[00:01:04] marx [00:01:20] where is that guy when I want to corner him? ;) [00:04:19] I was doing the same thing :) [00:23:18] 10Deployment-Systems, 10MediaWiki-Configuration: extension-list should live in the mediawiki branch rather than mediawiki-config - https://phabricator.wikimedia.org/T125678#1996447 (10Legoktm) We should just get rid of extension-list... [00:38:33] thcipriani: ghetto warmup script -- https://phabricator.wikimedia.org/P2560 [00:41:31] 10Deployment-Systems, 10MediaWiki-Configuration: extension-list should live in the mediawiki branch rather than mediawiki-config - https://phabricator.wikimedia.org/T125678#1996491 (10mmodell) @legoktm: I agree, it seems unnecessary? But update-l10n breaks without it, currently... I'll dig deeper into it. [00:43:14] !log Jenkins agent on deployment-bastion.eqiad doing the trick where it doesn't pick up jobs again [00:43:19] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [00:49:05] anybody remember a good way to break the goofy deadlock that makes deployment-bastion stop running Jenkins jobs? [00:50:11] * bd808 found https://phabricator.wikimedia.org/T72597 [00:54:01] 10Deployment-Systems, 10MediaWiki-Configuration: extension-list should live in the mediawiki branch rather than mediawiki-config - https://phabricator.wikimedia.org/T125678#1996528 (10mmodell) It looks like [[ /diffusion/MW/browse/master/maintenance/mergeMessageFileList.php;25dbd91513f1e55ee2ae9f01ef6abce64837... [00:54:25] 10Continuous-Integration-Infrastructure, 6Release-Engineering-Team, 7Jenkins, 7Upstream: [upstream] Jenkins Gearman plugin has deadlock on executor threads (was: Beta Cluster stopped receiving code updates (beta-update-databases-eqiad hung) - https://phabricator.wikimedia.org/T72597#1996532 (10bd808) Locke... [00:57:10] RECOVERY - Free space - all mounts on integration-slave-trusty-1016 is OK: OK: All targets OK [00:57:45] !log Got deployment-bastion processing Jenkins jobs again via instructions left by my past self at https://phabricator.wikimedia.org/T72597#747925 [00:57:50] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [01:20:59] 10Deployment-Systems, 10MediaWiki-Configuration: extension-list should live in the mediawiki branch rather than mediawiki-config - https://phabricator.wikimedia.org/T125678#1996626 (10mmodell) https://gerrit.wikimedia.org/r/#/c/268337/ [01:21:08] 10Deployment-Systems, 10MediaWiki-Configuration, 5Patch-For-Review: extension-list should live in the mediawiki branch rather than mediawiki-config - https://phabricator.wikimedia.org/T125678#1996627 (10mmodell) [02:57:19] Project beta-scap-eqiad build #88542: 04FAILURE in 10 min: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/88542/ [05:38:55] Project browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-os_x_10.9-chrome-sauce build #337: 04FAILURE in 22 min: https://integration.wikimedia.org/ci/job/browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-os_x_10.9-chrome-sauce/337/ [06:35:54] Can some one take a look at phab? [06:51:25] twentyafterfour: are you upgrading phab still? [07:41:10] 10Continuous-Integration-Config, 10Continuous-Integration-Infrastructure, 6Collaboration-Team-Backlog, 10Flow, and 2 others: Wikidata broken on branch REL1_25 causing other extensions to fail. - https://phabricator.wikimedia.org/T125722#1996660 (10Anomie) >>! In T125722#1995833, @Paladox wrote: > Scribunto... [07:41:14] 10Continuous-Integration-Config, 10Continuous-Integration-Infrastructure, 6Collaboration-Team-Backlog, 10Flow, and 2 others: Wikidata broken on branch REL1_25 causing other extensions to fail. - https://phabricator.wikimedia.org/T125722#1996662 (10Anomie) [07:56:55] 10Deployment-Systems, 10Fundraising-Backlog: make-wmf-branch doesn't work for named extension deployment branches - https://phabricator.wikimedia.org/T113428#1996992 (10JanZerebecki) Is there a requirement to always deploy on all wmf branches at the same time (or all wikis need to be the same CentralNotice ver... [07:59:25] 10Deployment-Systems: make-wmf-branch used master instead of the branch specified in special_extensions - https://phabricator.wikimedia.org/T125663#1996994 (10JanZerebecki) It seems ef2599a42526f2f3299bb05297a722f24b375af9 broke it. [08:04:14] 10Deployment-Systems: make-wmf-branch used master instead of the branch specified in special_extensions - https://phabricator.wikimedia.org/T125663#1996997 (10JanZerebecki) No, I misread that. Can't find any error in that commit. [08:09:59] (03CR) 10JanZerebecki: [C: 031] Make "recheck" magic word less restrictive for Code-Review votes [integration/config] - 10https://gerrit.wikimedia.org/r/267553 (owner: 10Florianschmidtwelzow) [08:13:53] Project browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-os_x_10.9-safari-sauce build #866: 04FAILURE in 3 min 53 sec: https://integration.wikimedia.org/ci/job/browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-os_x_10.9-safari-sauce/866/ [08:14:58] 10Deployment-Systems, 10Fundraising-Backlog: make-wmf-branch doesn't work for named extension deployment branches - https://phabricator.wikimedia.org/T113428#1997005 (10JanZerebecki) [08:15:01] 10Continuous-Integration-Config, 10Deployment-Systems, 10MediaWiki-extensions-CentralNotice, 10MediaWiki-extensions-Validator, and 2 others: [Task] make core wmf branches only use submodule branches that run with it in CI - https://phabricator.wikimedia.org/T113731#1997004 (10JanZerebecki) [08:15:23] 10Deployment-Systems, 10Fundraising-Backlog: CentralNotice: make-wmf-branch doesn't work for named extension deployment branches - https://phabricator.wikimedia.org/T113428#1997006 (10JanZerebecki) [08:19:42] (03CR) 10JanZerebecki: [C: 031] Bump tox minversion to 1.9 [integration/config] - 10https://gerrit.wikimedia.org/r/268207 (https://phabricator.wikimedia.org/T125705) (owner: 10Florianschmidtwelzow) [08:23:48] 10Beta-Cluster-Infrastructure, 6Services, 6operations: Move Node.JS services to Jessie and Node 4.2 - https://phabricator.wikimedia.org/T124989#1996707 (10mobrovac) [08:27:58] (03PS1) 10JanZerebecki: add repo wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268352 (https://phabricator.wikimedia.org/T111173) [08:36:01] 10Continuous-Integration-Config, 10Continuous-Integration-Infrastructure, 6Collaboration-Team-Backlog, 10Flow, and 2 others: Wikidata broken on branch REL1_25 causing other extensions to fail. - https://phabricator.wikimedia.org/T125722#1997021 (10Paladox) But why does it fail on REL1_25 branch but not on... [09:30:28] Yippee, build fixed! [09:30:28] Project browsertests-Gather-en.m.wikipedia.beta.wmflabs.org-linux-chrome-sauce build #408: 09FIXED in 11 min: https://integration.wikimedia.org/ci/job/browsertests-Gather-en.m.wikipedia.beta.wmflabs.org-linux-chrome-sauce/408/ [10:02:01] zeljkof: Gather browser tests work ! [10:02:34] !log https://integration.wikimedia.org/ci/view/Beta/job/beta-scap-eqiad/ is broken :( [10:03:12] !log beta-scap-eqiad fails with AttributeError: 'bool' object has no attribute 'encode' [10:03:20] I am sure I have seen that before [10:03:53] hashar: great news :) [10:04:23] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 3Scap3, 7WorkType-Maintenance: scap-rebuild-cdbs failed: 'bool' object has no attribute 'encode' - https://phabricator.wikimedia.org/T123366#1997168 (10hashar) 5Resolved>3Open That is happening again. Will nuke l10n cache and see wh... [10:04:47] !log beta: nuking the whole l10n cache ( https://phabricator.wikimedia.org/T123366 ) [10:08:41] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 3Scap3, 7WorkType-Maintenance: scap-rebuild-cdbs failed: 'bool' object has no attribute 'encode' - https://phabricator.wikimedia.org/T123366#1997172 (10hashar) 5Open>3 l10n cache nuked. New build going on at: https://integration.wiki... [11:11:25] Yippee, build fixed! [11:11:25] Project beta-scap-eqiad build #88588: 09FIXED in 16 min: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/88588/ [11:40:39] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 3Scap3, 7WorkType-Maintenance: scap-rebuild-cdbs failed: 'bool' object has no attribute 'encode' - https://phabricator.wikimedia.org/T123366#1997436 (10hashar) 5>3Resolved Fixed! https://integration.wikimedia.org/ci/job/beta-scap-eqi... [12:14:40] Yippee, build fixed! [12:14:41] Project browsertests-PdfHandler-test.wikipedia.org-linux-firefox-sauce build #17: 09FIXED in 40 sec: https://integration.wikimedia.org/ci/job/browsertests-PdfHandler-test.wikipedia.org-linux-firefox-sauce/17/ [12:30:46] 10Continuous-Integration-Infrastructure, 6Release-Engineering-Team, 7Jenkins, 7Upstream: [upstream] Jenkins Gearman plugin has deadlock on executor threads (was: Beta Cluster stopped receiving code updates (beta-update-databases-eqiad hung) - https://phabricator.wikimedia.org/T72597#1997547 (10hashar) [12:32:39] 10Continuous-Integration-Infrastructure, 6Release-Engineering-Team, 7Jenkins, 7Upstream: [upstream] Jenkins Gearman plugin has deadlock on executor threads (was: Beta Cluster stopped receiving code updates (beta-update-databases-eqiad hung) - https://phabricator.wikimedia.org/T72597#747903 (10hashar) Sorry... [12:46:12] Project browsertests-PdfHandler-test2.wikipedia.org-linux-firefox-sauce build #783: 04FAILURE in 12 sec: https://integration.wikimedia.org/ci/job/browsertests-PdfHandler-test2.wikipedia.org-linux-firefox-sauce/783/ [12:47:07] 10Beta-Cluster-Infrastructure, 6Labs: Disable /data/project for instances in deployment-prep that do not need it - https://phabricator.wikimedia.org/T125624#1997565 (10hashar) Last time I checked, a various backend services wrote their logs to /data/project so it can be read from anywhere. Though that might j... [13:34:49] (03CR) 10Hashar: [C: 032] "Danke Schön !" [integration/config] - 10https://gerrit.wikimedia.org/r/268207 (https://phabricator.wikimedia.org/T125705) (owner: 10Florianschmidtwelzow) [13:35:45] 10Continuous-Integration-Config, 5Patch-For-Review: Running tox on a new environment fails with "No such file or directory" - https://phabricator.wikimedia.org/T125705#1997656 (10hashar) a:3Florian Sorry for the delay yesterday, I was busy preparing deployment of a new MediaWiki version on the cluster. I am... [13:35:54] 10Continuous-Integration-Config, 5Patch-For-Review: Running tox on a new environment fails with "No such file or directory" - https://phabricator.wikimedia.org/T125705#1997658 (10hashar) 5Open>3Resolved [13:41:59] (03Merged) 10jenkins-bot: Bump tox minversion to 1.9 [integration/config] - 10https://gerrit.wikimedia.org/r/268207 (https://phabricator.wikimedia.org/T125705) (owner: 10Florianschmidtwelzow) [13:54:27] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10MediaWiki-extensions-Cards: Cards stuck on version 0.3.0 on Beta Cluster - https://phabricator.wikimedia.org/T125182#1997753 (10hashar) As I understand it, but @demon might correct me: We use a feature from Gerrit that let it magically update submodules... [13:55:24] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10Gerrit, 10MediaWiki-extensions-Cards: Gerrit does not update Cards in mediawiki/extensions (it is thus stuck on Beta Cluster) - https://phabricator.wikimedia.org/T125182#1997755 (10hashar) 5Resolved>3Open a:5thcipriani>3None [14:01:22] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10Gerrit, 10MediaWiki-extensions-Cards: Gerrit does not update Cards in mediawiki/extensions (it is thus stuck on Beta Cluster) - https://phabricator.wikimedia.org/T125182#1997787 (10hashar) Ah easy enough. There are two repositories named `Cards`: ```... [14:10:27] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10Gerrit, 10MediaWiki-extensions-Cards: Gerrit does not update Cards in mediawiki/extensions (it is thus stuck on Beta Cluster) - https://phabricator.wikimedia.org/T125182#1997832 (10hashar) > SELECT * FROM submodule_subscriptions WHERE submodule_project... [14:13:07] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10Gerrit, 10MediaWiki-extensions-Cards: Gerrit does not update Cards in mediawiki/extensions (it is thus stuck on Beta Cluster) - https://phabricator.wikimedia.org/T125182#1997838 (10hashar) For VisualEditor the breakage is: | submodule_project_name | su... [14:25:33] hashar, I tried updating submodule_subscriptions but still can't delete Cards.git [14:25:38] maybe it's cached somewhere [14:28:53] 5Continuous-Integration-Scaling, 6operations, 7HHVM: Provide a HHVM package for jessie-wikimedia matching version of trusty-wikimedia - https://phabricator.wikimedia.org/T125821#1997897 (10hashar) [14:29:27] 5Continuous-Integration-Scaling, 6operations, 7HHVM: Provide a HHVM package for jessie-wikimedia matching version of trusty-wikimedia - https://phabricator.wikimedia.org/T125821#1997890 (10hashar) [14:29:30] 10Continuous-Integration-Config, 5Continuous-Integration-Scaling, 3releng-201516-q3, 7WorkType-NewFunctionality: [keyresult] Migrate php composer (Zend and HHVM) CI jobs to Nodepool - https://phabricator.wikimedia.org/T119139#1997906 (10hashar) [14:30:47] 10Continuous-Integration-Config, 5Continuous-Integration-Scaling, 3releng-201516-q3, 7WorkType-NewFunctionality: [keyresult] Migrate php composer (Zend and HHVM) CI jobs to Nodepool - https://phabricator.wikimedia.org/T119139#1819085 (10hashar) We first need bits to be able to select which flavor of PHP (Z... [14:32:28] Krenair: I have no idea. And I am too afraid to tweak the database/delete project without Chad watching over my shoulders :D [14:32:53] nvm, think I got it [14:32:54] Krenair: I think he tried tweaking the database row manually for VisualEditor, but eventually Gerrit rewrote the entry [14:32:56] and broke it again [14:33:08] it is Gerrit being confused because it only use the base name apparently [14:33:22] and I have no idea what /Cards.git is for [14:33:35] nothing, anymore [14:33:38] it's gone [14:34:06] Project beta-scap-eqiad build #88604: 04FAILURE in 41 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/88604/ [14:36:06] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10Gerrit, 10MediaWiki-extensions-Cards: Gerrit does not update Cards in mediawiki/extensions (it is thus stuck on Beta Cluster) - https://phabricator.wikimedia.org/T125182#1997947 (10hashar) From https://www.mediawiki.org/wiki/Git/New_repositories/Reques... [14:38:10] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10Gerrit, 10MediaWiki-extensions-Cards: Gerrit does not update Cards in mediawiki/extensions (it is thus stuck on Beta Cluster) - https://phabricator.wikimedia.org/T125182#1997953 (10Krenair) I ran `update submodule_subscriptions set submodule_project_na... [14:40:23] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10Gerrit, 10MediaWiki-extensions-Cards: Gerrit does not update Cards in mediawiki/extensions (it is thus stuck on Beta Cluster) - https://phabricator.wikimedia.org/T125182#1997956 (10demon) >>! In T125182#1997953, @Krenair wrote: > I ran `update submodul... [14:41:45] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10Gerrit, 10MediaWiki-extensions-Cards: Gerrit does not update Cards in mediawiki/extensions (it is thus stuck on Beta Cluster) - https://phabricator.wikimedia.org/T125182#1997961 (10Krenair) >>! In T125182#1997956, @demon wrote: > You do realize that wo... [14:42:16] Krenair: No, updating the database. [14:42:41] I didn't just update the database [14:43:31] Text bleeding together. [14:43:37] Yippee, build fixed! [14:43:37] Project beta-scap-eqiad build #88605: 09FIXED in 7 min 37 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/88605/ [14:43:38] * ostriches wants to go back to bed instead [14:44:03] Krenair: "just need someone to delete it from git.wikimedia.org now" -- lol [14:44:08] "Need" is such a funny word [14:44:23] need may be the wrong word [14:46:48] in order to completely and utterly eliminate it from everywhere, git.wikimedia.org would still need to be dealt with [14:47:40] Meh, who cares if git.wm.o is crufty? [14:47:44] Lots of cruft on it already [14:48:11] well, your choice.. I don't care a huge deal about it either, just prefer to not leave such things lying about [14:48:26] Do what I do: pretend git.wm.o doesn't exist [14:48:27] :) [14:48:30] k [14:50:44] ostriches: good morning :-} [14:50:52] ostriches: thank you for all your help yesterday [14:50:52] ! [14:52:21] Morning! [14:52:25] And you're welcome :) [15:08:32] (03Abandoned) 10Hashar: Update template extension-gate [integration/config] - 10https://gerrit.wikimedia.org/r/267455 (owner: 10Paladox) [15:09:42] 10Beta-Cluster-Infrastructure, 6Services, 6operations: Move Node.JS services to Jessie and Node 4.2 - https://phabricator.wikimedia.org/T124989#1998036 (10Paladox) [15:09:58] (03Abandoned) 10Hashar: Revert "Revert all the php53/php55 recent additions" [integration/config] - 10https://gerrit.wikimedia.org/r/267911 (owner: 10Paladox) [15:10:00] 10Continuous-Integration-Config, 5Continuous-Integration-Scaling, 5Patch-For-Review, 7WorkType-NewFunctionality: Migrate javascript npm CI jobs to Nodepool - https://phabricator.wikimedia.org/T119143#1998038 (10Paladox) [15:10:35] (03PS2) 10Hashar: add repo wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268352 (https://phabricator.wikimedia.org/T111173) (owner: 10JanZerebecki) [15:11:51] (03CR) 10Hashar: [C: 032] "Straightforward! Feel free to self merge and deploy such trivial changes :-}" [integration/config] - 10https://gerrit.wikimedia.org/r/268352 (https://phabricator.wikimedia.org/T111173) (owner: 10JanZerebecki) [15:13:15] (03Merged) 10jenkins-bot: add repo wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268352 (https://phabricator.wikimedia.org/T111173) (owner: 10JanZerebecki) [15:56:56] hashar, twentyafterfour, I metioned you in https://phabricator.wikimedia.org/T125803 [15:57:00] maybe you can take a look [15:57:36] Luke081515: replied already! thank you to have taken the time to fill a bug hehe [15:57:36] I am out again [15:58:33] I'm not the author of this bug ;) [15:59:14] Krenair did [16:07:48] 10Beta-Cluster-Infrastructure, 10Deployment-Systems, 10Gerrit, 10MediaWiki-extensions-Cards, and 2 others: Gerrit does not update Cards in mediawiki/extensions (it is thus stuck on Beta Cluster) - https://phabricator.wikimedia.org/T125182#1998226 (10Krenair) 5Open>3Resolved a:3Krenair that did the trick [16:09:22] greg-g: will you be coming to the office? [16:12:23] nope [16:12:32] should I have? [16:13:36] greg-g: nah, just wondering if i can steal r32 from you [16:13:37] :) [16:16:16] ah, yeah, sorry [16:17:18] greg-g: stolen! :P [16:22:36] greg-g: euh, actually, mind removing it form your cal? it has just declined my double-book [16:22:52] "a room declined my invitation" [16:22:57] heh, yeah, one sec [16:22:59] i never thought i would say this [16:23:12] done [16:23:16] cheers! [16:29:53] 6Release-Engineering-Team, 5WMF-deploy-2016-02-02_(1.27.0-wmf.12): MW 1.27.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T125475#1998307 (10greg) [16:29:55] 10Deployment-Systems, 6Release-Engineering-Team, 6operations: /srv/mediawiki-staging broken on both scap masters - https://phabricator.wikimedia.org/T125506#1998305 (10greg) 5Open>3Resolved >>! In T125506#1994318, @greg wrote: > Keeping open until the report is posted, but for the purposes of pushing out... [16:42:41] Morning, zeljkof! I decided to spend the day on UploadWizard browser tests. I'm sorry in advance for bothering you :) [16:42:49] But not too far in advance. [16:43:24] 6Release-Engineering-Team: Blockers to deploying 1.27-wmf.11 - https://phabricator.wikimedia.org/T125143#1998341 (10Anomie) [16:44:52] zeljkof: I'm getting an error from phantomjs when I click a in an , that says "Click succeeded but Load Failed. Status: 'fail'" - I found a few bug reports that suggested phantomjs 1.9 or 1.92 would fix it, but 2.1.1 told me that it was "Unable to set cookie" which seems more problematic [16:47:58] The test passes in Firefox, so it *is* a phantomjs bug [16:48:19] Maybe 1.9.2 is still downloadable somewhere [16:50:13] Turns out 1.9.8 is [16:55:23] twentyafterfour: thanks for the quick fix! [16:55:59] cwd: no problem. I should have noticed last night and fixed it then. [16:57:25] 6Release-Engineering-Team: Blockers to deploying 1.27-wmf.11 - https://phabricator.wikimedia.org/T125143#1998405 (10Anomie) [16:58:32] OK, never mind, 1.9.8 works OK [17:00:05] MarkTraceur: sorry, I was in a meeting with greg-g [17:00:31] let me know if you have any questions, but it is 6pm here, I will probably go away soon [17:00:52] Dan (marxarelli) will also know, and Antoine (hashar) [17:01:20] are we still using phantomjs? I thought we have deprecated it [17:03:56] No problem [17:04:16] zeljkof: I didn't realize, I guess I'll start running them in firefox [17:04:25] Though I like not having to spawn a Firefox process to run tests... [17:05:21] MarkTraceur: why do you care? you start phantomjs to run the tests :) [17:05:27] are you on linux? [17:05:45] in that case you can run firefox headless, it will look like phantomjs :) [17:05:52] Ooh, neat. [17:05:54] zeljkof: How do? [17:07:17] let me check, it is documented [17:08:55] MarkTraceur: https://github.com/wikimedia/mediawiki-selenium/#headless-mode [17:09:12] HEADLESS=true bundle exec cucumber [17:09:23] that is all you need to do [17:09:43] you probably need to install xvfb too [17:09:55] have to go, send mail to qa mailing list if you have questions, or ping here, I will reply when I get back [17:10:33] KK, cheers [17:19:18] twentyafterfour: I created a stub incident report for phab, with timeline mostly there already [17:19:21] twentyafterfour: https://wikitech.wikimedia.org/wiki/Incident_documentation/20160204-Phabricator [17:19:41] greg-g: oh, I was doing the same [17:20:25] greg-g: thanks, I'll merge my summary into your timeline [17:20:53] 10Deployment-Systems, 3Scap3: refreshCdbJsonFiles should be rewritten in python - https://phabricator.wikimedia.org/T125685#1998544 (10EBernhardson) I don't know how relevant it is, but there is a longer term plan to use plain php files intead of cdb files (T99740). This use case has been specifically optimize... [17:22:09] 10Continuous-Integration-Infrastructure, 6Release-Engineering-Team, 7Jenkins, 7Upstream: [upstream] Jenkins Gearman plugin has deadlock on executor threads (was: Beta Cluster stopped receiving code updates (beta-update-databases-eqiad hung) - https://phabricator.wikimedia.org/T72597#1998549 (10bd808) >>! I... [17:25:30] twentyafterfour: sweet, wasn't sure, just trying to help :) [17:34:13] 10Deployment-Systems, 3Scap3: refreshCdbJsonFiles should be rewritten in python - https://phabricator.wikimedia.org/T125685#1998616 (10demon) It means we wouldn't //have// to rewrite it anymore, but we still //should// regardless. [17:35:04] 10Deployment-Systems, 3Scap3: refreshCdbJsonFiles should be rewritten in python - https://phabricator.wikimedia.org/T125685#1998628 (10bd808) Using PHP files for l10n cache is blocked on {T103886} which is in turn blocked on {T119637}. The refreshCdbJsonFiles script is fairly simple at its heart. It generates... [17:42:05] bd808: Thanks for the extra context too [17:42:12] (all in my head, but I didn't have the links :)) [17:42:33] I've become quite the phab and wiki gnome lately [17:43:05] * bd808 fixed all the broken tags on wikitech last night for "fun" [17:46:21] * greg-g chants "he's one of us, he's one of us" [17:46:33] gooble gobble [17:50:08] Have I mentioned recently how much I hate page-object [17:50:38] Hey marxarelli want to help a Ruby Nuby overcome his fear of pseudo-declarative language concepts? :D [17:50:57] MarkTraceur: weee! [17:51:06] you're not alone in your hate for page-object [17:51:08] or cucumber [17:51:09] or ... [17:51:22] I keep failing to find a friggin' text field [17:51:51] MarkTraceur: context? [17:51:55] The query is like browser.div(id: 'mwe-upwiz-detailsform0').div(index: 1).text_field [17:52:01] marxarelli: UploadWizard, details page [17:52:21] Ooh, maybe it should be page.div instead of browser.div [17:52:46] MarkTraceur: so, browser is a Watir::Browser [17:52:56] (another lovely dependency we're trying to get rid of) [17:52:59] Yeah [17:53:05] but yeah, it's not a page object [17:53:10] I had a good reason for using it before, I think, but I doubt I do now [17:57:32] MarkTraceur: you should still be able to find the element that way, but i recommend you put it in a page object [17:57:42] also, it the form dynamically loaded? [17:58:04] Yeah, it's all JavaScript [17:58:07] Trying something clever now [17:58:22] MarkTraceur: ok, you'll probably need a when_present in there then [17:59:14] It's in there later [17:59:29] since you're dealing with webdriver, the browser state isn't available to the client upon dom changes, so it needs to poll [18:00:31] Oh, christ, this is a class or something [18:00:40] I'm just totally incompetent, hold please [18:00:48] haha [18:01:04] talk to the duck and get back [18:01:49] (03PS3) 10Florianschmidtwelzow: Make "recheck" magic word less restrictive for Code-Review votes [integration/config] - 10https://gerrit.wikimedia.org/r/267553 [18:05:10] Project beta-scap-eqiad build #88622: 04FAILURE in 1 min 6 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/88622/ [18:06:12] marxarelli: OK, am I high? [18:06:15] Seriously [18:06:37] *** You are calling a method named text_field at /home/marktraceur/projects/mediawiki/extensions/UploadWizard/tests/browser/features/support/pages/describe_page.rb:39:in `field'. [18:06:37] *** This method does not exist in page-object so it is being passed to the driver. [18:06:46] https://github.com/cheezy/page-object/wiki/Elements#textfield [18:07:16] wah wah [18:07:39] what's the entire expression? [18:07:42] Like clearly, the method does exist, and is documented [18:08:04] self.next_parent_element.div_element(id: 'mwe-upwiz-detailsform' + index.to_s).div_element(index: 1).text_field [18:09:01] index is 0 in this case (because UW is 0-indexed even if Ruby is not [18:09:31] ruby is also 0 indexed [18:09:40] arrays in ruby are at least [18:10:11] I thought someone told me to use 1 when I passed indices to page-object or Watir or something [18:10:14] 3Scap3, 6Phabricator, 7WorkType-Maintenance: Refactor phabricator module in puppet to remove git tag pinning behavior - https://phabricator.wikimedia.org/T125851#1998830 (10mmodell) 3NEW a:3mmodell [18:10:32] So I have to fix that [18:10:52] well, i believe it uses nth-child under the hood [18:11:10] and nth-child(1) would be the first child node [18:11:32] it's a lot of unnecessary abstraction [18:11:40] i would just use css selectors to simplify it [18:12:07] Hnnnggg [18:12:14] so '#mwe-upwiz-detailsform div:nth-child(1) input[type=text]' or whatever [18:12:18] I was trying to be good and traverse the dom properly [18:12:38] trust me, it's not traversing the dom properly :)( [18:12:47] that's not how webdriver works [18:13:03] 3Scap3, 6Phabricator, 7WorkType-Maintenance: Move /srv/phab/repos to /srv/repos - https://phabricator.wikimedia.org/T125853#1998871 (10mmodell) 3NEW a:3mmodell [18:13:07] Which brings me back to I really, really hate browser tests [18:13:14] it uses selectors to find elements, so you might as well not fake the funk and use selectors to define the page elements [18:13:15] 3Scap3, 6Phabricator, 7WorkType-Maintenance: Refactor phabricator module in puppet to remove git tag pinning behavior - https://phabricator.wikimedia.org/T125851#1998830 (10mmodell) [18:13:30] marxarelli: The thing is, I have these specified dynamically [18:13:34] you _can_ define nested elements, but even then it's not really traversing [18:13:40] So I have a helper method that gets the nth upload's div [18:13:51] And then a helper method that gets the nth upload's, say, title field [18:14:38] that's ok, you can define your own methods in the page object [18:14:39] But I guess it should still work [18:15:49] I still don't understand why text_field, which is documented as existing, throws a deprecation warning [18:16:04] so from within a page object method, you can do `browser.element(css: 'some selector')` [18:16:05] Like, I suspect it has something to do with why my code won't work [18:16:09] Wai wai wai [18:16:15] I thought browser was Bad News™ [18:16:24] Yippee, build fixed! [18:16:25] Project beta-scap-eqiad build #88623: 09FIXED in 8 min 59 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/88623/ [18:16:25] not from within a page object :) [18:16:32] * MarkTraceur weeps [18:16:44] That's actually how it was defined before [18:17:23] it's not categorically bad news really, from anywhere. the gist of it is that you generally want to encapsulate your element selection within the page object somehow [18:17:27] Trying that now [18:17:39] OK, we have liftoff, thanks [18:17:47] Moving everything to use that now [18:17:55] w00t. np! [18:18:09] btw, you don't need to say `self.foo`, you can just say `foo` [18:18:23] TIL [18:18:27] self is implicit in ruby because you're always in the context of some object [18:18:49] And that brings me back to, I really hate ruby [18:19:11] But it will feel better when the browser tests don't email me every day. [18:19:20] and just about every expression is sending a message to an object, self being the default receiver [18:19:41] haha, you're getting a taste of it in the worst context possible [18:19:49] cucumber + page-object + watir [18:20:20] it's really a pretty simple language if someone isn't shoving a DSL down your throat [18:23:30] 5Gerrit-Migration, 6Phabricator, 6Repository-Admins, 10pywikibot-core: Migrate Pywikibot to Differential code review - https://phabricator.wikimedia.org/T95526#1998987 (10Luke081515) Guess in this case we will host the repo at phab instead of gerrit primary? [18:23:53] 3Scap3, 6Phabricator, 7WorkType-Maintenance: Refactor phabricator module in puppet to remove git tag pinning behavior - https://phabricator.wikimedia.org/T125851#1999001 (10mmodell) [18:23:56] 10Deployment-Systems, 3Scap3, 6Phabricator, 7WorkType-NewFunctionality: Deploy Phabricator with scap3 - https://phabricator.wikimedia.org/T114363#1998999 (10mmodell) [18:24:14] MarkTraceur: also, we're experimenting with various JS end-to-end testing libraries if you want to be an early adopter :) [18:24:21] 10Deployment-Systems, 3Scap3, 6Phabricator, 7WorkType-NewFunctionality: Deploy Phabricator with scap3 - https://phabricator.wikimedia.org/T114363#1693175 (10mmodell) Added subtasks related to https://wikitech.wikimedia.org/wiki/Incident_documentation/20160204-Phabricator [18:24:36] marxarelli: Hm, that sounds like it would mean I'd need to spend even more time on browser testing. [18:25:11] 10Deployment-Systems, 3Scap3, 6Phabricator, 7WorkType-NewFunctionality: Deploy Phabricator with scap3 - https://phabricator.wikimedia.org/T114363#1999008 (10mmodell) [18:25:25] MarkTraceur: inherently, yes. :) but hopefully less time banging your head against bad API documentation? [18:25:26] 10Deployment-Systems, 3Scap3, 6Phabricator, 7WorkType-NewFunctionality: Deploy Phabricator with scap3 - https://phabricator.wikimedia.org/T114363#1693175 (10mmodell) [18:25:40] and a language you don't like [18:25:40] That sounds optimistic and useful [18:25:44] True dat [18:25:48] marxarelli: Sign me up [18:27:02] (03CR) 10JanZerebecki: [C: 031] Set up php55lint jobs, adapt phplint macro for $PHP_BIN (re-do) [integration/config] - 10https://gerrit.wikimedia.org/r/268047 (owner: 10Legoktm) [18:27:09] ya hear that, zeljkof? we have a guinea pig! i mean ... MarkTraceur has graciously volunteered for most extreme ultimate happiness [18:28:06] marxarelli: Is this the sort of thing that could happen soon enough that I don't need to waste my time fixing these broken ass-tests? [18:30:12] Because as much as that last fix helped, it seemed to also break things a lot. [18:31:16] MarkTraceur: so, yes, and no. we could probably get an MVP for the library done in a few weeks, but we still need CI infra [18:31:41] it'd might be more worthwhile for you to pair on it with us, however [18:31:51] Hmm [18:32:12] (03PS1) 10Jforrester: [Citoid] Depend on the Cite extension too [integration/config] - 10https://gerrit.wikimedia.org/r/268435 [18:32:17] (03PS1) 10Aude: Run npm test for wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) [18:32:46] Hey, could someone push https://gerrit.wikimedia.org/r/268435 into CI for me quickly? It's blocking a hairy multi-repo merge. :_) [18:33:23] (03CR) 10Dduvall: [C: 032] [Citoid] Depend on the Cite extension too [integration/config] - 10https://gerrit.wikimedia.org/r/268435 (owner: 10Jforrester) [18:33:25] (03CR) 10Aude: "not sure this is correct..." [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [18:33:58] marxarelli: So I'm sold on using it, but I'll need to know more about the plans if I'm going to spend a significant amount of time and effort on it [18:34:13] (03CR) 10Aude: "also would like to run phplint, though us moving away from php 5.3, not sure what to do here now to enable that" [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [18:34:22] Thanks marxarelli. :-) [18:34:47] MarkTraceur: kk. i'll ping zeljkof about it and we can figure out a plan [18:34:57] James_F: np! [18:35:23] (03CR) 10jenkins-bot: [V: 04-1] Run npm test for wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [18:35:36] (03Merged) 10jenkins-bot: [Citoid] Depend on the Cite extension too [integration/config] - 10https://gerrit.wikimedia.org/r/268435 (owner: 10Jforrester) [18:35:46] MarkTraceur: i think we'll know which libraries to choose by next week [18:35:55] been trying out a few [18:37:28] !log Reloading Zuul to deploy Iccf4f48fe5bf964a4c4e6db3f404f152628a4a24 [18:39:57] MarkTraceur: fwiw, i don't think we'll need a considerable amount of your time, i just want to make sure we'll establishing something that's far less painful to use. end-to-end testing is hard enough without confusing APIs/DSLs getting you down [18:40:16] s/we'll/we're/ [18:40:24] OK [18:44:02] marxarelli: In that case sign me up again! [18:44:09] 100% on board with not using Ruby [18:47:43] 6Release-Engineering-Team, 10MediaWiki-Unit-tests, 10VisualEditor, 5WMF-deploy-2016-02-02_(1.27.0-wmf.12): VE QUnit tests fail on 1.27.0-wmf.12 branch due to Wikidata branch cut issue - https://phabricator.wikimedia.org/T125655#1999138 (10Jdforrester-WMF) [18:50:12] I guess I will stop trying to rescue the sinking ship that is the UW browsertests. Sigh. Wasted four hours. [18:50:23] Wonder what I'm going to waste the next 5 on [18:51:07] Tales of Wage Slavery with MarkTraceur [18:51:23] Heh. [18:52:37] sorry, was reading an article about The Protestant Ethic this morning :) [18:54:07] so, greg-g, i noticed i'm back on the train today. what's the scoop on group0? testwiki seems responsive at least [18:54:28] bah, one sec [18:55:05] marxarelli: Hmm. If I make VE depend on Cite will anything break? [18:55:37] James_F: i believe it ext_dependencies handles circular deps [18:55:40] i _believe_ [18:55:42] :) [18:55:58] Ooookay. [18:56:00] Let's give it a go. [18:56:03] marxarelli: https://wikitech.wikimedia.org/w/index.php?title=Deployments&type=revision&diff=287274&oldid=287053 [18:56:09] let me use my epicurean mind, hold on [18:56:45] James_F: hmm, maybe not [18:56:45] (03PS1) 10Jforrester: [VisualEditor] Depend on the Cite extension [integration/config] - 10https://gerrit.wikimedia.org/r/268440 [18:56:49] `TODO: Max recursion limit?` [18:56:51] :) [18:56:59] heh, that sounds ominous [18:57:04] OK, let's not. [18:57:14] * James_F sighs. [18:57:16] All these unit tests in VE. [18:57:35] greg-g: oh ha, sorry. i had the page up from yesterday [18:58:06] marxarelli: check that timestamp ;) (I just fixed it) [18:58:19] All I want is "when testing/merging patches in VE, use master of Cite; when testing/merging patches in Cite, use master of VE". [18:58:39] greg-g: :) [18:59:23] James_F: give me a minute and i'll implement the todo :) [18:59:40] marxarelli: Awesome. Thanks. [19:00:15] James_F: and by a minute i mean more like 20, cause it will need a test [19:00:29] :-) [19:00:39] psha [19:01:05] i know, 'nerd alert' [19:02:31] (03PS2) 10Aude: Run npm test for wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) [19:05:08] 10Beta-Cluster-Infrastructure: Setup a Swift cluster on beta-cluster to match production - https://phabricator.wikimedia.org/T64835#1999232 (10Gilles) [19:08:44] so, the train is not for another hour? [19:12:46] * aude back in ~an hour, just in (unlikely) case htere is a problem with wikibase [19:17:23] aude: thanks, and yeah, since antoine's doing it [19:18:49] ok, I have an annoying headache, I'm going to take a shower and be back in a little bit, hopefully in time for the deploy as well, but maybe not [19:19:03] #workingfromhomewins [19:23:19] marxarelli/greg-g: Can I get an emergency config change? I… umm, oopsed. [19:23:48] ostriches: ^ [19:25:23] James_F: Link? [19:26:16] https://gerrit.wikimedia.org/r/268446 [19:26:32] Sending enwiki etc. feedback to a MW.org Flow page isn't ideal in community relations. [19:27:48] doing [19:28:06] Thanks, ostriches. [19:29:29] good morning [19:29:57] kids are asleep, so going to prep the group 2 switch [19:30:18] * greg-g waves to hashar [19:30:25] 10Beta-Cluster-Infrastructure, 6Labs: Disable /data/project for instances in deployment-prep that do not need it - https://phabricator.wikimedia.org/T125624#1999364 (10yuvipanda) If anything is writing logs to NFS we must make sure it stops as soon as possible - I'll check with the parsoid people :) ALL insta... [19:33:53] ok, going afk for a short bit for real now [19:37:56] 10Beta-Cluster-Infrastructure, 6Labs: Disable /data/project for instances in deployment-prep that do not need it - https://phabricator.wikimedia.org/T125624#1999407 (10hashar) [19:39:35] 10Beta-Cluster-Infrastructure, 6Labs: Disable /data/project for instances in deployment-prep that do not need it - https://phabricator.wikimedia.org/T125624#1992944 (10hashar) Listed the NFS types via salt/df/magic: ``` root@deployment-salt:~ # salt -v --out=txt '*' cmd.run "df -t nfs 2>/dev/null|grep -v ^File... [19:40:40] (03PS1) 10Dduvall: Handle cyclical extension dependencies [integration/config] - 10https://gerrit.wikimedia.org/r/268451 [19:42:28] (03CR) 10jenkins-bot: [V: 04-1] Handle cyclical extension dependencies [integration/config] - 10https://gerrit.wikimedia.org/r/268451 (owner: 10Dduvall) [19:45:37] (03PS2) 10Dduvall: Handle cyclical extension dependencies [integration/config] - 10https://gerrit.wikimedia.org/r/268451 [19:45:46] ah gerrit [19:46:33] ostriches: is Gerrit http authentication using labs LDAP password? [19:47:09] hashar: it's a separate gerrit password [19:47:20] hashar: no, your http password needs to be generated in your settings [19:47:24] ... [19:47:32] hashar: https://gerrit.wikimedia.org/r/#/settings/http-password [19:47:33] hashar: go to settings -> HTTP Password [19:47:46] I want a title change [19:47:49] instead of "Software Engineer" [19:47:55] name "Software lame end user" [19:48:01] fixe [19:48:06] thank you! [19:48:08] :) [19:48:13] Phab does the same thing over https. [19:48:23] I sorta see why, but it's totally non-intuitive. [19:48:27] after years on Gerrit, I have discovered we can write/push over https [19:48:33] why keeping them separate? [19:48:57] greg-g: Assuming people don't have https, and are using plain http auth. [19:49:04] (PS1) Hashar: all wikis to 1.27.0-wmf.12 [mediawiki-config] - https://gerrit.wikimedia.org/r/268452 [19:49:05] !! [19:49:21] So in an http world you'd probably reset the vcs password pretty regularly if you're using untrusted networks [19:49:53] ostriches: good point, I guess that needs to be thought of (non-ssl) [19:50:16] and of course I broke deploy-promote :-} [19:55:46] hashar: broke it? [19:55:52] well [19:56:17] as usual, script doesn't deal with errors / CTR+C in middle of flow ;-D [19:56:23] na kidding really [19:56:36] i failed to authenticate on the first step [19:56:36] haha, yeah, doesn't handle that too well :) [19:56:48] then wondered whether the script ended up letting the staging area in some kind of half baked step [19:57:06] and I could nitpick about how: [19:57:09] function foo() { [19:57:11] cd blalal [19:57:17] is usually a terrible idea :-} [19:57:33] but yeah script does the job, should barely fail and is a nice addition. Well done! [19:57:54] pf [19:58:08] that script was mostly a twentyafterfour joint. I just made it needlessly complicated :D [19:58:23] I think I will have a policy for cherry pick to wmf/1.27.0-wmf.12 to clearly state why the cherry pick is proposed, what are the impacts, how it got tested and when it can land [19:58:46] thcipriani: na it is fine. I am just teasing you because it is written in shell :-} [19:58:50] it is all fine really [19:58:55] make the promote WAY easier [19:59:07] I think next time I will have my kid to handle the train [19:59:16] you guys streamline the whole process [19:59:22] streamlined [19:59:25] I mean [19:59:31] the more people that know how to do it, the better :P [19:59:33] we could even wrap it in a jenkins job [20:00:02] (03PS3) 10Dduvall: Handle cyclical extension dependencies [integration/config] - 10https://gerrit.wikimedia.org/r/268451 [20:00:15] * greg-g smiles [20:01:55] https://gerrit.wikimedia.org/r/#/c/268455/ <-- might need a CR+2 from #together team [20:02:18] * thcipriani looks [20:02:26] promote all [20:02:34] no occurrences of .10 nor .11 [20:02:46] that should properly kill the wikis [20:02:46] s/kill/update/ [20:02:54] danke [20:06:15] hashar: congrats on the train :) [20:06:45] :) [20:07:14] well [20:07:18] I am disappointed. No fatals [20:07:19] :( [20:08:03] there's always next time. [20:08:51] !log All wikis to 1.27.0-wmf.12 No troubles so far congratulations to everyone involved @wikimedia #wikimedia [20:10:35] (03CR) 10Jforrester: "Well, it seems sane to me…" [integration/config] - 10https://gerrit.wikimedia.org/r/268451 (owner: 10Dduvall) [20:11:31] !log ping [20:11:39] poor morebot [20:16:50] Woo everyone. [20:17:07] * James_F ceremonially marks wmf.10 as Archived in phab. [20:18:04] James_F: out with a wimper or a fight? [20:18:48] greg-g: Well, given that Kunal had to unmark it as Archived… a deathknell. ;-) [20:19:14] * greg-g grumbles [20:19:21] hashar: Please review https://gerrit.wikimedia.org/r/#/c/268306/ it is for wikidata on REL1_25 branch. [20:27:53] 10Deployment-Systems, 6Performance-Team, 10Traffic, 6operations, 5Patch-For-Review: Make Varnish cache for /static/$wmfbranch/ expire when resources change within branch lifetime - https://phabricator.wikimedia.org/T99096#1999652 (10Krinkle) [20:29:09] 6Release-Engineering-Team, 5WMF-deploy-2016-02-02_(1.27.0-wmf.12): MW 1.27.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T125475#1999654 (10greg) [20:34:28] 10Continuous-Integration-Infrastructure: MediaWiki core php test coverage not found - https://phabricator.wikimedia.org/T125876#1999669 (10Florian) 3NEW [20:38:08] 10Deployment-Systems, 3Scap3, 5Patch-For-Review: Make puppet provider for scap3 - https://phabricator.wikimedia.org/T113072#1999681 (10mmodell) [20:38:10] 3Scap3, 6Phabricator, 7WorkType-Maintenance: Refactor phabricator module in puppet to remove git tag pinning behavior - https://phabricator.wikimedia.org/T125851#1999680 (10mmodell) [20:39:35] paladox: it in my backlog yeah https://gerrit.wikimedia.org/r/#/q/is:open+reviewer:hashar+label:code-review%253D0%252Chashar+NOT+owner:hashar,n,z [20:40:05] which is roughly 85 patches to review [20:45:31] (03CR) 10JanZerebecki: "For phplint one can probably use jakub-onderka/php-parallel-lint from composer test with the added benefit of it running composer update. " (032 comments) [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [20:46:20] 10Continuous-Integration-Config, 10Continuous-Integration-Infrastructure, 6Collaboration-Team-Backlog, 10Flow, and 2 others: Wikidata broken on branch REL1_25 causing other extensions to fail. - https://phabricator.wikimedia.org/T125722#1999759 (10Mattflaschen) >>! In T125722#1997021, @Paladox wrote: > But... [20:47:51] (03PS3) 10Aude: Run npm test for wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) [20:49:50] 10Continuous-Integration-Infrastructure: MediaWiki core php test coverage not found - https://phabricator.wikimedia.org/T125876#1999771 (10hashar) The coverage is generated by https://integration.wikimedia.org/ci/job/mediawiki-core-code-coverage/ which in turns trigger the job publish-on-gallium. It is supposed... [20:50:45] 10Continuous-Integration-Config: MediaWiki test coverage kills Zend 5.3 engine (ex: zend_mm_heap corrupted), switch it to other instance - https://phabricator.wikimedia.org/T125876#1999780 (10hashar) [20:51:30] (03CR) 10Aude: "probably running composer would be good" [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [20:55:01] (03CR) 10JanZerebecki: [C: 031] Run npm test for wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [20:56:42] hashar: Do you know of a way to disable jsduck testing on the REL1_25 branch on Flow or would we have to disable it for all REL1_25 branches. [20:56:54] (03CR) 10JanZerebecki: "Yes the composer-test template will do just that." [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [20:57:48] (03PS4) 10JanZerebecki: Make "recheck" magic word less restrictive for Code-Review votes [integration/config] - 10https://gerrit.wikimedia.org/r/267553 (owner: 10Florianschmidtwelzow) [20:58:25] (03CR) 10JanZerebecki: [C: 031] [VisualEditor] Depend on the Cite extension [integration/config] - 10https://gerrit.wikimedia.org/r/268440 (owner: 10Jforrester) [20:59:17] paladox: depends [20:59:36] paladox: if is a generic job applied on various repo, we can't filter it out [20:59:38] (03PS2) 10Jforrester: [VisualEditor] Depend on the Cite extension [integration/config] - 10https://gerrit.wikimedia.org/r/268440 [20:59:51] paladox: why do you care about Flow on REL1_25 anyway? I dont think it receive any patch [21:00:16] hashar: Mattflaschen uploaded some patches to that branch and asked why the tests are failing. [21:00:51] hashar: Could we make it non generic so we can filter please for flow. [21:00:53] https://gerrit.wikimedia.org/r/#/q/is:open+branch:REL1_25,n,z [21:00:53] bah [21:01:19] ok. [21:01:40] 00:00:03.592 Error: You should specify some input files, otherwise there's nothing I can do :( [21:01:49] I am wondering why he wonders it does not work :-} [21:02:02] Maybe we could backport from flow. [21:02:16] rake-jessie dies as well [21:02:46] we would need to create more specific jobs that are prefixed with the repo name and then we could set branch filter [21:02:49] i.e. we would be able to do something like: [21:02:56] - job: [21:03:10] name: flow-jsduck [21:03:20] branch: !REL1_25 [21:03:27] syntax is off [21:04:07] and of course mediawiki-extensions-qunit is not going to work :D [21:06:18] hashar: Thanks. Ok i will try backporting first. [21:06:47] hashar: Since done here https://gerrit.wikimedia.org/r/#/c/268484/ [21:07:29] oh [21:07:51] so you back port stuff in less than 30 seconds :-} nice! [21:08:46] paladox: what we would really want is to get rid of the jsduck job [21:08:54] and have them run using bundler/rake [21:09:13] hashar: Oh. So i am better off just disabling it for that branch. [21:09:21] yeah but you can't really [21:09:30] cause the job 'jsduck' is generic and used on a wild range of repos [21:10:52] hashar: Ok i am going to create a non generic one. [21:11:08] hashar: Question what does this code do for core [21:11:10] - name: mediawiki-core-jsduck [21:11:10] branch: ^(REL1_23|REL1_24|master)$ [21:11:41] since it shows REL1_23 and 24 branch plus master and no 25 and 26 branch. [21:12:37] (03Merged) 10jenkins-bot: Make "recheck" magic word less restrictive for Code-Review votes [integration/config] - 10https://gerrit.wikimedia.org/r/267553 (owner: 10Florianschmidtwelzow) [21:12:42] 10Deployment-Systems, 6Performance-Team, 10Traffic, 6operations, 5Patch-For-Review: Make Varnish cache for /static/$wmfbranch/ expire when resources change within branch lifetime - https://phabricator.wikimedia.org/T99096#1999935 (10Krinkle) Deployment strategy: 1. [x] [mediawiki/core] Change MediaWiki... [21:20:55] 10Continuous-Integration-Config, 10Continuous-Integration-Infrastructure, 6Collaboration-Team-Backlog, 10Flow, and 2 others: Wikidata broken on branch REL1_25 causing other extensions to fail. - https://phabricator.wikimedia.org/T125722#1999988 (10Paladox) Thanks. [21:22:12] (03PS4) 10JanZerebecki: Handle cyclical extension dependencies [integration/config] - 10https://gerrit.wikimedia.org/r/268451 (owner: 10Dduvall) [21:22:37] (03CR) 10JanZerebecki: [C: 032] Handle cyclical extension dependencies [integration/config] - 10https://gerrit.wikimedia.org/r/268451 (owner: 10Dduvall) [21:25:11] 3Scap3, 6Phabricator, 7WorkType-Maintenance: Move /srv/phab/repos to /srv/repos - https://phabricator.wikimedia.org/T125853#2000011 (10mmodell) [21:26:18] paladox: sorry. The 'branch' for mediawiki-core-jsduck is to only run that job on the branch mentionned there i.e. REL1_23 REL1_24 master [21:26:50] paladox: looks like we forgot to update it for 1_25/1_26 or maybe it is broken on those branches [21:26:55] hashar: Oh. [21:27:21] paladox: another thing is that mediawiki core takes a while to clone, so we have more specific jobs and have the workspace kept between build so it does not have to reclone it entirely [21:27:35] hashar: I belive it would work for REL1_25 and REL1_26 since they would be off the master branch. [21:27:48] (03PS1) 10Paladox: Blacklist REL1_25 branch for jsduck and rake-jessie [integration/config] - 10https://gerrit.wikimedia.org/r/268557 [21:27:49] but really all that jsduck stuff should be migrate to an entry in Rakefile and thus run via rake-jessie [21:28:05] paladox: yeah that is a good assumption [21:28:09] maybe git blame the line [21:28:30] (03CR) 10jenkins-bot: [V: 04-1] Blacklist REL1_25 branch for jsduck and rake-jessie [integration/config] - 10https://gerrit.wikimedia.org/r/268557 (owner: 10Paladox) [21:28:42] hashar: Yes. Is there any easy migration because i could do it if it is just a couple of lines added to the Rake file. [21:29:10] paladox: i think we just forgot to update that branch filter [21:29:17] paladox: we can probably remove it entirely [21:29:36] been running on master for ages, so it should pass on REL1_25 and REL1_26 [21:29:58] hashar: I will upload a patch that removes the branch filter for that test in a min. [21:30:04] yeah [21:30:16] might want to run jsduck on each release branch to make sure it pass [21:30:19] will do tomorrow [21:30:21] (03PS2) 10Paladox: Blacklist REL1_25 branch for jsduck and rake-jessie [integration/config] - 10https://gerrit.wikimedia.org/r/268557 [21:30:27] to port it to rake ... I have no idea :-/ [21:30:58] but we have a task for that https://phabricator.wikimedia.org/T109005 Switch CI from jsduck deb package to a gemfile/bundler system [21:30:58] hashar: Ok. i will go and recheck some patches in those branches. But i would have to use check experimental since those branches are blacklisted. [21:31:05] (03CR) 10jenkins-bot: [V: 04-1] Blacklist REL1_25 branch for jsduck and rake-jessie [integration/config] - 10https://gerrit.wikimedia.org/r/268557 (owner: 10Paladox) [21:32:02] OHHH [21:32:04] HOLD ON [21:32:06] magic [21:32:20] paladox: so if you look at https://gerrit.wikimedia.org/r/#/c/268557/2/zuul/layout.yaml,cm [21:32:27] on line 609 there is a magic skip-if [21:32:31] so [21:32:40] I can't remember the doc off hand but I must have read it [21:32:46] and in theory [21:32:48] you can do something like: [21:32:52] - name: ^jsduck$ [21:32:55] skip-if: [21:33:08] - project: '^mediawiki/extensions/Flow$' [21:33:21] branch: (?:^REL1_25$) [21:33:22] hashar: Oh ok i will do that instead thanks hashar. [21:33:32] which would skip running jsduck for Flow on branch REL1_25 :-} [21:33:36] sorry :( [21:33:42] (03Merged) 10jenkins-bot: Handle cyclical extension dependencies [integration/config] - 10https://gerrit.wikimedia.org/r/268451 (owner: 10Dduvall) [21:33:46] (03PS3) 10Paladox: Blacklist REL1_25 branch for jsduck and rake-jessie [integration/config] - 10https://gerrit.wikimedia.org/r/268557 [21:33:57] Zuul learned a bit of workflow recently [21:34:14] Do you know why when I try to "git review" from some release branches, it fails because the master has been closed (! [remote rejected] HEAD -> refs/publish/master/268473 (change 219550 closed))? [21:34:23] I can work around it by changing defaultbranch in .gitreview, but is that just a bug in the branch setup, or am I missing something? [21:35:04] paladox: there are some indications on http://docs.openstack.org/infra/zuul/zuul.html#jobs (search for 'skip-if' ) . The doc is for the latest version so might not match our version though [21:35:20] matt_flaschen: you are using the same change-id ? [21:35:50] hashar: Ok thanks. I found skip-if in layout.xml but didnt occur to me we could use it until you said. [21:36:35] matt_flaschen: this change https://gerrit.wikimedia.org/r/#/c/219550/ has been merged. The change number correspond to a unique key made of (repo, branch, change-id) or (Scribunto, master, I9c3597e4fea669ff7ec5086dfc6a61d797fa48fe) [21:36:38] hashar, yes, which works fine if it's a different branch. But .gitreview seems to be using the wrong defaultbranch. [21:37:02] matt_flaschen: so if you try to propose a patch for Scribunto@master with change-id I9c3597e4fea669ff7ec5086dfc6a61d797fa48fe ... Gerrit considers it is change 219550 and tell you it is closed [21:37:15] (03CR) 10Florianschmidtwelzow: "Thanks :)" [integration/config] - 10https://gerrit.wikimedia.org/r/267553 (owner: 10Florianschmidtwelzow) [21:37:16] the trick is the unique key has the branch [21:37:34] hashar: Would we also need to do postmerge for flow. Since it has postmerge for jsduck. [21:37:37] so if you propose it to a different branch (eg: same repo but branch REL1_26), Gerrit creates a new change [21:37:45] I know, I'm trying to do REL1_25, but the REL1_25 defaultbranch is apparently wrong: https://git.wikimedia.org/blob/mediawiki%2Fextensions%2FScribunto.git/REL1_25/.gitreview#L5 [21:37:55] matt_flaschen: look at .gitreview maybe the branch parameter hasn't been updated [21:37:56] And I've seen this on at least one other repo too. [21:38:04] hashar, could this be a bug in the branch creation script? [21:38:19] who knows [21:38:51] $ git show origin/REL1_26:.gitreview|grep defaultbranch [21:38:51] defaultbranch=REL1_26 [21:39:09] looks like jenkins is contaminated for wikibase [21:39:10] https://gerrit.wikimedia.org/r/#/c/268558/ [21:39:14] matt_flaschen: Try changing it to REL1_25 branch see what it does. [21:39:21] with stuff from the release branch [21:39:40] matt_flaschen: maybe we made the branch cutter to update default branch after REL1_25 got cut [21:39:42] jzerebecki: ^ [21:40:09] eek [21:40:21] hashar, yeah, you're right, it's fixed: https://git.wikimedia.org/blob/mediawiki%2Fextensions%2FScribunto.git/REL1_26/.gitreview#L5 [21:41:11] (03PS3) 10JanZerebecki: [VisualEditor] Depend on the Cite extension [integration/config] - 10https://gerrit.wikimedia.org/r/268440 (owner: 10Jforrester) [21:41:16] (03CR) 10JanZerebecki: [C: 032] [VisualEditor] Depend on the Cite extension [integration/config] - 10https://gerrit.wikimedia.org/r/268440 (owner: 10Jforrester) [21:41:55] matt_flaschen: so craft another change to update the defaultbranch, rebase, done :-} [21:41:59] (03PS1) 10Thcipriani: Update specialExtensions submodule add [tools/release] - 10https://gerrit.wikimedia.org/r/268562 [21:42:11] aude wondering how npm ends up failling [21:42:18] hashar, yeah, noted, I just wanted to make sure there wasn't some subtlety I was missing. [21:42:26] i think the workspace needs to be purged [21:42:32] hashar: Ok. if we were to add jsduck to experimental in mediawiki core would those branch filter still block it from running on REL1_25. [21:42:37] (03Merged) 10jenkins-bot: [VisualEditor] Depend on the Cite extension [integration/config] - 10https://gerrit.wikimedia.org/r/268440 (owner: 10Jforrester) [21:42:48] phpunit tests complain about ORMTable not being there [21:42:51] We have been trying to fix Wikibase and Scribunto issues on REL1_25, don't know how that broke Jenkins though. [21:42:52] aude: https://integration.wikimedia.org/ci/job/mwext-Wikibase-repo-tests-mysql-hhvm/7993/consoleFull that one fails with: Fatal error: unknown class ORMTable in /mnt/jenkins-workspace/workspace/mwext-Wikibase-repo-tests-mysql-hhvm/src/extensions/Wikibase/lib/includes/ChangesTable.php [21:42:58] which we long since removed and removed usage in wikibase [21:43:16] so it has old code somewhere, such as from REL1_25 patches [21:43:38] (03PS4) 10Paladox: Blacklist REL1_25 branch for jsduck and rake-jessie in extension flow [integration/config] - 10https://gerrit.wikimedia.org/r/268557 [21:43:41] if we look at the top of that console output, the zuul-cloner info/debug notify which branches ends up being checked out [21:44:04] 00:00:15.630 INFO:zuul.Cloner:upstream repo is missing branch None [21:44:10] sounds bad [21:44:17] :( [21:44:31] hashar: Please review https://gerrit.wikimedia.org/r/#/c/268557/ [21:45:54] hashar: if we were to add jsduck to experimental in mediawiki core would those branch filter still block it from running on REL1_25. [21:46:24] (03PS5) 10Paladox: Blacklist REL1_25 branch for jsduck and rake-jessie in extension flow [integration/config] - 10https://gerrit.wikimedia.org/r/268557 [21:46:36] aude: wrong branch? [21:46:47] wmf/1.26wmf12 instead of 1.27... [21:47:19] oh, could be [21:47:40] yeah ... doh [21:47:42] jzerebecki: Could you review https://gerrit.wikimedia.org/r/#/c/268306/ for wikidata please. [21:48:47] ahhh [21:49:04] aude: yeah wrong branch. So there is None in mediawiki/core and zuul-cloner ends up fall back to 'master' [21:49:10] jzerebecki: you are awesome :-} [21:49:51] (03PS4) 10JanZerebecki: Run npm test for wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [21:49:56] (03CR) 10JanZerebecki: [C: 032] Run npm test for wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [21:50:39] qa-morebots: ping [21:50:39] I am a logbot running on tools-exec-1214. [21:50:39] Messages are logged to https://tools.wmflabs.org/sal/releng. [21:50:39] To log a message, type !log . [21:50:49] !log salt-key -d integration-slave-precise-1011.eqiad.wmflabs [21:50:51] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [21:51:16] !log salt-key -d integration-slave-jessie-1001.eqiad.wmflabs [21:51:19] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [21:57:22] (03PS1) 10Addshore: Remove OAuthAuthentication PluggableAuth dep [integration/config] - 10https://gerrit.wikimedia.org/r/268566 [21:57:38] jzerebecki: hashar it would be really cool to have the above deployed asap ;) [21:58:59] hashar: If i was to add jsduck to experimental in mediawiki/core would that run or would it not. [22:00:01] addshore: uh i think that was needed before. was that extension converted to the new auth thingie? what happens with release or old deployment branches? [22:00:28] jzerebecki: not converted [22:00:49] hence the tests for PlugableAuth are killing the OauthAuth ones :/ [22:01:09] which in turn of course means my core patch would be unmergable (hehee dependancy tree) [22:01:39] (03Merged) 10jenkins-bot: Run npm test for wikidata/build-resources [integration/config] - 10https://gerrit.wikimedia.org/r/268436 (https://phabricator.wikimedia.org/T125837) (owner: 10Aude) [22:01:55] addshore: but doesn't that make that extension broken to begin with? [22:02:08] which extension? :P [22:02:21] ostriches, thcipriani: I just realized that "we" probably need to rebuild/replace deployment-bastion with something running trusty [22:02:51] jzerebecki: both? [22:02:52] And split it from being fenari 4.0 [22:02:52] or the bump to PHP 5.5+ will melt beta cluster :) [22:02:53] :) [22:03:07] jzerebecki: OAuthAuthentication is not a PluggableAuth extension, I'm not sure what's the point of the dependency [22:03:19] thx [22:03:36] jzerebecki: PluggableAuth has not been changed since May 2015, OAuthAuthentication also since May [22:04:19] tgr: what about OpenIDConnect ? [22:04:24] looking at the commits it really doesn't look like the have ever had a dependency [22:04:39] addshore: to late for me to deploy anything sorry :-/ [22:04:44] jzerebecki: also not, I think [22:04:53] hashar: i'm already on it [22:04:58] \O/ [22:05:01] \o/ [22:05:19] oh [22:05:26] I thought it was a mediawiki-config change due [22:05:44] addshore: dont you have JJB access to update the jobs? [22:05:53] hashar: nope :P or not that I know of [22:05:56] D: [22:06:08] I may however have it and have forgiotten ;) [22:06:16] you mightn't be able to CR+2 [22:06:18] should ask the cabal [22:06:28] haha [22:06:56] which is probably all about offering jzerebecki his favorite bewerage [22:07:21] jzerebecki: still in the office? ;) I can get you a beverage ;) [22:07:30] addshore: you dont have access [22:07:35] wmf ldap group does [22:07:46] and Jan has a per-user override [22:07:58] jzerebecki: all PluggableAuth extensions implement a PluggableAuthSomething hook, I believe [22:08:30] !log reloading zuul for bed7be1..f57b7e2 [22:08:33] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [22:09:22] jzerebecki: https://gerrit.wikimedia.org/r/#/c/267684/ is the patch. Let me know when it is ready for a recheck! [22:10:11] (03PS1) 10Paladox: Allow jsduck in mediawiki/core to test branches REL1_25 and REL1_26 [integration/config] - 10https://gerrit.wikimedia.org/r/268569 [22:10:51] (03PS2) 10Paladox: Allow jsduck in mediawiki/core to test branches REL1_25 and REL1_26 [integration/config] - 10https://gerrit.wikimedia.org/r/268569 [22:11:13] so, I have some doubts about how to proceed regarding scheduling a mediawiki deployment [22:11:24] *asking* to schedule [22:11:25] hashar: Please merge https://gerrit.wikimedia.org/r/#/c/268569/ i would like to test jsduck quickly on REL1_25 and REL1_26 branch. [22:11:32] tgr: grep says OpenIDConnect depends on it [22:11:52] (03PS2) 10JanZerebecki: Remove OAuthAuthentication PluggableAuth dep [integration/config] - 10https://gerrit.wikimedia.org/r/268566 (owner: 10Addshore) [22:11:55] this is the context: https://phabricator.wikimedia.org/T69223#1877878 [22:12:08] and https://phabricator.wikimedia.org/T69223#2000165 [22:12:31] jzerebecki: could be, I'm probably mixing it up with the OpenID extension [22:12:59] should I add it to swat directly? [22:15:06] hey jynus!Yeh, normally you just add it to SWAT on https://wikitech.wikimedia.org/wiki/Deployments [22:15:16] (03CR) 10JanZerebecki: [C: 032] "Deployed to Jenkins: (['mwext-OAuthAuthentication-testextension-hhvm', 'mwext-OAuthAuthentication-testextension-php53'])" [integration/config] - 10https://gerrit.wikimedia.org/r/268566 (owner: 10Addshore) [22:15:38] paladox: branch: ^(REL.*|master)$ [22:15:42] addshore, what if it is a *new* feature, I read it follow the "normal deployment" [22:15:47] (03CR) 10Hashar: [C: 04-1] Allow jsduck in mediawiki/core to test branches REL1_25 and REL1_26 (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/268569 (owner: 10Paladox) [22:16:16] paladox: a side effect of the patch is the job will trigger on any branches. I think it is fine to have it run only on REL and master [22:16:30] jynus: is this a config change? [22:16:39] yes, enabling a new feature [22:16:48] hashar: Ok sorry i will do that now. [22:16:51] I asked to test it on testwiki first [22:17:01] after I deployed a schema change [22:17:34] jynus: that is fine in swat (a small config change, especially because it is on testwiki) [22:17:43] jzerebecki, thanks [22:18:05] I know the technical side, I never faced the procedure myself [22:18:14] (03Merged) 10jenkins-bot: Remove OAuthAuthentication PluggableAuth dep [integration/config] - 10https://gerrit.wikimedia.org/r/268566 (owner: 10Addshore) [22:18:24] (03PS3) 10Paladox: Allow jsduck in mediawiki/core to test branches REL1_25 and REL1_26 [integration/config] - 10https://gerrit.wikimedia.org/r/268569 [22:18:25] and then just be around for the point of deployment jynus :) [22:18:33] hashar: Ok done ive uploaded the new patch. [22:18:34] well, I am asking for a fried [22:18:45] so I will tell him! [22:19:20] addshore: done [22:19:31] rechecking [22:19:37] many thanks! :) [22:22:19] (03PS4) 10Hashar: Allow jsduck in mediawiki/core to test branches REL1_25 and REL1_26 [integration/config] - 10https://gerrit.wikimedia.org/r/268569 (owner: 10Paladox) [22:23:01] (03CR) 10Hashar: [C: 032] "That should do it. Once deployed we will want to recheck master, a REL branch and a wmf branch. Should trigger on master/REL and not on " [integration/config] - 10https://gerrit.wikimedia.org/r/268569 (owner: 10Paladox) [22:23:21] (03CR) 10Dduvall: [C: 032] Update specialExtensions submodule add [tools/release] - 10https://gerrit.wikimedia.org/r/268562 (owner: 10Thcipriani) [22:23:21] paladox: rolling it , will sleep after that one [22:24:11] hashar: Thanks. [22:24:31] (03Merged) 10jenkins-bot: Allow jsduck in mediawiki/core to test branches REL1_25 and REL1_26 [integration/config] - 10https://gerrit.wikimedia.org/r/268569 (owner: 10Paladox) [22:24:33] (03Merged) 10jenkins-bot: Update specialExtensions submodule add [tools/release] - 10https://gerrit.wikimedia.org/r/268562 (owner: 10Thcipriani) [22:25:34] hashar: Ive done recheck on https://gerrit.wikimedia.org/r/#/c/265230/ which is master branch. [22:26:04] paladox: doing rechecks [22:26:13] hashar: Yes. [22:26:29] REL1_25 https://gerrit.wikimedia.org/r/#/c/247337/1 [22:26:41] wmf https://gerrit.wikimedia.org/r/#/c/268096/1 [22:26:56] hashar: thanks. [22:27:03] and a fundraising branch https://gerrit.wikimedia.org/r/#/c/181251/1 [22:27:51] paladox: looks good to me :-} [22:28:01] hashar: Yes. [22:29:07] hashar: It seems that mediawiki/core [22:29:08] [22:29:08] 268096,1 [22:29:08] has frozen in test. [22:29:30] frozen ? [22:29:46] ah [22:29:47] bu [22:29:49] bug [22:29:51] hashar: Some of the tests arn't being run but it shows red. [22:29:57] hashar: Yes looks like it. [22:29:59] I have manually aborted the builds to free up [22:30:01] "aborted" [22:30:22] hashar: Ok. well we know it works now i think. jsduck wasent run on that branch. [22:30:57] hashar: Could you qucikly merge https://gerrit.wikimedia.org/r/#/c/268557/ for flow please. [22:36:08] Project beta-scap-eqiad build #88639: 04FAILURE in 2 min 53 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/88639/ [22:37:06] lots of: [22:37:08] 22:36:07 rsync: rename failed for "/srv/mediawiki/php-master/cache/l10n/upstream/l10n_cache-kl.cdb.MD5" (from php-master/cache/l10n/upstream/.~tmp~/l10n_cache-kl.cdb.MD5): No such file or directory (2) [22:38:12] (03CR) 10Legoktm: "-1, I would have objected that we should not have cyclical extension dependencies." [integration/config] - 10https://gerrit.wikimedia.org/r/268451 (owner: 10Dduvall) [22:38:15] paladox: would be fixed with https://review.openstack.org/276480 [22:38:15] did it break again because of the uid fix in prod? https://phabricator.wikimedia.org/T117016 [22:38:22] paladox: the abort thing I mean [22:38:33] why would that be now instead of yesterday? /me is making things up [22:38:40] hashar: Thanks. [22:38:43] marxarelli: what extensions need cyclical dependencies? [22:39:04] legoktm: cite and ve were what caused the conversation [22:39:06] legoktm: James_F requested it [22:39:34] i thought it was sane either way to not have get_dependencies blow up should it happen :) [22:39:42] hashar: Could you qucikly merge https://gerrit.wikimedia.org/r/#/c/268557/ for flow please. [22:39:44] oookay.... [22:39:44] legoktm: VE unit tests depend on Cite; Cite VE code depends on VE. [22:39:54] that's a bit weird. [22:39:55] legoktm: We need to factor out the code in VE-MW into Cite, but that'll take some time. [22:40:00] legoktm: Inorite. [22:40:17] legoktm: See https://gerrit.wikimedia.org/r/#/c/254082/ [22:40:25] (03CR) 10Legoktm: "Okay, on IRC it was explained this is needed for Cite/VE, which depend upon each other for tests." [integration/config] - 10https://gerrit.wikimedia.org/r/268451 (owner: 10Dduvall) [22:40:37] funnn :DD [22:40:49] legoktm: It gets better. [22:41:16] legoktm: There's code in VE-core that is extended in VE-MW which is mostly tested there, but used by the code that's (soon to be) in Cite. [22:42:12] legoktm: Also, if you're free, php55… ;-D [22:42:49] I'm not gonna do it until I'm back home, there are too many moving parts and right now I'm not going to be consistently around to fix it when it breaks [22:43:04] So Sunday or Monday probably [22:43:16] Kk. [22:43:37] In that case, want to consult on WTF is causing that VE patch to fail? The tests work perfectly locally. :-( https://integration.wikimedia.org/ci/job/mediawiki-extensions-qunit/29275/consoleFull [22:45:43] 22:42:13 Error: No class registered by that name: mwReference [22:45:44] 22:42:13 at VeDmNodeFactory.oo.Factory.create [22:46:15] (03CR) 10Hashar: [C: 04-1] "Almost there. See comment to slightly adjust the branch filter. Once taken in account I think it is good to be deployed (poke #wikimedia-" (032 comments) [integration/config] - 10https://gerrit.wikimedia.org/r/268557 (owner: 10Paladox) [22:46:17] paladox: reviewed [22:46:39] hashar: Thanks. I will go do that now. [22:46:41] dirty ops/puppet on puppetmaster ... [22:46:43] yet again [22:46:43] Yippee, build fixed! [22:46:44] Project beta-scap-eqiad build #88640: 09FIXED in 8 min 46 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/88640/ [22:46:46] paladox: now I really need to sleep. Thanks! [22:46:58] hashar: Ok. [22:47:23] mobrovac: that happens whenever a cherry picked patches conflict with origin/production . Needs to abort rebase (git rebase --abort) and rebase manually [22:47:38] k, will try [22:47:46] mobrovac: most of the time a cherry picked patch got merged but with slight/trivial changes [22:49:09] (03PS6) 10Paladox: Blacklist REL1_25 branch for jsduck and rake-jessie in extension flow [integration/config] - 10https://gerrit.wikimedia.org/r/268557 [22:49:46] legoktm: mwReference is registered by ext.cite.visualeditor though. https://github.com/wikimedia/mediawiki-extensions-Cite/blob/HEAD/modules/ve-cite/ve.dm.MWReferenceModel.js [22:50:01] legoktm: Is it being run without the Cite dependency? [22:50:13] is Cite in the mediawiki-extensions shared job? [22:50:43] (03PS7) 10Paladox: Blacklist REL1_25 branch for jsduck and rake-jessie in extension flow [integration/config] - 10https://gerrit.wikimedia.org/r/268557 [22:51:07] hashar: Ive uploaded a new patch now. I also added the branch filter to jsduck for flow. Which i forgot to do prevously. [22:51:07] paladox: poked team. Sleeping now [22:51:20] hashar: Ok thanks. [22:52:11] legoktm: Oh, does the dependency graph not get injected into that? [22:52:11] legoktm: I thought that that was its point? [22:52:20] no [22:52:26] mediawiki-extensions is special [22:54:08] mediawiki-extensions-qunit you mean? [22:54:08] It seems to define Cite and VisualEditor. [22:54:35] hmm, not sure then [22:54:45] I assume they pass in your browser? [22:55:16] Yes. [22:56:19] "Tests completed in 27559 milliseconds. 6620 assertions of 6620 passed, 0 failed.”