[00:21:42] http://mysite.science.uottawa.ca/rsmith43/Zombies.pdf [06:54:48] Good morning neilk_ [06:54:55] greetings [06:55:08] i'm ready to go if you are [06:55:33] maybe 5 seconds away from committing a bugfix, though. (decided to be productive...) [06:55:43] OK [06:55:54] I'll just look at what you guys have been up to in the past 12h [06:57:39] nothing core, just the extension [06:59:21] OK [07:00:24] I posted a comment on http://www.mediawiki.org/wiki/Special:Code/MediaWiki/85953#c15941 [07:07:40] reading [07:07:54] ok I just submitted r86029 which should be it for tonight :) [07:08:14] And I just OKed it [07:08:33] Everything is marked OK now except r85953 [07:15:36] neilk_: Could you look at the fixme on r85953, tell me how serious it is, and whether I should deploy anyway? [07:15:49] I am fixing it now [07:16:05] it's not really serious, in fact it dealt with an almost impossible error condition... :) [07:16:12] but I noticed another bug with it whilst testig [07:16:15] testing [07:16:49] Yeah I was wondering myself, how can a title both exist and be invalid [07:20:33] well, you could have a result that the title was existent, but the API might not return the URL of the other page [07:20:42] it is kind of crazy but I have seen worse [07:20:54] anyway, submitted a fix [07:21:20] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/86033 [07:21:57] Is there even a difference between toString and getNameText? [07:22:02] yes. [07:22:12] but I decided it didn't matter [07:22:20] getNameText would give you "Foo bar.jpg" [07:22:31] .toString() gives you PrefixNameText which is "File:Foo_bar.jpg" [07:22:47] Right [07:23:02] OK I'll proceed then [07:25:05] thanks [07:27:18] BTW quick comment about $.fn.collapseToggle: if you ever need to make it work with multiple elements at once, you'll want to wrap it in return $(this).each( function() { ... } ); [07:29:31] yes [07:29:36] I'll make a note [07:30:38] http://test.wikipedia.org/wiki/Special:UploadWizard [07:31:41] do I have to be an admin here? [07:31:47] that actually is a bit problematic [07:31:51] no? [07:31:53] Why? [07:32:05] admins are automatically allowed to overwrite files [07:32:13] but, actually it shouldn't matter now [07:32:20] autoconfirmed have reupload on test [07:32:22] it can be tested as admin [07:35:37] Whee, duplicate detection! [07:36:05] OK this is a usability comment [07:36:14] I uploaded a dupe, and it said it was already there [07:36:24] But the only button I saw was retry these uploads [07:36:36] I had to mouseover the entry to accidentally discover I could delete it [07:38:54] RoanKattouw: yeah. It's a bit of a fail right now. Brandon wants to put big close boxes & I agree [07:39:20] so if you have one failed upload, it's not obvious how to start over [07:39:43] Oh and remote repository and local dupes seem to not be differentiated, that's good IMO [07:39:59] Although for Commons deployment this whole remote repo business is irrelevant anyway [07:40:02] we really have to sit down and rethink some stuff. Ever since Oct or so I've just been making it up as problems arise. [07:40:09] Alright, everything's looking good to me [07:40:17] Wanna push this to the live site? [07:40:26] I'm feeling lucky! [07:40:29] No wait, that's the other company [07:40:32] um [07:40:35] what's our motto [07:40:46] In Scap We Trust. [07:40:48] please [07:40:57] "Hell let's just try it and see what happens" [07:41:02] yeah [07:41:12] (Actually scap is giving me nasty sudo errors right now :( gotta work around that) [07:41:14] I'm a bit silly, it's midnight here and I've been coding quite a bit [07:41:30] I was there 9 hours ago [07:41:30] NOS MORITURI TE SALUTAMUS. [07:41:49] Stayed up coding AFT stuff till 1:30 last night [07:41:54] lovely [07:42:00] how complex is that really? [07:42:11] nm, I know how complex THIS is [07:42:21] Less complex than UW, trust me [07:42:46] oddly, Erik just reminded me that my first take-home exam at the WMF was to design a star-rating system. [07:43:01] I assumed this was just an exercise in design for its own sake, guess not. [07:43:01] But I was writing a script that reads rows from the DB into a 4D array, converts that to another 4D array, then converts that to rows for an aggregate table and inserts them [07:43:22] That 4D->4D mapping shit gets kinda nasty after midnight [07:43:24] Heh, nice [07:43:40] Was this back in July 09? [07:43:43] yes [07:43:54] does our design use histograms or is there one row per vote [07:44:24] I think it uses histograms to visualize avg values [07:44:32] but not to store it? [07:44:40] by histogram I mean [07:44:55] each article has five new columns, 1 star, 2 star, 3 star, etc. [07:45:04] as opposed to articles -> votes <- users [07:45:24] Oh you mean how is it stored [07:45:34] the latter is the most obvious thing but it is super inefficient [07:45:40] The main table stores a 0-5 value per revision per user per rating type [07:46:04] yeah, that's what I was afraid of, I think. [07:46:09] Anyway is scap happening? [07:46:14] Yeah it's done [07:46:17] We should be live now [07:46:28] So we have aggregate tables for per-page and per-revision totals [07:53:21] RoanKattouw: looks all good to me. Thank you. [07:53:41] Sure [07:53:49] yeah, re: the 5-star thing, it was an odd coincidence since I designed something similar for another customer [07:54:13] do we need users to remember their own ratings? when they return to a page? [07:54:38] ^RoanKattouw [07:54:49] Depends [07:54:57] If they're logged-in, we remember them for them [07:55:15] If they're anon, it depends on them still having the same cookie-based anon token [08:00:24] right [08:00:43] well, I need to sleep now. Thanks again, talk with you tomorrow. [08:02:21] Good night [11:07:37] Good morning ^demon [11:08:06] I was wondering, have you ever broken the site before? Like HTTP-500-blank-page-on-all-pageviews broken? [11:08:20] If not, you have now [11:08:37] ;) [11:11:53] *werdna chuckles [11:11:59] also, DC! [11:12:00] \o/ [11:12:04] Yeah [11:12:45] I must say I was really hoping DC would beat SA just because of the distances involved [11:18:03] I know you're probably just gonna laugh at me because you do those trips all the time, but SF - South Africa is similar to Australia-Europe in terms of time spent in planes, and I wasn't looking forward to that much [11:28:41] Hello Roan & Andrew. [11:29:01] Hey guillom [11:31:18] ello guillom [11:31:34] SA is closer to AU [11:32:26] Yeah it's almost the opposite consideration for you [11:32:49] Sydney-DC should take about 20, 21 hours [11:45:20] I've never been to South Africa [11:45:23] it'd be cool for that reason [11:45:30] I've been to DC before [11:46:58] *Bryan is looking for ESA to get to fly him to SA [11:47:02] little chance though [11:47:59] Why would they fly you to SA? [11:50:03] I mean they're the /European/ Space Agency, right? :P [11:50:53] yeah [11:51:05] there is a congress over there [11:51:07] well, NASA won't do it [11:51:17] they don't have the equipment anymore [11:51:31] When are we finding out who WMF is sending to Haifa? [11:51:31] and they might want to have some rockets launched [11:51:39] If I have to book my own tickets, I should do it soonish :p [11:51:44] Heh [11:51:58] I wonder if OSCON has decided on our presentation proposal yet [11:52:53] If that gets accepted and we do a ResourceLoader talk in Haifa too, then given the fact that OSCON is in the week before Wikimania I'll probably end up doing AMS-SFO-PFX-SFO-TLV-AMS [11:53:04] *PDX [11:53:47] *werdna has to take time out of uni for wikimania this year [11:53:49] still worth it [11:54:16] I have to take time off for Berlin, also worth it :) [11:54:37] I'm missing Friday :( [11:55:37] I'll be launching rockets until 15:00 then hurry to the train and arrive around 23:40 in Berlin [11:56:02] "All proposers notified ??? Early March 2011" [11:56:05] Yeah right [11:56:12] These guys are worse than the Wikimania jury [11:56:17] (Which was remarkably timely this year) [12:00:41] :) ah well [12:00:43] you'll be right [12:00:46] hey Reedy [12:00:46] wassup [12:00:51] Morning [12:00:52] ish [12:01:03] :) [12:01:25] ^demon, about? [12:01:47] did the UK drift away so far that it is still morning over there? [12:02:28] Reedy: He hasn't responded to my taunts about him breaking the site earlier :) [12:02:35] lol [12:02:46] (He, or I'm assuming it was him anyway, left conflict markers in CodeReview.php that broke all of mediawiki.org when synced) [12:03:25] lol [12:03:31] Seems path browsing is broken [12:03:41] Seems like the maintenance script hasn't been run [12:03:57] He logged something about updating CR but it looks like the update was only half-done when he left [12:04:07] And I unsuspectingly ran sync-common-all this morning [12:04:14] lol [12:04:39] Ideally 2 maintenance scripts need running, one is for usefulness, one is to unbreak path searching [12:05:04] You have shell access to fenari, right? You can run them then [12:05:14] I was just thinking that [12:05:41] I'll get round to it in a little bit [12:06:55] oh, he only broke MW.org [12:06:58] well that's boring :)( [12:07:14] I overloaded the whole site for about 20 minutes once [12:07:17] you've got that to top [12:07:52] Was that the thing Domas bitched about? [12:07:58] You killing the site while on a plane? [12:08:18] no no, that was just replag [12:08:23] and that was a mysql bug [12:08:33] I'm pretty sure I overloaded the whole thing and brought it down once [12:08:42] I know my first act with commit access was to break blocking [12:08:51] (to be fair, Tim forgot to apply the DB patch, but still) [12:09:47] When my first extension was deployed, it overloaded the site, and completely broke it when it was disabled [12:09:50] (LocalisationUpdate) [12:10:05] To be fair, I didn't write it myself, but I did review and tweak it [12:12:27] you're also guilty for an SQL injection vulnerability once, I think :p [12:14:45] not I, said the little red hen [12:15:21] I don't recall introducing any SQL injection vulnerabilities [12:15:30] But it's quite possible I forgot [12:29:12] <^demon> I am now [12:31:16] <^demon> And I take the blame for never coming back to finish that deploy last night [12:31:29] Populate followup revisions all run [12:32:08] there's now a royal wedding drinking game [12:32:13] I hope you're playing, Reedy [12:32:17] hello ^demon [12:32:19] DC!!!!1 [12:32:27] <^demon> Yepyep! [12:32:49] Reedy: http://www.facebook.com/pages/The-Royal-Wedding-Drinking-Game-2942011/170990552921694 [12:33:40] ^demon, running populate script now [12:34:02] <^demon> Ooh, drinking games are always fun [12:34:22] <^demon> Reedy: populate followups? [12:34:29] <^demon> What about rebuilding paths? [12:34:35] Bah [12:34:37] I meant that [12:36:20] I had to reboot and use Windows to request a press accreditation for an event; the form wouldn't work on Ubuntu. I feel dirty. [12:36:56] "2. Any time Prince Harry appears all players must produce a Nazi salute." [12:36:59] fucking awesome [12:37:31] guillom: Install VirtualBox ;) [12:50:28] Script run [12:52:56] Reedy, RoanKattouw: I'm reading through bash [12:52:57] *werdna lols [12:53:16] :D [12:53:52] One more fucking-bar! [12:53:58] ? [12:54:17] Brussels [12:55:34] tomasz [12:57:34] Hey look, rev numbers is right :D [13:14:04] was there stats somewhere which says how many percent of users come to wikipedia trough search engines? [13:19:46] Nikerabbit: http://stats.wikimedia.org/wikimedia/squids/SquidReportOrigins.htm [13:19:50] No percentages though :( [13:22:12] <^demon> RoanKattouw: To answer your first question? Not 500-on-every-pageview, but certainly have broken stuff :) [13:22:22] heh [13:22:42] Well congrats then, you now have a 500 breakage to your name as well [13:23:01] *RoanKattouw mumbles something about rite of pasasge [13:23:11] <^demon> I'm really not sure how those markers crept in. [13:23:42] <^demon> I resolved the 1 or 2 minor merge conflicts. [13:24:50] http://stats.wikimedia.org/wikimedia/squids/SquidReportGoogle.htm [13:25:00] It would've never broken if scap had worked properly [13:25:04] scap checks syntax but s-c-a does not [13:25:26] However scap is currently mysteriously breaking for me with all kinds of [sudo] Password for catrope: prompts [13:26:06] *^demon blames puppet [13:26:20] <^demon> Puppet makes a good punching bag for server excuses these days. [13:26:23] Is there a way to use JS to put the cursor in the textbox, but not scroll down to it? [13:27:59] has anybody seen http://www.nico.schottelius.org/software/cdist/ already? [13:28:54] <^demon> saper: I haven't. I dunno if ops looked into that at any point. I doubt we're gonna drop puppet anytime soon though :) [13:29:26] ^demon: no idea how it compares to puppet, will be checking out soon, maybe. [13:30:21] <^demon> I still love that the controller for puppet is puppetmaster [13:32:13] Reedy: The scrolling behavior is browser-specific, but autofocus is generally nasty [13:32:24] yeah [13:32:25] e.g. for things like pgdn [13:32:26] I'll kill it [14:52:52] <^demon> I forgot we killed verified. [14:53:45] heh [14:55:37] I forgot we forgot [17:29:45] alolita: Who was the person halfway saying "Hi Roan" when I hung up? :) [17:49:48] 10 GSOC proposals to look at rating... [17:49:53] hi Reedy [17:49:57] hey [17:50:17] *Reedy stops procrastinating again for a while [17:53:53] ping alolita & mmerz [17:54:04] ping :-) [18:02:52] hi alolita ;) [18:03:04] hi alolita [18:21:54] bye alolita [18:26:27] bye alolita.. [19:28:12] awjr, I think the idea of averaging the gsoc stuff make sense [19:28:23] otherwise it's just arbitary numbers... [20:36:01] exams done, back to coding :p [20:38:45] wheee [20:49:51] Alright, so AFT deployment [20:49:56] Let's do this [20:50:42] 1) merge to 1.17wmf1 2) deploy to test, including schema update and script and all, will be a nice dry run 3) turn off AFT on enwiki 4) update AFT on enwiki and run the scripts 5) reenable [20:52:26] sounds good [20:57:55] Pfft it looks like I haven't updated AFT to trunk in a long long time [20:58:15] Oh, not so bad, March 14 [21:00:09] Hmm, bah [21:00:23] Too much cherrypicking in the past, I'm throwing it out and rebranching [21:03:01] Hmm, the only schema change is the new table, right? [21:03:58] yeah [21:04:03] should be [21:04:33] Reedy: glad you agree - particularly if we don't read /every/ single proposal [21:04:59] I added it in for reference [21:05:02] Diffing to make sure [21:05:14] Yup, diff confirms it [21:06:48] TrevorParscal: Did I need to change something in the af_ratings table, you said? [21:08:21] TrevorParscal: We should be live on test otherwise, please poke [21:13:01] yeah, it's not critical [21:13:14] it's 4 rows, and it affects the API output, but nothing else [21:13:18] and nothing actually depends on them [21:13:33] but it would be nice if they were updated to have the new i18n keys [21:13:45] I can do that, just tell me how [21:13:53] In the meantime, you and howief should bang test.wikipedia.org [21:14:03] Hmm, bang /on/ it maybe [21:14:11] *RoanKattouw pamphlets self [21:15:23] http://test.wikipedia.org/wiki/Article_Feedback_Test [21:16:55] just went through the flow and created an account [21:17:25] and edited the page [21:17:51] does someone have IE loaded? [21:17:58] *RoanKattouw doesn't [21:19:10] Looks alright on a quick test from IE8 [21:19:19] did the ratings number increment? [21:19:52] well, expanding the "I am knowledgeable..." thing goves over the categories and the ReaderFeedback stuff [21:20:14] so after you submit a rating [21:20:30] it also seems a bit delayed at collapsing back [21:20:31] and then click on "view ratings" [21:20:39] sorry "view page ratings" [21:20:46] do you notice the number of ratings go up by one? [21:21:14] aye [21:21:15] no worries if you haven't noticed -- you may have done this quickly and not noticed [21:21:25] you can delete your ratings and they go back down [21:21:48] ah [21:21:50] and then if you rerate [21:21:55] do they go back up in ie? [21:22:03] ya [21:22:07] aweseom [21:22:32] When you guys are ready to go to enwiki let me know [21:22:52] can someone check to see if the click tracking is working? [21:23:02] i created an account and edited a page after getting the CTA [21:23:11] just wondering if this info is actually being captured [21:25:47] TrevorParscal: What kind of clicktracking key should I look for [21:28:14] ext.articleFeedback.* [21:29:03] Yeah figured that out by myself [21:29:09] I see a few entries, all very recent [21:29:24] Looks good to me [21:29:25] is the actual account name logged? [21:30:05] and can we confirm that reedy's rating went through in to the db? [21:30:13] It's ... not [21:30:23] We just log loggedin/anon and some edit count data [21:30:49] (reedy, would it be possible to submit the rating in a way that would be uniquely identifyable?, e.g.., specific combo of the 4 ratings?) [21:30:56] I see three unique session IDs [21:30:58] ok [21:31:07] One of them logged in halfway [21:31:16] I did them all as anon i think [21:31:19] can you confirm we capture the fact that an account was created? [21:31:21] Oh, no I'm looking at the clicktracking data [21:31:24] Ahm [21:31:31] pitch-signup-accept [21:31:43] is there at least one record of an accept? [21:31:52] Presumably that means you clicked the "you can create an account" link [21:32:04] It doesn't look like we actually register the fact that an account was really created [21:32:09] And I don't remember any code going in for that [21:32:11] TrevorParscal: ? [21:32:11] ah [21:32:15] wait [21:32:20] so what does "accept" actually mean then? [21:32:40] also, are we showing more pitches than accepts? [21:32:43] I think it means they click the link in the CTA [21:32:53] ok [21:33:05] sorry [21:33:09] reading [21:33:13] Basically the flow is pitch-edit-show -> pitch-edit-accept -> pitch-edit-save-attempt -> pitch-edit-save-complete [21:33:13] k [21:33:18] (for edits) [21:33:23] ah [21:33:24] accept means you clicked the green button [21:33:26] For account creation we're missing the final two [21:33:35] I also see reject events [21:33:36] do you see at least one record for pitch-edit-save-complete [21:33:41] Yes [21:33:45] you have to update the click tracking stuff roan [21:33:45] great [21:33:59] well, hmm - it should be up to day [21:34:00] *date [21:34:02] TrevorParscal: Update what from where [21:34:15] nevermind [21:34:25] i misread you [21:34:38] Reedy: would you mind sending a uniquely identifiable rating via ie? [21:34:39] howief: For your session I see join-show, signup-accept, edit-show, edit-accept, edit-save-attempt, edit-save-complete [21:34:41] for account creation, we indeed, do not have hooks that conintue tracking users [21:34:55] Well Howie just created an account [21:34:58] yup [21:35:04] His logged_in bit went from 0 to 1 but his session ID stayed the same [21:35:10] yes [21:35:14] that makes sense [21:35:21] that's good [21:35:29] whats the successful account creation path? [21:35:34] 1, 2, 3, 5. Knowledgeable, 1 and 4 [21:35:41] great [21:35:48] do you guys see reedy's record? [21:35:53] Let's see [21:35:56] That's the ratings table [21:36:16] (Reedy what version of IE are you on?) [21:36:24] 8 [21:36:27] thx [21:36:27] Logged in [21:36:29] ? [21:36:36] nope [21:36:48] 77.86.68.220 [21:36:53] 77.86.68.220 [21:36:56] lol [21:36:59] Yeah I see your rating [21:37:02] fantastic [21:37:15] Reedy: would you mind logging in and resubmitting? [21:37:30] | 97218 | 77.86.68.220 | iKLIobxQCipkKUaphgBcLX486fkMPrMw | expertise | 0 | general|studies|other | [21:37:44] roan is watching you man! [21:39:14] (i'm going to rate an article on chrome with ratings 1,2,3,4 and "i am highly knowledgeable" (only) checked [21:39:28] on chrome [21:39:38] Logged in and moar ratings [21:40:33] i just submitted the 1,2,3,4; knowledgeable rating and subsequently edited the page [21:41:06] | 97218 | Reedy | | expertise | 0 | general|studies|profession|hobby|other | [21:41:18] Ratings 5,4,3,1 [21:41:35] howief: Username? [21:41:40] sounds right [21:41:41] anon [21:41:46] Oh [21:41:54] 216.38.130.162 [21:42:10] Ratings 1,2,3,4 [21:42:16] Checked expertise but no specific boxes [21:42:21] Yeah I recognized the office IP range [21:42:22] awesome [21:43:07] I see an edit pitch leading to a save at :46 , and a survey pitch at :49 [21:43:15] that's prob me [21:43:18] do you see an edit complete? [21:43:19] Yeah [21:43:22] Yes [21:43:28] awesome! [21:43:41] i'm going to go through th eflow and create an account now [21:43:53] (gonna open up a private chrome browser) [21:44:11] You already tested that didn't you [21:44:39] i'm testing again [21:46:02] and i'm actually not clear from the previous whether the successful account creation is actually captured [21:46:37] It's not [21:46:43] ok [21:46:45] Because that wasn't implemented [21:46:51] If you want it, bother Trevor [21:47:04] ok thanks for clarifying [21:47:18] i think we're good to go then [21:47:37] one last question [21:47:52] actually never mind [21:47:55] we never added that [21:48:09] it was on the bonus section, but it's not bonus time right now I guess [21:48:13] no worries [21:48:22] how many click-tracking sessions do you guys see/ [21:48:23]