[00:33:28] 10MediaWiki-Releasing, 10MW-1.31-release: Bundle Minerva Neue skin with MW 1.31 - https://phabricator.wikimedia.org/T191743#4132601 (10Legoktm) [00:34:09] legoktm: Drop because tests were running under 5 but aren't compatible with 7 yet so get skipped? Or fake tests that shouldn't have been counted in the first place? [00:35:38] James_F: PHP 5's xdebug was buggy and overcounted in some scenarios. PHP 7 is more accurate [00:36:01] Right. [00:36:06] "Good", I suppose. [00:37:33] There's going to one more large drop soon, because it's going to stop counting deprecated classes/functions as 100% covered (a really stupid idea in the first place) [00:38:27] Oh, good. [00:38:47] Of course if we go delete all those functions that'll "fix" that too. [00:38:54] ;) [00:40:08] (03CR) 10Krinkle: [C: 032] Expose wikiadmin user/pass via env and use a more complex password [integration/jenkins] - 10https://gerrit.wikimedia.org/r/415312 (https://phabricator.wikimedia.org/T188520) (owner: 10Reedy) [00:41:03] 10Continuous-Integration-Config, 10Release-Engineering-Team (Kanban), 10MediaWiki-Core-Tests, 10Patch-For-Review, 10User-zeljkofilipin: Use a more complex password for WikiAdmin in selenium tests - https://phabricator.wikimedia.org/T188520#4132613 (10Krinkle) >>! In T188520#4131548, @Reedy wrote: > Obvio... [00:41:13] Krinkle: Support with T190548 appreciated. ;-) [00:41:14] T190548: Update mediawiki-core-qunit-selenium-jessie/mediawiki-extensions-qunit-jessie jobs from PHP5 to PHP7/HHVM - https://phabricator.wikimedia.org/T190548 [00:45:07] James_F: I know the problem, but not how or where to solve it. I don't know where the apache config for the CI nodes lives at the moment. [00:45:18] I guess somewhere in puppet, if it's Nodepool [00:45:21] So that makes three of us? :-( [00:45:22] Antoine would know [00:45:28] * James_F nods. [00:47:15] (03CR) 10Krinkle: [C: 04-1] "According to https://integration.wikimedia.org/ci/job/mediawiki-core-qunit-selenium-jessie/19167/console it uses Nodepool (ci-jessie-wikim" [integration/config] - 10https://gerrit.wikimedia.org/r/426115 (https://phabricator.wikimedia.org/T190548) (owner: 10Jforrester) [00:49:55] (03CR) 10Krinkle: [C: 04-1] "I've submitted https://gerrit.wikimedia.org/r/426834 to help make it easier to confirm/test." [integration/config] - 10https://gerrit.wikimedia.org/r/426115 (https://phabricator.wikimedia.org/T190548) (owner: 10Jforrester) [00:53:19] James_F: I hope this helps ^ but beyond that, I'll leave it to Antoine. [00:53:27] * James_F nods. [02:49:18] (03PS1) 10Legoktm: Add experimental mwext-qunit-php70-jessie job to test with [integration/config] - 10https://gerrit.wikimedia.org/r/426837 [02:49:30] (03CR) 10Legoktm: [C: 032] Add experimental mwext-qunit-php70-jessie job to test with [integration/config] - 10https://gerrit.wikimedia.org/r/426837 (owner: 10Legoktm) [02:50:49] (03Merged) 10jenkins-bot: Add experimental mwext-qunit-php70-jessie job to test with [integration/config] - 10https://gerrit.wikimedia.org/r/426837 (owner: 10Legoktm) [02:51:48] !log deployed https://gerrit.wikimedia.org/r/426837 [02:51:50] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [02:56:42] (03CR) 10Legoktm: [C: 032] "> X-MediaWiki-PHP: 5.6.33-0+deb8u1" [integration/config] - 10https://gerrit.wikimedia.org/r/426837 (owner: 10Legoktm) [03:00:52] I think if we swap out mod_php5 with mod_php7.0 it'll just work [03:00:57] I have no idea how to test that though [03:02:56] legoktm: Push it, run it for core and mwgate, revert if it doesn't work? [04:33:34] James_F: it requires rebuilding the nodepool images, which makes me feel really iffy [06:24:45] legoktm: Run as a different image and switch back? Or is that harder? [06:25:12] I have no idea how to do that [06:26:01] maybe if hashar thinks quibble is ready for qunit we could just switch to that [06:51:08] good morning hashar [06:51:18] legoktm: hello :) [06:51:29] legoktm: I noticed your quibble patches, thanks! [06:51:41] :) [06:52:09] we're a bit stuck with how to move the qunit jobs over to PHP 7 https://phabricator.wikimedia.org/T190548 [06:52:33] regardless of what $PHP_BIN is set to, apache is using PHP 5.6 to serve MediaWiki [06:53:28] from what I can tell it looks like if we swap https://gerrit.wikimedia.org/r/plugins/gitiles/integration/config/+/master/dib/puppet/ciimage.pp#103 with php7 it might just work? [06:53:32] legoktm: ah yeah because the jessie apache uses mod_php5 which is 5.6 [06:53:54] and I am not sure sury get a mod_php7 but maybe it does, so that would need a bit of tuning in nodepool/puppet [06:54:02] it does [06:54:20] my other thought was whether quibble is ready for primtetime for qunit at least? [06:54:40] and yeah that puppt bit is what would enable the mod_php5 [06:55:02] for quibble, instead of using apache I am using mediawiki built-in webserver [06:55:13] well really that is php -S and maintenance/dev/router.php [06:55:23] which will hopefully make things easier I think :) [06:55:32] and when using hhvm that is hhvm -m server (which works good enough) [06:55:34] yup [06:56:21] and i wanna work on skipping some tests so that when two jobs run concurrently and the only difference si php, we can skip qunit on one them [06:56:24] so if https://gerrit.wikimedia.org/r/#/c/426741/ is merged, I think we could use quibble for qunit? [06:56:33] eg there is no point in running qunit tests against hhvm and zend php [06:56:37] uh [06:56:42] ( my lame table of doom is https://docs.google.com/spreadsheets/d/1Fmcj3gwbUj9r0M8WI2W1lRjUC4XaznWZJ4kOeELmcgg/edit?usp=sharing ) [06:57:18] unfortunately, tests without @Database still depend on the database implementation [06:57:35] and we also need Firefox in the mix [06:57:43] is it ok if I edit that? [06:57:50] please do :] [06:58:04] I just brain dumped my thoughts probably late at night [07:00:53] hashar: ok, that reflects my ideal matrix [07:01:16] I added nodejs so npm test would get represented somewhere [07:01:42] I don't think we need the composer variants of tests tbh [07:01:52] Or at least, extensions/core should pick either composer or vendor [07:11:42] legoktm: yeah by composer I just meant what is the source of php libraries [07:11:49] either composer install or vendor.git [07:11:55] mhm [07:12:01] nodejs, maybe we can it to the mix, but I am not sure we need to vary on it [07:12:11] oh yeah, I only put one version [07:12:35] so hopefully I will get to implement that this week [07:13:00] gotta find names, but most probably that will end up being a dict of some sort [07:13:03] have you looked at https://gerrit.wikimedia.org/r/#/c/426741/ yet? [07:13:25] and yeah your test filter, I like it :] [07:13:45] that should just allow us to run jenkins jobs --run qunit / etc. [07:14:01] but most probably in CI I would just pass the kind of environment being used or maybe just detect it [07:14:11] And I'm wondering whether we should try and make the existing qunit nodepool job work with PHP 7 or if we can just swap to quibble now [07:14:14] so we would have a reference (php7, mysql, vendor, chrome64) [07:14:18] hm [07:14:30] and if quibble runs on a different env, an option would make it skip unneeded tests [07:14:55] so a job that aims at sqlite would skip npm test/composer test/qunit [07:15:12] I am not entirely sure what I want to do to be honest [07:15:13] do we have to do it all in one job? [07:15:20] can't we just have a quibble-composer-test job? [07:15:27] quibble-php70-composer-test-etc [07:15:42] yup potentially. [07:16:23] as for quibble, I gotta verify a few more things, but I am definitely adding it to test pipeline this week [07:16:29] ok [07:16:31] announce on wikitech [07:16:42] and then most probably make it voting + add it to gate-and-submit [07:16:56] we were supposed to drop PHP < 7 for 1.31 (branching on Tuesday) but that doesn't seem likely at all [07:16:58] I am not sure about its memory usage nor how many additional instances I need in labs [07:17:12] 1.31 is the LTS? [07:17:16] yes [07:17:24] so yeah definirelt dont want to support php 5 [07:17:29] that could just be a note in the release notes [07:17:41] it already was approved through the RfC process [07:17:45] then we have until the hackathon/release to polish up the php7 support [07:17:52] https://phabricator.wikimedia.org/T172165 [07:17:54] or we delay the branching if we are not reay [07:17:56] it supports PHP 7 fine [07:18:10] yeah, I'm not sure [07:18:30] anyways, I need to sleep "early" tonight, good night ;) [07:18:39] worse thing: CI has to catch up and eventually we find out something is broken under php7 [07:18:47] then we can just fix it before the release happen :] [07:18:58] legoktm: thanks for all your thoughts and the php7 work \o/ [07:25:34] 10MediaWiki-Releasing, 10MW-1.31-release: Bundle CategoryTree extension with MW 1.31 - https://phabricator.wikimedia.org/T191735#4132769 (10Legoktm) [07:27:42] 10MediaWiki-Releasing, 10MW-1.31-release: Bundle CategoryTree extension with MW 1.31 - https://phabricator.wikimedia.org/T191735#4132770 (10Legoktm) >>! In T191735#4122296, @Legoktm wrote: > The main thing I'd like to get rid of is CategoryTree::initialize(). The hooks should mostly all be set in extension.jso... [07:44:50] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (Kanban), 10Lexicographical data, 10Wikidata, and 2 others: MediaWiki core's selenium tests flaky when run as part of mwext-mw-selenium-node-composer-jessie job - https://phabricator.wikimedia.org/T191537#4132783 (10WMDE-leszek) > It is pr... [08:20:32] <_joe_> legoktm: the blocker being dumps? [09:46:08] (03CR) 10Hashar: "We need to backport the tests/selenium/selemium.sh to all branches we support and have webdriver.io :] That is because selenium run-scri" [integration/config] - 10https://gerrit.wikimedia.org/r/424592 (https://phabricator.wikimedia.org/T179190) (owner: 10Zfilipin) [10:42:59] (03CR) 10Alexandros Kosiaris: "> thcipriani and I got it running! https://integration.wikimedia.org/ci/blue/organizations/jenkins/service-pipeline-test-only-debug/detail" [integration/config] - 10https://gerrit.wikimedia.org/r/425936 (https://phabricator.wikimedia.org/T188935) (owner: 10Dduvall) [12:06:28] (03CR) 10Hashar: "Some mock.patch calls will need to be adjusted indeed see https://gerrit.wikimedia.org/r/#/c/426745/1/quibble/backend.py@11" (032 comments) [integration/quibble] - 10https://gerrit.wikimedia.org/r/426745 (owner: 10Legoktm) [12:48:40] <_joe_> !log turning off deployment-mediawiki{04,05} T192071, will be deleted by EOW [12:48:43] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [12:48:43] T192071: Upgrade deployment-prep appserver fleet to Debian Stretch (using HHVM) - https://phabricator.wikimedia.org/T192071 [12:58:56] Project beta-scap-eqiad build #204085: 04FAILURE in 5 min 17 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/204085/ [13:09:09] Project beta-scap-eqiad build #204086: 04STILL FAILING in 5 min 29 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/204086/ [13:09:55] <_joe_> uhm might be my fault? [13:10:10] <_joe_> oh right I have to go add the new jobrunner to the ssh keys [13:12:10] <_joe_> !log manually fixing ssh hostkey for mediawiki-jobrunner03 in scap on deployment-prep [13:12:11] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [13:13:52] <_joe_> !log also shutting down deployment-jobrunner02 T192071 [13:13:55] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [13:13:55] T192071: Upgrade deployment-prep appserver fleet to Debian Stretch (using HHVM) - https://phabricator.wikimedia.org/T192071 [13:19:30] Yippee, build fixed! [13:19:31] Project beta-scap-eqiad build #204087: 09FIXED in 5 min 47 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/204087/ [13:19:40] <_joe_> you're welcome, wmf-insecte [13:19:47] <_joe_> I broke it, I fixed it :) [13:23:37] 10Beta-Cluster-Infrastructure, 10Operations, 10HHVM, 10Patch-For-Review, and 2 others: Upgrade deployment-prep appserver fleet to Debian Stretch (using HHVM) - https://phabricator.wikimedia.org/T192071#4133505 (10Joe) All the main servers have been substituted with stretch VMs; the only one remaining turne... [13:46:40] 10Beta-Cluster-Infrastructure, 10Operations: Beta cluster Obama page often responds with 503 - https://phabricator.wikimedia.org/T188913#4133571 (10Niedzielski) I think this task is still being worked on but in case it helps, here's another report from the Obama page this morning: ``` Request from 73.252.38.2... [13:51:10] 10Beta-Cluster-Infrastructure, 10Operations: Beta cluster Obama page often responds with 503 - https://phabricator.wikimedia.org/T188913#4133579 (10Joe) @Niedzielski interstingly, When requiring the `/summary/precambrian` page, I see a successful request to the API cluster, so the error is not a 503 on the par... [13:53:51] 10Beta-Cluster-Infrastructure, 10Operations, 10HHVM, 10Patch-For-Review, and 2 others: Upgrade deployment-prep appserver fleet to Debian Stretch (using HHVM) - https://phabricator.wikimedia.org/T192071#4133583 (10Joe) 05Open>03Resolved [14:12:21] 10Beta-Cluster-Infrastructure, 10Mobile-Content-Service, 10Operations, 10Page-Previews, and 3 others: [Bug] Beta cluster page summary endpoint sometimes reponds with 5xx - https://phabricator.wikimedia.org/T192287#4133655 (10Niedzielski) [14:15:50] (03CR) 10Hashar: [C: 032] "Good one :]" [integration/quibble] - 10https://gerrit.wikimedia.org/r/426740 (owner: 10Legoktm) [14:16:13] 10Beta-Cluster-Infrastructure, 10Operations: Beta cluster Obama page often responds with 503 - https://phabricator.wikimedia.org/T188913#4133687 (10Niedzielski) Could the summary endpoint issue be a network or caching problem related to this task? I was wondering because it seems like the Node.js service is is... [14:16:17] (03Merged) 10jenkins-bot: Skip test that requires HHVM if it isn't installed [integration/quibble] - 10https://gerrit.wikimedia.org/r/426740 (owner: 10Legoktm) [14:26:16] 10Beta-Cluster-Infrastructure, 10Operations: Beta cluster Obama page often responds with 503 - https://phabricator.wikimedia.org/T188913#4133745 (10Mholloway) @Niedzielski I'll look into this on the mobileapps side today. It's possible there's a problem in the config for the beta cluster. Just to be clear, i... [14:33:25] 10Beta-Cluster-Infrastructure, 10Operations: Beta cluster Obama page often responds with 503 - https://phabricator.wikimedia.org/T188913#4133786 (10Niedzielski) Thanks @mholloway. This task is specific to the [[ https://en.wikipedia.beta.wmflabs.org/wiki/Barack_Obama | Barack Obama ]] page. The page summary is... [14:37:04] no_justification https://gerrit-review.googlesource.com/c/gerrit/+/172414 :) [14:44:45] 10Beta-Cluster-Infrastructure, 10Mobile-Content-Service, 10Operations, 10Page-Previews, and 3 others: [Bug] Beta cluster page summary endpoint sometimes reponds with 5xx - https://phabricator.wikimedia.org/T192287#4133813 (10ovasileva) p:05Triage>03High [14:47:44] apparently they are adding opensearch support in gerrit. [14:47:51] https://gerrit-review.googlesource.com/c/gerrit/+/172534 [14:53:36] (03PS3) 10Hashar: Minor code and style cleanup [integration/quibble] - 10https://gerrit.wikimedia.org/r/426743 (owner: 10Legoktm) [14:53:38] (03PS1) 10Hashar: Enhance MySQL exceptions [integration/quibble] - 10https://gerrit.wikimedia.org/r/426938 [14:54:49] (03CR) 10Hashar: [C: 032] Minor code and style cleanup (032 comments) [integration/quibble] - 10https://gerrit.wikimedia.org/r/426743 (owner: 10Legoktm) [14:55:15] (03Merged) 10jenkins-bot: Minor code and style cleanup [integration/quibble] - 10https://gerrit.wikimedia.org/r/426743 (owner: 10Legoktm) [14:55:26] (03CR) 10Hashar: [C: 032] "(I have cherry picked your patch on tip of master)" [integration/quibble] - 10https://gerrit.wikimedia.org/r/426743 (owner: 10Legoktm) [14:56:01] (03PS2) 10Hashar: Enhance MySQL exceptions [integration/quibble] - 10https://gerrit.wikimedia.org/r/426938 [15:08:44] (03PS2) 10Hashar: Allow specifying what tests to run [integration/quibble] - 10https://gerrit.wikimedia.org/r/426741 (owner: 10Legoktm) [15:10:17] (03PS3) 10Hashar: Allow specifying what tests to run [integration/quibble] - 10https://gerrit.wikimedia.org/r/426741 (owner: 10Legoktm) [15:11:50] (03CR) 10Hashar: [C: 032] "I have lot of nitpicks and gotchas, but nothing that should prevent this to be merged. That is a nice improvement for human consumption / " (031 comment) [integration/quibble] - 10https://gerrit.wikimedia.org/r/426741 (owner: 10Legoktm) [15:12:17] (03Merged) 10jenkins-bot: Allow specifying what tests to run [integration/quibble] - 10https://gerrit.wikimedia.org/r/426741 (owner: 10Legoktm) [15:12:57] (03CR) 10Hashar: [C: 04-1] "I would hold that one. I would like the tests to fail as fast as possible and running the @Database test further delay the result of other" [integration/quibble] - 10https://gerrit.wikimedia.org/r/426742 (owner: 10Legoktm) [15:15:58] (03CR) 10Hashar: [C: 031] "I haven't tested." [integration/config] - 10https://gerrit.wikimedia.org/r/426578 (owner: 10Legoktm) [15:16:01] (03CR) 10Hashar: [C: 031] "I haven't tested." [integration/config] - 10https://gerrit.wikimedia.org/r/426548 (owner: 10Legoktm) [15:37:10] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (Kanban), 10Lexicographical data, 10Wikidata, and 2 others: MediaWiki core's selenium tests flaky when run as part of mwext-mw-selenium-node-composer-jessie job - https://phabricator.wikimedia.org/T191537#4133943 (10zeljkofilipin) [[ https... [15:44:44] 10Continuous-Integration-Infrastructure, 10Wikidata, 10Wikidata Query UI: Useless, mostly blank stylelint output in Jenkins CI - https://phabricator.wikimedia.org/T192300#4133994 (10Lucas_Werkmeister_WMDE) [16:00:48] 10Beta-Cluster-Infrastructure, 10MediaWiki-extensions-CentralAuth, 10Regression: [beta cluster] Persistently kicked out from a user session - https://phabricator.wikimedia.org/T142543#4134057 (10MarcoAurelio) [16:58:51] 10Diffusion, 10Striker: Diffusion repository creation fails via toolsadmin - https://phabricator.wikimedia.org/T182142#4134246 (10bd808) [18:25:47] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MW-1.31-release, 10Patch-For-Review: Expand the set of bundled extensions to achieve a default MediaWiki experience that's comparable to Wikimedia sites - https://phabricator.wikimedia.org/T178349#4134415 (10cicalese) [19:09:21] 10Beta-Cluster-Infrastructure, 10Mobile-Content-Service, 10Operations, 10Page-Previews, and 3 others: [Bug] Beta cluster page summary endpoint sometimes reponds with 5xx - https://phabricator.wikimedia.org/T192287#4134473 (10Mholloway) a:03Mholloway [19:59:21] 10Beta-Cluster-Infrastructure, 10Mobile-Content-Service, 10Operations, 10Page-Previews, and 3 others: [Bug] Beta cluster page summary endpoint sometimes reponds with 5xx - https://phabricator.wikimedia.org/T192287#4134550 (10Niedzielski) [20:28:35] 10Beta-Cluster-Infrastructure, 10Mobile-Content-Service, 10Operations, 10Page-Previews, and 3 others: [Bug] Beta cluster page summary endpoint sometimes reponds with 5xx - https://phabricator.wikimedia.org/T192287#4134626 (10Mholloway) @Niedzielski I restarted the beta cluster restbase and mobileapps servi... [20:31:15] 10Beta-Cluster-Infrastructure, 10Mobile-Content-Service, 10Operations, 10Page-Previews, and 3 others: [Bug] Beta cluster page summary endpoint sometimes reponds with 5xx - https://phabricator.wikimedia.org/T192287#4134640 (10Mholloway) [20:35:31] !log restarted restbase and mobileapps services for testing (T192287) [20:35:35] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [20:35:37] T192287: [Bug] Beta cluster page summary endpoint sometimes reponds with 5xx - https://phabricator.wikimedia.org/T192287 [20:52:51] 10Beta-Cluster-Infrastructure, 10Mobile-Content-Service, 10Operations, 10Page-Previews, and 3 others: [Bug] Beta cluster page summary endpoint sometimes reponds with 5xx - https://phabricator.wikimedia.org/T192287#4134666 (10Mholloway) For posterity: there was a recent change to the config variable that se... [21:31:38] (03PS2) 10Krinkle: WebPageTest: Remove the Windows test agent [integration/config] - 10https://gerrit.wikimedia.org/r/423653 (https://phabricator.wikimedia.org/T165626) (owner: 10Phedenskog) [22:13:39] 10Continuous-Integration-Config, 10AbuseFilter, 10Upstream: stylelint is just outputting dots and number of errors, making it impossible to fix - https://phabricator.wikimedia.org/T190072#4134796 (10Legoktm) 05Resolved>03Open Based on above reports, upgrading doesn't seem to help. [22:13:43] 10Continuous-Integration-Config, 10AbuseFilter, 10Upstream: stylelint is just outputting dots and number of errors, making it impossible to fix - https://phabricator.wikimedia.org/T190072#4134800 (10Legoktm) [22:13:45] 10Continuous-Integration-Infrastructure, 10Wikidata, 10Wikidata Query UI: Useless, mostly blank stylelint output in Jenkins CI - https://phabricator.wikimedia.org/T192300#4134802 (10Legoktm) [22:19:24] 10Beta-Cluster-Infrastructure, 10Mobile-Content-Service, 10Operations, 10Page-Previews, and 3 others: [Bug] Beta cluster page summary endpoint sometimes reponds with 5xx - https://phabricator.wikimedia.org/T192287#4134804 (10Niedzielski) I tried about 50 links and it seems to work. Thanks (and thanks for k... [22:22:04] (03PS2) 10Legoktm: Run phan for MultimediaViewer [integration/config] - 10https://gerrit.wikimedia.org/r/426548 [22:22:13] (03CR) 10Legoktm: [C: 032] Run phan for MultimediaViewer [integration/config] - 10https://gerrit.wikimedia.org/r/426548 (owner: 10Legoktm) [22:23:23] (03Merged) 10jenkins-bot: Run phan for MultimediaViewer [integration/config] - 10https://gerrit.wikimedia.org/r/426548 (owner: 10Legoktm) [22:25:21] !log kick stuch beta-update jenkins jobs [22:25:23] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [22:27:14] 10MediaWiki-Releasing, 10MW-1.31-release: Bundle Multimedia Viewer extension with MW 1.31 - https://phabricator.wikimedia.org/T191744#4134816 (10Legoktm) [22:28:10] (03PS1) 10Legoktm: Bundle MultimediaViewer extension [tools/release] - 10https://gerrit.wikimedia.org/r/427026 (https://phabricator.wikimedia.org/T191744) [22:30:54] (03PS1) 10Krinkle: Remove mwext-jsduck-publish from MinervaNeue postmerge [integration/config] - 10https://gerrit.wikimedia.org/r/427027 [22:31:02] (03CR) 10Krinkle: [C: 032] Remove mwext-jsduck-publish from MinervaNeue postmerge [integration/config] - 10https://gerrit.wikimedia.org/r/427027 (owner: 10Krinkle) [22:32:14] (03Merged) 10jenkins-bot: Remove mwext-jsduck-publish from MinervaNeue postmerge [integration/config] - 10https://gerrit.wikimedia.org/r/427027 (owner: 10Krinkle) [22:32:44] !log Reloading Zuul to deploy https://gerrit.wikimedia.org/r/427027 [22:32:45] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [22:34:18] (03PS1) 10Krinkle: fab: Promote use of [url] instead of [hash] in !log message [integration/config] - 10https://gerrit.wikimedia.org/r/427028 [22:39:46] (03CR) 10Legoktm: [C: 031] "I use URLs mostly because I can just copy them directly out of my IRC client." [integration/config] - 10https://gerrit.wikimedia.org/r/427028 (owner: 10Krinkle) [22:39:55] (03CR) 10Krinkle: [C: 032] fab: Promote use of [url] instead of [hash] in !log message [integration/config] - 10https://gerrit.wikimedia.org/r/427028 (owner: 10Krinkle) [22:40:18] legoktm: thx :) Just wanted to confirm [22:40:40] legoktm: btw, re: code coverage, draft/idea yet for how to switch the bin used for installation on php7? [22:40:57] Is that hardcoded somewhere that can (just) use PHP_BIN instead? Or is it harder? [22:41:15] (03Merged) 10jenkins-bot: fab: Promote use of [url] instead of [hash] in !log message [integration/config] - 10https://gerrit.wikimedia.org/r/427028 (owner: 10Krinkle) [22:41:38] not yet [22:42:04] I'm thinking we could switch /usr/bin/php to default to php 7 tbh [22:42:30] everything except for these weird jobs already should set an explict PHP_BIN I think... [22:42:59] once hhvm is gone, it will make updating in the future easier [22:43:05] ie going from 7.2 to 7.3 [22:43:08] or to 8.0 [22:43:10] legoktm: Hm.. are you saying setting PHP_BIn to php7 would make it use that for install, but the coverage job isn't setting that? [22:43:49] Krinkle: yes. normally zuul/parameter_functions.py sets PHP_BIN, but coverage doesn't get triggered by zuul. [22:43:57] is there a way to set a global env variable for a jenkins job? [22:43:58] rrrright [22:44:26] iirc doing export PHP_BIN in a shell snippet is local only to that part [22:44:41] Yeah, there's something about that. [22:44:48] Jenkins stores each chunk as a .sh file and runs that [22:44:56] as opposed to e.g. sourcing each chunk [22:45:05] so they don't persist, indeed. [22:45:21] but afaik there's nothing special about what zuul does [22:46:21] we should be able to do the same thing from a job config yeah [22:47:24] legoktm: Yeah, it's already there. [22:47:29] for the fake ZUUL parameters :) [22:47:33] 'parameters: ' [22:47:37] >.< [22:47:38] in mediawiki.yaml [22:47:46] ok so PHP_BIN = php7.0 there then [22:47:53] * Krinkle guesses [22:48:50] legoktm: I assume that means we can change 'php7.0' in the shell script for that job back to plain 'php', right? [22:49:01] Assuming that the bin wrapper is available everywhere [22:49:04] yes [22:59:42] Project mwext-phpunit-coverage-publish build #3495: 04FAILURE in 1 min 58 sec: https://integration.wikimedia.org/ci/job/mwext-phpunit-coverage-publish/3495/ [23:02:54] Project mwext-phpunit-coverage-publish build #3496: 04STILL FAILING in 3.6 sec: https://integration.wikimedia.org/ci/job/mwext-phpunit-coverage-publish/3496/ [23:05:18] Yippee, build fixed! [23:05:19] Project mwext-phpunit-coverage-publish build #3497: 09FIXED in 2 min 22 sec: https://integration.wikimedia.org/ci/job/mwext-phpunit-coverage-publish/3497/ [23:05:55] Krinkle: were you going to submit a patch for coverage or should I do it? [23:06:03] legoktm: go ahead :) [23:08:09] (03PS1) 10Legoktm: Set $PHP_BIN in MediaWiki coverage job to force everything onto PHP 7 [integration/config] - 10https://gerrit.wikimedia.org/r/427030 (https://phabricator.wikimedia.org/T147778) [23:10:51] 23:10:27 Using PHP 7.0.29-1+0~20180405085551.17+jessie~1.gbp7bc4a6 [23:10:51] 23:10:32 PHPUnit 6.5.8 by Sebastian Bergmann and contributors. [23:10:53] awesome [23:11:30] (03CR) 10Legoktm: [C: 032] "23:10:27 Using PHP 7.0.29-1+0~20180405085551.17+jessie~1.gbp7bc4a6" [integration/config] - 10https://gerrit.wikimedia.org/r/427030 (https://phabricator.wikimedia.org/T147778) (owner: 10Legoktm) [23:12:50] (03Merged) 10jenkins-bot: Set $PHP_BIN in MediaWiki coverage job to force everything onto PHP 7 [integration/config] - 10https://gerrit.wikimedia.org/r/427030 (https://phabricator.wikimedia.org/T147778) (owner: 10Legoktm) [23:13:45] !log removing obsolete coverage report legoktm@contint1001:/srv/org/wikimedia/doc/cover$ sudo -u jenkins-slave rm -rf mediawiki-core-php7 [23:13:47] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [23:20:55] 10MediaWiki-Codesniffer: Find cases where a function argument without a default is used after one with a default - https://phabricator.wikimedia.org/T192322#4134841 (10Huji) [23:33:25] legoktm: yay [23:38:27] 10MediaWiki-Codesniffer: Find cases where a function argument without a default is used after one with a default - https://phabricator.wikimedia.org/T192322#4134885 (10Legoktm) 05Open>03declined [[https://www.mediawiki.org/wiki/Continuous_integration/Phan|phan]] already checks for this, so I don't think it's... [23:43:05] 10MediaWiki-Codesniffer: DisallowSpaceIndent sniff fails to detect incorrect use of spaces - https://phabricator.wikimedia.org/T183880#4134887 (10Krinkle) [23:44:48] 10MediaWiki-Codesniffer, 10Patch-For-Review: Position of boolean operators inside an if condition - https://phabricator.wikimedia.org/T116561#1752267 (10Krinkle) >>! In T116561#1759336, @Legoktm wrote: > I think you're looking for https://www.mediawiki.org/wiki/Manual:Coding_conventions#Line_continuation ? The... [23:47:29] 10MediaWiki-Codesniffer: Treat skin templates differently - https://phabricator.wikimedia.org/T104511#4134893 (10Krinkle) 05Open>03Resolved a:03Krinkle Use of `endif` is allowed in our style configuration. The comparison to curly braces therefore is not as important. The addition of new lines seems harmle... [23:49:40] legoktm: The xml normalisation thing in jenkins is quite annoying and makes job config diffs almost useless. [23:49:49] Do you know of a task that tracks this and/or a workaround? [23:49:56] E.g. https://integration.wikimedia.org/ci/job/mediawiki-core-code-coverage/jobConfigHistory/showDiffFiles?timestamp1=2018-04-16_14-06-06×tamp2=2018-04-16_23-09-59 [23:49:58] omg yes. I don't [23:50:08] It is really annoying [23:50:10] The constant back-and-forth between /> and <> and new lines and stuff [23:50:18] I just rely on diffing JJB locally [23:50:23] Yeah [23:50:31] Oh well. [23:50:44] So, can we close https://phabricator.wikimedia.org/T147778 now :) [23:52:43] It would make us go from 36/44 resolved to 37/44 resolved on PHP7 sub tasks. That's almost 3% ! [23:52:44] :D [23:55:54] Krinkle: yeah, just waiting for the run I triggered to finish [23:55:58] k