[00:13:03] "You're holding Tim Starling's opinion too high. His opinion is why only now we're getting Lua script instead of years ago and the $100,000+/year in unnecessary hardware that's cost us." [00:13:08] http://commons.wikimedia.org/wiki/Commons:Bots/Work_requests [00:16:41] heh I just ignored that part [00:17:22] now I'll see if TimStarling responds on that page with a snarky comment ;) [00:18:04] AaronSchulz: troll [00:18:04] yeah I'm so anti-lua [00:18:09] I was dragged into the lua age kicking and screaming [00:21:26] The hell is that $100,000+/year claim? [00:21:42] * robla shrugs [00:22:50] Probably some unsubstantiated figure for how much we spent on additional servers and operating costs for running templates in wikitext [00:27:31] if you add in all the zhwiki pain we've had, i'd say it's more than that ;) [00:28:23] RoanKattouw: yeah, it's probably more like $93,009,203/yr [00:29:21] Right, we've totally spent 5x our annual budget on this crap annually :) [00:32:16] of course [03:59:26] TimStarling: http://www.mediawiki.org/wiki/Special:ExtensionDistributor/ConfirmEdit seems to be formatted wrong [04:00:32] you're saying that the biggest thing wrong with ExtensionDistributor is the formatting? [04:00:46] usually it is completely broken [04:02:42] Well, it also fails to download ConfirmEdit [04:05:56] right, so formatted wrong *and* completely broken [04:09:46] well, not completely, this problem seems to be that extension and only a few others [09:51:18] thedj[work]: how about https://commons.wikimedia.org/wiki/File:Figure_D-6_Ominidirectional_antenna_patter_%28FM_7-93_1995%29.gif ? [09:53:21] or https://commons.wikimedia.org/wiki/File:Figure_2-8_Enemy_intercept_and_direction_finding_%28FM_79-3_1995%29.gif [10:10:42] Nemo_bis: looks like those images are broken and that imagemagick became more strict. [10:11:47] seems they are impartial downloads of the originals or something.... [10:12:58] thedj[work]: really? they render completely on firefox for me [10:13:13] you can clearly see the 'interlacing' at the bottom of the originals, which is typical for how gif images are stored (interlaced). [10:13:13] unlike some other images which give similar errors (maybe I'm not looking closely enough) [10:13:21] ah [10:15:30] Nemo_bis: actually, if you look with firefox, you see that the bottom is 'vague'. Some interpolation has happened to fill in the missing lines. [10:15:50] in safari and chrome, i see this. [10:16:12] http://imagebin.org/233206 [10:17:32] http://imagebin.org/233207 [10:17:42] oh how smart firefox is [10:17:42] That's firefox for the other image, it's corrupted there. [10:18:02] yes, I got it when you mentioned the interlacing ^^' thanks [10:18:30] it seems to recover in firefox, but in Safari that image just stops rendering after that line. [10:20:19] please add the info to the relevat tickets in bugzilla, might help others with their reports. [10:21:38] thedj[work]: I've no idea what the relevant ticket might be :) [10:21:43] what about https://commons.wikimedia.org/wiki/File:IWH_Blau.tif ? [10:27:00] Nemo_bis: that might be a bug. we use identify to get some metadata about the file i believe and it seems that the warnings make it fail. It should ignore the warnings though. [10:27:00] please file a ticket for that. [10:27:21] thedj[work]: ok [11:26:43] NoScript complains about XSS here O_o https://upload.wikimedia.org/wikipedia/commons/f/f7/Lower_Big_Spring,_by_Thum,_M._%20Mandeville%20.png#0445619312902642286 [12:30:08] Hallo there wikimedian: Wikipedia has a disambiguation mechanism to resolve name clashes and url generation , Anybody know how it works on a database level... [13:39:01] anyone know why you cant move subpages with the parent page in the mediawiki namespace on meta? [14:33:56] Seddon I believe it's a configuration setting. [15:44:43] Seddon: it's because that is a namespace that officially doesn't have subpages. (just as File) [16:05:31] the gerrit downtime is scheduled, right? [16:05:57] when is it going to be back up? (approximately) [16:06:12] <^demon> Yes, it's scheduled. Shouldn't take more than about 20-30m. [16:06:25] <^demon> We already did the slave. [16:07:21] thanks [17:51:30] Reedy: pleeease, can you fix https://bugzilla.wikimedia.org/show_bug.cgi?id=40285 by deploying the config change? :) [17:53:17] spagewmf, mlitn: what's the status on the clicktracking changes in AFT. Are they ready for review and merging or are you guys still discussing? [17:55:43] kaldari: almost ready - I've sollicited feedback from Ori, who should be on his way to office now [17:56:48] Nemo_bis: I'll take a look [17:57:32] kaldari: did you mean my bug or (tab failed so) Niklas' clicktracking needs [17:58:58] Nemo_bis: the logo bug [17:59:25] Nemo_bis: Looks like the wo logo is bad... [17:59:42] Compare http://upload.wikimedia.org/wikipedia/commons/4/41/Wikipedia-logo-v2-wo.png to http://upload.wikimedia.org/wikipedia/en/b/bc/Wiki.png [17:59:42] oh thanks for looking [18:00:13] it's been scaled down for some reason (and is a bit blurry) [18:00:33] hm indeed [18:00:45] it doesn't even look like the same graphic [18:02:22] it's one of the few svg logos not produced by an expert user [18:02:59] it's still better than the old one... [18:03:10] Nemo_bis: might want to see if we can get someone to do a better job on that one. Shouldn't be too difficult [18:03:17] we can remove it [18:03:29] and do it later together with the other missing ones [18:04:05] kaldari: have you checked the others too or do you want me to check them again? [18:04:47] I haven't check them all yet. If you could compare them with the standard and let me know if any others are weird, that would help. [18:05:03] I can throw it in with my deploys later today if no one has any objections [18:05:38] kaldari: ok, nice; now I have to get dinner, then I'll re-check the logos [18:05:41] and amend the patch [18:05:50] thanks [18:46:36] mlitn: I'm going to try to squeeze in a CentralNotice deploy for fundraising during the window, but don't want to conflict with your deployment. Should I go ahead and start mine now or wait until you're done? [18:47:34] kaldari: you can go ahead [19:02:41] mlitn I'm trying to update the test server piramido to your patch, but if you're waiting I can +2 [19:03:11] if you'd like to test once more first, go ahead [19:03:50] mlitn did you see my comment about mw.user.id() and logged-in users? [19:06:16] I did, it's ok to have the usernames for logged in users :) [19:06:29] mlitn I filled out the form and one of the articlefeedbackvt5 calls it made was "found=1&comment=Testing+API+calls.&action=articlefeedbackv5&format=json&anontoken=Spage&pageid=6&revid=82&bucket=6&cta=1&experiment=6X&link=X" [19:07:11] So long as Dario knows that anontoken isn't always anonymous :) [19:08:12] I guess he does, although that shouldn't even matter since the goal is to test the anonymous users ^^ [19:08:12] gonna let him know about it just once more though [19:09:52] on piramido logged in and doing feedback on Sandbox page it said "Did you know you can edit this page", I clicked [Edit], and went to 404 page http://piramido.wmflabs.org/wiki/&action=edit . Reproducing... [19:10:06] spagewmf: that one call is the one to AFT api (to save the feedback content), all the other calls should be CT [19:12:05] yes, the focus is on anons only [19:12:39] we'll have to clean up the log and remove the few events coming from registered users [19:12:51] oh, he's here as well :) [19:12:54] howdy [19:12:57] he's everywhere [19:13:12] Yup on piramido.wmflabs.org , after providing feedback the [Edit this page] links to http://piramido.wmflabs.org/wiki/&action=edit , a 404. I assume it's a Piramido-only misconfiguration [19:16:30] mlitn, is ee-prototype.wmflabs.org configured with AFTv5? Logged in , I look at http://ee-prototype.wmflabs.org/wiki/Test and there's no AFT. [19:18:39] spagewmf: ee-prototype is not up-to-date [19:18:53] but the edit-thingy was a bug, just fixed it in the patchset [19:19:49] mlitn the best kind of bug, let me pull down new patch set. [19:24:59] I go to [Edit page] fine. Is the last CT call ext.articleFeedbackv5@10-option6X-submit_success ? [19:25:27] "A new wiki was created by demon at Thu, 25 Oct 2012 15:44:32 GMT for a Wikidata in English (en)." ^demon, is going to pop up within the next hours? :P [19:25:58] <^demon> hoo: Oh, did that get logged somewhere? I was testing. [19:25:59] <^demon> It wasn't setup yet :) [19:26:29] Thought so... just curios :) [19:26:44] hoo: what? the new projects list work? [19:26:48] *works [19:26:58] Nemo_bis: Sure, it always did... [19:26:59] orly [19:27:04] not quite [19:27:08] huh? [19:27:16] anyway, ncie to hear :D [19:27:50] spagewmf: if you clicked the button to edit page, you should have a button_click, is it there? [19:28:15] "wikidatawiki" is even on http://noc.wikimedia.org/conf/all.dblist :P [19:28:16] <^demon> hoo: It's in progress :) [19:28:20] <^demon> Other stuff needs setting up. [19:29:01] ^demon: Which means... hours, days, weeks, unknown? (Yes, I'm over excited) [19:29:16] <^demon> Hours/days, at most. [19:29:22] <^demon> Trying to have it done by the end of the week. [19:29:47] :) [19:30:03] mlitn, ext.articleFeedbackv5@10-option6X-click_yes , then ext.articleFeedbackv5@10-option6X-submit_attempt, then articlefeedbackv5 with my comment, then ext.articleFeedbackv5@10-option6X-cta_edit-impression , then ext.articleFeedbackv5@10-option6X-submit_success [19:30:39] mlitn and then editing the page and no more XHRs. [19:32:18] hrm [19:34:02] Also, I'm looking in the piramido log and I don't see the articlefeedbackv5 with my comment [19:35:36] mlitn: if your deployment is going to be held up for a while let me know. in that case, I'll seperate our 2 config changes so I can push mine without yours. [19:36:06] ok, go ahead :) [19:36:18] mlitn: ping me when the code is up for testing on testwiki [19:36:21] the server response to the api call is {"articlefeedbackv5":{"result":"Success","feedback_id":9,"aft_url":"\/srv\/mediawiki\/index.php?title=Special:ArticleFeedbackv5\/Hipster_ipsum&ref=cta","permalink":"\/srv\/mediawiki\/index.php?title=Special:ArticleFeedbackv5\/Hipster_ipsum\/9&ref=cta"}} , which is better than its usual response :) [19:36:30] DarTar: will do [19:41:13] spagewmf: sorry, the event should've be option6X-cta_edit-button_click, not plain button_click [19:41:20] but I assume that's not there as well [19:41:36] so I've just added stall() to all of those ^^ [19:42:27] if you could re-pull the updated patchset for another test-run, before going to the real test [19:45:46] or I could be not lazy and do it myelf, which I just did [19:46:05] mlitn: CentralNotice config change deployed [19:52:14] nighty people o/ [19:53:31] weird, why on earth is this the first result in a Google search on bugzilla? https://bugzilla.wikimedia.org/show_bug.cgi?id=912 [19:55:31] and following 4 bugs are by purodha, he must have googlebombed them :p [20:00:42] mlitn, with Patch 4 I'm still not seeing the articlefeedbackv5 event with comment appear in the ct.log on piramido :-( [20:01:19] spagewmf: oh, the one with the comment doesn't go to the CT log, but goes to AFT API, which stores to database [20:01:59] mlitn yay, one less thing to worry about [20:02:05] do you have an event called ctaXXX-button_click [20:02:44] mlitn yes I do! ext.articleFeedbackv5@10-option6X-cta_edit-button_click [20:04:40] ok, sounds good, that's on of those that have now changes [20:04:48] changed* [20:04:53] that used to use the redirect-thingy [20:05:42] another one: if you're on a page with the fork, and click on the "edit" button at the top of the page, you should see a edit_tab_link-click [20:05:57] clicking a section edit link should yield a section_edit_link-click event [20:07:44] and if you complete the form at piramido.wmflabs.org/wiki/Main_Page?aftv5_cta=4 and click either of both buttons at the end, and event cta_signup_login-button_signup_click or cta_signup_login-button_login_click [20:13:41] mlitn, give me a sec, I'm going to move the ct.log so it's viewable in the browser [20:13:51] take your time [20:21:39] is post-edit feedback function about to be enabled for all projects? when? [20:21:58] Steven (WMF) just posted a message about it (for ex. to pl.wiki), but it gives no real information... [20:22:21] https://pl.wikipedia.org/wiki/Wikipedia:Kawiarenka/Kwestie_techniczne#Og.C5.82oszenie_nowej_funkcji [20:25:05] mlitn, Fabrice, DarTar, http://piramido.wmflabs.org/mediawiki_logs/ct.log.txt , have fun QAing [20:25:35] great, thanks [20:25:45] thank you [20:26:31] I'll go ahead and make tests, the log is updated in real time, right? [20:26:35] the \n's were a temp issue [20:27:31] mlitn, DarTar, Fabrice, as anonymous I wasn't invited to create an account. Expected? [20:28:06] spagewmf: 50/50 account creation or edit wiki [20:28:29] you can manually manipulate that with ?aftv5_cta=1 (edit page) of =4 (create account) [20:30:00] LzIbigBulhJC7Z4cGRnSHuqyLagLlg6n was me [20:30:58] edit CTA funnel went as expected [20:30:58] mlitn, X9zkjaUio... was me anonymous. After clearing cookies I did get the invitation to create account [20:31:07] all events have the additional data stored as expected [20:31:44] and the campaign is there in the ACUX lines [20:32:00] mlitn: only concern: the aft impression event is missing [20:32:25] funnel starts with 10-option6X-click_yes [20:32:36] DarTar: that's only 1% [20:32:51] oh, I thought we were sampling inits [20:32:51] or do you want it at 100%? [20:33:01] both init and impression [20:33:16] not impressions, which require user action (scrolling to the bottom) [20:33:16] but if you want 100% impression, it's an easy change ;) [20:33:32] the previous set up had 1% init, 100% AFT impressions [20:33:54] i'll change it [20:33:57] sweet [20:34:03] mlitn and DarTar good luck, I'm moving to E3 work. Please put in a good word with Reed Hastings ;-) [20:34:23] thanks spagewmf (and ori-l) for your support [20:35:20] mlitn, let me know if you want me to fetch a Patch 5... 6. DarTar, don't thank me, tell Reed. http://www.slideshare.net/reed2001/culture-1798664/34 third bullet point :-) [20:35:51] :) [20:36:44] off to test cta4 [20:36:45] DarTar: should be at 100% now [20:36:55] might need to clear cached js though ;) [20:37:02] great, will restart with a clean cache [20:37:04] spagewmf: & ori-l: thanks indeed [20:37:10] much appreciated [20:39:30] mlitn: there's an issue with new lines, but impresisons are there now (first occurrence of myYnx0vZrWf0bYItEthbB8XRw8OdWVy7) [20:40:31] cta4 is next [20:44:27] DarTar, Ori added newline to some logging on piramido to make the ct.log easier to scan [20:44:44] cta4 test completed and successful, but I want to try once more [20:47:21] mlitn: everything is working as expected [20:47:49] tested each funnel twice after clearing the cache [20:47:57] great :) [20:48:36] --> testwiki is next? [20:50:26] Don't forget to let me know before you scap to live. [20:50:44] We should be ready soon [20:51:03] ok, I'm gonna +2 myself - k? [20:51:04] trying to do a partial roll-back [20:51:10] kaldari: it's ok to pull on testwiki in about 5 min? [20:51:30] sure, you can do whatever you need on test [20:54:26] i'm going to deploy a CommonSettings.php update to disable an experimental option in MobileFrontend [20:54:27] be warned [21:12:45] so, DarTar, the change is on test [21:13:02] yay [21:14:51] kaldari: logos will be ready tomorrow [21:15:52] Nemo_bis: cool [21:16:42] somehow the CentralNotice updates got pushed to the live wikis already [21:16:56] did someone run scap? [21:18:00] (http://imgur.com/ZzDLQ too, already reported) [21:18:28] kaldari: I can't click away the banner, it keeps bugging met at nlwp [21:18:43] Does this have to do with the fact that the level is "emergency"? [21:18:59] no, should still be able to close it... [21:19:03] This is starting to piss me off (and I'm not the only one) [21:19:14] works for me [21:19:29] blue banner on nl.wiki? [21:20:30] https://meta.wikimedia.org/w/index.php?title=Special:NoticeTemplate/view&template=B12_1025_facts_world <- this ugly bastard [21:20:47] I see.... [21:21:17] What is the centralnotice priority? That's new to me, what does it do? [21:24:07] multichill: I turned the campaigns off [21:25:12] multichill: the priority is similar to the old "preferred" setting, but with more levels. [21:26:44] So Zack has been planning this for months and he didn't even test the code on some tiny wiki? :-( [21:29:08] Reedy: Why's you sync the extensions dir? [21:29:13] Why'd [21:29:36] We were in the middle of trying to roll-back a change to CentralNotice [21:29:36] 1 command instead of 3 [21:32:41] i don't know how many years this needs repeating, but the dismissability of central notice banners should belike the most tested part of the whole freaking campaing. and somehow it never seems to be :D [21:33:27] multichill, thedj: That code wasn't supposed to go out yet, but Reedy accidently deployed it, thus breaking things [21:35:19] AbuseFilter might currently be broken as well [21:35:30] kaldari: https://commons.wikimedia.org/wiki/File:Wm-fundraiser2012.png <- nice [21:35:38] * jeremyb wants salt! (and not on my food) [21:38:22] mlitn: it looks like AbuseFilter got deployed prematurely. You may want to check that it's still working, as I don't know if the config change for that has gone out too [21:38:25] it's taking many scripts with it as well it seems. [21:39:12] kaldari: what has gone out? config or extension? [21:39:24] mlitn: the extension [21:39:55] ok, i'm going to revert 1 commit and sync-dir [21:40:28] Reedy deployed all the extensions, which includes all the ones that were still pending on test: AbuseFilter, ArticleFeedbackv5, CentralNotice, LandingCheck, etc. [21:41:15] oh aft too [21:41:15] k [21:44:20] Unfortunately, it happened right in the middle of a fundraising test :( [21:51:39] FYI: jenkins fails my extension update commit for a reason i'm pretty sure is unrelated: [21:51:43] ParserTests::testParserTest with data set #0 ('Blank input', '', '', '', '') [21:51:43] PHP_Invoker_TimeoutException: Execution aborted after 10 seconds [21:51:52] ^ anyone know what that's about? [21:52:50] full stack trace: https://integration.mediawiki.org/ci/job/MediaWiki-Tests-Parser/7946/testReport/(root)/ParserTests__testParserTest/testParserTest_with_data_set__0/ [22:06:43] mlitn: OK, I have the fixes for CentralNotice ready to push to test, but I want to wait until you guys are totally finished with your deployment [22:07:59] kaldari: go ahead! [22:08:05] thanks [22:23:35] I am using the override [22:23:43] (wrong window) [22:35:13] Just so everyone knows: I'm currently doing another deployment for CentralNotice to clean-up from earlier [23:10:44] about to run scap [23:14:01] running scap [23:27:55] is anyone aware of the problem with the fundraising banners? [23:27:56] https://en.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_(technical)#New_banner [23:28:12] helderwiki_: yes [23:28:13] Yes [23:29:05] ok [23:29:19] (I saw questions on ptwiki and then found it on enwiki...) [23:29:50] thanks for double checking [23:29:57] =) [23:38:57] kaldari: aftv5 rolled back as well ? [23:39:27] it was rolled back for a bit, but is live now [23:40:50] there is a report on vp/t about a JS error, and i can reproduce it [23:40:52] or [23:40:52] line: 16 [23:40:52] message: "'undefined' is not a function (evaluating 'link.text(mw.msg('articlefeedbackv5-watchlist-view-feedback')).html(link.html()+' »').attr('href',url).stall('click',function(){return $.aftTrack.trackClick(↵'watchlist_view_feedback-button_click');})')" [23:40:57] sourceURL: "https://bits.wikimedia.org/en.wikipedia.org/load.php?debug=false&lang=en&modules=ext.Experiments.experiments%2Clib%7Cext.UserBuckets%2CmarkAsHelpful%2CpostEdit%7Cext.articleFeedback.startup%7Cext.articleFeedbackv5.watchlist%7Cext.gadget.BugStatusUpdate%2CDRN-wizard%2CHotCat%2CNavigation_popups%2CNoAnimations%2CReferenceTooltips%2CTwinkle%2CUTCLiveClock%2Ccharinsert%2Cexlinks%2CmySandbox%2Cpurgetab%2Cteahouse%2Cwidensearch%7Cjquery.artic [23:41:03] stack: "@https://bits.wikimedia.org/en.wikipedia.org/load.php?debug=false&lang=en&modules=ext.Experiments.experiments%2Clib%7Cext.UserBuckets%2CmarkAsHelpful%2CpostEdit%7Cext.articleFeedback.startup%7Cext.articleFeedbackv5.watchlist%7Cext.gadget.BugStatusUpdate%2CDRN-wizard%2CHotCat%2CNavigation_popups%2CNoAnimations%2CReferenceTooltips%2CTwinkle%2CUTCLiveClock%2Ccharinsert%2Cexlinks%2CmySandbox%2Cpurgetab%2Cteahouse%2Cwidensearch%7Cjquery.articleF [23:41:14] woops, bit longer than it looked when i copied that. [23:44:30] so more readable, with debug=true on [23:44:30] ext.articleFeedbackv5.watchlist.js:34 [23:44:30] TypeError: 'undefined' is not a function (near '....stall( 'click', function(...') [23:49:49] mlitn: ^ [23:52:07] thedj: looking [23:53:02] ori-l: ^ [23:53:11] looks related to the stall functionality [23:53:17] but maybe that's cached from earlier? [23:53:27] i used debug=true [23:53:39] * ori-l looking [23:53:41] ok [23:55:12] specific on Special:Watchlist [23:55:17] i'm seeing a "Uncaught TypeError: Cannot read property 'name' of undefined " from bannerController.js:95 [23:55:43] i was using safari/chrome [23:55:58] ori-l: hm, that's also not good then :D [23:56:28] yeah, looks like it's CentralNotice... [23:56:39] just saw that myself [23:57:41] thedj: which page(s) are you seeing that error on? [23:58:01] https://en.wikipedia.org/wiki/Special:Watchlist?debug=true [23:59:03] hrm, i'm not getting that. might be a race condition; perhaps aft forgot to declare E3Experiments.lib as a dependency. taking a look.