante>	 good one
[19:38:16] <Bsadowski1>	 I don't like this change. What will happen to snitchbot?
[19:38:43] <Bsadowski1>	 I use https://github.com/mzmcbride/irc-bots/blob/master/snitch.py (well a modified version)
[19:39:03] <Bsadowski1>	 It gets it from irc.wikimedia.org
[19:39:33] <mutante>	 i made a bug for that redirect, you have mail
[19:40:35] <Krenair>	 Bsadowski1, what, you don't like the idea of moving away from our IRC+human-formatted system?
[19:41:53] <Bsadowski1>	 I do like the idea, but how will the bots we currently use get the feed?
[19:42:37] <Krenair>	 They'll be modified to use the new one
[19:42:41] <Krenair>	 By their developer(s)
[21:45:33] <mutante>	 what is this for?
[21:45:36] <mutante>	 http://en.wikipedia.org/wiki/Special:OAIRepository
[21:45:39] <mutante>	 it's search related
[21:45:42] <mutante>	 and asks for a login
[21:46:07] <hoo>	 brion: ^
[21:46:35] <Reedy>	 mutante: recent changes, essentially
[21:46:38] <brion>	 mutante: https://www.mediawiki.org/wiki/Extension:OAIRepository
[21:46:56] <mutante>	 thanks!
[21:46:58] <brion>	 i believe ‘testy’ ‘mctest’ works as a username/pass ;)
[21:47:27] <hoo>	 brion: nope :P
[21:47:48] <brion>	 darn someone must have removed that :D
[21:48:12] <hoo>	 what access data is that? Wiki one? Or some secret stuff that I can find on tin/ fenari?
[21:48:21] <hoo>	 * could (don't care thta much)
[21:48:28] <Reedy>	 username testing?
[21:48:43] <mutante>	 docs: "it will fail" :)
[21:48:53] <brion>	 Reedy: yes that’s it :D
[21:48:54] <Reedy>	 brion: testing/mctest it seems ;)
[21:49:22] <brion>	 but yeah it’s not super exciting, it’s just an export feed wrapped in OAI-PMH xml requests
[21:49:27] <Reedy>	 :)
[21:49:30] <hoo>	 Reedy: Works... but doesn't seem useful to have that restricted
[21:49:30] <brion>	 i think we want to deprecate it though
[21:49:37] <brion>	 in favor of more modern rcstream whatever
[21:49:53] <Reedy>	 I should probably re-enable auditing and see who is still using it
[21:50:10] <brion>	 ^demon|away: does cirrus search still use OAI for updates or is it on something else now?
[21:50:21] <brion>	 iw ould love to kill it and close any outstanding bugs ;)
[21:50:22] <Reedy>	 Why is it behind http auth?
[21:50:31] <^demon|away>	 We hook into MW all over the place like an octopus.
[21:50:33] <^demon|away>	 No OAI
[21:50:45] <brion>	 Reedy: originally we used it to offer real-time data feeds for a fee to some mirrors
[21:50:47] <^demon|away>	 Main source of updates is LinksUpdateCompleted but there are others.
[21:50:50] <brion>	 now we’ve phased them out
[21:50:56] <Reedy>	 ah
[21:50:57] <Reedy>	 MONIES
[21:50:58] <brion>	 we just don’t want anyone using OAI cause it’s old and unmaintained
[21:51:16] <hoo>	 undeploy, then?!
[21:51:21] <brion>	 so once old search is off, i think it may be safe to undeploy yeah
[21:51:24] <brion>	 but not just yet i suspect
[21:51:24] <Reedy>	 people are/were using it
[21:51:30] <Reedy>	 i'm going to enable auditing
[21:51:34] <Reedy>	 then we can start poking people
[21:51:36] <^demon|away>	 brion: So once we've decom'd lsearchd we can start killing OAI imho.
[21:51:39] <^demon|away>	 Yep.
[21:51:40] <brion>	 Reedy: the auditing had poor performance iirc…
[21:51:41] <^demon|away>	 Agreed.
[21:51:44] <Reedy>	 I think dbpedia (among others) is still using it
[21:51:49] <Reedy>	 brion: It was enabled for ages
[21:51:50] <mutante>	 https://gerrit.wikimedia.org/r/#/c/148992/1/templates/lucene/lsearch.conf.erb
[21:51:57] <Reedy>	 I disabled it not too long ago
[21:52:03] <brion>	 ah
[21:52:31] <Reedy>	 Sometime around the start of the year I think :)
[21:53:01] <brion>	 :D
[21:53:18] <^demon|away>	 brion: So yeah, I think we could be ready to stop using it internally inside the next quarter (and a half, tops)
[21:53:26] <brion>	 \o/
[21:53:29] <^demon|away>	 So if you wanna start trying to get our third parties off it, be my guest.
[21:53:32] <^demon|away>	 :)
[21:53:36] <Reedy>	 Actually
[21:53:40] <Reedy>	 Didn't I break it a few months ago
[21:53:43] <Reedy>	 And we had people complaining? :P
[21:54:00] <odder>	 best way to see who's using it?
[21:54:04] <brion>	 heh
[21:54:13] <Reedy>	 there's already 14 rows!
[21:54:17] <brion>	 definitely doublecheck if dbpedia’s still using it
[21:54:17] <mutante>	 it's always amazing how one thing leads to another :)
[21:54:24] <mutante>	 just mentioned it because of some lint change, hehe
[21:55:06] <^demon|away>	 brion: For dbpedia and any other cool people consuming it we don't just want to pull the rug out from under them though.
[21:55:06] <Reedy>	 Yup
[21:55:13] <Reedy>	 dbpedia is DEFINITELY still using it
[21:55:21] <Reedy>	 44 audit rows so far
[21:55:21] <Reedy>	 all them
[21:55:26] <^demon|away>	 So I'd hope we'd be able to service them from a combination of rcstream/pubsub/api stuffs.
[21:55:32] <brion>	 ok let’s plan to plan to migrate them to something more modern
[21:55:59] <Reedy>	 Do we know who to contact?
[21:56:03] <brion>	 rcstream is just metadata right? they might need the text as well
[21:56:08] <brion>	 no idea
[21:56:12] <^demon|away>	 admin@dbpedia.org?
[21:56:14] <^demon|away>	 :p
[21:56:31] <brion>	 root@, postmaster@, attention@ ;)
[21:56:59] <odder>	 help@
[21:57:06] <Reedy>	 stopusingfuckingoai@
[21:57:08] <^demon|away>	 See if we can get an apache error page with ContactAdmin set :p
[21:57:12] <mutante>	 http://wiki.dbpedia.org/Support
[21:57:28] <^demon|away>	 Oh they have a bug tracker.
[21:57:35] <^demon|away>	 File "Stop using oai"
[21:57:41] <Reedy>	 http://sourceforge.net/p/dbpedia/tracker/
[21:57:50] <Reedy>	 No activity for well over a year
[21:58:19] <Reedy>	 To Facebook! https://www.facebook.com/groups/4340232249/
[21:58:32] <^demon|away>	 "Hey friends i want some data of the banking for my big data experiments please provide me source ..."
[21:59:28] <mutante>	 try him http://dws.informatik.uni-mannheim.de/en/people/professors/prof-dr-christian-bizer/
[22:02:39] <Reedy>	 Ah
[22:02:45] <Reedy>	 I've got some old emails about this...
[22:03:25] <Reedy>	 Diederik emailed them.. Reply from Sebastian Hellmann 
[22:04:00] <Reedy>	 Of course, we are willing to change to the MediaWiki API, if necessary (and we also have to man power to achieve this within several months).
[22:04:00] <Reedy>	 There were two major reasons, why we didn't switch, yet:
[22:04:00] <Reedy>	 1. we have a running system, there is no real incentive to switch unless you tell us to.
[22:04:00] <Reedy>	 2. we didn't have a contact from Wikimedia. I wrote one or two emails in the past, but didn't get a response.
[22:04:00] <Reedy>	 3. We did not find any good documentation on how to get *all* updates from Wikipedia. Query RC and then do Special:Export requests?
[22:04:01] <Reedy>	 4. We were afraid to get blocked, since we would be over the 1 request per second limit.
[22:04:24] <Reedy>	 I forwarded this to Engineering back in Feb 2013
[22:04:25] <Reedy>	 "[Engineering] FW: Fwd: Use of OAI by Dbpedia"
[22:08:06] <brion>	 guess we should pull that one out again sometime
[22:08:32] <brion>	 and make sure we have a newer soluiton for them
[22:09:09] <Reedy>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=68538
[22:09:25] <Reedy>	 Just opened that.. Seems Sebastian has a bz account on a gmail email
[22:11:23] * Reedy emaisl him too
[22:11:46] <brion>	 thanks Reedy :D
[22:15:18] <Reedy>	 Hopefully gets the ball rolling at least
[23:14:26] <Reedy>	 Seems we have some OAI traffic from fresheye.com
[23:15:17] <mutante>	 wants "https://wikitech.wikimedia.org/wiki/Projects#Gerrit_repo_creation_through_wikitech