[00:04:49] umm [00:04:54] https://integration.wikimedia.org/ci/job/mediawiki-core-npm/2916/console [00:04:58] why is this in chinese? [00:05:17] (or some related language, that i definitely cannot read at all) [01:44:42] Yippee, build fixed! [01:44:42] Project browsertests-Flow-test2.wikipedia.org-linux-chrome-sauce build #278: FIXED in 58 分: https://integration.wikimedia.org/ci/job/browsertests-Flow-test2.wikipedia.org-linux-chrome-sauce/278/ [03:21:41] Yippee, build fixed! [03:21:41] Project browsertests-Echo-test2.wikipedia.org-linux-chrome-sauce build #170: FIXED in 15 分: https://integration.wikimedia.org/ci/job/browsertests-Echo-test2.wikipedia.org-linux-chrome-sauce/170/ [03:43:27] PROBLEM - CI: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: integration.integration-puppetmaster.diskspace._var.byte_avail.value (11.11%) [03:56:17] PROBLEM - BetaLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: deployment-prep.deployment-mediawiki02.diskspace._var.byte_avail.value (33.33%) [04:37:39] RECOVERY - CI: Low disk space on /var on labmon1001 is OK: OK: All targets OK [04:57:33] !log cleaned up coredump on mediawiki02 on deployment-prep [04:57:37] Logged the message, Master [04:58:40] greg-g: done [04:58:44] greg-g: also why aren't you asleep? [04:58:49] oh fuck, I woke up early today, that's why [05:06:11] greg-g: https://phabricator.wikimedia.org/T1259 is the phab task tracking this [05:06:22] greg-g: I've a workaround in place for toollabs, want me to set it up for deployment-prep too? [05:14:30] 3Wikimedia Labs / 3deployment-prep (beta): Puppet failures on deployment-pdf01 - 10https://bugzilla.wikimedia.org/73506 (10Yuvi Panda) 3NEW p:3Unprio s:3normal a:3None Error: Sysctl::Parameters[wikimedia base]: Could not evaluate: can't dup Symbol Notice: instanceproject: deployment-prep Notice: /Sta... [05:49:12] PROBLEM - BetaLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: deployment-prep.deployment-mediawiki01.diskspace._var.byte_avail.value (100.00%) [05:49:58] goddamnit [06:00:44] RECOVERY - BetaLabs: Low disk space on /var on labmon1001 is OK: OK: All targets OK [06:39:28] Project browsertests-CirrusSearch-test2.wikipedia.org-linux-firefox-sauce build #267: FAILURE in 2 分 59 秒: https://integration.wikimedia.org/ci/job/browsertests-CirrusSearch-test2.wikipedia.org-linux-firefox-sauce/267/ [06:48:41] Project browsertests-VisualEditor-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce build #413: FAILURE in 1 小时 10 分: https://integration.wikimedia.org/ci/job/browsertests-VisualEditor-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce/413/ [07:01:04] PROBLEM - CI: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: integration.integration-puppetmaster.diskspace._var.byte_avail.value (11.11%) [07:07:13] RECOVERY - CI: Low disk space on /var on labmon1001 is OK: OK: All targets OK [08:50:59] zeljkof: hi [08:51:07] hashar: good morning :) [08:51:13] zeljkof: I am on IRC today. Would be nice to process the rubocop patches you have proposed against oojs/ui [08:51:17] there is a few of them I -1ed [08:51:44] hashar: saw that, will do, I just have to prepare the papers for my accountant first :( [08:51:53] https://gerrit.wikimedia.org/r/#/q/is:open+label:code-review-1%252Cuser%253Dhashar+project:oojs/ui,n,z [09:00:59] PROBLEM - CI: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: integration.integration-puppetmaster.diskspace._var.byte_avail.value (11.11%) [09:02:47] Yippee, build fixed! [09:02:47] Project browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-linux-chrome-sauce build #292: FIXED in 34 min: https://integration.wikimedia.org/ci/job/browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-linux-chrome-sauce/292/ [09:24:32] !log moved *old* /var/log/eventlogging into /home/yuvipanda so puppet can run without bitching [09:24:36] Logged the message, Master [09:35:05] 3Wikimedia Labs / 3deployment-prep (beta): Puppet fails to run on deployment-eventlogging02 - 10https://bugzilla.wikimedia.org/73479#c2 (10Yuvi Panda) 5PATC>3RESO/FIX Fixed with https://gerrit.wikimedia.org/r/#/c/173758/ :) [09:36:07] 3Wikimedia Labs / 3deployment-prep (beta): Puppet fails on deployment-sca01 - 10https://bugzilla.wikimedia.org/73508 (10Yuvi Panda) 3NEW p:3Unprio s:3normal a:3None yuvipanda@deployment-sca01:~$ sudo puppet agent -tv Warning: Unable to fetch my node definition, but the agent run will continue: Warnin... [10:12:49] Project browsertests-Flow-test2.wikipedia.org-linux-chrome-sauce build #279: FAILURE in 57 min: https://integration.wikimedia.org/ci/job/browsertests-Flow-test2.wikipedia.org-linux-chrome-sauce/279/ [11:01:23] PROBLEM - CI: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: integration.integration-puppetmaster.diskspace._var.byte_avail.value (100.00%) [11:16:17] (03CR) 10Hashar: "Sounds fine. I am writing a test to ensure that all configured mediawiki repositories have the composer validate job." [integration/config] - 10https://gerrit.wikimedia.org/r/173457 (https://bugzilla.wikimedia.org/73444) (owner: 10Legoktm) [12:27:36] * zeljkof is out of lunch [14:05:00] (03PS1) 10Zfilipin: Prepare oojs/ui repository for RuboCop [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) [14:15:24] (03PS2) 10Zfilipin: Prepare oojs/ui repository for RuboCop [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) [14:15:33] (03PS3) 10Zfilipin: Prepare oojs/ui repository for RuboCop [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) [14:26:00] (03PS4) 10Hashar: Run php-composer-validate test for all extensions and skins [integration/config] - 10https://gerrit.wikimedia.org/r/173457 (https://bugzilla.wikimedia.org/73444) (owner: 10Legoktm) [14:26:02] (03PS1) 10Hashar: Rename test zuul_layout to zuul_project_in_gerrit [integration/config] - 10https://gerrit.wikimedia.org/r/173817 [14:26:04] (03PS1) 10Hashar: Test mw repos have php-composer-validate job [integration/config] - 10https://gerrit.wikimedia.org/r/173818 [14:26:39] (03CR) 10Hashar: [C: 032] Rename test zuul_layout to zuul_project_in_gerrit [integration/config] - 10https://gerrit.wikimedia.org/r/173817 (owner: 10Hashar) [14:30:43] (03CR) 10jenkins-bot: [V: 04-1] Test mw repos have php-composer-validate job [integration/config] - 10https://gerrit.wikimedia.org/r/173818 (owner: 10Hashar) [14:33:52] (03PS2) 10Hashar: Test mw repos have php-composer-validate job [integration/config] - 10https://gerrit.wikimedia.org/r/173818 (https://bugzilla.wikimedia.org/73444) [14:35:35] (03CR) 10Hashar: "I wrote an integration test to verify the mediawiki repositories have the composer validation job: https://gerrit.wikimedia.org/r/#/c/1738" [integration/config] - 10https://gerrit.wikimedia.org/r/173457 (https://bugzilla.wikimedia.org/73444) (owner: 10Legoktm) [14:36:51] (03CR) 10Hashar: "A first step toward having tests for the Zuul layout. A bit rough but it is already a base to be improved later on." [integration/config] - 10https://gerrit.wikimedia.org/r/173818 (https://bugzilla.wikimedia.org/73444) (owner: 10Hashar) [14:37:10] (03CR) 10Hashar: [C: 032] Run php-composer-validate test for all extensions and skins [integration/config] - 10https://gerrit.wikimedia.org/r/173457 (https://bugzilla.wikimedia.org/73444) (owner: 10Legoktm) [14:37:20] (03Merged) 10jenkins-bot: Run php-composer-validate test for all extensions and skins [integration/config] - 10https://gerrit.wikimedia.org/r/173457 (https://bugzilla.wikimedia.org/73444) (owner: 10Legoktm) [14:37:27] (03Merged) 10jenkins-bot: Rename test zuul_layout to zuul_project_in_gerrit [integration/config] - 10https://gerrit.wikimedia.org/r/173817 (owner: 10Hashar) [14:44:58] (03PS3) 10Hashar: Test mw repos have php-composer-validate job [integration/config] - 10https://gerrit.wikimedia.org/r/173818 (https://bugzilla.wikimedia.org/73444) [14:45:38] (03CR) 10Hashar: "PS2 fix flake8" [integration/config] - 10https://gerrit.wikimedia.org/r/173818 (https://bugzilla.wikimedia.org/73444) (owner: 10Hashar) [15:40:04] (03PS1) 10Hashar: tests: factor out code to find defined project [integration/config] - 10https://gerrit.wikimedia.org/r/173830 [15:40:06] (03PS1) 10Hashar: (WIP) test Zuul gate configuration (WIP) [integration/config] - 10https://gerrit.wikimedia.org/r/173831 [15:41:22] (03CR) 10jenkins-bot: [V: 04-1] (WIP) test Zuul gate configuration (WIP) [integration/config] - 10https://gerrit.wikimedia.org/r/173831 (owner: 10Hashar) [15:51:36] 3Wikimedia / 3Continuous integration: Add php-composer-validate jobs for all extensions and skins with a composer.json file - 10https://bugzilla.wikimedia.org/73444#c5 (10Antoine "hashar" Musso (WMF)) https://gerrit.wikimedia.org/r/173457 fix the current issues. It has been merged and deployed. I have writt... [15:53:41] (03CR) 10Hashar: [C: 031] "Seems fine :]" [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) (owner: 10Zfilipin) [15:58:39] (03PS2) 10Hashar: mediawiki-vendor-integration -> mediawiki-phpunit [integration/config] - 10https://gerrit.wikimedia.org/r/172970 (https://bugzilla.wikimedia.org/72787) [16:03:11] (03CR) 10Hashar: [C: 032] "Job is up-to-date. Deploying Zuul change." [integration/config] - 10https://gerrit.wikimedia.org/r/172970 (https://bugzilla.wikimedia.org/72787) (owner: 10Hashar) [16:07:39] (03Merged) 10jenkins-bot: mediawiki-vendor-integration -> mediawiki-phpunit [integration/config] - 10https://gerrit.wikimedia.org/r/172970 (https://bugzilla.wikimedia.org/72787) (owner: 10Hashar) [16:10:06] 3Wikimedia / 3Continuous integration: Jenkins: Rename mediawiki-vendor-integration to mediawiki-core-phpunit - 10https://bugzilla.wikimedia.org/72787#c4 (10Antoine "hashar" Musso (WMF)) 5PATC>3RESO/FIX Jenkins job pushed and Zuul configuration applied. [16:11:37] 3Wikimedia / 3Continuous integration: Jenkins: Rename mediawiki-vendor-integration to mediawiki-core-phpunit - 10https://bugzilla.wikimedia.org/72787 (10Antoine "hashar" Musso (WMF)) [16:11:39] 3Wikimedia / 3Continuous integration: Delete workspaces of mediawiki-vendor-integration (job has been deleted) - 10https://bugzilla.wikimedia.org/73515 (10Antoine "hashar" Musso (WMF)) 3NEW p:3Unprio s:3normal a:3None The mediawiki-vendor-integration job has been renamed mediawiki-phpunit (bug 72787)... [16:14:47] (03CR) 10Jforrester: [C: 031] "Let's do this?" [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) (owner: 10Zfilipin) [16:22:13] (03CR) 10Hashar: [C: 032] ":-)" [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) (owner: 10Zfilipin) [16:22:16] (03PS4) 10Hashar: Prepare oojs/ui repository for RuboCop [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) (owner: 10Zfilipin) [16:22:23] (03CR) 10Hashar: [C: 032] Prepare oojs/ui repository for RuboCop [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) (owner: 10Zfilipin) [16:23:17] (03CR) 10jenkins-bot: [V: 04-1] Prepare oojs/ui repository for RuboCop [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) (owner: 10Zfilipin) [16:26:00] (03Merged) 10jenkins-bot: Prepare oojs/ui repository for RuboCop [integration/config] - 10https://gerrit.wikimedia.org/r/173813 (https://bugzilla.wikimedia.org/72841) (owner: 10Zfilipin) [16:36:16] RECOVERY - CI: Low disk space on /var on labmon1001 is OK: OK: All targets OK [16:57:31] chrismcmahon: bd808 do you know what the lucid instance on deployment-prep was for? [16:57:33] can I delete it? [16:58:08] "the lucid instance"? /me looks [16:58:18] bd808: I emailed the qa list [16:58:21] but we have a lucid instance [16:58:26] that's not responding to ssh or anything [16:58:34] deployment-lucid-salt [16:58:56] YuviPanda: deployment-lucid-salt ? [16:59:01] yeah [16:59:08] http://shinken.wmflabs.org/host/deployment-lucid-salt [16:59:16] That was from apergos testing the latest salt version upgrade [16:59:18] ah [16:59:21] let me poke apergos [16:59:35] *nod* [16:59:41] thanks bd808 [17:02:38] 3Wikimedia Labs / 3deployment-prep (beta): Puppet failure on instance udplog - 10https://bugzilla.wikimedia.org/73516 (10Yuvi Panda) 3NEW p:3Unprio s:3normal a:3None Error: Failed to apply catalog: Could not find dependency File[/usr/lib/ganglia/python_modules] for File[/usr/lib/ganglia/python_module... [17:02:57] twentyafterfour: hey! [17:04:36] * bd808 shakes fist at that udplog instance in deployment-prep [17:07:44] bd808: is it still needed? [17:08:00] YuviPanda: https://bugzilla.wikimedia.org/show_bug.cgi?id=72275#c7 [17:08:29] Ori introduced it to make things in beta/prod more same same but it's never worked right [17:08:39] yeah [17:08:44] and puppet's been broken on it for a month now [17:09:01] mostly because the udplog code is intertwined with ganglia code and labs doesn't have ganglia no more [17:09:14] I think somebody should just add hiera to set the target host for logging and then we can send to deployment-bastion which at least used to work [17:09:30] file a bug in phab! :) [17:09:56] If I filed a bug for all the broken crap... [17:10:50] put it on phab! :D [17:10:54] someone might take it up. [17:10:56] (incl. me) [17:39:01] (03PS4) 10Legoktm: Test mw repos have php-composer-validate job [integration/config] - 10https://gerrit.wikimedia.org/r/173818 (https://bugzilla.wikimedia.org/73444) (owner: 10Hashar) [17:48:26] (03CR) 10Legoktm: [C: 032] Test mw repos have php-composer-validate job [integration/config] - 10https://gerrit.wikimedia.org/r/173818 (https://bugzilla.wikimedia.org/73444) (owner: 10Hashar) [17:49:05] (03Merged) 10jenkins-bot: Test mw repos have php-composer-validate job [integration/config] - 10https://gerrit.wikimedia.org/r/173818 (https://bugzilla.wikimedia.org/73444) (owner: 10Hashar) [17:58:03] (03PS1) 10Hashar: Remove cloudbees related doc [integration/config] - 10https://gerrit.wikimedia.org/r/173860 [17:58:06] 3Wikimedia / 3Continuous integration: Add php-composer-validate jobs for all extensions and skins with a composer.json file - 10https://bugzilla.wikimedia.org/73444#c7 (10Kunal Mehta (Legoktm)) 5PATC>3RESO/FIX https://gerrit.wikimedia.org/r/#/c/173429/ has a php-composer-validate job now :) [18:00:51] 3Wikimedia / 3Continuous integration: Add php-composer-validate jobs for all extensions and skins with a composer.json file - 10https://bugzilla.wikimedia.org/73444#c8 (10Antoine "hashar" Musso (WMF)) 5RESO/FIX>3PATC The test still needs to be tested/reviewed though :-d https://gerrit.wikimedia.org/r/17... [18:01:36] 3Wikimedia / 3Continuous integration: Add php-composer-validate jobs for all extensions and skins with a composer.json file - 10https://bugzilla.wikimedia.org/73444#c9 (10Kunal Mehta (Legoktm)) (In reply to Antoine "hashar" Musso (WMF) from comment #8) > The test still needs to be tested/reviewed though :-d... [18:23:23] 3Wikimedia Labs / 3deployment-prep (beta): Puppet failures on deployment-bastion - 10https://bugzilla.wikimedia.org/73520 (10Yuvi Panda) 3NEW p:3Unprio s:3normal a:3None Error: Could not retrieve catalog from remote server: Error 400 on SERVER: Must pass trusted_group to Class[Keyholder] on node i-00... [18:32:23] 3Wikimedia / 3Continuous integration: Make passing mediawiki-core-regression-hhvm-* a prerequisite for merging - 10https://bugzilla.wikimedia.org/73521 (10Tim Landscheidt) 3NEW p:3Unprio s:3normal a:3None With more and more traffic being served by hhvm backends, it feels prudent to make passing the h... [18:33:08] twentyafterfour: I've filed a bunch of bugs about puppet failures on betalas :) [18:35:18] YuviPanda: ok [18:44:37] YuviPanda: https://bugzilla.wikimedia.org/show_bug.cgi?id=73520 is Ori. https://bugzilla.wikimedia.org/show_bug.cgi?id=73516 is Ori. https://bugzilla.wikimedia.org/show_bug.cgi?id=73508 is a messed up puppet master cert. [18:45:12] bd808: I wouldn't say the udplog thing is ori, it's just general puppet fuckery [18:45:22] ganglia + udplog intertwined [18:45:35] if deployment-bastion also has udplog role, I expect it too will have same error once it gets past the current one [20:04:20] 3Wikimedia / 3Continuous integration: Make passing mediawiki-core-regression-hhvm-* a prerequisite for merging - 10https://bugzilla.wikimedia.org/73521#c1 (10Antoine "hashar" Musso (WMF)) I will add another non voting job that runs the phpunit test suite using HHVM. I found out last week that our Trusty ins... [20:17:52] 3Wikimedia / 3Continuous integration: Make passing mediawiki-core-regression-hhvm-* a prerequisite for merging - 10https://bugzilla.wikimedia.org/73521#c2 (10Antoine "hashar" Musso (WMF)) Bah the Trusty slaves are back to Zend php. So either we can: A) set the alternative to have /usr/bin/php point to /usr... [20:29:35] 3Wikimedia / 3Continuous integration: Make passing mediawiki-core-regression-hhvm-* a prerequisite for merging - 10https://bugzilla.wikimedia.org/73521#c3 (10Antoine "hashar" Musso (WMF)) Adding a couple dependencies that are blocking HHVM on CI slaves: bug 71413 [OPS] hhvm 3.3.0-20140925+wmf3 has some anno... [20:29:38] 3Wikimedia / 3Continuous integration: [OPS] hhvm 3.3.0-20140925+wmf3 has some annoying build dependency - 10https://bugzilla.wikimedia.org/71413 (10Antoine "hashar" Musso (WMF)) [20:29:39] 3Wikimedia / 3Continuous integration: CI Trusty slaves no more have hhvm installed - 10https://bugzilla.wikimedia.org/73356 (10Antoine "hashar" Musso (WMF)) [21:18:15] (03CR) 10Zfilipin: [C: 031] "Looks good to me. The point of this was to document how to set up Jenkins (views, plugins, environment variables...) if the existing insta" [integration/config] - 10https://gerrit.wikimedia.org/r/173860 (owner: 10Hashar) [21:19:42] (03CR) 10Hashar: [C: 032] "We have some documentation on wiki https://www.mediawiki.org/wiki/Continuous_integration look for the tutorials." [integration/config] - 10https://gerrit.wikimedia.org/r/173860 (owner: 10Hashar) [21:20:27] (03Merged) 10jenkins-bot: Remove cloudbees related doc [integration/config] - 10https://gerrit.wikimedia.org/r/173860 (owner: 10Hashar) [21:21:34] (03PS1) 10Hashar: Make oojs-ui-bundle-rubocop voting [integration/config] - 10https://gerrit.wikimedia.org/r/173939 (https://bugzilla.wikimedia.org/72841) [21:21:44] (03PS2) 10Hashar: Make oojs-ui-bundle-rubocop voting [integration/config] - 10https://gerrit.wikimedia.org/r/173939 (https://bugzilla.wikimedia.org/72841) [21:23:19] (03CR) 10Hashar: [C: 032] Make oojs-ui-bundle-rubocop voting [integration/config] - 10https://gerrit.wikimedia.org/r/173939 (https://bugzilla.wikimedia.org/72841) (owner: 10Hashar) [21:25:09] (03Merged) 10jenkins-bot: Make oojs-ui-bundle-rubocop voting [integration/config] - 10https://gerrit.wikimedia.org/r/173939 (https://bugzilla.wikimedia.org/72841) (owner: 10Hashar) [21:45:35] 3Wikimedia / 3Continuous integration: Add php-composer-validate jobs for all extensions and skins with a composer.json file - 10https://bugzilla.wikimedia.org/73444#c10 (10Antoine "hashar" Musso (WMF)) 5PATC>3RESO/FIX Bah I am not paying attention :-] Thank you very much for the python review! [21:52:33] (03PS1) 10Cscott: Fix `parsoidsvc-jslint` job. [integration/config] - 10https://gerrit.wikimedia.org/r/173974 [21:57:25] (03PS2) 10Cscott: Fix `parsoidsvc-jslint` job. [integration/config] - 10https://gerrit.wikimedia.org/r/173974 [21:58:13] (03CR) 10Hashar: "So I guess the package.json is a symbolic link to a file in a submodule." [integration/config] - 10https://gerrit.wikimedia.org/r/173974 (owner: 10Cscott) [21:58:21] cscott: replied :D [21:58:33] hashar: yeah, it's a symlink inside the submodule. [21:59:02] i'm trying to redefine the parsoidsvc-jslint job so that i don't have to update zuul as well, but maybe i should just bite the bullet and give the new job a slightly different name. [21:59:28] no, wait, my latest patch seems to have fixed it [22:00:02] hashar: it's so... "fun"... how jjb doesn't actually expand parameters in cases. so that {name}-jslint is different than parsoidsvc-jslint [22:00:08] ...anyway... [22:00:36] well [22:00:40] there is actually two jobs generated [22:00:51] one using the template (with parameter name = parsoidsvc ) [22:01:00] and another job using your job template :D [22:01:12] https://integration.wikimedia.org/ci/job/integration-jjb-config-diff/1644/console looks right to me [22:01:55] i probably don't strictly need to disable json-lint now that i'm enabling submodules, but i believe in a belt & suspenders approach [22:02:59] cscott: well if you enable submodules, jshint will complain [22:03:12] the job should probably be split between jshint and the json linter [22:03:26] I guess you want to ensure that all .json files are valid aren't you ? [22:03:33] maybe i can keep the submodules disabled and just remove the json linter [22:03:52] hashar: we added the job because our localsettings.js files were getting deployed with syntax errors, which was bad [22:04:03] so we really just want to syntax-check conf/**/*.js [22:04:37] cf https://gerrit.wikimedia.org/r/173968 [22:05:26] well you end up forcing jshint to pass on third party modules [22:05:33] that is going to be a mess [22:06:06] hashar: we have a .jshintignore to control that, it should be fine. [22:06:06] maybe you could create a job that just verify the validity of json files under /conf/ ? [22:06:16] again, we don't care about json files [22:06:24] ahaz [22:06:45] I forgot about jshintignore :D [22:07:12] https://integration.wikimedia.org/ci/job/parsoidsvc-jslint/2788/console worked just fine. [22:07:19] with the new parsoidsvc-jslint job [22:07:36] or you could have done : - shell: /srv/deployment/integration/slave-scripts/bin/json-lint.php conf/ [22:07:50] yeah, but there are no json files in conf/ ;) [22:08:15] i'm happy with the way it is now, if you are happy. it seems to work at least. [22:08:16] $jsonFiles = new RegexIterator( $iter, '/\.json$/' ); [22:08:19] stupid script :D [22:08:22] so it works anyway [22:08:38] so if you get the job refreshed --> +2 ! \O/ [22:08:57] yup, already refreshed the job so i could test it on https://gerrit.wikimedia.org/r/173968 [22:09:00] (03CR) 10Hashar: [C: 031] "All good per discussion with C Scott. They have a .jshintignore to control what is being passed to jshint." [integration/config] - 10https://gerrit.wikimedia.org/r/173974 (owner: 10Cscott) [22:09:11] which is now green, yay. so maybe i can finish the parsoid deploy today! [22:09:20] thanks for babysitting some new Jenkins jobs and making them green! [22:09:33] (not to say kudos on avoiding js errors in production hehe) [22:09:48] my travis integration stuff is still waiting on ops.. i'm hoping to get that working soon too [22:10:02] I don't have bandwidth to look at it though :( [22:10:03] (03CR) 10Cscott: [C: 032] "Tested, deployed, works!" [integration/config] - 10https://gerrit.wikimedia.org/r/173974 (owner: 10Cscott) [22:10:11] but Ori / Timo might [22:10:28] hashar: no worries, i've got a private puppet to test it out on [22:10:38] it's just the private keypair stuff that is gumming up the works [22:10:53] i might just borrow your jenkins-mw-sync identity, saves a bunch of trouble [22:10:56] too hard to do things right ;) [22:11:13] ah yeah you asked me about it [22:12:35] anyway bed crash [22:12:44] kudos on adding yet another check! [22:13:28] (03Merged) 10jenkins-bot: Fix `parsoidsvc-jslint` job. [integration/config] - 10https://gerrit.wikimedia.org/r/173974 (owner: 10Cscott) [22:24:35] cscott: Hi [22:28:39] (03CR) 10Krinkle: "As seen in the layoutdiff, it made other changes as well which may not have been on purpose" [integration/config] - 10https://gerrit.wikimedia.org/r/173974 (owner: 10Cscott) [22:29:10] Krinkle: it removed the jsonlint step and checked out the submodules. [22:29:21] (03CR) 10Krinkle: Fix `parsoidsvc-jslint` job. (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/173974 (owner: 10Cscott) [22:29:33] cscott: Why both remove the json lint and add sub modules? [22:29:44] Krinkle: please see backlog! [22:30:13] short story: belt & suspenders. [22:31:49] What does that mean [22:32:22] OK, I can guess, I get it (and comfirmed def) [22:32:23] the job was failing because it was trying to json-lint package.json, which was a symlink to a submodule. [22:32:37] I'm not sure I see how removing the json lint is an being extra safe [22:32:39] two problems with that: we don't want to json lint, and the symlink was broken. so i fixed both. [22:32:47] at first that sounds like opposite to me [22:32:53] Krinkle: there are no relevant json files in our project. [22:33:06] there is only package.json, which is implicitly checked already by the 'npm test' step. [22:33:31] what is json lint going to add to that? [22:34:34] I think that's asking the wrong question, especially if you want to be extra safe (belt & suspenders). The question is, if there are json files with syntax errors or broken symlinks, you'd want to know about it right? [22:34:57] Or at least have to explicitly opt out of jsonlinting for a known known at that point instead of just taking it away thinking it should be fine. [22:35:00] (03CR) 10Cscott: "Yes, the slave configuration now matches the other parsoid jobs, and the task just does jshint. I reviewed the diff before testing the jo" [integration/config] - 10https://gerrit.wikimedia.org/r/173974 (owner: 10Cscott) [22:35:31] Krinkle: i think it's a little ridiculous to run tests on content your project doesn't even have yet. [22:35:50] Unless your npm-test/grunt has a jsonlint task, you're not linting them elsewhere (except package.json itself) [22:35:58] and the jsonlinter has no .*ignore file or any way to keep it from running on random .json files in node_modules. [22:36:12] there are no json files in our project (except package.json itself) [22:36:12] 3Wikimedia / 3Continuous integration: Publish cdb documentation on doc.wikimedia.org - 10https://bugzilla.wikimedia.org/73530 (10Kunal Mehta (Legoktm)) 3NEW p:3Unprio s:3normal a:3None The cdb code in MediaWiki was recently split out into a new library ("cdb" in gerrit). Can we publish the doxygen d... [22:36:15] Sure, but -jslint is for .js and .json at the moment. Many projects don't have json files but still run it implicitly. [22:36:18] if we ever add some, maybe we'll need a linter. [22:36:33] It's not explicitly deviating from the parent macro, that's not the same as removing something you don't wnat to opt in on . [22:36:40] Krinkle: right, that's why were are not using the {name}-jslint job any more. it's not appropriate for parsoid. [22:36:49] it's now * [22:37:22] Krinkle: would you have preferred I rename the task? what name would you prefer? [22:37:28] No [22:37:47] i think it's actually the {name}-jslint task that should be renamed/split, since it's doing two things, neither of which are jslint. [22:38:32] I would've preferred we named -jslint, -jshint (that's been wrong since day 1, we never used jslint). And that json-lint wasn't thrown into -jslint, but instead we discourage use of -jslint alltogether and switch to -npm and have people add their own jshint and json-lint [22:38:35] via Grunt [22:38:59] Anything not -npm or -phpunit is overal imho considered discouraged and unmaintainable. [22:39:17] especially those that use global slave scripts such as jshint and jslint (as opposed to local installs from package.json) [22:39:57] we can't upgrade slave-script/bin/jshint (which the -jslint macro uses) without breaking random shit in existing repo's pipelines. [22:40:11] cscott: I'd recommend parsoid to stop using this job and use -npm instead. [22:40:26] that way you can also trivially run it locally [22:41:46] Krinkle: this isn't parsoid, it's parsoid's deploy repo [22:42:23] 3Wikimedia / 3Continuous integration: Make passing mediawiki-core-regression-hhvm-* a prerequisite for merging - 10https://bugzilla.wikimedia.org/73521 (10Tim Landscheidt) [22:42:59] Krinkle: i can write a custom job that does 'node_modules/.bin/jshint .' in parsoid's deploy repo if you prefer. [22:43:21] if it's the use of slave-script/bin/jshint which is the problem. [22:58:14] npm jobs have node_modules/.bin in their PATH by default [22:58:29] (that's part of npm itself actually) [22:59:06] cscott: Ideally the repo (whatever repo it is, don;t see why a "deploy" repo would be any different) specifies its test in packag.ejson scripts.test [22:59:16] Which repo is this on gerrit? [22:59:27] because the deploy repo borrows the package.json from the source repo. [22:59:56] mediawiki/services/parsoid/deploy.git [23:00:55] feel free to propose patches, i'm a little overworked today just trying to get this parsoid deploy out the door [23:04:51] cscott: Hm.. package.json itself is a symlink, why? [23:05:08] Does that even work? Don't the paths go in a different context and not work as expected? [23:05:18] Can you run npm install and npm test in the root and it works? [23:05:22] so that `npm install` in the deploy repo puts the right packages in deploy/node_modules [23:05:43] also see the ocg service repo, which actually builds its package.json as a union of four different submodule package.jsons [23:06:24] i will grant you that this setup could stand to be rethought, but today i'm not so much concerned with rearchitecting everything, as i am with fixing jenkins so that people can get a deploy done and patches merged. [23:08:28] i have some other problem with this lint job cf https://gerrit.wikimedia.org/r/173319 where it's failing on the source repo now. [23:08:49] but i think i will probably take your advice and just disable it entirely for the source repo, and add jshint to our npm test rule in parsoid. [23:10:26] cscott: hm.. so the package.json symlink in -deploy is only used to populate the hardcoded node_modules directory, right? [23:10:34] In production I assume we execute from deploy/src [23:10:43] and it works because node falls back to parent directory [23:10:44] right? [23:26:57] Krinkle: yup [23:27:29] we don't want to have node_modules in the source repo, but need a static copy to avoid downloading stuff from the internet on deploy [23:28:33] some other places have solved the same issue by setting up their own npmjs.org mirror/proxy [23:30:28] Sure, I get that. [23:30:48] the package.json there is just to make it easier to update [23:30:52] not actually used for anything [23:31:12] gwicke: you wouldn't run npm-test or npm-start inside the deploy/ directory directly, right? [23:31:25] I assume in prod we cd into /src [23:31:59] I hope we do run tests on -deploy so that we verify the hardcoded node_modules dir is good? [23:32:04] gwicke: [23:32:45] Krinkle: I'm not 100% sure if the tests are called in src/ or not [23:33:09] we are generally testing the deploy repo [23:33:21] as well as a new npm install in travis [23:51:30] we run 'npm test' inside the src directory. [23:51:47] we split our tests in two, so we run the same tests in the 'src' and 'deploy' repo. [23:52:08] Krinkle: there's another reason for the {name}-jslint test (as opposed to folding it into 'npm test') -- security. [23:52:23] the {name}-jslint test is the only test which is run on many repos when patches are submitted by untrusted users [23:52:50] because you can execute arbitrary code via npm test with a malicious patch submission [23:53:35] so there's a benefit to running a hardcoded version of jshint which isn't under the control of the repo code