[09:37:56] <zeljkof>	 hi hashar
[09:38:55] <zeljkof>	 could you please check if all repos listed here (including the linked repo) have ruby linters running after every commit?
[09:39:09] <zeljkof>	 (I will also create a bug in bugzilla, so this does not get lost)
[09:58:25] <Nikerabbit>	 zeljkof: can I just use @random_string in my tests?
[09:58:35] <zeljkof>	 Nikerabbit: yes :)
[09:58:45] <zeljkof>	 pairing with Amir, will reply in detail later
[10:33:18] <Nikerabbit>	 zeljkof: https://gerrit.wikimedia.org/r/#/c/91841/ is the patch where I used it
[11:23:42] <zeljkof>	 Nikerabbit: I will take a look right now
[11:30:05] <zeljkof>	 Nikerabbit: I gave it +1 but looks like Kartik gave it +2 before me :)
[11:37:41] <Nikerabbit>	 zeljkof: hehe
[11:38:05] <Nikerabbit>	 zeljkof: had a good session?
[11:38:27] <zeljkof>	 I think it was fun
[11:38:33] <zeljkof>	 I will send weekly e-mail report now
[11:38:52] <Nikerabbit>	 what is code climate?
[11:39:09] <zeljkof>	 Take a look at e-mail that I have just sent to wikitech :)
[11:39:14] <Nikerabbit>	 oh
[11:39:35] <Nikerabbit>	 zeljkof: I really like how the tests are finding regressions
[11:39:45] <zeljkof>	 Nikerabbit: that is what they do :)
[11:40:02] <zeljkof>	 I see you were not a believer up until now :)
[11:40:49] <Nikerabbit>	 zeljkof: I wouldn't say I was non-believer, but I'm glad we finally started writing tests
[11:41:23] <zeljkof>	 I am also glad somebody else is writing the tests :)
[11:41:56] <zeljkof>	 Chris and I can not cover a lot of ground
[11:43:20] <Nikerabbit>	 zeljkof: wee code climate looks nice, although a lot is WIP
[11:43:40] <zeljkof>	 Nikerabbit: it is stable for Ruby I think
[11:43:48] <zeljkof>	 they have just launched js support
[11:44:14] <Nikerabbit>	 yeah
[11:45:12] <Nikerabbit>	 zeljkof: and it's complaining about my name: https://codeclimate.com/github/wikimedia/mediawiki-extensions-UniversalLanguageSelector/lib/jquery.uls/src/jquery.uls.languagefilter.js
[11:45:16] <Nikerabbit>	 ;D
[11:45:42] <zeljkof>	 "This character may get silently deleted by one or more browsers. (Line 5)" :)
[11:47:08] <Nikerabbit>	 now if I only could get mainpage tests green...
[11:50:20] <zeljkof>	 It should be easy to fix
[11:50:28] <zeljkof>	 we just ran out of time this week
[11:50:35] <zeljkof>	 we have to leave something for next week
[11:51:47] <Nikerabbit>	 zeljkof: I thought you were working on Translate, not TwnMainPage
[11:52:09] <zeljkof>	 we were just fixing jobs
[11:52:13] <zeljkof>	 I did not care which repo
[11:52:25] <zeljkof>	 probably twn looked easier to fix
[11:53:04] <Nikerabbit>	 zeljkof: I'm starting a job manually to see if it works now
[11:53:29] <zeljkof>	 Nikerabbit: go ahead
[11:53:40] <zeljkof>	 and good luck
[12:18:47] <Nikerabbit>	 zeljkof: not enough, there is still one failing test which is not failing locally
[12:19:04] <zeljkof>	 Nikerabbit: oh fun :)
[12:19:12] <Nikerabbit>	 https://wmf.ci.cloudbees.com/view/r-tw/job/TwnMainPage-sandbox.translatewiki.net-linux-firefox/lastCompletedBuild/testReport/(root)/Main%20page%20in%20the%20source%20language/Selecting_another_language_for_project_tiles/
[12:25:50] <Nikerabbit>	 ahha
[12:25:57] <Nikerabbit>	 was able to produce it locally
[12:28:00] <Nikerabbit>	 zeljkof: https://gerrit.wikimedia.org/r/91851
[12:28:35] <zeljkof>	 looks good to me
[12:28:39] <zeljkof>	 should I merge it?
[12:29:57] <Nikerabbit>	 zeljkof: thanks
[12:30:06] <zeljkof>	 merged
[12:30:08] <Nikerabbit>	 that should make r-tw green
[12:30:37] <Nikerabbit>	 after that all language team tests should be green
[12:30:43] <zeljkof>	 I have started the job
[12:30:52] <zeljkof>	 we will know in 12 minutes or so
[12:31:11] <zeljkof>	 hashar: are you around?
[12:35:51] <zeljkof>	 hashar: created https://bugzilla.wikimedia.org/show_bug.cgi?id=56156
[12:36:00] <zeljkof>	 but I think I know how to fix it myself, will try
[12:37:19] <hashar>	 hello
[12:37:38] <hashar>	 zeljkof: got to add them manually :/
[12:37:53] <zeljkof>	 hashar: here, right? https://gerrit.wikimedia.org/r/#/admin/projects/integration/jenkins-job-builder-config
[12:37:57] <hashar>	 ideally we would have a linting job able to run any possible linting for us :D
[12:37:59] <zeljkof>	 like this https://gerrit.wikimedia.org/r/#/c/57730/
[12:37:59] <hashar>	 yeah
[12:38:04] <zeljkof>	 ok, on it
[12:38:31] <hashar>	 yeah exactly
[12:38:46] <hashar>	 though we have different ruby version :-]
[12:39:05] <hashar>	 the current jobs are:
[12:39:06] <hashar>	 mediawiki-selenium-ruby1.9.3lint
[12:39:07] <hashar>	 mwext-MobileFrontend-ruby1.9.3lint
[12:39:08] <hashar>	 mwext-UniversalLanguageSelector-ruby1.9.3lint
[12:39:09] <hashar>	 qa-browsertests-ruby1.9.3lint
[12:42:00] <zeljkof>	 hashar: thanks, will add the rest
[12:43:27] <hashar>	 !!
[12:51:52] <grrrit-wm>	 (PS1) Zfilipin: Ruby lint should run after every patch set submission for all repositories that contain Ruby code [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/91853
[12:52:07] <zeljkof>	 hashar: ^
[12:52:36] <zeljkof>	 Nikerabbit: green!!!111oneoneone https://wmf.ci.cloudbees.com/view/r-tw/
[13:14:55] <grrrit-wm>	 (PS1) Cmcmahon: prevent false failure like https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.beta.wmflabs.org-windows-internet_explorer_9/338/testReport/ [qa/browsertests] - https://gerrit.wikimedia.org/r/91861
[13:18:08] <hashar>	 mwext-CirrusSearch-ruby1.9.3lint
[13:18:09] <hashar>	 mwext-Flow-ruby1.9.3lint
[13:18:10] <hashar>	 mwext-Translate-ruby1.9.3lint
[13:18:12] <hashar>	 mwext-TwnMainPage-ruby1.9.3lint
[13:18:13] <hashar>	 mwext-VisualEditor-ruby1.9.3lint
[13:18:15] <hashar>	 mwext-Wikibase-ruby1.9.3lint
[13:18:16] <hashar>	 \O/
[13:19:10] <hashar>	 creating them
[13:19:11] <grrrit-wm>	 (PS2) Hashar: ruby1.9.3 lint jobs for some repositories [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/91853 (owner: Zfilipin)
[13:20:12] <grrrit-wm>	 (CR) Hashar: [C: 2] "Deployed on Jenkins, will add the Zuul triggers." [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/91853 (owner: Zfilipin)
[13:20:32] <grrrit-wm>	 (Merged) jenkins-bot: ruby1.9.3 lint jobs for some repositories [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/91853 (owner: Zfilipin)
[13:25:21] <grrrit-wm>	 (PS1) Hashar: ruby1.9.3 lint jobs for some repositories [integration/zuul-config] - https://gerrit.wikimedia.org/r/91862
[13:25:55] <grrrit-wm>	 (CR) Hashar: "Zuul triggers with same changeid: https://gerrit.wikimedia.org/r/91862" [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/91853 (owner: Zfilipin)
[13:26:15] <grrrit-wm>	 (CR) Hashar: [C: 2] ruby1.9.3 lint jobs for some repositories [integration/zuul-config] - https://gerrit.wikimedia.org/r/91862 (owner: Hashar)
[13:26:20] <grrrit-wm>	 (Merged) jenkins-bot: ruby1.9.3 lint jobs for some repositories [integration/zuul-config] - https://gerrit.wikimedia.org/r/91862 (owner: Hashar)
[13:39:07] <grrrit-wm>	 (PS1) Hashar: trigger rubylint for VisualEditor and Wikibase [integration/zuul-config] - https://gerrit.wikimedia.org/r/91867
[13:39:32] <grrrit-wm>	 (CR) Hashar: [C: 2] trigger rubylint for VisualEditor and Wikibase [integration/zuul-config] - https://gerrit.wikimedia.org/r/91867 (owner: Hashar)
[13:39:37] <grrrit-wm>	 (Merged) jenkins-bot: trigger rubylint for VisualEditor and Wikibase [integration/zuul-config] - https://gerrit.wikimedia.org/r/91867 (owner: Hashar)
[13:58:14] <grrrit-wm>	 (CR) Hashar: [C: 2] gitmodules: Consistently use .git in the urls [integration/jenkins] - https://gerrit.wikimedia.org/r/91647 (owner: Krinkle)
[13:58:18] <grrrit-wm>	 (Merged) jenkins-bot: gitmodules: Consistently use .git in the urls [integration/jenkins] - https://gerrit.wikimedia.org/r/91647 (owner: Krinkle)
[14:02:45] <grrrit-wm>	 (CR) Hashar: "deployed on all slaves. I basically ran:" [integration/jenkins] - https://gerrit.wikimedia.org/r/91647 (owner: Krinkle)
[14:05:03] <grrrit-wm>	 (CR) Hashar: [C: 2] "the ant build file is still valid, no clue how git will update its modules though :/" [integration/jenkins] - https://gerrit.wikimedia.org/r/91648 (owner: Krinkle)
[14:05:06] <grrrit-wm>	 (Merged) jenkins-bot: Remove obsolete testswarm submodule and ant task [integration/jenkins] - https://gerrit.wikimedia.org/r/91648 (owner: Krinkle)
[14:09:09] <grrrit-wm>	 (CR) Hashar: "Updated." [integration/jenkins] - https://gerrit.wikimedia.org/r/91648 (owner: Krinkle)
[14:24:02] <grrrit-wm>	 (CR) Zfilipin: [C: 2] prevent false failure like https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.beta.wmflabs.org-windows-internet_explorer_9/338/testR [qa/browsertests] - https://gerrit.wikimedia.org/r/91861 (owner: Cmcmahon)
[14:37:13] <hashar>	 zeljkof: I am leaving
[14:37:22] <zeljkof>	 hashar: have fun
[14:37:23] <hashar>	 zeljkof: but might deploy additional jobs tomorrow
[14:37:32] <zeljkof>	 thanks for merging my commit
[14:37:39] <zeljkof>	 there is no rush
[15:11:21] <James_F>	 zeljkof: Do you need me to merge 91873?
[15:11:40] <chrismcmahon>	 James_F: I just did
[15:11:43] <zeljkof>	 James_F: chrismcmahon will merge it
[15:11:45] <James_F>	 zeljkof: … apparently not, chrismcmahon just did. :-)
[15:11:47] <James_F>	 Yeah.
[15:12:47] <chrismcmahon>	 James_F: we refactor, and fix a mistake where we're logging in twice sometimes and trying to shut down twice.
[15:13:49] * James_F  nods.

[15:40:55] <grrrit-wm>	 (PS1) Cmcmahon: no need to log in twice [qa/browsertests] - https://gerrit.wikimedia.org/r/91881
[15:41:23] <grrrit-wm>	 (PS2) Zfilipin: no need to log in twice [qa/browsertests] - https://gerrit.wikimedia.org/r/91881 (owner: Cmcmahon)
[15:41:49] <grrrit-wm>	 (CR) Zfilipin: [C: 2] no need to log in twice [qa/browsertests] - https://gerrit.wikimedia.org/r/91881 (owner: Cmcmahon)
[15:42:42] <James_F>	 zeljkof: BTW, the CodeClimate thing is awesome - thanks! Is it a shared log-in I can fix up the VE query (to exclude some libraries) in?
[15:43:21] <zeljkof>	 James_F: there is no log in
[15:43:29] <zeljkof>	 I have just provided link to the repo
[15:43:39] <zeljkof>	 I do not think I can customise it
[15:44:04] <James_F>	 zeljkof: "To exclude files from the analysis go to your Repository Settings area by clicking the gear icon next to your repository name on your dashboard."
[15:44:25] <James_F>	 zeljkof: I don't have a dashboard, but presumably you do?
[15:44:26] <zeljkof>	 James_F: I think that is for paid accounts
[15:44:38] <zeljkof>	 we are using open source account, there is no login
[15:44:46] <James_F>	 Ah.
[15:44:57] <James_F>	 So we can't exclude jQuery/etc.? Oh well.
[15:45:46] <James_F>	 "High total complexity	modules/qunit/qunit.js" doesn't tell us that much. :-)
[16:00:14] <zeljkof>	 :)
[16:00:23] <zeljkof>	 You could contact them directly
[16:00:33] <zeljkof>	 maybe there is a way
[16:09:37] <James_F>	 zeljkof: They say that they don't offer it for OSS projects "right now".
[16:09:48] <James_F>	 zeljkof: Maybe we should sub-module all our libraries so they don't show up. :-)
[16:40:12] <chrismcmahon>	 James_F: are you in the office on Monday, and will you be meeting Rummana?
[16:47:14] <James_F>	 chrismcmahon: Yes and yes.
[16:47:57] <chrismcmahon>	 James_F: great, thanks, I have a training/charter agenda for the rest of the week for her when you're gone.
[16:48:09] <James_F>	 chrismcmahon: Excellent.
[17:08:04] <chrismcmahon>	 James_F:  VE build is fixed, single failure just now is a timeout because saving an edit took too long:  https://wmf.ci.cloudbees.com/job/VisualEditor-en.wikipedia.beta.wmflabs.org-linux-firefox/lastCompletedBuild/testReport/   now to rejigger the cursor tests a bit.
[17:15:29] <James_F>	 chrismcmahon: Brilliant. :-)
[17:16:40] <James_F>	 chrismcmahon: Worth discussing whether we can speed up some of the things we're testing in the browser tests by combining them?
[17:19:41] <chrismcmahon>	 James_F: depends on what you mean by "speed up".  the VE tests support an environment variable REUSE_BROWSER that is useful when running locally, especially e.g. the Headings test which is really painfully slow without it.   Zeljko does not want to use REUSE_BROWSER in the Jenkins job; I'm ambivialent, I don't think it makes much difference.
[17:26:13] <James_F>	 chrismcmahon: Most of VE core actions are tied together; if one item fails, it brings down everything. This means that testing separately (triggering a new VE instance each time) slows the browser test running down massively without actually telling us anything new, I feel.
[17:27:22] <chrismcmahon>	 James_F: not sure I get you.  got an example?
[20:05:09] <grrrit-wm>	 (PS1) Cmcmahon: wait for the textarea to appear, this fixes occasional flaky failures [qa/browsertests] - https://gerrit.wikimedia.org/r/91989
[20:12:01] <grrrit-wm>	 (PS1) Cmcmahon: wait for image, this fails sometimes in IE without it [qa/browsertests] - https://gerrit.wikimedia.org/r/91992
[20:22:35] <grrrit-wm>	 (CR) Zfilipin: [C: 2] wait for image, this fails sometimes in IE without it [qa/browsertests] - https://gerrit.wikimedia.org/r/91992 (owner: Cmcmahon)
[20:22:54] <grrrit-wm>	 (PS1) Cmcmahon: wait for errorbox, this causes flaky IE failures [qa/browsertests] - https://gerrit.wikimedia.org/r/91994
[20:23:01] <grrrit-wm>	 (CR) Zfilipin: [C: 2] wait for the textarea to appear, this fixes occasional flaky failures [qa/browsertests] - https://gerrit.wikimedia.org/r/91989 (owner: Cmcmahon)
[20:26:44] <grrrit-wm>	 (CR) Zfilipin: [C: 2] wait for errorbox, this causes flaky IE failures [qa/browsertests] - https://gerrit.wikimedia.org/r/91994 (owner: Cmcmahon)
[23:41:36] <James_F>	 chrismcmahon: The timeouts waiting for VE to load from different classes are unnecessary, and put undue strain on the test runner to set up and tear down a VE instance each time when it's not needed - the "VisualEditor" test is sufficient to determine if VE laods at all, we don't need the implicit test repeated for each of "Visualeditor Headings", "VisualEditor Links", etc.
[23:41:58] <James_F>	 [He says, hours later, finding this in his send queue.]
[23:49:28] <chrismcmahon>	 James_F: np.  There are a lot of tradeoffs and constraints.  For now I'm OK with inefficiency in the service of coverage, and things like tests written with REUSE_BROWSER in mind make local exploration much less painful.
[23:49:50] <James_F>	 OK.
[23:50:20] <James_F>	 But we could use the server time for more coverage, rather than repeated coverage. :-)
[23:53:21] <James_F>	 chrismcmahon: On the plus side, only the cursor tests failed in the latest labs/FF VE test run - https://wmf.ci.cloudbees.com/view/r-ve/job/VisualEditor-en.wikipedia.beta.wmflabs.org-linux-firefox/lastCompletedBuild/testReport/
[23:55:02] <chrismcmahon>	 James_F:  yeah, and probably because VE changed cursor behavior.  that's what we like to see.
[23:55:35] <James_F>	 chrismcmahon: Yeah, we just landed a major change in the cursor behaviour. Ideally we'd also have changed the browser tests as part of the commit.
[23:55:46] <chrismcmahon>	 James_F: after my research this week I am less confident that cursors are correct in Firefox and more confident that they are correct in Chrome.
[23:55:58] <James_F>	 chrismcmahon: Right. :-(
[23:56:41] <James_F>	 chrismcmahon: In that case, we need to find out why Chrome is loading the gadgets in the iframe. :-(
[23:56:46] <chrismcmahon>	 and then there is IE.  I am actually provisioning some Windows VMs to run tests in ie10 and ie9.  Windows is a drag.
[23:57:01] <James_F>	 chrismcmahon: Don't bother for VE; we haven't even got it loading right, right now.
[23:57:47] <chrismcmahon>	 James_F: VE is not my only test suite, I am trying to get the signal vs. noise better, it's been a struggle in recent times.
[23:57:53] * James_F  nods.