[00:00:36] <bd808>	 The hhvm logs in beta cluster are full of "Notice: Undefined variable: wgMFPageActions in /srv/mediawiki/wmf-config/mobile.php on line 95
[00:01:06] <bd808>	 That line is the only place in the config that wgMFPageActions is mentioned.
[00:01:20] <bd808>	 Is that something in the extension config?
[00:01:31] <bd808>	 and if so why is it missing in beta?
[00:01:48] <bd808>	 jdlrobson, bmansurov: ^
[00:02:14] <bmansurov>	 let me check
[00:02:21] <jdlrobson>	 hmm we just moved to extension.json
[00:02:27] <jdlrobson>	 possibly related
[00:02:35] <bd808>	 https://github.com/search?q=%40wikimedia+wgMFPageActions&type=Code&utf8=%E2%9C%93
[00:02:41] <bd808>	 did it get lost?
[00:03:08] <jdlrobson>	 yuk
[00:03:08] <jdlrobson>	 $wgMFPageActions = array_diff( $wgMFPageActions, $wmgMFRemovePageActions );
[00:03:12] <jdlrobson>	 in wmf-config/mobile.php
[00:03:20] <jdlrobson>	 will that even work when loading via extension.json ?
[00:03:23] <bd808>	 https://github.com/wikimedia/mediawiki-extensions-MobileFrontend/blob/master/extension.json#L2029-L2034
[00:03:31] <bd808>	 apparently not
[00:03:44] <jdlrobson>	 bd808: i'd suggest killing that line in wmf-config/mobile.php
[00:03:59] <jdlrobson>	 and bmansurov i'd recommend dropping "upload" from the config in extension.json
[00:04:04] <jdlrobson>	 (from the default value)
[00:04:35] <bd808>	 the setting for wmgMFRemovePageActions removed upload
[00:04:42] <jdlrobson>	 that's a hacky way old school way of removing the upload button which we haven't supported in a long time
[00:04:48] <jdlrobson>	 it doesn't even exist in MobileFrontend anymore
[00:05:05] <bd808>	 cool. somebody want to make a patch and I'll get it deployed?
[00:05:10] <jdlrobson>	 bmansurov: can you write a patch for MobileFrontend or do you want me to write the patch ?
[00:07:23] <bmansurov>	 jdlrobson: if you're not in the middle of something
[00:07:23] <jdlrobson>	 ok ill do it. Will need a merger thouh
[00:07:23] <bmansurov>	 thanks, i'll review it
[00:07:49] <bd808>	 so is our fix the double whammy of taking "upload" out of the defaults and also removing the config hack?
[00:08:05] <jdlrobson>	 bd808:  https://gerrit.wikimedia.org/r/267808    and https://gerrit.wikimedia.org/r/267807
[00:08:07] <jdlrobson>	 review in that order :)
[00:08:14] <jdlrobson>	 ^ bmansurov 
[00:08:23] <bmansurov>	 ok
[00:11:18] <bd808>	 sweet. thanks jdlrobson 
[00:27:13] <Krinkle>	 jdlrobson: What is the "HTML" view (as opposed to "base" and "beta").
[00:28:01] <jdlrobson>	 Krinkle: to my knowledge HTML is something legacy - I was actually writing a patch to try removing it :)  I'm not sure if it's needed but was scared to tinker with it in same patch :)
[00:28:12] <Krinkle>	 jdlrobson: OK
[00:29:15] <Krinkle>	 jdlrobson: something unrelated, I see MF still has dev code for -qunit and the old non-plain, non-karma version of the QUnit entry point in corre.
[00:29:19] <Krinkle>	 I'd like to get rid of that
[00:29:22] <Krinkle>	 Is that used by anything?
[00:29:31] <Krinkle>	 Is anyone consuming the istanbul report?
[00:29:43] <jdlrobson>	 Krinkle: somewhat related => https://phabricator.wikimedia.org/T112568
[00:29:49] <jdlrobson>	 Krinkle: i dont think the istanbul report is used any more
[00:30:00] <jdlrobson>	 it was used ad-hoc to get an idea of where we sucked at writing tests
[00:30:04] <Krinkle>	 OK
[00:30:10] <jdlrobson>	 feel free to remove it
[00:30:34] <jdlrobson>	 bd808: let me know when those patches are merged so i can test
[00:31:00] <bd808>	 jdlrobson: there's a scap running right now :(
[00:31:01] <jdlrobson>	 (just want to verify no uploady things are popping up anywhere unexpectedly)
[00:31:08] <jdlrobson>	 SCAPMANNNN
[00:31:19] <bd808>	 not for our patches though
[00:31:36] <bd808>	 maybe we should just put it up for swat in the morning?
[00:31:39] <jdlrobson>	 MaxSem: assume you saw https://www.goldstar.com/tr/nc/46629/4107835/e/111082 ?
[00:32:21] <MaxSem>	 yawn. politics.
[00:32:25] <MaxSem>	 but thanks :)
[00:32:37] <jdlrobson>	 :-)
[00:33:16] <Krinkle>	 jdlrobson: Looks like dev-scripts/pre-commit mentions 'make qunit'
[00:33:26] <Krinkle>	 That's interesting.
[00:33:30] <Krinkle>	 It's not a bad practice :)
[00:33:59] <Krinkle>	 If people use that, that could be rewritten to simply "cd $MW_INSTALL_PATH" and run "grunt qunit" there
[00:34:01] <jdlrobson>	 yeh it was handy to run qunit tests on every commit up until it broke - i had to comment mine out :/ but in the README we do recommend people use make installhooks so they get up and running with best practices asap
[00:34:10] <Krinkle>	 OK
[00:46:58] <jdlrobson>	 Krinkle: are you able to make sense of these graphs? http://wpt.wmftest.org/video/compare.php?tests=160201_SY_AQ,160201_E2_AH,160201_71_AF,160201_2H_AJ they show first paint is radically better on the stable mode of the beta cluster compared to beta mode on beta cluster and production (beta and stable)
[00:47:08] <jdlrobson>	 first paint is around 5s on 2G for beta cluster and 10s on production
[00:47:27] <jdlrobson>	 we'd expect beta cluster to be as close to production as possible no?
[00:47:55] <Krinkle>	 beta cluster is generally an order of magnitude slower due to latency and virtualised resources
[00:48:14] <Krinkle>	 It's useless for absolute performance measures, only useful for relative measures (compare its own data to the past, not to prod)
[00:48:27] <Krinkle>	 jdlrobson: 
[00:49:34] <jdlrobson>	 okay i guess that makes sense, i was just surprised to see that beta cluster beta was very similar to production beta
[00:53:03] <jdlrobson>	 Krinkle: am getting a strangeTypeError: sockets.forEach is not a function when I run `make qunit`
[00:53:11] <jdlrobson>	 known issue in karma?
[00:53:17] <Krinkle>	 jdlrobson: Yes, 2 months ago.
[00:53:22] <Krinkle>	 jdlrobson: Be sure to npm install in mwcore dir
[00:53:32] <jdlrobson>	 i just did
[00:53:37] <jdlrobson>	 i'm gonna nuke by node_modules just in case
[00:53:38] <Krinkle>	 rm -rf node_modules and try again
[00:53:40] <Krinkle>	 Yeah
[00:53:51] <Krinkle>	 npm doesn't support long survived npm modules dirs
[00:53:54] <Krinkle>	 it's a known upstream issue
[00:54:06] <Krinkle>	 They break it every few weeks in some unforeseen way
[00:56:56] <jdlrobson>	 Krinkle: any chance the module loading work will get merged this week?
[01:09:59] <Krinkle>	 jdlrobson: I hope so yes.
[01:10:52] <Krinkle>	 jdlrobson: But in general I'm trying to align performance work more with quarterly goals. E.g. adding new bits mid-quarter without prior communication is going to be harder to prioritize with less notice. I know you have new tickets rise within each sprint but that's not entirely compatible with there are dependencies outside the team.
[01:11:06] <Krinkle>	 But for now it's fine as I have time set aside for reading.
[01:11:21] <jdlrobson>	 yeh i completely understand. TBH i'm disappointed the rest of the frontend standards team isn't interested in that patch
[01:11:43] <bmansurov>	 jhobs: hey, is zero using the 'mobile.overlays/LanguageOverlay' module, do you know?
[01:12:44] <jdlrobson>	 bmansurov:  any chance of a merge on https://gerrit.wikimedia.org/r/265779 ?
[01:12:59] <jdlrobson>	 bmansurov: i haven't heard much from jhobs and haven't seen any patches fly back so i think he's gone for the day
[01:13:09] <bmansurov>	 ok, i'm reviewing your patch
[01:13:19] <bmansurov>	 ok on jhobs' status too
[01:13:29] <jdlrobson>	 bmansurov: just keen to see that JS drop slightly in next deploy
[01:13:49] <bmansurov>	 yes, that'd be good
[01:13:51] <jdlrobson>	 it's mostly documentation updates
[01:17:41] <jdlrobson>	 wrapping up for day
[01:17:44] <jdlrobson>	 anything you need bmansurov ?
[01:18:02] <bmansurov>	 no, I'm good, thanks
[01:18:29] <bmansurov>	 have a good one
[01:20:02] <bd808>	 bmansurov: I put that config patch for MFE up for swat tomorrow morning with your name and mine on it
[01:20:19] <bmansurov>	 bd808: cool, thanks, i'll be online
[01:20:20] <bd808>	 not sure I know what to test other than en.m still works after
[01:20:40] <bmansurov>	 bd808: I'll try to look into it too before then
[01:21:13] <bd808>	 I was going to sneak it out tonight but swat is taking a while due to the first full scap from mira being involved :/
[01:21:27] <bmansurov>	 I see
[01:23:10] <bmansurov>	 bd808: you're right, I don't think we can test anything with that change. The mobile upload feature was removed some time ago if I'm not mistaken.
[01:23:18] <jdlrobson>	 see ya bmansurov -btw you can test the zero thing via jenkins - zero tests run via jenkins and if nothing complains that's zero's fault :) (also you can do a simple grep if you need to of ZeroBanner extension)
[01:23:41] <jdlrobson>	 bd808: bmansurov just make sure Special:Uploads doesn't appear in the left menu
[01:23:44] <bmansurov>	 jdlrobson: cool
[01:23:51] <jdlrobson>	 and no upload icon on articles and you are golden
[01:24:12] <jdlrobson>	 right cya all - off for a swim
[01:24:22] <bmansurov>	 later
[03:00:44] <josephine_l>	 Hi niedzielski :)
[03:01:32] <niedzielski>	 o/ josephine_l
[03:03:44] <josephine_l>	 niedzielski, I was hoping to ask about what we should do next. But I think that might have to wait for Nicolas to arrive? 
[03:04:51] <niedzielski>	 josephine_l: yeah, i was actually wondering if he wanted you to pick up one of the enhancements or issues in github
[03:05:15] <josephine_l>	 niedzielski, there's still the optional task left - https://phabricator.wikimedia.org/T119288
[03:05:28] <niedzielski>	 josephine_l: i wasn't sure if he had kept adding those just for his own tracking or not
[03:06:05] <niedzielski>	 josephine_l: is there a specific task or refactoring that you were particularly fond of?
[03:06:49] <josephine_l>	 niedzielski, I think I could work on the optional task that I linked above, since it was part of the original plan but just 'optional' (probably as a buffer in case p1 or p2 took longer than expected)
[03:07:24] <niedzielski>	 josephine_l: that sounds good to me. no one can fault you for working on an optional part of the plan :)
[03:07:55] <josephine_l>	 niedzielski, haha, sounds good. :) I suppose when Nicolas arrives I could ask him if he had anything else in mind to be done first.
[03:08:25] <josephine_l>	 niedzielski, if we go the optional task route, do you think it's better to enable or disable the GPS preference by default?
[03:09:51] <niedzielski>	 josephine_l: i think i would leave it disabled. there's some neat things you can do with location awareness without even turning the gps on called passive location.
[03:10:11] <niedzielski>	 josephine_l: there's an old blog post from retro meier that's still relevant on the topic http://android-developers.blogspot.com/2011/06/deep-dive-into-location.html
[03:10:20] <josephine_l>	 niedzielski, oh? Cool, will read that up :)
[03:10:51] <niedzielski>	 josephine_l: i think there's a repo for the code snippets to somewhere (maybe on google code :(). if you can't find it, let me know. i might have a copy floating around here
[03:11:09] <josephine_l>	 niedzielski, so if we do use that, you think we don't need to put in an extra preference for users to disable/enable it?
[03:14:05] <niedzielski>	 josephine_l: i think you still want a preference whether you use just gps or gps + passive listener
[03:14:22] <niedzielski>	 josephine_l: some folks really prefer to maintain their location privacy
[03:14:28] <niedzielski>	 in my opinion :)
[03:14:31] <josephine_l>	 niedzielski, ah, okay, gotcha. :)
[03:16:06] <josephine_l>	 niedzielski, I'll read the article you linked later and see if I can implement it, then. I think we only want the app to do that (use current location) if the picture itself has no geolocation info, right? So we bypass it entirely if the picture does
[03:17:06] <niedzielski>	 josephine_l: that's my vote unless we want to expose a button to the user that says something like "use current location"
[03:17:31] <josephine_l>	 niedzielski, aight, thanks.
[03:18:05] <josephine_l>	 niedzielski, also, do you think I should message all those mailing lists telling them that the p2 version of the app is out, or would it be a bit spammy?
[03:18:19] <josephine_l>	 niedzielski, since I did mail all of them about p1 only a few weeks ago, heh.
[03:19:01] <niedzielski>	 josephine_l: i think i'd wait a bit since presumably most of the folks who downloaded it will get the update automatically
[03:19:58] <josephine_l>	 niedzielski, ah, yeah, good point. So just leave it til the end of the program?
[03:20:33] <niedzielski>	 josephine_l: maybe not the end but the last couple weeks in case any bugs come up
[03:21:38] <josephine_l>	 niedzielski, sounds good. Hopefully I would've been able to implement the optional task by then, too.
[03:22:02] <niedzielski>	 :)
[03:22:55] <josephine_l>	 niedzielski, ah, not strictly program-related.... but I was hoping to continue working on the app after the internship is over. Maybe via an IEG grant application.
[03:23:05] <josephine_l>	 Do you think that might be feasible?
[03:23:25] <josephine_l>	 niedzielski, I've looked through the IEG rules and such and it seems like it might qualify in theory, but I might be missing something.
[03:23:27] <niedzielski>	 josephine_l: hm, i don't know anything about IEG grants. maybe quim would know?
[03:23:56] <josephine_l>	 niedzielski, ah, okay, thanks. I'll have a chat with Quim then.
[03:27:30] <josephine_l>	 niedzielski, hmm, I've messaged Nicolas on Google Hangouts, but I guess he's busy. I figure I'll just wait here for another 10 min or so in case he shows up?
[03:28:31] <niedzielski>	 josephine_l: yeah, if he can't make it, we can just follow up on email with the task
[03:31:51] <josephine_l>	 niedzielski, sounds good. Anything you wanted to discuss on your end? :)
[03:33:25] <niedzielski>	 josephine_l: no, but I'll be happy to check out the new GPS PR when it comes out :)
[03:34:23] <josephine_l>	 niedzielski, great, thanks. :) I've learnt a lot from your feedback so far.
[03:36:19] <niedzielski>	 josephine_l: woo! :)
[03:39:29] <josephine_l>	 I'll follow up with Nicolas on the Phab task. Cya niedzielski-afk :)
[05:51:22] <nico_>	 Hi josephine_l niedzielski-afk I forgot my phone at home today so without tethering I could not attend the meeting this morning sorry :-/ The selected task is great and a good way to further take advantage of the implemented features. "Touch the real location" would also be great if time allows, for instance when the subject is a bit for away from the camera (for instance when shooting the Eiffel Tower from a kilometer away, the lo
[05:51:22] <nico_>	 cation should ideally be the location of the tower, not of the shooting point)
[06:00:25] <josephine_l>	 Hi nico_ , no problems. :) I'll start on that ASAP.
[13:18:16] <phuedx>	 joakino: i've queued up a config patch to be swatted this afternoon
[13:18:34] <joakino>	 👍
[13:18:37] <phuedx>	 i might enqueue 267813 as it's trivial and we'll know if the window is open at the time
[13:18:47] <joakino>	 sure
[13:18:59] <joakino>	 i just wrote it down for standup report
[13:19:15] <joakino>	 i thought we shouldn't need a window for that?
[13:20:44] <phuedx>	 we shouldn't
[13:20:49] <phuedx>	 actually, i'll ask during the window
[15:24:30] <mdholloway>	 dbrant: hey, question for you about the w0 eventlogging schema here: https://meta.wikimedia.org/wiki/Schema:WikipediaZeroUsage  ... it looks like they want some actions timed and not others, so i'm debating the best way to implement that. 
[15:24:54] <mdholloway>	 dbrant: (1) make TimedFunnel timing optional, or (2) just use two separate funnels, one timed and one not.
[15:25:33] <mdholloway>	 dbrant: or (3) see if they actually don't mind getting times for all actions.
[15:27:14] <dbrant>	 mdholloway: yes, I was just about to suggest (3). Definitely use a TimedFunnel, and send the "time" field unconditionally. I'm guessing they won't mind ignoring it in whatever queries they'll make.
[15:27:51] <mdholloway>	 dbrant: ok, great!  makes life easier as well. 
[16:27:14] <thedj>	 ps, since yesterday i played with the ios beta app a bit. I think i already found 5 bugs :)
[16:27:19] <thedj>	 i'll file reports later today.
[17:01:57] <mobrovac>	 bearND: mdholloway: there's a hold up on ops' end, we'll need to wait
[17:02:13] <mdholloway>	 mobrovac: ok
[17:02:20] <bearND>	 ok
[17:02:55] <mdholloway>	 bearND: would you like to do the deployment when the time comes, or should i?
[17:05:34] <bearND>	 mdholloway: no sure yet. I'd like to get node 4.2.1 installed on my VM if I do it. Do you already have that one installed?
[17:06:05] <mdholloway>	 bearND: i have 4.2.4
[17:06:32] <bearND>	 mdholloway: i guess that's probably close enough. right mobrovac ?
[17:06:49] <mobrovac>	 that's the exact prod version
[17:06:51] <mobrovac>	 4.2.4
[17:07:12] <bearND>	 mobrovac: mdholloway: great! then I shall get that one, too.
[17:08:57] <bearND>	 mobrovac: ok, i tried but I got 4.2.6 following your instructions
[17:09:44] <bearND>	 mdholloway: maybe it's best if you do the prep of the server package this time. I could do the rest if you'd like
[17:13:15] <mdholloway>	 bearND: ok!  i'll build the deploy repo here and +2 in gerrit when the time comes, and you can do the rest.  
[17:13:30] <mdholloway>	 bearND: one thing to mention, in case you didn't see the email about this, is that you'll want to connect to deployment.eqiad.wmnet to do the deployment, rather than connecting to tin.
[17:13:51] <bearND>	 mdholloway: why not tin anymore?
[17:14:57] <Krenair>	 it was moved to mira
[17:15:37] <Krenair>	 while tin was being reimaged
[17:15:38] <mdholloway>	 bearND: will forward the email.  short answer is that tin is getting a software upgrade. 
[17:15:57] <bearND>	 ah, ok. thanks mdholloway and Krenair 
[17:16:00] <Krenair>	 you should be getting an email about list subscriptions while having deployment access later
[17:17:32] <mdholloway>	 bearND: as i understand it, it'll go back to tin later but it's better to connect to the service name deployment.eqiad.wmnet in any case
[17:22:00] <mobrovac>	 bearND: 4.2.6 is ok as well
[17:22:17] <bearND>	 mobrovac: k, thanks
[17:22:46] <mobrovac>	 correct mdholloway, connecting to tin is a no-go currently
[17:34:47] <mobrovac>	 bearND: mdholloway: puppetswat moved :/
[17:34:53] <mobrovac>	 no deploy today, sorry
[17:34:56] <mobrovac>	 moved to tomorrow
[17:35:18] <mobrovac>	 there are issues with syncing mediawiki, so ops are on it
[17:35:27] <mdholloway>	 mobrovac: ok, thanks for the heads-up.
[17:35:56] <bearND>	 mobrovac: thanks. good to know
[17:36:09] <mobrovac>	 yeah, sorry about that guys
[17:37:33] <bearND>	 mobrovac: any idea when that will happen? I don't see the deployments calendar updated yet
[17:38:01] <mobrovac>	 bearND: they should update it soon, but my guess is same time as today
[17:38:09] <bearND>	 k
[17:38:43] <mobrovac>	 bearND: mdholloway: i updated appservice.wmflabs.org to node 4.2
[17:39:00] <bearND>	 mobrovac: cool. thanks!
[17:39:05] <mdholloway>	 mobrovac: thanks!
[18:30:37] <joakino>	 i'm back
[19:47:39] <jdlrobson>	 dr0ptp4kt: can you reply to https://phabricator.wikimedia.org/T125278#1990237 ? as am keen to get it merged
[19:54:28] <phuedx>	 later
[19:55:26] <jdlrobson>	 dr0ptp4kt: also https://phabricator.wikimedia.org/T124511#1988122
[23:16:22] * niedzielski just discovered a 👍 comment causes a Gerrit internal server error
[23:37:54] <jdlrobson>	 bd808: able to advise about whether I can SWAT a config change to beta cluster?
[23:39:07] <mobrovac>	 jdlrobson: afaik, no swat is needed for beta
[23:39:30] <jdlrobson>	 mobrovac: yeh that's what i thought but given the issues today i wasn't sure if beta cluster merges were off the table
[23:39:59] <mobrovac>	 don't think so, no
[23:40:00] <jdlrobson>	 I want to merge https://gerrit.wikimedia.org/r/267812
[23:40:04] <mobrovac>	 greg-g: ^ ?
[23:40:07] <jdlrobson>	 I don't have +2 though :/
[23:40:48] <greg-g>	 changes to BC are fine
[23:41:23] <mobrovac>	 changes BeforeChrist are fine
[23:41:28] <greg-g>	 :)
[23:41:54] <greg-g>	 jdlrobson: worst case it'll go out tomorrow morning (hopefully, god forbid we see more fuckups)
[23:42:18] <greg-g>	 (I have no +2 either, fwiw)
[23:42:48] <jdlrobson>	 bd808: any chance you could merge the above for me? I'm keen to make a start on this since it's time sensitive
[23:43:47] <bd808>	 greg-g: it's a prod no-op but I would need to sync-file it to make the universe happy
[23:44:04] <thedj>	 oh right. ios team. I'll just create tickets for the testflight issues i encountered, but some might be doubles. please forgive. :)
[23:44:21] <bd808>	 jdlrobson: If I merge it without fetching and syncing Roan will find me an beat me with a rubber hose
[23:46:02] <mobrovac>	 bearND: mdholloway|afk: time's set for puppetswat for tomorrow for same time as today (17:00 utc)
[23:46:28] <bearND>	 mobrovac: thanks
[23:46:49] <jdlrobson>	 :-/
[23:46:55] <jdlrobson>	 today has not been a productive day
[23:47:16] <greg-g>	 oh right...
[23:47:25] <greg-g>	 le'ts not and wait until tomorrow jdlrobson bd808 
[23:49:49] <bd808>	 jdlrobson: could have been worse. _joe_ started his day at 08:00 local time and finished about 01:00 local time
[23:51:00] <jdlrobson>	 poor _joe_