[00:08:58] Maryana: try mailing now. i've been talking with joseph and we just got a message through [00:09:23] i see messages! [00:11:55] ok, looks like we're back up & running [00:13:15] [Commons-iOS] brion pushed 2 new commits to master: http://git.io/RWhbbA [00:13:15] Commons-iOS/master af25d7e Brion Vibber: toolbar hack... [00:13:15] Commons-iOS/master b391b1b Brion Vibber: quick hack to fix broken selections [00:21:39] Maryana: yup, were all good [00:21:43] i worked it out with joseph [00:22:02] sweet. thank you! [00:28:28] tfinc what was the problem with the list? [00:30:26] bah. webm quicktime encoder is single-threaded. waste of coooores [00:33:17] awjr: chip will respond and tell us [00:33:21] word [00:57:54] New review: awjrichards; "Patch Set 1: Verified+2 Code-Review+2" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/48366 [00:57:55] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/48366 [01:09:10] New review: awjrichards; "Patch Set 3: Code-Review-1" [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/48475 [01:16:14] New review: awjrichards; "Patch Set 1: Code-Review-1" [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/48583 [01:56:31] New patchset: JGonera; "Fix photo upload bug when the same image selected twice" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/49195 [01:59:08] New review: JGonera; "Patch Set 6: Code-Review-1" [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/43002 [08:28:29] New review: MaxSem; "Patch Set 1:" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/48583 [11:42:47] New patchset: Krinkle; "Get rid of home-made RL links generation" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/49076 [12:51:18] Project Android-Commons (mobile) - Nightly builds build #56: SUCCESS in 46 sec: https://integration.mediawiki.org/ci/job/Android-Commons%20(mobile)%20-%20Nightly%20builds/56/ [12:51:18] yuvipanda: Initial cut of EventLogging support [15:28:42] notnarayan: how do you like the updates so far on iOS? :D [15:29:03] brion: saw it today morning and i must tell you it made my day. :) [15:29:11] :D [15:29:14] iv been showing it to everybody and his brother [15:29:24] hehe nice [15:30:10] brion: the upload icon seemed a little of though. loved the transition that you gave the icons. just as i imagined them to be [15:30:56] i'm thinking of getting rid of the 'upload' button on the main screen [15:31:09] maybe integrate it into the queued items [15:31:39] brion: i dint understand the upload button actually. [15:32:09] well i could switch it to start uploading in the background... [15:32:20] …but if we cancel or network dies, we need to be able to restart [15:32:29] probably tapping on the items should be able to do that tho [15:33:17] notnarayan: also crazy idea time :D [15:33:36] when we push the add media button, should the photo & library buttons *slide out from its position*? [15:33:40] cause that would be awesome [15:33:55] yes yes [15:34:03] brion: thats what i imagined it to be like [15:34:07] :DDD [15:34:18] excellent i'll figure out how to do that :D [15:34:29] they both have their center aligned to "upload image" icon [15:34:45] on tapping they fan out and the upload image icon becomes a lil transperent [15:35:09] brion: that would be great! [15:35:12] sweeeeeeet [15:36:04] :D [15:39:14] brion: tapping anywhere after the icons fan out could send them back where they came from. after this the gallery mode is active [15:39:41] *nod* i think i can do that [15:39:51] brion: \m/ [17:46:30] YuviPanda|awaaay: is your new account all set? [18:14:25] YuviPanda|awaaay: https://upload.wikimedia.org/wikipedia/commons/c/c6/Commons_media_discovery_app.png [18:17:45] awjr: RE [Ops] Puppetizing squid confs [18:18:04] awjr: you can just use Varnish for now [18:18:22] preilly: well, the idea is to as clolsey mirror production as possible [18:18:29] awjr: you could port the redirector to Varnish as well if you'd like — that would really help us out [18:18:33] preilly, we're trying to fully reproduce production [18:18:37] i dunno if it's better to use a stripped-down squid conf or varnish though [18:18:53] MaxSem: you can't [18:18:54] i mean, we will need to port the redirector to varnish at some point anyway [18:18:56] MaxSem: well at least not easily [18:19:19] preilly: do you know what the timeline is like for switching to varnish for text? [18:19:19] hehe, I kinda know it=) [18:19:56] awjr: I'm not sure what the exact timeline is for the text squids but it isn't too far off as Mark is actively working on it [18:20:35] preilly: if the confs mark sent out work, or are close to workable, my inclination is to start with that and then work on getting things ported to varnish while mark is working on it [18:20:54] preilly, I heard approximately the same during the all-staff;) [18:21:00] that is under the assumption it'll be faster to get up and running with squid [18:21:21] yeah the squid -> varnish migration has been fabled for a few years now [18:21:32] awjr: I don't really understand the beta labs set-up [18:22:27] awjr: and frankly if it's in our labs it's technically impossible to replicate production as the networking in labs doesn't support LVS correctly [18:22:34] totally [18:22:38] awjr: so I guess I've got no real opinion [18:23:03] yeah - preilly i'll make sure to pitch your suggestion to the folks working on this and the team can make the call [18:23:27] awjr: okay thanks [18:23:35] nono sir, thank you [18:24:27] * preilly — OMG awjr just called me sir, now I feel really old [18:24:40] jorm: are you in the office today? [18:24:45] heh [18:30:08] hello [18:30:14] i'm here, yes. [18:30:37] thank god, i need your help for mobile fronted [18:31:27] the error is when i visit to my wiki main page through desktop it is of mobile version [18:32:13] jorm, are you familiar with these types of problem [18:32:26] i'm sorry, no. i was answering someone else's question from earlier. [18:32:46] i am trying to install mobile fronted [18:32:47] awjr brion preilly may be able to help you. [18:33:10] are they here [18:33:19] yes, but they may be idle/lurking. [18:33:32] so give them some time to see this and answer. [18:33:48] also maxsem may be able to help [18:34:38] buddy if they donot help me then can i again consult you. [18:34:49] sure, what's the problem vp__ [18:35:26] Thanks maxsem, the problem is when i visit to main page through desktop it is of mobile version [18:36:00] in more than one browser? [18:36:07] [Commons-iOS] brion pushed 1 new commit to master: http://git.io/4CSo-A [18:36:07] Commons-iOS/master b3587aa Brion Vibber: some more localization [18:36:41] yep [18:37:10] how did you install MobileFrontend? [18:37:41] and when i try to point it to a subdomain through mobilesettings template like %h1... it does not connect with indexx.php [18:38:09] i have verified the installed special version [18:38:53] * MaxSem slaps his KVIrc hard [18:39:09] [Commons-iOS] brion pushed 1 new commit to master: http://git.io/83EYnQ [18:39:09] Commons-iOS/master d1c4383 Brion Vibber: durrr [18:39:25] maxsem, check the url punjabworks.org [18:39:41] running wikiwmf9 [18:40:33] current status error of mobile bersion is that when i click on it it refershes but it is desktop version [18:40:33] it shows the desktop version to me [18:40:49] but now click on mobile view [18:40:59] http://a.punjabworks.org/w/index.php/Main_Page [18:41:58] i have edited local setting.php many times so there are different errors [18:42:01] apparently, Apache is not configured to handle this subdomain [18:42:32] awjr, now that I analyze that better, I'm not sure how we log things [18:42:42] we log HTTP response headers? [18:43:22] now listen, installation verified. and i have correctly added localsettings.php. [18:43:45] okay forgot about subdomain [18:44:03] awjr, have you seen my pesponse at https://gerrit.wikimedia.org/r/#/c/48583/1/includes/MobileContext.php ? [18:44:18] can we do that on a domain both mobile and desktop version runs [18:44:26] not yet MaxSem - will look shortly, just trying to wrap something up now [18:45:43] jgonera: that meeting is going to be canceled, asher and diederik sorted it out :) [18:45:52] jgonera: have you met diederik in person yet? [18:45:54] maxsem, i have doubt about url template settings [18:46:02] awjr, yes, I have [18:46:18] so is there any place where they sorted it out? [18:46:22] guide me to write [18:46:27] I mean, is the decision public? ;) [18:46:37] brion, jcmish & i are planning out some tasks for a community QA event on the 25th - is the apps team going to have the ability to add categories to uploads by then, or is that for future iterations? [18:46:51] oh, maybe the same thread, I still haven't reached the end [18:47:04] Maryana: i could probably rig it in… YuviPanda|awaaay ? [18:47:12] jgonera: they did it over private email, but we'll make sure it gets out probably on the wikitech thread [18:47:23] ok, good [18:47:28] maxsem, [18:47:37] brion - no rush. just making sure we have the right set of things to test :) [18:47:56] ok it might be in or might not. no guarantee but hopefully :D [18:48:03] lemme make sure it's listed on the backlog [18:48:28] ok add categories is on iteration 3 [18:48:37] so hopefully we'll have it done by then :D [18:48:49] vp__, what's your $wgMobileUrlTemplate? [18:48:54] hehe, k. i'll just flag it as a maybe in case you don't get to it [19:01:54] MaxSem: did you see the meteor? [19:02:36] maxsem buddy [19:03:54] Can we do to have a mobile and desktop view on a domain [19:04:11] maxsem are u busy [19:04:55] vp__: i might be able to help but i dont totally understand your question - can you please rephrase it? [19:06:01] i want to view mobile view and desktop view on a domain name but not on a subdomain bcoz i donot use appache [19:06:17] like http://wikimediafoundation.org [19:06:48] awjr [19:07:10] vp__: are you using any kind of caching? [19:07:25] no [19:07:51] vp__: and you probably want automatic mobile device detection? [19:08:47] awjr, and you still haven't merged automatic detection?:P [19:08:49] the problem if i do urltemplatesettings like%h0%h1 to view both [19:09:09] currently, MobileFrontend does not support automatic mobile device detection (at least not until https://gerrit.wikimedia.org/r/#/c/48583/1 gets merged, which it likely will in a few mintues ;) [19:09:09] the main page is desktop [19:09:36] vp__: if you are using a single domain (eg you do not have separate sub domains for desktop/mobile) you should not set the template variable [19:09:46] tell me more about it, ajwr [19:10:31] vp__: so the $wgMobileUrlTemplate is only used if you have separate domains. since you do not, leave it as an empty string. [19:10:44] you mean i should update and then check it [19:10:52] the link [19:11:09] vp__: give me a minute - i might be able to merge that change now :) [19:11:15] i just need to check it first [19:11:32] okay [19:12:14] MaxSem: i do not totally understand your comment in https://gerrit.wikimedia.org/r/#/c/48583/1/includes/MobileContext.php [19:12:40] MaxSem: wont that change cause isMobileDevice() to always return false in our production environment? [19:13:00] (assuming we keep $wgMFAutodetectMobileView = false) [19:13:09] awjr, our x-device detection detects various desktop browsers as various mobile ones [19:13:38] because we use URL rewrites to set useformat=mobile, it will not break WMF [19:14:06] MaxSem: are you planning to set $wgMFAutodetectMobileView = true in WMF config? [19:14:13] no [19:14:19] varnish does that [19:14:26] err, squid [19:15:27] yeah i know - but that means your change will cause isMobileDevice() to return false when it should not in prod [19:15:31] ajwr [19:15:36] MaxSem: ^^ [19:15:59] it will not - useformat has precendence [19:16:22] i am confused - where is useformat being used? [19:16:51] awjr, do you know ho mobile domains are set up? [19:17:05] i thought i did, MaxSem but apparently i am missing something [19:17:27] m.en.wikipedia.org/wiki/foo --> en.wikipedia.org/wiki/foo?useformat=mobile [19:17:38] wat [19:17:48] x-device is irrelevant for mobile detection [19:18:25] where does that happen MaxSem? the only time i'm aware of that happening is if you make a request for en.m.wikipedia.org/ [19:18:29] wait maxsem sorry to interrupt its langcode.m.wikipedia.org [19:19:03] (line 548 of mobile-frontend.inc.vcl.erb) [19:19:29] ugh [19:19:32] lol [19:19:43] I got lost in it myself [19:19:46] hehehe [19:19:53] that is very easy to do [19:20:27] i will be here but for god sake please fix it [19:22:12] ajwr got any fix [19:23:04] are u here buddies [19:23:06] vp__: we have not supported automatic detection of mobile devices without special configration and infrastructural set up (our primary focus is on this software being used on the WMF infrasturcture [that hosts Wikipedia, etc]). max hacked this in the other day on his own, but we need to make sure it will not cause problems for us in production before we can merge his changes with the codebase. please be patient while we get it taken care of. [19:24:39] okay bro, i will be here but tell it will be in today only [19:27:00] New patchset: JGonera; "Fix margins in uploads dashboard" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/49271 [19:28:03] best of luck max and awjr [19:28:17] thanks vp__ we'll let you know, hopefully it'll get merged today. [19:28:48] awesome [19:42:36] New patchset: MaxSem; "Bug 36894: Mobile device detection for third parties" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/48583 [19:42:45] awjr_lunch, , ^^^ [19:42:48] ajwr_lunch when completed, send the links to vpsinghpunjab@gmail.com if i quit. [19:43:45] ori-l, Maryana told me that event logging can fill in some fields by itself, is it documented somewhere? [19:44:36] i think maybe this? http://meta.wikimedia.org/wiki/Schema:EventQueryMeta [19:44:38] jgonera ^ [19:45:49] I was wondering if userAgent for example is something that I need to fill in myself [20:13:00] ajwr_lunch bro fixing [20:16:16] jgonera: in a meeting, will reply in a bit. [20:16:30] ori-l, ok, let me know when you have time [20:22:38] Maryana, I made some small changes to https://meta.wikimedia.org/wiki/Schema:MobileWebUploads, can you check if it's OK? [20:22:55] ajwr bud how much more time [20:22:57] also, I'm still not sure how to figure out isEditable, but I'll aks Ori after his meeting [20:22:59] * Maryana looks [20:23:54] how much more time [20:24:40] Maryana, also, I noticed we log userId for watchlist, but we don't want that for photos? [20:25:23] i waver between wanting more redundancy and less [20:25:32] technically we can dedupe users by the user token [20:25:38] but then we don't know who they actually are [20:27:04] i suppose if we're collecting userids for watchlist, we should for uploads, too [20:27:24] but in that case we should really fast-track this story: https://mingle.corp.wikimedia.org/projects/mobile/cards/422 [20:27:37] it's absolutely no difference on my end, so it's your call ;) [20:27:56] heh. yeah, it's really murky territory, legally speaking [20:28:15] yeah, I was surprised we log this for watchlist actually ;) [20:28:55] indeed - it seemed ok at the time, but now that we're adding more logging and expanding from the tiny beta userbase to the huge stable userbase... [20:29:06] getting more concerned about user privacy [20:29:30] let's start off not collecting userids. we can always add them back in [20:30:04] sure [20:31:10] hey bud sorry to distrub but are you working on my problem [20:32:51] vp__, sorry to be blunt, but you're behaving like we owe you an eternal debt and you're in a position to demand that awjr_lunch drops his food and runs to review my commit [20:33:12] don't worry - everything will be done soon:) [20:34:13] no bud. its just a wiki for charity and i have not sorted this problem for a month [20:34:45] so, i was being a bit quicker. sorry [20:35:33] i just had a big laugh [20:39:04] MaxSem: im taking a look now [20:44:53] New review: awjrichards; "Patch Set 2: Code-Review-1" [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/48583 [20:45:04] MaxSem: almost - failing unit test ^ [20:47:34] seems to work otherwise though, max :) [20:49:49] New patchset: MaxSem; "Bug 36894: Mobile device detection for third parties" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/48583 [20:50:20] does anyone know if there's an easy way of testing EventLogging on a local dev instance? I probably should have asked Jon... [20:50:33] jgonera: ori-l probably does [20:50:36] you should ask ori-l [20:51:08] he's still in a meeting I guess so it was worth trying if anyone else knows ;) [20:51:30] doh [20:51:44] sorry jgonera - i think he or jon actually showed me once but i can't remember :-/ [20:58:02] hey [20:58:31] um, so the tooling around debugging/development is not amazing, but it exists [20:59:01] there's scripts/DevServer.php, which is a CLI tool that launches a tiny HTTP server that can receive events and output them to stdout [20:59:32] New review: awjrichards; "Patch Set 3: Verified+2 Code-Review+2" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/48583 [20:59:33] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/48583 [20:59:38] MaxSem: ^^ [20:59:45] whee [20:59:54] thanks [20:59:58] vp__: that change has been merged; so if you check out the extension at the current master, you should have the new functionality [21:00:25] all you'll need to do is set $wgMFAutodetectMobileView = true; in your LocalSettings.php [21:00:51] if you ever implement a caching solution, you may need to take a different approach but from what you mentioned earlier, sounds like that should work just fine for you [21:01:02] thanks for hacking that in MaxSem, i think you just made a lot of people very happy :) [21:01:19] Thankyou very much [21:01:40] vp__: let us know how it goes [21:02:53] i am checking it [21:03:47] I got disconnected *again*. jgonera, did you get any of that? [21:03:53] re: DevServer.php? [21:04:20] ori-l, yeah, I got it [21:04:29] MaxSem: did you see the meteor? [21:04:44] will the EventLogging extension by default send events to DevServer? [21:04:48] do you really think Russia is so small?;) [21:04:51] or should I set some config options? [21:04:56] if you configure $wgEventLoggingBaseUri to be "//127.0.0.1:8080/event.gif' and then run it in the background, you'll see the events that you're logging. [21:05:16] oh, ok, didn't get this one [21:05:36] hahaha i read it was like 1000km away from moscow or something when it exploded, but figured you might still be able to see it in the sky [21:05:51] If you wait 10-15 mins I'll update it to be nicer to work with. [21:06:07] also, Maryana told me that you were using a 'isEditable' field in GettingStarted, but I'm not quite sure where to get that information from [21:06:28] I can certainly wait 15 mins although it seems to be good enough for now too [21:06:53] its work great. i wish you my blessings. [21:07:32] :) [21:07:38] jgonera, this is why e3 just hired a technical writer [21:09:24] Maryana, I still hope you won't force them to comment every line of code ;) [21:10:23] no, but every field used in a schema, hopefully! [21:10:54] ori-l, I can't seem to get it working, should I load EventLogging extension before MobileFrontend or after? and $wgEventLoggingBaseUri should be before I load it or after [21:10:55] ? [21:11:43] $wgEventLoggingBaseUri should be configured in LocalSettings.php. It is only used to construct URLs in JavaScript that the events are sent off to, so you need to ensure that it is connectable by your browser [21:11:44] New patchset: Zfilipin; "Minor wording improvement for readme file" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/49347 [21:12:15] for example, if you are running your browser on your laptop, but mediawiki in vagrant, the devserver needs to run on your host environment, or you need to forward the port. [21:12:48] but because the python library exposes a fairly high-level API, i'm thinking it'll only take me a few minutes to whip up something using the built-in HTTP server, so hang on a bit and I'll give you something easier to use :) [21:13:44] I'm not using vagrant (I had some bad experience some time ago) [21:13:55] but I just noticed that mw.eventLog in JS is undefined [21:14:52] can you compare your LocalSettings.php to jon's? [21:15:09] is Jon's LocalSettings out there somewhere? [21:15:17] in a galaxy far, far away [21:15:28] um, dunno. I thought he was around but I now notice he isn't. [21:15:37] New review: Cmcmahon; "Patch Set 1: Verified+2 Code-Review+2" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/49347 [21:15:38] probably not that far, in Costa Rica I suppose... [21:15:39] Change merged: Cmcmahon; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/49347 [21:15:45] heh [21:16:16] jgonera: let me look through my configs, sec. [21:16:46] ok [21:18:41] munaf_, I remember vaguely that you were reviewing some commits on gerrit, do you have a local instance of mediawiki on which you can test a commit? [21:19:09] jgonera i had a vagrant instance at one point and i'm setting it up again now. [21:19:52] ok, I just though that if you guys had an instance maybe this would allow faster feedback when it comes to design changes [21:20:07] yeah i'll let you know when it's up and running [21:20:11] ok [21:20:45] jgonera: betalabs should help with that :) [21:20:59] in general it might be better to deploy commits to a labs instance (or provide screenshots) for design review. i think vibha doesn't know gerrit/git/vagrant yet and it'd definitely slow her down in terms of design review [21:21:09] are we going to run the most recent master on beta labs? [21:21:16] jgonera: yeah [21:21:32] oh, ok, if it'll work like that then it's great [21:21:44] jgonera: im not sure if it's that way right now for MF, but some of the other extensions that rely on betalabs get automagically git pull'd [21:22:02] that would be awesome [21:22:11] dude i know, i cant wait [21:22:15] ;) [21:22:20] actually it might do that already - MaxSem do you know? [21:22:29] (if betalabs automatically runs latest master for MF) [21:22:34] em, I'm not sure [21:22:39] michelle would know for sure [21:22:42] what's the URL? [21:22:43] but she's not around apparently [21:22:47] ah chrismcmahon might know [21:23:06] beta.wmflabs.org [21:23:13] which currently looks hella broken. [21:24:58] heh [21:25:10] chrismcmahon: betalabs looks busted atm - any idea what's going on? [21:25:52] awjr: it's been busted all day. problem is with gluster as I understand [21:25:59] gahh [21:26:07] thanks chrismcmahon [21:26:19] fwiw, sorry [21:27:23] jgonera: http://dpaste.de/0Vqc7/raw/ [21:27:30] ori-l, I can see there's also $wgEventLoggingFile, can I also use that? [21:28:07] jgonera: the meaning of the various configurations _is_ actually documented fairly well in EventLogging.php. $wgEventLoggingFile specifies where events logged from PHP code are routed; it has no impact on JS-generated events. [21:28:44] chrismcmahon: no worries, i assume it's not your fault :p gluster seems to be a problem child. [21:28:48] yes, sorry, I wasn't clear enough, I meant, can I use it instead of the server? [21:28:51] DevServer [21:30:04] DevServer receives events via HTTP; it's unrelated to $wgEventLoggingFile, which is related to a PHP API for logging events (using UDP or file output) from within MediaWiki. [21:31:16] yes, I do understand that DevServer receives events via HTTP. I was just wondering if I have to run it to see if event logging actually works or I could just log events to a file [21:31:33] How would JavaScript write to a file? [21:31:34] either way is good though [21:32:03] oh, ok, that was stupid, I thought that the PHP layer of the extension actually sends it to the server [21:32:11] but that makes no sense, it would just add one layer more [21:32:31] Is [204 No Content] to be expected in the DevServer? [21:32:39] Yes. [21:32:40] I assume it's the response [21:32:43] ok [21:33:06] alright, I guess I have it set up [21:33:18] I was basically missing $wgMFLogEvents = true; [21:34:47] I'm sorry some of this is complex -- I tried to keep things simple but there were five hundred fiefdoms of code and infrastructure to work with and around to get this working and performant. The documentation gaps are my fault entirely, but you'd be happy to know that we got a technical writer to help and I will work with her much of my time over the next few weeks. [21:34:53] what's the $wgEventLoggingSchemaIndexUri in production? [21:35:36] sure [21:36:01] $wgEventLoggingSchemaIndexUri = 'http://meta.wikimedia.org/w/index.php'; [21:36:17] awjr: http://toro.wmflabs.org/srv/mediawiki/index.php?title=Special:UserLogin&returnto=Main+Page&type=signup [21:36:19] you can check such things in operations/mediawiki-config repo. [21:36:27] check out that sweet captcha refresh button that e3 is building [21:36:29] I appreciate the effort, because without this we would be launching features blindly, not knowing if they work or if they're useful [21:36:34] whoa [21:36:39] ok, good to know that repo [21:36:41] it's going into core... [21:36:56] cool Maryana, that's awesome [21:37:19] Maryana: the captcha thing is going into core? [21:37:31] yep, and the e3 experiments extension [21:37:35] that's cool, BUT, most of the captchas I've seen in my life already have this so don't expect the users to get equally excited ;) [21:37:44] right now i think captcha stuff is powered by another extension (ConfirmEdit i believe) [21:37:48] hehe. yeah, but the mobile one doesn't, currently [21:37:59] sweet Maryana [21:37:59] true [21:38:04] we'll change that ;) [21:38:08] :D [21:38:26] i hate captchas and i hate the captchas we use even more [21:38:47] they are particularly evil on mobile devices [21:38:55] yeah. sigh. [21:39:00] awjr: +1,000,000,000 [21:39:03] at least the captcha generation we use often comes up with great word combos [21:39:19] ori-l, I'm still getting Unknown schema: MobileBetaWatchlist after setting $wgEventLoggingSchemaIndexUri [21:39:23] I thought it would go away [21:39:29] for instance, i just got 'bogusbooze' [21:39:41] haha [21:39:46] niiiice [21:40:01] you know that's how oliver keyes got his username on enwiki, right? [21:40:07] his captcha was ironholds [21:40:16] hehe [21:40:17] lol [21:40:19] awesome [21:40:27] i got this one a few days ago: http://i.imgur.com/kzO9JkO.png [21:40:39] ! [21:40:45] … i lol'ed [21:40:48] nobody's updated the captcha generator in years, i'm sure it could use some love [21:40:54] hah [21:40:58] awjr: don't blame me; blame the captcha [21:41:11] hehehehe no blame, only praise [21:41:39] nobody's updated the captcha generator in years, i'm sure it could use some love [21:41:41] *crickets* [21:41:42] *silence* [21:41:45] * ori-l coughs. [21:42:14] * brion awkward silence [21:42:18] haha [21:42:23] ooooh look http://arstechnica.com/gadgets/2013/02/ubuntu-phone-developer-preview-will-arrive-on-february-21st/ [21:42:25] that's like asking who wants to fix centralauth [21:42:49] centralauth was the bane of my existence for like 2 years. i don't ever want to touch it again ;) [21:43:01] hahaha [21:43:18] ooo [21:43:31] i wish you could dualboot your phone [21:43:33] searching amazon on mobile has never been better [21:43:37] * ori-l ducks. [21:43:47] * MaxSem touches brion with CentralAuth [21:43:54] lol [21:43:55] get it offfffff meeeeeee [21:44:06] ori-l, can I still bother you with a couple of questions? ;) [21:44:25] jgonera: I think I should come downstairs and try to look at this with you [21:44:40] ok, although I think I have it almost figured out [21:44:50] or I can come upstairs since it's me having a problem ;) [21:45:03] and reconcile myself to the fact that EventLogging constitutes an oral tradition [21:45:16] heh [21:45:36] ok, I'm coming up [21:45:44] cool. [21:46:00] * ori-l hides in the office supply closet. [21:46:21] ;) [21:53:22] hmm, https://graphite.wikimedia.org/render/?target=alias(color(dashed(drawAsInfinite(deploy.sync-common-file)),%22c0c0c080%22),%22sync-common-file%22)&target=alias(lineWidth(color(drawAsInfinite(deploy.sync-common-all),%22gold%22),2),%22sync-common-all%22)&target=alias(lineWidth(color(drawAsInfinite(deploy.scap),%22white%22),2),%22scap%20deploy%22)&title=en.m.wikipedia.org%20Pageview%20Request%20L [21:54:07] our emergency fixage indeed improved the situation [22:29:02] New patchset: awjrichards; "(mingle 397) Append configurable string to photo upload description" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/49354 [22:33:29] New patchset: MaxSem; "Display thumbnails in watchlist using PHP" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/48475 [22:33:30] jgonera, can you take a look why images don't have enough real estate in ^^^? [22:33:58] MaxSem, sure, after I'm done eating [22:34:06] thanks [22:41:10] no Yuvi, no brion.... meh [23:20:43] MaxSem, is there any good reason why this https://gerrit.wikimedia.org/r/#/c/48060/ is not in our codebase? [23:22:17] jgonera, no idea [23:23:06] awjr, maybe you remember what was the reason behind it? [23:26:02] jgonera im not sure exactly why jon did that, but i presume it was to make it easier for upkeep/avoid creating a depdencny [23:26:06] *dependency [23:26:46] i think it was a good idea - it would be annoying to have to update MF code whenever the schema changes, and folks who might be using event logging + mobilefrontend might not want to be using that particular schema [23:27:08] I talked with ori-l about this, and we concluded that we should probably move it back to our code and remove the $wmgUseEventLogging check [23:27:23] hm [23:27:36] but it's hardcoded in other places anyway [23:27:41] in JS files [23:27:52] why do you guys want to move it back? [23:28:39] it seems weird to have this one bit in other place and all the other event logging things in our codebase [23:28:58] what are the others jgonera? [23:29:09] I agree that we should decouple it somehow, but currently it's not decoupled anyway [23:29:31] yah - what are the others? i grep'd for 'schema' and didnt come up with anything [23:29:38] M.log() function, then logWatchEvent() function in mf-watchlist.js [23:30:31] so my reasoning is decoupled and in two different places => good, not decoupled and in two different places => confusing [23:30:53] yeah agreed [23:31:28] the logWatchEvent() function isn't necessarily eventlog specific though [23:31:38] jgonera, MF is a generic extension, while this monitoring if extremely WMF-specific [23:32:20] MaxSem, I know, but we still have this WMF-specific logging in our code, no matter where we put the schema module definition [23:32:45] well the only wmf-specific thing is really the mw.eventLog check [23:32:48] because we're sending the data in logWatchEvent() following the schema anyway [23:33:22] oh i take that back, 'schema name' :-/ [23:33:29] well, not really, the logWatchEvent, although doesn't contain any programmatic references to WMF stuff, follows the schema [23:33:35] yeah [23:33:36] yeah [23:33:37] jinx [23:34:12] ok, let's keep it the way it is for now, we can get back to it when Jon's back [23:34:14] i guess you could abstract it and define external loggers along with formats [23:34:22] but that is probably overly complicated for our needs right now [23:34:26] I'll just clone the operations repo and add the uploads schema there too [23:34:56] yeah, abstracting that might be not worth the effort [23:35:32] I'm wondering if we should just tell people who use mediawiki to either use the eventlogging with our schemas if they want to and that's it [23:35:36] uh, what a mess ;) [23:35:42] heh [23:36:01] * either use it with our schemas or not use it [23:36:03] let's cross that bridge when and if we ever get to it ;) [23:37:01] ok, but if we're not abstracting that then I'd opt for moving the schema modules back to our code ;) [23:37:03] our primary mission is to support the software in production on the WMF architecture - making it generic/supporting for 3rd party use is totally secondary [23:37:14] I see [23:37:43] i disagree on that one jgonera, because updating the config is easier than having to update code (if/when the schemas change) [23:38:14] in the end it's not a huge deal either way, i guess [23:38:23] when would we have a situation where we update a schema without the need of updating the code that uses it? [23:38:34] (which is most often JS) [23:38:36] dunno [23:38:51] i'd have to look at the schemas i guess, but maybe never [23:39:19] ok, let's leave this discussion for some other day ;) [23:39:29] ok how about this - needing to roll back to a previous schema (when maybe the only thing that's ver changed was adding new params)? [23:39:35] heh [23:39:53] i guess my preference is config but it probably doesn't really matter as long as we can keep it all in one place [23:40:02] I mean, look at the logWatchEvent() function, if anything changes in the schema, this function also has to change [23:41:33] meh [23:42:15] jgonera, awjr: I was away earlier, but https://gerrit.wikimedia.org/r/#/c/49360/ [23:42:32] I have to step out for a bit, but let me know if you have a question for me. [23:42:47] The change I linked to above is designed to help express soft dependencies in JS. [23:42:51] MaxSem may have some opinions, but i dont think i really care as long as it's all in one place. [23:42:56] jgonera: ^ [23:43:11] cool ori-l, i'll take a look in a bit [23:43:45] I'll have a look