[00:00:45] <TimStarling>	 it was too clever
[00:01:57] <AaronSchulz>	 TimStarling: maybe a ternary with an explicit true is fine
[00:02:10] <AaronSchulz>	 I whole if/else if kind of klunky
[00:02:23] <TimStarling>	 you know, humans have to read this as well as computers
[00:02:59] <AaronSchulz>	 $x = ( $a == $b ) ? true : z();
[00:02:59] <TimStarling>	 if ( $source === $dest ) {
[00:03:03] <TimStarling>	 $ok = true;
[00:03:04] <TimStarling>	 } else {
[00:03:11] <TimStarling>	 $ok = copy( $source, $dest );
[00:03:18] <TimStarling>	 }
[00:03:29] <TimStarling>	 crystal clear, doesn't take long to type
[00:04:09] * AaronSchulz  gathers Tim wouldn't like functional programming

[00:04:33] <AaronSchulz>	 I think an explicit true is plenty readable
[00:04:42] <AaronSchulz>	 unless one doesn't like ternaries in general
[00:05:20] <TimStarling>	 I wrote in the style guide that I will begrudgingly accept their existence as long as they fit on one line
[00:06:00] <TimStarling>	 so I suppose $ok = ( $source === $dest ) ? true : copy( $source, $dest ); has to be accepted under that rule
[00:06:43] <AaronSchulz>	 I was running 'git review -f' as you said that ;)
[00:06:56] <TimStarling>	 my problem with them is that they don't read like english, they read like a mathematical formula
[00:07:38] <AaronSchulz>	 gwicke: what do you think? :)
[00:08:46] <gwicke>	 AaronSchulz: hehe- don't try to drag me into this ;)
[00:09:20] <gwicke>	 readable is always good IMO- the only trouble is that it means different things to different people
[00:09:29] <TimStarling>	 functional programming really is designed for people who love maths more than english
[00:10:02] <AaronSchulz>	 gwicke: or pseudo-math :)
[00:10:13] <TimStarling>	 you know how recursion is favoured over iteration in mathematical proofs
[00:10:21] <gwicke>	 both functional and imperative programming are acquired tastes
[00:11:16] <gwicke>	 AaronSchulz: so you read that blog post ;)
[00:12:34] <TimStarling>	 but I think there are more people in the world who know english than people who are familiar with proof by induction
[00:12:42] <gerrit-wm>	 New patchset: Adamw; "CentralNotice supports SQLite now" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/55830
[00:12:55] <TimStarling>	 so I'm biased
[00:30:58] <AaronSchulz>	 TimStarling: I'm sure you had lots of fun with math doing physics work
[00:32:26] <gwicke>	 TimStarling: the 'similarity with English' thing did not work out so well for COBOL
[00:32:37] <TimStarling>	 AaronSchulz: sure, but it's not for everyone
[00:34:10] <TimStarling>	 I like to think of programming as being a kind of poetry
[00:34:35] <TimStarling>	 or a kind of restrained writing, like a lipogram
[00:35:03] <TimStarling>	 you're writing something which explains what the program does to a human
[00:35:38] <TimStarling>	 and simultaneously, the same text also actually does that thing when run on a computer
[00:36:45] <TimStarling>	 and I think the human audience is primary, and the computer audience is secondary
[00:37:55] <TimStarling>	 making a computer do a thing is almost trivial, compared to explaining to a human how you did it
[00:38:45] <AaronSchulz>	 gwicke: ah, COBOL :)
[00:39:36] <AaronSchulz>	 TimStarling: I'm more bothered by poorly named functions and no functions docs than a ternary here and there
[00:39:39] * AaronSchulz  shrugs

[00:39:54] <gwicke>	 TimStarling: I agree that humans are the primary audience, but that does not mean that we can't use some abstract concepts
[00:42:01] <gwicke>	 just have to choose the right set of concepts
[00:53:44] <xyzram>	 I like the ternary/conditional expressions because the factor out the common part making it more readable.
[00:53:53] <xyzram>	 the --> they
[00:54:28] <xyzram>	 especially when the variable being assigned to is a long name perhaps with subscripts etc.
[08:48:56] <gerrit-wm>	 New review: Hashar; "Congratulations :-]" [integration/zuul-config] (master); V: 2 C: 2;  - https://gerrit.wikimedia.org/r/55830
[08:49:07] <gerrit-wm>	 Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/55830
[09:54:03] <hashar>	 zeljkof: gooood morning :]
[09:54:19] <zeljkof>	 hashar: morning :)
[09:55:33] <hashar>	 zeljkof: yesterday with rob we were wondering whether selenium is running against test2 or beta :)
[09:55:45] <zeljkof>	 hashar: all around the place :)
[09:55:57] <zeljkof>	 both test2 and beta and translatewiki, maybe something else too
[09:56:05] <hashar>	 whaouuu
[09:56:14] <zeljkof>	 and it runs locally too
[09:56:38] <zeljkof>	 and with all possible browser/os combinations, some work better than others
[09:56:43] <hashar>	 does it mail you whenever something dies ?
[09:56:48] <zeljkof>	 yes
[09:57:01] <hashar>	 \O/
[09:57:17] <zeljkof>	 the reason it runs all over the place is that we did not set up one machine that has all the extensions that we need
[09:57:27] <hashar>	 for beta, we thought it might be a good idea to train you to fix the beta cluster too.  I am not sure whether you have enough time for that.
[09:57:28] <zeljkof>	 and when it fails it also complains here
[09:57:44] <zeljkof>	 hashar: I am interested
[09:57:55] <zeljkof>	 but I do have a lot of things going on
[09:58:12] <zeljkof>	 it could be a good thing for level up next quarter, right?
[09:58:13] <hashar>	 I can imagine :-]
[09:58:32] <zeljkof>	 I think it is a good idea, actually
[09:58:43] <hashar>	 I guess next time something is broken, ping me and I can share my screen to give you a quick overview
[09:58:57] <hashar>	 I am rewriting the beta documentation this afternoon
[09:59:11] <zeljkof>	 great
[09:59:36] <hashar>	 also need to make my beta jobs to report to -labs :-]
[09:59:41] <zeljkof>	 the biggest thing missing for selenium tests to be really effective is good and stable virtual machines
[10:00:26] <hashar>	 I would love to make beta more stable, but I usually don't know when it is broken
[10:01:41] <zeljkof>	 oh, we always know when something is broken :)
[10:01:48] <zeljkof>	 our test scream and shout
[10:02:02] <zeljkof>	 the next time I see something broken, I will let you know
[10:02:20] <hashar>	 maybe fill in a bug under Wikimedia Labs -> deployment-prep (beta)
[10:02:28] <zeljkof>	 will do
[10:02:30] <hashar>	 that will give us a nice log of actions taken to fix each issue
[10:14:58] <gerrit-wm>	 Change abandoned: Hashar; "no point in keeping this around for now." [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/44384
[12:03:22] <gerrit-wm>	 New patchset: Hashar; "move mw core code coverage" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/55876
[12:04:06] <gerrit-wm>	 Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/55876
[12:50:07] <gerrit-wm>	 New patchset: Zfilipin; "Switch to new RVM setup files" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/55878
[13:44:00] <hashar>	 zeljkof: do you use any specific beta wiki beside enwiki ?
[13:44:18] <zeljkof>	 hashar: I do not think so, for now
[13:44:42] <hashar>	 zeljkof: I am going to remove a few of them.  500 is way too more for such a small cluster :-]
[13:45:03] <zeljkof>	 hashar: go ahead as far as I am concerned :) leave the english one
[13:48:31] <hashar>	 I will leave the de en and wikivoyages wikis
[13:48:34] <hashar>	 maybe a few others
[13:57:09] <gerrit-wm>	 New patchset: Zfilipin; "Enable AFTv5 tests since bug 46382 is fixed" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/55885
[14:07:08] <anomie>	 hashar- Are you ready to deploy the wikiversions-labs.cdb change (gerrit change 47564)?
[14:11:22] <hashar>	 !g 47564
[14:11:22] <wm-bot>	 https://gerrit.wikimedia.org/r/#q,47564,n,z
[14:11:27] <hashar>	 anomie: helllo :-]
[14:12:32] <anomie>	 hashar- It looks good to me, if you're ready to deploy it I'll +2 it now. Otherwise I'll +1 it and let someone else +2 later.
[14:13:36] <hashar>	 anomie: lets do it so
[14:14:14] <anomie>	 hashar- Does Jenkins not merge that repo, or is it just that everyone doesn't bother to let Jenkins do it?
[14:14:45] <hashar>	 jenkins should be merging it on a CR+2
[14:15:24] <anomie>	 Just did. I had looked through recent merges, and everyone seemed to do it manually instead.
[14:15:41] <hashar>	 while you are looking at mediawiki-config you might want to look at https://gerrit.wikimedia.org/r/55889 as well
[14:15:44] <hashar>	 it clean up the dblist files for beta to get less wiki there
[14:18:56] <anomie>	 hashar- Do we care about https://gerrit.wikimedia.org/r/#/c/33388/ anymore? If so, I'll rebase it, otherwise I'll just abandon it.
[14:19:39] <hashar>	 anomie: I have no idea
[14:19:44] <hashar>	 it is no more on beta for sure
[14:20:02] <hashar>	 maybe it can be useful, I have no idea
[14:20:04] <qgil>	 zeljkof, GSOC. Do you understand my point, and will you have time to discuss today after the ECT call?
[14:20:09] <zeljkof>	 qgil: do you have time to chat now? both irc and hangout are good for me
[14:20:15] <qgil>	 heh
[14:20:25] <qgil>	 ok, now
[14:20:37] <qgil>	 hangout is faster
[14:20:38] <hashar>	 anomie: trying out the dumpInterwiki.php
[14:21:22] <qgil>	 btw zeljkof I enjoyed your stats diagrams, especially the one showing the andre__ hashar mortal kombat.  ;)
[14:21:33] <hashar>	 anomie: I would go and abandon it :-]
[14:21:42] <anomie>	 hashar- works for me
[14:21:43] <hashar>	 qgil: which stats ? :-]
[14:22:04] <hashar>	 anomie: jenkins merge mediawiki-config changes on beta automatically https://integration.wikimedia.org/ci/job/beta-mediawiki-config-update/ ;-D
[14:22:29] <qgil>	 hashar, http://lists.wikimedia.org/pipermail/wikitech-l/2013-March/067914.html
[14:24:24] <andre__>	 but hashar fixes bugs at least. I just close them. :P
[14:26:28] <hashar>	 oh my god
[14:26:34] <hashar>	 I need to close more :-]
[14:27:17] <^demon>	 zooooookeeper :)
[14:29:02] <hashar>	 anomie: testing out in beta
[14:29:48] <hashar>	 anomie: if you have testing use case,   you might want to test thelm on beta too :)
[14:32:21] <hashar>	 deploying on production
[14:36:01] <hashar>	 hmm unrelated but:
[14:36:11] <hashar>	 2013-03-26 06:51:02 mw1032 commonswiki: [53b9d00a] /w/index.php?title=MediaWiki:Filepage.css&action=raw&maxage=2678400&usemsgcache=yes&ctype=text%2Fcss&smaxage=2678400   Exception from line 637 of /usr/local/apache/common-local/php-1.21wmf12/includes/cache/MessageCache.php: Message key 'Filepage.css' does not appear to be a full key.
[14:38:43] <hashar>	 MessageCache->get('Filepage.css', true, true, true);
[14:39:36] <qgil>	 andre__, true! Your practice could be considered the equivalent of doping in a Bugzilla context
[14:40:37] <andre__>	 Well, most of the time when I'm triaging now, I'm clean. I've become older, you know. :)
[14:41:20] <andre__>	 hashar: I remember a few days ago you had some thoughts on splitting the "Testing Infrastructure" component in Bugzilla, or creating a new component for something related, or so? (I can't find the log, hence my vague question.) Which current problem do you see that I can start thinking about and draft an email? :)
[14:44:09] <chrismcmahon>	 andre__: wild guess on the "test infrastructure" split, it might be useful to distinguish between beta cluster support for outward-facing tests and Jenkins/gallium for unit-level "technology-facing" tests.
[14:44:38] <andre__>	 chrismcmahon: ah. Who would be parties to discuss this by email?
[14:44:43] <andre__>	 I'm hapy to start that.
[14:45:25] <chrismcmahon>	 andre__: Antoine, me, Zeljko, Michelle, Timo to start I would think.
[14:45:52] <andre__>	 thanks, alright. will do so in the next days.
[14:46:10] <chrismcmahon>	 andre__: hashar might have some more refined ideas than that.
[14:46:18] <andre__>	 yepp :)
[14:49:45] <hashar>	 !g Iaaf6ccebd8c40c9602748c58c3a5c73c29e7aa4d
[14:49:45] <wm-bot>	 https://gerrit.wikimedia.org/r/#q,Iaaf6ccebd8c40c9602748c58c3a5c73c29e7aa4d,n,z
[14:55:35] <hashar>	 anomie: thank you for the review/merges. Seems to work fine on beta and production :-]
[14:55:46] <anomie>	 hashar- Good!
[14:57:05] <hashar>	 andre__: ah yeah splitting testing infrastructure.  So that is currently used to fill bugs for several softwares   Jenkins (and its jobs),  Zuul, browser tests and whatever misc stuff we are in charge of.
[15:03:38] <^demon>	 Whee. http://8983.solr0.instance-proxy.wmflabs.org/solr/#/~cloud
[15:04:14] <hashar>	 oh
[15:04:25] <^demon>	 (Can swap solr0 for solr[0-3], they're clustered via zookeeper)
[15:04:47] <hashar>	 I don't even know what zookeeper is , but that sounds good
[15:05:38] <^demon>	 That's so awesome. You just add another node by spinning it up, pointing it at zookeeper, and it syncs the part of the core its supposed to shard.
[15:05:44] <^demon>	 Solr 4.x rocks.
[15:12:22] <rachel99>	 zeljkof: Hi! do you have any time to help me program the search.feature file?
[15:12:40] <zeljkof>	 rachel99: sure, but in 10-20 minutes, I am in a meeting now
[15:12:57] <zeljkof>	 rachel99: is that a good time for you?
[15:12:58] <rachel99>	 Ok.  Just get in touch when you are free.
[15:13:15] <rachel99>	 zeljkof: yes, that is fine.
[15:13:30] <zeljkof>	 rachel99: great, will let you know
[15:15:44] * hashar  tries its first Jenkins matrix job

[15:45:56] <zeljkof>	 rachel99: the meeting is over, I can help you now
[15:46:29] <guillom>	 robla: if you had to give a time estimate on that tagging functionality (regardless of whether if it's in gerrit itself or bugzilla or somewhere else), what would you say?
[15:47:10] <robla>	 ^demon: ^ see guillom's question
[15:48:13] <zeljkof>	 rachel99: let me know when you are around, I will check in 5-10 minutes, I am going to get something to eat
[15:50:10] <^demon>	 guillom: I'd hoped to be further along than this, but about 2 weeks?
[15:50:34] <guillom>	 ^demon: 2 weeks would be awesome :)
[15:50:45] <guillom>	 Thanks for the answer.
[15:50:48] <^demon>	 yw.
[15:51:03] <rachel99>	 zeljkof: Ok, I'm here.whenever you get back.
[15:51:05] <^demon>	 robla: I love solr. Can we just kill lsearchd like tomorrow? ;-)
[15:51:44] <robla>	 ^demon: we're all booked up tomorrow.  Thursday?
[15:52:20] <^demon>	 Heh. I wish.
[15:52:52] <^demon>	 All I know is, setting up a 4-node solr cluster was dead easy in 4.x+, with no spof (no master!)
[15:53:08] <^demon>	 If this thing scales, and we can come up with the MW glue, I'm sold.
[15:53:27] <robla>	 I saw your note about using zookeeper.  I don't recall if the Analytics folks are merely planning on using zookeeper or already have it up
[15:53:41] <qgil>	 guillom, robla ref havig statuses in homepage: PS: Quim doesn't mind to leave this for a potential Wikitech homepage based on statuses filled using Semantic Forms. No need to develop a hole system based on our legacy if we are going to change this in a few months.
[15:53:43] <robla>	 I think they're already using it.
[15:53:47] <^demon>	 zookeeper used to be a part of hadoop, now its own project.
[15:53:52] <^demon>	 So yeah, they might be using it.
[15:54:05] <qgil>	 guillom, robla meaning: don't start coding Lua templates right away unless you really want to do it.
[15:54:16] <guillom>	 qgil: got it; thanks
[15:54:44] <robla>	 qgil: got it
[15:55:08] <guillom>	 qgil: however I may steal Rob's Lua talents to make the events on the home page more automated :) I'll follow up by e-mail.
[15:56:33] <chrismcmahon>	 ^demon: fwiw, we're not the first good-size wiki to put solr in place:  https://www.socialtext.net/st-rest-docs/solr_schema_guide
[15:57:22] <^demon>	 We've already got the export schema too, which is probably a good starting point.
[16:00:00] <robla>	 the Analytics folks are planning to eventually replace udp2log with Kafka.  Kafka makes messages flow from producers->brokers->consumers, and they were planning to use zookeeper for broker coordination
[16:00:46] <robla>	 they're using kafka now (udp2log is sending to kafka), but I don't think they're using zookeeper yet
[16:01:53] <^demon>	 Gotcha.
[16:02:32] <^demon>	 I got it working using the standard ubuntu package :)
[16:02:44] <zeljkof>	 rachel99: ping me when you are around
[16:02:59] <^demon>	 Solr's another story. Even up through raring, it's only on 3.x. You need 4.x+ for the cloudy stuff.
[16:03:00] <rachel99>	 I am here
[16:09:07] <zeljkof>	 rachel99: I am back :)
[16:09:22] <rachel99>	 ok,
[16:09:46] <rachel99>	 I opened up private chat - do you see it?
[16:10:08] <zeljkof>	 rachel99: I see it, will reply there
[16:15:28] <hashar>	 holyyyywooood
[16:15:49] <hashar>	 upgrading the database on the beta cluster via jenkins … https://integration.wikimedia.org/ci/job/beta-update-databases/14/  :)D  <--- chrismcmahon
[16:16:17] <chrismcmahon>	 \o/ hashar nice
[16:22:46] <Krinkle>	 James_F: robla: greg-g: Hi there, I believe I've been summoned?
[16:22:57] <Krinkle>	 What's up?
[16:23:04] <robla>	 hiya
[16:23:16] <robla>	 bug 46401
[16:23:20] <robla>	 !b 46401
[16:23:20] <wm-bot>	 https://bugzilla.wikimedia.org/46401
[16:24:11] <robla>	 we're getting a lot of complaints on Village Pumps (commons, enwiki, possibly others)
[16:25:08] <robla>	 It *may* be isolated to specific Gadgets/scripts/extensions, but we're not sure
[16:25:19] <DanielK_WMDE>	 ^demon: hi! would you have time to run a script for wikidata? the rebuildTermSearchKey thing we discussed last week.
[16:25:37] <^demon>	 Oh which wikis?
[16:25:49] <DanielK_WMDE>	 ^demon: wikidata.org
[16:25:51] <DanielK_WMDE>	 bug 46378
[16:25:57] <DanielK_WMDE>	 !b 46378
[16:25:57] <wm-bot>	 https://bugzilla.wikimedia.org/46378
[16:25:59] <^demon>	 Ah, couldn't remember if it was client or base :)
[16:26:03] <^demon>	 ty, reading.
[16:26:34] <^demon>	 Yeah, I can run this. robla: Asher signed off?
[16:27:00] <Krinkle>	 robla: reading now, I got a hunch. I think this'll be an easy fix. Taking a few seconds to confirm.
[16:27:33] <DanielK_WMDE>	 ^demon: the script may run for quite a long time, but it can be killed and re-started later, so it doesn't get in the way of deployment. just save the log with the last row id.
[16:27:54] <^demon>	 *nod*
[16:28:27] <robla>	 DanielK_WMDE: ^demon: let's switch over to #wikimedia-operations and ask woosters
[16:28:54] <Krinkle>	 robla: Just to be clear, this is about the uncaught exception "cssText.indexOf is not a function", correct?
[16:29:21] <robla>	 Krinkle: yup
[16:29:58] <Krinkle>	 robla: Have any actions been taken against it that I should know about? Or is everything still the same in master and wmf-branches regarding this as 24 hours ago?
[16:30:14] <robla>	 Krinkle: same afaik
[16:30:19] <Krinkle>	 ok
[16:32:43] <Krinkle>	 robla: en.wiki and commons both running on the version that should include this bug? (should i be able to reproduce it on commons or enwiki?)
[16:32:50] <Krinkle>	 I can't reproduce it in neither Chrome or Firefox.
[16:32:59] <Krinkle>	 Trying with various different gadgets but nothing yet.
[16:34:21] <robla>	 Both commons and enwiki are on the branch with the bug, so yeah, it should be possible.  I'll admit I haven't been able to repro.  Erik had a quasi repro
[16:38:19] <robla>	 let's see if Kaldari can help with the repro case when he comes online
[16:40:43] <robla>	 one other option:  since we know which line of code has the actual error, perhaps we can deploy some debug code (guarded with a feature flag) that we can enable on test2 and ask people to try to repro there
[16:45:06] <gerrit-wm>	 New patchset: Ottomata; "Adding debian packaging. linking against libdb, adding upstart conf file" [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/55900
[16:47:43] <Krinkle>	 robla: Yeah, I'm considering that right now.
[16:48:02] <Krinkle>	 I've looked at the code and there's only one way it can throw "cssText.indexOf is not a function"
[16:48:18] <Krinkle>	 and that's if it is given something that is truthy- but not a string
[16:48:28] <greg-g>	 truthy?
[16:48:40] <greg-g>	 is that a Stephen Colbert-ism?
[16:48:45] <Krinkle>	 A javascript value that would cast to boolean true
[16:48:48] <greg-g>	 ah
[16:48:55] <Krinkle>	 eg. non-empty strings, numbers above zero and objects
[16:49:11] <Krinkle>	 (numbers below zero, too, actually)
[16:49:18] <Krinkle>	 which is odd, since css can only be a string
[16:49:20] <gerrit-wm>	 Change merged: Ottomata; [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/55900
[16:49:28] <greg-g>	 hrm
[16:49:40] <Krinkle>	 So something is feeding it garbage input, and I can't figure out where that could come from.
[16:49:52] <robla>	 Krinkle: I was assuming that it would be because cssText isn't initialized yet
[16:49:52] <Krinkle>	 I suspect a user script, since the server always feeds it strings, guruanteed
[16:50:09] <Krinkle>	 robla: No, cssText is a function parameter.
[16:50:40] <Krinkle>	 robla: https://github.com/wikimedia/mediawiki-core/blob/master/resources/mediawiki/mediawiki.js#L463
[16:50:55] <Krinkle>	 Called only from line 483
[16:51:32] <robla>	 but is it possible that addEmbeddedCSS isn't initialized yet for some reason?
[16:51:42] <robla>	 err....
[16:51:52] <robla>	 *what's passed into addEmbeddedCSS
[16:53:08] <Krinkle>	 robla: Not really, it is a private function (inaccessible outside mw.loader's scope) only called from mw.loader.implement->execute
[16:53:47] <Krinkle>	 which in turn is what it called from load.php to load the css
[16:54:03] <Krinkle>	 hence some users are seeing absence of stylesheets for some of these additional components
[16:54:36] <robla>	 what's the scope of cssBuffer on line 483?
[16:54:43] <Krinkle>	 robla: mw.loader's private scope
[16:55:01] <Krinkle>	 though the error is on cssText, not cssBuffer
[16:55:35] <robla>	 I'm trying to figure out how cssBuffer gets initialized prior to your "!cssBuffer" check
[16:56:03] <Krinkle>	 cssBuffer is initialised to an empty string at the top of the scope. And it is fed the new cssText after each invocation that isn't scheduled
[16:56:15] <Krinkle>	 robla: Anyhow, I already know the problem (not the cause yet)
[16:56:37] <Krinkle>	 something somewhere is manually invoking mw.loader.implement outside regular resourceloader flow with an invalid stylesheet array
[16:56:57] <Krinkle>	 Id like to push a change to gerrit and check it out on test.wikipeida only
[16:57:07] <Krinkle>	 however I"m not sure how to do that without forbidding sync for a while
[16:57:14] <Krinkle>	 it can't be done behind a feature flag, it just can't.
[16:57:28] <robla>	 hrm
[16:57:36] <^demon>	 !log it, and yell at anyone who tries?
[16:57:36] <Krinkle>	 well, I guess there is.
[16:57:45] <Krinkle>	 I'll try something
[16:57:59] <robla>	 greg-g: ^
[16:59:55] <robla>	 mlitn: Krinkle is going to need to do some deploy work
[17:00:20] <robla>	 I'll let the two of you coordinate so you don't step on each other
[17:00:58] <robla>	 anomie: grabbing a conf room and then I'll be ready
[17:01:02] <mlitn>	 Krinkle: go ahead for now - I'm waiting on code review, so not yet doing anything
[17:01:04] <anomie>	 robla- ok
[17:01:27] <mlitn>	 Krinkle: I'll poke you once I'm ready to go and/or you ping me once you're done?
[17:02:02] <Krinkle>	 mlitn: What are you ready to do?
[17:02:59] <mlitn>	 Krinkle: right now, nothing; I'm waiting on code to be reviewed before I'll start to enable AFTv5 on frwiki
[17:03:33] <Krinkle>	 mlitn: ok, I'll let you know when I'm done.
[17:03:44] <mlitn>	 great, thanks
[17:05:26] <Krinkle>	 robla: debug code up here https://gerrit.wikimedia.org/r/55903 merging now to pull on fenari
[17:12:28] <Krinkle>	 robla: matthiasmullie: "git status" on fenari in /h/w/c/php-1.21wmf12 tells me there have been updates in extension submodules that have been comitted to the wmf branch but not yet updated in the local submodule
[17:12:35] <Krinkle>	 #	modified:   extensions/GettingStarted (new commits)
[17:12:35] <Krinkle>	 #	modified:   extensions/GuidedTour (new commits)
[17:12:40] <Krinkle>	 Just FYI
[17:13:02] <Krinkle>	 whomever wants to deploy that needs to do "git submodule update extensions/GettingStarted" as well.
[17:13:06] <MatmaRex>	 hi
[17:13:07] <MatmaRex>	 where can i get a hold on the fundraising guys and shout at them?
[17:13:10] <Krinkle>	 Anyway, not deploying that for now.
[17:13:16] <MatmaRex>	 https://pl.wikipedia.org/wiki/Loren_Acton is displaying three banners, all the same, one under another (when logged out)
[17:13:54] <Krinkle>	 !log Change I34a488247 is now merged in wmf/1.21wmf12 and pulled on fenari for test.wikipedia.org, do not sync.
[17:13:55] <MatmaRex>	 and if this isn't fixed prompty, i'm going to hide the banners in site CSS. (i'm a sysop at pl.wp)
[17:14:08] <Krinkle>	 (in operations, of course)
[17:15:07] <Krinkle>	 robla: I've commented at https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#CSS_errors_with_search_box.2C_etc..3F now we have to get someone who can reproduce it to do so in a browser and have their console open and give us the stracktrace they get
[17:15:16] <robla>	 thanks!
[17:15:16] <Krinkle>	 robla: Is Erik around?
[17:17:01] <robla>	 Krinkle: lemme see if he's at his desk.
[17:17:18] <Krinkle>	 He's not on IRC it seems.
[17:20:02] <robla>	 Krinkle: I pinged Praveena on GTalk.  He's got a headset on, so I think he may be in a meeting
[17:21:32] <robla>	 Krinkle: just send him a mail.  I need to get back to my 1:1 with Brad
[17:31:31] <greg-g>	 thanks Krinkle
[17:33:08] <greg-g>	 chrismcmahon: ^^^ I know you can't reproduce either, but, just FYI, Krinkle has a fix up on test.wiki, see Krinkle's comment on the Village pump above
[17:34:43] <guillom>	 qgil: is there a deadline on the trilingual hackathon post, or can it wait tomorrow? (I may do it today, but I want to know if I have to :)
[17:36:09] <greg-g>	 AaronSchulz: The mesage to wikitech-l about page moving from pv isn't related to the issue we had last week, right? :)
[17:37:48] <greg-g>	 Aaron|home: ^^
[17:41:18] <Krinkle>	 ugh, firebug is such a horribe outdated debugger.
[17:41:26] <Krinkle>	 I can reproduce it in firefox < 14
[17:41:32] <greg-g>	 huh
[17:41:42] <Krinkle>	 But there is no stracktrace
[17:41:47] <Krinkle>	 doesn't firebug have that? wtf
[17:42:11] <robla>	 oh, repro \o/
[17:42:40] <Krinkle>	 robla: http://cl.ly/image/3d1C1v2m382c
[17:43:06] <MatmaRex>	 can you just add some guard to initialize() for now?
[17:56:07] <Krinkle>	 robla: Figured it out on my own, prepping a fix now,
[17:56:19] <robla>	 excellent!
[17:56:23] <robla>	 thanks!
[17:56:57] <greg-g>	 Krinkle: wonderful
[18:05:20] <Aaron|home>	 greg-g: no
[18:06:01] <greg-g>	 Aaron|home: good deal. thanks.
[18:06:06] <sumanah>	 robla: so the JavaScript-related regression you mentioned earlier -- it sounds like it's on its way to being fixed.  Should I be looking at bug 46401 or 46575?
[18:06:25] <robla>	 !b 46575
[18:06:25] <wm-bot>	 https://bugzilla.wikimedia.org/46575
[18:07:04] <sumanah>	 (biab)
[18:07:17] <greg-g>	 Krinkle says 46401 isn't the same issue that is being reported on the Village Pumps
[18:07:18] <matthiasmullie>	 Krinkle: I assume Krinkle just to be sure - you're no longer deploying anything?
[18:07:31] <Krinkle>	 matthiasmullie: I am still
[18:07:58] <robla>	 sumanah: 46575 I guess
[18:08:10] <matthiasmullie>	 Krinkle: alright, take your time :)
[18:08:10] <greg-g>	 yeah
[18:08:20] <greg-g>	 robla: sumanah yeah, 46575
[18:09:58] <Krinkle>	 hashar where are you
[18:10:09] <Krinkle>	 jenkins is fucked again, it can't find the git revision.
[18:10:13] <sumanah>	 andre__: ^ keeping you aware -- bug 46575 is the one that's being reported on the various VPs, it sounds like
[18:10:13] <Krinkle>	 https://integration.wikimedia.org/ci/job/mediawiki-core-phpcs-HEAD/6112/console
[18:10:18] <Krinkle>	 Will bypass for now
[18:12:00] <andre__>	 thanks
[18:13:06] <sumanah>	 guillom: it sounds like bug 46575 is the one that's being reported on the various VPs, but I'm not getting the sense from greg-g that this is something we need to send out a notice to wikitech-ambassadors about
[18:14:00] <greg-g>	 sumanah: it's not a major high impact issue that we can tell, by the absolute number of people reporting problems
[18:14:21] <greg-g>	 but, a note about a regression in functionality might be nice
[18:18:48] <Krinkle>	 greg-g: There is no regression in functionality
[18:19:24] <Krinkle>	 greg-g: That is bug 46401, which is still open and turns out to be unrelated.
[18:20:00] <greg-g>	 gotcha
[18:20:31] <greg-g>	 Krinkle: so, about that bug, is that something you can take a look at as well?
[18:20:55] <robla>	 Krinkle: I'm guessing that maybe some of what is showing up on the VPs is 46401, but that doesn't seem quite as urgent to address
[18:21:00] <Krinkle>	 greg-g: robla: James_F: Filed under bug 46575. Summary: A recent update in ResourceLoader caused an uncaught exception in older versions of Firefox due to non-standard behaviour of setTimeout(). This has been fixed and deployed.
[18:21:34] <greg-g>	 thank you
[18:21:44] <robla>	 clearly, anyone who was getting that error is affected by 46575
[18:21:47] <Krinkle>	 robla: No, none of it is bug 46401 as far as I can tell. The code that I committed that caused bug 46401, is also the commit in which I triggered that FIrefox bug, but that's just a coincendence.
[18:22:07] <awjr>	 aude: there's a PHP fatal in wikibase that appears to be affecting mobile watchlists
[18:22:16] <Krinkle>	 matthiasmullie: I'm ready.
[18:22:18] <gerrit-wm>	 New patchset: Yuvipanda; "Add Share attempts to dashbaord" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55924
[18:22:23] <matthiasmullie>	 Krinkle: ok, thanks!
[18:22:41] <Krinkle>	 Left a short notice on https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#CSS_errors_with_search_box.2C_etc..3F to close up
[18:23:07] <awjr>	 aude: PHP fatal error in /usr/local/apache/common-local/php-1.21wmf12/extensions/Wikibase/client/WikibaseClient.hooks.php line 311: <br/>
[18:23:07] <awjr>	 11:19
[18:23:07] <awjr>	         Argument 5 passed to Wikibase\ClientHooks::onSpecialWatchlistQuery() must be an array, none given
[18:28:53] <robla>	 awjr: we should converge on #wikimedia-wikidata and figure out a workaround
[18:29:12] <awjr>	 robla yeah, looks like a quick fix - mobile's in the middle of planning meetings :-/
[18:35:31] <gerrit-wm>	 New review: Milimetric; "(1 comment)" [analytics/limn-mobile-data] (master); V: 2 C: 2;  - https://gerrit.wikimedia.org/r/55924
[18:40:09] <Ryan_Lane>	 sumanah: hey. I'm in NYC
[18:40:32] <Ryan_Lane>	 sumanah: I'm going to be giving a talk at Columbia university on Thursday at 8pm, if you're interested
[18:40:37] <Ryan_Lane>	 it's going to be a talk + workshop
[18:40:39] <Ryan_Lane>	 on devops
[18:41:48] <Ryan_Lane>	 wikimedia focused, of course
[18:41:57] <gerrit-wm>	 Change merged: Milimetric; [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55924
[18:44:27] <sumanah>	 Hi Ryan_Lane! are you in NYC Thursday night and/or Friday?
[18:44:37] <Ryan_Lane>	 I'm here till the 1st
[18:45:01] <sumanah>	 I'm in Pittsburgh right now but I should be back in time to make your talk. Link?
[18:45:14] <Ryan_Lane>	 umm. not sure if there's a link
[18:45:16] <Ryan_Lane>	 let me see
[18:46:10] <Ryan_Lane>	 sumanah: http://www.google.com/calendar/event?eid=MXBucjI2YmIwdDc1aTh0NXNiMTlua2JmdmMgYWRpY3UuY29tX3R1ZDVldG1tbzVtZm11dmRmYjU0dTczM2k0QGc&ctz=America/New_York
[18:46:58] <Ryan_Lane>	 salt stack founder is also in town and may be there
[18:47:12] <sumanah>	 oh!
[18:47:22] <sumanah>	 Ryan_Lane: so, I presume it's ok to spread the word about this far & wide, right?
[18:47:42] <Ryan_Lane>	 I'm not sure. I think it may be for columbia students
[18:48:02] <Ryan_Lane>	 I'm able to bring folks, but I'm not sure if it's ok to invite all of NYC :)
[18:48:34] <Ryan_Lane>	 probably ok for some wikimedians to show up, especially tech ones
[18:48:52] <Ryan_Lane>	 since there'll be a workshop, and we'd like to get people involved in some open source projects
[18:49:25] <sumanah>	 it starts at 8pm and goes till 9; that's pretty late for a workshop to go, in my experience
[18:49:33] <sumanah>	 (I got my grad degree at Columbia)
[18:49:39] <Ryan_Lane>	 ah. cool
[18:49:45] <sumanah>	 9pm is pretty late for stuff around there, IME
[18:49:48] <Ryan_Lane>	 it's a talk + workshop, too ;)
[18:49:56] <Ryan_Lane>	 no clue why the scheduling is that way
[18:50:10] <sumanah>	 Do you think the workshop is really meant to start at 9, after the talk?
[18:50:18] <Ryan_Lane>	 probably so
[18:50:18] <sumanah>	 or is your talk more like 20 min and then 40 min workshop?
[18:50:22] <sumanah>	 btw who's your contact there?
[18:51:10] <Ryan_Lane>	 Dina Lamdany with ADI
[18:51:23] <Ryan_Lane>	 and Elaine Mao, who used to intern with Wikimedia
[18:54:29] <sumanah>	 Oh I don't remember Elaine.
[18:54:52] <sumanah>	 OK, so, I've arranged things appropriately so I think I can come -- thank you for the heads-up!
[18:55:17] <sumanah>	 Ryan_Lane: do you have stickers with you?
[18:55:21] <Reedy>	 Warning:  array_map() [<a href='function.array-map'>function.array-map</a>]: An error occurred while invoking the map callback in /usr/local/apache/common-local/php-1.21wmf12/includes/resourceloader/ResourceLoaderFileModule.php on line 611
[18:55:38] <Ryan_Lane>	 sumanah: lemme see
[18:56:32] <Ryan_Lane>	 30 or so
[18:56:39] <Ryan_Lane>	 and about 15 or so buttons
[18:59:18] <qgil_>	 sumanah, hi how you want to communicate?
[18:59:36] <qgil_>	 I mean in this meeting now  :)
[18:59:54] <sumanah>	 qgil_: Hi! I'm sorry, it's just taking a moment for me to Hangout
[18:59:58] <sumanah>	 will be there in a moment
[19:00:39] <sumanah>	 there you go
[19:00:49] <sumanah>	 Ryan_Lane: cool! that should go far :)
[19:09:57] <Aaron|home>	 anomie: https://gerrit.wikimedia.org/r/#/c/55650/
[19:22:54] <anomie>	 Aaron|home- I'm not seeing how that change actually changes anything. For line 1299, in PG $dbw->insertId() should be returning the same value $old_id was assigned on line 1291, and for line 1349 it seems like the old and new code should be the same as long as $rev_id isn't somehow 0, false, etc. What am I missing?
[19:23:18] <Aaron|home>	 in Postgres insertId() returns null
[19:25:16] <anomie>	 Aaron|home- It looks like it returns null until after nextSequenceValue() is called. Then it returns whatever was returned by the most recent call to nextSequenceValue() on that database object.
[19:29:00] <hashar>	 good evening
[19:29:07] <Aaron|home>	 anomie: I don't think it will make a difference, yes
[19:29:37] <Aaron|home>	 half of the code using it assumes it is null for some DBs and uses the seq val (with would not be null for those DBs) and the docs say it is always null
[19:29:58] <Aaron|home>	 of course the code indeed uses the last sequence number incremented
[19:30:26] <Aaron|home>	 so the usage/code/documentation is incoherent
[20:26:18] <gerrit-wm>	 New review: Krinkle; "Reported upstream as https://github.com/gruntjs/grunt-contrib-qunit/issues/27." [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/55408
[20:27:56] <gerrit-wm>	 New patchset: Hashar; "(bug 45084) beta: matrix job to update all databases" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/55949
[20:28:18] <gerrit-wm>	 New review: Hashar; "Depends on upstream patch https://review.openstack.org/#/c/25432/" [integration/jenkins-job-builder-config] (master) C: -2;  - https://gerrit.wikimedia.org/r/55949
[20:52:15] <gerrit-wm>	 New patchset: Yuvipanda; "Minor copy modifications" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55957
[20:55:16] <gerrit-wm>	 Change merged: Milimetric; [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/55957
[21:36:54] <marktraceur>	 qgil: So if Parsoid wanted to pull off a GSoC project, and we wanted to submit a proposal for it, when would be the deadline for doing it?
[21:37:21] <qgil>	 marktraceur, soon enough for me to go to your desk  :)
[21:55:03] <Reedy>	 marktraceur: That sounds slightly backward...
[21:55:17] <marktraceur>	 Reedy: How do you mean?
[21:55:44] <Reedy>	 Usually, the participant submits the proposal, not the organisation
[22:12:07] <marktraceur>	 rfaulkner: So you're the only one who responded - unless I can scrape up some other participants we can just leave whenever you're ready.
[23:01:46] <rfaulkner>	 marktraceur: gwicke is in and tfinc_ is a maybe (he has his gear with)
[23:01:53] <marktraceur>	 Oh cool
[23:05:39] * marktraceur  needs to maybe ask more explicitly for RSVPs next time