[00:28:13] Hi, i have a problem with phabricator install on labs project instance [00:29:26] The phabricator can not connect to mysql by the puppet. I using a local mysql [00:29:50] See: https://wmve.wmflabs.org/ [00:30:45] Hello, does anyone have tips on how I could create and destroy mediawiki VMs programmatically? I was thinking of making a tool/bot that automatically made screenshots of skins etc. for mw.org. [00:31:08] and I figured using VMs (or some other kind of isolation) would help prevent security issues [00:31:45] Maybe something with vagrant? [00:32:23] Abin_Sur: Why are you installing phab on labs? [00:35:40] Reedy, for the purposes of coordination and organization of the Wikimedia Venezuela' Engineering Technology Committee [00:35:54] Abin_Sur: You're going about it wrong [00:36:03] Request a project (or similar) on the wikimedia phabricator [00:36:16] Reedy, I have a project [00:36:21] in labs [00:36:22] Ehm.. [00:36:24] https://www.mediawiki.org/wiki/Phabricator/Creating_and_renaming_projects [00:36:35] you don't need your own labs-based instance of phab [00:36:39] You don't need.. [00:36:41] ^ [00:38:33] We have several things to which we require root access. So we request the creation of the project [00:39:02] The instance was created for the purpose of organization. [00:39:24] You're not going to get support for your own instance of phab for these purposes. [00:39:37] Otherwise, I would have done what you suggest. [00:40:25] What do you need root access for? [00:41:04] Well, thanks for your help [01:39:55] Anybody know if our phab servers are on trusty or jessie? [08:33:51] Yippee, build fixed! [08:33:51] Project browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-os_x_10.9-safari-sauce build #844: 09FIXED in 23 min: https://integration.wikimedia.org/ci/job/browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-os_x_10.9-safari-sauce/844/ [11:21:49] PROBLEM - Host deployment-ms-fe02 is DOWN: CRITICAL - Host Unreachable (10.68.21.118) [12:54:52] Project browsertests-GettingStarted-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce build #720: 04FAILURE in 51 sec: https://integration.wikimedia.org/ci/job/browsertests-GettingStarted-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce/720/ [12:59:31] Project browsertests-MobileFrontend-en.m.wikipedia.beta.wmflabs.org-linux-chrome-sauce build #909: 04FAILURE in 27 min: https://integration.wikimedia.org/ci/job/browsertests-MobileFrontend-en.m.wikipedia.beta.wmflabs.org-linux-chrome-sauce/909/ [15:39:49] hashar: hello! i had a question about jenkins for you. i am currently working on implementing some android screenshot tests. we currently run the emulator with the "-nowindow" option. it kind of works but i think we really need an X11 display to get the functionality we want. when i use xvfb, the emulator dies complaining that GLX is not supported. [15:39:49] are there any jobs that would be a good starting point? [15:40:14] oh [15:40:24] yeah we have Xvfb running on all slaves [15:40:33] no clue about GLX though :( [15:40:47] niedzielski: hello :) [15:41:05] hashar: hey! [15:41:35] hashar: hm, so i was wondering if I can install xorg or if this is a bad idea [15:41:42] it is run as : Xvfb :94 -screen 0 1280x1024x24 -ac -noreset [15:41:57] hashar: i was wondering where that 94 display came from [15:42:20] arbitrary assigned :-} [15:42:21] hashar: Xdummy seems to be designed to support GLX but i think i need xorg for it [15:42:33] http://xpra.org/trac/wiki/Xdummy [15:44:49] so that is really Xorg but with a dummy driver, I guess you can try it out :-} [15:45:10] though I dont think Trusty would got it [15:48:57] hashar: so is it ok to install xorg on trusty-1023? [15:49:14] I guess so [15:49:28] make sure to write down the step you did so they can be puppetized [15:50:36] hashar: makes sense. thanks! i'm trying to mock this out locally as much as possible first [15:59:25] niedzielski: good luck :-} [15:59:44] :) [17:57:58] bd808: When did we buy logstash100[1-3]? They're not nearing EOL yet are they? [17:58:21] 1-3 were random spares [17:58:32] I don't know how old they are [17:58:49] * bd808 looks for an rt ticket [18:03:39] legoktm: hi! :) Do you have the time to take a look at https://integration.wikimedia.org/ci/job/mediawiki-phpunit-zend/10252/console ? [18:03:47] error: /srv/deployment/integration/slave-scripts/bin/mw-phpunit.sh: line 42: 23093 Segmentation fault [18:03:59] ostriches: https://rt.wikimedia.org/Ticket/Display.html?id=6325 -- they apparently started life as testsearch100[1-3] [18:04:56] ouch, those are old [18:05:21] ostriches: logstash100[1-3] do not hold elasticsearch data anymore. Technically they could be replaced with gannetti vms or something I think [18:05:58] logstash100[4-6] are the elasticsearch data nodes. They are only ~1 year old [18:06:31] Hmm, non-data nodes to gannetti, interesting [18:07:19] today 1-3 run logstash itself and apache for Kibana. [18:07:59] they also have Elasticsearch installed but without data to just make it really easy for each logstash process to connect to the cluster [18:11:15] ostriches: 4-6 were purchased in April 2015 so not even a year yet [18:11:25] * bd808 can't keep track of these things apparently [18:11:27] Yeah I knew they were new. [18:23:40] 10Continuous-Integration-Infrastructure, 10MediaWiki-General-or-Unknown, 7Jenkins: mediawiki-phpunit-zend segfaults (mediawiki-phpunit-hhvm works fine) - https://phabricator.wikimedia.org/T123644#1934790 (10Florian) 3NEW [18:44:49] 10Continuous-Integration-Infrastructure, 10MediaWiki-General-or-Unknown, 7Jenkins: mediawiki-phpunit-zend segfaults (mediawiki-phpunit-hhvm works fine) - https://phabricator.wikimedia.org/T123644#1934903 (10Florian) Last successful commit from translatewiki.net): https://github.com/wikimedia/mediawiki/commit... [18:45:58] FlorianSW: I know why [18:46:25] legoktm: I suspect because of https://gerrit.wikimedia.org/r/#/c/263813/5/tests/phpunit/phpunit.php ? [18:46:43] However, I don't see any special things there :/ [18:49:36] FlorianSW: https://gerrit.wikimedia.org/r/#/c/264067/ ? [18:49:50] FlorianSW: it passed and didn't segfault... [18:49:59] legoktm: I'm waiting for jenkins now and merge it then :) (I've already saw it in wm-dev) [18:50:23] legoktm: damn, false change, sorry [18:55:02] 10Continuous-Integration-Infrastructure, 10MediaWiki-General-or-Unknown, 7Jenkins: mediawiki-phpunit-zend segfaults (mediawiki-phpunit-hhvm works fine) - https://phabricator.wikimedia.org/T123644#1934947 (10Florian) 5Open>3Resolved a:3Florian Fixed by: https://gerrit.wikimedia.org/r/#/c/264067/ [18:55:12] 10Continuous-Integration-Infrastructure, 10MediaWiki-General-or-Unknown, 7Jenkins: mediawiki-phpunit-zend segfaults (mediawiki-phpunit-hhvm works fine) - https://phabricator.wikimedia.org/T123644#1934954 (10Florian) a:5Florian>3thiemowmde [18:56:47] thanks FlorianSW [18:57:07] legoktm: thanks for your quick response! :) [19:11:30] (03PS1) 10Legoktm: Move Graph extension to generic jobs [integration/config] - 10https://gerrit.wikimedia.org/r/264108 [19:12:56] (03CR) 10Jforrester: Move Graph extension to generic jobs (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/264108 (owner: 10Legoktm) [19:13:54] (03CR) 10Legoktm: Move Graph extension to generic jobs (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/264108 (owner: 10Legoktm) [19:14:31] (03CR) 10Jforrester: [C: 031] Move Graph extension to generic jobs (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/264108 (owner: 10Legoktm) [19:14:56] (03CR) 10Legoktm: [C: 032] Move Graph extension to generic jobs [integration/config] - 10https://gerrit.wikimedia.org/r/264108 (owner: 10Legoktm) [19:18:10] (03Merged) 10jenkins-bot: Move Graph extension to generic jobs [integration/config] - 10https://gerrit.wikimedia.org/r/264108 (owner: 10Legoktm) [19:18:20] !log deploying https://gerrit.wikimedia.org/r/264108 [19:18:25] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [19:29:08] (03PS1) 10Legoktm: Set up CI for WikimediaPageViewInfo extension [integration/config] - 10https://gerrit.wikimedia.org/r/264114 [19:31:25] (03CR) 10Legoktm: [C: 032] Set up CI for WikimediaPageViewInfo extension [integration/config] - 10https://gerrit.wikimedia.org/r/264114 (owner: 10Legoktm) [19:32:40] (03Merged) 10jenkins-bot: Set up CI for WikimediaPageViewInfo extension [integration/config] - 10https://gerrit.wikimedia.org/r/264114 (owner: 10Legoktm) [19:32:56] !log deploying https://gerrit.wikimedia.org/r/264114 [19:33:00] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [19:39:28] lo [19:40:01] * greg-g waves [19:41:20] I am experimenting new time schedules :p [19:41:47] did 10am - 5pm straight with a snack [19:42:07] end result: spend almost 4 hours with kids! [19:44:41] well done! [19:55:34] hashar: hm, so i did some stuff and it seems to work. however, it raised some new questions :) 1) i'm currently using xpra to start an X display. it sounds like i could use the Jenkins Xvnc Plugin instead which would probably be cleaner. is it ok for me to just install this plugin for the whole jenkins instance willy nilly? [19:56:15] niedzielski: for Xvfb what we did is craft an upstart job that ensure it is always running [19:56:24] and for Jessie there is the equivalent systemd managed service [19:56:29] what is xpra ? [19:56:55] hashar: that's part of that Xdummy thing [19:57:03] ...i don't know more than i need to about it :) [19:57:21] hehe [19:57:32] * niedzielski feels like an ex-Xdummy [19:57:37] what is the Xvnc plugin about? [19:57:51] I guess it handles spawning the X server for you then add a vnc server for it isn't it ? [19:58:07] hashar: that's what i was thinking. right now i have to manually do something like this Xorg -logfile x.log -config /etc/xpra/xorg.conf :49374& [19:58:31] 49374 being the display port right ? [19:58:37] hashar: right [19:58:44] hashar: i guess we could call it 95 [19:59:12] hashar: but anyway, i _think_ or hope that Xvnc might let us avoid managing it [19:59:25] hashar: Please review https://gerrit.wikimedia.org/r/#/c/263390/ and https://gerrit.wikimedia.org/r/#/c/263825/ [19:59:27] hashar: not really sure and would ideally just try out the plugin [20:00:36] Also https://gerrit.wikimedia.org/r/#/c/263827/ please. [20:00:43] niedzielski: in puppet you can have a look at the modules/xvfb dir which has the necessary bit to wrap around xvfb [20:00:58] niedzielski: but feel free to install the Xvnc plugin on jenkins and see what happens :-} [20:01:09] Also https://gerrit.wikimedia.org/r/#/c/263936/ please. [20:02:00] hashar: my personal inclination is to try Xvnc first. however, i don't normally work in this area so i'm happy to just "make it like xvfb". would you prefer i do the latter? [20:02:37] Hi guys! there are two new ops guys in SF at the moment so I was wondering if anybody of you would be free to talk about deployments and how to release code in Wikimedia (informal brownbag, even 30 mins) [20:03:36] niedzielski: yeah it is probably better to have the xorg/xpra to be managed by the system directly so it is used by everyone [20:03:58] hashar: ok cool. i can do that [20:04:16] paladox: the bluespiceextensions change, seems it is being reviewed and just about to be landed / CR+2 [20:04:17] (03PS1) 10Legoktm: Mark WikimediaShopLink as archived [integration/config] - 10https://gerrit.wikimedia.org/r/264122 [20:04:33] (03CR) 10Legoktm: [C: 032] Mark WikimediaShopLink as archived [integration/config] - 10https://gerrit.wikimedia.org/r/264122 (owner: 10Legoktm) [20:04:38] paladox: Semantic extensions, honestly I have no clue how they pass :-/ [20:04:45] hashar: Ok thanks. And yes. [20:05:21] hashar: ok so question 2) the android emulator plugin only supports one device per job which breaks the use case we're going for (we want to test a few devices simultaneously). there's a number of ways we could tackle it, each one sounds ugly to me :) currently, i'm leaning towards starting a chain of jobs where each one fires up a different emulat [20:05:21] or. any better ideas? [20:05:23] niedzielski: and maybe we can even stop using Xvfb and migrate to the Xorg / Xdummy whatever thing [20:05:55] (03Merged) 10jenkins-bot: Mark WikimediaShopLink as archived [integration/config] - 10https://gerrit.wikimedia.org/r/264122 (owner: 10Legoktm) [20:06:18] !log deploying https://gerrit.wikimedia.org/r/264122 [20:06:22] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [20:07:09] (03PS1) 10Paladox: [SemanticFormsInputs] Add dependance on SemanticForms extension [integration/config] - 10https://gerrit.wikimedia.org/r/264123 [20:07:12] niedzielski: so for multiple devices, it all depends what you wanna tes t:-} [20:07:34] hashar: I will upload a patch now that we could use to check weather that extension passes the test. [20:07:38] niedzielski: in jenkins world one usually create a multi-configuration project with the device names as parameters [20:07:46] but those jobs are not properly handled by Zuul [20:07:46] (03PS2) 10Paladox: [SemanticFormsInputs] Add dependance on SemanticForms extension [integration/config] - 10https://gerrit.wikimedia.org/r/264123 [20:07:59] niedzielski: so I would craft a Jenkins job per device [20:08:09] hashar: i basically need a way to spawn multiple instances of the android emulator simultaneously. they can't be tested one after another [20:08:17] niedzielski: and Zuul has a feature that let you run a job first, then only if it pass spawn the others [20:08:21] syntax is something like: [20:08:35] hashar: Please review https://gerrit.wikimedia.org/r/#/c/264123/ it should allow us to run check experimental and see if it passes the test since that extension requires SemanticForms be installed. [20:08:38] - job-required-fast: [20:08:45] - job-required-slow [20:08:54] - job-brokendevice-slow [20:09:07] Zuul will trigger job-required-fast first [20:09:09] https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm-composer/690/console [20:09:15] then once it is completed and successful, it will trigger the two others [20:10:55] paladox: we have two different ways to injects dependencies :-( [20:11:21] paladox: on your change https://gerrit.wikimedia.org/r/#/c/264123/2/jjb/mediawiki-extensions.yaml,cm you are changing the Jenkins job mwext-SemanticFormInputs-xxxx [20:11:23] hashar: Oh you mean the ext_dependencies.py file. [20:11:47] paladox: but the experimental pipeline triggers the generic job mwext-testextension-hhvm , so yeah ext_dependencies.py [20:11:47] hashar: Should i do it in the ext_dependencies.py file instead. [20:11:56] paladox: ideally everything should be moved to ext_dependencies.py [20:11:56] hashar: Ok [20:12:48] paladox: and all the Semantic extensions should probably depends on the SemanticMediaWiki extension [20:13:18] (03PS3) 10Paladox: [SemanticFormsInputs] Add dependance on SemanticForms extension [integration/config] - 10https://gerrit.wikimedia.org/r/264123 [20:13:32] hashar: Ok i could probaly migrate the extensions. [20:13:53] Hashar: Ive uploaded the new patch with the change done in the ext file. [20:14:08] grrrit-wm: ping [20:16:12] (03CR) 10Hashar: [C: 032] [SemanticFormsInputs] Add dependance on SemanticForms extension [integration/config] - 10https://gerrit.wikimedia.org/r/264123 (owner: 10Paladox) [20:16:14] paladox: :-} [20:16:49] (03CR) 10Paladox: "Thanks." [integration/config] - 10https://gerrit.wikimedia.org/r/264123 (owner: 10Paladox) [20:17:09] (03CR) 10Hashar: "You are welcome." [integration/config] - 10https://gerrit.wikimedia.org/r/264123 (owner: 10Paladox) [20:17:11] hashar: i think that still runs each job sequentially. i'm verifying right now but i think the tool we're using only really works if all the devices are present at once :( [20:17:15] I need a bot for that [20:17:37] niedzielski: so spawning multiple emulators ? [20:17:57] hashar: right [20:18:37] (03Merged) 10jenkins-bot: [SemanticFormsInputs] Add dependance on SemanticForms extension [integration/config] - 10https://gerrit.wikimedia.org/r/264123 (owner: 10Paladox) [20:18:48] niedzielski: are you using gradle/java to spawn the emulator nowadays ? [20:18:59] or: have you got rid of the android emulator plugin? [20:19:00] hashar: no [20:19:07] hashar: no we still need the plugin [20:19:13] ok ok [20:19:31] hashar: our tests all run through gradle [20:19:50] so the emulator is spawned by the jenkins job but gradle drive the tests right? [20:19:55] hashar: righ [20:19:56] t [20:20:07] and you need the gradle test suite to be able to run against multiple emulators in parallel [20:20:12] or at least in the same gradle test run session [20:20:13] hashar: yes [20:20:30] paladox: deployed [20:20:33] hashar: yeah, all emulators must be present before running gradle for tests [20:20:52] hashar: Thanks. I will go and check experimental now thanks for deploying it. [20:21:58] niedzielski: so the thing is that the jenkins android plugin only spawn a single emulator so you are out of luck on that front [20:22:07] niedzielski: would it be feasible to have a job per device ? [20:22:20] and have the grade test suite to run only tests for the given device ? [20:22:27] for browser tests, we split tests using tags [20:22:41] so you get some tests flagged @test.wikipedia.org others @firefox or @chrome [20:22:48] and we can then run combinaison of them [20:23:10] all the jobs are similar, they just filter tests differently ( eg: @test.wikipedia.org AND @chrome ) [20:23:14] hashar: it would be an unhappy compromise. one of the main benefits of the tool we're using is that it aggregates test results from all attached devices [20:23:57] but you can't spawn (yet?) emulators right ? [20:24:31] does your test tool generate Junit test results ? Jenkins supports aggregating tests results from multiple jobs somehow [20:25:08] thus you have a parent job spawning various combination. Each sub jobs generates its test results. The parent jobs waits for all of them to complete and aggregate the results for you [20:26:58] hashar: this tool collections screenshots and makes pretty animations and things :) [20:27:07] collects* [20:27:27] kind of a video capture of what happens on the device ? [20:28:53] hashar: right [20:29:30] hashar: It seems to fail the test. It says that semanticforms must be installed first but looking at https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm/1050/console it is done after loading SemanticFormsInputs first. Maybe it should be changed in the ext file so that dependacy are first loading. Or and option to allow per extension to decide weather to load dependacy first or after. [20:29:37] hashar: i'm going to keep poking around. maybe there's another way [20:30:08] niedzielski: yeah sorry I dont have any trivial / out of the box solution [20:30:17] hashar: it's kind of a weird use case i think [20:30:37] paladox: the extension should be adjusted to not die() abruptly [20:30:41] hashar: i think the tool we're using basically expects a USB hub to be connected to CI with a bunch of devices hanging off it [20:31:06] paladox: I have made CI to load them in alphabetical order, and I dont plan to handle load order dependency via ext_dependencies.py -:) [20:31:23] hashar: Would you suggest using $wgExtensionFunctions[] = function () { like you did in $wgExtensionFunctions[] = function () { or what should i do. [20:31:26] paladox: that being said, I think the extension registration (extensions.json files) have some way to specificy the load order [20:31:28] https://gerrit.wikimedia.org/r/#/c/263338/1/Spreadsheet.php [20:31:51] paladox: yeah $wgExtensionFunctions would do I guess, and yeah spreadsheet is what I had in mind. But I am not sure that is the proper way to fix the issue [20:32:14] paladox: so the error on Semantic should be fixed the same way we are going to fix the Spreadsheet issue [20:32:25] hashar: Yes. [20:33:04] Maybe we could use the wikimedia jenkins config so that the error isent displayed in jenkins. That would be a temporarily soloution untill we did find one. [20:33:07] Reedy: hola. Since you are an extension guru, would you have a couple minutes to offer your insights on a change to delay an extension die() when a dependency is missing please ? :-} [20:33:07] Reedy: the patch is https://gerrit.wikimedia.org/r/#/c/263338/ [20:33:22] Reedy: Spreadsheet die() whenever PHPExcel() is included after [20:34:08] paladox: I would rather fix the extensions so they no more depends on the load order [20:34:18] hashar: Ok. [20:34:21] paladox: but I am pretty sure that is what extension registration is solving already [20:35:15] It seems that $wgExtensionFunctions[] = function () { dosent stop the error showing in SemanticFormsInput. It might be because extension.json is used in SemanticForms and since it dosent support defines yet it wont detect it. [20:35:15] paladox: and apparently the Jenkins extension loader invokes wfLoadExtensions() for all extensions having the extension.json [20:35:27] paladox: then does require_once for those only having the .php entry point [20:35:41] so maybe if we migrate the semantic extensions to extension.json that will solve it :-} [20:35:49] hashar: Yes. I have a patch that should support defines but wont support using define in the version number. [20:35:50] Yes [20:35:54] I will do that now. [20:36:10] niedzielski: also, if you ever figure out how to spawn the android emulator without Jenkins, that will offers you much more freedom [20:36:31] niedzielski: since this way everything will be handled in your source code repo and you would no more have to care about Jenkins (kind of) [20:36:51] paladox: you might want to poke semantic mediawiki folks. They would surely be interested in migrating every extensions [20:37:07] paladox: not sure where they hangout though. But they are nice people. [20:37:14] hashar: we could do it but i think it would be easy to make a total hackjob [20:37:37] niedzielski: yeah just spurting random ideas :-} [20:37:51] niedzielski: what ever works for you is fine to me :-} [20:38:04] hashar: the android command line clients are kind of lame and we'd probably want to interact with the sdk-manager plugin somehow [20:38:46] hashar: Maybe where ever the doc on this is to advise against using the version number as a way to tell weather the extension is installed. [20:38:51] niedzielski: yeah sounds like new interesting challenges [20:39:17] niedzielski: hey if you have access to some java developers, maybe the Jenkins android plugin could be taught to spawn multiple emulators :D [20:40:28] hashar: maybe... i'm actually not a jenkins believer :) [20:46:26] hashar: i presume it works? :P [20:46:33] It looks vaguely fine (minor style issues) [20:47:16] Reedy: I have no clue whether SpreadSheet would do stuff before wgExtensionFunctions kicks in that would require PHPExcel to be loaded [20:47:27] Reedy: but then, since probably nobody use that extension .... [20:47:36] Isn't it actively developed? [20:47:55] Reedy: doesn't look like [20:48:07] mostly bulk changes and the usual l10n span [20:48:09] spam [20:48:12] lol [20:48:20] Where is the spreadsheet stuff autoloaded? [20:49:10] autoloaded ? [20:49:13] from Jenkins [20:49:54] Where does the class live? [20:50:38] PHPExcel ? it is an include in the autoloader [20:50:43] oh [20:54:24] ? [20:54:37] In theory, it should work [20:54:45] But like I say, I can't see where the code even appears etc [20:55:12] well that surely fix the issue I had for the tests :D [20:55:13] but then [20:55:43] I am wondering whether we should all keep investing time in maintaining extensions that are not active and are most probably not used :-) [20:55:49] but that is a different topic [20:55:52] Yeah [20:55:56] The amount of work FlorianSW and I did [20:56:00] Perfect example of that [20:56:12] https://gerrit.wikimedia.org/r/#/q/status:merged+project:mediawiki/extensions/Spreadsheet+-owner:l10n-bot,n,z [20:56:13] If the maintainer doesn't care, why should we? [20:56:33] No real patches since May 2013 [20:56:34] If we they just commited and abandoned it... [20:57:55] legoktm: Do we get any extension distributor stats? [20:57:58] Same for gerrit... [20:58:04] Mostly anonymous clonses I guess? [20:59:02] Leah: Yes. [20:59:05] Err. Reedy: Yeah. [20:59:17] Curious finger-o. [20:59:36] Updating all the repos takes an age [20:59:49] Reedy: I remember, that ext dist has a statistic about downloads, somewhere... [21:00:05] "Updating all the repos takes an age" <- and isn't really funny! [21:00:33] https://grafana.wikimedia.org/dashboard/db/extension-distributor-downloads [21:01:35] Then it's stupid stuff like Semantic extensions only being in gerrit because of i18n updates [21:01:51] yeah :( [21:02:17] [20:56:13] If the maintainer doesn't care, why should we? [21:02:20] That's the crux of it, for me [21:02:20] and because it's in prod [21:02:27] Krenair: Well, true [21:03:02] Is there a timeline for burning SMW off the face of labswiki? [21:04:12] James_F: Andrew wants a couple of months for Horizon stuff, ideally [21:04:45] But also, we don't seem to know what exactly we'll lose by undeploying it [21:04:57] If we kill it now and Horizon comes later, all that's lost is a small amount of data not being shown to non-project-admins on the Nova: pages, right? [21:04:58] Is avoiding that loss worth the pain? [21:05:00] Oh. [21:05:16] * James_F sighs. [21:05:17] We've gotta move tool labs signup to phab [21:05:21] I think that's the biggest blocker [21:05:24] But it's not much work [21:05:49] "Not much" as in three months? [21:06:19] 3 months work? [21:06:55] Well, different people have different scales for "not much work". :-) Do you have a feeling? [21:06:58] FlorianSW: Do you know if you can do dependacys in extension.json. I thought i saw a patch in gerrit that did that but carn't rember what it is called and was a while back. I think you were the owner of the patch. [21:07:03] For what? [21:07:07] Moving tool labs signup? [21:07:37] paladox: https://gerrit.wikimedia.org/r/#/c/250060/ [21:07:39] Yeah. [21:08:00] Reedy: Thanks. [21:08:03] Couple of hours, topside [21:08:07] We'd lose stuff like the instance tables on project pages [21:08:25] which you otherwise need to be a projectadmin to get a table generated for [21:08:30] Reedy is faster and finding my changes as me reading the recent messages to get not lost in the discussion :D *cool* [21:08:32] Unless, I'm missing something obvious [21:08:51] * and -> in [21:09:28] FlorianSW: Get everyone to speak German and see how quickly I follow then ;) [21:09:47] Reedy: also ich hätte damit kein Problem :D [21:09:56] ja mein liebe [21:10:07] ein gross beer bitte [21:10:10] genau [21:10:22] hashar: ? :D Can you say the first one in english? :/ [21:10:37] FlorianSW: thanks for the ext-dist graph. Completely forgot about it [21:10:45] hashar: np :) [21:10:57] FlorianSW: i have no clue what it means, maybe "yes my darling" ?? [21:11:05] yeah [21:11:08] but it's great that there are some people, who speaks a "little bit german" :D [21:11:11] I should take german lessons really [21:11:13] or love, for the direct translation [21:11:19] ein bisschon [21:11:22] ignore the sp [21:11:51] FlorianSW: in France schools you take either english or german as 2nd language, then follow up with german or spansih [21:11:52] hashar: possible, it's not a complete sentence, and some pharses in german requires to know the context :) [21:12:04] *yay* :) [21:12:27] hashar: I had french for 4 years :P [21:12:37] !!!! [21:13:14] oh that grafana page is so nice [21:13:28] can get the extensions that havent been download over a month [21:13:44] but unfortunately I can't speak any really helpful sentence :( Just something like "Je habite Germany" or "Je apelle Florian" [21:14:01] yeah you gotta practice :( [21:14:19] hashar: but you still can download extensions from another place, e.g. gerrit (git clone e.g.) or github (our mirror) :) [21:14:37] I learned most of my english by playing online games [21:14:40] FlorianSW: Could i add you as a reviewer to https://gerrit.wikimedia.org/r/#/c/264136/ please. It is todo with converting the extension to extension.json. [21:15:00] In my opinion, french is a fairly difficult language :) [21:15:49] hashar: I think I learned most of my english here :P (Ok, I had the basics in school, but that's not the same as writing with "real" people in "real" life about "real" topics) :) [21:16:49] paladox: Sorry, I try to don't touch any extension named "Semantic*" anytime soon :P So, sure, you can add me, but I will not respond, sorry :) [21:17:05] FlorianSW: Oh ok. [21:37:22] Reedy: that grafana dashboard is the stats we collect [22:39:53] 5Testing-Initiative-2015: Put link to onboarding Phab board in New tech employee orientation wiki - https://phabricator.wikimedia.org/T123683#1935642 (10ggellerman) 3NEW [22:41:12] 5Testing-Initiative-2015: Put link to onboarding Phab board in New tech employee orientation wiki - https://phabricator.wikimedia.org/T123683#1935661 (10ggellerman) [22:54:07] 5Testing-Initiative-2015: Define way to clone Onboarding board - https://phabricator.wikimedia.org/T123686#1935719 (10ggellerman) 3NEW [23:32:43] (03PS1) 10Niedzielski: Update Android test target [integration/config] - 10https://gerrit.wikimedia.org/r/264209