[00:04:19] lol [00:06:19] RainbowSprinkles how many films do you actually have? I own more then 10 on itunes then i have access to sky go and amazon prime (used to be lovefilm) [00:06:44] Probably ~250 or so [00:06:53] Oh wow thats alot [00:07:32] I have access to a ton of new films through sky go. But netflix i found had old films compared to amazon prime. [03:58:34] PROBLEM - Free space - all mounts on deployment-fluorine02 is CRITICAL: CRITICAL: deployment-prep.deployment-fluorine02.diskspace._srv.byte_percentfree (<30.00%) [04:07:21] Project selenium-MultimediaViewer » safari,beta,OS X 10.9,BrowserTests build #311: 04FAILURE in 11 min: https://integration.wikimedia.org/ci/job/selenium-MultimediaViewer/BROWSER=safari,MEDIAWIKI_ENVIRONMENT=beta,PLATFORM=OS%20X%2010.9,label=BrowserTests/311/ [04:23:50] Eh, post-merge builds for mediawiki-core just randomly started failing. Because the git-clone of mediawiki/core is timing out after 10min. [04:24:03] It's fetching from contint2001 instead of contint1001 as of a few minutes ago [04:24:15] https://integration.wikimedia.org/ci/job/mediawiki-core-jsduck-publish/10197/console [05:41:39] 10Scap, 06WMF-Legal, 07Documentation, 07Software-Licensing: Scap is lacking a license - https://phabricator.wikimedia.org/T94239#3052296 (10Slaporte) I recommend the GPLv3. If you have a preference for Apache 2.0, please shoot me an email. [06:25:42] Yippee, build fixed! [06:25:42] Project selenium-Wikibase » chrome,test,Linux,BrowserTests build #281: 09FIXED in 1 hr 45 min: https://integration.wikimedia.org/ci/job/selenium-Wikibase/BROWSER=chrome,MEDIAWIKI_ENVIRONMENT=test,PLATFORM=Linux,label=BrowserTests/281/ [06:45:51] 10Gerrit, 07Regression: Automatic reviewer assignment does not work for certain repos - https://phabricator.wikimedia.org/T158844#3052317 (10Osnard) Thanks for the quick response. I will check if the issue still occurs [06:58:34] RECOVERY - Free space - all mounts on deployment-fluorine02 is OK: OK: All targets OK [08:29:51] 10Gerrit, 06Release-Engineering-Team: Update gerrit to 2.13.6 - https://phabricator.wikimedia.org/T158946#3052390 (10Paladox) [08:35:43] 10Gerrit: jenkins-bot ran gate-and-submit and voted V+2, but did not submit (merge) the change - https://phabricator.wikimedia.org/T155558#3052403 (10Paladox) This looks like a zuul problem rather then a gerrit problem as it seemed you could press the submit button to merge. Has this problem happened since repor... [08:56:43] 10Gerrit: jenkins-bot ran gate-and-submit and voted V+2, but did not submit (merge) the change - https://phabricator.wikimedia.org/T155558#3052433 (10hashar) Paladox it is definitely an issue in Gerrit, due to a lock on the branch tip which prevents Gerrit from updating it and thus failing the submit: ``` Error... [09:08:12] 10Gerrit: jenkins-bot ran gate-and-submit and voted V+2, but did not submit (merge) the change - https://phabricator.wikimedia.org/T155558#3052443 (10Paladox) Oh, it may be fixed in gerrit 2.14 as I did see some commits related to performance improvements and others to do with fixing duplicating opening the repo. [09:10:19] 10Gerrit: jenkins-bot ran gate-and-submit and voted V+2, but did not submit (merge) the change - https://phabricator.wikimedia.org/T155558#3052462 (10Paladox) I found someone else reported this against gerrit 2.13 https://groups.google.com/forum/m/#!topic/repo-discuss/lvLRH8bmvl4 [09:19:16] 10Gerrit: jenkins-bot ran gate-and-submit and voted V+2, but did not submit (merge) the change - https://phabricator.wikimedia.org/T155558#3052489 (10Paladox) I see this https://gerrit-review.googlesource.com/#/c/87512/ change upstream which turns the error into a 500. But that's on the master branch. [09:57:10] Tobi_WMDE_remote: just stumbled upon this piece of data, I remember we have talked about it yesterday [09:57:37] Sauce Labs: the last version of Chrome with Linux support is 48, and for Firefox it's version 45 [09:58:01] sauce labs support said: [09:58:02] The reason why the later versions of these browsers are not listed under Linux is because they require the use of the 64-bit version of Linux. Currently, our Linux image is 32-bit, but one of our action items is to update the image so that we can support the latest versions of Chrome and Firefox when paired with Linux. [09:58:12] We don't have a date of when this will occur, but I highly recommend subscribing to our Product Announcements and Release Notes blog to get all the up-to-date information of our service as it comes: https://wiki.saucelabs.com/display/DOCS/Product+Announcements+and+Release+Notes+Blog [10:20:10] (03PS1) 10Umherirrender: [WPtouch] Add npm job [integration/config] - 10https://gerrit.wikimedia.org/r/339617 [10:29:26] zeljkof: thanks for the info! [10:51:32] addshore: if you are around :} I think I found a nice way to include mediawiki autoload-dev to vendor.git [10:51:41] ohhaloo [10:51:43] addshore: after hours of debugging it. I got a one liner fix! [10:51:57] basically we grab the list of dev dependencies from mediawiki [10:52:03] composer require --dev them inside /vendor/ [10:52:34] then we can use the composer merge plugin to grab the autoload-dev sections (and others) and update /vendor/composer.json with the one line fix: [10:52:36] + composer config extra.merge-plugin.include "$MW_INSTALL_PATH/composer.json" [10:52:49] then running composer dump-autoload --optimize does inject the section [10:52:50] \O/ [10:53:27] what piss me off is I am pretty sure that bd808, Krinkle, legoktm would have fine that solution in a few minutes since they know how composer merge plugin works [10:53:31] so I am gonna feature flag that behavior and deploy [10:54:21] hahaa, sounds like it would work :P [10:55:27] though I am not sure what kind of mess composer-merge-plugin will randomly add / merges [10:55:27] worth giving it a shot though [11:04:45] (03PS6) 10Hashar: Option to merge mw composer into vendor [integration/jenkins] - 10https://gerrit.wikimedia.org/r/339202 (https://phabricator.wikimedia.org/T158674) (owner: 10Aude) [11:06:06] (03CR) 10Hashar: "I think I managed to dramatically simply the fix. Before autoload dump I just enable the composer-merge-plugin in$ vendor/composer.json wi" [integration/jenkins] - 10https://gerrit.wikimedia.org/r/339202 (https://phabricator.wikimedia.org/T158674) (owner: 10Aude) [11:06:40] 10Gerrit: jenkins-bot ran gate-and-submit and voted V+2, but did not submit (merge) the change - https://phabricator.wikimedia.org/T155558#3052671 (10Paladox) This may be fixed by https://gerrit-review.googlesource.com/#/c/98530/ but it is really only a guess fix. [11:44:51] (03CR) 10Hashar: "Want a big changelog/release notes about the firefox now requiring geckodriver/Marionette." (031 comment) [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [11:45:24] hashar: worth a try [11:45:36] aude: yeah will do on one of the jenkins slaves [11:45:38] unpool it [11:45:41] ok [11:45:51] copy paste the wikibase jobs and point them to the unpooled slave [11:45:55] k [11:45:57] then run different kind of tests [11:46:03] wikibase but also mediawiki core branches [11:46:11] and see what happens. If all fine I guess I will just merge it [11:46:26] I am doing a few more tests locally first [11:46:49] sounds good :) [11:46:58] (03CR) 10Hashar: [C: 032] "Changelog / release notes will be done in a different patch when the gem is about to be released." [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [11:48:09] (03CR) 10jerkins-bot: [V: 04-1] Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [11:59:47] zeljkof: fails the gate :( [12:00:10] comparison of Symbol with 0 failed (ArgumentError) [12:00:16] (03CR) 10Hashar: [C: 032] "recheck" [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [12:00:25] hashar: huh [12:00:27] looking [12:00:48] comparison of Symbol with 0 failed (ArgumentError) [12:00:57] no clue what that means [12:01:11] I am out for lunch [12:02:08] (03CR) 10jerkins-bot: [V: 04-1] Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [12:02:15] hm, the job was fine a day or two ago :( [12:04:26] (03CR) 10jerkins-bot: [V: 04-1] Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [12:05:32] (03PS7) 10Hashar: Option to merge mw composer into vendor [integration/jenkins] - 10https://gerrit.wikimedia.org/r/339202 (https://phabricator.wikimedia.org/T158674) (owner: 10Aude) [12:09:23] will take a look after lunch [13:21:16] (03PS10) 10Zfilipin: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) [13:22:18] Hello [13:23:07] (03CR) 10jerkins-bot: [V: 04-1] Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [13:23:33] (03PS11) 10Zfilipin: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) [13:24:05] hashar: what has changed on the autoload-dev thing [13:24:07] Anything? [13:25:56] (03CR) 10jerkins-bot: [V: 04-1] Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [13:31:13] (03CR) 10Zfilipin: "Argh, can not reproduce the failure locally. :(" [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [13:34:41] (03PS12) 10Zfilipin: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) [13:37:09] (03CR) 10jerkins-bot: [V: 04-1] Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [13:39:29] (03PS8) 10Hashar: Option to merge mw composer into vendor [integration/jenkins] - 10https://gerrit.wikimedia.org/r/339202 (https://phabricator.wikimedia.org/T158674) (owner: 10Aude) [13:40:11] (03PS13) 10Zfilipin: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) [13:42:09] (03CR) 10jerkins-bot: [V: 04-1] Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [13:42:25] zeljkof: could it be that a new water or selenium-driver got released in between ? [13:42:49] hashar: both [13:42:56] but downgrading did not help :( [13:42:56] ;D [13:43:05] checking all gem versions [13:43:19] it's frustrating that I can not reproduce on my machine [13:44:51] Running `bundle install --verbose` with bundler [-1.14.4-] {+1.14.5+} [13:44:57] a new bundler version as well [13:45:14] hm [13:45:21] bundler should not be the problem [13:45:26] but who knows [13:45:27] :( [13:46:09] I am looking at the diff of the consoles [13:46:17] for the last passing build and the first failing one [13:46:22] Yippee, build fixed! [13:46:22] Project selenium-VisualEditor » firefox,beta,Linux,BrowserTests build #318: 09FIXED in 2 min 20 sec: https://integration.wikimedia.org/ci/job/selenium-VisualEditor/BROWSER=firefox,MEDIAWIKI_ENVIRONMENT=beta,PLATFORM=Linux,label=BrowserTests/318/ [13:46:34] Installing selenium-webdriver [-3.1.0-] {+3.2.0+} [13:46:34] Installing watir [-6.1.0-] {+6.2.0+} [13:46:48] Installing i18n [-0.8.0-] {+0.8.1+} [13:46:57] Using bundler [-1.14.4-] {+1.14.5+} [13:49:09] PROBLEM - Puppet run on buildlog is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [0.0] [13:49:57] yes, both watir and selenium updated, just checked the diffs [13:50:07] (03CR) 10Hashar: [C: 032] "Tested it locally against few branches and that looks all good now." [integration/jenkins] - 10https://gerrit.wikimedia.org/r/339202 (https://phabricator.wikimedia.org/T158674) (owner: 10Aude) [13:50:15] let's see if locking both down helps, I think I have tried locking each separately [13:51:07] (03Merged) 10jenkins-bot: Option to merge mw composer into vendor [integration/jenkins] - 10https://gerrit.wikimedia.org/r/339202 (https://phabricator.wikimedia.org/T158674) (owner: 10Aude) [13:51:41] :) [13:52:14] (03PS14) 10Zfilipin: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) [13:52:35] !log deployed slave script update to be able to merge mediawiki/composer.json into vendor/composer.json 6527f49..a7728a5 https://gerrit.wikimedia.org/r/#/c/339202/ T158674 [13:52:39] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [13:52:39] T158674: Fatal error: Call to undefined function Wikibase\Client\Tests\RecentChanges\both() - on jenkins - https://phabricator.wikimedia.org/T158674 [13:53:02] aude: gotta refresh the nodepool instances now :) [13:53:12] ok [13:53:28] let me know how it works [13:54:11] (03CR) 10jerkins-bot: [V: 04-1] Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [13:56:03] !log Log refresh Nodepool instances to deploy slave script update to be able to merge mediawiki/composer.json into vendor/composer.json 6527f49..a7728a5 https://gerrit.wikimedia.org/r/#/c/339202/ T158674 [13:56:07] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [13:59:39] BAH i broke it :( [14:00:16] hashar, you do have a backup... right? [14:06:17] (03PS1) 10Hashar: mw-fetch-composer-dev: shallow [ -f exit code [integration/jenkins] - 10https://gerrit.wikimedia.org/r/339633 [14:06:35] (03CR) 10Hashar: [C: 032] mw-fetch-composer-dev: shallow [ -f exit code [integration/jenkins] - 10https://gerrit.wikimedia.org/r/339633 (owner: 10Hashar) [14:06:37] Zppix: it is git [14:06:38] so yeah [14:06:44] it has the whole history of all changes [14:07:11] thats not what i thought you broke [14:09:33] (03CR) 10Hashar: [C: 032] "Various composer related files now get attached to the build. Example https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm/38" [integration/jenkins] - 10https://gerrit.wikimedia.org/r/339202 (https://phabricator.wikimedia.org/T158674) (owner: 10Aude) [14:17:11] (03PS1) 10Hashar: Set MW_COMPOSER_MERGE_MW_IN_VENDOR for Wikidata build [integration/config] - 10https://gerrit.wikimedia.org/r/339634 (https://phabricator.wikimedia.org/T158674) [14:21:42] 2017-02-24 14:21:31,133 INFO nodepool.image.build.wmflabs-eqiad.snapshot-ci-trusty: Error: Removing mount "files": /etc/puppet/files does not exist or is not a directory [14:21:56] the never ending fight against error messages [14:22:26] * Zppix hands hashar a hammer [14:22:30] try hitting with that a few times [14:30:26] (03CR) 10Hashar: [C: 032] "Lets see what kind of magic happens on the build that failed https://gerrit.wikimedia.org/r/#/c/338940/" [integration/config] - 10https://gerrit.wikimedia.org/r/339634 (https://phabricator.wikimedia.org/T158674) (owner: 10Hashar) [14:31:15] (03Merged) 10jenkins-bot: Set MW_COMPOSER_MERGE_MW_IN_VENDOR for Wikidata build [integration/config] - 10https://gerrit.wikimedia.org/r/339634 (https://phabricator.wikimedia.org/T158674) (owner: 10Hashar) [14:37:14] hashar: we can't put psy/psysh in the WMF vendor repo due to some PHP 7 test code :P [14:39:48] aude: Setting extra.merge-plugin.include does not exist or is not supported by this command [14:39:50] :((((((((((((((((( [14:40:10] I did my tests using a newer composer version :(((((((((((((((((((((((((((((((((((((((((((((((((((( [14:40:39] Reedy: get whatever runs the php linting to excludes those files? or whatever [14:41:02] Reedy: on the mw/core change that introduced psysh Gergo mentionned it was probably not worth the trouble to add it in vendor [14:41:08] Heh [14:41:16] Thought it might be useful to have on the WMF cluster [14:41:27] Isn't that a bigger problem we've still not managed to solve for the autoload files generated by newer composer? [14:41:36] Even though php lint can exclude specific files [14:44:41] we should really fix that specific bug on the composer lint job [14:45:10] well the linter works as intended [14:45:15] it bails out because the code is invalid [14:45:26] -;} [14:46:21] well, invalid on that version of php [14:46:25] not invalid on newer versions [14:47:22] "You should not be linting test fixtures of the vendor library." [14:47:25] Thanks symfony [14:47:46] CHANGELOG.md: * Added support for editing all top-level properties (name, minimum-stability, ...) as well as extra values via the `config` command [14:47:47] neat [14:50:14] (03PS15) 10Zfilipin: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) [14:51:17] 10Continuous-Integration-Config, 10Wikidata, 13Patch-For-Review: Fatal error: Call to undefined function Wikibase\Client\Tests\RecentChanges\both() - on jenkins - https://phabricator.wikimedia.org/T158674#3053125 (10hashar) Gave it a try on a Wikidata change and the job fails https://integration.wikimedia.or... [14:53:04] addshore: aude: so that fails. Our composer 1.0.3 doesn't let us set arbitrary settings such as extra.include :( [14:53:32] all of that cause I have used locally a new composer version grrrr [14:53:46] hashar: Well... [14:53:55] If we fix the problem caused by the newer one [14:54:07] ie stop the parrallel linter from linting that autoload file... [14:54:11] We can use newer composer [14:54:16] (03PS16) 10Zfilipin: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) [14:54:48] Reedy: well maybe the parallel linter has some exclude/ignore option ? [14:54:57] It does [14:55:07] It was (wrongly) documented as excluding dirs [14:55:12] It actually excludes files too [14:55:24] https://github.com/JakubOnderka/PHP-Parallel-Lint/commit/3c7a3c4330ccf8507f98111d495d89f6dfe28ce0 [14:55:50] But nothing we actually need to bump versions for [14:56:09] (03PS17) 10Zfilipin: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) [14:59:39] (03CR) 10Hashar: [C: 032] "PS9 was only locked to major version. PS15 does it properly: ~> 3.1.0" [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [15:00:39] zeljkof: +2ed [15:00:51] (03Merged) 10jenkins-bot: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [15:00:55] hashar: thanks! [15:01:00] hashar: we should just need to add a --exclude=composer/autoload_static.php in the right place [15:01:08] I was looking at both watir and selenium code, did not notice anything [15:01:11] I guess I have missed it [15:01:47] (03CR) 10jenkins-bot: Upgrade to Selenium 3 [selenium] - 10https://gerrit.wikimedia.org/r/336824 (https://phabricator.wikimedia.org/T157695) (owner: 10Zfilipin) [15:09:49] hashar: :( [15:10:37] and composer does not even git tag their releases [15:10:38] * aude thinks we should stop supporting php 5.5 [15:10:56] oh no [15:10:57] I forgot to git fetch --tags [15:12:17] git log --oneline --no-merges 1.0.3..1.1.0-RC|wc -l > 74 patch [15:12:18] es [15:12:41] the thing is Jan Zerebecki tried to upgrade composer and it failed to run on PHP5.5. :( [15:13:49] 5.5 has passed end of life [15:13:50] http://php.net/supported-versions.php [15:14:00] no security fixes, etc [15:14:17] we can't use latest stable phpunit, etc [15:17:08] and given we are about to cut a LTS [15:17:14] yeah probably want to open that can of worms [15:17:41] though 5.6 is supported till December 2018 [15:17:52] yeah [15:17:58] i think 5.6 is not so much a problem yet [15:18:08] so if we cut our LTS in May 2017 and claim we support it for three years [15:18:31] we will have to keep using php 5.5 or 5.6 until May 2020. Which 6 months after EOL of PHP 7.1 [15:18:59] https://github.com/sebastianbergmann/phpunit/blob/master/ChangeLog-6.0.md [15:19:11] so we won't be able to use phpunit 6.0 [15:19:17] but can use 5.x [15:20:08] imho [15:20:14] we should stop mediawiki LTS [15:20:18] ;D [15:20:21] :) [15:20:40] cause clearly the whole world is moving toward releasing often and using a short life span [15:20:41] And/or not support LTS' for quite so long [15:20:57] I mean [15:21:00] then it is no more **L**TS [15:21:04] lol [15:21:07] How much of a fight did we have to bump from 5.3 to 5.5 [15:21:18] And we knew 5.5 really wasn't good enough in terms of long term PHP support [15:21:27] that dragged for a while [15:21:40] cause folks complained their current distro of choice was still 5.5 [15:21:47] IIRC we had people saying we should have "one more" LTS with 5.3 support etc [15:21:48] they wanted the new mediawiki on their old setup [15:22:34] and we still have 1.23 to support in CI :/ [15:23:02] Right, indeed [15:23:14] oh no [15:23:15] I really don't get all that. Want to use newest of some random things, with ooold of other things [15:23:27] it is our 1.27 form last year which is LTS :} [15:25:20] hashar: Where's the config/whatever for the php linting of the composer repo? [15:25:29] Let's see if we can get that fixed, and start using newer composer :) [15:25:30] I dont know [15:25:36] rofl [15:25:57] the only thing I know is that composer 1.1.0-RC was breaking builds for some reason https://gerrit.wikimedia.org/r/#/c/286305/ [15:26:11] > Because it is not compatible with PHP 5.5. [15:26:35] Right [15:26:45] the commit that bump it is https://gerrit.wikimedia.org/r/#/c/283852/ [15:26:46] IT was the one autoload_static.php file [15:26:55] Which needs excluding from our linting on php 5.5 [15:26:56] and maybe that got generated using PHP 5.6++ [15:27:10] which would cause some files to be generated in a way that is supported by 5.6 but not by 5.5 [15:27:26] maybe we can try updating again and making sure we use 5.5 for all the operations [15:28:08] Using the previous composer version after the next one in [15:28:08] 8a176a28da0d967fdfc2f3426428941b59db6d68 produced an autoloader that was not [15:28:08] PHP 5.5 compatible. [15:28:09] [15:28:10] yeah [15:28:45] The annoying one, would be making sure people had a local php 5.5 instance to update composer stuff [15:28:46] oh Reedy sorry [15:28:56] now I got what you mean https://phabricator.wikimedia.org/T135161 Composer v1.1.0 generated vendor dirs will fail lint by PHP <5.6 [15:28:59] It may still generate the offending file though [15:29:11] heh :) [15:29:26] if I look at the history in https://gerrit.wikimedia.org/r/#/q/project:integration/composer [15:29:45] I would guess the attempt to update to 1.1.0-RC generated the files using 5.6 + [15:29:55] That sounds about right [15:30:12] it is not like I have any clue what I assuming :} [15:33:47] I would suggest we make a commit using latest composer, watch it fail, exclude that file, see if it passes etc [15:34:24] and the doc I wrote in integration/composer README.md is off [15:39:13] we should update https://phabricator.wikimedia.org/diffusion/CIJJ/ to upstream. :) [15:42:58] (03PS1) 10Hashar: Update composer to 1.1.0 [integration/composer] - 10https://gerrit.wikimedia.org/r/339645 (https://phabricator.wikimedia.org/T125343) [15:44:15] Parse error: parse error, expecting `')'' in ./vendor/composer/autoload_static.php on line 10 [15:44:17] ... [15:45:18] hashar: yeah, autoload static was the problem file [15:45:53] (03CR) 10Hashar: "Bah:" [integration/composer] - 10https://gerrit.wikimedia.org/r/339645 (https://phabricator.wikimedia.org/T125343) (owner: 10Hashar) [15:46:12] I have used composer 1.3.1 being run under PHP 5.5 to generate that one [15:48:38] So I guess composer just statically writes the file out [15:49:09] I don't even got why the syntax is wrong on line 10 of https://gerrit.wikimedia.org/r/#/c/339645/1/vendor/composer/autoload_static.php [15:49:10] :/ [15:49:56] it'll be some referred error [15:49:56] ah string contatenation [15:50:27] you can't do: ✘ public static $files = 'a' . 'b'; [15:50:27] :/ [15:54:39] fun [15:57:18] (03PS1) 10Hashar: Update composer to 1.1.0 with composer 1.1.0/PHP 5.5 [integration/composer] - 10https://gerrit.wikimedia.org/r/339650 [15:57:56] grrr [15:58:06] (03Abandoned) 10Hashar: Update composer to 1.1.0 with composer 1.1.0/PHP 5.5 [integration/composer] - 10https://gerrit.wikimedia.org/r/339650 (owner: 10Hashar) [15:58:26] (03PS2) 10Hashar: Update composer to 1.1.0 with composer 1.1.0/PHP 5.5 [integration/composer] - 10https://gerrit.wikimedia.org/r/339645 (https://phabricator.wikimedia.org/T125343) [16:01:19] hashar i think we can fix ^^ if we find a way to ignore file autoload_static.php on the phplint and composers phpcs and composers phplint :) [16:01:31] yeah sure [16:01:34] we have done it for scap [16:01:40] then when one runs composer and it tries require that file [16:01:43] build explodes [16:02:03] yep, but i mean linting the file. [16:02:06] upstream commit is apparently fd2f51cea8e5f1ef978cd8f90b87b69dc1778976 [16:02:24] src/Composer/Util/Filesystem.php has a method findShortestPathCode() [16:02:37] which has a bool parameter $staticCode = false [16:02:48] yeh upstream wont budge in allowing us to disable generating the file. [16:02:56] and when using static code it returns the full path on the file system [16:03:09] else the default is to return __DIR__ . ... [16:03:19] yeh. [16:03:33] so when generating the autoload_static.php file at some place findShortestPathCode() is invoked without the $staticCode=true [16:03:34] grr [16:04:03] oh [16:05:49] https://github.com/composer/composer/commit/fd2f51cea8e5f1ef978cd8f90b87b69dc1778976 [16:05:56] will update the task [16:08:43] ok [16:09:12] hashar now only if we have a way to ajust content in that file using the script field in composer.json. [16:10:10] oh wait [16:10:55] hashar i have an idea. What about removing the autoload_static.php file and replacing it with the one made for php 5.5 or lower. so we will have autoload_static.php but with the same content as the other files. [16:11:06] Yeh its duplicating but it will fix our issue. [16:12:05] oh it does this call_user_func(\Composer\Autoload\ComposerStaticInit_mediawiki_vendor::getInitializer($loader)); too [16:12:16] so we will need to add that function but make sure it does nothing. [16:12:39] that idea wont work. [16:18:29] Reedy: now I get what you were saying :D [16:18:44] (03CR) 10Paladox: Update composer to 1.1.0 with composer 1.1.0/PHP 5.5 (036 comments) [integration/composer] - 10https://gerrit.wikimedia.org/r/339645 (https://phabricator.wikimedia.org/T125343) (owner: 10Hashar) [16:19:23] hashar we could always customise it for ci. It wont stop it breaking for users who use it but will for when ci runs composer update. [16:20:12] (03CR) 10Paladox: "Always we should upgrade to 1.3.1." [integration/composer] - 10https://gerrit.wikimedia.org/r/339645 (https://phabricator.wikimedia.org/T125343) (owner: 10Hashar) [16:20:38] 10Continuous-Integration-Infrastructure, 07Composer, 13Patch-For-Review: Upgrade integration/composer to 1.3.2 stable - https://phabricator.wikimedia.org/T125343#3053249 (10Paladox) [16:24:35] hashar we only need to figure out how to ignore that file here https://github.com/wikimedia/integration-jenkins/blob/master/bin/git-changed-in-head#L48 [16:24:43] no [16:24:47] I am not changing that script [16:24:55] 10Gerrit, 07artificial-intelligence: Patch-wrangler -- suggests the best reviewers for a patch - https://phabricator.wikimedia.org/T155851#3053275 (10Aklapper) **[Kind of offtopic comment: Posting too many links to academic papers about (mostly pretty technical) bug assignee recommendation approaches - feel fr... [16:24:59] but that's phplint. [16:25:09] I would rather drop phplint entirely [16:25:17] and solely use the PHP PArallal-lint package [16:26:28] Oh ok :) [16:26:55] So we should drop the phplint that dosent use composer and use the composer one? [16:27:07] yeah [16:27:18] the phplint jobs are doing: [16:27:25] We can do that :), i am unsure if mw core has a test that uses it. [16:27:26] something like: [16:27:34] find . -name *.php -0|xargs -0 -n1 php -ll [16:27:45] yep [16:28:34] the thing is that if we start adding --exclude / ignore to git-changed-in-head it is going to be messy very fast [16:28:46] and yep [16:28:50] I will go and look at mw core now. I will switch it to using phplint from composer if it dosent use it. [16:29:46] ah nope it dosen't https://github.com/wikimedia/integration-config/blob/8d6d146d833a548f5fe0426c31ca29f47324f9c0/jjb/mediawiki.yaml#L327 [16:29:56] no [16:30:02] because that takes ages to lint [16:30:06] oh [16:32:24] try on your mediawiki/core composer install && composer lint . [16:32:29] hashar https://github.com/overtrue/phplint [16:32:29] but we could end up with something like: [16:32:30] ~/projects/integration/jenkins/bin/git-changed-in-head php|xargs composer lint [16:32:33] that one caches it [16:32:43] and oh [16:32:47] ah [16:32:54] did'nt know that. [16:35:14] so for mediawiki/core [16:35:26] we would change the phplint jobs by some new job [16:35:51] yep, we could do what you suggested git-changed-in-head php|xargs composer lint [16:36:11] maybe even git-changed-in-head php|xargs composer install && composer lint [16:36:12] bonus [16:36:24] we could then also phase out the mediawiki-core-phpcs job [16:36:28] yeh [16:36:30] that runs the code style against every single files [16:36:33] put it into one job [16:36:35] so basically merge both [16:36:40] yeh [16:36:43] and have them only run against HEAD [16:37:03] i will see if i can create a patch now. [16:37:13] well [16:37:15] gotta test it [16:37:20] yep [16:37:26] and have that supported on all branches [16:37:28] and on both mediawiki/core and mediawiki/vendor [16:37:57] extensions will be impacted as well [16:38:06] so all of them would need to ignore the static file :/ [16:38:16] though they probably already ignore vendor anyway [16:38:22] yep [16:38:36] well we could do vendor/bin/phpcs for example [16:38:51] you know instead of composer lint [16:38:58] so we can merge some tests even furhtur [16:39:02] furthur like phplint [16:39:44] but then we want phpcs. [16:40:26] actually in mediawiki/core [16:40:30] the composer test would run: [16:40:33] composer lint [16:40:35] composer phpcs [16:40:41] yep [16:40:44] ;-) [16:40:51] ah [16:40:56] so we can drop the phplint jobs and the mediawiki-core-phpcs [16:41:01] and replace them both with a new: [16:41:10] mediawiki-composer-test-HEAD [16:41:15] which would install the dependencies [16:41:16] what about php53? [16:41:27] then git-changed-in-head | xargs composer test [16:41:42] :) [16:41:46] probably restricting to php files [16:41:50] I don't know really :( [16:42:23] oh wait [16:42:31] doint we want to get rid of php53 tests? [16:42:50] we could just keep them until we want to drop those as not many will be running anyways [16:43:31] anyway, I need a break and get some fresh air [16:43:36] ok [16:53:39] paladox: actually I would even drop our script git-changed-in-head :} [16:53:50] Oh :) [16:53:56] looking at PHP CodeSniffer, it supports running some PHP files to bootstrap it [16:54:03] * paladox is creating the testing patch [16:54:05] so maybe that can be used to detect whether we run on CI [16:54:17] find the file that changes by shelling out to git just git-changed-in-head [16:54:24] oh [16:54:50] then only keep files that have an extension mentioned in phpcs.xml: [16:54:56] [16:55:03] and feed the resulting list to phpcs as files to process [16:55:04] bah [16:55:10] i would need help with dropping that one as i would not know how to use git to detect file changes. and oh [16:55:56] I am just having random thoughts out loud [16:56:20] :) [17:01:46] yeah [17:01:50] we can hack it :} [17:02:38] (03PS1) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 [17:02:39] hashar ^ [17:03:41] woops i added a cd to the script [17:03:46] * paladox removes it [17:04:06] (03PS2) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 [17:05:50] (03CR) 10jerkins-bot: [V: 04-1] Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (owner: 10Paladox) [17:07:18] (03PS3) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 [17:08:54] (03CR) 10jerkins-bot: [V: 04-1] Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (owner: 10Paladox) [17:10:03] (03PS4) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 [17:10:29] 10Continuous-Integration-Config, 10MediaWiki-Unit-tests: MediaWiki core PHPCS job should only run against files changed in HEAD - https://phabricator.wikimedia.org/T158974#3053371 (10hashar) [17:11:27] (03PS5) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (https://phabricator.wikimedia.org/T158974) [17:13:16] (03CR) 10jerkins-bot: [V: 04-1] Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (https://phabricator.wikimedia.org/T158974) (owner: 10Paladox) [17:13:22] (03PS6) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (https://phabricator.wikimedia.org/T158974) [17:15:05] (03PS7) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (https://phabricator.wikimedia.org/T158974) [17:16:02] (03PS8) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (https://phabricator.wikimedia.org/T158974) [17:17:24] (03CR) 10jerkins-bot: [V: 04-1] Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (https://phabricator.wikimedia.org/T158974) (owner: 10Paladox) [17:17:52] (03PS9) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (https://phabricator.wikimedia.org/T158974) [17:33:26] PROBLEM - Puppet run on deployment-tin is CRITICAL: CRITICAL: 44.44% of data above the critical threshold [0.0] [17:33:33] 10Gerrit, 10MediaWiki-Vagrant: "index-pack failed" when installing new MediaWiki-Vagrant box - https://phabricator.wikimedia.org/T152801#3053418 (10bd808) [17:46:30] paladox: https://gerrit.wikimedia.org/r/339675 PHP CodeSniffer on CI should only lint HEAD :)} [17:46:51] oh [17:46:53] :) [17:47:03] it is running at https://integration.wikimedia.org/ci/job/mediawiki-core-phpcs-trusty/11970/console [17:47:14] so potentially we can later add a composer-test job to core [17:47:20] and it should be fast [17:47:26] we will see [17:47:29] I am off for now. Been a long and tiring day :( [17:47:33] Ok [17:47:48] i was about to test my patch on gerrit-jenkins [17:48:03] but zuul-merger wont start [17:48:07] but it connects properly to jenkins as gerrit-zuul picks up the change. [17:48:11] Time: 415ms; Memory: 2.69Mb [17:48:12] haha [17:48:19] lol, thats quick :) [17:48:24] it still takes 96 seconds [17:48:27] due to clones and composer [17:48:28] :D [17:48:41] happy week-end [17:48:49] yep and you too :) [18:02:20] PROBLEM - Puppet run on deployment-tmh01 is CRITICAL: CRITICAL: 22.22% of data above the critical threshold [0.0] [18:02:30] whoo [18:02:32] it worked [18:02:32] http://gerrit-jenkins.wmflabs.org/job/composer-php-55/8/console [18:03:30] (03CR) 10Paladox: "@hashar i've tested this patch and it worked http://gerrit-jenkins.wmflabs.org/job/composer-php-55/8/console" [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (https://phabricator.wikimedia.org/T158974) (owner: 10Paladox) [18:03:43] (03PS10) 10Paladox: Add new mediawiki-core-composer-test-HEAD-php55lint-trusty test [integration/config] - 10https://gerrit.wikimedia.org/r/339666 (https://phabricator.wikimedia.org/T158974) [18:04:11] Reedy ^^ :) [18:08:29] RECOVERY - Puppet run on deployment-tin is OK: OK: Less than 1.00% above the threshold [0.0] [18:42:18] RECOVERY - Puppet run on deployment-tmh01 is OK: OK: Less than 1.00% above the threshold [0.0] [18:44:16] What is the modern equivalent to 'sync common'? [18:58:52] andrewbogott is that for scap? [18:59:06] Yeah, I just needed to do 'scap pull' [19:00:17] thcipriani ^^ [19:01:30] yup scap pull is the new sync-common :) [19:29:56] Is this something? [19:29:57] Class 'MediaWiki\\Logger\\LoggerFactory' not found in /srv/mediawiki/php-1.29.0-wmf.13/includes/GlobalFunctions.php on line 1097 [19:30:07] (That's happening on the labtest wikitech box) [19:30:39] oh, it's from a hack to turn on local logging. Guess that hack doesn't work anymore [19:32:06] thcipriani, have a moment to help me with this? I'm trying to use this but it has clearly rotted... [19:32:09] https://www.irccloud.com/pastebin/jeTMScGz/ [19:32:16] Is there a simple updated equivalent? [19:32:21] * thcipriani looks [19:33:53] andrewbogott: oh boy — I'm not sure about that one. I'm not the right person for that question :) [19:34:06] ok :) [19:34:11] maybe ask in -core? or maybe RainbowSprinkles has some idea about it... [19:35:48] cc bd808 [19:37:08] andrewbogott: I can probably help... I build most of that fragile crap [19:37:32] bd808: ok… see paste above for the current state of the code that's failing [19:41:46] andrewbogott: where do you set that? [19:42:05] it's in wmf-config/wikitech.php [19:42:14] so loaded during the config phase, let me see where exactly... [19:42:27] It's loaded in CommonSettings [19:44:58] I remember working on this with Alex a long time ago... [19:45:16] It's not just that 'MediaWiki\\Logger\\LoggerFactory' has been renamed? [19:45:19] trying to figure out if it still works with changes that have happened in logging.php [19:46:15] no, that's the name of the class [19:46:30] but... this code all happens before the class loader I think [19:46:56] no, that doesn't make sense [19:48:00] there is a 'use MediaWiki\Logger\LoggerFactory' at the top of commonsettings [19:49:15] I think you have a login on this box, if you want to poke [19:49:20] labtestweb2001.wikimedia.org [19:50:48] You could also tell me to give this up and to read the central logs, and then tell me where the central logs are :) [19:53:27] andrewbogott: I can see the wikitech-ldap handler in the live logger using eval.php [19:53:47] I made a tiny tweak to the wikitech.php that maybe fixed something? [19:54:23] indeed it seems better [19:54:30] want to make a patch, or paste your change so I can? [19:54:34] ok.. I'll put up a patch for the config change [20:39:37] twentyafterfour: i think this ticket is resolved, isnt it? https://phabricator.wikimedia.org/T151544 [20:39:55] i mean .. you did restrict the size .. so done [21:10:32] 06Release-Engineering-Team (Deployment-Blockers), 05Release: MW-1.29.0-wmf.15 deployment blockers - https://phabricator.wikimedia.org/T158996#3053926 (10greg) [21:10:52] 06Release-Engineering-Team (Deployment-Blockers), 05Release: MW-1.29.0-wmf.14 deployment blockers - https://phabricator.wikimedia.org/T157639#3011581 (10greg) [21:11:42] 06Release-Engineering-Team (Deployment-Blockers), 05Release: MW-1.29.0-wmf.16 deployment blockers - https://phabricator.wikimedia.org/T158997#3053950 (10greg) [21:11:57] 06Release-Engineering-Team (Deployment-Blockers), 05Release: MW-1.29.0-wmf.15 deployment blockers - https://phabricator.wikimedia.org/T158996#3053926 (10greg) [21:40:22] 06Release-Engineering-Team (Deployment-Blockers), 05Release: MW-1.29.0-wmf.13 deployment blockers - https://phabricator.wikimedia.org/T157638#3054088 (10demon) 05Open>03Resolved [22:06:56] aude: how did that patch go [22:20:07] (03PS1) 10Urbanecm: Whitelist user Matěj Suchánek [integration/config] - 10https://gerrit.wikimedia.org/r/339788 [22:34:29] PROBLEM - Puppet run on deployment-tin is CRITICAL: CRITICAL: 55.56% of data above the critical threshold [0.0] [22:39:53] (03CR) 10Zppix: [C: 031] Whitelist user Matěj Suchánek [integration/config] - 10https://gerrit.wikimedia.org/r/339788 (owner: 10Urbanecm) [23:09:28] RECOVERY - Puppet run on deployment-tin is OK: OK: Less than 1.00% above the threshold [0.0]