[07:10:06] (03CR) 10Tobias Gritschacher: [C: 04-1] "(2 comments)" [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/88770 (owner: 10Daniel Kinzler) [07:12:24] (03CR) 10Tobias Gritschacher: [C: 032] Update dependency version [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88740 (owner: 10Jeroen De Dauw) [07:16:03] (03Merged) 10jenkins-bot: Update dependency version [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88740 (owner: 10Jeroen De Dauw) [08:33:04] hashar: ? :P [08:36:05] addshore: yeah lets do the jenkins stuff [08:36:23] :D [08:40:03] so is wikidata another extension? [08:40:27] Wikidata is the 'build' that will eventually be deployed, rather than having to deploy multiple extensions [08:40:37] so it got submodules I guess ? [08:41:10] hashar: its not all entirly figured out yet :) [08:41:33] please please make sure Sam Reed is aware of it [08:41:46] and probably want to get Ryan Lane in the loop as well (he is working on Sartoris, a new deployment system) [08:41:57] having submodules in an extension potentially have some impacts [08:42:07] submodules = good or bad? :P [08:42:48] it is irrelevant [08:42:51] if you need them so be it [08:43:01] but make sure people doing the wmf branches / deploying new releases are aware of it :-] [08:43:05] just for safety [08:43:09] okay, will make sure we discuss today and work out which we decide to go for [08:43:25] you don't want your deployment of Wikidata extension to be cancelled because impacts have not be evaluated beforehand [08:43:31] I imagine everyone doing deploys will be aware when we switch from many extensions to a single extension [08:43:39] most probably [08:43:55] my point is merely to make people aware of it beforehand [08:44:12] anyway, the wikidata built extension probably wont be deployed for months as we tweak stuff while talking to wmf :) [08:50:08] (03CR) 10Hashar: "You might want to use the Classmap autoloading system in composer. That would generate a flat list of classname -> files." [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 (owner: 10Addshore) [08:51:15] I have another jenkins job change to check first [08:51:19] then will do the Wikidata one [08:51:58] (03CR) 10Addshore: "Will do." [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 (owner: 10Addshore) [08:52:33] ahh hashar I actually need 2 jobs changes ;p Not just the wikidata one, also the one on wikibase splitting the repo and api jobs :) [08:55:02] (03CR) 10Daniel Kinzler: "(2 comments)" [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/88770 (owner: 10Daniel Kinzler) [09:16:58] addshore: ok task completed [09:17:01] lets do the wikidata thing [09:18:14] addshore: DanielK_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=47022 [09:27:00] addshore: I have deployed basic jobs for Wikidata extension [09:27:49] (03PS4) 10Hashar: Wikidata Build from 12:00 09.OCT.2013 DNM [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 (owner: 10Addshore) [09:28:13] (03CR) 10jenkins-bot: [V: 04-1] Wikidata Build from 12:00 09.OCT.2013 DNM [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 (owner: 10Addshore) [09:28:23] addshore: and that fails :( https://integration.wikimedia.org/ci/job/mwext-Wikidata-testextensions-master/1/ [09:30:12] hmm, okay, will look at that shortly :) [09:30:49] heh, while running update.php :P [09:33:12] I imagine thats something to do with the change of strucute then [09:36:18] php maintenance/eval.php [09:36:18] PHP Notice: Undefined index: file in /srv/ssd/jenkins-slave/workspace/mwext-Wikidata-testextensions-master/includes/Exception.php on line 740 [09:36:19] PHP Stack trace: [09:36:20] yeahhh [09:37:21] I'm slightly confused.. Exceptions is erroring :P [09:37:47] yeah [09:37:58] here is the real stack trace [09:37:58] http://paste.openstack.org/show/48198/ [09:38:04] *goes to try and run update aginst this on his test instance [09:38:14] and I need to inject $wgShowExceptionDetails = true [09:38:20] addshore: fyi: https://www.wikidata.org/wiki/Wikidata:Project_chat#Recent_changes [09:39:05] * hashar facepalm [09:39:16] :> [09:39:55] ahhhhh! [09:39:58] I know what this is [09:40:19] need to use the default config files for wikibase I imagine [09:40:43] *wonders where this is added when running the tests in wikibase repo.. [09:41:39] oh yay, looks like we need some fun entry point hacking.. [09:44:21] hashar: Lets just leave that one for now, we should be able to fix the tests our end, jenkins seems to be trying to do the right thing, just the build isnt quite setup for the tests to run it seems [09:46:25] Tobi_WMDE: did you already see https://bugzilla.wikimedia.org/show_bug.cgi?id=48962#c5 ? [09:47:23] Lydia_WMDE: yes, saw it but according to Henning_WMDE this wont change much in solving the bug [09:47:52] ok [09:48:05] documenting discussions in bugs ++ ;-) [09:58:35] addshore: i am adding api_breaking_change to the whiteboard of that bug so we can keep track of the breaking api changes we have lined up and can do them all at once - feel free to add to more bugs [09:59:23] addshore: yeah Wikibase has a default entry point with some weird switch according to JENKINS_JOB env variable [10:01:56] coding dojo, be back in a couple hours [10:03:50] (03PS1) 10Addshore: Fix claim weirdness [extensions/Wikibase] (mw1.22-wmf21) - 10https://gerrit.wikimedia.org/r/88949 [10:04:46] (03CR) 10Tobias Gritschacher: [C: 032] Fix claim weirdness [extensions/Wikibase] (mw1.22-wmf21) - 10https://gerrit.wikimedia.org/r/88949 (owner: 10Addshore) [10:05:14] (03PS2) 10Tobias Gritschacher: Fix claim weirdness [extensions/Wikibase] (mw1.22-wmf21) - 10https://gerrit.wikimedia.org/r/88949 (owner: 10Addshore) [10:05:50] (03PS3) 10Tobias Gritschacher: Fix claim weirdness [extensions/Wikibase] (mw1.22-wmf21) - 10https://gerrit.wikimedia.org/r/88949 (owner: 10Addshore) [10:06:18] (03CR) 10Tobias Gritschacher: [C: 032] Fix claim weirdness (cherry picked from commit 46948189c275acd63429f6e42e26d3bc551e9ee4) [extensions/Wikibase] (mw1.22-wmf21) - 10https://gerrit.wikimedia.org/r/88949 (owner: 10Addshore) [10:10:16] (03CR) 10Addshore: "(1 comment)" [extensions/WikibaseDatabase] - 10https://gerrit.wikimedia.org/r/88429 (owner: 10Addshore) [10:12:17] (03Merged) 10jenkins-bot: Fix claim weirdness (cherry picked from commit 46948189c275acd63429f6e42e26d3bc551e9ee4) [extensions/Wikibase] (mw1.22-wmf21) - 10https://gerrit.wikimedia.org/r/88949 (owner: 10Addshore) [10:16:01] [travis-ci] wikimedia/mediawiki-extensions-Wikibase#1005 (mw1.22-wmf21 - 7028ace : addshore): The build has errored. [10:16:01] [travis-ci] Change view : https://github.com/wikimedia/mediawiki-extensions-Wikibase/compare/a2056e5f7a01...7028ace5809a [10:16:01] [travis-ci] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-Wikibase/builds/12361532 [10:21:45] (03PS1) 10Addshore: Refactor SQLiteTableDefinitionReader [extensions/WikibaseDatabase] - 10https://gerrit.wikimedia.org/r/88953 [10:22:41] (03PS4) 10Addshore: Update docs and @since tags. [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88739 (owner: 10Jeroen De Dauw) [10:25:32] (03CR) 10jenkins-bot: [V: 04-1] Update docs and @since tags. [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88739 (owner: 10Jeroen De Dauw) [11:33:07] (03PS5) 10Daniel Kinzler: (bug 52916) Baseline for Quantity data type. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/84120 [11:33:23] (03CR) 10jenkins-bot: [V: 04-1] (bug #55513) Baseline for Quantity data type. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/84120 (owner: 10Daniel Kinzler) [11:33:32] (03CR) 10Daniel Kinzler: [C: 04-1] "Blocked on QuantityType implementation (bug #55512)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/84120 (owner: 10Daniel Kinzler) [11:33:53] (03PS6) 10Daniel Kinzler: (bug #55513) Baseline for Quantity data type. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/84120 [11:34:02] (03CR) 10jenkins-bot: [V: 04-1] (bug #55513) Baseline for Quantity data type. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/84120 (owner: 10Daniel Kinzler) [11:48:29] (03PS1) 10Tobias Gritschacher: Fix encoding bug in sitelinks browsertest [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88960 [12:04:01] (03PS1) 10Tobias Gritschacher: setClaim autosummary should not be overwritten by ChangeOp autosummary [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88964 [12:04:17] addshore: https://gerrit.wikimedia.org/r/#/c/88964/ [12:04:33] (03CR) 10Addshore: [C: 032] setClaim autosummary should not be overwritten by ChangeOp autosummary [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88964 (owner: 10Tobias Gritschacher) [12:05:50] (03PS1) 10Tobias Gritschacher: setClaim autosummary should not be overwritten by ChangeOp autosummary [extensions/Wikibase] (mw1.22-wmf21) - 10https://gerrit.wikimedia.org/r/88965 [12:06:18] (03CR) 10Addshore: [C: 032] setClaim autosummary should not be overwritten by ChangeOp autosummary [extensions/Wikibase] (mw1.22-wmf21) - 10https://gerrit.wikimedia.org/r/88965 (owner: 10Tobias Gritschacher) [12:10:23] (03Merged) 10jenkins-bot: setClaim autosummary should not be overwritten by ChangeOp autosummary [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88964 (owner: 10Tobias Gritschacher) [12:11:40] (03PS1) 10Jeroen De Dauw: Update gitignore [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88966 [12:12:33] (03PS2) 10Jeroen De Dauw: Install and update the store when needed [DNM, DRAFT] [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88753 [12:12:40] (03CR) 10Tobias Gritschacher: [C: 032] Fix encoding bug in sitelinks browsertest [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88960 (owner: 10Tobias Gritschacher) [12:14:33] (03CR) 10jenkins-bot: [V: 04-1] Update gitignore [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88966 (owner: 10Jeroen De Dauw) [12:16:17] (03Merged) 10jenkins-bot: setClaim autosummary should not be overwritten by ChangeOp autosummary [extensions/Wikibase] (mw1.22-wmf21) - 10https://gerrit.wikimedia.org/r/88965 (owner: 10Tobias Gritschacher) [12:17:27] (03CR) 10jenkins-bot: [V: 04-1] Install and update the store when needed [DNM, DRAFT] [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88753 (owner: 10Jeroen De Dauw) [12:19:06] (03Merged) 10jenkins-bot: Fix encoding bug in sitelinks browsertest [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88960 (owner: 10Tobias Gritschacher) [12:22:19] *goes to decode what all of our jenkins jobs are doing even more.. [12:24:45] addshore: looks like WikibaseDatabase has a failing test or smth https://integration.wikimedia.org/ci/job/mwext-WikibaseQuery-testextensions-master/176/consoleFull [12:24:55] Getting this when changing the gitignore of WikibaseQuery [12:25:20] mhhhm [12:25:41] will take a look, thats rather odd [12:26:09] JeroenDeDauw: >> https://gerrit.wikimedia.org/r/#/c/88953/ [12:26:24] is jenkins running the integration tests? [12:26:52] addshore: it should be [12:27:07] addshore: not the ones that need mysql ofc [12:27:21] mhhm, but the ones you linked to above are in the sqlite classes [12:27:36] try merging my above refacotring? [12:27:53] addshore: I did find a bug in the code... dont think it is causing this tho [12:28:08] (03CR) 10Jeroen De Dauw: [C: 032] Refactor SQLiteTableDefinitionReader [extensions/WikibaseDatabase] - 10https://gerrit.wikimedia.org/r/88953 (owner: 10Addshore) [12:28:36] wheres the bug? [12:28:48] addshore: will commit fix in a sec [12:28:57] =] [12:29:00] addshore: looks like the bug is not causing any problems ATM though [12:33:02] (03PS1) 10Jeroen De Dauw: Add tableExists to QueryInterface as well [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88967 [12:33:35] (03CR) 10Jeroen De Dauw: [C: 04-2] Add tableExists to QueryInterface as well [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88967 (owner: 10Jeroen De Dauw) [12:34:05] (03PS1) 10Jeroen De Dauw: Add tableExists to QueryInterface as well [extensions/WikibaseDatabase] - 10https://gerrit.wikimedia.org/r/88968 [12:34:25] (03CR) 10Addshore: [C: 032] Add tableExists to QueryInterface as well [extensions/WikibaseDatabase] - 10https://gerrit.wikimedia.org/r/88968 (owner: 10Jeroen De Dauw) [12:34:27] heh [12:34:40] I spotted that a while ago but have never added it.. [12:34:42] (03PS3) 10Jeroen De Dauw: Install and update the store when needed [DNM, DRAFT] [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88753 [12:34:45] (03Abandoned) 10Jeroen De Dauw: Add tableExists to QueryInterface as well [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88967 (owner: 10Jeroen De Dauw) [12:35:03] .. [12:35:24] addshore: so this does not explain the issue seen in Jenkins [12:36:58] Hi. I'm new to wikidata. plz direct me to a start page to help me about adding inter wiki to a page. [12:37:10] (03PS1) 10Henning Snater: Allow passing index when adding a Reference to a ReferenceList [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/88969 [12:38:40] (03CR) 10jenkins-bot: [V: 04-1] Install and update the store when needed [DNM, DRAFT] [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88753 (owner: 10Jeroen De Dauw) [12:38:54] Big big props to Lydia_WMDE for cleaning up Bugzilla like mad! :P [12:38:58] addshore: so what happens is that the table definition reader is not reading in any fields... [12:39:07] addshore: works for me locally though [12:39:31] andre__: ;-) [12:39:45] andre__: you were away when i answered to your question re keywords the other day [12:39:53] JeroenDeDauw: I guess that is untested behaviour and we should probably expand things to account for it [12:40:07] Lydia_WMDE, I think I saw it [12:40:21] andre__: i think witing for feedback should be a status as i can't do anything with these bugs [12:40:27] *waiting [12:40:34] (03PS1) 10Jeroen De Dauw: Add extra check to test to get a better idea of what is failing [extensions/WikibaseDatabase] - 10https://gerrit.wikimedia.org/r/88972 [12:40:53] addshore: might be that the table creation is failing [12:41:00] Lydia_WMDE, feel free to read up/add your thoughts to https://bugzilla.wikimedia.org/show_bug.cgi?id=36064 [12:41:05] JeroenDeDauw: ... [12:41:06] or that there is an issue with the table name [12:41:13] andre__: ok [12:41:29] addshore: if at one place the name is formatted and not at the orher, this might happen [12:41:29] you just added an asertion which is already there.. [12:41:46] (03Abandoned) 10Jeroen De Dauw: Add extra check to test to get a better idea of what is failing [extensions/WikibaseDatabase] - 10https://gerrit.wikimedia.org/r/88972 (owner: 10Jeroen De Dauw) [12:41:56] Lydia_WMDE, as I'm not entirely sure if flag or status is better myself. both have pros and cons. recently I had a tendency to prefer the flag extension (if I ever manage to make it work) [12:42:12] k [12:42:17] will add after meeting [12:43:14] (03PS6) 10Daniel Werner: Got rid of "private" callback parameter in jQuery.ui.inputextender's draw [extensions/DataValues] - 10https://gerrit.wikimedia.org/r/79153 [12:44:55] (03PS3) 10Addshore: Refactor Jenkins Job Entry File [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88486 [12:47:50] addshore: I'll go back to doing some more abstract stuff for now, if I cant manage to commit stuff to the right repo, then debugging this is also not going to work :D [12:51:59] addshore: http://stackoverflow.com/questions/19253673/pdosqlite-does-not-run-multiple-statements-in-one-query/19255916? [12:52:14] addshore: so the same is probably true for MySQL then [12:53:12] addshore: I'm also wondering if doing the array thing you suggested is not better then the newline thing, assuming we do it for all the sql building functions [12:53:36] addshore: that is something we can change later though [12:57:18] (03PS1) 10Jeroen De Dauw: Rename method to match interface name [extensions/WikibaseQueryEngine] - 10https://gerrit.wikimedia.org/r/88975 [12:57:38] (03PS1) 10Jeroen De Dauw: Match https://gerrit.wikimedia.org/r/88975 in QueryEngine [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88976 [12:57:44] addshore: anything else ? [12:58:06] hashar: give me 2 ticks [12:58:49] do you see the 3 changes I have up? [12:58:54] *adds you as a reviewer [12:59:09] aude: ONOEZ! You caused build 1000 to error! https://travis-ci.org/wikimedia/mediawiki-extensions-Wikibase/builds [12:59:14] (03PS6) 10Henning Snater: Implemented claimgrouplistview widget [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88048 [13:00:04] hashar: mainly https://gerrit.wikimedia.org/r/#/c/88974/ (the other 2 are linked to) [13:00:42] (03CR) 10jenkins-bot: [V: 04-1] Match https://gerrit.wikimedia.org/r/88975 in QueryEngine [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88976 (owner: 10Jeroen De Dauw) [13:01:03] (03PS10) 10Daniel Werner: Introduces GenericSet.js [extensions/DataValues] - 10https://gerrit.wikimedia.org/r/76079 [13:01:03] error: too many tabs open [13:01:05] wtf [13:01:07] :> [13:01:50] (03CR) 10Jeroen De Dauw: [C: 032] Refactor Jenkins Job Entry File [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88486 (owner: 10Addshore) [13:01:59] addshore: deploying the repoapi jenkins job [13:02:04] okay :) [13:03:27] the change to the zuul-config should have the biggest effect [13:04:33] we can probably deploy it right now, that would run the tests using the fallback case in Wikibase: [13:04:34] require repo/Wikibase.php [13:04:35] require repo/ExampleSettings.php [13:04:39] though not sure whether that will fail or not [13:04:58] hashar: the change to change the entry point has just been merged [13:05:05] https://gerrit.wikimedia.org/r/#/c/88486/ [13:05:14] so that is why I like our community [13:05:19] you don't have to fight to get stuff done [13:05:20] (03PS2) 10Daniel Werner: Follows Core change Ib7a8a8f, implements EntityContent::getActualContent [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/83648 [13:05:23] :) [13:05:30] (03Abandoned) 10Daniel Werner: Follows Core change Ib7a8a8f, implements EntityContent::getActualContent [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/83648 (owner: 10Daniel Werner) [13:05:31] as soon as you are made aware of a proposal, it is already implemented and available [13:05:40] :d [13:05:42] :d [13:05:46] :D [13:05:54] I really havn't got used to this keyboard yet... [13:06:51] how do you do the reversed b ? [13:07:06] its a d.. :< [13:07:11] yeah that one [13:07:48] (03PS1) 10Jeroen De Dauw: Add update method to QueryStoreSetup [extensions/WikibaseQueryEngine] - 10https://gerrit.wikimedia.org/r/88977 [13:08:07] I type on a german keyboard with german keys but my computer layout in UK, hence If I ever look at the keyboard i tend to get things [13:08:08] wrong... [13:09:17] (03PS2) 10Addshore: Remove use of 2 deprecated js methods [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88046 [13:09:57] addshore: question asked on https://gerrit.wikimedia.org/r/#/c/88974/ [13:09:59] I think mwext-Wikibase-testextensions-master should either be triggered in test and gate-and-submit, or not triggered at all. [13:13:50] hmm [13:14:03] exactly [13:15:12] hashar: the annoying thing is the testextensions job takes twice the ammount of time as the next biggest job [13:15:32] but idealy should still be run as it runs the tests with both the client and the repo loaded at the same time [13:16:21] I could split it slightly further [13:16:38] you can use the split tests on patch submission (that is test pipeline) [13:16:44] so people get feedback faster [13:16:58] and run the slow one on gate-and-submit only, ensuring everything works one [13:16:58] (03PS1) 10Henning Snater: Removed movetoolbar widget's listView option [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88979 [13:17:09] imho gating can take several minutes, it is not really preventing you from developing [13:17:27] hashar: indeed, but the getsubmit should also check that they all work when loaded seperatly [13:17:36] yeah [13:17:37] *gate [13:17:51] so on gate-and-submit, trigger all jobs (the split one + the slow one that does client&repo) [13:18:09] thats what is in the patch :P [13:18:10] this way you get the best of both world: fast replies that covers almost all tests [13:18:17] and merged changes that pass any combinaison [13:18:20] ok [13:18:26] I think [13:18:36] unless I am reading the layout.yml wrong :) [13:18:43] na na it is fine [13:18:46] just wanted to make sure [13:21:20] (03PS4) 10Jeroen De Dauw: Install and update the store when needed [DNM, DRAFT] [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88753 [13:23:24] addshore: Zuul reloading. [13:23:48] :) [13:24:07] (03CR) 10jenkins-bot: [V: 04-1] Install and update the store when needed [DNM, DRAFT] [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88753 (owner: 10Jeroen De Dauw) [13:24:25] reloaded [13:24:34] submitted a test patch so I can make sure they all run as expected :) [13:24:37] 84527,16 is being tested right now [13:24:42] yup :0 [13:24:44] :) [13:24:46] https://integration.wikimedia.org/ci/job/mwext-Wikibase-repoapi-tests/1/ [13:26:30] and now watching the console output you understand why we get annoyed at tests taking forever :P [13:27:29] mwext-Wikibase-repoapi-tests: SUCCESS [13:27:32] where is the fun seriously [13:27:37] :d [13:27:45] !g 84527,16 [13:27:45] https://gerrit.wikimedia.org/r/#q,84527,16,n,z [13:27:50] * addshore is happy [13:28:08] so the slower is now repoapi :-D [13:28:10] well done! [13:28:30] it can probably be optimized [13:28:30] testextensions is still the slowest :) but we dont get to see it as much :) [13:28:43] the mw/core parser tests used to take like 3+ minutes [13:28:53] I think they are down to 1 minute after I optimized them [13:29:04] I guess we will have to try and mock api requests in a better way [13:29:07] there was lot useless initiazliation / deinitialization [13:29:22] well you are hitting core API aren't you ? [13:29:27] based on the API test helpers ? [13:29:33] yup [13:29:36] that might need some optimization in core itself [13:29:45] which would let every other folks to benefit from the optimziation [13:29:58] probably need to profile them and find out what is slowing them them [13:30:07] that might even be an issue in the API itself :-((( [13:30:12] I was taking a look the other day, problem is the api stuff is all linked together so tightly its hard to know where to start sometimes.. [13:30:26] if it turns out to be the case and it is fixable, that will lower the load on the wmf cluster \O: [13:30:28] \O/ [13:30:32] :P [13:30:39] profile EVERYTHING! [13:30:43] paying of by itself for the time spent on it [13:30:47] yeah profiling [13:31:10] on the cluster we do full profiling of a sample of requests (maybe 1 out a million ?? ) [13:31:32] ApiTestCase->doApiRequest [13:31:37] the wfProfieIn() / out timing ends up being sent to graphite, a metrics storing system [13:31:50] and tada : dashbbbooooaaard http://gdash.wikimedia.org/dashboards/apimethods/ [13:32:15] I have never seen that before... :O [13:32:39] basically every single wfProfiled function has a metric in that storage system [13:32:50] with # of occurrences / duration [13:32:59] we can then query the storage engine and graph whatever we want [13:33:08] for example the maximum time of a given function for a period of 1 hour [13:33:19] or the 95 percentile duration [13:33:24] interesting! [13:33:36] hmm, but this is still only say 1 in a million? [13:33:42] a pity is that the metrics list is only accessible to person in the `wmf` ldap group [13:33:44] or is this every run of the method? [13:33:50] I got a ticket pending to get wmde folks added to it [13:33:58] I have no idea what the sampling rate is [13:34:05] maybe 1/1000 [13:34:14] sounds about right [13:34:30] * addshore thinks you should up the priority of your RT ticket ;p [13:34:56] I see only of our api modules appearing rather allot on those graphs :P [13:35:31] I wonder where the others are :> [13:36:43] } elseif ( PHP_SAPI !== 'cli' && $wmfDatacenter == 'eqiad' && ( mt_rand() % 50 ) == 0 ) { [13:38:12] so 2% of requests I guess [13:38:16] roughly [13:40:16] damn [13:40:24] I am tired of finding bugs over and over [13:40:40] <: [13:44:13] (03PS5) 10Jeroen De Dauw: Install and update the store when needed [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88753 [13:47:20] (03CR) 10jenkins-bot: [V: 04-1] Install and update the store when needed [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88753 (owner: 10Jeroen De Dauw) [13:47:50] (03PS2) 10Jeroen De Dauw: Update gitignore [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88966 [13:50:35] (03CR) 10jenkins-bot: [V: 04-1] Update gitignore. [extensions/WikibaseQuery] - 10https://gerrit.wikimedia.org/r/88966 (owner: 10Jeroen De Dauw) [13:54:56] * JeroenDeDauw tried to run Wikibase tests [13:54:58] Failures: 21, Errors: 1333 [13:55:39] (03PS5) 10Addshore: Wikidata Build from 12:00 09.OCT.2013 DNM [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 [13:56:21] (03CR) 10jenkins-bot: [V: 04-1] Wikidata Build from 12:00 09.OCT.2013 DNM [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 (owner: 10Addshore) [14:06:04] (03PS6) 10Addshore: Wikidata Build from 12:00 09.OCT.2013 DNM [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 [14:06:41] (03CR) 10jenkins-bot: [V: 04-1] Wikidata Build from 12:00 09.OCT.2013 DNM [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 (owner: 10Addshore) [14:07:10] (03PS7) 10Addshore: Wikidata Build from 12:00 09.OCT.2013 DNM [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 [14:07:53] (03CR) 10jenkins-bot: [V: 04-1] Wikidata Build from 12:00 09.OCT.2013 DNM [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 (owner: 10Addshore) [14:08:38] hashar: still around? [14:10:24] addshore: yeah [14:10:38] addshore: heavily context switching, you get some interrupt cycles right now [14:10:42] damn jenkins [14:10:43] :( [14:10:49] ah no that is wikidata [14:10:52] with the above change, which includes submodules now as a test, [14:11:09] will jenkins automatically get the submodules or do we need to do something funky? [14:11:12] checking [14:11:17] https://integration.wikimedia.org/ci/job/mwext-Wikidata-testextensions-master/ws/ [14:11:20] workspace of last build [14:11:42] I need a login :P [14:11:49] your labs login [14:11:50] should do [14:11:52] :> [14:12:02] now that is very handy [14:12:06] isn't it ? [14:12:13] single user login (kind ofà [14:12:26] okay it does seem to pull everything :) [14:12:29] (03PS1) 10Jeroen De Dauw: Add updating capability to SQLStore setup [DNM, DRAFT] [extensions/WikibaseQueryEngine] - 10https://gerrit.wikimedia.org/r/88988 [14:25:01] addshore: looking at job conf [14:25:55] it is apparently supposed to fetch submodules [14:26:18] yep it looks like it does :) [14:26:28] *goes to find the next issue* :) [14:40:58] ahhhhh /srv/ssd/jenkins-slave/workspace/mwext-Wikidata-testextensions-master/extensions/Wikidata/Wikibase/repo/maintenance/../../../../maintenance/Maintenance.php [14:41:14] we are deeper inside the directory tree now so this fails :D [14:44:27] $basePath = getenv( 'MW_INSTALL_PATH' ) !== false ? getenv( 'MW_INSTALL_PATH' ) : __DIR__ . '/../../../..'; yay!" [14:46:16] (03CR) 10Addshore: [C: 04-1] "FAILS due to" [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/88701 (owner: 10Addshore) [14:53:34] addshore: http://www.lieferando.de/lieferservice-world-of-pizza-berlin#!cart [14:54:52] (03PS1) 10Henning Snater: Improved toolbarcontroller's event namespace assignment [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88994 [14:55:54] JeroenDeDauw: BBQ Rib Burger und Pommes Frites :) [14:55:55] Abraham_WMDE: what was the filling you had for the http://www.lieferando.de/lieferservice-world-of-pizza-berlin#!cart "Pizzabrötchen nach Belieben zum selber Füllen" [15:00:40] sounds like serious developers talk in here right now [15:00:55] (03PS1) 10Henning Snater: Implemented index in ChangeOpReference [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88997 [15:01:04] (03CR) 10jenkins-bot: [V: 04-1] Implemented index in ChangeOpReference [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88997 (owner: 10Henning Snater) [15:01:50] JeroenDeDauw: yep [15:05:02] (03PS1) 10Henning Snater: Implemented index in SetReference API module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88999 [15:05:13] hashar: still here? :> [15:05:17] (03CR) 10jenkins-bot: [V: 04-1] Implemented index in SetReference API module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88999 (owner: 10Henning Snater) [15:08:12] (03PS2) 10Henning Snater: Implemented index in SetReference API module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88999 [15:09:00] (03CR) 10jenkins-bot: [V: 04-1] Implemented index in SetReference API module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/88999 (owner: 10Henning Snater) [15:15:17] (03PS1) 10Henning Snater: Keep track of reference's index in referenceview [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/89004 [15:15:41] (03CR) 10jenkins-bot: [V: 04-1] Keep track of reference's index in referenceview [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/89004 (owner: 10Henning Snater) [15:16:32] (03PS1) 10Henning Snater: Implemented referenceview movetoolbar definition [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/89007 [15:16:42] (03CR) 10jenkins-bot: [V: 04-1] Implemented referenceview movetoolbar definition [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/89007 (owner: 10Henning Snater) [15:17:31] addshore: yeah [15:18:00] is it possible to set the MW_INSTALL_PATH environment var on jenkins? :) [15:18:27] thats the next thing that is tripping me up :) [15:18:57] * addshore has only just noticed you can pass shell commands in the yaml file :) [15:19:58] it should probably be set for all extension tests :) [15:22:34] oh really ? [15:22:40] yup :) [15:22:42] I don't think I ever needed it [15:22:56] the maintenance scripts basically all start with something like [15:22:57] $basePath = getenv( 'MW_INSTALL_PATH' ) !== false ? getenv( 'MW_INSTALL_PATH' ) : __DIR__ . '/../../../..'; [15:23:10] which is fine if you are a regular extension [15:23:25] oh my [15:23:41] but not in the wikidata repo the maint scripts are 2 directories deeper in the tree, so the '/../../../..' would need an extra ../../ ;p [15:23:52] we could: export MW_INSTALL_PATH="$WORKSPACE" [15:24:52] should be sufficient :) [15:25:52] addshore: you could hack that in Wikibase entry point meanwhile [15:26:03] whenever you are in the jenkins job context, you could setenv [15:26:19] something like: setenv( 'MW_INSTALL_PATH', getenv( 'WORKSPACE' ) ) [15:26:19] hashar: does not work, need this set before the entry point [15:26:24] ah man :( [15:26:36] hashar: else update.php fails [15:27:27] hashar: whats the problem with doing what addshore suggested? Ought to be trivial to do [15:27:29] Is it not? [15:27:48] it is [15:27:51] just have to think about it [15:27:59] and need to leave in 3 minutes to go grab my daughter :-D [15:28:05] I sent myself a reminder, will add that tonight [15:28:07] sorry :-((((((( [15:28:19] whenever we use travis , that would solve those issues [15:28:26] giving the power to code authors [15:30:02] :> [15:30:18] hashar: no rush :d ping me when you manage it and link me to your change? :) [15:32:00] hashar: yeah, rest assured that any finger pointing for stuff being slow will be at addshore :D [15:32:14] :D [15:32:23] hashar: what, plans to just use Travis? ;p [15:32:32] heh, I'm the one that refactored all of the api tests and added more test cases, so yes, all my fault :D [15:32:57] addshore: exactly, you are the last person that tauched it, so now the code is yours [15:33:16] addshore: please go click http://www.nooooooooooooooo.com/ [15:33:32] hah [15:34:12] JeroenDeDauw: the evil plan is to reuse travis mechanism to run tests instead of maintaining a soup of shell scripts in jenkins jobs [15:34:28] JeroenDeDauw: so potentially all jobs will end up with something like: phar travis $WORKSPACE/.travis.yaml [15:34:33] that is the evil plan [15:34:42] Travis Worker manages VMs, runs your tests, and does your washing :> [15:34:46] but we need to have the job run in isolated virtual machine first. that is being worked on [15:35:01] the VM we need to maintain a pool of them in labs [15:35:01] hashar: cool [15:35:20] which in turns need to write some vagrant files to boot vm in a labs isolated project and having the VM on dedicated hardware [15:35:27] I should write down all that stuff somewhere [15:35:28] hashar: then we could drop a lot of stupid hacks and just use our current travis stuff which uses Composer <3 [15:35:53] :-D [15:35:59] and drop Jenkins by the way [15:36:10] which is evil plan 2 [15:36:16] anyway I am off for real [15:36:18] gotta be lat [15:36:20] e [15:36:23] see ya :) [15:37:49] drop Jenkins? [15:38:05] heh - hes so going to be assasinated now [15:38:09] :> [15:38:41] <(^.^)> [15:38:44] JeroenDeDauw: ^^ [15:38:52]