[03:58:48] anybody around? [03:59:00] I need to know what files in UsabilityInitiative control your fancy toolbar. [11:48:32] RoanKattouw: HI. [11:48:42] Hi [11:48:54] do you know what issues are still remaining with the multi-line expansion for text areas? [11:48:56] Apparently the Opera table header bug is still not fixed [11:49:04] also, good day to you. [11:49:06] Could you check that on a prototype soonish [11:49:10] oh want me to take another look? [11:49:20] Yeah just confirm that Calcey isn't crazy [11:49:24] sure, check prototype and not localhost? [11:49:37] And see if $('#checkboxID').is(':checked') is true [11:49:39] Yes [11:53:50] ok so it's returning true, but it's still not working until i uncheck and recheck the heading option [11:56:04] RoanKattouw: saying your name in case you didnt see my response. [11:56:27] You mean that one also returns false initially, even though the checkbox is checked? [11:56:40] it returns true initially [11:56:44] like your code should be working [11:56:55] BUT it's still not [11:56:58] Hmmm [11:57:27] it still fails to insert the header row unless you recheck the header option [11:57:42] RoanKattouw: want to reassign that bug to me? [11:57:59] since i've got opera and then you can worry about other things [11:58:26] Sure [12:01:21] RoanKattouw: you're gonna kick yourself :) [12:01:42] see anything you may have wanted to remove from this line? [12:01:43] $j( '#wikieditor-toolbar-table-dimensions-header:checked' ).is( ':checked' ) [12:02:22] ARGH [12:02:39] *RoanKattouw starts kicking himself [12:03:01] nkomura: I never got a definitive answer to this: are we gonna do today's team meeting at 10am PDT or 11am PDT? [12:03:44] 10am [12:04:20] OK [12:04:39] I hope I can make that, I don't know yet [12:05:45] Hm seems like I will make it [12:10:37] RoanKattouw: thank you for accommodating the oddness for two weeks [12:10:57] :) [12:11:07] I'm still confused all the time [12:11:10] pushing out meetings here have schedule conflicts with other meetings and availability of the meeting rooms [12:11:19] I'm sure I'll manage to adapt just before having to switch back again [12:11:30] Yeah I understand [12:12:06] i really don't understand what we gain by changing time two weeks earlier [12:12:44] I think you change back at the same time though [12:12:54] do we? [12:13:00] *RoanKattouw looks it up [12:14:59] Hm actually you don't, you go back one week later [12:29:22] crazy isn't it [13:00:42] whoa, the hold music is extra good today [13:06:42] Should I call in yet? [13:08:40] I think they started, yes [13:09:28] RoanKattouw: yes [13:31:49] http://eiximenis.wikimedia.org:9000/oSMFYtrAWU [13:35:07] https://bugzilla.wikimedia.org/show_bug.cgi?id=22815\ [13:44:51] nkomura: hi [13:45:04] the meeting starts in 15 minutes, right? [13:45:52] No, it started 45 mins ago [13:46:10] I did request for it to be at 7 but I asked Naoko earlier and she said it'd be at 6 [13:46:35] shoot [13:46:42] summertime [13:46:52] Yep [13:47:09] so are u still at the meeting? [13:47:15] than I ll call in [13:47:15] It'll be at 6 again next week, and the week after that we'll be in summertime too so it's back to 7 [13:47:17] Yes [13:47:23] Go ahead and call in [13:47:52] nkomura: http://usability.wikimedia.org/wiki/Usability_study_features --> please get folks to update this when the decisions have been made [13:48:51] k [13:49:09] hi hannes-_- [13:49:41] hi nkomura I just joined the conference [13:57:30] http://collab.wikimedia.org/wiki/Edit_Page_Vacuuming [13:58:02] thanks [14:02:07] http://upload.wikimedia.org/wikipedia/usability/7/7f/100120_vacuuming_06.jpg [14:02:15] we are looking at this page [14:08:58] What the hell [14:09:03] hey [14:09:08] "The host has ended the conference" [14:09:09] secret talking now [14:09:11] maybe they ran out of $$$ on the conference line [14:09:14] same here [14:09:17] "You will now be disconnected. Bye. [click]" [14:09:30] nkomura: ---^^ [14:09:55] or maybe thats just a feature of the conference line. you get ten minutes after the end of the meeting, then it ends it for you [14:10:18] haha great feature [14:10:30] my pc needs that feature too [14:10:32] I would supplement that with a 10-min and 5-min warning [14:10:49] the line was cut off [14:10:57] are you guys there [14:10:58] ? [14:11:06] we re here [14:11:06] adam_miller and RoanKattouw [14:11:22] shall we call back in? [14:11:28] yes [14:11:30] ok [14:13:37] parutron: http://usability.wikimedia.org/wiki/Usability_study_features [14:13:44] nkomura is gonna update that now-ish [14:30:19] RoanKattouw: has http://prototype.wikimedia.org/deployment-en/ been updated since the 12th? [14:30:38] Yes, I think it was updated yesteray [14:30:45] I'll update it right now if you like [14:31:29] thats ok [14:31:48] i just couldn't replicate the first bug i tried fixing [14:32:29] RoanKattouw: so features per release is captured in this doc [14:32:35] and it has been the case always [14:33:32] Ah yes the Google doc [14:35:20] RoanKattouw: did S&R pass the test? [14:35:33] You mean Calcey's? [14:35:35] I think so [14:35:46] They reported an Opera bug, which prompted us to blacklist Opera [14:36:12] And they reported some vague bug where Find next doesn't highlight the word, while I think it's probably just behind the dialog [14:36:33] Although maybe it needs scrolling too, hmmm [14:36:37] *RoanKattouw annotates on the bug [14:38:12] i suppose you are only referring to S&R wrt blacklisting opera [14:38:14] ? [14:38:26] Yes [14:38:27] Of course [14:38:44] And there was this JS error in IE that I assigned to Adam [14:38:54] adam_miller: That IE JS error should be your top priority IMO, just FYI [14:39:15] RoanKattouw: that's the one i started with and couldn't replicate [14:39:25] Hmm [14:39:29] the link dialog is working fine in IE 7/8 now [14:39:42] Oh is that why you asked about updates to the code? [14:39:47] yes [14:40:01] *RoanKattouw updates code again just to be sure [14:40:11] they logged it on the 12, so i'm thinking something that was deployed since then fixed it [14:47:39] RoanKattouw: good catch http://www.mediawiki.org/wiki/Special:Code/MediaWiki/63823 [14:54:17] RoanKattouw: so - performance [14:54:44] TrevorParscal: http://eiximenis.wikimedia.org:9000/oSMFYtrAWU [14:55:00] I just locally reverted Nimish's rev reverting my offset smartness [14:55:13] With the aim of fixing said smartness so it doesn't cause stuff to jump all over the place [14:55:25] right [14:56:11] RoanKattouw: sounds good...you can use the SF article to check out some of the bounce out stuff b/c I'm certain that afterwards it'll be way faster [14:58:12] could someone verify section copy/paste issues with line breaks are specific to FF3 on linux? [14:59:55] Hm for some reason all my templates got unwrapped right after I edited the comment on top of the page [15:00:05] Does this happen on a clean HEAD as well? [15:01:06] nkomura: for sure it works on Mac and Windows [15:01:16] just tested, again on Mac - all good [15:01:24] cool [15:01:28] the issue is fairly well known now [15:09:22] TrevorParscal: can we blacklist at browser plus OS level? [15:09:57] not right now [15:10:03] k [15:10:06] but the issue is only for iframe [15:10:50] but templates are dependent on that [15:15:44] linux users can't participate remotely anyways [15:16:11] and the screen recording software in the lab doesn't support it either [15:16:31] so, study-wise, this doesn't matter / otherwise we need to get this solved for sure [15:25:45] nimish_g: Could you help me trying to reproduce the text jumping out bug? [15:25:54] sure [15:26:04] I ran svn merge -c -63796 . so my offset black magic is bac [15:26:13] But I'm not seeing anything [15:26:38] RoanKattouw: you know, running jquery.wkiEditor.templateEditor - line #69 once per template isn't exactly "quick" [15:27:09] It creates a damn array [15:27:11] How is that slow [15:27:26] I know that .push() is about five times slower as [i++], I'll poke at that if needed [15:27:31] well, it seems like there's allot of function creation there too [15:27:35] what's the overhead there like? [15:28:05] I think these functions don't need specific scope, maybe breaking them out will be faster [15:28:10] and it also means that there are possibly hundreds of copies of those function [15:28:15] right [15:28:23] i'm going to poke [15:28:34] TrevorParscal: While you're at it, can you make another change [15:28:39] sure? [15:29:04] It's a fix for http://www.mediawiki.org/wiki/Special:Code/MediaWiki/63797 [15:29:12] I'd commit it but my local copy isn't clean [15:29:34] RoanKattouw: I'm gonna gmail you the SF article text minus the giant infobox on top, put that in your local wiki as something, then open it and check changes [15:29:48] OK [15:29:52] TrevorParscal: [15:29:54] what's the fix? [15:29:59] I'll PM it [15:30:03] k [15:43:39] nimish_g: The wikitext you e-mailed has extra linebreaks in it, can you grab it using http://localhost/w/index.php?title=San_Francisco&action=raw and pastebin it? [15:59:24] RoanKattouw: http://pastebin.com/EWxCYE96 [15:59:44] Thanks [16:08:18] RoanKattouw: what browsers is search and replace enabled on? [16:11:46] The usual minus Opera [16:11:57] Grep search-and-replace in Toolbar.js to find the browser map [16:45:54] RoanKattouw: we need your eyes on something [16:46:33] what we are seeing is that any time there are 2 consecutive templates, everything after that is broken (toc, additional template wrapping) [16:47:29] if they are divided by a \n they will be individually capsulized, but if one is directly after the other they get merged in the wrapping [16:47:40] can you reproduce this / help us debug it? [16:56:59] I'm back now, sorry for the absence [16:57:14] I'll start debugging that now [16:59:32] Interesting [16:59:52] For short templates this produces an empty preceding the template rather than wrapping it, and adds a newline (dirty diff) [17:00:05] pre-r63796 things were working better [17:00:14] Or wait [17:00:18] It doesn't introduce a newline [17:00:20] TrevorParscal, RoanKattouw: how well was the search and replace dialog working before? it's giving me problems on long articles in everything but firefox [17:00:35] it seems this-> http://svn.wikimedia.org/viewvc/mediawiki/trunk/extensions/UsabilityInitiative/js/plugins/jquery.wikiEditor.highlight.js?r1=63679&r2=63796&pathrev=63823 caused the consecutive template merging issue [17:00:49] adam_miller: Yeah I can imagine performance on IE sucking, esp. for replace all [17:01:16] TrevorParscal: Really? ^^ That's Nimish's revert of my rev [17:01:19] I have that reverted locally [17:02:00] right - so what was the purpose of r3796? [17:02:06] *TrevorParscal is sure there was one [17:02:24] oh [17:02:25] To not make text jump out in weird places [17:02:27] dirty diffs [17:02:30] Yes [17:02:39] Haven't been able to reproduce yet, will use Nimish's provided pastey for that later [17:02:43] oh - yes, 96, cursor jumping issues [17:03:11] well, I was hoping to sit down and optimize but it seems like there's some other things that need to be fixed first [17:03:13] no, it's actually inside the mark function...if you change 'realchange' on line 122 of template editor you get normal behavior [17:03:32] :| [17:04:03] RoanKattouw: I reverted my changes that you had me make [17:04:13] why? [17:04:26] *TrevorParscal forgot to save a diff before running svn update -r##### [17:04:32] i still have the PM [17:04:32] Ah [17:04:36] OK [17:04:41] I can make them and commit them now [17:04:49] if that's easier/faster for you [17:04:51] basically the issue is this: if you comment out line 43 of highlight, templates work great (no cursor jumping etc) but if you comment it out TOC is totally broken. So I was trying to find a way to go "if templates, mark do nothing" [17:04:57] but that doesn't work either [17:05:30] if templates and keypress, rather [17:08:07] RoanKattouw: ok - I just committed those fixes [17:13:57] RoanKattouw: scrollToCaretPosition(). Any ideas for implementing that? [17:14:02] right now i've got context.fn.scrollToTop( context.fn.getElementAtCursor(), false ); [17:14:14] You mean iframe-wise? [17:14:24] Yeah that seems sane [17:14:28] yes sir [17:14:40] well it doesn't always seem to work [17:14:56] What does the false parameter mean? [17:16:00] oh i guess i could change that to true and It would. It's wether it should force the scroll or not if the element is already in view [17:16:35] the problem was with parent elements which were already on screen, but contained more than a line of text. the matched text could be offscreen still [17:19:30] Ah [17:19:42] So that detection needs to be more fine-grained [17:21:12] also, in safari, our setSelection method seems to get stuck somewhere around 15200 characters when the end and start containers are left null [17:21:27] Hm [17:21:53] its really bizarre, works fine up until that point then just keeps selecting the same thing [17:33:22] hello usabikity people [17:34:37] Hm I just discovered a bug in getOffset() [17:34:48] Hopefully that solves some tihngs [17:52:56] RoanKattouw: hi [17:53:08] I was away from my desk for a minute [17:56:12] How do I disable the *just* the iframe? [17:56:45] er, ignore the extra 'the' in that question [17:57:45] Uncheck all 'labs' prefs [17:58:00] Essentially you're disabling all features that rely on the ifraem [18:01:46] RoanKattouw: you still have allot of out of sync code? [18:02:00] A bit, yes [18:02:02] *RoanKattouw diffs [18:02:25] Calcey is 1/4 here on bugs that I can replicate [18:03:09] I have local modifications to getOffset(), saveHistory() and the highlighter [18:03:38] right on [18:03:50] how long till you commit them? [18:03:54] Not long [18:04:01] I have to get up at 7:15 tomorrow and it's 11pm [18:04:07] *TrevorParscal wants to get his grubby hands on them [18:04:10] owch [18:04:41] there's people working on the roof at my house right now, they showed up at 6:30am and started pounding on the roof with hammers and crap [18:06:43] Ugh [18:08:55] OMG [18:08:58] MY CODE WORKS [18:09:19] *TrevorParscal praises RoanKattouw for his programming swagger [18:10:17] Holy shit: "Approximately 1.9TB of text storage, mostly revisions compressed individually with gzip, was recompressed to about 140GB, a saving of 93%." [18:10:42] yeah [18:10:45] we were just talking about that [18:10:47] go tim! [18:11:42] That's a factor of 14 [18:11:54] I think it uses 50-rev blobs [18:13:24] TrevorParscal: remember how you asked me to improve textarea expansion on the template editor dialogs? [18:13:34] i was poking at it this morning and couldn't see anything wrong with it [18:13:52] it might have been breaking because of something else [18:14:08] it does seem to be working, I was just thinking of asking what you did [18:14:18] haha [18:14:25] magic [18:14:30] black magic [18:15:25] well if you see anything not working as it should, let me know and i'll work on it. I'm gonna mark that one as done on our spreadsheet for now though [18:15:57] Oooh wait [18:16:16] It seems there's a 'req': [ 'iframe' ] line for the s&r dialog in Toolbar.js [18:16:36] Unless you kill that line it'll load the iframe even when you think it's loading the textarea [18:17:30] RoanKattouw: so does that mean i lied on that message on bug 22494? [18:17:51] Probably [18:17:53] *RoanKattouw slaps himself [18:17:57] BUT [18:18:03] We need to recheck some bugs after this [18:18:04] how are they testing it without the iframe then? [18:18:08] TrevorParscal, or nimish_g , would you have 5 minutes to help me get http://meta.wikimedia.org/w/index.php?title=Special:NoticeTemplate/view&template=ethnioFordenwp to show for both logged-in and logged-out users please? [18:18:58] adam_miller: They were until I accidentally uncommented that line [18:19:07] So some of their testing as well as some of ours has been compromised [18:20:01] *RoanKattouw shakes fist at nimish_g [18:20:08] nimish_g: Your pastie *also* contains double newlines [18:20:23] What's the difference between your pastie and the SF article? Just the missing infobox? [18:21:47] pretty much, yeah [18:23:20] So what does r63840 do? Suppress generation of markers in realChange and recycle the old ones instead? [18:23:27] Why does it do that? Performance? [18:23:38] TrevorParscal and nimish_g - if we can make the flow 1/6th of what it is now - we're good to go [18:23:41] You know that'll totally break when mark gets called on something other than load [18:24:42] Oh wait maybe it won't [18:24:47] Because the marker itself is also skipped [18:24:51] Still, why do that? [18:24:56] RoanKattouw: yeah [18:25:22] Yeah so Trevor already suggested a performance improvement to the function you're shortcutting [18:25:23] basically a convoluted system so that 'mark' effectively does nothing with templates while still doing something with TOC [18:25:28] Yeah [18:25:40] I don't like it architecturally, but it's fine for now [18:25:41] so no skipping issues and no breaking if you stop the script and hit 'enter' [18:26:00] yeah over the last 2 days I've written lots of code that disturbs me architechturally [18:26:01] *RoanKattouw types svn up and holds breath [18:26:41] might need to plug your nose - the code might reek of bad practice [18:26:44] As a general rule, could you admit to such ugliness in a code comment or commit summary? So reviewers (me) will know you know it was ugly [18:26:45] RoanKattouw: it also speeds things up a little bit b/c alot of code is skipped [18:26:54] Yes [18:26:54] hahaha ok [18:27:32] TrevorParscal and nimish_g if we do 4% of 40% instead of the 20% of 40%, we're good to go ;) [18:27:34] It's kinda pointless to come up with this critique of why something sucks and get a response like "I know, I thought of all these things myself, but I just needed a quick hack" [18:28:16] hmm [18:29:13] nimish_g: So the template you killed is Infobox settlement? [18:29:29] RoanKattouw: yes [18:29:40] parutron: stop with your math, you know I can't add [18:30:02] want me to come over there an do it?!?!? [18:30:07] i'm your number woman! [18:32:43] nimish_g: i r(Θ) = 1 + 2cos(Θ) [18:32:52] nimish_g: i r(Θ) = 1 + 2cos(Θ) MATH [18:32:53] oh damn [18:33:33] guillom: look over here [18:38:07] hehe [18:38:40] *RoanKattouw makes mental note to let parutron glance at one of his signal processing homework assignments involving complex numbers [18:39:14] oooh RoanKattouw i <3 signal processing and am down to help whenever needed! [18:39:26] OK attention developers ( adam_miller , TrevorParscal , nimish_g ) [18:39:34] I just committed some random code that should fix quite a few issues [18:39:41] It produced a clean diff for me on Nimish's test page [18:39:43] updating [18:39:51] adam_miller: were there critical dialogue bugs i should pay attention to? [18:40:13] It also makes sure that the iframe really is disabled when you think it was, PREVIOUSLY IT WAS NOT, so some of our test results are probably COMPROMISED [18:40:15] nkomura: search and replace is a bit wonky [18:40:26] in what way? [18:40:39] nkomura: other than that we are good i believe. fixed the insert table dialog stuff this morning [18:40:42] In the way that we thought we were testing the textarea while the iframe was enabled [18:40:50] So we'll need to redo some tests [18:40:57] nkomura: https://bugzilla.wikimedia.org/show_bug.cgi?id=22801#c3 [18:41:13] Or we... you [18:41:15] I'm off to bed [18:41:19] you rock [18:41:21] thanks! [18:41:29] good night RoanKattouw [18:41:53] yay! well done RoanKattouw [18:42:02] adam_miller: iframe was enabled even when NTOC was disabled? [18:42:07] Yes [18:42:10] yep [18:42:12] It was always enabled [18:42:17] I'm not sure between when and when exactly [18:42:19] oh [18:42:20] I'll look that up real quick [18:42:47] so it was a config thing that was not right on deployment? [18:43:05] No, it was the toolbar config in the software [18:43:14] The S&R dialog claimed it needed the iframe, so it got it [18:43:15] k [18:43:35] RoanKattouw: all my templates on my supercrazy template test page are working right too, woohoo! [18:44:34] OK the change that compromised test results was made yesterday (Monday) at 12:30pm PDT [18:46:20] And pushed to prototype at an unknown time after that [18:46:34] I'll push the recent revs to prototypes, then go to get my 7 hours of sleep [18:46:55] i'll send out the test instruction to calcey for another round of testing [18:47:11] OK [18:47:22] you mean deployment by prototype right? [18:47:27] Yes all of them [18:47:29] Including that one [18:48:02] may i ask you to do one config change to the prototypes? [18:48:59] never mind [18:49:20] What's that? Or do you no longer need it? [18:49:28] i thought it over [18:49:38] vector is enabled as default on prototypes [18:49:54] i think it'd be good to disable it so that we can verify opt-in logic too [18:50:05] Right [18:50:14] I think it's that way on the deployment prototype, right? [18:50:36] vecor is always on regardless opt-in status [18:50:44] which is different from production environment [18:51:16] oh wait [18:52:01] RoanKattouw: sorry for going back and the forth [18:52:09] i want to test beta opt-in logic tonight [18:52:29] will you not make vector and other beta features default? [18:52:40] OK so they are default on deployment right now? [18:52:43] I thought they weren't [18:52:45] They shouldn't be [18:52:58] I'll fix that right after I sort out this nasty merge conflict between .png files [18:53:03] beta features are default on prototype [18:53:10] prototypes [18:53:19] thanks [18:53:27] adam_miller, TrevorParscal: I'm sorting out a nasty SVN conflict between .png files. If you see old versions of the templateEditor icons on the deployment prototype, tell me [18:53:36] nkomura: Normal prototypes or deployment prototypes? [18:53:52] i'm checking normal prototypes now [18:54:19] deployment looks fine [18:54:36] the beta is not enabled as default on deployment [18:55:09] OK [18:55:14] So do you still want me to make the config change? [18:55:31] I personally feel this stuff should be on by default on the regular prototypes as that's our showcase, and the deployment prototypes are for testing [18:55:32] yes please [18:55:48] good point [18:56:00] it is for testing [18:56:11] so let's keep as they are then [18:56:17] i'm indecisive today [18:57:39] OK really running out now [18:57:49] thank RoanKattouw [18:57:56] sorry for holding you back [18:57:59] Apparently we want the deployment to be done by 11am PDT [18:58:08] i'll send out the test instruction to calcey [18:58:14] Or 10:45am PDT, whatever [18:58:25] (I'll be away between 10:45 and 11:30ish) [18:58:56] i'll ask them to focus on verification of dialogue bugs [18:58:56] Either way, TrevorParscal: please be at the office before 10 tomorrow so I actually have a chance of making that [18:59:01] Yes, good [18:59:23] I'll be doing as much deployment prep work as I can before office hours [19:00:07] I will be in at 9 sharp if at all possible [19:18:07] i need to head home. goodnight ladies and gents. talk to you tomorrow [19:33:08] hello howief