[00:07:12] <stwalkerster>	 csteipp, are you around?
[06:20:30] <Nemo_bis>	 There are no more bits issues, are there?
[08:46:43] <legoktm>	 Nemo_bis: ^
[08:46:51] <twkozlowski>	 yay legoktm
[08:47:12] <legoktm>	 I probably should have checked that the site was actually up before doing that :P
[08:48:31] <Nemo_bis>	 twkozlowski: the overload lasted perhaps 20 min, but network stayed higher than normal for few hours I'd say https://ganglia.wikimedia.org/latest/?r=day&cs=1%2F9%2F2014+18%3A11&ce=1%2F9%2F2014+21%3A14&m=cpu_report&s=by+name&c=Bits+application+servers+eqiad&h=&host_regex=&max_graphs=0&tab=m&vn=&hide-hf=false&sh=1&z=small&hc=4
[08:48:31] <twkozlowski>	 legoktm: You would know if it was down.
[08:48:43] <Nemo_bis>	 not necessarily
[08:49:42] <twkozlowski>	 Nemo_bis: He /would/ know.
[08:57:55] <Sir_Designer_>	 odder gratulacje za naswietlenie tematu edytowania za forsę w WMF
[08:59:34] <twkozlowski>	 Sir_Designer_: to tylko anglojęzyczne media chyba
[08:59:46] <twkozlowski>	 Sir_Designer_: plus nie wiedziałem, że jesteś Polakiem / znasz polski :)
[08:59:52] <twkozlowski>	 niespodzianki niespodzianki
[08:59:54] <Sir_Designer_>	 ale Ars Technica,  heh heh
[09:00:01] <Sir_Designer_>	 albo arse, as I call it
[09:00:13] <twkozlowski>	 ah, /whois is your friend :)
[09:00:35] <twkozlowski>	 Sir_Designer_: aż takie złe że arse?
[09:00:36] <Sir_Designer_>	 twkozlowski nie wyg-ʉpiaj się,a lbo wyjmij język z policzka :/  /ns info
[09:00:51] <twkozlowski>	 Sir_Designer_: właśnie sprawdziłem :-)
[09:00:58] <Sir_Designer_>	 twkozlowski na odwrǿt.  b. dobre.  celujące, nawet.
[09:03:28] <Sir_Designer_>	 twkozlowski a to w temacie wyjaśniania mojego nowego nicka  http://share.shutterfly.com/share/received/welcome.sfly?fid=8716ad8f412f3a56&sid=0AasXDJu0as27mY
[09:04:32] <twkozlowski>	 Sir_Designer_: :-)
[09:05:09] <twkozlowski>	 Sir_Designer_: well, na pewno artykuł na Ars jest dużo bliższy prawdy niż ten na Daily Dot
[09:13:35] <Sir_Designer_>	 Ars jest pisana, nawet kiedy blogowato, rzetelnie i fachowo przez (u)znane autorytety, tak więc...
[11:44:00] <mutante>	 !summon Thehelpfuloe
[11:44:07] <mutante>	 one
[11:52:12] <Nemo_bis>	 WARNING: after wandering whole heaven and hell, we confirm that you tried to summon a non-existing soul. Satan is not amused.
[11:52:36] <mutante>	 Nemo_bis: *gg*
[12:09:07] <arav93_>	 Could someone tell me where I can find the bug mentioned in https://bugzilla.wikimedia.org/show_bug.cgi?id=45771 in the code?
[12:12:17] <mutante>	 arav93_: since it says Component:Special Pages, a rough guess would be  [mediawiki/core.git] /includes /SpecialPageFactory.php
[12:13:05] <mutante>	 maybe https://git.wikimedia.org/blob/mediawiki%2Fcore.git/HEAD/includes%2FSpecialPageFactory.php  but not a dev
[12:13:26] <mutante>	  Factory for handling the special page list and generating SpecialPage objects.
[12:28:26] <arav93_>	 Could you please type the exact command? I'm a newbie here.
[12:31:51] <mutante>	 arav93_: git clone https://gerrit.wikimedia.org/r/p/mediawiki/core.git
[12:32:04] <twkozlowski>	 What does https://gerrit.wikimedia.org/r/#/c/106537/ mean?
[12:32:54] <mutante>	 twkozlowski: looks like it means that commons is temp. back to old search system
[12:33:19] <twkozlowski>	 Does it have any consequences for me as a user?
[12:33:30] <twkozlowski>	 wondering if we should feature this in Tech News #3
[12:33:54] <mutante>	 no wait, looks more complicated. see "commons is special" comment.. hrmm
[12:34:13] <mutante>	 i don't know, let's ask search people
[12:34:32] <twkozlowski>	 they're asleep :)
[12:35:18] <mutante>	 so there is no extra index for the File: namespace
[12:35:29] <mutante>	 is what it says, right
[12:35:57] <mutante>	 $wgCirrusSearchExtraIndexes[ NS_FILE ] = array( 'commonswiki_file' );
[12:36:30] <mutante>	 that would mean searching for stuff in File: NS would be slower than before
[12:36:48] <twkozlowski>	 okay...
[12:36:52] <twkozlowski>	 thanks mutante :)
[12:37:47] <twkozlowski>	 also mutante, while I have you here :-P
[12:38:00] <twkozlowski>	 any ideas what caused the bits problems yesterday?
[12:39:40] <mutante>	 sorry no, i'm not up2date on that
[12:39:52] <mutante>	 do you mean ULSFO?
[12:39:54] <mutante>	 or other
[12:40:17] <twkozlowski>	 eqiad
[12:40:31] <twkozlowski>	 https://ganglia.wikimedia.org/latest/?r=day&cs=1/9/2014+18:11&ce=1/9/2014+21:14&m=cpu_report&s=by+name&c=Bits+application+servers+eqiad&h=&host_regex=&max_graphs=0&tab=m&vn=&hide-hf=false&sh=1&z=small&hc=4
[12:41:22] <mutante>	 hmm, no, i don't
[12:43:24] <mutante>	 twkozlowski: unrelated note. PDF has issues
[12:43:30] <mutante>	 pdf1 died
[12:43:55] <twkozlowski>	 Yes, I noticed in the git log
[12:44:08] <mutante>	 "pdf2 and pdf3 are taking a heavy beating" hrmmm
[12:45:13] <twkozlowski>	 the Wikipedian in me cries there is no source to use for the bits load spike :-(
[12:45:58] <mutante>	 twkozlowski: re: PDF we still have hope the dc tech might save it.  
[12:46:17] <mutante>	 yes
[12:47:15] <mutante>	 wanna bring it up on list? bbl .. lunch
[12:47:52] <twkozlowski>	 mutante: not sure it's newsworthy for a public like ours on Tech News
[15:19:13] <Tpt>	  hashar: Hi! ProofreadPage has now 1.23wmfX branches created automatically. https://gerrit.wikimedia.org/r/#/admin/projects/mediawiki/extensions/ProofreadPage,branches Does it change anything to the backport procedure?
[15:24:17] <hashar>	 Tpt: I have no idea
[15:24:26] <hashar>	 Tpt: I guess you have to cherry-pick your changes from master to the wmf branch
[15:24:37] <hashar>	 and I have no idea how the extensions are now deployed in production
[15:26:38] <Tpt>	 hashar: Ok. Thanks. I have two questions about testing of file based features.
[15:26:39] <Tpt>	 1) About https://gerrit.wikimedia.org/r/#/c/106507/ I haven't managed to create a fake RepoGroup based on a file directory. Do you know if there is an easy way to do it?
[15:27:47] <Tpt>	 The test file is: https://gerrit.wikimedia.org/r/#/c/106507/1/tests/includes/FileProviderTest.php
[15:33:08] <hashar>	 ahhh
[15:33:29] <hashar>	 the fake filesystem is a bit crazy, I think I coded them to avoid hitting disk
[15:35:13] <saper>	 hashar: isn't operating system vm machinery responsible for deciding what goes ondisk or into the memory those days?
[15:41:06] <hashar>	 saper: we don't use vm :(
[15:43:22] <hashar>	 Tpt: my use case was preventing parser tests from creating files on disk https://gerrit.wikimedia.org/r/#/c/61276/
[15:43:59] <hashar>	 and I do  new MockFileBackend( )   instead of new FSFileBackend()
[15:45:06] <Tpt>	 hashar: Thanks :-) I'll use it.
[15:46:00] <hashar>	 Tpt: AaronSchulz can help there
[15:46:03] <Tpt>	 I've also tried to add to core a parser test for the [[File:|page=Xx]] parameter but I haven't managed to load the test DjVu file: https://gerrit.wikimedia.org/r/#/c/102475/
[15:46:11] <hashar>	 I must say I not fully understand how our FS system works
[15:49:24] <Tpt>	 hashar: I'll ask Aaron. Thanks for you help :-)
[18:10:47] <mutante>	 http://guillaumepaumier.com/timelines/wikipedia-in-2013/
[18:43:40] <Nemo_bis>	 +1
[20:51:39] <effeietsanders>	 hello
[20:52:02] <effeietsanders>	 when logged in, my mw-content-text is maxed on 715px (Chrome, Windows)
[20:52:42] <effeietsanders>	 does anyone know what could cause that?
[20:53:16] <se4598>	 effeietsanders: beta feature - Typography refresh
[20:54:00] <effeietsanders>	 se4598: what does that mean? will it disappear, do I have to do something to make it go back to normal?
[20:54:10] <effeietsanders>	 (I'm guessing it's a bug and not a feature)
[20:55:00] <Excirial>	 effe: have a look at https://en.wikipedia.org/wiki/Special:Preferences#mw-prefsection-betafeatures
[20:55:04] <effeietsanders>	 it causes the page content to be squeezed into a small column :)
[20:55:16] <Excirial>	 "Typography refresh" is one if the features listed there :)
[20:55:21] <se4598>	 effeietsanders: It's "by design" (original quote) since last update. You have to disable the rule in your user-css or the beta feature itself
[20:56:38] <effeietsanders>	 ok, so I just have to disable the typography thing then
[20:56:58] <se4598>	 effeietsanders: see https://www.mediawiki.org/wiki/Talk:Typography_Update#Max_width_.2F_Narrow_pages_opt_out_option for a css-rule to disable the width
[20:57:30] <effeietsanders>	 se4598: I just disabled the whole thing. For beta features, I rather experience the real feature, or not at all :)
[20:57:36] <effeietsanders>	 otherwise I'd just be fooling myself
[20:57:50] <effeietsanders>	 (the reason I enabled them at all is to see what is coming up for new users)
[20:58:11] <effeietsanders>	 thanks for helping me find the source! :)
[21:19:01] <Tpt>	 may someone approve a new oAuth consumer, please? https://www.mediawiki.org/wiki/Special:OAuthListConsumers/view/eec12bb221b2e8e4e5dafbbacf949505
[21:19:16] <Tpt>	 It's only a modified version of https://www.mediawiki.org/wiki/Special:OAuthListConsumers/view/efb0f3953364514e15ff4d4ac0dc881a
[23:14:02] <dschwen>	 hi ori! I have just drafted a Schema for an EventLogging run
[23:14:24] <dschwen>	 at https://meta.wikimedia.org/wiki/Schema:CommonsCategoryTreeUse
[23:14:33] <dschwen>	 how should I proceed?
[23:16:51] <ori>	 hey dschwen, that's pretty cool
[23:18:16] <ori>	 so, eventlogging will provide a means for you to dispatch the data to our servers for analysis, but you still need to have a bit of javascript code that generates the event object on some codition
[23:20:41] <ori>	 for impression, it'll be something like: if ( mw.config.get( 'wgCanonicalNamespace' ) === 'Cateogory' ) { mw.eventLog.logEvent( 'CommonsCategoryTreeUse', { isAnon: mw.user.isAnon(), action: 'impression', pageNs: mw.config.get( 'wgNamespaceNumber' ) } ); }
[23:21:05] <ori>	 though you don't want to generate an event on *every* impression, that could be a lot
[23:21:24] <dschwen>	 sure
[23:21:38] <ori>	 the code needs to live somewhere; for one off data-collection jobs the natural place is the WikimediaEvents extension
[23:21:48] <dschwen>	 I would have to bind click handlers to the links as well?
[23:22:13] <ori>	 capturing link-clicks is trickier. there's no nice way of doing it so i never added it to eventlogging
[23:22:37] <dschwen>	 Couldn't I just put the code into Mediawiki:Common.js (on commons)
[23:22:48] <ori>	 no, because there's a php snippet too
[23:22:53] <dschwen>	 Oh
[23:22:59] <ori>	 if you click on the red icon in the top right on the schema page you'll see it
[23:23:12] <dschwen>	 ok, but I cannot do that, can I?
[23:23:23] <ori>	 you can submit a patch and ask to have it merged
[23:23:24] <dschwen>	 with gerrit access maybe?
[23:23:26] <dschwen>	 ok
[23:23:31] <dschwen>	 hm
[23:23:53] <ori>	 you can also log document.referrer on category pages instead
[23:23:58] <ori>	 if you want to know how people are getting to them
[23:24:12] <dschwen>	 The problem with link clicking is that asynchronous operations are killed as soon as the link is followed, right?
[23:24:35] <dschwen>	 I could add a URL parameter to each link and fire the event upon arrival at the new page
[23:25:00] <dschwen>	 or the referrer
[23:25:08] <ori>	 you could; you could also cancel the click event, log the link-click, then change window.location
[23:25:19] <dschwen>	 yeah
[23:25:20] <ori>	 but it's easy to screw up and if you do it breaks things rather profoudnly for the user
[23:25:25] <dschwen>	 yep
[23:25:51] <ori>	 but in general i recommend that you write up a short description of what you want to do and e-mail it to the wikimedia analytics mailing list
[23:26:08] <ori>	 because while we don't have the full request log anywhere just yet we can capture specific slices of it
[23:26:09] <dschwen>	 the stuff that is on the schema talk page?
[23:26:31] <ori>	 ah yeah, that's great!
[23:27:09] <ori>	 i have to run in a moment, but if you get stuck you can email me directly at ori@wikimedia.org
[23:27:45] <MatmaRex>	 "<ori> you could; you could also cancel the click event, log the link-click, then change window.location" please don't
[23:28:08] <MatmaRex>	 this breaks middle-clicks, shift-clicks, ctrl-clicks, right-clicks and god knows what else when not done right
[23:28:19] <ori>	 dschwen: MatmaRex is pretty knowledgable about front-end code too so I'd be inclined to take his advice :)
[23:28:23] <MatmaRex>	 and it's hard to get right, it seems
[23:28:35] <MatmaRex>	 since i've seen it broken a few times already :D
[23:28:56] <ori>	 i think you can get this via document.referer on the actual category page or by filtering the request log on the server side
[23:29:07] <ori>	 probably best to discuss on the analytics list