[04:48:32] <tgr>	 are search suggestions handled by CirrusSearch?
[05:50:28] <quiddity>	 tgr, https://www.mediawiki.org/wiki/Search#Timeline  (iiuc, that short list is the remaining wikis, where it's still a Beta Feature)
[05:51:19] <tgr>	 quiddity: what I meant is, does Cirrus handle search suggestions specifically?
[05:51:36] <tgr>	 as there used to be separate extensions for suggestions and the actual search
[05:51:41] <quiddity>	 oic, sorry. idk.
[05:51:43] <tgr>	 but the answer seems to be yes
[06:53:19] <DanielK_WMDE>	 hm, anyone around how knows what the first column in the *-multistream-index.txt files mean?
[06:53:47] <DanielK_WMDE>	 it doesnt seem to be an offset in the corresponding xml file, not a page id... 
[15:06:13] <Nemo_bis>	 Reedy: http://noc.wikimedia.org/conf/commonsuploads.dblist lies
[15:06:48] <Nemo_bis>	 It doesn't contain the 220 or so new lines, though the config was applied on the wikis...
[15:10:13] <hoo>	 Nemo_bis: You're right about this... looking
[15:11:49] <Reedy>	 Fenari is broken
[15:11:53] <Reedy>	 sync-common doesn't work properly on it
[15:11:57] <Reedy>	 So it's not completely up to date
[15:12:05] <hoo>	 yeah
[15:12:15] <hoo>	 that's because nobody ever deleted /a/common on it -.-
[15:12:49] <Reedy>	 It was working fine
[15:12:58] <Reedy>	 Just as of thursday it was being a PITA
[15:13:03] <Reedy>	 presumably due to some other change
[17:41:07] <Betacommand>	 Reedy: do you know where the irc feed for mediawiki.org is on irc.wikimedia.org?
[17:41:32] <Krenair>	 #mediawiki.wikipedia
[17:41:41] <Betacommand>	 thanks
[21:11:17] <legoktm>	 Nemo_bis: https://www.mediawiki.org/wiki/Admin_tools_development/SUL_Audit/Metrics_20140704
[21:11:32] <legoktm>	 I should probably add in commas
[21:11:34] <Nemo_bis>	 uuuuuuuh
[21:11:43] <Nemo_bis>	 or formatnum
[21:12:41] <legoktm>	 hmm, my total is "5,829,957"
[21:12:45] <legoktm>	 that's not right.
[21:13:01] <dtm>	 Nemo_bis: hellos.  did you see my reply to your discussion about wikilinking
[21:13:29] <Nemo_bis>	 dtm: I saw quiddity is handling it
[21:13:38] <Nemo_bis>	 legoktm: yes, numbers look irrealistic
[21:13:46] <dtm>	 oic
[21:14:15] <Nemo_bis>	 legoktm: the queries are done per-wiki aren't they
[21:14:23] <legoktm>	 yes
[21:14:34] <legoktm>	 I'm iterating through each wiki
[21:14:56] <Nemo_bis>	 heh, I wish we really had only 261072 unattached accounts (down from 4 millions)
[21:16:50] <Nemo_bis>	 legoktm: why do you have user_id twice in the selecty
[21:17:01] <Nemo_bis>	 https://gist.github.com/legoktm/98a4ea1b217ce9463f7d#file-thingy-py-L151
[21:17:19] <legoktm>	 oh, weird.
[21:17:30] <legoktm>	 probably when we were debugging
[21:17:51] <legoktm>	 Also, I put a more documented copy in gerrit https://gerrit.wikimedia.org/r/144181
[21:19:52] <Nemo_bis>	 is https://gist.github.com/legoktm/98a4ea1b217ce9463f7d#file-thingy-py-L120 safe enough? will aa_lastlogin be NULL or what when undefined
[21:19:55] <Nemo_bis>	 ah
[21:20:54] <legoktm>	 aa_lastlogin would be python's None when undefined
[21:21:00] <legoktm>	 which the script handles
[21:25:21] <legoktm>	 ohhhhh
[21:25:23] <legoktm>	 I found it.
[21:26:14] <Nemo_bis>	 last time I did this, I made a mistake with normalisation of usernames
[21:26:52] <legoktm>	 I was specifying the specific month for everyone, like, "last active 15 months ago", but only counting those in the columns that were being outputted
[21:27:31] <Nemo_bis>	 Ah, instead of summing. Right :)
[21:28:01] <Nemo_bis>	 I was still checking all the rows would be fetched in the first place
[21:28:32] <dtm>	 James_F|Away: i'm glad to see that the full citation metadata is supported in visualeditor's 'cite' menu.  there's the whole list of 'cite web' options and such.  
[21:28:58] <dtm>	 James_F|Away: so you're going to make citoid fill those in, from just a url?  is there anywhere i can use citoid now?
[21:29:14] <legoktm>	 I don't think citoid has been written yet.
[21:29:50] <Nemo_bis>	 dtm: fyi there's more https://wikimedia.mingle.thoughtworks.com/projects/language_engineering/cards/4235
[21:31:30] <dtm>	 legoktm: that could be a damper
[21:32:03] <dtm>	 legoktm: but it looks like it has https://www.mediawiki.org/wiki/Citoid#VisualEditor_and_Citation_Tool
[21:32:52] <legoktm>	 looks like there's some code https://github.com/wikimedia/mediawiki-services-citoid
[21:33:09] <dtm>	 yeah i dont want to have to install it though :/
[21:33:27] <dtm>	 i didn't know if it actually works yet
[21:57:01] <legoktm>	 Nemo_bis: ok, fixed it and confirmed with a test on envoy, started running again...should be done in an hour or two
[21:57:56] <Nemo_bis>	 legoktm: how comes it's so much faster now
[21:58:09] <legoktm>	 because Roan helped me optimize the query :P
[21:59:23] <legoktm>	 I put a time in front this time so we'll know exactly how fast
[22:06:58] <dtm>	 legoktm: be sure to account for the overhead of 'time'
[22:07:12] <dtm>	 at some point in time.
[22:09:59] <Reedy>	 I suspect that level of accuracy probably isn't needed
[22:10:19] <dtm>	 noted.
[23:02:31] <legoktm>	 real    61m34.944s
[23:02:50] <hoo>	 legoktm: Sul audit?
[23:02:52] <legoktm>	 ok, these numbers look more reasonable :)
[23:02:53] <legoktm>	 yup!
[23:05:33] <legoktm>	 https://www.mediawiki.org/wiki/Admin_tools_development/SUL_Audit/Metrics_20140704
[23:05:44] <legoktm>	 Nemo_bis: ^
[23:05:57] <legoktm>	 only ~100k new unattached accounts
[23:07:43] <legoktm>	 hi csteipp 
[23:08:00] <csteipp>	 hey legoktm 
[23:08:12] <legoktm>	 https://www.mediawiki.org/wiki/Admin_tools_development/SUL_Audit/Metrics_20140704 :)
[23:08:23] <hoo>	 hi csteipp :)
[23:08:37] <hoo>	 Not sure you noticed, but we landed global rename
[23:08:50] <hoo>	 and people yet tested (without any problems)
[23:09:03] <csteipp>	 legoktm: Awesome!
[23:09:30] <csteipp>	 hoo: Definitely saw that... and saw it didn't get reverted after the bug in beta
[23:09:47] <legoktm>	 bug in beta?
[23:10:24] <csteipp>	 Oh, Chris Mcmahn reported that the db schema was missing... I think he hit it before the update.php cron ran
[23:10:37] <hoo>	 oh, that's not a thing
[23:10:38] <csteipp>	 So no problems now that I know of
[23:11:15] <hoo>	 I had a steward test various cases on beta and it went smooth (only some enhancement suggestion, like the confirm thing)
[23:11:24] <legoktm>	 CA doesn't support update.php, I created the tables manually
[23:12:35] <Krenair>	 Yeah... Would be nice if we could have MW's update system deal with external DBs
[23:12:51] <Krenair>	 I think globalblocking doesn't either...?
[23:12:52] <hoo>	 legoktm: csteipp: Already done for production, btw
[23:13:00] <csteipp>	 Ah, I forgot we still had that problem. I thought we fixed it..
[23:13:35] <legoktm>	 Krenair: no, it doesn't, but a hook was added to do it on jenkins
[23:13:42] <legoktm>	 and use the local db
[23:33:29] <Nemo_bis>	 legoktm: "who do not clash with another account" should be 0 now, right? They should all have got a global account
[23:33:46] <Nemo_bis>	 Well, except local accounts moved out of a global account
[23:34:12] <legoktm>	 no, there are about 10k of those left
[23:34:30] <Nemo_bis>	 Those may be incomplete account creations?
[23:34:45] <legoktm>	 possibly, I'm not sure.
[23:34:54] <legoktm>	 I have a list in my home on terbium, I was going to run the script to migrate them sometime next week
[23:35:17] <Nemo_bis>	 Most interesting is "who clash with a global account"; "who clash with 1 or more local accounts" in theory should be dominated by inactive accounts as we've been creating global accounts on login for a year
[23:35:30] <Nemo_bis>	 ok
[23:36:11] <Nemo_bis>	 How are you going to merge the autocreated detached accounts?
[23:38:47] <legoktm>	 I don't know what to do about those yet
[23:42:48] <Nemo_bis>	 It would also be ok to just delete them/move them out
[23:43:33] <Nemo_bis>	 (Or at least I wouldn't care too much for mine; even though I would lose the lower timestamp and user_id, they're just unusable and unused accounts after all)
[23:44:01] <legoktm>	 on what wiki is your unattached account?
[23:44:41] <Nemo_bis>	 legoktm: I don't remember, I have two
[23:45:03] <legoktm>	 http://tools.wmflabs.org/sulinfo/ is down :/
[23:45:16] <hoo>	 If these have 0 edits, no password and no email, I guess we can automerge them with a global account of the same name
[23:45:16] <Nemo_bis>	 https://meta.wikimedia.org/wiki/Special:CentralAuth/Nemo_bis
[23:45:57] <Nemo_bis>	 fr.voy, gu.wikt
[23:46:00] <hoo>	 not going to do that per hand now, though...
[23:46:09] <hoo>	 a lot of people would then also ping me, so meh :P
[23:46:22] <Nemo_bis>	 However some of those have either password or email
[23:46:35] <hoo>	 the unattached?
[23:46:48] <legoktm>	 if they have either password or email, the user should be able to login or get a password reset
[23:46:50] <hoo>	 the ones I saw from that long lasting bug didn't AFAIR
[23:47:02] <hoo>	 yeah, what legoktm said
[23:47:08] <Nemo_bis>	 Nothing at all? Didn't some have email but unverified
[23:47:22] <legoktm>	 both your accounts have no email nor password
[23:47:27] <Nemo_bis>	 ok
[23:47:43] <legoktm>	 if it has an unverified email, they should still be able to get a password reset
[23:47:55] <hoo>	 yeah, taht works
[23:48:30] <Nemo_bis>	 So you wouldn't even need to check the autocreation log?
[23:49:01] <legoktm>	 probably not, it's impossible to create an unattached account with no password
[23:49:08] <Nemo_bis>	 In theory it could be an ancient account with no edits attached and a blank password of the times where min length was 0, renamed to a global username by a rogue bureaucrat? 
[23:49:20] <Nemo_bis>	 Not worth checking :P
[23:49:24] <hoo>	 Yep