[00:00:16] <frksand89>	 Thanx for the link. Anyway, this is all new to me..Just created the account.. Bout an hour ago actually. The thing is.. It says that my account is active on 5 project sites and I have no idea how to change that..
[00:00:31] <kaldari>	 of course, it may be different for SVGs (what I'm looking at currently)
[00:01:39] <kaldari>	 frksand89: you might want to try asking on wikimedia-tech or wikipedia-en
[02:46:32] <Ryan_Lane>	 TimStarling: there are scripts for everything you need to do in ldap
[02:46:43] <Ryan_Lane>	 TimStarling: it's *very* rare to need to make an ldif
[02:46:50] <Ryan_Lane>	 modify-ldap-group will work
[02:47:09] <Ryan_Lane>	 if there's a command to do it locally on the system, there should be a script to do it too
[06:20:36] <gerrit-wm>	 New review: Tim Starling; "void append_char(char* s, char c) {" [analytics/udp-filters] (master); V: 0 C: -2;  - https://gerrit.wikimedia.org/r/3222
[08:12:44] * RoanKattouw  grabs a drink and sets out to do the largest rebase in his life

[08:13:56] <RoanKattouw>	 RL2 and master have diverged by 219 and 118 commits respectively
[08:22:09] <Nikerabbit>	 wii
[08:46:57] <RoanKattouw>	 Krinkle-away: You here yet?
[09:57:28] <RoanKattouw>	 Krinkle-away: ping?
[10:41:26] <gerrit-wm>	 New review: Hashar; "Marking do not submit for now till I have time to review and actually configure it on Jenkins." [integration/jenkins] (master); V: 0 C: -2;  - https://gerrit.wikimedia.org/r/2289
[10:59:08] <hashar>	 RoanKattouw: I have seen your mail about direct push to non master branches
[10:59:15] <hashar>	 I guess if people want it them so be it
[10:59:30] <hashar>	 but we probably do not want that on mediawiki/core.git
[10:59:45] <hashar>	 will review the access rights this afternoon
[10:59:45] <RoanKattouw>	 Why not?
[10:59:48] <RoanKattouw>	 OK
[11:00:15] <hashar>	 we still have to polish up the workflow with branches
[11:00:21] <hashar>	 I think Rob has been working on it
[11:00:22] <hashar>	 and Chad
[11:00:27] <RoanKattouw>	 OK
[11:00:32] <RoanKattouw>	 I'll add RobLa to CC then
[11:00:38] <hashar>	 might be a good idea
[11:00:45] <hashar>	 and sumanah :-)
[11:00:51] <hashar>	 since she is the project manager somehow
[11:01:16] <RoanKattouw>	 meh too late
[11:01:18] <hashar>	 ahah
[11:01:27] <hashar>	 it is ok
[11:01:39] <hashar>	 she is just going to complain about the platform engineering Cabal
[11:01:46] <hashar>	 ;-D
[11:01:51] <RoanKattouw>	 I was mostly just wondering if there was a technical reason not to do this
[11:01:59] <hashar>	 none I know of
[11:02:05] <RoanKattouw>	 From a social perspective I don't think it's a problem
[11:02:25] <hashar>	 merely a question of process. I think we want any submissions to be associated to a change in gerrit
[11:02:34] <RoanKattouw>	 Well still
[11:02:44] <RoanKattouw>	 In order to submit it to master, there has to be a Gerrit change somewhere
[11:03:10] <hashar>	 yup
[11:03:52] <RoanKattouw>	 So people could submit a merge commit (this is not ideal because Gerrit doesn't display diffs for those, so you can't do inline commenting), or they could squash-merge their branch
[11:04:15] <hashar>	 I think operations/puppet is configured like you did;  Aka you can not push directly to production branch but you can to the test branch.
[11:04:38] <hashar>	 yeah we disallow merge commit
[11:04:57] <RoanKattouw>	 Are you sure?
[11:05:03] <RoanKattouw>	 https://gerrit.wikimedia.org/r/3854
[11:05:04] <hashar>	 overall, we have no idea how to handle working with multiple branches
[11:05:16] <hashar>	 ohhhhh
[11:05:32] <RoanKattouw>	 To be fair, I am a project owner
[11:05:38] <RoanKattouw>	 So maybe I have special powers in that regard
[11:05:59] <RoanKattouw>	 (FYI, RL2 is a gated working branch for now)
[11:06:49] <hashar>	 hmm
[11:06:53] <hashar>	 that is indeed a merge commit
[11:07:11] <hashar>	 which looks clean to me
[11:07:36] <hashar>	 couldn't you just rebase RL2 on top of master?
[11:07:40] <hashar>	 (just wondering)
[11:08:40] <hashar>	 anyway that is a perfect example to build on
[11:09:00] <hashar>	 I guess that would help us define a process (or polish the process if it is being written)
[11:10:16] <hashar>	 lunch time anyway I am starving
[11:11:37] <RoanKattouw>	 I've done that too
[11:11:40] <RoanKattouw>	 Was very tedious
[11:11:58] <RoanKattouw>	 But then I'd have to submit it as a new branch, because we disable force push (with good reason)
[12:12:15] <Krinkle-away>	 RoanKattouw: Hi there
[12:12:29] <RoanKattouw>	 Ohai
[12:12:32] <Krinkle>	 I slept in a bit, still have to get used to nl timezone (how ironic)
[12:12:36] <RoanKattouw>	 hehe
[12:12:43] <RoanKattouw>	 And I was up at 7:45 today
[12:12:51] <Krinkle>	 wow
[12:13:01] <RoanKattouw>	 I've been up early all week, it's a remnant of my jetlag
[12:13:09] <RoanKattouw>	 And I'm not too eager to get rid of it either, it's useful :)
[12:13:16] <RoanKattouw>	 (otoh I fall over at 9:30pm)
[12:13:47] <Krinkle>	 yeah, I would've gone to bed earlier but I didn't fully realized how early 10am was.
[12:15:21] <Krinkle>	 checking the task management page now
[12:15:31] <^demon>	 RoanKattouw: Can we upgrade gerrit-dev?
[12:15:43] <Krinkle>	 +* (RK) Per-module config in core, apparently Andrew needs it
[12:15:46] <Krinkle>	 hm..
[12:16:09] <Krinkle>	 interesting, per-module config for non-gadgets as well I suppose. Configurable from LocalSettings?
[12:16:58] <RoanKattouw>	 ^demon: Feel free
[12:17:08] <^demon>	 2.3rc0 :D
[12:17:23] <RoanKattouw>	 ^demon: Are you on the project already? If not, tell me and I'll add you
[12:17:31] <^demon>	 Yeah I am.
[12:17:48] <Krinkle>	 ^demon: There is a good chance that upgrading gerrit even to the slightest newer version will render the stylesheet useless. I suspect the classnames in gerrit are completely rendered automatically based on whatever is the most common color on the page etc. (not kidding)
[12:17:49] <^demon>	 Now, where is Antoine...
[12:18:05] <Krinkle>	 but just on labs
[12:18:06] <Krinkle>	 that should be fine
[12:18:58] <^demon>	 Yeah, I really really wanna start playing with 2.3 though :)
[12:19:12] <^demon>	 It's got some must-have features and I plan to upgrade production as soon as we're confident.
[12:21:57] <Krinkle>	 k
[12:22:46] <^demon>	 https://gerrit-review.googlesource.com/#/c/33220/9/ReleaseNotes/ReleaseNotes-2.3.txt - planned release notes, worth reading.
[12:24:56] <RoanKattouw>	 issue 203 Create project through web interface
[12:26:03] <^demon>	 There's that :)
[12:26:16] <^demon>	 There's also "Allow superprojects to subscribe to submodules updates"
[12:26:18] <^demon>	 My pet issue
[12:26:33] <RoanKattouw>	 Krinkle: OK, so, RL@
[12:26:34] <RoanKattouw>	 *RL2
[12:26:43] <Krinkle>	 Yes
[12:26:54] <Krinkle>	 I was reading Erik's mail yesterday
[12:27:05] <RoanKattouw>	 Did you see... yes that
[12:27:08] <RoanKattouw>	 And https://gerrit.wikimedia.org/r/3854
[12:27:17] <Krinkle>	 Perhaps we should go for an earlier release that doesn't change much for the user
[12:27:26] <Krinkle>	 get the repo and RL-only stuff out first in a stable fashion
[12:27:31] <Krinkle>	 (and Special:Gadgets)
[12:27:53] <RoanKattouw>	 A lot of changes are user-facing though
[12:27:59] <RoanKattouw>	 Like the Gadget definition: namespace, Special:Gadgets, etc
[12:28:04] <Krinkle>	 Yes
[12:28:07] <Krinkle>	 I meant more like no procedural changes
[12:28:10] <Krinkle>	 just where what is what
[12:28:14] <RoanKattouw>	 Or well I guess they're administrator-changes
[12:28:29] <RoanKattouw>	 Oh you mean the things that Erik is suggesting?
[12:28:31] <RoanKattouw>	 Or ... what?
[12:28:41] <RoanKattouw>	 Could you give an example of a change that should be excluded?
[12:28:56] <Krinkle>	 no not like that
[12:29:16] <Krinkle>	 I meant something we might be going to implement, that we should perhaps move to a Gadgets 2.5 plan
[12:29:19] <Krinkle>	 like module prefs
[12:29:54] <Krinkle>	 which is a fairly major change that is going to add a new feature for gadget writes and also change the interface for users
[12:30:25] <Krinkle>	 and I am not fully sure we can get that into a state that we are willing to support into "infinity"
[12:30:33] <Krinkle>	 within reasonable time
[12:31:10] * Krinkle  checks gerrit link

[12:31:41] <RoanKattouw>	 Oh, OK
[12:31:43] <RoanKattouw>	 Yeah that's fair
[12:31:44] <Krinkle>	 what ? "splitting of Gadget modules in a scripts module and a style module"
[12:31:53] <Krinkle>	 I know you didn't merge it but… that was reverted in trunk, not?
[12:31:58] <RoanKattouw>	 No, it's still in master
[12:32:02] <RoanKattouw>	 Wait
[12:32:06] <RoanKattouw>	 No, I reverted that in master
[12:32:15] <RoanKattouw>	     9009f36e81b3189ecb0cf77de63924eab25d66a7 (Revert r106007 (Split Gadget modules into two, one for scripts and one for styles): I like the concept of having multiple modules, and I'll probably implement this in the RL2 branch, but this specific implementation has issues (subclassing OutputPage, building its own head items) that were pointed out on CR but never fixed.)
[12:32:28] <Krinkle>	 Right, I remember us being fairly clear on that it wasn't done "right"
[12:32:37] <RoanKattouw>	 Yes, but I would like us to do it
[12:32:48] <RoanKattouw>	 Similar thing for skin-specific gadgets
[12:33:02] <RoanKattouw>	 That was implemented right, but the underlying code has been rewritten so much that we'll have to redo it
[12:33:10] <Krinkle>	 Yeah
[12:33:30] <Krinkle>	 that should be fairly simple to support in the front-end, it's only an addition to Special:Gadgets editor
[12:34:02] <Krinkle>	 RoanKattouw:
[12:34:02] <Krinkle>	 294abd1e32b365e1d84e4f7448a8694a1a989b6e (Move styles in gadgets to the top of pages to avoid flash of unstyled content.)
[12:34:20] <Krinkle>	 did that end up in 1.19wmf1? I read a bug the other day that seemed like that happened
[12:34:29] <RoanKattouw>	 I don't think so
[12:35:01] <^demon>	 gerrit-dev now running 2.3rc0
[12:35:02] <^demon>	 :)
[12:35:07] <Krinkle>	 !b 35471 | RoanKattouw
[12:35:07] <mw-bot>	 RoanKattouw: https://bugzilla.wikimedia.org/show_bug.cgi?id=35471
[12:35:35] <^demon>	 Sweet! You can disable/hide projects :D
[12:35:39] <Krinkle>	 RoanKattouw: the @import bug with concatenating <style> isn't new after al
[12:35:51] <Krinkle>	 it is new for <style> but it was already in only=styles
[12:36:07] <RoanKattouw>	 Ah of course
[12:36:23] <RoanKattouw>	 Ah, I think the module split isn't in 1.19wmf1, but 1.19wmf1 may be using addModuleStyles()
[12:36:42] <Krinkle>	 basically only=styles should never be used for more than 1 user-created module
[12:37:22] <Krinkle>	 ah, yeah, I see where that separation can come in handy
[12:37:24] <Krinkle>	 but ahm,..
[12:37:26] <Krinkle>	 isn't it redundant with
[12:37:34] <Krinkle>	 isn't it redundant with position=top being supported in Gadgets 2.0?
[12:38:15] <Krinkle>	 If the gadget has JS and CSS the CSS likely applies to dynamically created content, so a bottom load with css right before js is fine then.
[12:38:32] <Krinkle>	 but "style"-only gadgets need to be given position=top, that's how we do it in core and extensions
[12:38:48] <RoanKattouw>	 Yes, you're right, it is
[12:38:56] <RoanKattouw>	 Only problem is, RL2 doesn't actually support position=top yet :D
[12:39:07] <Krinkle>	 Are you sure?
[12:39:14] <Krinkle>	 Hm..
[12:39:20] <RoanKattouw>	 That's OK
[12:39:23] <Krinkle>	 Oh, we just din't add it yet.
[12:39:25] <RoanKattouw>	 I have to add skin anyway
[12:39:28] <Krinkle>	 that's a simple addition though
[12:39:42] <Krinkle>	 ah, I lost RL2 in svn update, its gone in svn completely
[12:40:00] <RoanKattouw>	 Yes
[12:40:10] <Krinkle>	 RoanKattouw: I know I can switch branches with git, but I'd like to keep master available as well. Should I make a separate clone then?
[12:40:16] <RoanKattouw>	 No
[12:40:34] <RoanKattouw>	 Clone the Gadgets repo, then run git checkout -b RL2 origin/RL2
[12:40:51] <Krinkle>	 ah, right Gadgets has its own repo
[12:41:09] <Krinkle>	 I'm fine not having RL2 and master available at the same time
[12:41:44] <RoanKattouw>	 Well you can only have one of the two checked out at the same time, unless you have two clones
[12:42:15] <^demon>	 You can fake it with submodules ;-)
[12:42:18] <RoanKattouw>	 git clone ssh://gerrit.wikimedia.org:29418/mediawiki/extensions/Gadgets.git
[12:42:23] <RoanKattouw>	 cd Gadgets
[12:42:26] <Krinkle>	 RoanKattouw: Yeah, if it were about core I would like to be able to access both from http://localhost/
[12:42:26] <^demon>	 Probably more effort than it's worth
[12:42:27] <RoanKattouw>	 git checkout -b RL2 origin/RL2
[12:44:40] <RoanKattouw>	 Meanwhile, I'm looking at https://www.mediawiki.org/wiki/ResourceLoader/V2_Task_management#Todo_list . Looks like nothing in the "This week" section got done :D
[12:45:07] <Krinkle>	 I think I did do "(TT) Fix preference tables display bug causing varying "gap" at some point
[12:45:15] <Krinkle>	 but not in the mysterious "this week"
[12:45:29] <Krinkle>	 fixed in core
[12:45:35] <RoanKattouw>	 OK
[12:45:38] <RoanKattouw>	 Striking that then
[12:46:04] <Krinkle>	 yep, confirmed
[12:46:13] <RoanKattouw>	 I may or may not have done my thing
[12:46:52] <Krinkle>	 So delete addModuleStyle from master?
[12:47:16] <RoanKattouw>	 Oh for that bug
[12:47:18] <RoanKattouw>	 yeah
[12:48:12] <Krinkle>	 By the way, I wrote up some very rough notes the other night for what I think we might do in 3.0, https://www.mediawiki.org/wiki/User:Krinkle/Gadgets_3.0
[12:48:20] <Krinkle>	 not for the here and now, but perhaps something to read at some point :)
[12:48:34] <RoanKattouw>	 Ok
[12:48:52] <^demon>	 RoanKattouw: I replied to your e-mail re: branches.
[12:49:01] <RoanKattouw>	 OK
[12:49:22] <Krinkle>	 basically allowing a user to create a repository manifest in JSON on a user sub page, and users who trust that user can add that repo url to their Gadgets dashboard when searching for gadgets.
[12:49:47] <RoanKattouw>	 aha
[12:49:48] <Krinkle>	 ohwell, lets ignore that for now :/
[12:50:13] <Krinkle>	 A6 | Research async-behavior of <script> and <link> when inserted to HEAD a
[12:50:20] <Krinkle>	 that's all done and taken care of in core, right?
[12:50:42] <RoanKattouw>	 Yes, the async loading thing
[12:50:59] <Krinkle>	 was it implemented this year?
[12:51:03] <RoanKattouw>	 (That variable was a PR disaster BTW, the WMF EPMs blamed all frontend-related issues in 1.19 on that)
[12:51:05] <RoanKattouw>	 Yes, that was in '12
[12:51:29] <Krinkle>	 Yeah I read it afterwards
[12:52:19] <Krinkle>	 "Shared gadgets in preferences" (and related ones). So the main decision we'll have to make is wether we should primarily organize by category or by repository.
[12:52:47] <Krinkle>	 I'm thinking category and mention the repository wiki-id somewhere for each gadget
[12:54:29] <RoanKattouw>	 Yeah I think that's what we decided on
[12:54:46] <RoanKattouw>	 I need to reconfigure my Apache and point it to my git clones
[12:55:56] <Krinkle>	 I always make symlinks in /var/www to wherever stuff is
[12:56:54] <RoanKattouw>	 I have Alias rules
[12:57:10] <Krinkle>	 yeah I figured
[12:59:43] <Krinkle>	 RoanKattouw: So when looking over the preferences panel again, and noticing descriptions are still broken I realized this may be a good thing. Rendering all those advertisements as wikitext blocks in the <label> is going to get messy. I have an idea just tell me if it's crazy
[13:00:02] <RoanKattouw>	 "advertisements"?
[13:00:11] <Krinkle>	 Add {Enable|Disable} buttons on Special:Gadgets and make ajax POST requests to Special:Preferences
[13:00:20] <Krinkle>	 (since there is no API)
[13:00:23] <RoanKattouw>	 What the hell are you talking about, I don't follow
[13:00:24] <Krinkle>	 (and it works)
[13:00:25] <^demon>	 Hahaha, I managed to change the -1 message :p
[13:00:26] <^demon>	 http://gerrit-dev.wmflabs.org/r/#/c/3/
[13:00:34] <^demon>	 "Patch Set 1: Oh snap!!"
[13:00:42] <RoanKattouw>	 ^demon: Dude, Sumana will love you
[13:00:55] <Krinkle>	 RoanKattouw: well, rendering descriptions on Special:Preferences for gadgets is going to become ugly and long. on Special:Gadgets it looks  nice
[13:00:58] <^demon>	 I don't know if it changes for old comments if they're hardcoded.
[13:01:00] <^demon>	 Gonna test that now
[13:01:08] <RoanKattouw>	 Now if you could configure it so that instead of V:+1 and C:+2 we have V:+1, C:+1 and A:+1 (approved instead of +2)
[13:01:35] <RoanKattouw>	 Krinkle: OK, so you want just the titles there?
[13:01:42] <Krinkle>	 so I thought we keep the checkboxes on Preferences clean with just the titles and on Special:Gadgets the descriptions are rendered nicely in a dedicated page. Only thing is, we don't want users to jump back and forth.
[13:01:51] <RoanKattouw>	 Tooltips
[13:02:07] <Krinkle>	 but there is no preferences API yet, so I'm not sure we should abuse that in production
[13:02:28] <RoanKattouw>	 We should not abuse that
[13:02:33] <RoanKattouw>	 And yes there should be a preferences API
[13:02:41] <Krinkle>	 Tooltips would become annoying imho. Users shouldn't have to hover to get primary information
[13:03:02] <RoanKattouw>	 Well what else
[13:03:26] <RoanKattouw>	 We could use JS-based tooltips with a hover timeout of zero
[13:03:27] <Krinkle>	 ok, well have to hope that until Gadgets 3.0 people won't find out that gadget descriptions can now be full blown wikipages with tables, images and such :)
[13:03:47] <RoanKattouw>	 OK, so leave it as is then?
[13:03:50] <Krinkle>	 yeah
[13:03:53] <RoanKattouw>	 That goes back to the question I asked in Gerrit
[13:03:55] <Krinkle>	 or rather, fix it so it will be as it was
[13:04:19] <RoanKattouw>	 There is a $1: $2 message somewhere, with the description injected as the second parameter, right?
[13:04:24] <Krinkle>	 " Fix JS to not touch markup of local gadgets preferences"
[13:04:28] <Krinkle>	 Yes
[13:04:31] <RoanKattouw>	 'gadgets-preference-description' => '$1: $2',
[13:04:50] <^demon>	 RoanKattouw: Simple as a one-line database change, stop+start gerrit. Does not change old values, those have been hardcoded in submitted comments.
[13:04:53] <RoanKattouw>	 In my merge I changed that to '$1: <span dir="auto">$2</span>', but I guess that should be a div then
[13:05:00] <RoanKattouw>	 ^demon: Sweet
[13:05:09] <RoanKattouw>	 But it's in the DB so it's not configurable?
[13:05:09] <^demon>	 I changed it to "Oh snap!!" then "THIS IS THE WORST CHANGE EVER!!!"
[13:05:23] <^demon>	 It's in the database, so it's not configurable from the UI, no.
[13:05:25] <RoanKattouw>	 OTOH that means we can do it in production too, right?
[13:05:37] <^demon>	 Yeah, should be able to. I'll run it by Ryan.
[13:05:37] <RoanKattouw>	 My concern was that it's not configurable from puppet either
[13:05:54] <Krinkle>	 RoanKattouw: yes, div it is
[13:05:56] <RoanKattouw>	 OK
[13:06:40] <Krinkle>	 RoanKattouw: I find no span in i18n.php though
[13:06:48] <Krinkle>	 or "auto"
[13:07:11] <Krinkle>	 	'gadgets-preference-description' => '$1: $2',
[13:07:18] <RoanKattouw>	 Krinkle: https://gerrit.wikimedia.org/r/3854
[13:07:28] <Krinkle>	 oh,r ight. it's not reviewed yet
[13:07:30] <RoanKattouw>	 I merged master into RL2 and then submitted the merge for review
[13:07:45] <Krinkle>	 review would merge it into the RL-branch, right?
[13:07:47] <RoanKattouw>	 Gerrit won't show you a diff there because it's a merge commit, see instructions in the comment
[13:07:49] <RoanKattouw>	 Yes
[13:08:56] <Krinkle>	 ok, reviewing now
[13:09:42] <Krinkle>	 RoanKattouw: Andrew's need, can you elaborate on that?
[13:10:24] <RoanKattouw>	 Let's see
[13:10:35] <RoanKattouw>	 Per-module *config* not *preferences8
[13:10:40] <RoanKattouw>	 It's something I've wanted for a long long time
[13:11:01] <Krinkle>	 wow, that diff is full of stuff that appears to be utf-8 failures?
[13:11:21] <RoanKattouw>	 That's what came in from master yeah
[13:11:33] <RoanKattouw>	 It's adding a character before and after
[13:11:50] <Krinkle>	 http://i.imgur.com/WtGK3.png
[13:12:03] <Krinkle>	 ai, transparent Terminal windows..
[13:12:07] <RoanKattouw>	 -/** Kazakh (Latin) (Қазақша (Latin)) */
[13:12:09] <RoanKattouw>	 +/** Kazakh (Latin script) (<U+202A>Qazaqşa (latın)<U+202C>) */
[13:12:30] <Krinkle>	 yeah, I haven't seen those <> in git diffs before
[13:12:33] <RoanKattouw>	 You removed some of the U+202A characters at one point, but apparently those things are in master
[13:12:36] <^demon>	 Krinkle: I prefer beige :)
[13:12:44] <RoanKattouw>	 Krinkle: That's a terminal thing I thnik
[13:12:51] <Krinkle>	 ok
[13:13:09] <Krinkle>	 RoanKattouw:
[13:13:09] <Krinkle>	 +       'gadgets-preference-description' => '$1: <div dir="auto">$2</div>',
[13:13:11] <Krinkle>	 :)
[13:13:20] <Krinkle>	 so, it already was a div in your merge
[13:13:30] <Krinkle>	 as you said you would do in the comment, but thought you didn;t?
[13:14:40] <RoanKattouw>	 No, I changed my merge
[13:14:46] <RoanKattouw>	 It was a span at first, I changed that 5 mins ago
[13:15:09] <Krinkle>	 ok
[13:15:29] <Krinkle>	 I'm confused though, the diff I got from the cmd you gave only shows i18n changes and whitespace+type change in other files
[13:15:40] <Krinkle>	 that's all it is supposed to be?
[13:16:49] <RoanKattouw>	 Ah, let's see here
[13:17:13] <RoanKattouw>	 There are some typo fixes and doc comment additions
[13:17:24] <RoanKattouw>	 Removal of deprecated wgExtensionAliasesFiles
[13:17:39] <Krinkle>	 yes I spotted that one too
[13:17:41] <RoanKattouw>	 s/category/categories for a member variable
[13:17:51] <RoanKattouw>	 Unused $type in foreach
[13:17:58] <RoanKattouw>	 Typo fix (possble)
[13:18:09] <Krinkle>	 merging global statement
[13:18:14] <RoanKattouw>	 Global consolidation in the .... yeah
[13:18:16] <RoanKattouw>	 And that's it
[13:18:39] <RoanKattouw>	 There were more conflicts but I resolved those in RL2's favor, so you don't see those
[13:18:52] <Krinkle>	 ok
[13:19:00] <Krinkle>	 I only have to review Patch Set 3, right?
[13:19:32] <RoanKattouw>	 Yes
[13:19:50] <RoanKattouw>	 1 and 2 are obsolete
[13:20:30] <RoanKattouw>	 If you approve, you'll need to give it a Verified +1 (because we don't have Jenkins integration yet) and a Code Review +2, then click "Publish and submit"
[13:21:14] <Krinkle>	 Local testing turned out OK
[13:21:16] <Krinkle>	 done
[13:21:25] <RoanKattouw>	 Re per-module config: the feature I had in mind is that each module has a list of variables it wants to export, and they're exported in a namespaced manner, not globally
[13:21:32] <RoanKattouw>	 I don't remember why Andrew wanted this
[13:21:41] <RoanKattouw>	 It's a good idea but not really related to this project
[13:21:42] <RoanKattouw>	 Thanks
[13:22:06] <Krinkle>	 so a Map instance per module
[13:22:10] <RoanKattouw>	 yeah
[13:22:20] <RoanKattouw>	 Anyway, that can wait
[13:22:21] <Krinkle>	 makes perfect sense, that's how I imagined it for gadget preferences as well
[13:22:35] <RoanKattouw>	 I'd like to identify what large things still need to be done before this is semi-releaseable
[13:24:17] <Krinkle>	 RoanKattouw: I suppose we go through V2TM and either move them to 2.1 or keep them, and if keeping whether today or later
[13:24:27] <RoanKattouw>	 OK
[13:25:12] <Krinkle>	 Shared gadgets in preferences: Marking as done
[13:25:17] <Krinkle>	 (visual idea)
[13:25:31] <RoanKattouw>	 OK
[13:25:39] <RoanKattouw>	 I've honestly forgotten what the state of that is
[13:26:02] <RoanKattouw>	 We'll have to move the test install to labs, too
[13:26:25] <RoanKattouw>	 Or, *I* will have to
[13:26:41] <Krinkle>	 Layout similar to local gadgets, but with potentially multiple sources. we merge categories and I'll add a little badge for "source-wiki-id" to the output of a shared gadget.
[13:27:55] <Krinkle>	 C6
[13:28:05] <Krinkle>	 that's you
[13:28:21] <RoanKattouw>	 "Make sure page title is also a valid module name before insertion to database, and return custom error to UI or API when invalid."
[13:28:24] <RoanKattouw>	 Yeah that's me
[13:28:28] <Krinkle>	 yeah, when creating pages
[13:29:18] <Krinkle>	 "eny edits to .js pages in NS_GADGET_DEFINITION that result in invalid JSON "
[13:29:25] <RoanKattouw>	 Yes, all the validation stuff
[13:29:33] <RoanKattouw>	 And normalization
[13:30:13] <RoanKattouw>	 So 1) validation of module names / page titles , 2) validation of submitted JSON, 3) normalization of submitted JSON
[13:31:05] <Krinkle>	 yep, and prettify json save to wikipage revision/text
[13:31:16] <RoanKattouw>	 aye
[13:31:36] <Krinkle>	 D1
[13:31:44] <Krinkle>	 that's me
[13:31:52] <Krinkle>	 error handling, and AJAX-ifying delete
[13:32:03] <Krinkle>	 delete already works, but links to wiki page action=delete
[13:32:46] <Krinkle>	 Hm.. not sure that is a bad thing actually
[13:32:55] <Krinkle>	 we don't have a gadget API to delete anymore, right?
[13:33:01] <Krinkle>	 it's hooked into the wikipage actions
[13:33:06] <RoanKattouw>	 Yeah, that's right
[13:33:25] <RoanKattouw>	 Oh, right, unit tests
[13:33:32] <RoanKattouw>	 THAT's what I did last time I worked on this codebase
[13:33:41] <Krinkle>	 linking to native interface for editing is bad, but of deletion I think it makes sense. Allowing local stuff that customized delete page (and delete reason drop down) to exist
[13:33:51] <RoanKattouw>	 Yeah
[13:33:57] <RoanKattouw>	 Sounds good
[13:34:47] <RoanKattouw>	 So from this I gather that the largest TODO items are these:
[13:34:57] <RoanKattouw>	 1) Validation/normalization of Gadget definition pages (Roan)
[13:35:09] <RoanKattouw>	 2) Error handling and general polishing of the AJAX stuff (Timo)
[13:35:10] <Krinkle>	 so, @import bug in core for <style> tags, that's a regression. Unless we come up with a miracle, I think wikis with a large number of extensions with modules are going to have bad luck on IE again?
[13:35:26] <RoanKattouw>	 rawr, IE
[13:36:05] <Krinkle>	 I don't like it we may not have a choice, if anything it wouldn't be a regression from 1.19 since the bug has been in RL since day 1
[13:36:17] <RoanKattouw>	 @import is user-authored modules only, right? Can we make that assumption?
[13:36:24] <Krinkle>	 Hm..
[13:36:40] <Krinkle>	 Well, we could banish it from RL
[13:36:47] <RoanKattouw>	 3) Unit tests (Roan)
[13:36:50] <Krinkle>	 only in site and user level
[13:36:57] <RoanKattouw>	 4) Misc bug fixes and tweaks (both)
[13:37:10] <RoanKattouw>	 5) Integration of Salvatore's code, but this can happen later
[13:37:32] <Krinkle>	 simply document "Thou shalt not use @import in core  & extensions"
[13:37:36] <RoanKattouw>	 What I'm saying is I think it's reasonable to say that if you're core or an extension, you cannot use @import
[13:37:39] <RoanKattouw>	 ....jinx
[13:38:36] <Krinkle>	 What do you have in mind?
[13:39:02] <RoanKattouw>	 Well it reduces the problem
[13:39:13] <RoanKattouw>	 But I guess we'd still have to split up user-authored stylesheets somehow
[13:39:41] <Krinkle>	 are there any that aren't in only=styles?
[13:39:55] <Krinkle>	 I wonder what happens if I have css in my common.css and a @import in my vector.css
[13:40:08] <RoanKattouw>	 So yeah that is a problem
[13:40:14] <RoanKattouw>	 And there's Gadgets of coures
[13:40:20] <Krinkle>	 yes
[13:40:25] <Krinkle>	 but those should never be in only=styles
[13:40:36] <RoanKattouw>	 No, that's right
[13:40:43] <RoanKattouw>	 But they'll still be in <style> tags
[13:40:44] <Krinkle>	 evnethough they are in 1.19wmf1 now
[13:40:47] <Krinkle>	 right
[13:41:00] <Krinkle>	 can we assume a certain order?
[13:41:24] <Krinkle>	 core/extensions depending on gadgets?
[13:41:29] <Krinkle>	 core not, for sure
[13:41:47] <Krinkle>	 extensions… I do encourage that, although I haven't seen it so far.
[13:42:31] <Krinkle>	 RoanKattouw: We could also resort to string matching, if contains @import, create a new one after the current one and concat to that from now on, and so on whenever there is one
[13:42:42] <RoanKattouw>	 Yeah for <style> tags that could work
[13:42:57] <RoanKattouw>	 I was gonna suggest moving @import up to the start, but that changes the cascading order I guess
[13:43:01] <Krinkle>	 It does mean that it could still hit the 31 limit, but there is no way around that
[13:43:11] <Krinkle>	 indeed
[13:44:00] <Krinkle>	 going to check on a wmf wiki if skin-css and common-css are concatenated after another right now for user
[13:44:18] <RoanKattouw>	 I think they aer
[13:45:09] <Krinkle>	 indeed
[13:45:13] <Krinkle>	 https://bits.wikimedia.org/en.wikipedia.org/load.php?debug=false&lang=en&modules=user&only=styles&skin=vector&user=Krinkle&version=20120328T134440Z&*
[13:45:20] <Krinkle>	 (right, you don't see that, or do you?)
[13:45:40] <RoanKattouw>	 No
[13:45:44] <Krinkle>	  debug{hello:1 } @import url('//meta.wikimedia.org/w/index.php?title=User:Krinkle/global.css&action=raw&ctype=text/css');
[13:46:10] <Krinkle>	 common.css + skin.css
[13:47:00] <Krinkle>	 so the only reason we're loading those through only=styles is to avoid having to wait 5 minutes (users want instant result)
[13:47:08] <Krinkle>	 or no, it can't be in the startup module
[13:48:01] <RoanKattouw>	 That's right
[13:48:16] <Krinkle>	 what about a regular load.php load?
[13:48:20] <Krinkle>	 without only=
[13:48:20] <RoanKattouw>	 User JS does that I think
[13:48:36] <Krinkle>	 we wanted to move user js to the top?
[13:48:52] <RoanKattouw>	 IIRC user JS is loaded with a server-provided timestamp baked into the HTML (which is OK because logged-in page views aren't cached)
[13:49:03] <RoanKattouw>	 User CSS isn't loaded this way because we want it to be available without JS
[13:49:08] <Krinkle>	 yes, that way they are also instantly refreshed (the timestamps)
[13:49:18] <RoanKattouw>	 Yeah
[13:49:37] <RoanKattouw>	 We do this for user CSS too I think
[13:49:51] <RoanKattouw>	 Probably not for site CSS though
[13:49:52] <Krinkle>	 see url 10 lines up
[13:49:59] <RoanKattouw>	 Yeah, it contains a timesatmp
[13:50:04] <Krinkle>	 yes, it is the same for all <link> and <script> tags except startup module
[13:51:49] <Krinkle>	 so the only reason nobody has complained about it so far I guess is because since Vector introduction people use common.css most of the time
[13:52:00] <Krinkle>	 both for site and user css
[13:52:10] <Krinkle>	 (since it is common.css + <skin>.css)
[13:52:43] <drecodeam>	 JeroenDeDauw: hi
[13:57:09] <Krinkle>	 RoanKattouw: filed a bug about it
[13:57:15] <Krinkle>	 !b 35562
[13:57:17] <mw-bot>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=35562
[14:00:57] <Krinkle>	 RoanKattouw: Are you working on something?
[14:01:08] <RoanKattouw>	 I just replied to Terry
[14:01:10] <RoanKattouw>	 Now replying to Erik
[14:01:19] <Krinkle>	 ok
[14:01:21] <RoanKattouw>	 I'll finish that and then I'll have stuff to talk to you about
[14:04:43] <Krinkle>	 RoanKattouw: Hm.. "Hello Krinkle, I'd like you to do a code review.  Please visit .." I guess you didn't write that yourself?
[14:05:24] <RoanKattouw>	 No, that's Gerrit
[14:05:31] <RoanKattouw>	 You already reviewed that one ,right?
[14:05:33] <Krinkle>	 (mail from Roan Kattouw (catrope
[14:05:34] <Krinkle>	 yes
[14:05:43] <Krinkle>	 ) <gerrit@wikimedia.org>
[14:05:50] <Krinkle>	 ah, right!
[14:06:42] <RoanKattouw>	 OK, so
[14:07:00] <RoanKattouw>	 When you read the e-mails I just sent to Terry and Erik, you'll see that I suggested blocking out 1 day per week for you and I to work on RL2
[14:07:15] <Krinkle>	 "…and Timo has some evil plans for Gadgets 3.0 on a wiki page somewhere."
[14:07:15] <Krinkle>	 xD
[14:07:26] <RoanKattouw>	 I think I pretty much have to do that, otherwise I'll never have time to work on this project at all
[14:08:21] <RoanKattouw>	 So do you like that idea, and do you have a specific day in mind?
[14:09:03] <RoanKattouw>	 (Keep in mind that I'm usually in the office 10am-6pm PDT so we're talking evenings; but I can do stuff like 7am-3pm PDT one day per week if that works better for you)
[14:09:14] <RoanKattouw>	 (brb)
[14:10:00] <Krinkle>	 RoanKattouw_away: As of last month until (give or take) the date go my graduation I have a very nice college schedule
[14:10:28] <Krinkle>	 mon: college all day, tue: NONE, wed: NONE, thu: morning only, fri: all day
[14:10:38] <Krinkle>	 so Tuesday and Wednesday would be ideal
[14:11:20] <Krinkle>	 but even other days I can spend about 3-4 hours since you get in around the time I get back from an allday- collage day
[14:13:30] <RoanKattouw_away>	 hmm
[14:13:39] <RoanKattouw_away>	 So it sounds like Tuesday would work best then
[14:15:40] <RoanKattouw>	 Hmm
[14:16:01] <RoanKattouw>	 Krinkle: What time would you be up until on Wednesday, given that you have a class the next morning?
[14:16:47] <Krinkle>	 probably midnight
[14:17:03] <RoanKattouw>	 OK
[14:17:10] <Krinkle>	 usually a little later, but that should be free to wrap things up, not as part of scheduled time
[14:17:18] <RoanKattouw>	 Yeah
[14:17:51] <RoanKattouw>	 Alright, so let's schedule Wednesdays for RL2 work then
[14:18:16] <Krinkle>	 Just FYI, I nodded mostly during the reading of your last 2 mails to Erik/Terry
[14:18:29] <Krinkle>	 sounds good
[14:18:30] <RoanKattouw>	 I'll try to come in early on those days (shouldn't be a problem next week thanks to jetlag -- after that we'll see, I'm used to coming in at 10 so 9 is already early)
[14:18:52] <RoanKattouw>	 And then you sign off around 3pm PDT
[14:19:26] <Krinkle>	 what time is 9am here? noon? 1pm ?
[14:19:33] <Krinkle>	 no, can't be
[14:19:33] <RoanKattouw>	 Oh, sorry
[14:19:39] <RoanKattouw>	 9am PDT = 6pm CEST
[14:19:42] <Krinkle>	 right
[14:19:54] <RoanKattouw>	 Ideally I'd come in at 8 or 7
[14:20:04] <RoanKattouw>	 Well, 7 is probably not realistic :D
[14:20:27] <Krinkle>	 If we keep the task management, I'll be able to work a few hours before, and you'll be able to work a few hours after
[14:20:38] <RoanKattouw>	 Yeah, that should be fine
[14:21:10] <Krinkle>	 RoanKattouw: Remember that Smarterer thing I sent a few weeks ago?
[14:21:19] <RoanKattouw>	 No?
[14:21:26] <Krinkle>	 http://smarterer.com/
[14:21:48] <Krinkle>	 testing knowledge of stuff (js, mediawiki, php, 80s music, whatever)
[14:21:53] <RoanKattouw>	 Oh that
[14:22:00] <RoanKattouw>	 You showed me that a few months ago yeah
[14:22:17] <Krinkle>	 well, I've been given the sysop-flag for JavaScript test (well they call it "Professor")
[14:22:24] <Krinkle>	 just now
[14:22:26] <RoanKattouw>	 Sweet
[14:22:42] <Krinkle>	 which means I can't do the test anymore :P
[14:23:43] <Krinkle>	 they are still working on it, but their score calculator is pretty sweet. New questions are added all the time and you need to keep up every now and then to proof you weren't lucky. So it gets better over time, but due to me being unable to enter the test anymore (which makes sense, since I can peek the answers)
[14:23:48] <Krinkle>	 my score has dropped to Newbie
[14:24:16] <RoanKattouw>	 So now you're a newbie who's also a professor
[14:24:29] <Krinkle>	 yeah
[14:24:30] <Krinkle>	 xD
[14:24:36] <Krinkle>	 but the latter isn't visible anywhere yet
[14:26:48] <hexmode>	 before i commit: are we naming branches YYYY/bug/# or just bug/# ?
[14:27:00] <Reedy>	 however you want
[14:27:04] <hexmode>	 k
[14:27:04] <Reedy>	 i don't think anyone cares
[14:27:23] <RoanKattouw>	 hexmode: git-review will automatically set the topic to bug/12345 if there is a bug number in your commit msg, unless you manually override that
[14:27:36] <hexmode>	 just didn't want to get slapped
[14:27:40] <RoanKattouw>	 Krinkle: OK, scheduled
[14:27:47] <hexmode>	 RoanKattouw: a, k thanks
[14:27:50] <RoanKattouw>	 hexmode: We haven't decided on a convention yet so it's a free for all now
[14:28:51] <Krinkle>	 hexmode: some have vouched for YYYY, but I think for bugs the year is irrelevant and only  harder to search for
[14:28:58] <Krinkle>	 for non-bugs it may make sense
[14:29:10] <Krinkle>	 so that it's easy to find past things related to the same bug
[14:35:57] <Krinkle>	 RoanKattouw: do you know how to poke if I need a lists.wikimedia.org list admin password re-sent? I can't figure out a way to reset it
[14:36:25] <Krinkle>	 CVN list was created last year by cbrown who forwarded a mail containg a random password, but I lost that mail.
[14:36:51] <RoanKattouw>	 I think you need to poke ops
[14:36:53] <Krinkle>	 I figured there would be a way from the online thing, but don't see any
[14:37:07] <RoanKattouw>	 No, I don't think there is
[14:37:10] <JeroenDeDauw>	 drecodeam: hey
[14:42:42] <drecodeam>	 JeroenDeDauw: hi, the updating script worked for me
[14:42:59] <JeroenDeDauw>	 drecodeam: you'd hope so :)
[14:43:12] <JeroenDeDauw>	 Else the extension would have been broken pretty badly :p
[14:43:27] <drecodeam>	 :P.
[14:44:26] <drecodeam>	 the other day i was asking about your opinions over the work that i intend to do
[14:45:27] <drecodeam>	 about implementing flickr and geolocation support in UploadWizard
[14:46:32] <hexmode>	 w00! my first real git commit ;)
[14:48:57] <sumanah>	 yay!
[14:49:29] <JeroenDeDauw>	 drecodeam: there already is some geolocation support in it
[14:49:41] <JeroenDeDauw>	 I added inputs for lat lon and alt fields a while back
[14:49:53] <JeroenDeDauw>	 And there is some stuff that in some cases automatically populates those
[14:50:01] <JeroenDeDauw>	 Not sure how that works though
[14:50:50] <JeroenDeDauw>	 drecodeam: flickr integration seems cool, but I'm not sure how useful this is for the foundation. Have you looked into that?
[14:51:16] <JeroenDeDauw>	 drecodeam: I'm thinking that if you go add support for uploading to remote services, then you might want to start with uploading to remote MediaWikis
[14:51:26] <JeroenDeDauw>	 That'd allow for uploading to commons from your wiki
[14:51:36] <JeroenDeDauw>	 Making the instant commons feature all that nicer
[14:51:46] <drecodeam>	 Ya it is there, but i was reading the design documents for the geolocation part, i would be implementing a way to use a map interface by connecting it to google maps and openlayers
[14:52:22] <drecodeam>	 JeroenDeDauw: thats a nice suggestion, i would look into that. I think that would be more useful as of now than the flickr integration.
[14:53:17] <drecodeam>	 also flickr integration was the idea suggested by kaldari, so i was just going in by his suggestion
[14:56:08] <drecodeam>	 actually looking more into it, kaldari told me that mediawiki is planning to have another WikiLovesMonuments campaign
[14:56:26] <drecodeam>	 i think that flickr and geolocation would be very useful for that
[15:04:16] <drecodeam>	 JeroenDeDauw: the integration would not be to upload images to flickr, it would be to automatically transferring image, image's data and license from flickr
[15:05:11] <JeroenDeDauw>	 drecodeam: ok, an "import" feature would be cool :)
[15:06:26] <JeroenDeDauw>	 drecodeam: Obviously I can't tell you much about this feature, but if you have specific questions, I'd be happy to help :)
[15:06:44] <drecodeam>	 one more thing, i was just seeing the commons API. Does it support uploading media also ? as i can only find docs to retrieve information about media
[15:07:39] <sumanah>	 drecodeam: "mediawiki is planning to have another WikiLovesMonuments campaign" that is worded strangely.  MediaWiki is the software.  Wikimedia is the movement
[15:07:53] <sumanah>	 !upload
[15:07:54] <mw-bot>	 File uploads are disabled by default, set $wgEnableUploads=true; to enable it. See <https://www.mediawiki.org/wiki/Manual:Configuring_file_uploads> for configuration details, <https://www.mediawiki.org/wiki/Help:Managing_files> and <https://www.mediawiki.org/wiki/Help:Images> for using uploaded images and other files.
[15:08:13] <sumanah>	 drecodeam: https://www.mediawiki.org/wiki/API:Upload
[15:08:14] <drecodeam>	 sumanah: sorry, my mistake !
[15:08:27] <sumanah>	 drecodeam: I see you didn't read the API docs terribly carefully :)
[15:09:29] <drecodeam>	 sumanah: no i have read these. These are used in the UploadWizard. I though that wikimedia commons had a different API
[15:28:38] <drecodeam>	 sumanah: found it now. The pages i was reading were the further pieces of information that were made available. Silly from my part, as i missed reading that it inherits the MediaWiki API
[15:35:56] <Krinkle>	 RoanKattouw_away: I just found out on fy.wikipedia NS_SPECIAL is localized as "Wiki"
[15:36:02] <Krinkle>	 interesting
[15:36:33] <RoanKattouw>	 Really
[15:36:35] <RoanKattouw>	 That's strange :)
[15:36:52] <Krinkle>	 Wiki:Wat_is_hjirmei_keppele/Berjocht:Commons (Special:WhatLinksHere/Template:NowCommons)
[15:37:22] <RoanKattouw>	 Wait, Template is Berjocht?
[15:37:27] <Krinkle>	 yeah..
[15:37:30] <RoanKattouw>	 That's a strange choice
[15:37:34] <Krinkle>	 indeed
[15:37:39] <Krinkle>	 although historically correct
[15:37:50] <Krinkle>	 in a way
[15:38:00] <RoanKattouw>	 It's probably the most common use of templates for them
[15:38:14] <Krinkle>	 it should be the the only use of templats
[15:38:44] <Krinkle>	 if you eliminate infoboxes and programming structures into parser functions
[15:38:47] <RoanKattouw>	 Ahm, infoboxes?
[15:38:49] <Krinkle>	 the templates left are messages
[15:38:49] <RoanKattouw>	 Yeah
[15:39:02] <RoanKattouw>	 Well, and navigation templates
[15:39:06] <Krinkle>	 True
[15:39:32] <Krinkle>	 although we can hope for those to be automated some day as well (based on categories)
[15:40:25] <Krinkle>	 I got to this because I just did my monthly cross-wiki watchlist (which takes 20 minutes to generate on my server)
[15:41:34] <Krinkle>	 api request to meta, take cookie, then manipulate the cookiebin and do what centralauth does post-login (but including *.wikimedia.org) and then about 1400 (2x700) api requests, and then a bunch of php to merge them together into a synchronous time line
[15:41:54] <Krinkle>	 and then echo it :P
[15:42:47] <Krinkle>	 (I believe I showed it to you once on a hackathon)
[15:43:19] <RoanKattouw>	 Yeah
[15:44:23] <Krinkle>	 RoanKattouw: btw, what is your expectation on CORS happening on WMF? Any estimate? It's really blocking a lot of things
[15:44:57] <Krinkle>	 including some things Huggle wants to do from a gadget (the only alternative is requiring user credentials be entered on a stand-alone server to make the api request for them and send it through JSONP)
[15:45:36] <RoanKattouw>	 Hmm
[15:46:08] <Krinkle>	 the origin whitelist thingy, it is blocked on a cache issue?
[15:46:30] <Krinkle>	 !b 20814
[15:46:30] <mw-bot>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=20814
[15:46:36] <Krinkle>	 !b 30881
[15:46:36] <mw-bot>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=30881
[15:54:16] <Nikerabbit>	 don't we already have CORS for webfonts?
[15:55:25] <Krinkle>	 I don't know, but how does that help?
[15:55:34] <Krinkle>	 upload.wikimedia.org is all static
[15:55:49] <RoanKattouw>	 WebFonts doesn't need CORS
[15:56:05] <RoanKattouw>	 upload.wm.o should have CORS but doesn't yet
[15:56:09] <RoanKattouw>	 API needs CORS but that needs coding
[15:56:14] <RoanKattouw>	 Which I need to JFDI
[15:56:20] <Krinkle>	 that's what I meant, API
[15:56:33] <Krinkle>	 (above bug ids)
[15:58:04] <RoanKattouw>	 It needs a couple of hours of coding
[15:58:53] <RoanKattouw>	 Maybe I'll do it on a train or plane ride
[16:31:30] <RoanKattouw>	 hashar: Could you review&merge https://gerrit.wikimedia.org/r/#change,3863 please?
[16:33:09] <hashar>	 I am on 3824 right now
[16:34:10] <hashar>	 RoanKattouw: hmm I do not know anything about ClickTracking.  Isn't Arthur around ?
[16:34:48] <RoanKattouw>	 I don't think Arthur knows anything about it either
[16:34:55] <RoanKattouw>	 But that shouldn't matter, it's a two-line fix for an earlier change
[16:37:56] <JeroenDeDauw>	 sumanah: "that is worded strangely.  MediaWiki is the software.  Wikimedia is the movement" Heh, who has been doing that for SMW lately? :D
[16:38:35] <sumanah>	 JeroenDeDauw: :) doing semantics about MediaWiki, you mean?
[16:39:00] <JeroenDeDauw>	 ahhh!
[16:39:38] <sumanah>	 JeroenDeDauw: :-) I might not have gotten your initial witticism, actually
[16:39:43] <sumanah>	 if that's not what you meant :)
[16:39:55] <RoanKattouw>	 sumanah: Remember when a few days ago you asked "is SMW planning to do X" ?
[16:40:07] <sumanah>	 AHAAH Oh right
[16:40:28] <sumanah>	 ok, I could get all pedantic about how what I did was TOTALLY FINE and DIFFERENT but I think it's better for me to take my lumps :-)
[16:40:41] <RoanKattouw>	 hehe
[16:41:02] * RoanKattouw  pats Ms Special Snowflake on the back

[16:41:05] <sumanah>	 the answer is probably for us to insert some AI into MW and into SMW so they can indeed plan to do things
[16:41:26] <sumanah>	 the head would be more patronizing :D
[16:42:56] <RoanKattouw>	 I guess that's what I was looking for
[16:43:39] <RoanKattouw>	 OMG now I'm having a visual of Domas patting Sumana on the head in a patronizing way
[16:43:53] <RoanKattouw>	 He's at least a foot taller than you :)
[16:44:24] <sumanah>	 RoanKattouw: is this in your visual catalog of "hypothetical things that would lead to raised eyebrows"?
[16:44:52] <RoanKattouw>	 Of course
[16:45:03] <RoanKattouw>	 The "wouldn't it be weird if ... " catalog
[16:46:14] <sumanah>	 at least
[16:46:14] <sumanah>	 but to me most people are just "tall"
[16:46:16] <sumanah>	 you know, to clumsy people like me, most people are graceful
[16:46:16] <sumanah>	 to short people, everyone else is tall
[16:47:09] <RoanKattouw>	 Domas makes me feel like a dwarf
[16:47:54] <RoanKattouw>	 Although maybe that's just because I'm used to being able to see over people's heads when standing on my toes
[16:54:39] <Nemo_bis>	 maybe also because he's twice as big as you
[16:54:52] <Nemo_bis>	 if not more
[16:55:45] <sumanah>	 btw Nemo_bis thanks for adding those Gerrit bugs into Bugzilla
[16:56:10] <Nemo_bis>	 thank you to those who look at them :)
[16:56:24] <Nemo_bis>	 RoanKattouw should be able to add some info to some of them for instance
[16:58:18] <RoanKattouw>	 rsterbin: Could you verify https://www.mediawiki.org/wiki/Special:Code/MediaWiki/114557 is complete ?
[16:58:33] <RoanKattouw>	 (There's no reason it wouldn't be, I'm just paranoid)
[16:58:36] <rsterbin>	 looking now...
[17:00:51] <Nikerabbit>	 wooof
[17:00:54] <Nikerabbit>	 does it start now?
[17:01:19] <RoanKattouw>	 Didn't they say 19:00 UTC?
[17:01:22] <RoanKattouw>	 It is now 17:01 UTC
[17:01:53] <Nikerabbit>	 RoanKattouw: 20% checkin
[17:01:58] <RoanKattouw>	 Oh
[17:02:05] <Nikerabbit>	 ah it's 15 over
[17:02:08] <rsterbin>	 RoanKattouw: I think you already deployed https://gerrit.wikimedia.org/r/#change,3501
[17:02:14] <rsterbin>	 but everything else looks perfect
[17:02:22] <RoanKattouw>	 Sweet, thanks
[17:03:10] <Krinkle>	 Can someone poke Trevor to join IRC?
[17:03:21] <fabriceflorin>	 Howdy, RoanKattouw and rsterbin. Ready  to help in the next couple hours with the deployment of bug fixes for these new AFT5 features:
[17:03:21] <fabriceflorin>	 * Feedback links (A & E with X box)
[17:03:22] <fabriceflorin>	 * Bucketing fixes
[17:03:22] <fabriceflorin>	 * Clicktracking fixes + AFT4 cookie change
[17:03:44] <RoanKattouw>	 I merged the code from git to 1.19wmf1 and Reha just verified I did it correctly
[17:04:02] <RoanKattouw>	 Krinkle: Here he is
[17:04:18] <TrevorParscal>	 I'm on Trevor Standard Time
[17:04:36] <TrevorParscal>	 it's slightly shifted west from PST
[17:04:41] <RoanKattouw>	 So I'm gonna deploy this code to test now
[17:04:51] <TrevorParscal>	 I'm actually early in TST
[17:05:05] <TrevorParscal>	 hello Krinkle
[17:05:46] <rsterbin>	 fabriceflorin: features look good (except for "AFT4 cookie change", which i assume someone else did -- it's not in my changes)
[17:06:14] <RoanKattouw>	 Yeah the AFT4 changes are next on my list, and OmniTI isn't responsible for them
[17:06:27] <fabriceflorin>	 Cool, sounds good, RoanKattouw and rsterbin. Look forward to testing this on testing shortly.
[17:06:31] <Krinkle>	 brb to bathroom
[17:07:02] <rsterbin>	 do we know what to expect re: caching?
[17:07:43] <Krinkle>	 TrevorParscal: Timo Sebastiaan Tijhof ?
[17:07:48] <rsterbin>	 hey, elizabeth
[17:07:55] <rsterbin>	 roan just started deploying to test
[17:07:55] <auroraeosrose>	 hiya
[17:07:56] <fabriceflorin>	 Also, this may be a good time to switch the oversight email address from the test aft@wikimedia.org address to the correct address for oversight requests.
[17:08:01] <RoanKattouw>	 In non-debug mode the changes should show up within 10 minutes of me announcing I've deployed them to testwiki
[17:08:04] <RoanKattouw>	 Same for enwiki
[17:08:24] <rsterbin>	 cool
[17:08:24] <RoanKattouw>	 In debug mode you may experience caching issues, but regular users aren't affected by that
[17:08:25] <TrevorParscal>	 Krinkle and RoanKattouw #mw-bug-34885
[17:08:28] <fabriceflorin>	 Hi aurorarose and yoni_omniti.
[17:08:30] <RoanKattouw>	 Krinkle: Trevor Standard Time
[17:08:51] <Krinkle>	 haha
[17:09:38] <RoanKattouw>	 fabriceflorin, rsterbin, auroraeosrose: Changes now live on testwiki. Caching issues may occur but should be resolved within 5-10 mins
[17:09:45] <rsterbin>	 thx
[17:10:07] <auroraeosrose>	 thanks ;)
[17:10:19] <yoni_omniti>	 thanks
[17:11:37] <DarTar>	 morning everybody
[17:12:06] <fabriceflorin>	 Thanks, RoanKattouw. We will test these features on test.wikimedia.org: 1) new feedback links A and E; 2) bucketing for A, E and X; 3) Clicktracking for all above events, plus changes requested by Dario. Yoni_omniti and rsterbin, what else should we test?
[17:12:46] <yoni_omniti>	 one minute...
[17:13:06] <rsterbin>	 when the comment box gets a scrollbar, it shouldn't look quite as bad -- i adjusted the padding down to about half of what it was
[17:13:22] <sumanah>	 hi DarTar
[17:13:26] <DarTar>	 hey sumanah
[17:13:27] <yoni_omniti>	 thats about it
[17:14:02] <rsterbin>	 i don't think i have any way of testing clicktracking on test -- who's looking at that?
[17:14:15] <DarTar>	 I don't have access either
[17:14:21] <rsterbin>	 (or bucketing, for that matter)
[17:14:57] <yoni_omniti>	 RoanKattow: any advice on how could we test click tracking on testwiki?
[17:15:16] <RoanKattouw>	 Hmm, doesn't this log to the normal place?
[17:15:18] * RoanKattouw  looks

[17:15:34] <DarTar>	 with testwiki as a project id?
[17:15:37] <RoanKattouw>	 DarTar: This should just be going to emery
[17:15:39] <RoanKattouw>	 Yeah
[17:15:44] <DarTar>	 k let me look it up
[17:15:46] <RoanKattouw>	                 $wgClickTrackingLog = 'udp://208.80.152.184:8421/' . $wgDBname; // 208.80.152.184 = emery
[17:16:30] <DarTar>	 yes we do have data coming through from testwiki
[17:16:54] <DarTar>	 now looking into the specific issues I reported last week
[17:17:47] <sumanah>	 hexmode: so, it's just raindrift for today (unless MaxSem wants to get in on things -- he's listed as "monday-friday")
[17:17:48] <DarTar>	 I still see a ext.articleFeedbackv5@1-option2X-impression-bottom meaning that someone is seeing the AFT Option2
[17:17:53] <Nikerabbit>	 does it start now?
[17:18:13] <DarTar>	 rsterbin: did we decide to disable entirely O2 and O3?
[17:18:23] <hexmode>	 Nikerabbit: "it"?
[17:18:46] <RoanKattouw>	 Krinkle: Could you approve https://gerrit.wikimedia.org/r/3864 for me please?
[17:18:51] <rsterbin>	 DarTar: they're still available if you manipulate the url -- which is what was in my browser history
[17:19:00] <DarTar>	 ic
[17:19:16] <DarTar>	 so that explains that hybrid log line
[17:19:23] <Nikerabbit>	 hexmode: 20% checkin
[17:19:38] <DarTar>	 -option1X is now correctly captured
[17:19:44] <DarTar>	 trying a CTA
[17:19:51] <hexmode>	 Nikerabbit: yes
[17:20:30] <hexmode>	 hrm... Nikerabbit you seem pretty busy w/ git ... need something else?
[17:21:08] <Nikerabbit>	 hexmode: actually I'm traveling tomorrow so I don't need anything :)
[17:21:08] <MaxSem>	 sumanah, later maybe
[17:21:36] <Nikerabbit>	 just wanting to know what's the focus point
[17:21:50] <rsterbin>	 RoanKattouw: the x image on Option E is out of date -- how long would that be cached?
[17:22:03] <DarTar>	 rsterbin: we are now removing entirely edit_attempts and edit_success for blank edits, correct?
[17:22:22] <rsterbin>	 DarTar: no, that's not been changed
[17:22:31] <RoanKattouw>	 rsterbin: Depends on whether you are using an @embed or an <img> tag
[17:22:38] <rsterbin>	 it's an @embed
[17:22:45] <RoanKattouw>	 That should have cleared by now
[17:22:47] <RoanKattouw>	 Oh wait
[17:22:49] <DarTar>	 resterbin that's strange, I just submitted a CTA edit and nothing came thorough in the log
[17:22:51] <rsterbin>	 er, whatever it was doing before, it's doing now
[17:22:52] <RoanKattouw>	 It wasn't in my changes
[17:23:06] <RoanKattouw>	 I used git diff --patch | patch  , so obviously the binary file changes didn't come through
[17:23:22] <hexmode>	 hrm... Bugzilla's api could be better.  There is no straightforward way to search for new attachments/patches
[17:23:27] <hexmode>	 But I can hack it
[17:24:01] <rsterbin>	 DarTar: i misspoke.  i haven't made any changes to it, but it does not record edit_attempt or edit_success when the edit is blank.  the change would have been to aft4.
[17:24:01] <DarTar>	 good grief, I cannot save edits on prototype: database error
[17:24:17] <rsterbin>	 yeah, the jobs table is missing there for some reason
[17:24:24] <Krinkle>	 RoanKattouw: I can't, don't have rights in AFT
[17:24:50] <RoanKattouw>	 Hmph
[17:24:52] <DarTar>	 rsterbin, the way it worked so far for blank edits in AFT5 is: log edit attempt and log an edit success with 0 as a rev_id
[17:24:53] <RoanKattouw>	 Could you +1 it then?
[17:25:03] <RoanKattouw>	 If you approve of the change, I feel less bad about +2ing my own code
[17:25:13] <Krinkle>	 gotcha
[17:25:56] <hexmode>	 RoanKattouw: Krinkle: could you guys check my tiny js commit?
[17:26:02] <RoanKattouw>	 hexmode: Link?
[17:26:02] * sumanah  goes to meeting

[17:26:06] <rsterbin>	 DarTar: i haven't made any changes.  AFT5 doesn't test anything, it just logs every time the hook hits.  if something's different, it's a core change, not an AFT5 one.
[17:26:40] <hexmode>	 RoanKattouw: https://gerrit.wikimedia.org/r/#change,3860
[17:26:41] <DarTar>	 yeah, that makes sense, I was just describing how things work so far for AFT5 blank edits
[17:27:05] <rsterbin>	 are you seeing the right stuff in the log for test?
[17:27:12] <Krinkle>	 nice, Gerrit allows changing the diff parameters
[17:27:13] <DarTar>	 rsterbin, fabriceflorin: I cannot test edit attempt/cpmplete events on prototype because I cannot complete an edit
[17:27:14] <Krinkle>	 (whitespace etc.)
[17:27:43] <rsterbin>	 DarTar: are you sure prototype is what you want to be testing right now?
[17:28:04] <rsterbin>	 that's running off the omniti branch, not the master
[17:28:07] <RoanKattouw>	 DarTar: Please test on test.wikipedia.org
[17:28:08] <Krinkle>	 hm.. doesn't seem to work
[17:28:18] <DarTar>	 ah stupid me, I was logged on prototype
[17:28:31] <DarTar>	 ok, CTA test take 2
[17:28:50] <RoanKattouw>	 hexmode: Done
[17:29:01] <hexmode>	 Krinkle: real diffs are colored differently than w/s only
[17:29:15] <hexmode>	 but yea, looking at my own, it fails
[17:29:19] <fabriceflorin>	 Hi guys, I have been testing the new feedback links A, E and X on testing and everything is generally working well. However, I am getting a weird graphic artifact on the feedback button E on Firefox, where a small 'X' graphic appears to overlap another, larger 'X' graphic, at this URL: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?bucket=1&debug=true&aftv5_link=E (I will send you a screenshot via email)
[17:29:29] <Krinkle>	 hexmode: when I change the "Ignore whitespace" drop down and {update}, nothing changes
[17:29:37] <RoanKattouw>	 fabriceflorin: I'm working on the X image
[17:29:40] <rsterbin>	 fabriceflorin: that's the new-image issue.  Roan's looking into it
[17:29:41] <hexmode>	 Krinkle: same
[17:29:47] <rsterbin>	 lol, jinx
[17:30:10] <RoanKattouw>	 I just updated the image on the server, may take up to 10 minutes to change on your end
[17:30:20] <rsterbin>	 thanks
[17:30:33] <fabriceflorin>	 Thanks, RoanKattouw and rsterbin, much appreciated. Please let me know what else I could be testing in the meantime, besides options A, E and X.
[17:30:55] <rsterbin>	 there really wasn't very much in this push :)
[17:31:04] <DarTar>	 ok AFT5 CTA edit_attempt / edit_success works as expected for blank edits
[17:31:06] <rsterbin>	 once Dario
[17:31:16] <rsterbin>	 's confirmed clicktracking is good, i think we're done
[17:31:30] <rsterbin>	 (stupid apostrophe right next to enter)
[17:31:55] <yoni_omniti>	 fabriceflorin: it looks like a "dashed border"?
[17:32:01] <fabriceflorin>	 Thanks, DarTar. Let us know if everything else in this deployment works for you.
[17:33:30] <DarTar>	 RoanKattouw: just to double check, did you change the AFT4 code to actually not log edit_success events for blank edits at all? Or is it still matching what AFT5 does, namely edit_attempt event are logged with parent rev_id and edit_success are also logged but with a rev_id = 0
[17:33:56] <RoanKattouw>	 DarTar: I'm doing that right now
[17:33:58] <DarTar>	 if it's the latter, we're good
[17:34:00] <DarTar>	 ah
[17:34:05] <RoanKattouw>	 Oh, wait
[17:34:13] <RoanKattouw>	 Yeah I didn't change it in that regard at all
[17:34:20] <RoanKattouw>	 I'm in the process of applying my earlier changes on test.wp.o
[17:34:29] <RoanKattouw>	 They bring AFT4 in line with AFT5
[17:34:46] <DarTar>	 ok - I actually don't care how we treat them as long as AFT4 and AFT5 have consistent logs
[17:35:23] <DarTar>	 I should be able to test if this is the case by making an AFT4 blank edit on testwiki
[17:35:52] <RoanKattouw>	 HAng on
[17:35:57] <RoanKattouw>	 I haven't deployed to testwiki just yet
[17:36:10] <DarTar>	 hmm no I cannot (AFT4 is not enabled on testwiki)
[17:36:14] <DarTar>	 ok
[17:36:27] <RoanKattouw>	 It's not?
[17:37:17] <DarTar>	 not globally, maybe there's a specific category where it's enabled?
[17:37:30] <DarTar>	 can anyone give me a link with an AFT4 article on testwiki?
[17:37:43] <RoanKattouw>	 DarTar: [[Category:Article Feedback Pilot']]
[17:37:47] <RoanKattouw>	 Sans the quote
[17:37:51] <fabriceflorin>	 yoni_omniti, I just sent you, rsterbin, auroraerose and RoanKattouw a screenshot of the problem with Option E graphic. I can still reproduce on Firefox on testing, at this URL: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?bucket=1&debug=true&aftv5_link=E  Do I need to clear my cache?
[17:37:53] <DarTar>	 good
[17:37:54] <RoanKattouw>	 Also, [[Category:Really big category]]
[17:38:09] <RoanKattouw>	 fabriceflorin: Try Shift+Refresh button
[17:38:11] <RoanKattouw>	 ?
[17:38:31] <rsterbin>	 fabriceflorin: the image hasn't cleared for me yet.  i'm periodically refreshing.
[17:38:41] <yoni_omniti>	 fabriceflorin: its reha just confirmed that its an artifact from caching
[17:38:50] <yoni_omniti>	 could be cached on a proxy though
[17:38:53] <rsterbin>	 the image size will change from 12x12 to 16x16 when it clears.
[17:39:42] <rsterbin>	 ha, there it goes
[17:39:48] <RoanKattouw>	 hah
[17:39:54] <RoanKattouw>	 I guess I did that then
[17:40:00] <RoanKattouw>	 All hail touch(1)
[17:40:01] <DarTar>	 ok tested an AFT4 CTA edit, events are logged as expected. Waiting for RoanKattouw to push the changes (additional data appended to the log) to see what I get
[17:40:06] <yoni_omniti>	 cleared for me as well
[17:40:08] <fabriceflorin>	 OK, I tried yoni_omniti's suggestion of using Shift_Refresh and it worked for me. The graphic artifact on Option E is gone now. Yay!
[17:40:21] <RoanKattouw>	 DarTar: AFT4 changes applied
[17:40:24] <yoni_omniti>	 Roan's suggestion actually :)
[17:40:39] <DarTar>	 RoanKattouw: cool, off to test again
[17:40:40] <rsterbin>	 hm, i'm seeing the 12x12 tab_x_on.png still
[17:41:30] <yoni_omniti>	 rsterbin: the rollover cleared for me now
[17:41:32] <rsterbin>	 nm, got the new one
[17:41:59] <rsterbin>	 yay
[17:42:04] <DarTar>	 hmm, no CTA served upon submission of a rating - RoanKattouw I guess this is because we haven't yanked the AFT4 CTA throttling cookie yet, right?
[17:42:05] <fabriceflorin>	 DarTar, I think if you create a new article on test, it should display AFT4 behavior, but I cannot guarantee that. It appears that testing doesn't have copies of English Encyclopedia articles, unless I am misunderstanding that environment.
[17:42:15] <RoanKattouw>	 DarTar: I did make that change
[17:42:23] <RoanKattouw>	 DarTar: But you might still have a residual cookie
[17:42:28] <fabriceflorin>	 I had to create the Golden Sparrow article on testing in order to test it.
[17:42:40] <DarTar>	 fabriceflorin: it actually looks like AFT4 is enabled by a whitelist category on test
[17:42:51] <DarTar>	 RoanKattouw: ok
[17:44:18] <DarTar>	 hmm, cleared cookies, still don't see a CTA
[17:44:30] * RoanKattouw  tries

[17:44:55] <DarTar>	 trying to empty cache + all other cookies
[17:46:00] <RoanKattouw>	 DarTar: I get a CTA just fine
[17:46:17] <DarTar>	 even if you try again?
[17:46:27] <RoanKattouw>	 Yes
[17:46:32] <RoanKattouw>	 I removed the code that sets the cookie
[17:46:38] <RoanKattouw>	 However, I think any existing cookies may still be honored
[17:46:52] <DarTar>	 yes, works for me now
[17:47:00] * DarTar  checking the logs

[17:49:37] <fabriceflorin>	 BTW, I also checked the feedback page and it still seems to work as intended on testing. I also checked the edit flow and it displays this code when I make an edit from the CTA after posting feedback:
[17:49:43] <fabriceflorin>	 BTW, I also checked the feedback page and it still seems to work as intended on testing. I also checked the edit flow and it displays this code when I make an edit from the CTA after posting feedback:
[17:50:18] <DarTar>	 ok AFT4 CTAs work as expected, the additional data (article_title|rev_id) is still not appended to the log line, this has still to be pushed?
[17:50:19] <fabriceflorin>	 BTW, I also checked the feedback page and it still seems to work as intended on testing. I also checked the edit flow and it displays this long string in the URL when I make an edit from the CTA after posting feedback - I assume this is OK, based on Dario's earlier comments.
[17:50:21] <fabriceflorin>	 http://test.wikipedia.org/w/index.php?title=Golden-crowned_Sparrow&action=edit&articleFeedbackv5_click_tracking=1&articleFeedbackv5_ct_token=IgytJsrQpF8YnAYEiufL7KbGGZ2mqLro8&articleFeedbackv5_bucket_id=1&articleFeedbackv5_cta_id=1&articleFeedbackv5_f_link_id=A&articleFeedbackv5_link_id=A&articleFeedbackv5_location=overlay
[17:50:44] <RoanKattouw>	 DarTar: Additional data should appear AFAIK
[17:50:52] <DarTar>	 fabriceflorin: yes AFT5 works as expected
[17:51:25] <RoanKattouw>	 Oooh wait a second
[17:52:10] <RoanKattouw>	 No, that's correct
[17:52:12] <RoanKattouw>	 hm
[17:52:23] <RoanKattouw>	 So what are you seeing? You are getting the events but without the additional info?
[17:52:35] <fabriceflorin>	 DarTar: good, so that leaves AFT4, then. Let us know when you are done with that.
[17:53:01] <DarTar>	 RoanKattouw: just the old logs, no additional data
[17:53:33] <DarTar>	 e.g.
[17:53:34] <DarTar>	 testwiki ext.articleFeedback@10-pitch-edit-save-save-complete	20120328174550	1	U8lDI08B351uU5vYG0ZrllMUlStD30Qf3	0	0	2	2	2
[17:53:47] <DarTar>	 well, including the tabs ;)
[17:53:54] <RoanKattouw>	 Try now?
[17:54:00] <DarTar>	 k
[17:54:50] <DarTar>	 ouch
[17:55:30] <DarTar>	 I removed all cookies but I am still unable to get a CTA when submitting a rating in AFT4
[17:56:01] <DarTar>	 just getting a confirmation that the ratings have been submitted
[17:56:16] <RoanKattouw>	 Strange
[17:56:24] <RoanKattouw>	 Are you logged in and allowed to edit the page you're looking at?
[17:56:48] <DarTar>	 yes and yes
[17:56:52] <DarTar>	 trying http://test.wikipedia.org/wiki/Pela_Seri
[17:57:57] <DarTar>	 fabriceflorin: can you help me reproduce this?
[17:58:00] <fabriceflorin>	 RoanKattouw, when you are done with AFT4, I would like to ask your advice about the oversight request emails, which are now going to aft@wikimedia.org. Our experience is that these emails have a long lag time, which can take as much as an hour to get to this email address (I just did this 48 mins. ago, and the email still hasn't arrived). Ironholds suggested that this may be related to the fact that the messaging action is not in the mediawi
[17:58:00] <fabriceflorin>	 message queue, which means that sending is not prioritized by the servers. Is there anything that could easily be done to address this?
[17:58:31] <RoanKattouw>	 fabriceflorin: I have a fix for this, but to apply it I need to restart the job runners
[17:58:39] <fabriceflorin>	 Hi DarTar, I am happy to try to reproduce this. Give me a few minutes, and I will report back.
[17:58:40] <DarTar>	 1) go to an AFT4 page on testwiki 2) rate the page 3) click through (edit attempt) 4) go to another AFT4 page 5) rate that page: do you get an edit CTA?
[17:58:45] <RoanKattouw>	 And I discovered that the job runners are in a state where they wouldn't be able to recover from a restart
[17:58:57] <DarTar>	 fabriceflorin: thx
[17:58:58] <RoanKattouw>	 I submitted a fix for this but it hasn't been applied yet AFAIK
[17:59:20] <fabriceflorin>	 Thanks, DarTar. I will go to Pela Seri page on Firefox, Chrome and Safari and let you know what I find out.
[17:59:37] <DarTar>	 make sure you're logged in too
[17:59:45] <DarTar>	 and clear your cookies/cache before you start
[17:59:54] <DarTar>	 the other way round actually :)
[18:00:01] <DarTar>	 clear cookies, log in, start
[18:03:49] * DarTar  trying to reproduce cookie problem on a different browser

[18:12:14] <fabriceflorin>	 HI DarTar, I'm sorry I had to do a short standup with the NPT team. I'm back at testing the use case you gave me.
[18:12:22] <DarTar>	 np
[18:15:16] <fabriceflorin>	 Hi DarTar, I just did the test sequence you gave me on Safari, and I get a second edit CTA on the second page after rating the first one. I will past below the URL I got for the first edit, then the second URL I got for the second edit.
[18:15:18] <fabriceflorin>	 http://test.wikipedia.org/w/index.php?title=Pela_Seri&action=edit&clicktrackingsession=TdcEwlskKcKDnrUDYEWZhqnwjZrY9K2Pl&clicktrackingevent=ext.articleFeedback%4010-pitch-edit-save
[18:15:25] <fabriceflorin>	 http://test.wikipedia.org/w/index.php?title=Main_Page&action=edit&clicktrackingsession=TdcEwlskKcKDnrUDYEWZhqnwjZrY9K2Pl&clicktrackingevent=ext.articleFeedback%4010-pitch-edit-save
[18:16:00] <DarTar>	 yeah, I tried again from a different browser with a clean cache and no cookies and the throttling seems to be disabled
[18:16:18] <fabriceflorin>	 DarTar, I was logged in for the Safari test. Would you like me to do the same test on Firefox or Chrome? Would you like me to clear cookies first or not?
[18:17:27] <DarTar>	 please do, I'd like to have someone else to try and reproduce it, if it passes I am confident it was a temporary glitch because I am now getting CTAs every time I submit AFT4 ratings
[18:17:38] <DarTar>	 someone else = you :)
[18:18:02] <DarTar>	 RoanKattouw: the cookie issue seems to be fixed, still no additional info coming through tho
[18:18:14] <RoanKattouw>	 That's very strange
[18:18:28] <fabriceflorin>	 RoanKattouw, thanks for confirming that you have a fix for the delay in sending emails to oversighters. It's been more than an hour since I first requested oversight and I have still not received an email at aft@wikimedia.org, even though the request was immediately logged in the suppression log, as so: 09:09, 28 March 2012 Aft5hide (talk | contribs) requested oversight on the feedback Special:ArticleFeedbackv5/Barack Obama/69495(Offensive
[18:18:28] <fabriceflorin>	 language, f-word.)
[18:19:10] <fabriceflorin>	 DarTar, I'm on it! Hang on another minute. I will clear cookies first on both Firefox and Chrome, to be safe.
[18:20:58] <DarTar>	 RoanKattouw: I should be seeing the last fields formatted as a pipe separated value: article_title|rev_id on all AFT4 events, correct?
[18:21:10] <DarTar>	 s/fields/field
[18:22:41] <DarTar>	 the last 5 fields I see are numeric values, matching the old AFT4 log format (ns, lifetime edits, 6-month edits, 3-month edits, 1-month edits)
[18:26:32] <Krinkle>	 I don't think he read that
[18:29:15] <fabriceflorin>	 Hi DarTar. I have performed the test you gave me on Chrome (with my pre-existing afthide account) and Firefox (with a brand new account). I cleared the cache on both browsers prior to doing the test. For the Chrome test with an existing account, I got the edit CTA twice. For the Firefox test with a new account, I did not get the edit CTA the second time: all I got was a green message on the form saying "Saved successfully". If I change my
[18:29:15] <fabriceflorin>	 ratings and click again, I get the same green message, but no Edit CTA is displayed. I took a screenshot of that, in case you need it.
[18:30:39] <fabriceflorin>	 RoanKattouw, I see that we are approaching the end of our deployment window for AFT5. Should we plan to deploy these changes to production now, or do you want to deploy both AFT4 and AFT5 at the same time?
[18:31:17] <RoanKattouw>	 Oh, right
[18:31:18] <RoanKattouw>	 I'd almost forgotten about AFT, oops :S
[18:31:20] <RoanKattouw>	 I am just going to deploy to production now
[18:31:25] <RoanKattouw>	 Only half of Dario's AFT4 problems are fixed but I'll poke at that later
[18:31:31] <RoanKattouw>	 Sound good?
[18:31:49] <rsterbin>	 word
[18:32:00] <rsterbin>	 i, personally, would like to have lunch ;)
[18:32:01] <DarTar>	 fine with me, but I haven't actually looked at the other CTA issues, can I do this for 5 mins?
[18:32:23] <DarTar>	 I can review them real quick
[18:32:36] <DarTar>	 other issues as in AFT5-specific
[18:32:44] <rsterbin>	 ok
[18:32:54] <RoanKattouw>	 Oh wow, did we hold up your lunch for 2.5 hours?
[18:32:55] * RoanKattouw  fels bad

[18:33:10] <rsterbin>	 my own fault -- i forgot to eat earlier
[18:33:15] <fabriceflorin>	 OK, sounds good. Let's wait 5 mins, then plan to deploy AFT5. I'd like to release the OmniTI developers so they can have lunch and get back to work. ;o)
[18:33:23] <DarTar>	 same here - rsterbin I think we don't need you here
[18:33:50] <fabriceflorin>	 DarTar, we need rsterbin here in case something goes wrong with the deployment.
[18:33:52] <DarTar>	 only one thing before you take off: have we tested the actual data stored in the AFT5 table?
[18:34:14] <DarTar>	 fabriceflorin: right (sorry rsterbin)
[18:34:34] <DarTar>	 AFT5 table, meaning the af_link_id field
[18:34:37] <rsterbin>	 DarTar: i don't have access to the db on test
[18:34:50] <DarTar>	 I don't have access either
[18:35:12] <rsterbin>	 Roan deployed that fix last week, so it should be fine
[18:35:34] <fabriceflorin>	 Didn't we already test this on prototype? If so, I recommend we deploy to production and test it there.
[18:35:35] <DarTar>	 no as far as I know no fix has been deployed for this
[18:36:05] <DarTar>	 rsterbin: you mean to production?
[18:36:15] <rsterbin>	 that's what i thought he did
[18:36:21] <DarTar>	 RoanKattouw ?
[18:36:24] <rsterbin>	 perhaps i'm wrong
[18:36:48] <DarTar>	 I don't think we've had any data bugfix pushed to production since last week's deployment
[18:37:17] <RoanKattouw>	 I don't see it
[18:37:23] <RoanKattouw>	 Last push was Mar 21
[18:37:24] <DarTar>	 RoanKattouw: we're referring to the fact that all records where stored in the DB with the same af_link_id
[18:37:34] <RoanKattouw>	 yeah
[18:37:39] <Krinkle>	 RoanKattouw: k, I'm going to continue with continuous integration plan now, we'll continue RL2 next week?
[18:38:08] <RoanKattouw>	 Sure
[18:38:21] <RoanKattouw>	 "    Bug fix: setLinkId() still had a reference to the old linkId property; closeAsModal() updated to use 'X' rather than '0'"
[18:38:24] <RoanKattouw>	 That rev, right?
[18:38:30] <DarTar>	 RoanKattouw: can you check whether we're getting the correct af_link_id values in the testwiki DB?
[18:38:48] <RoanKattouw>	 Sure hang on
[18:39:06] <DarTar>	 rsterbin: what are the values for OptionA and OptionE ?
[18:39:46] <RoanKattouw>	 I see two zeros and a one in the past hour
[18:39:49] <rsterbin>	 A = 1; E = 5
[18:40:34] <rsterbin>	 just submitted feedback on test for option E
[18:40:44] <RoanKattouw>	 Yup, got a 5
[18:40:57] <rsterbin>	 cool
[18:41:19] <RoanKattouw>	 Am I good to deploy now?
[18:41:37] <rsterbin>	 all clear on my end.  DarTar?
[18:41:53] <fabriceflorin>	 I just tested E a couple more times, so you should now see 2 more 5's
[18:42:15] <rsterbin>	 fyi - DarTar -- Roan's push on March 21 was for af_link_id
[18:42:16] <DarTar>	 good for me, looking into the other clicktracking issues right now
[18:43:02] <rsterbin>	 so we should have correct af_link_id in production now as well
[18:43:09] <DarTar>	 rsterbin: I am confused, the patch was written and pushed on the 21?
[18:43:22] <DarTar>	 (the one for af_link_is, that is)
[18:43:26] <RoanKattouw>	 I don't know when it was pushed, and frankly I don't care. Can I deploy or not?
[18:43:32] <DarTar>	 yes :)
[18:43:55] * RoanKattouw  doesn't want to sound too grumpy but all this talking is taking time

[18:43:58] <RoanKattouw>	 OK, deploying
[18:44:10] <fabriceflorin>	 RoanKattouw, you have my go-ahead to deploy now, unless Dario has serious concerns about any remaining AFT5 data integrity issues.
[18:44:32] <RoanKattouw>	 scap is running, this will take a while
[18:44:42] <RoanKattouw>	 Please do not attempt to test on the live site until I say it's done
[18:44:43] <DarTar>	 no, we're ok to go for af_link_id, I haven't been able to look at the other clicktracking issues, but I am doing it now
[18:45:16] <DarTar>	 i.e. we're ok to push to production and if I catch any still open issue with clicktracking I'll report back
[18:45:43] <yoni_omniti>	 DarTar: can you look into other CT issues before we push to production?
[18:46:02] <DarTar>	 I am doing it right now, but I don't know if Roan has much time left
[18:46:11] <yoni_omniti>	 ok, thanks!
[18:46:19] <fabriceflorin>	 Thanks, RoanKattouw, I appreciate your patience with all this. Would you prefer to continue to work on AFT4 issues with Dario? or can you address the question about email delays to oversighters? (that one is not a super-high priority item, so it could be postponed until later, if need be)
[18:47:02] <RoanKattouw>	 fabriceflorin: I told you about the hold-up on the e-mail issue. I'd poke Tim but it's 6am there. I'll file an RT ticket I guess
[18:47:12] <RoanKattouw>	 But half the ops dept is on vacation, it's gonna be slow probably
[18:47:30] <DarTar>	 CT issue (1): All bucket X events are misnamed: FIXED
[18:48:13] <fabriceflorin>	 RoanKattouw, thanks for filing an RT ticket about the email issue. We can wait until next week to do the email address switch, as I want to confirm that issue is solved while we can still test it on our end.
[18:48:40] <DarTar>	 CT issue (2): Bucket IDs disappear at the end of the Edit CTA funnel: FIXED
[18:49:21] <DarTar>	 CT issue (3): Blank edit saves should not be logged as edit_success events: NOT FIXED (but we agreed to not fix it to make sure AFT5 and AFT4 behave in the same way)
[18:49:59] <DarTar>	 CT issue (4): Version number reset to 0 - don't know, have we changed the version number ?
[18:50:17] <DarTar>	 it shows up as 1 in the testwiki logs
[18:50:27] <rsterbin>	 if it says "ext.articleFeedbackv5@1", it's correct
[18:50:35] <DarTar>	 ok
[18:50:38] <DarTar>	 FIXED
[18:50:41] <rsterbin>	 :)
[18:50:45] <yoni_omniti>	 :)
[18:50:52] <yoni_omniti>	 music to my ears
[18:51:06] <fabriceflorin>	 Thanks, DarTar, much appreciated. I am glad that all issues are FIXED except for the 3rd one, which we agreed to postpone.
[18:51:38] <DarTar>	 CT issue (5): inconsistent clicktracking events: not really fixed because we can still serve Option2 or Option 3, but I don't care at this stage
[18:52:27] <DarTar>	 ideally I would have disabled entirely O2 and O3 but I'll keep an eye on the data and see if we're still getting this inconsistency in production a few days from now
[18:52:40] <DarTar>	 that's all for AFT5 CT issues
[18:52:42] <rsterbin>	 i'd like to remove the other options entirely, but i think that's fabrice's call
[18:52:58] <DarTar>	 oh Fabrice, if that's a simple config change can we do that?
[18:53:10] <DarTar>	 rsterbin: is it a config change?
[18:53:19] <rsterbin>	 no, it's a js change
[18:53:37] <DarTar>	 hmm ok, is that likely to break anything or can we assume we can safely push it
[18:53:48] <fabriceflorin>	 Now let's go test on production and make sure there are no glitches there either. RoanKattouw, how long do you think we should wait to test on production? I am still seeing the old graphics when I test on production, even after I clear my caches. Do you need to purge any server caches on your end?
[18:53:49] <fabriceflorin>	 http://en.wikipedia.org/wiki/Golden-crowned_Sparrow?bucket=1&debug=true&aftv5_link=E
[18:54:39] <rsterbin>	 DarTar: not today.  we should talk about it on the call on monday.
[18:54:45] <DarTar>	 ok fine with me
[18:55:00] <fabriceflorin>	 DarTar, if keeping options 2 and 3 doesn't corrupt the data, I don't see a problem with keeping them there a bit longer.
[18:55:12] <rsterbin>	 fabriceflorin: i don't think Roan said the push was done.
[18:55:17] <RoanKattouw>	 I did not
[18:55:24] <RoanKattouw>	 It's still running
[18:55:48] <RoanKattouw>	 fabriceflorin: *grumble* Did you listen to me at all, this week or last week?
[18:55:51] <fabriceflorin>	 Thanks, rsterbin, good point. We'll wait to hear from RoanKattouw then.
[18:55:59] <RoanKattouw>	 RoanKattouw	scap is running, this will take a while
[18:56:01] <RoanKattouw>	 	RoanKattouw	Please do not attempt to test on the live site until I say it's done
[18:56:21] <fabriceflorin>	 RoanKattouw: sounds good, thanks.
[18:56:51] <DarTar>	  fabriceflorin: we did have corrupt data (mixed phase 1 and phase 2 events) over the last few days, possibly due to cookies/caching - I assume I shouldn't be seeing other o2 and o3 at all other than our tests from now on
[18:57:52] <DarTar>	 as long as everybody gets option 1 and there's no lurking cookie that can override that, I am good with that
[18:57:57] <RoanKattouw>	 fabriceflorin: OK, deploy is done
[18:58:11] <RoanKattouw>	 Caches may take an additional 5-10 minutes to clear
[18:58:11] <rsterbin>	 ten min till cache clears, right?
[18:58:34] <rsterbin>	 simultaneous comment action!
[19:00:16] <hexmode>	 git triage!!!!
[19:00:49] <hexmode>	 http://etherpad.wikimedia.org/BugTriage-2012-03
[19:01:16] <Nikerabbit>	 hola
[19:01:19] <hexmode>	 ugh
[19:01:25] <hexmode>	 where is ^demon!
[19:01:44] <hexmode>	 Nikerabbit: hey... so who else can we harass?
[19:02:02] <hexmode>	 Krinkle-away: come back!
[19:02:09] <DarTar>	 Golden-crowned Sparrow is bound to become a 3-month trending article on enwiki
[19:02:13] <fabriceflorin>	 OK, I'll wait 5 more mins to test AFT5 on production. I have emptied all my caches and shut down all 3 browsers, just in case.
[19:02:42] <Nikerabbit>	 hexmode: dunno who else is here
[19:02:57] <hexmode>	 hashar: yo!
[19:03:07] <hashar>	 :)
[19:03:13] <hashar>	 oh19:00UTC already
[19:03:19] <hexmode>	 hashar:     http://bugzilla.wikimedia.org/33773:       [TestSwarm] fetcher need git support
[19:03:19] <hashar>	 do we do the triage here ?
[19:03:30] <hexmode>	 is that ^^ something you're doing?
[19:03:38] <hexmode>	 hashar: yes, triage is here
[19:03:53] <hashar>	 hexmode: 33773 we are probably going to wontfix
[19:04:09] <hashar>	 lets wont fix right now
[19:04:24] <hexmode>	 yes, lets get hashar to wontfix it
[19:04:26] <hexmode>	 :)
[19:04:28] <Nemo_bis>	 sigh
[19:04:43] <Nemo_bis>	 grr
[19:04:45] <hexmode>	 "NO"?
[19:04:59] <hashar>	 just did hexmode
[19:05:08] <hexmode>	 Nemo_bis: sounds like you adamently don't want one ;)
[19:05:21] <hashar>	 I wish we had chad
[19:05:24] <Nemo_bis>	 no, sounds like there's too much crap in the topic
[19:06:04] <hexmode>	 hashar: how about     http://bugzilla.wikimedia.org/34138:       Stage source tree on fenari with git instead of svn
[19:06:21] <hashar>	 Reedy: ^^^
[19:06:30] <hexmode>	 hashar: and yes, chad would be nice to have, I did invite him.
[19:06:36] <hexmode>	 :(
[19:06:37] <sumanah>	 I'm sorry for the delay - net problems
[19:06:41] <RoanKattouw>	 That's going to happen later, but not immediately
[19:06:43] <hashar>	 I do not know anything about the progress of staging git on fenari
[19:07:01] <RoanKattouw>	 We need to have the deployment branch set up properly
[19:07:06] <hashar>	 that is probably high priority but not trivial :-)
[19:07:20] <RoanKattouw>	 This will probably happen naturally as a consequence of deploying 1.20wmf1
[19:07:31] <hexmode>	 RoanKattouw: could you update comments on the fenari stuff?
[19:07:37] <hexmode>	 nothing since the 12th
[19:07:42] <RoanKattouw>	 But for details I defer to Robla (not here?!) and Chad (racking servers in eqiad today)
[19:07:50] <RoanKattouw>	 hexmode: Get robla or ^demon to
[19:08:07] <hexmode>	 will do
[19:08:18] <hexmode>	 robla is out this week
[19:08:20] <hashar>	 so Chad is ops today ? :-D  wonderful, I know who I will poke next time I have a physical server issue
[19:08:54] <hexmode>	     http://bugzilla.wikimedia.org/35139:       Create the WMF deployment branch
[19:09:24] <hexmode>	 is that a pre-req for fenari stuff?
[19:09:27] * hexmode  checks

[19:09:37] <hashar>	 hexmode: yes blocks it
[19:09:49] <rsterbin>	 fabriceflorin: do you need me any longer?
[19:09:54] <hashar>	 that is the deployement branch RoanKattouw was referring to
[19:10:07] <sumanah>	 We may just need to do a first pass today, and then run it past Chad when he returns
[19:11:06] <DarTar>	 RoanKattouw: shall we review the AFT4 log problems later today? Tomorrow? I'll keep an eye on the enwiki logs anyway to see if the additional data is stored there
[19:11:42] <rsterbin>	 welp, i think we're good, so i'm going to lunch
[19:11:46] <rsterbin>	 \nick rsterbin_away
[19:12:02] <hexmode>	 no reedy?
[19:12:09] <RoanKattouw>	 DarTar: Tomorrow?
[19:12:59] <Nikerabbit>	 :(
[19:13:11] <hexmode>	 hashar: https://bugzilla.wikimedia.org/34141 Integrate Jenkins with Git
[19:13:15] <hexmode>	 status?
[19:13:31] <RoanKattouw>	 That's the most important thing right now IMO
[19:13:39] <hashar>	 hexmode: high priority, I will work again on it soon(tm)
[19:13:48] <hashar>	 I agree with Roan, that is the top one
[19:13:54] <hashar>	 I have been busy with 1.19 then git migration
[19:14:01] <DarTar>	 RoanKattouw: sounds good
[19:14:05] * hexmode  knocks it up to "HIGHEST"

[19:14:06] <hashar>	 git seems good nowadays since most people are up to my level with gerrit
[19:14:13] <hashar>	 and seems most people are using it now
[19:14:16] <DarTar>	 do you want to block a slot? What time works for you?
[19:14:21] <hashar>	 so I will switch from git back to jenkins
[19:14:23] <RoanKattouw>	 I would like to at least have PHP lint working for all mediawiki/* projects
[19:14:30] <hexmode>	 sumanah: note for new HIGHEST ^^
[19:14:33] <sumanah>	 thank you!
[19:14:37] <RoanKattouw>	 And then unit tests for core
[19:14:38] <Nikerabbit>	 btw gerrit is loading very slow for me atm (30+ second)
[19:14:58] <hashar>	 RoanKattouw: we already have some unit tests for core in jenkins. Had since the switch to git wednesday 21st
[19:15:01] <sumanah>	 oh did we turn the usability flag on? that might slow things down (j/k)
[19:15:11] <hashar>	 RoanKattouw: but no reporting since jenkins does not aggregate children jobs results :-/
[19:15:31] <hexmode>	     http://bugzilla.wikimedia.org/34426:       installer: ability to override LocalSettings.php filename
[19:15:57] <hashar>	 hexmode: I will eventually need that one to write the job that tests postgre / mysql
[19:16:07] <sumanah>	 aiee, someone feel free to undo my weird bullet madness in Etherpad.
[19:16:17] <hexmode>	 \o/ pg testing
[19:16:21] <hexmode>	 sumanah: c-z
[19:16:27] <hexmode>	 control-z
[19:16:29] <hexmode>	 for undo
[19:16:43] <sumanah>	 does not work for me
[19:16:51] <hexmode>	 :(
[19:17:00] * sumanah  refrains from editing

[19:17:04] <hexmode>	 well, I'll fix it.
[19:17:06] <hexmode>	 1s
[19:17:08] <sumanah>	 thanks
[19:17:36] <sumanah>	 hashar: http://bugzilla.wikimedia.org/34426 and that is a bug for you to fix, right?
[19:17:44] <sumanah>	 hashar: as part of the CI stuff
[19:17:54] <hashar>	 yes I opened it
[19:17:58] <sumanah>	 ok
[19:17:58] <hashar>	 as a reminder
[19:18:03] <DarTar>	 RoanKattouw: does 10AM PDT work for you tomorrow?
[19:18:13] <sumanah>	 !b 34349
[19:18:13] <mw-bot>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=34349
[19:18:14] <DarTar>	 nothing in your calendar
[19:18:16] <hashar>	 it is not a big change anyway, just have to chat about it which Chad for a proper implementation.
[19:18:31] <RoanKattouw>	 DarTar: Sure
[19:18:32] <hashar>	 I am adding a comment to 34426 bug
[19:18:40] <DarTar>	 cool, sending you an invite
[19:18:43] <sumanah>	 Bug 34349 - Triage Git conversion issues wiki pages
[19:18:54] <sumanah>	 thanks hashar
[19:19:20] <sumanah>	 so basically Bug 34349 - Triage Git conversion issues wiki pages refers to a continuing activity that I should in theory be doing -- not triaging those issues but moving them into BZ!
[19:19:40] <sumanah>	 since people will just naturally be asking questions and mentioning complaints on wikitech-l and on the wiki
[19:19:55] <DarTar>	 RoanKattouw: invite sent and off to grab lunch unless fabriceflorin wants me to help with something?
[19:20:26] <sumanah>	 Assigning to self.
[19:21:01] <fabriceflorin>	 Hi RoanKattouw, rsterbin and yoni_omniti, I am still not able to test the latest versions of the AFT5 feedback links on the various browsers. So I think we should test a bit later, since it can take several hours for all the caches to clear. Roan, is there any server-side cache that needs to be purged, or have you done that already?
[19:21:40] <hexmode>	 how about: http://bugzilla.wikimedia.org/35144:       Autolink to new Gerrit / Git changesets and SHA-1 commits
[19:21:55] <hashar>	 sumanah: 34349 is about moving https://www.mediawiki.org/wiki/Git/Conversion/issues into bugzilla
[19:22:06] <hashar>	 sumanah: I have ranted about how a wiki is not a bug tracker :-]
[19:22:08] <RoanKattouw>	 fabriceflorin: It doesn't take several hours. What's your issue?
[19:22:22] <sumanah>	 hashar: now it is.
[19:22:23] <Nikerabbit>	 I've been just using full link to gerrit
[19:22:26] <sumanah>	 hashar: I just changed the summary
[19:22:32] <hashar>	 This afternoon, I have added a {{warn}} box on [[Git/Conversion/issues]]
[19:22:41] <fabriceflorin>	 I can confirm that Option E works on Safari: http://en.wikipedia.org/wiki/Golden-crowned_Sparrow?bucket=1&debug=true&aftv5_link=E (so that's good)
[19:23:07] <RoanKattouw>	 You're using &debug=true again?
[19:23:11] <RoanKattouw>	 That might be the source of your problem
[19:23:17] <RoanKattouw>	 Please try with &debug=false instead
[19:23:18] <DarTar>	 fabriceflorin: I am logged in and bucketed with Option A but I don't see the close button any more, did we kill it?
[19:23:50] <fabriceflorin>	 Yes, I am seeing a graphic artifact on Option A on Safari, where the old 'X' graphic is superimposed on top of the new text' X' box when I hover over "Improve this page" text link. http://en.wikipedia.org/wiki/Golden-crowned_Sparrow?bucket=1&debug=true&aftv5_link=A
[19:24:08] <hashar>	 I think that bug 35144 should be assigned to Chad.  It was IIRC
[19:24:19] <RoanKattouw>	 fabriceflorin: Please try with debug=false
[19:24:20] <fabriceflorin>	 DarTar, the "X" button only displays if you are logged in.
[19:24:26] <DarTar>	 I am logged in
[19:24:27] <Nikerabbit>	 it was what?
[19:24:29] <hashar>	 grr
[19:24:31] <DarTar>	 both on Safari and Chrome
[19:24:36] <hashar>	 we need a new channel
[19:24:38] <RoanKattouw>	 fabriceflorin: This is exactly the same thing as last week. debug=true has caching issues but you shouldn't be using it unless you know what you're doing
[19:24:39] <DarTar>	 tried also Chrome stealth mode
[19:24:53] <DarTar>	 logged in, I get OptionA, no close button
[19:24:53] <hashar>	 because I can't follow that many conversations overlapping in my IRC client
[19:25:16] <sumanah>	 !b 35144
[19:25:17] <mw-bot>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=35144
[19:25:49] <fabriceflorin>	 Hi, RoanKattouw and rsterbin, I tried 'false' but then the "X" button doesn't display at all on Safari, even though I am logged in. http://en.wikipedia.org/wiki/Golden-crowned_Sparrow?bucket=1&debug=false&aftv5_link=A
[19:26:37] <hexmode>	 sumanah: lowering priority on this one since we don't seem to be motivated to get it done and will need to rely on volunteers: Bug 35145 - Make git-review more easy to install on Windows
[19:27:08] <hashar>	 35145 should be down to lowest priority IMHO
[19:27:26] <hashar>	 the fallback is to use a push to refs/for/master
[19:27:34] <hexmode>	 you're wish, my command
[19:27:36] <hashar>	 and use the commands provided by gerrit
[19:27:39] <hexmode>	 your
[19:27:45] <sumanah>	 hexmode: yes, I think you're right. hashar not LOWEST but low
[19:27:50] <sumanah>	 Sid and other Windows volunteers think it's important
[19:27:50] <sumanah>	 so, to balance that: low
[19:27:58] <hexmode>	 kk
[19:27:59] <DarTar>	 fabriceflorin, RoanKattouw, rsterbin_away: confirmed here, even  without forcing bucketing/debugging
[19:28:04] <hashar>	 Raymond_ proposed to set up a wiki page that would hold multiple screenshots explaining how to install it
[19:28:17] <hexmode>	 "priority is HARD"
[19:28:18] * DarTar  off to grab a sandwich, brb

[19:28:21] <hashar>	 haha
[19:28:39] <fabriceflorin>	 RoanKattouw, DarTar, yoni_omniti and rsterbin, I can reproduce the problem with Option A on Chrome, in logged in mode. If I use debug=true, I get the old "X" graphic on top of the new "X" text link. If I use debug=false, I don't see the close button at all. http://en.wikipedia.org/wiki/Golden-crowned_Sparrow?bucket=1&debug=false&aftv5_link=A
[19:29:09] <hashar>	 I just do not understand how installing a python package under Windows could be so hard
[19:29:18] <hashar>	 that does not really make sense to me but there might be an issue :-D
[19:29:18] <yoni_omniti>	 fabriceflorin: the issue is already fixed in code, we can only wait for the various caches to update
[19:29:26] <hashar>	 anyway, next! ;-D
[19:29:29] <Nikerabbit>	 hashar: it's not easy on linux either
[19:29:44] <Nikerabbit>	 have to download via some weird stuff outside of package manager
[19:29:45] <hashar>	 Nikerabbit: I will be glad to guide anyone who have a trouble installing git-review on linux
[19:29:58] <hashar>	 Nikerabbit: that is like SUPER easy and there is a oneliner to do it
[19:30:25] <Nikerabbit>	 can someone explain what 35460 is about?
[19:30:32] <hexmode>	 re     http://bugzilla.wikimedia.org/35462:       Tell git on gerrit.wikimedia.org not to generate spurious terminal sequences
[19:30:42] <yoni_omniti>	 fabriceflorin: the missing button, however, is interesting…. its missing for me too
[19:30:50] <hexmode>	 could there be a .profile setting TERM to dumb on gerrit?
[19:31:01] <hashar>	 hexmode: that is an upstream issue we can remove it as a blocker
[19:31:13] <yoni_omniti>	 when reha comes back from lunch, i'll ask her to check what is being served from the production servers - maybe a revision didn't make it in
[19:31:19] <hashar>	 it is just a corner case when using `script`
[19:31:26] <hexmode>	 hashar: yeah, not a blocker, just thinking of possible solutions
[19:31:44] <hashar>	 we do not want to keep tracking it I guess.  saper is probably to patch upstream git to have the issue fixed
[19:31:53] <hashar>	 so I would say it is a reminder for him and we can forget about that bug
[19:32:34] <hexmode>	     http://bugzilla.wikimedia.org/35469:       Let LocalisationUpdate fetch from Git
[19:32:34] <hexmode>	 Nikerabbit: does it yet?
[19:32:34] <Nikerabbit>	 hexmode: apparently there is fix but ops must approve it (hearsay)
[19:32:43] <fabriceflorin>	 Yoni_omniti: OK, I recommend that we all take an hour break, then reconvene in an hour, at 4:30pm ET, to see if the caches have been cleared. Please let us know what you find out  about the missing button. RoanKattouw: I understand it's past your bedtime, so we will contact you by email tomorrow if we need any assistance. Thank you very much for all your good work today! I realize how much of your time all this is requiring, and will try to
[19:32:43] <fabriceflorin>	 streamline the process so that we don't require so much in the future. Be well.
[19:35:03] <Nikerabbit>	 RoanKattouw: ^^ didn't you do something about that?
[19:35:27] <hashar>	 we had the change merged and deployed yesterday
[19:35:35] <hashar>	 I think the cron job runs during the european night
[19:35:45] <RoanKattouw>	 2:00 UTC IIRC
[19:35:45] <Nikerabbit>	 nice
[19:35:46] <hashar>	 so we should be able to verify if it ran correctly :-)
[19:36:27] <hexmode>	 hashar: you can do that, right?
[19:36:41] * hexmode  is enjoying his birthday meeting

[19:36:43] <hashar>	 2:00 UTC correct
[19:36:43] <hexmode>	 :)
[19:36:51] <hashar>	 HAPPY BIRTHDAY !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
[19:37:36] <hexmode>	 http://bugzilla.wikimedia.org/35497:       Two-way syncing between Gerrit and GitHub
[19:37:45] <hashar>	 I am going to bring back 35469 back in pool
[19:38:13] <hashar>	 hmm no
[19:38:14] <hexmode>	 sumanah: if I can find an api for github, submitting branches from there shouldn't be an issue
[19:38:19] <hashar>	 will check with i18n team soon
[19:38:20] <sumanah>	 hexmode: great!
[19:38:39] <sumanah>	 hexmode: I hope you have a happy birthday.
[19:38:44] * hashar  keeps mouth shut to avoid ranting about github / gerrit :-D

[19:38:48] <Nikerabbit>	 (the i18n team is here :)
[19:38:58] <hexmode>	 I'm smiling now :)
[19:39:16] <hashar>	 Nikerabbit: do you have a way to check that l10nupdate correctly run from git this night ?  https://bugzilla.wikimedia.org/35469 ?
[19:40:17] <Nikerabbit>	 hashar: pick a new/changed message that has only been committed to git and check whether it exists in relevant language project
[19:40:51] <hashar>	 you know, I don't even know what l10nupdate does :-D
[19:41:38] <Nikerabbit>	 sigh
[19:41:43] <Nikerabbit>	 how do I filter by owner in gerrit
[19:41:44] <hexmode>	 Nikerabbit: How would we find ones that have only been committed to git?
[19:41:49] <hexmode>	 heh
[19:41:50] <hexmode>	 jinx
[19:42:01] <RoanKattouw>	 Nikerabbit: owner:catrope
[19:42:01] <alolita>	 hexmode: happy bday!
[19:42:08] <Nikerabbit>	 hmph, must have made a typo
[19:42:14] <Nikerabbit>	 https://gerrit.wikimedia.org/r/#q,owner:l10n-bot,n,z
[19:42:21] <hashar>	 Nikerabbit: I could probably find such a messages in maintenance/messages.php :D
[19:42:30] <hexmode>	 alolita: tyvm :)
[19:42:39] <hexmode>	 now, if only I was younger :P
[19:43:14] <alolita>	 hexmode: aaah - we need the growing younger magic potion!
[19:43:50] <hexmode>	 next birthday, I'll magically be 10 or 20 years younger
[19:43:56] <hexmode>	 ok, back to git
[19:44:02] <alolita>	 hexmode: yup ; back to git!!
[19:44:06] <hashar>	 Nikerabbit: can you take care of the verification ?
[19:44:17] <Nikerabbit>	 https://gerrit.wikimedia.org/r/#patch,sidebyside,3650,1,Narayam.i18n.php vs. http://ta.wikipedia.org/wiki/?uselang=da# I don't see the updated translation there
[19:44:19] <hashar>	 Nikerabbit: I could do it but will probably spend like 2 hours figuring out how to verify it
[19:44:20] <hexmode>	 Nikerabbit: looks like it hasn't had time to be merged?
[19:44:26] <hashar>	 ooh
[19:45:27] <hashar>	 Nikerabbit: could it be a cache issue ?
[19:46:16] <Nikerabbit>	 hashar: nope: http://da.wikipedia.org/wiki/MediaWiki:Narayam-menu-tooltip
[19:46:26] * hexmode  tries to superstitiously purge

[19:47:25] <RoanKattouw>	 02:10 logmsgbot: LocalisationUpdate completed (1.19) at Wed Mar 28 02:10:34 UTC 2012
[19:47:28] <Nikerabbit>	 another example: http://eo.wikipedia.org/wiki/MediaWiki:Webfonts-preview-sampletext vs. https://gerrit.wikimedia.org/r/#patch,sidebyside,3654,1,WebFonts.i18n.php
[19:47:28] <RoanKattouw>	 So it ran, supposedly
[19:47:32] <hexmode>	 k, will update the bug
[19:47:56] <RoanKattouw>	 Hmm
[19:48:41] <RoanKattouw>	 This might have been broken by manualRecache instead
[19:48:53] <RoanKattouw>	 But then again I ran scap today
[19:49:35] <hashar>	 Nikerabbit: I have added your test case on https://bugzilla.wikimedia.org/35469 ;)
[19:50:21] <hexmode>	     http://bugzilla.wikimedia.org/35429:       Setup external git mirrors/clones
[19:50:48] * hashar  keeps mouth shut just like the Github rant :-]

[19:50:56] <Nikerabbit>	 high prio? really?
[19:51:09] <hashar>	 I am not sure why we have two bugs about that though
[19:51:25] <Nikerabbit>	 wasn't the other one about two-way changes
[19:51:26] <hexmode>	 evidently sam thinks so ;)
[19:51:28] <Nikerabbit>	 so this would block that one
[19:51:40] <hashar>	 oh no that one is to setup the mlirror somewhere, the previous bug (35497) is to have them synced
[19:52:15] <hashar>	 <rant>   so we should just close boths as NOT_NEEDED_HAVE_GERRIT ;D  </rant>
[19:52:38] <Nikerabbit>	 I think that can be done later
[19:52:55] <Nikerabbit>	 yup
[19:53:02] <hashar>	 that might be part of some outreach program I am not aware of
[19:53:03] <hexmode>	 hashar: ... hush!
[19:53:09] <hexmode>	     http://bugzilla.wikimedia.org/35508:       List of Gerrit users
[19:53:11] <hashar>	 should ask sumanah but she left ;-)
[19:54:11] <Nikerabbit>	 would be awesome indeed
[19:54:12] <hashar>	 I think that one should be reported upstream as a feature request
[19:54:30] <MaxSem>	 hexmode, any bugs for me?
[19:54:46] <hexmode>	 for 20%...?
[19:54:50] <hashar>	 we might be able to hack something though by generating a user page based on the Gerrit database
[19:55:27] <Nikerabbit>	 user page?
[19:55:47] <sumanah>	 hi
[19:55:48] <hashar>	 anyway long term, we need that in upstream ;-D
[19:55:52] <sumanah>	 I'm sorry, my net connection died
[19:55:55] <sumanah>	 you wanted to ask me about something?
[19:55:59] <MaxSem>	 hexmode, yep
[19:56:23] <hexmode>	 MaxSem: three here: https://bugzilla.wikimedia.org/buglist.cgi?resolution=---&query_format=advanced&target_milestone=1.19.0%20release
[19:56:28] <sumanah>	 "<hashar>	we might be able to hack something though by generating a user page based on the Gerrit database" --- yes
[19:56:32] <sumanah>	 that is the request for now
[19:56:38] <hashar>	 so make it clearer
[19:56:45] <hashar>	 I would also want to be unassigned from https://bugzilla.wikimedia.org/35508
[19:56:47] <sumanah>	 hashar: https://groups.google.com/group/repo-discuss/msg/c426b6a83400b58e?dmode=source&pli=1 has a description of how to do that
[19:56:51] <Nikerabbit>	 user page where?
[19:56:53] <hexmode>	 MaxSem: so ignore 34885, but other than that...
[19:57:06] <hashar>	 cause I will never be able to work again on jenkins if I keep getting new stuff added ;-)
[19:57:17] <MaxSem>	 35046 is also not mine
[19:57:50] <sumanah>	 hashar: unassigned you and added a clarifying commnet
[19:57:52] <hexmode>	 hashar: done
[19:57:52] <sumanah>	 comment*
[19:57:58] <sumanah>	 !b 35046
[19:57:58] <mw-bot>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=35046
[19:58:13] <MaxSem>	 I'll take a look at the autopatrolling bug a bit later
[19:58:40] <hexmode>	 MaxSem: excellent
[19:58:55] <hexmode>	 I think krinkle was gonna look at 35046 anyway
[19:59:13] <sumanah>	 MaxSem: if you want to help  out reviewing extensions that are awaiting deployment review, that would also be a cool thing to do for 20% time
[19:59:21] <Krinkle>	 ?
[19:59:50] <sumanah>	 I think we have 9 more git-related bugs to triage
[20:00:01] <sumanah>	 and we are going overtime for our European contributors
[20:00:02] <MaxSem>	 stabstabstab
[20:00:07] <sumanah>	 so let's go on
[20:00:11] <MaxSem>	 hate you, git
[20:00:11] <Krinkle>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=35565	[Regression] Special:Log/patrol doesn't show "(automatic)" for autopatrols anymore
[20:00:23] <RoanKattouw>	 hehe, I think I found the other half of the l10nupdate bug
[20:00:31] <RoanKattouw>	 The extensions repo uses submodules
[20:00:37] <hexmode>	 could this be just add it to a wiki somewhere?    http://bugzilla.wikimedia.org/35510:       Make Gerrit project list more discoverable and available
[20:01:53] <RoanKattouw>	 So git pull doesn't work
[20:01:57] <sumanah>	 hexmode: the Gerrit project list needs to be available where people will look for it
[20:02:12] <sumanah>	 hexmode: let's leave that for Chad
[20:02:13] <sumanah>	 and move on
[20:02:31] <hexmode>	 sumanah: don't people look at the git pages on mw.o before they get to gerrit?
[20:02:34] <hexmode>	 but moving on
[20:02:45] <hexmode>	     http://bugzilla.wikimedia.org/35455: Change Gerrit database schema to support utf8
[20:02:48] <hexmode>	 upstream?
[20:02:54] <hexmode>	 or switch to pg?
[20:02:58] <hashar>	 PG
[20:03:07] <sumanah>	 Postgres?
[20:03:09] <hashar>	 or embed db :-D
[20:03:20] <hexmode>	 who thought mysql was a good db, anyway :P
[20:03:30] <hashar>	 but since I am pretty sure everyone will want to stick to MySQL, we need to find a way to have it fixed
[20:03:51] <hashar>	 possibly by changing mysql charset, but there might be some side effects
[20:04:06] <hashar>	 so we probably want to test it on some labs machine first
[20:04:07] <sumanah>	 !seen ^demon
[20:04:32] <hashar>	 one possible issue is with the length of file path and the size of the column that hold full paths
[20:04:55] <hashar>	 that "might" cause issues such as "key too long"
[20:04:56] <hashar>	 not sure
[20:04:59] <hexmode>	 hrm :( is that an issue for pg?
[20:05:05] <hashar>	 no idea
[20:05:17] <Nikerabbit>	 can't we just make it binary like mediawiki does?
[20:05:18] <hashar>	 anyway ops will want to stick to mysql
[20:05:23] <hashar>	 to handle backup and so on
[20:05:24] <hexmode>	 ok, upstream if mysql, switch otherwise. http://bugzilla.wikimedia.org/35531: Send email notifications of all MediaWiki commits to mediawiki-cvs list
[20:05:45] <hexmode>	 email notifications ... that is just chad, right?
[20:05:59] <hashar>	 and upstream
[20:06:10] <hashar>	 that should be an option in Gerrit so we want to report it there
[20:06:12] <sumanah>	 hexmode: there are a few people who want that to happen
[20:06:20] <hashar>	 meanwhile we can use a hook to send email whenever something is done
[20:06:27] <hashar>	 the request is unclear though
[20:06:32] <sumanah>	 like, this might help with saper's request to get better offline review capability
[20:06:40] <hashar>	 should we send notification on any patch submission or only on merge ?
[20:06:50] <sumanah>	 Nemo_bis: can you clarify ^^^ https://bugzilla.wikimedia.org/show_bug.cgi?id=35531 /
[20:06:51] <sumanah>	 ?
[20:07:08] <Nikerabbit>	 I would at least want submissions
[20:07:22] <hexmode>	 for listmail, submissions
[20:07:22] <Nikerabbit>	 there could be other list which included all actions?
[20:07:54] <sumanah>	 hexmode: I can collate these questions into a clarification request on the bug -- let's move on
[20:08:12] <Nikerabbit>	 sumanah: it originated from my request
[20:08:32] <hashar>	 Nikerabbit so you might want to add a comment to 35531
[20:08:45] <Nikerabbit>	 I can add what I just said
[20:08:51] <hashar>	 Nikerabbit: that surely can be fixed by hooks while waiting for an upstream fix
[20:08:54] <Nikerabbit>	 otoh, without diffs the emails are not very useful
[20:09:12] <Nikerabbit>	 but better than nothing
[20:09:14] <hexmode>	 for bug 35535 - No way to mark merged commits as needing fixing -- are we happy with the bz solutions?
[20:09:31] <Nikerabbit>	 it's suboptimal
[20:09:38] <hashar>	 Nikerabbit: as I said on wikitech-l , you might want another bug blocked by 35531 that would request full diff
[20:09:48] <Nikerabbit>	 hashar: there is one
[20:09:58] <Nikerabbit>	 few lines down the pad
[20:10:39] <hashar>	 coo
[20:10:40] <hashar>	 cool
[20:10:45] <hashar>	 good
[20:10:48] <hashar>	 well that is fine
[20:10:59] <hexmode>	 35537 and 35538 also require upstream work, I think
[20:11:25] <Nikerabbit>	 well
[20:11:30] <hexmode>	 or maybe just configuration?
[20:11:35] <hashar>	 355537 auto merge can be handled by a hook
[20:11:43] <Nikerabbit>	 chad and hashar has been working those for a week now...I'm still waiting for results
[20:12:10] <hexmode>	 kk
[20:12:12] <hashar>	 I actually haven't worked on it
[20:12:26] <hexmode>	 hashar: get to work
[20:12:29] <hexmode>	 :)
[20:12:32] <hashar>	 :D
[20:12:49] <hashar>	 I will probably just avoid any git / gerrit / whatever work
[20:13:00] <hashar>	 and focus on jenkins
[20:13:02] <sumanah>	 well, except that CI hooks into it
[20:13:35] <Nikerabbit>	 I consider 35537 critical for continued regular i18n updates
[20:13:54] <hashar>	 the way we do i18n updates need to be rethought entirely
[20:13:55] <RoanKattouw>	 I am working on fixing l10nupdate
[20:13:55] <sumanah>	 !b 35537
[20:13:55] <mw-bot>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=35537
[20:14:00] <Nikerabbit>	 hashar: yes but not now
[20:14:00] <RoanKattouw>	 I found out why it broke
[20:14:08] <hashar>	 but that is a subject for later
[20:14:33] <Nikerabbit>	 hashar: btw is there bug for that? :D
[20:14:43] <hexmode>	 hashar: later can be now if you want ... we've plowed through the git bugs :)
[20:14:53] <sumanah>	 OK, thanks all
[20:15:07] <sumanah>	 thanks Nemo_bis, Nikerabbit, hexmode, RoanKattouw, hashar
[20:15:13] <hashar>	 as for the auto merge, that can be handled in the gerrit hook  patchset-created  which is in operations/puppet.git : files/gerrit/hooks/patchset-created
[20:15:53] <Nikerabbit>	 hmm
[20:16:10] <Nikerabbit>	 did we actually look at 35532?
[20:16:42] <sumanah>	 !b 35532
[20:16:42] <mw-bot>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=35532
[20:17:06] <sumanah>	 I think we decided it should be a dependency of another -- hexmode is that right?
[20:17:10] <hexmode>	 Nikerabbit: it fell under the email for git commits conversation
[20:17:11] <hexmode>	 yep
[20:17:20] <sumanah>	 I have emailed Chad to ask when we can go over there
[20:17:22] <sumanah>	 these*
[20:17:56] <hashar>	 35532 and 35533 are conflicting :-D
[20:17:58] <Nikerabbit>	 hexmode: right. I just want to point that it is preventing me doing CR as I used to do it
[20:18:28] <hashar>	 they should probably both be blocker of 35531 (send email notifications)
[20:19:00] <Nikerabbit>	 I need /some/ solution that allows me to have quick overlook at the incoming submissions
[20:19:21] <sumanah>	 Nikerabbit: in comments on those bugs, could you point to your mailing list posts or other places where you've described your ideal CR workflow? I know you have emailed and said "this is what I'm used to" and described it step by step
[20:19:22] <hexmode>	 use the web, laxtrom
[20:20:01] <Nikerabbit>	 hexmode: +s, it's stream not tream :)
[20:20:26] <hexmode>	 :) I tried
[20:21:00] <hexmode>	 could someone review this: https://gerrit.wikimedia.org/r/#change,3874
[20:22:31] <hashar>	 oh man
[20:22:38] <hashar>	 off to some forex trading
[20:23:24] <gerrit-wm>	 New patchset: Diederik; "Rewrote append_char function (apologies to Tim) Removed exotic for/loop Replaced all strtok with strchr." [analytics/udp-filters] (master) - https://gerrit.wikimedia.org/r/3222
[20:31:27] <rsterbin>	 RoanKattouw: the missing close button appears to be because of this rule (only on en.wiki) -- #articleFeedbackv5-bottmrighttabclose, .articleFeedbackv5-close-trigger-link {    display: none;}
[20:31:37] <RoanKattouw>	 Ooh OK
[20:31:40] <RoanKattouw>	 I'll remove that rule
[20:31:42] <rsterbin>	 thanks
[20:31:50] <RoanKattouw>	 For future reference: don't misspell your CSS classes ;)
[20:32:06] <RoanKattouw>	 I had put it in in the assumption that whoever would fix the rendering of that link would also correct the typo in the class name
[20:32:10] <sumanah>	 ha
[20:32:27] <rsterbin>	 the problem is your comma
[20:32:53] <RoanKattouw>	 Wait, did I do that?
[20:32:59] <rsterbin>	 yup
[20:33:27] * RoanKattouw  shakes fist at Kalda

[20:33:29] <RoanKattouw>	 *Kaldari
[20:33:44] <RoanKattouw>	 Someone edited my hack :S
[20:34:17] <RoanKattouw>	 OK rule removed
[20:35:01] <rsterbin>	 thanks :)
[20:39:18] <rsterbin>	 RoanKattouw: this rule seems to exist on prod and not on test -- http://pastebin.com/H0iXKhxn
[20:39:43] <RoanKattouw>	 What's the origin URL?
[20:39:59] <rsterbin>	 http://en.wikipedia.org/w/index.php?title=Golden-crowned_Sparrow&bucket=1&debug=omg&aftv5_link=A
[20:40:18] <rsterbin>	 (figured i'd try any old value for debug)
[20:40:35] <RoanKattouw>	 Please try 'false' specifically
[20:40:54] <rsterbin>	 does the same thing
[20:41:27] <rsterbin>	 until just now, curiously
[20:41:29] <rsterbin>	 weird
[20:42:08] <RoanKattouw>	 I don't see that rule when using false
[20:43:06] <rsterbin>	 i don't either, as of a few minutes ago
[20:43:24] <rsterbin>	 when can we expect the cache to be cleared when using debug=true?
[20:43:45] <RoanKattouw>	 Hmm I have to manually clear that
[20:44:15] <RoanKattouw>	 Should be done now
[20:44:27] <rsterbin>	 thanks
[20:45:12] <rsterbin>	 awesome, there it is :)
[20:45:45] <RoanKattouw>	 Nikerabbit: Could you check LocalisationUpdate again? I believe it should be fixed now
[20:57:35] <sumanah>	 3 min to a meeting! time to prep!
[21:15:28] <Nikerabbit>	 RoanKattouw: yup looks better now
[21:15:35] <RoanKattouw>	 Yay
[21:15:38] <RoanKattouw>	 It's still syncing on my end
[21:28:28] <hashar>	 bed time, have a good night
[21:28:45] <RoanKattouw>	 Inez_, TrevorParscal: I detect a PM triangle, maybe we should just talk in here?
[21:35:42] <TrevorParscal>	 what's your github username?
[21:35:53] <RoanKattouw>	 Mine?
[21:35:57] <TrevorParscal>	 yes
[21:36:02] <RoanKattouw>	 catrope
[21:37:54] <RoanKattouw>	 Wait, we have a wikimedia account at github?
[21:38:09] <RoanKattouw>	 hah, nice
[21:39:09] <RoanKattouw>	 TrevorParscal: Where the hell is your repository
[21:39:13] <RoanKattouw>	 I can't find it
[21:39:41] <RoanKattouw>	 Oh, there it is
[21:39:46] <RoanKattouw>	 You added me to it before you created it or somethnig
[21:39:54] <Nikerabbit>	 is wikipedia-mobile going to come wmf now what we support git?
[21:40:30] <TrevorParscal>	 https://github.com/wikimedia/VisualEditor
[21:40:50] <RoanKattouw>	 Are you pushing the repo in or shall I?
[21:41:12] <TrevorParscal>	 i will
[21:41:16] <RoanKattouw>	 OK
[21:41:18] <TrevorParscal>	 or
[21:41:20] <TrevorParscal>	 crap, can you?
[21:41:26] <TrevorParscal>	 i just did it wrong already lolz
[21:41:48] <RoanKattouw>	 Their example is broken
[21:41:54] <RoanKattouw>	 'add origin' should be 'add github'
[21:41:54] <TrevorParscal>	 is it?
[21:42:07] <RoanKattouw>	 Well, it's broken for our specific use case at least
[21:42:11] <TrevorParscal>	 i see
[21:42:33] <RoanKattouw>	 Generally, replace origin with github, in the second command as well
[21:42:46] <TrevorParscal>	 interesting
[21:43:28] <TrevorParscal>	 are  you following the "Existing Git Repo?" instructions?
[21:43:36] <TrevorParscal>	 so it should be:
[21:44:03] <TrevorParscal>	 cd mediawiki/extensions/VisualEditor
[21:44:04] <TrevorParscal>	 git remote add github git@github.com:wikimedia/VisualEditor.git
[21:44:05] <TrevorParscal>	 git push -u github master
[21:44:17] <TrevorParscal>	 assuming you already have it checked out at that path
[21:44:29] <TrevorParscal>	 right?
[21:44:37] <RoanKattouw>	 Yes
[21:45:55] <TrevorParscal>	 are you going to do that? or should I?
[21:46:22] <TrevorParscal>	 i gotta walk to wikia
[21:46:38] <TrevorParscal>	 so, if you can do that - that would be sweet, I know you will do it right
[21:46:41] <TrevorParscal>	 gtg
[21:47:04] <RoanKattouw>	 OK
[22:47:24] <Nemo_bis>	 Reduced crap
[22:49:48] <preilly>	 TimStarling: ping
[22:50:04] <TimStarling>	 tes?
[22:50:07] <TimStarling>	 yes?
[22:51:45] <preilly>	 TimStarling: pm?
[23:05:26] <MaxSem>	 hexmode, I've fixed 35565
[23:28:51] <kaldari>	 so once I've addressed a negative code review in gerrit with a follow-up, how do I mark it for re-review? Or do I just hope they notice?
[23:29:38] <TimStarling>	 you just hope, I think
[23:29:44] <kaldari>	 fair enough :)
[23:30:21] <TimStarling>	 when you're a reviewer on a commit, the commits don't actually disappear from your "my changes" list when you review them
[23:30:44] <TimStarling>	 but some of them show up bold, possibly if there are new changes since you looked at them last
[23:31:12] <TimStarling>	 so maybe when you submit a new patchset the change will go bold in the review list
[23:31:59] <kaldari>	 hey MexSem, do you see my revised commit as bold in Gerrit?
[23:32:48] <kaldari>	 when you go under My Changes