[00:00:56] marktraceur: maybe "Review" could be just renamed to that... You can "Reply" on inline comments but still this needs to be pushed via "Review" [00:03:04] ori-l, spagewmf forgot to mention - we're all done [00:04:21] awjr thanks. Do you want me to mention it if/when we update http://wikitech.wikimedia.org/view/Software_deployments with our shenanigans ? [00:04:52] nah that's ok spagewmf, we have a regular deployment slot but we had to push out an emergency fix :) [00:05:31] i only bugged you guys today because i didnt want to step on your toes if you weren't done [00:06:02] awjr no prob, and http://wikitech.wikimedia.org/view/Server_admin_log sees all [04:56:52] Shirley there's some way to instruct Firebug to breakOnFunctionCall( "FuncName" )? (It's after a redirect to compressed JS, so I can't point to the line of code and set a break.) [06:30:33] spagewmf: You could indicate to firebug that it should break on errors, and throw a dummy error on the first line of the function [06:32:24] marktraceur, thanks for getting back to me. Yes, I could do that or add the magic word "debugger;" at the start of the function. I was hoping for something more like gdb's `break FuncName`. [12:23:05] New patchset: Hashar; "perllint on wikimedia/bugzilla/wikibugs" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/37417 [12:23:22] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/37417 [12:24:22] New patchset: Hashar; "check-voter for wikimedia/bugzilla/wikibugs" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37418 [12:24:37] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37418 [12:39:17] ah [12:41:47] New patchset: Hashar; "Jenkins now autosubmit mediawiki/core changes" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37420 [13:22:46] ping DanielK_WMDE [13:23:37] Platonides: hi [13:23:53] hi Daniel [13:24:11] do you happen to know which blacklist is being used by http://wikidata-test-client.wikimedia.de ? [13:26:12] no. the default, i suppose. [13:26:16] you could ask Silke_WMDE [13:26:36] she's in #wikimedia-wikidata [15:09:06] New patchset: Hashar; "jobs for mw/ext/OpenStackManager" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/37431 [15:09:18] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/37431 [15:15:39] New patchset: Hashar; "enable reporting for OpenStackManager" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37432 [15:15:55] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37432 [15:20:19] New patchset: Hashar; "disable unit test on OpenStackManager extension" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37434 [15:20:30] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37434 [15:23:47] New patchset: Hashar; "Jenkins now autosubmit mediawiki/core changes" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37420 [15:56:52] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37420 [15:58:20] Jenkins now auto submit changes in mediawiki/core.git whenever tests are successful ç [16:39:41] /wc [17:36:20] New patchset: Hashar; "jobs for mw/ext/ParserFunctions" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/37457 [17:36:33] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/37457 [17:38:07] New patchset: Hashar; "enable reporting for ParserFunctions" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37458 [17:39:32] New patchset: Hashar; "enable reporting for ParserFunctions" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37458 [17:39:43] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/37458 [17:58:27] off to home see you later [18:56:52] yay, Special:TemplateSandbox is deployed on all WMF wikis! [19:32:40] valeriej: I just realized https://www.mediawiki.org/wiki/User:Valeriej includes a Regina Spektor quote [19:38:43] hi aharoni got a moment? [19:58:23] You don't need to escape ampersands in regex search strings, right? [20:00:42] on the right hand side only [20:00:55] (when replacing) [20:06:22] saper: that's what I though, thanks! [21:05:31] hello [21:05:43] argl, my nickname is still "in use" [21:05:54] kaldari: I'm not sure this is exactly your beat, not terribly urgent, but if this is in your area, it would be nice to get a +2, it's causing an annoying little browser test failure https://gerrit.wikimedia.org/r/#/c/37210/ [21:06:31] kaldari: it's also been reported from at least one user fwiw [21:06:52] a question about Lua/Scribunto: it seems that I met some cache effects using modules (on test2.wikipedia.org) [21:08:02] I have a test module (that just generate some text). I also have a submodule, that is included by the main menu. It gives a simple function that generates an other text. The main module add this text to its own text and return the result. [21:08:35] chrismcmahon: sure I'll take a look [21:08:40] it works fine, but if I change the text in the submodule I can't see the changes when reloading my test page (test page -> template -> main module) [21:08:53] even after cleaning my browser cache [21:09:42] after "a while" I can see the changes, but it takes several minutes. [21:10:14] chrismcmahon: ack, AFT is using a class called "buttons" [21:10:19] that's very dangerous [21:10:37] hahaha [21:11:07] ah! I tested. What I said is not totaly true: it is not updated. But if I change the main module all is then updated. [21:11:20] merged the change [21:11:32] thanks! [21:12:03] I'm sure there a number of peculiar corners if the AFT code [21:12:04] so it seems that changing a module called by a template is always detected (even for what it includes itself), but changing a module that is only included by an other module is not "seen" [21:12:10] in [21:12:23] btw, is it the good place to talk about that? :) [21:12:50] hexasoft1: Yes, but there are only two people who know much if anything about Scribunto [21:13:21] oh, who are they? :) [21:13:22] vvv isn't here (I imagine he has final exams), and Tim-away is in Australia where it's Saturday morning [21:14:39] ok, thanks. You know if/where they have a WP account? I can maybe put them a message on their talk page (better for asyncronous discussions :)) [21:14:55] RoanKattouw: do we have any jquery lightbox-like thing that's already used by us? [21:16:02] I want to load something in an iframe inside of a lightbox [21:16:34] andrewbogott made an awesome openstackmanager change that links the puppet classes in instance configuration to puppet docs [21:17:00] but it navigates you away to the other site [21:17:10] hexasoft1: https://www.mediawiki.org/wiki/User_talk:Tim_Starling but per the header on that page he prefers e-mail [21:17:16] Ryan_Lane: What's a lightbox? [21:17:31] here's an example: http://lokeshdhakar.com/projects/lightbox2/ [21:17:35] Ryan_Lane: The current code sort of abuses the htmlform widget and uses raw html. I'm not sure if that limits our options. [21:17:41] it's an overlay [21:18:24] andrewbogott: https://gerrit.wikimedia.org/r/#/c/37250/2/special/SpecialNovaInstance.php,unified this? [21:18:57] RoanKattouw: thanks. [21:19:52] Yeah, you can see I make the Html::element and then jam it into a dict index. So any kind of other html tag would work, but any higher-level mediawiki widget would require refactoring quite a bit... [21:20:04] RoanKattouw: obviously it doesn't need to be a lightbox, but I want to load the content on the same page (in an iframe) [21:20:12] And probably changing the htmlform widget. Which I wouldn't mind doing *shrug* [21:20:30] andrewbogott: that's actually ok [21:21:20] you're just modifying the label, which is ok [21:21:24] and the html is being escaped [21:21:30] (do php people say 'dict' or do they call everything arrays?) [21:21:44] it's not dict [21:21:48] associative array [21:21:53] same same, though [21:22:15] I guess it's ordered [21:24:17] andrewbogott: either way, I merged your change [21:24:37] great, thanks. [21:24:57] we'll make it friendlier via js at some point [21:25:37] Ryan_Lane: I don't quite understand what you want. You're saying you want to load *something* in an iframe *somewhere* but I'm failing to pick up on any specifics beyond that [21:25:43] I need to add a header to that site along the lines of "How can I add fixes to these terrible docs" [21:25:46] RoanKattouw: sorry [21:25:56] there's another site which has puppet documentation [21:26:05] when you click on the link, it currently navigates you there [21:26:14] I'd like to load that page in an iframe [21:26:33] in an overlay on the page with the link [21:26:38] +3 for links to our puppet classes documentation :-] [21:26:44] (now we have to write moaar doc) [21:26:55] I see [21:26:58] so, in the lightbox example, it makes an overlay with a picture [21:27:05] I'd like that to be documentation :) [21:27:07] I don't think we have any existing stuff that does that [21:27:10] ok [21:27:14] we can add a plugin [21:27:22] But presumably, if you have an overlay, that's just a

that you can stick anything into [21:27:27] yeah [21:27:37] there's jquery plugins for this [21:27:41] that make it really easy [21:27:56] just wanted to make sure I didn't add one if we already had one [21:31:12] evening [21:31:55] Nikerabbit: howdy [21:34:09] not managed to tame phpunit yet [22:15:53] Platonides: I will review / amend your check-var change on monday probably [22:16:00] Platonides: for now bed + weekend [22:16:34] hashar: btw, the cr+1 and v+1 makes sense now [22:16:53] still not sure if it is better or not, but at least it makes sense now [22:17:05] ? [22:17:16] I haven't changed anything :-D [22:17:23] nonsense [22:17:27] you've changed everything [22:17:37] past 2 weeks [22:17:40] ahhh [22:18:39] Krinkle: anyway you might want to look at http://lists.wikimedia.org/pipermail/wikitech-l/2012-December/064917.html [22:18:53] we will probably change the votes again [22:19:10] :-D [22:19:20] well, my concern isn't naming [22:19:22] it is function [22:19:30] it is important that there is a distiction [22:19:37] and that both are required before "submit" can happen [22:20:34] hashar: are both required now? [22:20:43] both what ? [22:20:57] C+2 and V+1 ? That always have been the case [22:21:06] it is in Gerrit workflow [22:26:17] hashar: V+1 and CR1/CR2 [22:26:29] you talked about this [22:26:37] about requireing 2 "reviews" [22:27:08] I prefer the proposal of V for bots and CR for humans [22:27:34] Platonides: what proposal? [22:27:51] somewhere in the thread [22:27:55] k [22:28:04] feel free to propose different votes [22:28:08] about having jenkins only do V, as its CR +1 is confusing [22:28:18] maybe I should write a RFC to collect different models [22:28:18] hashar: wait so core has switched now? [22:28:23] no unit tests on patch set submissions [22:28:24] then we could vote on the one we want [22:28:32] I would require +2V, and have them given by jenkins independently [22:28:47] Krinkle: core has been switched on wednesday I think [22:28:56] +1: linted, +2 tests run [22:29:05] I know that, I designed that with hashar last year [22:29:06] and jenkins been auto merging for the past few hours [22:29:38] what I mean is I'm surpised we did it already for core, given that there were some second thoughts on it, especially for core, especially with vagrant coming up [22:29:38] Platonides: yup most probably what we are going to setup [22:29:52] Krinkle: we can't wait for vagrant [22:30:01] hashar: what can't wait? [22:30:21] the idea was to setup that new workflow quickly [22:30:29] then we can start working on vagrant :-] [22:30:33] we don't need vagrant to run tests, we've been running them for about a year now [22:30:43] and I also wanted to start using Jenkins Job Builder to automate jobs generation [22:30:49] why don't we keep running them until that is ready [22:30:51] and finally lint all extensions [22:31:23] I think it is a bad idea to wait for human review to run tests, that's going to push back the CR queue. [22:31:33] Erik also raised this I think [22:31:41] yeah I did to [22:32:12] and my feeling is that it is not going to slowdown our CR anyway [22:32:24] most patches submitted to mw/core already pass unit tests anywaay [22:32:29] Afaik there's short term high priority and long-term main priority. main is: vagrant and labs free for all, short term: get CR happening still, and don't regress the workflow. [22:32:45] hashar: what is that supposed to mean? [22:33:08] what I mean is that the unit test result are irrelevant to our ability to CR stuff. [22:33:47] anyway, next week we get some new votes :-] [22:33:57] and I am heading bed :-] [22:34:06] please cast your voice on the wikitech-l thread! [22:34:08] cya all [22:34:14] What I mean is we work better and need less human intervention by having tests run automatically. Having Zuul and all is nice. But why would we enable it with tests in the gate instead of on check-in? [22:34:22] What is holding us back right now from just running tests on submission? [22:34:24] Crap [22:34:49] As if there is a problem, there isn't. It was workinf fine. [22:35:28] and now it isn't.