[02:20:13] New review: Jeroen De Dauw; "Looks legit :)" [integration/zuul-config] (master) C: 1; - https://gerrit.wikimedia.org/r/53535 [03:20:32] DanielFriesen: dummy resources? [03:20:50] hmm, nobody seems to sleep around these parts. [03:21:24] YuviPanda: it's evening over here [03:21:50] 8:20 I guess [03:21:56] ori-l: Basically you create a resource that does nothing as a central point. Then you can have it depend on multiple things. And multiple things can depend on the dummy. But each of those resources only needs to depend on one. [03:22:47] Apparently the common way to do it is abusing exec as a dummy `exec { 'dummy-...': command => "/bin/true" }` [03:23:16] ugh, puppet is really awful [03:23:44] Personally I'd like to write an actual Dummy type [03:24:07] Another name you could call it might be state [03:24:11] wouldn't that effectively be a run stage? [03:24:36] Sure... ;) but without the plethora of bugs and limitations [03:24:37] lol [03:25:35] i don't know about you, but about 90% of the puppet bugs i've written have to do with ordering depedencies [03:26:37] and most of it stems from puppet's absolutely bizarre decision to treat server configuration steps as non-sequential by default [03:26:56] it's not like it uses this to parallelize things, btw [03:28:14] That's why packaging up deps into groups with common dummies indicating that a number of things you need done first are ready helps. Less dependencies to keep track of across boundaries. [03:29:30] Say have a `State['package-dependencies-installed']` make it depend on all the php, apache, etc... packages you need installed. Then just make MW stuff depend on that instead of piles of individual packages, etc... [03:33:55] well, a package resource can take an array of package names [03:34:19] package { [ 'foo', 'bar', 'buzz' ]: ensure => present } [03:34:44] and you can store the array of package names in an array to make it convenient to use as a dependency [03:35:14] so $mediawiki_deps = [ 'apache2', 'php5', 'mysql', /* ... */ ] [03:35:36] then package { $mediawiki_deps: ensure = present } [03:36:12] and then you can do things like require => Package[$mediawiki_deps] [03:39:57] You can but that kind of thing will quickly start to fall apart as you start growing the stuff you manage. [03:41:55] yeah, i see the benefit in the approach you suggest [03:42:28] ^_^ But it would look prettier with a dedicated type [03:42:53] Which theoretically should be easier to write than Gitclone was [03:43:17] Oh right... ori-l https://github.com/dantman/gareth-vagrant [14:34:39] YuviPanda, hello [18:48:03] YuviPanda, hey [19:40:54] Change abandoned: Hashar; "Nothing new for the last ten days, reopen whenever you get some peers approval." [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54823 [19:44:03] hey Rahul_21 [19:44:08] been intermittently away [19:44:41] are you the right person for a jquery doubt? [19:45:11] I can try [19:45:14] it's just a bit rusty [19:48:32] any particular reason why there is a plugin of collapsibleTabs and not for anything else like collapsibleNav etc? [19:54:01] YuviPanda, ^ [19:57:46] that's not a jQuery doubt :P [19:57:51] and no, I have no idea [20:24:22] YuviPanda, any idea what is bartosz Dziewonski's nickname? [20:24:30] MatmaRex: ^ [20:25:22] yup, that's me. [20:25:32] Rahul_21: ^ [20:25:34] :) [20:29:48] MatmaRex, i made all the necessary changes in skins/vector/vector.js but theres this i18n message that the extension vector creates ,how do i change that? [20:30:19] MatmaRex, This is for the editwarning component. [20:30:29] Rahul_21: there's also the preferences entry [20:31:17] Rahul_21: just look at existing code which does something similar and work from there [20:32:15] MatmaRex, ok but where are the basic i18n.php files? [20:34:11] Rahul_21: in languages/ [20:35:01] MatmaRex, ok so thats where i should add the error message? [20:49:44] Project browsertests-windows-internet_explorer_9 build #267: UNSTABLE in 18 min: https://wmf.ci.cloudbees.com/job/browsertests-windows-internet_explorer_9/267/ [23:10:38] New patchset: MarkAHershberger; "refactor a bit and add support for .tarignore" [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/28956 [23:10:38] New patchset: MarkAHershberger; "Move extension generation so that diffs work better" [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/56883 [23:49:02] Yippie, build fixed! [23:49:03] Project browsertests-windows-internet_explorer_9 build #268: FIXED in 17 min: https://wmf.ci.cloudbees.com/job/browsertests-windows-internet_explorer_9/268/ [23:49:47] why is this bot here? [23:50:22] is it not supposed to be? [23:50:46] or do you mean that the messages are spammy? [23:51:12] i'm mostly wondering [23:51:29] there was a bug about moving bots from #mediawiki to a separate channel [23:52:02] and i though that it was said #-dev is supposed to be more for talking than lurking [23:52:15] (but tbh i'd rather kill it and stay in #mediawiki) [23:54:51] it would be better if wmf-selenium-bot announced the build result and wmf-watcher-bot added the "Yippie" bit. [23:55:26] in keeping with "more for talking", i suppose [23:55:38] also, what is it actually reporting? i keep seeing it mentioning broken and fixed builds [23:55:45] does that mean that the tests randomly fail and pass? [23:55:55] or am i misunderstanding? [23:56:48] MatmaRex: i'm not sure. i was just looking at: https://wmf.ci.cloudbees.com/job/browsertests-windows-internet_explorer_9/267/testReport/(root)/Math/Display_simple_math/ [23:57:01] * ori-l looks up Watir::Exception::UnknownObjectException [23:58:56] It's not Watir! [23:59:19] which leads us to https://saucelabs.com/jobs/302b2cf196374819b39470b2b3e55918# [23:59:26] which leads us to: "NoSuchElement: Unable to find element with id == wpTextbox1 "