[09:49:08] 3Wikimedia / 3Continuous integration: Add pre-commit hook to check executable rights - 10https://bugzilla.wikimedia.org/68902 (10Andre Klapper) p:5Normal>3Lowest s:5normal>3enhanc [11:59:39] hello [12:15:09] 3Wikimedia / 3Continuous integration: [OPS] Jenkins: Package ruby-jsduck missing in Trusty - 10https://bugzilla.wikimedia.org/68257#c4 (10Antoine "hashar" Musso) 5NEW>3RESO/FIX a:3Alexandros Kosiaris Alexandros handled the packaging work of the missing dependencies \O/ integration-slave1006-trusty:~$... [13:00:03] (03PS1) 10Hashar: mwext-ApiExplorer-testextensions-master voting [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153396 [13:00:31] (03CR) 10Hashar: [C: 032] mwext-ApiExplorer-testextensions-master voting [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153396 (owner: 10Hashar) [13:00:40] (03Merged) 10jenkins-bot: mwext-ApiExplorer-testextensions-master voting [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153396 (owner: 10Hashar) [13:24:30] Krinkle: Can you deploy the unicodejs jjbc change? [13:29:56] James_F: will do :) [13:30:04] hashar: Thanks! :-) [13:30:16] Oh hi hashar :) [13:30:32] James_F: is the i18n team aware of that new unicodejs thing ? [13:30:53] James_F: there might be some overlap with other things they have developed but I am not familiar with that field [13:30:58] hashar: No, but there are no i18n changes. [13:31:15] hashar: Oh, the code's existed for years, this is just it being moved to its own repo from inside VisualEditor. [13:31:20] James_F: you should poke them about it :] [13:31:23] hashar: And yeah, the LE team know. :-) [13:31:46] * hashar digs in his Gerrit dashboard [13:32:12] addshore: the changes on extensions at https://gerrit.wikimedia.org/r/#/q/I0cbfadb4363df19f3a3ff6819280e5ccbedc50b9+is:open,n,z all have failing tests :-] [13:32:21] hashar: But I'll mention to them, no worries. [13:32:28] hashar: https://gerrit.wikimedia.org/r/#/c/153330/ :-) [13:32:36] addshore: some extensions do not support sqlite or are missing entry points. There are some other random failures as well :-D [13:32:46] addshore: I am keeping the changes opened until they are all figured out [13:32:51] James_F: awesome [13:33:03] hashar: Argh, wait. [13:33:13] * James_F didn't git review the fix. [13:33:40] (03PS2) 10Jforrester: Add the new unicodejs library to the testing system [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153330 [13:33:42] There. [13:33:54] (Krinkle wanted me to add jsduck stuff.) [13:34:03] yeah that is worth it :) [13:34:27] you will probably need to amend the Zuul config change [13:34:31] to trigger jsduck stuff [13:35:54] - '{name}-jsduck' [13:35:54] - 'unicodejs-jsduck-publish' [13:35:56] Yup, hashar I have an interesting spreadsheet :) got some other stuff I am working on too. Some phpcs sbiffs that look for usages of removed core features in extensions [13:35:56] Done? [13:36:26] addshore: ah the poor code sniffer. You will need to seek help from some other folks [13:36:37] addshore: it is unlikely I can devote anytime reviewing PHPCS changes :/ [13:37:05] That's okay, they come with tests ;p [13:37:10] (03CR) 10Hashar: "Deployed the jobs unicodejs-jsduck unicodejs-jsduck-publish unicodejs-npm" [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153330 (owner: 10Jforrester) [13:37:16] (03CR) 10Hashar: [C: 032] Add the new unicodejs library to the testing system [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153330 (owner: 10Jforrester) [13:37:36] I'm sure I can get some others to take a look :) [13:38:54] hashar: Thanks! [13:39:29] (03PS3) 10Hashar: Add the new unicodejs library to the testing system [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153330 (owner: 10Jforrester) [13:39:54] (03CR) 10jenkins-bot: [V: 04-1] Add the new unicodejs library to the testing system [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153330 (owner: 10Jforrester) [13:40:03] (03CR) 10Hashar: [C: 032] Add the new unicodejs library to the testing system [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153330 (owner: 10Jforrester) [13:41:28] (03Merged) 10jenkins-bot: Add the new unicodejs library to the testing system [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153330 (owner: 10Jforrester) [13:41:40] Yay. [13:41:40] hashar: if you fancy it take a breif look at https://gerrit.wikimedia.org/r/#/c/153399/ and just see what I'm trying to do :) [13:42:09] Ill set it up as a nightly job probably for all extensions and skins [13:42:44] as core changes we can update the cs :p [13:43:36] (03PS1) 10Hashar: Triggers for unicodejs repository [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153400 [13:43:45] James_F: zuul triggers are https://gerrit.wikimedia.org/r/153400 [13:44:16] (03CR) 10Hashar: [C: 032] Triggers for unicodejs repository [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153400 (owner: 10Hashar) [13:44:17] hashar: Aha, forgot that repo. [13:44:21] hashar: Sorry. [13:44:27] (03Merged) 10jenkins-bot: Triggers for unicodejs repository [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153400 (owner: 10Hashar) [13:45:06] dummy test [13:45:07] https://gerrit.wikimedia.org/r/153401 [13:45:21] and both jobs fail bah [13:45:39] Error: The config file /srv/ssd/jenkins-slave/workspace/unicodejs-jsduck/jsduck.config.json doesn't exist [13:45:40] hehe [13:46:07] James_F: What's the "code" in docs/code for? [13:46:18] hashar: that's fine, the repo is empty right now. [13:46:20] Krinkle: As opposed to test-coverage. [13:46:31] Krinkle: The repo isn't empty, FWIW. [13:46:31] hashar: https://gerrit.wikimedia.org/r/#/c/153078 [13:46:40] James_F: It's empty for the purpose of testing [13:46:47] James_F: you can comment in Gerrit "recheck" to retrigger the job :-] [13:46:49] for there is no gitreview, jshintrc, grunt or anything [13:46:54] James_F: saves you from uploading a new patchset / rebase [13:47:01] hashar: Who needs 'recheck' when you can cherry-pick to master to re-trigger? ;-) [13:47:10] James_F: yeah that works as well :-] [13:47:19] so npm works [13:47:21] jsduck doesn't [13:47:33] but the jsduck one seems quite easy to handle: just add in the config file :] [13:47:40] should I let the job fail in Zuul ? [13:47:45] you can always force merge :] [13:47:46] hashar: James_F: Hold on, I'm gonna fix jsduck in a different way [13:47:51] Krinkle: OK. Thanks! [13:47:59] oh hi Krinkle :-] [13:48:14] hashar: There is no need to force merge, this is a new repo. Merging is in no way priority over passing tests. [13:48:17] hashar: No force merging! :-) [13:48:26] * James_F agrees with Krinkle. [13:48:33] agrees as well [13:48:40] I disagree. [13:48:48] But there may be a bootstrapping issue. [13:49:04] In 153078 there /is/ a jsduck.config.json file… [13:49:42] Oh, damn, no. [13:49:50] I screwed up the zuul config. [13:49:52] * James_F sighs. [13:50:00] It's now jsduck.json not jsduck.config.json per Krinkle. [13:50:11] Krinkle: Want me to fix Zuul or just use the (old) standard file name? [13:50:21] Hold on. [13:50:38] James_F: s/zuul/jenkins job builder/ [13:50:41] Something got incorrectly changed without accounting for existing jobs and forgetting to deploy dependent jobs [13:51:08] unicodejs should use jsduck.json so that we don't introduce more legacy overrides. [13:51:09] hashar: jjbc yeah. [13:51:22] Krinkle: So… the former? [13:53:07] most repos use jsduck.conf.json, we're standardising on jsduck.json, I'm changing jenkins to default to jsduck.json, and add overrides for the other repos. The commit will be effective no-op and not actually deploy anything to Jenkins and only change the templates. The output will remain the same. [13:53:12] Except for the unicode job [13:58:10] (03PS1) 10Krinkle: [WIP] Change default jsduck config to jsduck.json [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153404 [13:58:50] (03PS2) 10Krinkle: [WIP] Change default jsduck config to jsduck.json [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153404 [13:59:20] (03PS3) 10Krinkle: [WIP] Change default jsduck config to jsduck.json [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153404 [14:07:27] Krinkle: I have pooled a Jenkins slave that uses Trusty :-D [14:07:43] Krinkle: akosiaris has ported the ruby-jsduck package there! Though it is the same version as Precise [14:08:04] we also have Doxygen 1.8.6 there :] [14:09:24] addshore: those PHP CodeSniffer patches could use some announcements on wikitech-l to attract some alpha testers :] [14:10:49] (03PS1) 10Zfilipin: Update README.md to mention MEDIAWIKI_API_URL [selenium] - 10https://gerrit.wikimedia.org/r/153405 [14:14:00] (03CR) 10Zfilipin: "Moved from https://github.com/wikimedia/mediawiki-selenium/pull/2" [selenium] - 10https://gerrit.wikimedia.org/r/153405 (owner: 10Zfilipin) [14:14:51] (03PS4) 10Krinkle: [WIP] Change default jsduck config to jsduck.json [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153404 [14:15:57] (03PS2) 10Zfilipin: Update README.md to mention MEDIAWIKI_API_URL [selenium] - 10https://gerrit.wikimedia.org/r/153405 [14:16:46] (03CR) 10Zfilipin: [C: 031] Update README.md to mention MEDIAWIKI_API_URL [selenium] - 10https://gerrit.wikimedia.org/r/153405 (owner: 10Zfilipin) [14:18:48] (03PS5) 10Krinkle: [WIP] Change default jsduck config to jsduck.json [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153404 [14:19:27] (03PS6) 10Krinkle: [WIP] Change default jsduck config to jsduck.json [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153404 [14:27:26] 3Wikimedia / 3Quality Assurance: Chrome not populating text fields correctly - 10https://bugzilla.wikimedia.org/69399 (10Željko Filipin) 3NEW p:3Unprio s:3normal a:3None Reported by S Page at QA mailing list[1]. Reported to Sauce Labs support[2]. 1: http://lists.wikimedia.org/pipermail/qa/2014-Augus... [14:27:54] 3Wikimedia / 3Quality Assurance: Chrome not populating text fields correctly - 10https://bugzilla.wikimedia.org/69399 (10Željko Filipin) 5NEW>3ASSI a:3Željko Filipin [14:27:58] (03PS7) 10Krinkle: Change jsduck macro to default to jsduck.json [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153404 [14:29:59] (03CR) 10Krinkle: [C: 032] "Compiled and deployed unicode-jsduck and unicode-jsduck-publish." [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153404 (owner: 10Krinkle) [14:31:31] (03Merged) 10jenkins-bot: Change jsduck macro to default to jsduck.json [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153404 (owner: 10Krinkle) [14:35:40] 3Wikimedia / 3Quality Assurance: Chrome not populating text fields correctly - 10https://bugzilla.wikimedia.org/69399#c1 (10Željko Filipin) Sauce Labs support answer: Allison Wilbur (Sauce Labs Help Desk) Aug 09 02:03 AM Hi there, I'm sorry this issue suddenly popped up for you! Sauce Labs just recently u... [14:54:17] (03PS1) 10Zfilipin: Chrome 28 is the last version that works properly on Sauce Labs [integration/jenkins-job-builder-config] (cloudbees) - 10https://gerrit.wikimedia.org/r/153408 (https://bugzilla.wikimedia.org/69399) [15:11:08] Krinkle: What's the way to ensure a directory exists for jsduck to be happy? It deletes a .gitignore file… [15:11:24] 3Wikimedia / 3Quality Assurance: Chrome not populating text fields correctly - 10https://bugzilla.wikimedia.org/69399#c3 (10Željko Filipin) I have started builds for all Chrome jobs. The jobs should be more green now. https://integration.wikimedia.org/ci/view/BrowserTests/job/browsertests-Flow-en.wikipedia.... [15:12:25] James_F: The target directory must not exist and not be used by anything else because it is re-created since it can't know about past runs and what to delete and not. [15:12:43] (03CR) 10Hashar: [C: 031] "Straightforward feel free to (self)merge at anytime :)" [integration/jenkins-job-builder-config] (cloudbees) - 10https://gerrit.wikimedia.org/r/152246 (owner: 10Zfilipin) [15:14:38] 3Wikimedia / 3Quality Assurance: rubocop should run after patch set submission to every repo that has Ruby code - 10https://bugzilla.wikimedia.org/69245#c4 (10Antoine "hashar" Musso) rubocop will have to be packaged for Debian so we can easily run it. Zeljkof filled https://rt.wikimedia.org/Ticket/Display.ht... [15:14:48] (03CR) 10Hashar: [C: 04-1] "Change looks fine. It depends on rubocop being made available on the Jenkins slaves which is https://rt.wikimedia.org/Ticket/Display.html?" [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/152918 (https://bugzilla.wikimedia.org/69245) (owner: 10Zfilipin) [15:16:40] Krinkle: But… the directory doesn't exist – that's what jsduck is complaining about. [15:17:01] James_F: The sub or parent directory? [15:17:22] Krinkle: The parent directory exists. The sub-directory didn't exist, and that was the complaint, I think. [15:17:28] * James_F looks. [15:17:36] Oh, hmm. "14:30:23 Error: The parent directory for /srv/ssd/jenkins-slave/workspace/unicodejs-jsduck/docs/code doesn't exist " [15:17:37] Hm... [15:17:46] Exactly [15:17:51] Is it using HEAD rather than target when running? [15:17:56] No [15:18:00] Bootstrapping issue? [15:18:04] The parent directory doesn't exist [15:18:10] But /docs definitely exists. [15:18:12] Did you add it to the commit? [15:18:14] No, it doesn't. [15:18:24] Yes. [15:18:24] It's created by npm. [15:18:42] Is jsduck running after npm? [15:18:43] npm deletes it and re-creates it. [15:18:48] Not safe to run in parallel, maybe? [15:18:56] They're unrelated. [15:18:59] Indeed. [15:19:06] Also, run on different servers entirely [15:19:32] Well, that's a bootstrapping issue then. :-) [15:19:35] * James_F fiddles. [15:19:37] No. [15:19:42] Directory should exist in git. [15:20:07] they can be made dependent, but it still wouldn't help. workspaces (by design) get wiped in between jobs, and may or may not run on the same server. [15:20:24] Where does npm create docs/ ? [15:20:26] I don't see that. [15:23:49] Krinkle: karma:phantomjs [15:24:27] James_F: That creates test-coverage, and implicitly auto-creates the parent directory. The same that jsduck could potentially do [15:24:33] Neither should rely on that [15:24:39] dist/.gitignore [15:24:43] Krinkle: Now fixed. See latest patchset. [15:24:43] docs/.gitignore [15:46:27] Krinkle: Want to merge https://gerrit.wikimedia.org/r/153078 then? :-) [15:57:25] 3Wikimedia / 3Quality Assurance: Use dotenv ruby gem for configuration management - 10https://bugzilla.wikimedia.org/69405 (10Željko Filipin) 3NEW p:3Unprio s:3normal a:3None We have a lot of environment variables. It is becoming hard to document them all. Should we start using a tool like dotenv[1]?... [16:01:53] 3Wikimedia / 3Quality Assurance: create API level tests to monitor services on beta/test2wiki and also on production (tracking) - 10https://bugzilla.wikimedia.org/58353#c9 (10Željko Filipin) Jenkins jobs: https://integration.wikimedia.org/ci/job/UploadWizard-api-commons.wikimedia.beta.wmflabs.org/ https://... [16:08:17] Krinkle: … [16:58:03] (03PS1) 10Addshore: Add job for codecoverage phplint [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153417 [16:59:29] (03PS1) 10Addshore: Add zuul trigger for codesniffer phplint [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153418 [17:03:53] 3Wikimedia / 3Continuous integration: Jenkins: Overhaul the phpcs macro - 10https://bugzilla.wikimedia.org/48420#c25 (10Addshore) Is there anything else to overhaul here? It looks pretty good to me :) (although I might add a few more options in the coming weeks! [17:23:23] 3Wikimedia / 3Quality Assurance: Cucumber step should fail if pending RSpec expectation no longer fails - 10https://bugzilla.wikimedia.org/56243#c9 (10Dan Duvall) We could reimplement `#pending` in a world helper module and rescue/re-raise the `Cucumber::Pending` exception as something else. module Mediawik... [18:54:04] (03CR) 10Dduvall: [C: 032] Update README.md to mention MEDIAWIKI_API_URL [selenium] - 10https://gerrit.wikimedia.org/r/153405 (owner: 10Zfilipin) [18:58:03] what's the best way to set a known-failing test so that CI browsertests show as green? Mark it as @wip, comment it out, add a Pending step somewhere, ?? [19:08:23] 3Wikimedia / 3Quality Assurance: Use dotenv ruby gem for configuration management - 10https://bugzilla.wikimedia.org/69405#c1 (10Dan Duvall) I like the idea of using this in tandem with mw-vagrant since it would allow for extension-specific variables, but I'm not sure we should actually commit the .env file;... [19:13:16] what's the best way to set a known-failing test so that CI browsertests show as green? Mark it as @wip, comment it out, add a Pending step somewhere, ?? [19:39:54] (03CR) 10Hashar: [C: 032] Add job for codecoverage phplint [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153417 (owner: 10Addshore) [19:40:07] cheers hashar :) [19:41:09] addshore: and if you are in the mood for it , you could add a standalone phpunit testsuite to assert our style recommendations :D [19:41:24] (03Merged) 10jenkins-bot: Add job for codecoverage phplint [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/153417 (owner: 10Addshore) [19:41:35] hashar: I dont follow ;p [19:42:07] (03CR) 10Hashar: [C: 032] "Well done :)" [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153418 (owner: 10Addshore) [19:42:13] we should definatly have some sort of voting phpcs though [19:42:19] our current situation sucks :> [19:42:53] deployed [19:43:05] I have setup the basic for phpcs [19:43:11] but never committed to have it used by everyone [19:43:19] some folks have been fixing phpcs issues though [19:43:48] before making it voting, we will probably want to write some guide about how to install phpcs / our style and how to run it [19:43:52] or folks will be annoyed :-] [19:44:19] indeed [19:45:03] what I might aim for it making the lenient one pass everywhere in core then make it voting, even if it requires me to shift around what is in lenient and strict [19:45:28] iirc lenient discards warnings [19:45:50] yup, I almost feel like it might be nicer to write our own sniffs almost from scratch [19:46:43] it would give is very fine grain control, but ofc some things like Generic.CodeAnalysis.ForLoopShouldBeWhileLoop etc should be fine [19:46:50] (03PS2) 10Hashar: Add zuul trigger for codesniffer phplint [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153418 (owner: 10Addshore) [19:46:55] (03CR) 10Hashar: [C: 032] Add zuul trigger for codesniffer phplint [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153418 (owner: 10Addshore) [19:47:10] (03Merged) 10jenkins-bot: Add zuul trigger for codesniffer phplint [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/153418 (owner: 10Addshore) [19:47:37] addshore: yeah I have just pick some "safe" sniffs [19:47:44] and timo later on evilly modified the list [19:47:47] could probably be revisited [19:48:04] :P [19:48:59] (03CR) 10Hashar: [C: 031] "Change is fine. Please update the jobs / self merge when done :)" [integration/jenkins-job-builder-config] (cloudbees) - 10https://gerrit.wikimedia.org/r/153408 (https://bugzilla.wikimedia.org/69399) (owner: 10Zfilipin) [19:49:07] hashar: I finally got access to graphite ;p [19:50:23] I spent a fair ammount of time talking to Timo yesterday evening [19:51:41] hashar: jenkins doesnt currently have a run of phpcs over the whole code anywhere does it? [19:52:56] addshore: I think I removed it [19:53:02] it was run on postmerge [19:53:10] but was really a waste of CPU cycles :] [19:53:18] How does a nightly one sound for a while? [19:55:09] if you figure out the conf sure [19:55:19] trying to figure out what happens with mediawiki/tools/codesniffer not triggering changes [19:55:38] :D [19:55:47] maybe recheck wont work in that case? [19:55:54] ie. there was no trigger before? [19:56:12] I haven't deployed the change :D [19:56:32] all good now [19:57:20] addshore: we already have the jobs :-] [19:57:32] :D [19:57:40] https://integration.wikimedia.org/ci/job/mediawiki-core-phpcs-strict/ and https://integration.wikimedia.org/ci/job/mediawiki-core-phpcs-lenient/ [19:57:49] could be as easy as triggering them on post merge in zuul:-] [19:58:00] ERRR [19:58:02] not on postmerge [19:58:23] :D [19:58:24] cron! [19:58:32] oh, or postmerge.? [19:58:43] yup, just found them in jjbc [20:00:51] heh, and now I have to try and figure out what job groups are in jjbc :P [20:15:08] addshore: that is a bit of a mess [20:15:13] addshore: we could reuse those jobs [20:15:19] and have Zuul trigger them daily [20:15:36] but I never played with Zuul timer :-D [20:16:03] addshore: will sleep in a few. Will review tomorrow :-) [20:16:09] :) [20:16:18] ill have a play ;p [22:18:25] James_F: edia/integration-docroot/commit/0abc7d13dd268ea2a4b4585a483ea4f59c3c0083 [22:18:32] James_F: https://github.com/wikimedia/integration-docroot/commit/0abc7d13dd268ea2a4b4585a483ea4f59c3c0083