[00:01:34] mlitn: checking [00:05:21] mlitn: yeah, i still see too many 1 second queries [00:06:14] it looks like the issue is with the sheer number of potential page ids in the IN clause for [00:06:19] SELECT /* DataModelBackendLBFactory::getList */ *,aft_relevance_score AS offset_value FROM `aft_feedback` WHERE aft_page IN ('27412213','1831030','36645139','27894611'…… [00:06:42] many thousands in some cases [00:07:44] that one should use an index on aft_page, but nothing else [00:08:08] mlitn: in some cases it is resulting in full table scans based on rows_examined [00:08:41] if you have enough non consecutive values in an IN clause, the index will no longer be used [00:09:25] # Query_time: 2.423105 Lock_time: 0.009054 Rows_sent: 25 Rows_examined: 580729 [00:09:31] oh ok; the values are watchlisted page ids, so that's quite the common case [00:15:49] I don't really see an alternative right away though [00:16:14] mlitn: just discovered another issue: DiscussOnTalk:

will be html-encoded inserted, so doesn't work. Could be temp-fixed by changeing the l10n-msg? [00:16:33] Hey se4598 or mlitn : Is there a test page on de.wiki where I could post some feedback without causing any issues? Should I use http://de.wikipedia.org/wiki/Studentendoppelgrab ? [00:17:13] fabriceflorin: that page is pretty much watched, I guess [00:17:17] fabriceflorin: I'd suggest bugzilla :p [00:17:23] denis_o: hi again ;) [00:17:30] hi all [00:17:33] Change merged: Krinkle; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/53997 [00:17:51] se4598: dont't think we can temporarily fix the message; it'll still be escaped [00:18:12] s/escaped/htmlencoded/ [00:19:14] fabriceflorin: choose http://de.wikipedia.org/wiki/Sar_%28Eukaryoten%29 [00:19:44] there you are for your own [00:19:46] :) [00:20:01] denis_o, mlitn, se4598: Thank you! I will post a comment on Sar_%28Eukaryoten%29, then immediately mark it as inappropriate. [00:20:08] mlitn: I mean by removing the html-tag, but that can also be fixed by hand for 2 days; feedback-text should be still encoded though [00:20:17] se4598: that would work indeed! [00:24:19] binasher: I've added a patch that will - for now - completely disabled AFTv5 on enwiki (where it's currently not visible for users anyway) - https://gerrit.wikimedia.org/r/#/c/54807 - just to make sure nothing bad happens because of these full table scans [00:26:41] binasher: while I have not yet any clue how that could be improved, it should definitely keep the load much lower for now [00:30:34] mlitn: thanks [00:30:37] Aaron|laptop and i were just talking about the watchlist issue [00:32:16] se4598: unfortunately, the blockquote-html is not part of the i18n message, so we can't temp-fix it [00:32:57] binasher: did you come up with any possible solution, I'm currently failing to see an alternative [00:33:29] mlitn: i actually think the watchlist feature needs to be completely redesigned [00:34:31] I think if we limited watchlist size that would be fine for sp:watchlist for a while...it would be of limited help to aft though [00:37:12] mlitn denis_o se4598 : Have you found any more serious issues, or are you generally comfortable with the application as it stands, despite its known bugs? I have been able to test it with anonymous accounts as well as with my own accounts and Aft5test, and it is generally working OK for me -- except on my Chrome browser, where I seem to have a bad cookie that doesn't let me moderate, as well as shows posts in the wrong filters. [00:37:13] find the cookie and remove it on that browser, if I can. [00:37:55] mlitn: need to run, talk to you more tomorrow re: watchlist [00:38:55] fabriceflorin: I don't think it's a cookie; chrome is known for (overly) agressive caching, I assume you'll have some outdated js or something [00:39:57] fabriceflorin: as I have only been intermittent part of this evenings work regrettably, i put my trust in the hard work you three have done so dedicated [00:40:15] fabriceflorin: I'm generally comfortable with how it is right now: 3 bugs is all I know of so far [00:40:20] Hmmm. Will that go away over time? Or should I stop testing AF5 with Chrome? For now, I will test on Firefox and Safari. [00:43:08] denis_o: Well, the folks that really deserve the credit here are mlitn and se4598 : big round of applause to you both for a job well done! [00:43:26] it should clear over time (or you may be able to clear the browser's cache?) [00:43:34] * denis_o co-applaudes heavily [00:43:40] fabriceflorin, did you know that to 'set up' an IRC channel, you just attempt to join one that doesn't currently exist? :) [00:44:06] fabriceflorin: there's also another potential issue with feedback for watchlisted pages that I'll need to discuss with Aaron|laptop & binasher (best prior to deploying on enwiki - the feedback size is much larger there) [00:44:07] denis_o, se4598 : have you announced this to your community yet? If so, can I see the link? [00:44:56] krenair: thanks for the tip, will start the ee-channel tomorrow, see how that works. Right now, we're busy with the German deployment of AFT5, so we will stay here for the rest of the night. [00:44:58] fabriceflorin: currently only at https://de.wikipedia.org/wiki/Wikipedia_Diskussion:Artikel-Feedback [00:45:40] A few people are already in #wikimedia-ee, #wikimedia-e2 and #wikimedia-e3 [00:46:08] fabriceflorin: we will place it on the Kurier with a day delay [00:48:05] se4598 and denis_o : Thanks, that announcement really looks great! I am glad that you are listing all the known issues on this page, so people get a fair warning that this is still being debugged as we speak. You might also want to consider linking to your editor help page from the announcement, so people can get an overview of all the tool's features. [00:49:24] mlitn: what is the issue with watchlist feedback? Is it a new problem, or a general request for improvements? [00:50:28] fabriceflorin: as people's watchlists are "unique", it makes no sense caching the data [00:50:50] fabriceflorin: and quite a number of watchlists are extremely large, making it not so optimal to retrieve the data [00:52:30] mlitn: I see. That watchlist caching problem sounds like it could be a serious issue, particularly if it's causing performance problems. Let's investigate in the morning. [00:53:06] yes - we'll definitely need to talk this over with Aaron|laptop & binasher [00:54:04] mlitn: Sounds good. Is the en-wiki data transfer going well? Will we be able to complete it by end of day Wednesday, for Thursday's en-wiki deploy? Any concerns about possibly impacting the main servers? [00:55:49] fabriceflorin: I expect the data to be merged in time, but would be cautious about the date: [00:56:11] * let's see how well dewiki unfolds and hope no big issues arise [00:56:29] denis_o, se4598: If you get feedback from the community tomorrow, it would be great if you could summarize the key points by the end of your day on Wednesday, with any suggestions for next steps. [00:56:49] * I'll want to at least talk about the watchlist thingy before enabling it on such large feedback sample, to make sure the shard can handle the load [00:57:40] fabriceflorin: ok [00:58:54] mlitn: Point well taken. We will not make any promises for Thursday's deployment. If it needs to be pushed to next Tuesday, I expect it would be OK. Let's think of some possible solutions regarding the watchlist issue: is there anything we can do from a feature design standpoint to reduce the load on the servers. Until we have a solution, maybe we can simply disable the watchlist feature on en-wiki for now. [01:00:45] fabriceflorin: sounds like a solid worst-case plan, let's see where it takes us the next couple of days :) [01:00:50] fabriceflorin, mlitn: as everything seems to be ok I will go offline to get some sleep. Bye [01:01:04] se4598: sure, have a good night; thanks for the help [01:01:16] much appreciated! [01:01:34] se4598 mlitn : Thank you so much for doing such wonderful work today. It's a true pleasure to be working with you both! [01:02:38] mlitn: Shouldn't you be catching some zzz's too? (I will be offline for 90 minutes shortly myself, but can resume testing this evening.) [01:03:13] goodnight fabriceflorin mlitn [01:03:34] denis_o: goodnight & thanks! [01:03:49] denis_o: Thanks so much for all the great work -- and good night! [01:03:56] fabriceflorin: indeed, I'll be leaving too; talk to you tomorrow :) [01:04:38] mlitn: Sounds great. Congrats again for getting this puppy out, and kudos for all your awesome work today! Good night ... [03:23:30] TimStarling: have you seen https://gerrit.wikimedia.org/r/#/c/52994/ ? [03:23:56] possibly [03:25:57] I was just reviewing some of Brad's stuff [03:26:21] # Make zero-based, without screwing up named args [03:26:23] $args = array_merge( array(), $args ); [03:26:38] do you like that? sometimes code review makes me happy [03:26:45] very occasionally [03:27:37] how would that work? [03:28:11] ah, nvm [03:29:03] sometimes I forget how it works differently depending on int/string keys, heh [03:29:30] right, but that property is all he's using it for [03:31:34] if a hammer works... [03:42:59] your change looks pretty straightforward [03:48:40] that new implementation of doDeduplicateRootJob() is not reachable at present, right? [03:49:08] and there are no immediate plans to add JobQueue subclasses that might use it? [03:49:40] well there is https://gerrit.wikimedia.org/r/#/c/52994/2 [03:49:56] and an aws extension [03:50:08] nothing we have would hit it atm though [03:50:17] 52994 is what I'm looking at [03:50:42] right [03:50:52] it's not a problem at a code review level, it's just questionable from the perspective of time management [03:51:01] and 52994 is a subclass [03:51:13] gah [03:51:19] 51823 [03:51:52] yeah, I was just looking through your changes to work out what you meant ;) [03:51:59] clipboard fail [03:52:10] as Reedy would probably call it [03:52:37] so it would be reachable if that were merged [04:08:06] TimStarling: I had to manually rebase that [04:08:45] did you give it +2? [04:09:15] no [04:09:18] * Aaron|home reads http://www.businessinsider.com/marissa-mayers-slow-bureaucratic-hiring-process-is-driving-some-people-nuts-2013-3 [04:15:32] TimStarling: maybe we should move gerrit to a server with a bunch of ssds in raid0 or something [04:15:53] * Aaron|home is only half joking [04:16:36] if you mean to improve submit time, it is mostly jenkins [04:17:02] if you mean to improve fetch time, I'm not sure if that is I/O, I think it might be mostly CPU [04:17:07] is the merge action getting stuck in some queue somewhere? [04:17:46] * Aaron|home needs to figure out more about what jenkins is doing [04:20:52] * Aaron|home looks at http://ganglia.wikimedia.org/latest/?c=Miscellaneous%20eqiad&h=gallium.wikimedia.org&m=cpu_report&r=hour&s=by%20name&hc=4&mc=2 [04:21:27] yeah, maybe gallium needs some SSDs [04:21:52] but there is also high CPU usage [04:22:12] well, not high as in capacity, but maybe high as in single threaded latency [04:22:28] maybe it needs to run tests in more processes in parallel [05:38:00] New patchset: Ori.livneh; "Add PleaseStand to trusted users group on zuul" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54823 [05:59:20] New patchset: PleaseStand; "Add PleaseStand to trusted users group on zuul" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54823 [06:49:16] Hello bug squad! [07:41:56] hello [08:32:06] New patchset: Hashar; "whitelist Kevin Israel (PleaseStand)" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54823 [08:33:44] New review: Hashar; "I have slightly adjusted the commit message. The change is good and will actually whitelist Kevin em..." [integration/zuul-config] (master); V: 1 - https://gerrit.wikimedia.org/r/54823 [08:36:12] New patchset: Hashar; "Revert "placeholder for zuul.wikimedia.org"" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54829 [08:36:26] New patchset: Hashar; "Revert "placeholder for zuul.wikimedia.org"" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54829 [08:36:35] Change merged: Hashar; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54829 [09:28:44] New review: Hashar; "(1 comment)" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53995 [09:33:13] New patchset: Zfilipin; "Updated Ruby gems" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54834 [10:27:39] New patchset: Hashar; "replace mediawiki logo with the wikimedia one + link" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54841 [10:27:53] Change merged: Hashar; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54841 [12:56:34] New patchset: Hashar; "Zuul status page" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54838 [12:59:35] New review: Hashar; "Yeah that is not nice, but good enough for now." [integration/docroot] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54838 [12:59:36] Change merged: Hashar; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54838 [13:00:26] ^demon: https://integration.wikimedia.org/zuul/ :-D [13:01:50] <^demon> whee. [13:02:11] <^demon> Use https for the image. [13:02:13] ^demon: all credits to James E. Blair @ Openstack : http://zuul.openstack.org [13:02:14] <^demon> I'm getting mixed content :p [13:03:21] New patchset: Hashar; "rm check-silent pipeline, no more used" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54857 [13:03:37] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54857 [13:04:28] ^demon: I swear I double checked it uses a // url [13:04:29] grrr [13:05:47] New patchset: Hashar; "use // as protocol for images" [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54858 [13:06:02] Change merged: Hashar; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54858 [13:06:44] ^demon: fixed :] [13:06:54] <^demon> yay, thanks. [13:07:50] the changes to integration/docroot are self deploying on merge :-] [13:08:15] <^demon> I wonder if zuul/status should just redirect to the nicer page now. [13:08:23] <^demon> Same info, much prettier :) [13:08:33] I will update zuul configuration later tonight [13:08:42] <^demon> *nod* [13:08:46] zuul.conf has a status_url parameter for that [13:09:00] also we use twitter bootstreap [13:09:13] maybe timo will ends up writing a WMF theme for bootstrap :-] [13:09:14] <^demon> Yeah [13:09:41] https://integration.wikimedia.org/nightly/ <--- ugly [13:23:05] $ sudo -u mwdeploy -- 'whoami; whoami;' [13:23:05] sudo: whoami; whoami;: command not found [13:23:06] i hate myself [13:32:48] Project _debug-template build #75: FAILURE in 34 sec: https://wmf.ci.cloudbees.com/job/_debug-template/75/ [13:37:45] zeljkof: hey there :-] [13:38:05] hashar: hi :) [13:38:10] zeljkof: I noticed you have opened a bunch of bug reports under Wikimedia > Testing Infrastructure, I can create you another component if you want [13:38:21] zeljkof: though I am fine having everything under the same component :-] [13:38:43] hashar: it would actually be nice to have Selenium component [13:38:47] maybe I will end up creating a "Continuous Integration" product [13:38:57] probably make more sense to have a product [13:39:21] hashar: what ever makes sense, browser automation, CI, Selenium… [13:39:25] andre__: ^demon: what do you think about creating a "Continuous Integration" product under bugzilla ? [13:39:55] andre__: ^demon: that would hold the selenium / jenkins / zuul / nightlies … components [13:39:59] <^demon> To replace Wikimedia -> Testing Infrastructure? [13:40:07] yeah [13:40:08] <^demon> That sounds sane. [13:40:24] Project _debug-MobileFrontend-template build #1: FAILURE in 4 min 56 sec: https://wmf.ci.cloudbees.com/job/_debug-MobileFrontend-template/1/ [13:40:42] I am wondering if I should just do it or mail engineering to bikesh^W get more input [13:41:29] <^demon> Well, maybe wait for andre__ to weigh in, but I don't think it needs bikeshedding. [13:41:46] waitfor -p `pidof andree__` [13:41:58] <^demon> If andre__ and maybe the QA guys say it's a good idea, Just Do It :) [13:43:44] <^demon> Ok, running to the airport to pick up someone. Back in ~1h. [13:43:53] *wave* [13:47:29] at least one QA (quantum armor) guy is fine with it [13:49:10] ^demon|away, hashar: I don't have any opinion. Whatever categorization works best for you and QA folks to find issues of interest I'm happy to set up. [13:49:55] andre__: taking that for a Yes :-] [13:50:03] zeljkof: so a product "Continuous Integration" ? [13:50:10] or is CI part of QA ? :-] [13:51:16] hashar: I do not think we need CI for quantum armour (QA) :) [13:51:30] ;-D [13:51:38] continuos integration is fine with me [13:52:05] will you just rename the existing testing infrastructure? or create a separate thing? [13:53:02] will create a new thing [13:53:05] and move the bugs I guess [13:55:46] whatever is done, please file a bug report about it. [13:55:53] in Bugzilla, against Wikimedia>Bugzilla. [13:56:07] oh sure [13:56:15] CI isn't part of Testing Infrastructure? [13:56:26] we could move testing infra under QA [13:56:26] hmm. [13:56:38] hmm, hmm. [13:56:39] I have been using Testing infrastructure for bug relates to Jenkins / Zuul [13:56:46] Maybe this needs more discussion and comments by stakeholders. [13:56:47] I am not really sure where to put all that crap :-] [13:56:51] yeah [13:57:07] File a bug report, define the problem, list some potential solutions, CC all involved / affected parties? :) [13:57:11] should write down something about it and propose a new layout [13:58:08] +1, highly welcome [13:59:00] Yippie, build fixed! [13:59:00] Project MobileFrontend-linux-android build #6: FIXED in 8 min 7 sec: https://wmf.ci.cloudbees.com/job/MobileFrontend-linux-android/6/ [13:59:01] * zeljko.filipin: Updated Ruby gems [13:59:01] * zeljko.filipin: Workarond for Sauce Labs Android driver bug [13:59:42] wmf-selenium-bot: good dog [13:59:42] zeljkof you may not issue bot commands in this chat! [14:00:19] what an attitude, wmf-selenium-bot :) [15:27:35] New patchset: Zfilipin; "Update Firefox to version 19" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54865 [15:31:56] New review: Cmcmahon; "maintenance" [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54865 [15:31:56] Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54865 [15:32:15] New review: Cmcmahon; "maintenance" [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54834 [15:32:15] Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54834 [15:52:57] mlitn: DiscussOnTalkpage-missing signature is fixed in translatewiki [15:53:06] ok great :) [16:13:30] mlitn: to provoke the layout issue you only have to reduce the width of the browser window, with german msg for me it starts at 1240px and lower, also because then the reader-toolbar has a line break [16:14:16] but IMO this is somewhat low priority [16:14:58] hi chrismcmahon [16:15:51] se4598: the comment and the text below that start to overlap each other? [16:16:10] hi sumanah [16:16:53] chrismcmahon: I'm sorry for the schedule juggling. I can meet in 15 min for the full hour. My other appointment hasn't shown up and so I deemed it cancelled by forfeit [16:17:10] sumanah: works for me [16:17:32] mlitn: yes [16:17:53] ok, got it [16:18:15] will definitely not be the first thing I work on though ^^ [16:48:17] hashar: https://integration.wikimedia.org/nightly/ is ugly? This is even uglier: https://integration.wikimedia.org/WikipediaMobile/nightly/ [16:59:02] New patchset: Zfilipin; "Disable a test in Firefox" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54874 [17:14:40] New patchset: Zfilipin; "modified search.feature to add test for accents search" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54719 [17:15:52] New patchset: Zfilipin; "modified search.feature to add test for accents search" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54719 [17:16:57] Change abandoned: Zfilipin; "I will merge this change in the original commit: https://gerrit.wikimedia.org/r/#/c/54719/" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54724 [17:19:41] New review: Zfilipin; "Merged this with the commit that implemented suggested changes: https://gerrit.wikimedia.org/r/#/c/5..." [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54719 [17:19:42] Change merged: Zfilipin; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54719 [17:20:55] New review: Zfilipin; "Looks like we are merging scenarios without implementations[1], so I am merging this one too." [qa/browsertests] (master) C: 2; - https://gerrit.wikimedia.org/r/54097 [17:24:21] New patchset: Zfilipin; "find tokens in the captions of uploaded files" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54097 [17:24:45] Change merged: Zfilipin; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54097 [17:37:15] mlitn: don't like to interrupt you, but we have another wrong calculation/negative-number-bug at helpful-votes: 1) 1xUP/3xDOWN => 0% helpful, 2) 2xDOWN => -100%; now 3xDOWN => -50% [17:37:19] https://de.wikipedia.org/wiki/Spezial:Artikelr%C3%BCckmeldungen_v5/Hamster/14334 [17:37:50] https://doc.wikimedia.org/mediawiki-core/master/js/ [17:38:08] ori-l: superm401: Initial run ^ Not automated yet. [17:38:34] Krinkle, at the security training, but very cool. [17:39:54] mlitn: but tooltip says -1xUP, 3xDOWN, which doesn't match log [17:43:15] New review: Cmcmahon; "disable failing test for FF until https://bugzilla.wikimedia.org/show_bug.cgi?id=46367 is fixed" [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54874 [17:43:16] Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54874 [18:05:19] YuviPanda, hey https://bugzilla.wikimedia.org/show_bug.cgi?id=34557 ,where do i start this one with [18:06:46] Krinkle, hey [18:07:26] Krinkle, can you help me out with a bug? [18:07:28] Rahul_21: Hi [18:07:39] Maybe, I'm a bit busy, but I'll read here and answer when I can. [18:07:53] https://bugzilla.wikimedia.org/show_bug.cgi?id=34557 [18:08:24] Krinkle, https://bugzilla.wikimedia.org/show_bug.cgi?id=34557 which file do i peep into ? [18:13:21] kaldari, hello [18:13:32] howdy [18:13:34] Rahul_21, please see https://bugzilla.wikimedia.org/show_bug.cgi?id=45775#c7 [18:14:25] Krenair, i am on it!i cant change the status [18:14:34] Rahul_21: I'm not sure what you're asking for bug 34557 [18:14:39] Rahul_21, try logging out and logging back in again? [18:14:44] Rahul_21: It is a general bug, not about a specific element [18:15:05] we shouldn't have any elements that just do width: 100% and on an element with a border [18:15:11] that's the bug. [18:15:18] Krinkle, where do i make the changes but?which file [18:15:24] if those are all phased out, the bug can be marked fixed [18:15:32] Rahul_21: I don't know, ask Daniel, he filed the bug. [18:15:43] Leave a comment on the bug or try to ping him on IRC [18:16:03] qgil, Wikimedia/IRC in Bugzilla is for issues with the RC feeds on irc.wikimedia.org [18:16:04] thedj: Dantman: ^ [18:16:08] completely separately - looking at http://meta.wikimedia.org/wiki/Schema:PageContentSaveComplete and thinking "should this be on meta?" [18:16:32] Rahul_21: thedj = Derk-Jan Hartman, Dantman = Daniel Friesen (from bug 34557) [18:17:25] Krenair, ah ok. I read the description of the component but it wasn't clear to me. I was about about ask andre__ but... [18:17:27] Rahul_21, did it work? [18:18:06] Rahul_21, yes,i set the status to resolved,fixed [18:18:06] Krenair, I'm supposed to move it to Wikimedia - General, or ask in Meta? andre__ [18:18:06] qgil, uh what I said was basically just a copy+paste of the component description [18:18:29] Well, part of it [18:18:38] Try "General" for other bots [18:18:45] Krenair, you just cnfirmed a suspicion [18:19:06] Rahul_21, great [18:19:29] * andre__ missing context [18:19:32] superm401: all EventLogging schemas are on meta [18:19:38] err [18:19:40] i mean sumanah [18:19:41] andre__: https://bugzilla.wikimedia.org/show_bug.cgi?id=46377 [18:19:41] ^ [18:19:43] YuviPanda, I know. [18:19:47] Except in dev of course. [18:19:56] thanks [18:20:31] superm401: sumanah asked about it in the channel a little bit back, and autocompletefail. sorry. [18:20:54] No problem [18:21:35] YuviPanda: ok. So, is that good? is that where they ought to live? [18:21:46] sumanah: yes. No data, only schemas [18:21:57] sumanah: mediawiki.org/wiki/Extension:EventLogging/Schemas has technical details about those [18:22:31] sumanah: so that is where they *do* live. I also think that that is perhaps where they *should* live, to make it easy for wikimedians to know what is being tracked [18:23:17] ok, that argument helps; otherwise I would ask myself "why should this stuff not be on mediawiki.org?" [18:24:07] sumanah: yeah, these are wmf specific. The only place other than meta I could think of is wikitech, but sicne this is a lot more for the 'community' as such than programmers... [18:24:08] meta makes sense to me [18:24:16] qgil: Yeah, goes to "General" (though no idea who to assign to this yet)... [18:24:41] okay! thanks for the explanation YuviPanda - the Wikimedia specificity is another aspect I had not considered [18:25:34] yw sumanah. [18:25:57] Hey YuviPanda, can I somehow trick you into answering my email from last Tuesday? :) [18:26:03] ow [18:26:04] (nothing urgent, just pinging) [18:26:05] sorry andre__ [18:26:05] forgot [18:26:10] no problem, really :) [18:34:37] YuviPanda, hello [18:49:26] New patchset: Krinkle; "Zuul: Clean up html and css." [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54887 [18:50:44] Change merged: Krinkle; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54887 [19:06:14] New patchset: Krinkle; "Doc: Add mediawiki-core (PHP + JS) and VisualEditor." [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54890 [19:06:37] Change merged: Krinkle; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54890 [19:06:50] hashar: https://doc.wikimedia.org/ [19:07:15] hey Rahul_21 [19:07:18] sorry, been busy [19:07:38] Rahul_21: as Krinkle said, you should poke Dantman [19:07:38] Rahul_21: another option would be to grep for 'width: 100%' [19:07:44] and check if any of them have borders [19:07:52] in core, vector, monboook, and the extensions [19:08:22] thedj suggestd i should look in skins/common/ [19:08:34] yeah [19:08:36] grep through core [19:08:40] (since skins/common is in core) [19:09:16] New patchset: Krinkle; "Doc: Fix lack of margin when window is narrow." [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54891 [19:09:23] yes i am adding box-sizing:border-box [19:09:35] New patchset: Krinkle; "Doc: Fix lack of margin when window is narrow." [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54891 [19:09:41] Change merged: Krinkle; [integration/docroot] (master) - https://gerrit.wikimedia.org/r/54891 [19:16:46] YuviPanda, do i look for other selectors apart from textarea? [19:16:56] I am unsure. [19:16:59] comment on the bug, ask Dantman? [19:17:01] he filed it [19:17:17] Dantman, is sadly not available [19:19:57] Rahul_21: leave a comment on the bug? [19:20:08] YuviPanda, sure [19:29:24] New patchset: Krinkle; "mwext-VisualEditor-docgen: Don't set --verbose flag for rsync." [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/54895 [19:29:48] Change merged: Krinkle; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/54895 [19:38:28] YuviPanda, I was going through the extension building manual yesterday,and i thought i should give it a try [19:53:26] Fatal error: require_once(): Failed opening required '/opt/lampp/htdocs/mediawiki/extensions/AllTimeZones/AllTimeZones.php' (include_path='.:/opt/lampp/lib/php') in /opt/lampp/htdocs/mediawiki/LocalSettings.php on line 141 [19:53:36] can any1 help me out [19:56:56] Krinkle: you rock :-] [19:57:14] Krinkle: might be time to make an announce on wikitech-l / mediawiki-l ? [19:57:19] still waiting for one final fix [19:57:22] jsduck is still broken [19:57:35] I duck punched it locally on gallium to make it work [19:57:44] I duck punched jsduck [19:57:44] ok :-] [19:58:13] hashar: the debian package incorrectly embeds a broken/old version of some dependency. [19:58:26] Krinkle: feel free to send the email whenever you want and skip me from the review :-] [19:58:27] hashar: Which it outputs at https://doc.wikimedia.org/VisualEditor/master/extjs/ext-all.js [19:58:46] So I hijack that url in htaccess [19:58:52] ugly hack :-] [19:59:08] and redirect to the "right" version I dumped in https://doc.wikimedia.org/extjs/ext-all.js [19:59:15] yeah, temporary [19:59:22] Krinkle: I also integrated OpenStack portal for Zuul at https://integration.wikimedia.org/zuul/ :D [19:59:26] Faidon says he'll publish the new package later [19:59:31] I noticed [19:59:32] very nice [20:00:05] I removed the simpler, couldn't figure out how to make it work despite using your fix [20:00:16] "the simpler" [20:00:18] what do you mean? [20:00:27] arh [20:00:31] I removed the FOOTER [20:00:38] maybe we should get a WMF theme for twitter bootstrap :] [20:00:53] There isn't much theme to it. Just html. [20:01:24] we might want to do some basic php templating in integration/docroot/shared to avoid repeating the general html structure [20:01:52] hashar: btw, you're going to be in AMS in May, right ? [20:01:54] hackathon [20:06:08] yeah [20:06:29] to both: php glue to avoid repletion + AMS [20:06:45] repetition [20:06:50] damn apple autocorrecter [20:07:24] I admit, I basically did it already to keep them in sync. I'll polish it up and push it later this week. [20:07:25] dinner [20:09:13] Project browsertests-linux-chrome build #223: UNSTABLE in 7 min 21 sec: https://wmf.ci.cloudbees.com/job/browsertests-linux-chrome/223/ [20:09:13] * zeljko.filipin: Updated Ruby gems [20:09:14] * zeljko.filipin: Update Firefox to version 19 [20:09:14] * zeljko.filipin: Disable a test in Firefox [20:09:15] * zeljko.filipin: modified search.feature to add test for accents search [20:09:15] * zeljko.filipin: find tokens in the captions of uploaded files [20:25:28] csteipp: any reason to be wary about https://gerrit.wikimedia.org/r/#/c/54634/ going out with the CentralAuth deploy? [20:25:58] pgehres: Nope. It would be good to deploy that [20:26:08] awesome. i will simply deploy master [20:28:20] Project browsertests-windows-internet_explorer_9 build #243: UNSTABLE in 10 min: https://wmf.ci.cloudbees.com/job/browsertests-windows-internet_explorer_9/243/ [20:28:21] * zeljko.filipin: Updated Ruby gems [20:28:21] * zeljko.filipin: Update Firefox to version 19 [20:28:22] * zeljko.filipin: Disable a test in Firefox [20:28:22] * zeljko.filipin: modified search.feature to add test for accents search [20:28:23] * zeljko.filipin: find tokens in the captions of uploaded files [20:35:44] cndiv: ping [20:37:15] 15 02:21:35 < jeremyb_> cndiv: you have some directions documented for how to start hangout on air properly? (there was another meetup this evening where they were doing it differently than y'all do) [20:37:19] 15 02:21:52 < jeremyb_> ironically said meeting was at google [20:37:21] 15 02:22:07 < jeremyb_> and the guy i was discussing the wrong link with was a googler. :) [20:49:29] YuviPanda, :There? [20:49:53] Rahul_21: a little, yes [20:50:27] YuviPanda, ok,take your time,ping me later without fail [20:50:42] Rahul_21: nah, tlel me [20:50:47] just expect higher latency responses :) [20:51:20] i have an extension idea,thought i'd share it with you [20:52:55] YuviPanda, something called the user-info [20:53:24] okay? [20:53:26] what does it do? [21:05:33] <^demon> qchris: Soo, I've been waiting and waiting for that hooks-its change to go in. [21:05:37] <^demon> I'm thinking of just merging ;-) [21:05:48] You have merge rights? [21:05:54] <^demon> Yeah [21:05:58] :-))) [21:06:17] <^demon> Heh, done. [21:06:20] By next week I would have asked you whether we should try to get you merge rights :-) [21:06:22] \o/ [21:06:25] Yippie! [21:07:04] Thanks, ^demon [21:07:17] <^demon> yw. [21:07:27] So everything that wmf requires from hooks-is has been merged I guess [21:07:42] <^demon> Ugh, I wish I could get jenkins to work sanely for builds. [21:07:58] Anything I can help with? [21:08:20] <^demon> Maybe :p [21:08:49] <^demon> So, I want the various plugins to build with the artifacts from the latest core build. So before I do `mvn install`, I clear out the installed versions from .m2/repository. [21:09:04] RoanKattouw: can you tell me anything about the order in which a ResourceLoader module applies CSS vs JS -- IE: I have a module which loads CSS and JS and then the JS does something which relies on the CSS having been applied (because it's looking for a width of an element) [21:09:07] <^demon> But I don't think it's picking up the latest builds--some of these failures are weird. [21:09:20] <^demon> https://integration.mediawiki.org/ci/view/Java/ - here's the gerrit builds (you can login with your ldap account) [21:09:22] RoanKattouw: but the CSS has not yet been applied so the JS does funky things [21:09:38] * qchris takse a look [21:10:30] <^demon> Plus, having to mvn package && mvn install is a waste of time, if there was some way to shorten that I'd go for it too :p [21:13:48] mwalker: Asynchronously [21:13:59] mwalker: There is no event for "your CSS is applied now", in any browser [21:14:07] :'( [21:14:14] The qunit test asserting the CSS is applied simply checks every 100ms until it's applied [21:19:47] hey a non-spammy gerrit notification bot [21:20:30] thanks ^demon/qchris [21:20:47] <^demon> Just need a puppet change live, then we'll be all set :) [21:21:27] hm? I'm seeing it working already [21:21:47] <^demon> Yeah, it's a little too spammy right now. [21:21:54] <^demon> Should be getting notifs on all comments. [21:22:08] <^demon> (But we want that off, hence the puppet change) [21:25:12] mwalker: If you need to measure based on the CSS, then something that might work is doing it inside of a setTimeout [21:25:31] ^demon: I cannot find the jenkins job definitions neither in puppet nor in integration/jenkins* Are they available through a repository? [21:25:33] RoanKattouw: ya; that's what I ended up doing [21:25:37] but it's a terrible hack [21:25:48] <^demon> qchris: No, just in gerrit I'm afraid :\ [21:25:58] <^demon> s/gerrit/jenkins/ [21:26:08] Oh. Ok. [21:34:04] ^demon: Probably it's just me being stupid, but how can I access the config.xml files? I find no way to access them through the web UI, nor can I connect to the machine directly using ssh. [21:34:43] <^demon> Hmm, I'll grant you some rights. Figured everyone could at least view the files. [21:34:54] Ok. Thanks. [21:38:58] greg-g: please sir, may I have another 15-minute window to fix a core bug https://gerrit.wikimedia.org/r/#/c/54973/ [21:39:35] hah [21:40:32] pgehres: sure thing, either now or during LIGHTNING DEPOYS! [21:41:00] i'll push it out as soon as jenkins bothers to merge :-) [21:41:07] thanks! [21:41:13] <^demon> qchris: Have you logged into jenkins? [21:41:21] Yes [21:41:28] Logging out and logging in again. [21:41:48] <^demon> Hmm, it's not finding you :\ [21:42:08] ] [21:42:15] ^demon: Damn Gerrit is really slow right now [21:42:17] I am logged in now (QChris) [21:44:11] each module invocation runs in its own sandbox so does not have an influence on other stuff, right? [21:46:05] <^demon> qchris: Ah, there we go. [21:46:20] Jenkins found me now? [21:46:37] <^demon> Yep [21:46:39] <^demon> hehe. [21:46:41] <^demon> Pressing save helps. [21:46:51] :-) [21:47:01] Now I see "configure". Great! [21:47:02] Thanks [21:47:15] <^demon> yw. [21:50:05] <^demon> RoanKattouw_away: Not really sure why. CPU usage is high, but nothing's weird in the logs and the queue's like < 5 items. [22:01:44] greg-g: all done. thanks again [22:02:20] <^demon> RoanKattouw_away: Ahh, actually one of the git-upload-packs was stuck and pegging the cpu. Should be better now. [22:03:04] ^demon: In Jenkins, we're using OpenJdk 7 for gerrit and OpenJdk 6 for plugins. While that's not obviously wrong, it at least rings a bell. [22:03:19] Aha [22:03:37] I'll have a look at the failing plugins tomorrow (the also fail locally against current master) [22:03:38] <^demon> Ah, I was testing 6 vs 7 awhile ago. [22:03:49] <^demon> Weird, they work for me locally :\ [22:04:13] <^demon> RoanKattouw: Also, if you want to look at the queue, `ssh -p 29418 gerrit.wikimedia.org gerrit show-queue -w` [22:04:50] and for the package / install ... As we do not do a clean in between, it should not be a problem. But we could do astraight 'mvn clean install' instead 'mvn clean package' and later mvn install. [22:05:00] (we also do the same thing for the plugins) [22:05:15] Of course we still have to install the API separately [22:10:39] Krinkle: https://bugzilla.wikimedia.org/show_bug.cgi?id=46367 [22:11:16] kaldari: Can you be more specific? [22:11:29] sorry, see the last comment [22:11:30] pgehres: good deal. [22:11:51] kaldari: before anything, this is what I see on first sight: http://cl.ly/image/383r1V271U0t [22:12:03] kaldari: (Chrome, stable, Mac) the "Sort by" text is cut off. [22:12:14] Something has a fixed height that shouldn't and has overflow hidden that shouldn't. [22:13:27] No, actually, someting is being floated without clearing [22:13:33] yes, the css is getting all munged by resourceloader [22:13:41] not sure why yet [22:13:55] I doubt this is rl related [22:14:08] that element needs clear both at the end or overflow-hack to clear it [22:14:36] floating an element inside a non-fixed height element means the floated element will stick out under [22:14:37] that's general css [22:14:58] Krinkle: http://test2.wikipedia.org/wiki/Special:NewPagesFeed?debug=true [22:14:59] anyway, checking the bug now [22:15:08] everything looks fine with debug=true [22:16:40] Krinkle: thanks for the CSS lesson [22:16:57] * Krinkle is trying to figure out where that line break is coming from in debug mode [22:19:02] kaldari: in non-debug mode when I touch the element in the console (not changing anything) it jumps to the right place [22:19:11] where it is in debug mode. [22:19:12] weird [22:19:19] looks like a webkit bug [22:19:31] yeah, the behaviour in webkit is especially weird [22:20:13] although the layout is also changing based on debug mode in Firefox, but not in exactly the same way [22:20:51] haven't gotten to the bottom of it, just just thought you might want to be aware of the problem [22:21:03] in case it could be affecting other things [22:21:58] it's probably something weird in my CSS [22:24:05] Yeah, something that isn't bad, but diferent enough to trigger unpredictable behaviour. [22:24:17] thing is, its usually something tiny [22:24:50] I'd try process of elimination. might take a while, but worth it. I learned some of the deepest thnings with css this way [22:25:17] Sometimes you spend half a day on a bug like this, keep simplifying it in the local working copy. until that one aha moment arives. [22:25:37] when the page is almost blank, almost no css left and there the bug is buzzing right in front of you. [22:25:39] I'll let you know if I find out what's triggering it. Sorry to bug you, just wanted to have you in the loop on the issue. [22:25:49] no worries [22:26:04] It seems firefox specific, the box is fine in Chrome [22:26:11] but no, I wouldn't know what it is right away. [22:26:43] kaldari: I'd look through the other browser tests but they ran at an inopportune time "Error: invalid magic word 'property'" causing all tests to fail. If I run across any other occurrences I'll add them to the BZ ticket. Re-running some browsers now. [22:27:17] thanks [22:28:02] is there any way for a stock IE10 to collect response headers? [22:28:24] I don't have a copy so it's hard to check [22:31:47] TimStarling, open a sniffer at the same time? [22:32:11] this is for https://ticket.wikimedia.org/otrs/index.pl?Action=AgentTicketZoom&TicketID=6908175&ArticleID=8135944&ZoomExpand=0#8135944 [22:32:32] opening a sniffer is not really an option [22:32:56] TimStarling: F12 developer tools [22:33:12] i've got a win8 VM i can fire up if you need to test something [22:33:54] wow, it seems so strange to see 'developer tools' in IE... [22:34:03] thanks brion [22:34:21] I thought it would probably have it [22:34:23] IE is muuuuuuch better than it used to be… until you have to debug shit under ie6/7 ;) [22:34:33] it would be hard for MS to ignore FF and Chrome combined [22:34:59] if only they supported webgl we could make wikipedia into a magic 3d wonderland ;) [22:36:00] hmm looks like i lost my OTRS access, i can't get in to read the ticket [22:37:21] well, it's really short: "Today’s Welcome Page (19th March) is showing yesterday’s Welcome page (18th [22:37:21] March). This also happened about a month ago [22:37:35] ah ….. well that might be hard to test :) [22:37:36] running Windows 8 with IE10 from the UK [22:38:33] I have [22:38:33] done the purges as requested with no effect and also fired up another [22:38:33] computer which hasn't been used for a week so its cache is empty and it [22:38:33] still shows the 18th. [22:38:41] that followup is more interesting [22:39:16] maybe you can point your IE10 to the european proxys? [22:39:18] nothing unusual in response headers when i hit the main page here… gives me a current page from today [22:39:25] hmmmmm lemme try that [22:39:50] wikipedia-lb.knams.wikimedia.org i assume? [22:40:08] wikipedia-lb.esams.wikimedia.org. [22:40:19] * brion durrrr [22:40:28] i'm trapped in the past ;) [22:40:56] well, I didn't remember the datacenter name, either :) [22:41:05] that's why I said 'European' [22:41:45] I hope British Telecomm didn't add a proxy of its own [22:41:47] ugh [22:42:15] no surprises in my response headers from esams either [22:42:42] X-Cache shows hits on amssq43 and amssq35, the date and lat-modified are up to date, shows today's article [22:42:53] for reference the magic recipe is: [22:43:11] * open developer tools (F12 or select 'Developer tools' from menu) [22:43:15] * click to 'Network' tab [22:43:20] * hit 'Start capturing' [22:43:44] then it works much like in chrome or firefox, you pick the request to look into and there's tabs for request headers, body, response headers, etc [22:44:14] can you continue the recipe with the steps to send us those contents? [22:44:36] I just sent a reply [22:44:54] * right-click in the 'response headers' section, hit 'select all' then 'copy'. paste into email [22:45:12] oh, you even found a msdn description :) [22:45:17] oh nice :D [22:45:21] http://msdn.microsoft.com/en-us/library/gg130952%28v=VS.85%29.aspx [22:45:38] * brion bookmarks for future reference [23:32:20] is it possible to find out in nested module by which param it was called in parent template? [23:32:24] like this: [23:33:01] {{parent | foo = {{child}} }} ---> child would output "foo" [23:41:38] Danny_B: i don't think so... [23:42:56] that's bad, that would simplify many things