[00:08:04] we need an emergency configfile change to disable archiving of feedback on enwiki [00:08:08] its going absolutely crazy [00:09:13] jackmcbarn: Crazy how? [00:09:23] look at recent changes on enwiki [00:10:03] They're all grouped [00:10:17] it's archiving the same posts over and over [00:10:24] and recent changes is completely unusable [00:10:52] Hmm, you're right [00:10:55] Didn't look that closely [00:10:57] give me a minute [00:11:55] So it looks like archiving is actually broken then [00:12:03] yes [00:12:14] Disabled for starters [00:12:18] and by the way, i mistyped a url while looking for the setting and discovered this: https://noc.wikimedia.org/conf/highlight.php?file=InitializeSettings.php [00:12:22] i dont think i should be able to see that [00:12:35] It's a joke [00:12:43] It's not a password ofr anything [00:12:57] ohj [00:12:58] oh [00:15:43] jackmcbarn: https://bugzilla.wikimedia.org/show_bug.cgi?id=54197 [00:16:15] thanks [00:47:39] https://en.wikipedia.org/wiki/Special:CentralAuth/Article_Feedback_V5 <-- it's being used as a system account but some user already owns it on de.wiki... [00:47:54] wtf [00:48:45] lmfao [00:48:48] AWESOME [00:52:46] Hmm [00:53:15] eh [00:53:22] The account doesn't even exist on enwiki.. [00:57:50] yeah [00:57:59] but i cant create it since its taken in SUL [00:58:25] when there's a will, there's a way [00:59:59] you can create it! [01:01:05] they should be using https://www.mediawiki.org/wiki/Manual:Hooks/UserGetReservedNames [01:01:28] can someone delete Basketball Manitoba? its one-week grace period expired a long time ago [01:01:46] wrong channel [01:01:59] Hmm [01:02:07] I wonder if they've variablised it [01:02:51] if ( $row->log_user_text == 'Article Feedback V5' [01:02:53] That's a no [01:03:12] legoktm: Of course, it only works in preventing it on wikis with AFTv5 [01:03:17] Until it's in CA [01:03:31] hm, right. [01:03:56] so that hook would need to be enabled globaly [01:05:55] I've added the hook subscriber for local useage at least [01:14:57] * Reedy looks up the email address [01:16:06] Oh look, not on set [01:16:24] *not one set [01:17:12] I'm going to change the password and log into enwiki with the account then [01:21:36] !log Changed password for global user "Article Feedback v5" and logged into enwiki to create an account [01:21:41] Logged the message, Master [01:23:15] I suspect that's "enough" to fix that bug [01:47:37] heh [01:47:52] are you sure that it was owned/created by a staff member and not some user? [01:48:38] https://en.wikipedia.org/wiki/Special:Log/Article_Feedback_V5 i guess the log entries need to be reassigned? [09:09:22] sigh, do we lack even a math stuff tracking bug? [09:09:47] Make one! [09:11:30] most of the maths stuff should be under the one area... [09:16:46] it's all one extension [09:25:11] thedj: but not all problems are in the extension [09:25:59] e.g. there are the MathJax adoption things [09:26:04] or https://bugzilla.wikimedia.org/show_bug.cgi?id=27574 [09:26:21] and about 300 more bugs in bugzilla which I have no intention to look at today [09:31:27] or this ancient bug which is in Math but from the summary seems Wikimedia only https://bugzilla.wikimedia.org/show_bug.cgi?id=4915 [09:32:46] we have so many requests of new formats that we obviously don't bother listing them https://bugzilla.wikimedia.org/show_bug.cgi?id=29461 [09:34:17] * Nemo_bis closes the tabs whose URL were donated to the commons aka IRC logs [10:28:03] back feature be blessed [10:46:27] currently, the disambiguations implementation is a bit broken, right? [10:46:37] i can't access any of the special pages [10:46:47] but they are still listed [10:51:41] I wonder if the extension is setup... [11:01:40] apergos, parent5446: hello [11:01:44] hey [11:02:10] so I should report publically the results of the testing I've been doing [11:02:53] converting the 3rd de pages meta history bz2 to the new format (writing stubs and pages) took 3 days 4 hours [11:03:13] with nothing else running on the box, but the reads and writes were (as all dumps are) to an nfs-mounted filesystem [11:03:36] the ganglia graphs showed quite a lot of wio for the nfs server during that time [11:04:09] network is not the issue, as atop shows a peak of only 23% for either interface (it's got two eth interfaces bonded) [11:04:46] so it's disks. and it's reads, afaict, because an attempt to run a smaller wiki, with reads from a local file and writes to the net, looks pretty normal for disk usage [11:05:21] I need to see what can be done about that, as it will adversely impact other dumps jobs if we run like that [11:05:25] What's the usual time for that dump using the old program? [11:05:30] in the meantime the smaller test (local read, nfs mounted write) on fiwiki failed, some bug [11:05:39] (exception), svick is looking into that [11:05:52] parent5446: can't compare because this is cnversion, not dump [11:05:57] no db access [11:06:12] Oh woops nevermind [11:06:14] do you know why would it behave this way with the disks? is there something i could do to fix that? [11:06:39] bz2 to 7z of de wiki part 3 took 1 day and 18 hours [11:06:49] but it just recompresses so that's not a fair comparison either [11:07:01] well I have to see what we can do [11:08:00] for instance, if we used veery large buffers for the reads maybe that would help (dunno)... if that's true, putting a program in between just for that purpse might show a difference [11:08:40] ok [11:09:04] anyways I need to play with that some, and probably ask folks with more knwledge about disk performance for help t [11:09:05] oo [11:13:29] ok, i found the problem in fiwiki: it's about parsing IPv6 addresses [11:13:40] oh my [11:14:27] i read Manual:IPv6_support and followed that "MediaWiki expands all IPv6 addresses to their entire length (however, the number of digits in a single group between colons is always the minimum. For example 2001:db8:: expands to 2001:db8:0:0:0:0:0:0)" [11:14:58] but the address i got was 2002:3e8e:71b::3e8e:71b (notice the ::) [11:15:29] ah [11:15:48] sounds like something for the talk page to me [11:17:51] hmm, Manual talk:IPv6 support doesn't even exist, so it's probably not watched much, but i can try [11:18:52] you culd bz it and there ask if that's deliberate... up to you [11:21:49] hmm, i think it is some sort of bug in MW, the revision is http://fi.wikipedia.org/w/index.php?oldid=72074, but when i click on the IP address, it says the user never made any contributions [11:22:23] the revision is from 2004, maybe IPv6 support wasn't completely done back then? [11:23:29] ah who knows [11:24:19] well, bug or not, your code will have to be able to handle that format since those numbers are in there :-/ [11:24:20] so, i'll ask on the talk page and i think i'll also post it to bugzilla; in the meantime, i will modify my code so that it accepts addresses in this format too (so that you can test fiwiki) [11:24:25] yep [11:24:38] well so that we can eventually have fiwki dumps :-D [11:25:11] because the odds of use going back through and retroactively converting all ips in the logs and wherever else to be correct... pretty slim [11:25:13] *of us [11:25:29] right [11:27:59] so, after i take care of IPv6, i'll get back to compressing comments; if i can't finish that today, i'll put it on hold for now and get to writing documentation [11:28:08] ok [11:31:05] anything else? [11:31:19] hmm not from me, I need to figure out the disk issue [11:31:45] parent5446: ? [11:31:53] I don't have anything else as well, although I'm curious as to the cause of the disk issue. [11:32:13] well the cause of the disk issue is they are overworked, but the q is how I can work around it [11:32:31] as in iostat shows that the reads % shoots way up there [11:33:08] OK well I'll see you guys tomorrow [11:33:15] ok, see folks [11:33:31] see you [12:16:29] and at last, after some 4 hours of bug triaging, I'm ready do file my simple enhancement request :( [12:16:37] (on behalf of someone else even) [16:02:28] Per http://enwp.org/User_talk:EpochFail#Snuggle_login_issue I'm wondering what the login attempt limit is set to. How many fails in what time frame? [16:05:53] T13|needsCoffee: https://www.mediawiki.org/wiki/Manual:$wgPasswordAttemptThrottle [16:11:49] Nemo_bis: is enwp using the default? [16:11:57] check yourself :P [16:13:14] IDR where that is (I had a link once from a question about email throttle limits, but seem to have misplaced it.) [16:14:24] noc.wikimedia.org [16:15:24] Thank you MatmaRex [18:23:18] running scap in a second [19:11:48] hi what would this be the right channel for question about the roll out of echo on NL wiki? [19:12:05] depends on the question [19:12:58] should you get a notification if you someone uses your nick on a page? [19:13:20] Larsnl: usually, the "channel" is "mail to fabrice florin" ;) [19:13:44] Larsnl: i'm not sure what you are asking about? [19:13:56] haha to much effort for me, i'll just try out [19:14:06] Larsnl: you get mention notifications if somebody links your userpage on a talk page and signs what they said with ~~~~ [19:14:08] MatmaRex: in that case there's #wikimedia-ee [19:14:18] (all three parts are necessary) [19:14:24] MatmaRex: great that's what i meant thx [19:14:27] MatmaRex: also depends on your prefs, right? [19:14:42] jeremyb: Larsnl: and yes, you only get them if you didn't disable that [19:15:00] k, is it auto-enabled? [19:15:08] jeremyb: fflorin is on irc only during office hours in my experience [19:15:18] Larsnl: yes, it's enabled by default [19:15:30] great thx for you help MatmaRex and Jer [19:15:32] jeremyb: [19:15:33] actually, i'm not sure. i tihnk it is [19:15:47] oh [19:15:49] it's documented somewhere probably [19:15:50] MatmaRex: i think right now is office hrs in SF. or you mean [[m:IRC office hours]] ? [19:16:07] jeremyb: yeah, i mean the "office" hours :) [19:16:41] well he hasn't been in #-ee as far back as my scrollback goes [19:17:48] Larsnl: i'm trying to find this out, one second :) [19:18:03] great thx a lot [19:18:34] Larsnl: got it. https://www.mediawiki.org/wiki/Echo/Feature_requirements#Defaults_by_User_Group [19:18:52] Larsnl: so yes, mention notifications are enabled for everyone by default [19:19:00] it doesn't really matter, does it? the current default is not necessarily the same as it was a few months ago. just log in to your prefs and check your own settings [19:19:03] notifications tab [19:19:36] jeremyb: there are different default for "new" and "current" users [19:19:39] allready figured that but couldn't find it this quick and letting other find it out for you is easier :P [19:19:40] defaults* [19:19:50] but thx again (i keep on thanking you) [19:20:00] :) [19:30:17] MatmaRex: btw just discovered that -ee is now -corefeatures [19:30:51] Nemo_bis: huh. well, that's at least less punny than -e3 becoming -growth [19:32:26] dunno [19:32:43] they are not channels designed to be found anyway I guess [19:37:47] heh [19:40:25] MatmaRex: oh, hah. so i didn't read the scrollback in -ee then i guess! [19:40:33] and fabrice is online [19:40:37] in -corefeatures [19:43:02] he's never on the channels real people frequent (as opposed to engineers :P) [19:43:09] i think Nemo_bis put it best above [19:43:42] well he can at least be found [19:43:47] anyway, bbl [20:03:16] Aaron|home: so i probably can't convince you to get rid of tables on the grants list, right? [20:04:44] maybe, I ended up changing so much I stopped there as commit [20:04:55] * Aaron|home got stuck with various bug fix diversions [20:06:10] if it would be quick, would you mind doing that? I think that will get it 80% of the way towards whatever ux will want [20:06:48] I'm not quite sure how it should look exactly, so that would be another commit [20:07:36] Hmm... let me play a bit.... [20:08:39] another/separate [20:13:28] Aaron|home: Probably best as a followup change. But I'd say just

Publisher:%username%

Allowed on.... [20:13:39] Reedy: so much wikibase spam in the exception log [20:16:00] Half of the lines in the file have wikibase [20:16:30] 1016 exceptions in the file [20:17:17] 814 contain wikibase in the top line [20:17:30] Handful of deserialise exceptions [20:17:38] The rest being No such revision found [20:18:53] https://bugzilla.wikimedia.org/show_bug.cgi?id=54178 [20:55:18] about to scap [20:55:50] WAIT NOOOOOOOOOO [20:55:57] just kidding, go ahead [21:15:10] kaldari: would you be comfortable taking a look at a patch relating to centralnotice / translation (mostly tagging) today? I can have adam look at it tomorrow but niklas is on a deadline to solve the growing translation performance problem by the end of the week [21:15:47] I can take a look, but I may need you or Niklas to walk me through it [21:15:57] awesome: https://gerrit.wikimedia.org/r/#/c/83980/4 [21:16:28] the basic refresher is that we tag a banner with a revision tag -- which is then picked up by the banner group to submit it into translation [21:16:41] what we werent doing is deleting the old revtags for old revisions of banners [21:16:53] we also had cases where we had revtags but no banners [21:17:19] I fixed the first problem; and the maintenance script will take care of the second when it runs [21:21:48] kaldari: ah; adam is apparently working today; so he looked at it [21:21:50] thanks though :) [21:22:10] mwalker: cool [21:22:30] yeah, looks like it's merged now :) [21:22:51] greg-g: may I sneak in a deployment at 1500? looks like its free on the calendar [21:24:11] mwalker: re the above? [21:24:16] yep yep [21:24:45] mwalker: yessir, you know the drill [21:26:14] awesome; adding to calendar [21:36:37] Hi. Whom could I add to a bug's CC for database query performance-stuff? I am looking for someone who would be able to "fix it" not for someone who is able to profile or is only operating it. [21:37:32] rillke: which bug? [21:37:40] https://bugzilla.wikimedia.org/show_bug.cgi?id=54122 [21:38:17] https://bugzilla.wikimedia.org/show_bug.cgi?id=54122#c4 [21:38:30] is most interesting in this regard [21:39:34] The issue is that the query seems to be that slow that requesting https://commons.wikimedia.org/wiki/User:CommonsDelinker/commands?action=edit often fails with a 504 timeout error [21:39:54] rillke: springle, already cc'd, is the person you want [21:39:59] ... but only sometimes [21:40:09] Thank you. [21:40:23] why don't you want him now? [21:40:23] :) [21:40:35] oh, I see, mis-timed responses [22:02:22] awjr, MaxSem: just to make sure -- y'all done with your mobile deployment stuff? [22:02:46] mwalker, kaldari was managing deployment, i'll let him answer :) [22:03:01] yep, all done [22:03:46] awesome; in that case, to the deploy machine for me!