[00:00:00] RoanKattouw: Thoughts? [00:00:07] spagewmf: actually, this is the first new db i've seen here :) there may be a system i don't know about [00:00:10] (03CR) 10jenkins-bot: [V: 04-1] Make header use *ActionPermissions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99416 (owner: 10Matthias Mullie) [00:01:04] Ohai [00:01:08] springle: well, there are cross-wiki databases, and there are databases on the extension1 cluster, but not I think the intersection of those [00:01:09] Is anyone else in this LD winodw? [00:01:12] Me. [00:01:14] RoanKattouw: marktraceur is. [00:01:56] OK go ahead then [00:02:00] I will go after you [00:02:02] 'kay [00:02:28] spagewmf: wikiuser will need grants to flowdb too, right [00:02:34] (03CR) 10MarkTraceur: [C: 032 V: 032] Update MultimediaViewer to work in conjunction with VisualEditor [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100718 (owner: 10MarkTraceur) [00:02:59] (03PS5) 10EBernhardson: Make header use *ActionPermissions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99416 (owner: 10Matthias Mullie) [00:03:05] springle, what bugzilla Product / Component for a new production DB [00:03:15] (03CR) 10Bsitu: Dont allow empty titles (031 comment) [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100601 (owner: 10EBernhardson) [00:04:06] (03CR) 10EBernhardson: Dont allow empty titles (031 comment) [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100601 (owner: 10EBernhardson) [00:04:26] (03CR) 10jenkins-bot: [V: 04-1] Make header use *ActionPermissions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99416 (owner: 10Matthias Mullie) [00:05:20] gwicke: got a moment to review https://gerrit.wikimedia.org/r/#/c/100542/ ? i'm looking to run it for a few hours and then roll it back [00:05:33] (03CR) 10EBernhardson: "Not sure whats wrong here, jenkins is complaining about non-existant issues afaict." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99416 (owner: 10Matthias Mullie) [00:05:44] (03PS1) 10Cmcmahon: run the test in beta labs [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100719 [00:06:47] (03PS2) 10Cmcmahon: run the test in beta labs [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100719 [00:07:08] (03CR) 10Bsitu: [C: 032] Dont allow empty titles [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100601 (owner: 10EBernhardson) [00:07:18] (03Merged) 10jenkins-bot: Dont allow empty titles [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100601 (owner: 10EBernhardson) [00:07:27] (03PS3) 10Cmcmahon: run the test in beta labs [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100719 [00:07:29] (03CR) 10jenkins-bot: [V: 04-1] run the test in beta labs [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100719 (owner: 10Cmcmahon) [00:07:39] spagewmf: i don't think we have such a category. Extensions/Flow and assign to me [00:07:46] (03CR) 10Krinkle: [C: 04-1] Site logo loaded using ResourceLoaderSiteModule instead of inline (033 comments) [core] - 10https://gerrit.wikimedia.org/r/98356 (owner: 10Tholam) [00:07:48] (03CR) 10Cmcmahon: [C: 032] "adding beta as target" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100719 (owner: 10Cmcmahon) [00:07:58] (03Merged) 10jenkins-bot: run the test in beta labs [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100719 (owner: 10Cmcmahon) [00:08:29] spagewmf: or use RT [00:08:37] which might actually make more sense [00:10:00] https://www.mediawiki.org/wiki/Talk:Download_from_Git badly needs git sourcerors [00:15:29] (03CR) 10EBernhardson: [C: 032] Introduce tipsy [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98820 (owner: 10Matthias Mullie) [00:15:57] (03CR) 10jenkins-bot: [V: 04-1] Introduce tipsy [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98820 (owner: 10Matthias Mullie) [00:17:04] (03PS1) 10Catrope: Update VisualEditor to wmf6 for cherry-picks [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100720 [00:18:48] (03CR) 10Catrope: [C: 032 V: 032] Update VisualEditor to wmf6 for cherry-picks [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100720 (owner: 10Catrope) [00:22:40] ori-l, did you test that code? [00:22:53] I am not sure that the Function actually modifies the global scope [00:23:55] (03CR) 10Jhall: [C: 032] login still failing occasionally in IEs [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100694 (owner: 10Cmcmahon) [00:25:05] any val declarations for example should behave differently when wrapped in a Function [00:25:14] (03Merged) 10jenkins-bot: login still failing occasionally in IEs [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100694 (owner: 10Cmcmahon) [00:25:48] gwicke: module code is always wrapped in a closure anyway -- mw.loader.implement( 'moduleName', function () { ... } ) [00:26:00] and yep, tested [00:26:28] I see, then there should not be much of a diff [00:26:49] probably not, but testing is cheap [00:26:54] that's basically http://jsperf.com/eval-vs-function-fib [00:27:09] yes, except on a representative sample of our traffic :) [00:27:30] (03CR) 10Spage: "Chad, flowdb will be a cross-wiki database (on the extension1 cluster). We need some way to point maintenance operations at it. Using --w" [core] - 10https://gerrit.wikimedia.org/r/100707 (owner: 10Aaron Schulz) [00:29:16] (03CR) 10GWicke: [C: 032] "Did not test this, but am taking your word for it that the source is always wrapped in functions anyway so that var scoping does not matte" [core] - 10https://gerrit.wikimedia.org/r/100542 (owner: 10Ori.livneh) [00:30:03] springle: I filed RT 6486 , thanks in advance. We're in #wikimedia-corefeatures if you have any questions [00:30:06] how do you figure out which was used btw? [00:30:17] (03CR) 10Chad: [C: 032] Added --wikidb param to sql.php [core] - 10https://gerrit.wikimedia.org/r/100707 (owner: 10Aaron Schulz) [00:30:18] ori-l ^^ [00:30:37] gwicke: mw.loader.store.useFunction is public; I'll log it as part of the event [00:30:40] thanks, btw [00:31:37] (03PS1) 10Ori.livneh: Module storage: randomly choose between Function and $.globalEval [core] (wmf/1.23wmf5) - 10https://gerrit.wikimedia.org/r/100721 [00:32:01] ori-l: yw [00:32:41] (03PS1) 10Ori.livneh: Restore module storage logging for I7183778cb [extensions/WikimediaEvents] - 10https://gerrit.wikimedia.org/r/100722 [00:32:57] (03Merged) 10jenkins-bot: Module storage: randomly choose between Function and $.globalEval [core] - 10https://gerrit.wikimedia.org/r/100542 (owner: 10Ori.livneh) [00:33:38] (03PS1) 10MaxSem: Allow customizing classes inside of diffs [core] - 10https://gerrit.wikimedia.org/r/100723 [00:34:42] (03Merged) 10jenkins-bot: Added --wikidb param to sql.php [core] - 10https://gerrit.wikimedia.org/r/100707 (owner: 10Aaron Schulz) [00:48:58] legoktm, what version of wikitools is your script built for? [00:49:10] the latest one [00:49:14] I'm getting 'APIRequest instance has no attribute '__getitem__' when I try to iterate through the URLs. [00:49:21] o.O [00:49:22] https://github.com/alexz-enwp/wikitools [00:49:33] Yeah, just cloned the latest. [00:49:37] oh [00:49:50] i see [00:50:18] superm401: https://gist.github.com/legoktm/7781996/revisions [00:51:25] Thanks, that gets it farther. [00:55:26] (03CR) 10Krinkle: Module storage: randomly choose between Function and $.globalEval (032 comments) [core] - 10https://gerrit.wikimedia.org/r/100542 (owner: 10Ori.livneh) [00:56:21] Krinkle: I'm going to revert it anyway after it spends an hour or two in prod [00:56:41] (03PS1) 10MegaAlex: Add personal tools messages for messages login, userlogout, createaccount [core] - 10https://gerrit.wikimedia.org/r/100724 [00:57:47] ori-l: It's in master, new Function() syntax is easy to get wrong, it affects an integral part of module execution, can affect lots of users, it should have tests. So that someone else can verify it. [00:58:26] (next time anyway) [00:59:16] And I would somewhat virtually frown on gabriel merging it on the author's sayso for such a central thing in resourceloader. I'm busy and don't care right now, but just saying :) [00:59:20] (03PS1) 10Ori.livneh: Revert "Module storage: randomly choose between Function and $.globalEval" [core] - 10https://gerrit.wikimedia.org/r/100725 [00:59:37] marktraceur: wanna talk now? [00:59:54] Krinkle: the intent was to deploy it today and revert it from both the production branch and master today [01:00:03] it turns out I can't do the former, so I'l revert it from master [01:00:08] I think you're overreacting a little [01:00:31] (03PS2) 10Ori.livneh: Revert "Module storage: randomly choose between Function and $.globalEval" [core] - 10https://gerrit.wikimedia.org/r/100725 [01:00:35] I'm not, I've seen this go sideways at least 3 times over the last 3 years. one of which was by myself. [01:00:37] (03CR) 10Ori.livneh: [C: 032] Revert "Module storage: randomly choose between Function and $.globalEval" [core] - 10https://gerrit.wikimedia.org/r/100725 (owner: 10Ori.livneh) [01:01:31] It's a central piece, it can break stuff, it can easily have tests, it should have tests. I mean, we'd test it if we keep it. And we actually do have tests for it, which would randomly pass/fail depending on math.random literally. [01:01:55] and aside from the tests, the way it was merged. [01:02:10] Anyway, moving on. [01:03:24] (03PS1) 10Tim Starling: Send QueryPage recache queries to a separate query group [core] - 10https://gerrit.wikimedia.org/r/100727 [01:03:25] (03PS1) 10Tim Starling: Fix QueryPage transaction plan [core] - 10https://gerrit.wikimedia.org/r/100728 [01:05:54] (03Merged) 10jenkins-bot: Revert "Module storage: randomly choose between Function and $.globalEval" [core] - 10https://gerrit.wikimedia.org/r/100725 (owner: 10Ori.livneh) [01:10:11] DarTar: Sorry, James_F|Away dragged me off for coffee torture, free now? [01:10:26] hey, yes [01:10:38] wanna come over? [01:25:05] (03CR) 10Springle: [C: 031] Send QueryPage recache queries to a separate query group [core] - 10https://gerrit.wikimedia.org/r/100727 (owner: 10Tim Starling) [01:27:52] (03CR) 10Aaron Schulz: [C: 032] Send QueryPage recache queries to a separate query group [core] - 10https://gerrit.wikimedia.org/r/100727 (owner: 10Tim Starling) [01:30:54] (03Merged) 10jenkins-bot: Send QueryPage recache queries to a separate query group [core] - 10https://gerrit.wikimedia.org/r/100727 (owner: 10Tim Starling) [01:37:04] (03PS1) 10BryanDavis: Use select to chose wikis [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100730 [01:37:11] (03CR) 10jenkins-bot: [V: 04-1] Use select to chose wikis [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100730 (owner: 10BryanDavis) [01:38:52] (03PS2) 10BryanDavis: Use select to chose wikis [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100730 [01:39:09] (03CR) 10Springle: [C: 031] Fix QueryPage transaction plan [core] - 10https://gerrit.wikimedia.org/r/100728 (owner: 10Tim Starling) [02:09:21] Krinkle: with which commit did the continuously-updating-mtime for mediawiki.language.init issue originate? was it https://gerrit.wikimedia.org/r/#/c/81884/ ? [02:09:48] ori-l: no, I don't think so [02:10:00] it predates that [02:10:09] languagedata module was the first module to use hash based versioning [02:10:17] everything else is derived from the way i implemented it there [02:10:28] except that it didn't include context properly as we've learned :) [02:10:51] can you help me track it down? [02:11:00] https://github.com/wikimedia/mediawiki-core/blame/master/includes/resourceloader/ResourceLoaderLanguageDataModule.php [02:11:15] checking [02:11:18] yeah, that's what I'm looking at [02:11:27] https://github.com/wikimedia/mediawiki-core/commits/master/includes/resourceloader/ResourceLoaderLanguageDataModule.php [02:11:53] https://github.com/wikimedia/mediawiki-core/blame/6c278b6d7eb3c451/includes/resourceloader/ResourceLoaderLanguageDataModule.php [02:11:59] https://github.com/wikimedia/mediawiki-core/commit/931c31c7bddc2781702fb9ca3ea2626681231a78 [02:12:07] line 57 [02:12:32] $key = wfMemcKey( 'resourceloader', 'langdatamodule', 'changeinfo' ); [02:13:25] so we've been systematically busting the RL cache for two years straight, wow [02:13:39] well, only for that one module. [02:14:00] yeah, but it was updating continuously, so its mtime was the max mtime [02:14:29] Which means we were also regenerating the startup module excessively [02:14:30] so it was that module's mtime that would most often end up as the suffix of multi-module urls [02:14:31] yes, but because of the strong caching it was basically just another startup module [02:14:34] and a cheap version at that [02:14:37] ANd a whole lot of other cache busting [02:14:48] Because as Ori said it gets combined in with a lot of other stuff [02:15:20] so every 5-10 minutes the first person who hits startup gets the new timestamp, that response is then server-side and client-side caches. Then when the first client requests the new module queue with that new max timestamp, apache regenerates the langdata module and composes the rest from minification cache. [02:15:39] True [02:15:40] so that's (aside from race conditions) 2 cache misses every 5 minutes, instead of 1. [02:15:48] (1 being the startup module) [02:16:00] more client-side of course [02:16:08] The minification cache composition is not free [02:16:18] But it's not super expensive either [02:16:28] client-side impact has been pretty substantial: http://ganglia.wikimedia.org/latest/graph.php?r=month&z=xlarge&hreg[]=client-side&mreg[]=%5Ebrowser.pageSpeed.%28desktop%7Cmobile%29_median%24>ype=line&title=pageSpeed%3A+domInteractive+to+loadEventEnd&aggregate=1 [02:16:29] Ori is right that it's shockingly wasteful for a shockingly long period of time [02:16:34] Yeah, but it only happens once for each 5-minute loop. After that its varnished [02:16:44] ori-l: Wow that's good [02:16:58] even clearer here: http://ganglia.wikimedia.org/latest/graph.php?r=year&z=xlarge&hreg[]=client-side&mreg[]=%5Ebrowser.pageSpeed.%28desktop%7Cmobile%29_median%24>ype=line&title=pageSpeed%3A+domInteractive+to+loadEventEnd&aggregate=1 [02:17:04] ori-l: Yeah, the impact client-side is mostly in bandwidth (not in number of requests) because that cache miss now results in a request for 1 module instead of the entire bottom load queue. [02:17:18] See that's why we have engineers like you ori-l [02:17:21] To fix mistakes we made before you even showed up [02:17:28] Timo fixed it, mind you [02:17:34] and find other random crap like that that ends up hurting things [02:17:37] That's true [02:17:45] no, you fixed that big dent in bandwidth with localStorage [02:17:47] But it would've gone undiscovered for another two years if it wasn't for you [02:17:53] (03PS1) 10Legoktm: Update README [extensions/MediaWikiAuth] - 10https://gerrit.wikimedia.org/r/100734 [02:18:04] OK, I'll take it :P [02:18:14] I then fixed the underlying issue that saved the remaining bit (lang data module), which, if it werent for localStorage, would've made the same dent in the graphs.. [02:18:33] I think I might've been the one to find which module was the offending one? :) But you were the one that started asking questions and got me to pay attention in the first place [02:18:34] well, we don't know, and that's frustrating [02:18:51] Yeah, what we also don't know is what dent lS would have made if it weren't for this bug [02:19:27] I can't parse that sentence. [02:19:53] (03CR) 10Legoktm: [C: 032 V: 032] Update README [extensions/MediaWikiAuth] - 10https://gerrit.wikimedia.org/r/100734 (owner: 10Legoktm) [02:19:58] the impact of module storage was (at minimum) exaggerated by the existence of that bug [02:20:11] We don't know how much of the impact of the module storage change was due to module storage itself vs due to the bug [02:20:20] (03CR) 10Legoktm: "(Was pulled from github: https://github.com/wikimedia/mediawiki-extensions-MediaWikiAuth/pull/1)" [extensions/MediaWikiAuth] - 10https://gerrit.wikimedia.org/r/100734 (owner: 10Legoktm) [02:20:24] So we don't know what the module storage deployment would have done on its own, [02:20:48] it could have been a performance regression, for all we know [02:20:52] nor do we know what just deploying a fix to this bug in the pre-localStorage era would have done [02:20:53] right, by primary design localstorage would've only saved 304 hits to the browser cache. the secondary gain (requesting fewer modules) was presumably anticipated, but should not've been this big, and now that that bug is fixed, is no longer. [02:21:06] Well we did run a trial and the data from that was good [02:21:28] yeah, but all the wikis we ran the trial on were affected by that bug, afaik [02:22:36] which makes the data essentially worthless :-/ [02:23:29] Right, right [02:23:43] so, this brings me back to: we should disable module storage and see if latency goes up [02:23:56] So, should we run a similar trial for disabling module storage for a small number of users? [02:24:36] i was thinking just disable it site-wide, but we can't, argh [02:25:16] well, no, we could [02:29:04] RoanKattouw: btw, the impact is visible on the bits varnishes too: http://ganglia.wikimedia.org/latest/graph.php?r=month&z=xlarge&c=Bits+caches+eqiad&m=cpu_report&s=by+name&mc=2&g=network_report [02:29:49] doesn't help resolve the ls / language module ambiguity but it's at least another confirmation that things really did improve for users [02:29:58] Yeah [02:30:02] Well that's good [02:36:37] what about disabling it on eswiki for 24h? [02:37:08] eswiki is pretty big so we'd get a good amount of data with the current 1:1000 nav timing sampling [02:37:22] and spanish speakers are geographically distributed [02:41:01] Could do that [02:41:13] Well Spanish speakers are but I don't know if users are [02:41:28] Given that there's this other Spanish-language on line encyclopedia, I forget the nae [02:42:39] Anyway, I gotta go [02:42:47] see ya [02:49:22] (03PS5) 10Shirayuki: i18n: consistency tweaks [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100564 [03:27:15] springle: thanks for your RT DB work [03:29:20] spagewmf: yw [03:35:54] (03PS1) 10BryanDavis: Sanitize csv export data [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100745 [03:38:58] (03PS1) 10BryanDavis: Validate application id [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100746 [03:48:59] (03PS1) 10BryanDavis: Do not store IP address [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100748 [04:01:27] ori-l: greg-g ping [04:01:36] does anyone want to sync https://gerrit.wikimedia.org/r/#/c/100701/ ? [04:01:50] dang. If someone was on Toolserver, where are they supposed to migrate to? [04:01:55] seems folks were too busy with other stuff earlier [04:02:12] Amgine: tools.wmflabs.org [04:02:23] thanks mutante [04:02:59] Amgine: yw, also see #wikimedia-labs [04:04:09] Well... keep in mind this is a wiktionarian. There's a fairly negative attitude re Labs in that (tiny) community due to past experiences. [04:04:52] labs has already managed to piss someone off? [04:05:14] "already"... [04:05:59] tool labs has only been around for a few months [04:06:26] yes, but wikimedia labs has been around longer. [04:06:48] ah [04:07:01] Inheritance is such a joy. [04:08:56] Amgine: i think we know each other from wiktionary:) [04:09:07] and i'll gladly help out with labs questions [04:09:23] Yes, but will you help Hippietrail? < grin > [04:09:31] hehee [04:09:46] what tools are you moving? [04:10:15] I don't know if xe is interested in moving xyr stuff. I'm trying to convince xyr. [04:13:41] Amgine: ever heard of talks to connect wikidata and wiktionary yet? [04:14:40] Yes. I failed to successfully communicate that there were costs to the community which might ... limit its implementation, and that I wanted to help reduce those costs. [04:14:42] legoktm, I've got an updated version of the script that seems to work: https://gist.github.com/mattflaschen/7904894 [04:14:53] Thanks very much for giving me somewhere to start. [04:15:01] awww, you renamed it :( [04:15:02] and np :) [04:16:11] if login_result != True: <-- erk. if login_result is not True [04:16:12] (03PS1) 10BryanDavis: Validate redirect destination on login [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100750 [04:16:31] Is there a difference, or is it style? [04:17:22] there is a technical difference I think, but in this case it doesn't matter [04:17:44] I think 'is' is object identity, but I don't think it matters for booleans. [04:17:53] its what PEP8 says :) [04:18:14] also, if you intend to have proper error checking you need to check for edit_result['edit']['result'] = 'Failure' since it won't throw an API error [04:18:17] ==* [04:18:44] [04:20:02] Amgine: hmmm.. i see, are people worried it would become omegawiki? [04:20:20] legoktm, can there be a failure without the 'error' key set? [04:20:26] yup! [04:20:42] Some, yes. But mostly it's a case of not seeing the value, without a large investment of nearly non-existent community coder resources. [04:20:50] action=edit is different from all the other API modules in that respect [04:23:29] legoktm, do you know where the error info generally goes for those cases ('Failure', but no 'error' key). [04:23:31] '*'? [04:24:06] not off the top of my head [04:24:12] lets see if its documented... [04:24:23] https://www.mediawiki.org/wiki/API:Edit#CAPTCHAs_and_extension_errors [04:24:44] assertedit also returns in that format, and I think AF might too [04:25:34] In theory, none of those should happen, so I'll just print the whole thing in that case. [04:45:18] (03CR) 10Aaron Schulz: [C: 032 V: 032] Avoid "Error: 1193 Unknown system variable table_type" error [extensions/WikimediaMaintenance] - 10https://gerrit.wikimedia.org/r/100715 (owner: 10Aaron Schulz) [04:57:54] (03PS6) 10Shirayuki: i18n: consistency tweaks [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100564 [05:00:54] mutante: if you want to attempt to seduce HT, xe is in #wiktionary at the moment. [05:15:16] (03PS8) 10Legoktm: Fire Echo event when revision is approved [extensions/FlaggedRevs] - 10https://gerrit.wikimedia.org/r/79775 (owner: 10Bartosz Dziewoński) [05:20:21] (03CR) 10Legoktm: "So when I tested it now it seems to work..." [extensions/FlaggedRevs] - 10https://gerrit.wikimedia.org/r/79775 (owner: 10Bartosz Dziewoński) [05:25:39] siebrand, around? could you +2 https://gerrit.wikimedia.org/r/#/c/100731/ [05:52:32] (03PS1) 10KartikMistry: phpcs fixes. [extensions/LocalisationUpdate] - 10https://gerrit.wikimedia.org/r/100754 [06:03:13] (03PS3) 10Yurik: Changed ZeroRatedMobileAccess to Zero [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/99158 [06:03:32] (03CR) 10Yurik: [C: 032] Changed ZeroRatedMobileAccess to Zero [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/99158 (owner: 10Yurik) [06:03:34] (03Merged) 10jenkins-bot: Changed ZeroRatedMobileAccess to Zero [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/99158 (owner: 10Yurik) [06:20:43] http://kingjamesprogramming.tumblr.com/ [06:20:46] this is the best thing ever [06:39:44] I have a really silly question about wikipedia usage if anyone's around... [06:40:09] nope, no one [06:40:19] oh, nevermind then. [06:40:21] :p [06:40:37] (03PS1) 10Santhosh: Correct the plural forms for Manx (Gaelg) [core] - 10https://gerrit.wikimedia.org/r/100757 [06:40:39] btw ori-l eventually I'll tackle you about vagrant on windows [06:41:35] but not right now -- so, if a user wrote a feedback complaint about something not working in a "private Sandbox" or "sandbox under the user namespace" -- is that only User:whatever/sandbox or is that valid to any page under the username namespace? [06:41:50] that is, is there any special difference to a /Sandbox page as opposed to any other subpage [06:43:39] The English Wikipedia adds a "Sandbox" link to the personal tools section. [06:43:47] Previously "My sandbox". [06:43:56] hm. There's a VE bug I'm trying to replicate [06:44:35] I'm upgrading to 10.8.5. [06:44:36] Allegedly. [06:44:42] well, a complaint about a bug. The user mentioned the 'private sandbox' and I was wondering if it has to be User:whatever/Sandbox or any subpage. [06:44:59] Sounds like user confusion. MediaWiki doesn't have private pages. [06:45:14] User page, not private page [06:45:32] You said private. [06:45:34] sorry, I'm trying to translate the complaint (it's in hebrew) [06:46:10] Indeed I did. I meant user page, the sandbox in the user namespace. [06:46:46] The distinction between a random user subpage and /sandbox (or /Sandbox) is that sandboxes have related custom JavaScript on some wikis. [06:47:18] I think the user meant somethign like this: https://en.wikipedia.org/w/index.php?title=User:Mooeypoo/Sandbox&veaction=edit [06:47:32] Probably. [06:47:33] and he's right, actually. Interesting. [06:48:10] Hah. Hmm. [07:26:56] (03PS13) 10Wikinaut: make SpecialChangePassword form text depending on login status of $wgUser [core] - 10https://gerrit.wikimedia.org/r/96651 [07:31:39] yurik: Not around at 06:25. We have dailies at 08:00 local time, or 23:00 PDT. [07:32:34] yurik: Reviewed and +2-ed [07:36:36] (03CR) 10Siebrand: [C: 032] "Nice work. This shows that even after Raimond and I reviewed the file and made over 80 updates, there still remain issues. Thanks for bein" [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100564 (owner: 10Shirayuki) [07:36:41] (03Merged) 10jenkins-bot: i18n: consistency tweaks [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100564 (owner: 10Shirayuki) [07:39:05] (03CR) 10Siebrand: [C: 04-1] phpcs fixes. (031 comment) [extensions/LocalisationUpdate] - 10https://gerrit.wikimedia.org/r/100754 (owner: 10KartikMistry) [07:39:39] (03CR) 10Siebrand: phpcs fixes. (031 comment) [extensions/LocalisationUpdate] - 10https://gerrit.wikimedia.org/r/100754 (owner: 10KartikMistry) [07:41:51] (03CR) 10Siebrand: "FYI: There are many more PHPCS warnings and errors in this extension. See https://dpaste.de/Fm2n." [extensions/LocalisationUpdate] - 10https://gerrit.wikimedia.org/r/100754 (owner: 10KartikMistry) [08:04:27] (03PS2) 10Siebrand: Fix CodeSniffer errors and warnings [extensions/LocalisationUpdate] - 10https://gerrit.wikimedia.org/r/100754 (owner: 10KartikMistry) [08:20:26] hello Niharika [08:26:13] (03CR) 10KartikMistry: [C: 032] "LGTM." [extensions/LocalisationUpdate] - 10https://gerrit.wikimedia.org/r/100754 (owner: 10KartikMistry) [08:26:15] (03Merged) 10jenkins-bot: Fix CodeSniffer errors and warnings [extensions/LocalisationUpdate] - 10https://gerrit.wikimedia.org/r/100754 (owner: 10KartikMistry) [08:55:55] Nemo_bis: Hello! :) [09:21:06] (03CR) 10Zfilipin: [C: 032] Add Translate-meta.wikimedia.org-linux-firefox [selenium] - 10https://gerrit.wikimedia.org/r/99054 (owner: 10Nikerabbit) [09:21:08] (03Merged) 10jenkins-bot: Add Translate-meta.wikimedia.org-linux-firefox [selenium] - 10https://gerrit.wikimedia.org/r/99054 (owner: 10Nikerabbit) [10:09:34] (03PS1) 10UltrasonicNXT: Add ability to show who has voted for a comment [extensions/Comments] - 10https://gerrit.wikimedia.org/r/100767 [10:21:34] MaxSem: thanks for the Geodata stacktrace fix ( https://bugzilla.wikimedia.org/show_bug.cgi?id=57450 ) [10:22:40] !logs [10:22:40] channel logs: http://bots.wmflabs.org/~wm-bot/logs/%23mediawiki/ html version: http://tools.wmflabs.org/wm-bot/logs/index.php?display=%23mediawiki [10:42:44] (03PS1) 10UltrasonicNXT: Allow the disabling of comment blocks [extensions/Comments] - 10https://gerrit.wikimedia.org/r/100770 [10:51:14] (03PS4) 10UltrasonicNXT: Allow the removal of various features [extensions/SocialProfile] - 10https://gerrit.wikimedia.org/r/97383 [10:52:39] (03CR) 10UltrasonicNXT: "Right, this has all got into a bit of a mess on my part, sorry about that :)" [extensions/SocialProfile] - 10https://gerrit.wikimedia.org/r/97383 (owner: 10UltrasonicNXT) [11:07:36] Hi everyone [11:27:02] Hi we have some problems on non-wikipedias, like it.wikivoyage.org (switched to 1.23wmf6 yesterday): now if I use @import url(...) in global common.css (or in my User:/common.css) it doesn't work anymore. Browser console returns a 404. Is it normal? [11:36:53] (03PS6) 10Matthias Mullie: Make header use *ActionPermissions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99416 [11:37:00] (03CR) 10jenkins-bot: [V: 04-1] Make header use *ActionPermissions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99416 (owner: 10Matthias Mullie) [11:38:11] (03PS7) 10Matthias Mullie: Make header use *ActionPermissions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99416 [11:40:01] (03CR) 10Matthias Mullie: [C: 032] "There was 2 times use Flow\Container; in Header.php, and lint wasn't too happy about that, apparently." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99416 (owner: 10Matthias Mullie) [11:40:11] (03Merged) 10jenkins-bot: Make header use *ActionPermissions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99416 (owner: 10Matthias Mullie) [11:42:15] (03CR) 10Matthias Mullie: "Can't be merged because it "depends" on a parent change that has not yet been merged." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98820 (owner: 10Matthias Mullie) [11:43:41] (03PS7) 10Matthias Mullie: Introduce tipsy [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98820 [11:44:04] (03CR) 10Matthias Mullie: [C: 032] Introduce tipsy [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98820 (owner: 10Matthias Mullie) [11:45:04] (03Merged) 10jenkins-bot: Introduce tipsy [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98820 (owner: 10Matthias Mullie) [11:48:47] I have found It stopped working with relative paths, with full path it works again. Thanks anyway, cya. [11:55:10] (03PS4) 10Matthias Mullie: Use addError() to add errors, instead of adding to array ourselves [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100387 [13:38:27] (03CR) 10Liangent: Create block log entries on HideUser blocks too (031 comment) [core] - 10https://gerrit.wikimedia.org/r/49234 (owner: 10Liangent) [13:44:05] (03PS5) 10Liangent: Create block log entries on HideUser blocks too [core] - 10https://gerrit.wikimedia.org/r/49234 [13:44:37] (03PS6) 10Liangent: Create block log entries on HideUser blocks too [core] - 10https://gerrit.wikimedia.org/r/49234 [14:26:40] (03PS1) 10Shirayuki: i18n tweaks [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100783 [14:30:18] (03PS1) 10Siebrand: Enable voting phpcs-strict-HEAD for LocalisationUpdate [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100785 [14:30:24] (03CR) 10jenkins-bot: [V: 04-1] Enable voting phpcs-strict-HEAD for LocalisationUpdate [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100785 (owner: 10Siebrand) [14:32:14] (03CR) 10Siebrand: "Am I missing something? Please let me know what else needs doing." [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100785 (owner: 10Siebrand) [14:34:28] (03PS1) 10Dan-nl: job-delay [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 [14:40:18] (03CR) 10Hashar: job-delay (034 comments) [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 (owner: 10Dan-nl) [14:43:27] (03PS5) 10Zfilipin: WIP Make it possible to run tests on Cloudbees using PhantomJS [selenium] - 10https://gerrit.wikimedia.org/r/100375 [14:44:08] (03CR) 10Zfilipin: "The first job on cloudbees: https://wmf.ci.cloudbees.com/view/fs/job/free-style-UniversalLanguageSelector-commons.wikimedia.beta.wmflabs.o" [selenium] - 10https://gerrit.wikimedia.org/r/100375 (owner: 10Zfilipin) [14:46:50] (03CR) 10Dan-nl: "addressing inline comments. will upload another patch in a moment." (034 comments) [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 (owner: 10Dan-nl) [14:53:26] (03PS2) 10Dan-nl: job-delay [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 [14:54:06] (03CR) 10Dan-nl: "addressing hashar‘s comment re:sanitized user params." [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 (owner: 10Dan-nl) [14:59:41] (03PS1) 10Matthias Mullie: Refactor header JavaScript [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100788 [15:00:05] (03CR) 10Bartosz Dziewoński: "I think the current linking behavior makes sense. Thanks!" (031 comment) [extensions/FlaggedRevs] - 10https://gerrit.wikimedia.org/r/79775 (owner: 10Bartosz Dziewoński) [15:00:38] (03CR) 10Bartosz Dziewoński: Site logo loaded using ResourceLoaderSiteModule instead of inline (031 comment) [core] - 10https://gerrit.wikimedia.org/r/98356 (owner: 10Tholam) [15:04:15] thedj: thedj[work]: https://gerrit.wikimedia.org/r/78640 could you have a final look? If you're ok with it, I'll +2 it [15:04:28] really want to have this in with the next code freeze tomorrow [15:16:51] (03PS1) 10Tpt: Fix yet another issue with old IE and Safari versions [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/100791 [15:18:37] (03CR) 10Hashar: [C: 031] "Sounds good feel free to merge." [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 (owner: 10Dan-nl) [15:19:23] (03CR) 10Gergő Tisza: [C: 031] job-delay [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 (owner: 10Dan-nl) [15:19:30] (03CR) 10Phe: [C: 031] "needed for IE <= 8 and safari 4.1/5.0 at least" [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/100791 (owner: 10Tpt) [15:21:20] YuviPanda: There? [15:21:59] (03CR) 10Bartosz Dziewoński: [C: 04-1] make SpecialChangePassword form text depending on login status of $wgUser [core] - 10https://gerrit.wikimedia.org/r/96651 (owner: 10Wikinaut) [15:28:32] (03CR) 10Tpt: [C: 032] Fix yet another issue with old IE and Safari versions [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/100791 (owner: 10Tpt) [15:28:42] (03Merged) 10jenkins-bot: Fix yet another issue with old IE and Safari versions [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/100791 (owner: 10Tpt) [15:36:51] hoo: ah, yes I'm good ! [15:38:04] :) [15:38:36] Good to here that graham liked it [15:38:42] hear [15:39:06] (03CR) 10Hoo man: [C: 032] "Worked fine during testing (tested by Graham and me). Also TheDJ is fine with the style changes I've amended." [core] - 10https://gerrit.wikimedia.org/r/78640 (owner: 10TheDJ) [15:39:12] Here you go :) [15:39:17] Another thing done \o/ [15:42:21] (03Merged) 10jenkins-bot: Preferences: Improve accessibility of the JS tabs of Preferences [core] - 10https://gerrit.wikimedia.org/r/78640 (owner: 10TheDJ) [15:45:30] hoo: thx for the ping [15:47:50] (03PS1) 10Tpt: Allows to keep user as null throw API edit [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/100793 [15:55:03] (03CR) 10Anomie: Clean up existing code and pass jshint (032 comments) [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/100709 (owner: 10Krinkle) [15:55:08] I can't login to my GCI account :/ [15:55:22] andre__: ^ [15:55:34] liangent, what did you try? [15:56:10] andre__: I pressed login on https://www.google-melange.com/gci/homepage/google/gci2013 , then chose my mail address and continue [15:56:23] then it replies http 500 [15:56:32] is it happening for everyone? [15:56:39] bawolff: Care to reply on https://gerrit.wikimedia.org/r/#/c/99738/ ? [15:57:54] (03CR) 10Phe: [C: 031] "Fine for me, it allow to api edit page with user="", there is a lot of such page, and protect against a corner case to disable changing th" [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/100793 (owner: 10Tpt) [15:58:12] (03CR) 10UltrasonicNXT: "Actually, on second thoughts, is this actually worth it? I won't abandon it yet, but wiki-wide CSS would do virtually the same job a littl" [extensions/Comments] - 10https://gerrit.wikimedia.org/r/100770 (owner: 10UltrasonicNXT) [15:58:38] anomie: What about making it so that &redirects=true causes file redirects (including foreign ones) to always be followed - then local and foreign redirects would be consistent [15:59:10] and have the redirect handling code of the api pass the final title to imageinfo, and have imageinfo act like any other module [15:59:12] ? [15:59:43] It seems weird to have file redirects behave so differently than normal redirects [16:00:13] bawolff: Because the redirect-handling code of the API doesn't know if there's a redirect in a ForeignDBRepo or ForeignAPIRepo repo. [16:00:18] ori-l: around? thoughts on https://bugzilla.wikimedia.org/show_bug.cgi?id=57567#c11 ? [16:00:45] anomie: But it could be modified to check for that if the page is in NS_FILE [16:01:12] (03CR) 10Tpt: [C: 032] "Reviewed by Phe" [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/100793 (owner: 10Tpt) [16:01:19] (03Merged) 10jenkins-bot: Allows to keep user as null throw API edit [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/100793 (owner: 10Tpt) [16:01:19] bawolff: Are you volunteering? ;) [16:01:39] I can, if you think this would be a good approach to take [16:03:12] Just make sure it properly (for whatever definition of "properly" matches the UI) handles weird cases like having a local page with no file and a redirect on the remote repo. And having a local redirect for a title where there's a file at that title in the foreign repo. And any other weird cases like that. [16:03:28] And also make sure it isn't horribly slow. [16:05:06] anomie: It could be a little slow if using ForeignAPIRepo - since that may require an http request to determine redirect status. Shouldn't be slow for someone with a setup like wmf's [16:05:36] bawolff: Make sure it's no slower than necessary, i.e. one query and not 500. [16:05:47] ok [16:09:24] bawolff: Am I correct in thinking that your change to the redirect handling in the API means that 99738 can be merged as more-sensible behavior for the keys from findFiles? [16:09:51] !gerrit 99738 [16:09:51] https://gerrit.wikimedia.org/r/#q,99738,n,z [16:10:36] my comment on that commit was more of an after thought, it wasn't supposed to block merging [16:11:28] Ok [16:11:47] (03PS1) 10Mayankmadan: Adding tests for collapsing and expanding sidebar sections [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100795 [16:11:54] (03CR) 10Brian Wolff: [C: 032] "This looks fine. I'll still look into changing the api, but that shouldn't block this change" [core] - 10https://gerrit.wikimedia.org/r/99738 (owner: 10Aaron Schulz) [16:12:22] (03CR) 10jenkins-bot: [V: 04-1] Tweaked FileRepo::findFiles() behavoir to match LocalRepo [core] - 10https://gerrit.wikimedia.org/r/99738 (owner: 10Aaron Schulz) [16:12:37] (03PS9) 10Legoktm: Fire Echo event when revision is approved [extensions/FlaggedRevs] - 10https://gerrit.wikimedia.org/r/79775 (owner: 10Bartosz Dziewoński) [16:17:13] bawolff: Are you on the API mailing list? I'm going to announce the fact that the imageinfo behavior with redirects is changing and might change again once your patch is ready, and if anyone replies it'd be good if you saw it. [16:17:36] I'm not generally on the api mailing list [16:18:04] but I'll pay attention to the archives for the time being [16:20:21] (03CR) 10Zfilipin: [C: 04-1] "please see inline comments" (032 comments) [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100795 (owner: 10Mayankmadan) [16:24:34] (03PS3) 10MaxSem: Add new one-column diff mode [php/wikidiff2] - 10https://gerrit.wikimedia.org/r/99541 [16:26:50] (03CR) 10Anomie: "I'm sad that this broke typing "vi incspSpecialFoo.php" to edit a special page. Now I'll have to get used to "vi incsp" [core] - 10https://gerrit.wikimedia.org/r/100394 (owner: 10Bartosz Dziewoński) [16:29:24] anomie: "( [16:29:25] :( [16:29:38] :/ [16:35:06] man, writing api modules is so boring [16:35:29] heh [16:35:30] what are you writing? [16:35:56] a module to call User#clearNotification for https://bugzilla.wikimedia.org/show_bug.cgi?id=51958 [16:36:02] i'll have like three lines of real code [16:36:09] 30 lines of error chekcing [16:36:14] 30 lines of boilerplate [16:36:23] and like a hundred lines of in-code documentation [16:36:52] there's just so many different errors that can happen when performing such a simple operation D: [16:38:08] MediaWIki PHP coding is no longer fun, it's like enterprise Java [16:38:10] :) [16:38:26] depends on the area [16:38:34] some are more like enterprise brainfuck [16:38:57] but some are somewhat okay [16:39:56] i definitely prefer JS though, less legacy junk [16:39:56] (03CR) 10BryanDavis: [C: 032] job-delay [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 (owner: 10Dan-nl) [16:40:02] (03Merged) 10jenkins-bot: job-delay [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 (owner: 10Dan-nl) [16:44:25] (03PS1) 10Hashar: make browsertest runner a macro [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100798 [16:44:31] (03CR) 10jenkins-bot: [V: 04-1] make browsertest runner a macro [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100798 (owner: 10Hashar) [16:45:05] (03PS2) 10Brian Wolff: Tweaked FileRepo::findFiles() behavoir to match LocalRepo [core] - 10https://gerrit.wikimedia.org/r/99738 (owner: 10Aaron Schulz) [16:48:40] (03CR) 10Brian Wolff: [C: 032] Tweaked FileRepo::findFiles() behavoir to match LocalRepo [core] - 10https://gerrit.wikimedia.org/r/99738 (owner: 10Aaron Schulz) [16:51:53] (03Merged) 10jenkins-bot: Tweaked FileRepo::findFiles() behavoir to match LocalRepo [core] - 10https://gerrit.wikimedia.org/r/99738 (owner: 10Aaron Schulz) [16:56:52] (03PS1) 10Hashar: (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 [16:56:56] (03CR) 10jenkins-bot: [V: 04-1] (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 (owner: 10Hashar) [16:58:02] (03PS1) 10Reedy: Revert "(bug 54951) Add Malayalam aliases for 828 and 829" [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/100801 [16:58:17] (03PS2) 10Reedy: Revert "(bug 54951) Add Malayalam aliases for 828 and 829" [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/100801 [16:58:51] (03PS1) 10Raimond Spekking: Add new message key to maintenace/messages.inc [core] - 10https://gerrit.wikimedia.org/r/100802 [16:59:05] (03CR) 10Raimond Spekking: [C: 032 V: 032] Add new message key to maintenace/messages.inc [core] - 10https://gerrit.wikimedia.org/r/100802 (owner: 10Raimond Spekking) [16:59:49] (03CR) 10Raimond Spekking: "New message key added to maintenace/messages.inc Ifed3c19dc1733c6454358f82ccd68dfd1f6c1e9a" [core] - 10https://gerrit.wikimedia.org/r/78640 (owner: 10TheDJ) [17:07:44] (03PS1) 10Raimond Spekking: Add message for 1.22 now that it is in Special:ExtensionDistributor [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/100803 [17:08:28] (03PS2) 10Hashar: (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 [17:08:29] ori-l: rh is too much struggle to start with mediawiki-vagrant ... going to use your stuff and adapt it later if I need to [17:08:52] (03CR) 10jenkins-bot: [V: 04-1] (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 (owner: 10Hashar) [17:12:05] (03PS2) 10Mayankmadan: Adding tests for collapsing and expanding sidebar sections [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100795 [17:13:35] what is the proper way to check if a given Revision belongs to a given Title? [17:14:11] $rev->getTitle()->equals( $myOtherTitle ); [17:14:19] I would guess anyhow [17:14:21] (03PS3) 10Hashar: (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 [17:14:25] (03CR) 10jenkins-bot: [V: 04-1] (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 (owner: 10Hashar) [17:16:02] will do i guess [17:18:16] (03PS1) 10Reedy: Revert "(bug 54951) Add Malayalam aliases for 828 and 829" [extensions/Scribunto] (wmf/1.23wmf5) - 10https://gerrit.wikimedia.org/r/100804 [17:18:34] (03PS1) 10Reedy: Revert "(bug 54951) Add Malayalam aliases for 828 and 829" [extensions/Scribunto] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100805 [17:23:21] anomie: ^ Could you review those and we can see about getting them deployed! [17:23:57] (03CR) 10Anomie: [C: 032] Revert "(bug 54951) Add Malayalam aliases for 828 and 829" [extensions/Scribunto] (wmf/1.23wmf5) - 10https://gerrit.wikimedia.org/r/100804 (owner: 10Reedy) [17:24:08] (03CR) 10Anomie: [C: 032] Revert "(bug 54951) Add Malayalam aliases for 828 and 829" [extensions/Scribunto] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100805 (owner: 10Reedy) [17:24:46] wtf is $wgRegisterInternalExternals?? [17:25:10] !wg RegisterInternalExternals [17:25:11] https://www.mediawiki.org/wiki/Manual:%24wgRegisterInternalExternals [17:25:43] (03Merged) 10jenkins-bot: Revert "(bug 54951) Add Malayalam aliases for 828 and 829" [extensions/Scribunto] (wmf/1.23wmf5) - 10https://gerrit.wikimedia.org/r/100804 (owner: 10Reedy) [17:25:50] lol, that's stupid. [17:25:59] MatmaRex: yay for meaningful names! [17:25:59] i bet someone had a chuckle [17:26:51] that was domas :P [17:26:55] (03CR) 10Zfilipin: Adding tests for collapsing and expanding sidebar sections (031 comment) [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100795 (owner: 10Mayankmadan) [17:27:06] <^d> Heh, I remember that one. [17:27:27] (03Merged) 10jenkins-bot: Revert "(bug 54951) Add Malayalam aliases for 828 and 829" [extensions/Scribunto] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100805 (owner: 10Reedy) [17:29:04] (03PS3) 10Mayankmadan: Adding tests for collapsing and expanding sidebar sections [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100795 [17:29:30] (03PS1) 10Reedy: Update Scribunto to fix ml namespace aliases [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100806 [17:30:30] (03PS1) 10Reedy: Update Scribunto to fix ml namespace aliases [core] (wmf/1.23wmf5) - 10https://gerrit.wikimedia.org/r/100807 [17:34:14] (03CR) 10Zfilipin: [C: 031] Adding tests for collapsing and expanding sidebar sections [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100795 (owner: 10Mayankmadan) [17:38:45] (03PS1) 10Cmcmahon: update for new Actions button and elements [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100808 [17:42:14] (03PS25) 10Mayankmadan: Added a test for downloading pdf from a random page [qa/browsertests] - 10https://gerrit.wikimedia.org/r/98160 [17:43:57] marktraceur: do we have an API method to look up media licenses from commons by providing the media filename? Or is that only going to happen when media metadata is on wikidata? [17:45:09] DarTar: Sort of yes. [17:45:25] We support some limited number of CC licenses based on tgr's work on CommonsMetadata [17:45:57] Maybe some PD licenses too, but I don't think anything but CMD supports those yet [17:46:03] (03CR) 10Zfilipin: "Looks like all new tests pass! :)" [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100795 (owner: 10Mayankmadan) [17:46:12] linky? [17:46:59] (03PS4) 10Cmcmahon: Adding tests for collapsing and expanding sidebar sections [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100795 (owner: 10Mayankmadan) [17:47:01] MatmaRex: we could just touch the module in all production branches to bump its version [17:47:08] DarTar: https://en.wikipedia.org/w/api.php?action=query&titles=File:MediaViewer_Sept9.png&prop=imageinfo&iiprop=extmetadata&format=json for example [17:47:13] MatmaRex: but what assurance do we have that the bug won't recur? [17:47:38] DarTar: License: { value: "cc-by-sa-3.0", source: "commons-templates" } [17:48:00] marktraceur: excellent, thanks [17:48:08] My pleasure [17:48:23] and (2) is there any way to resolve an image/media file deeplink to its canonical URL? eg: https://upload.wikimedia.org/wikipedia/commons/2/24/Loweswater_in_April_-_geograph.org.uk_-_268591.jpg -> https://commons.wikimedia.org/wiki/File:Loweswater_in_April_-_geograph.org.uk_-_268591.jpg [17:48:38] ori-l: see my last comment, we could try detecting mangled modules (there seems to be a clear pattern) or just blacklist opera [17:48:39] via an API method [17:49:23] (also, this seems to affect modules at random) [17:49:33] (not just one or anything like that) [17:49:50] (03CR) 10Zfilipin: [C: 031] update for new Actions button and elements (031 comment) [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100808 (owner: 10Cmcmahon) [17:50:02] how bizarre [17:50:06] DarTar: No, but mw.Title.newFromImg should do that on the client side [17:50:11] You can probably just steal that logic [17:50:48] right [17:50:48] thank you [17:50:52] ori-l: isn't it [17:51:00] opera has *the* weirdest bugs. [17:51:47] So it's basically the new IE? [17:52:06] have i already lamented how sometimes, a .clone()-d DOM node is an empty object, but only when you're not debugging, because then it is something else until you call a method on it? [17:52:38] DarTar: Quid pro quo - I'm seeing a bug in my local Limn instance where I can't save a new datasource, is this known? Specifically it mentions a circular object structure being saved into JSON. [17:52:58] or maybe a new document, not an element. something like that [17:53:18] ah [17:53:26] if ( !newDocument.documentElement || newDocument.documentElement.cloneNode() === undefined ) { [17:53:26] // Surprise! The document is not a document! Only happens on Opera. [17:53:26] // (Or its nodes are not actually nodes, while the document [17:53:26] // *is* a document. This only happens when debugging with Dragonfly.) [17:53:39] marktraceur: weird, I've never seen this [17:53:47] (another fun fact, the check will not actually trigger unless you place a breakpoint on this line in the debugger. [17:53:51] ) [17:53:54] DarTar: Maybe it's because I'm using CSV or somethhing [17:54:31] hang on, coming over to your desk [17:56:00] (03CR) 10Jhall: [C: 031] "Looks good - great work!" [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100795 (owner: 10Mayankmadan) [17:56:44] MatmaRex: http://my.opera.com/community/forums/topic.dml?id=172503 [17:57:13] (03CR) 10Jhall: [C: 031] Make it possible to run tests on Cloudbees using PhantomJS [selenium] - 10https://gerrit.wikimedia.org/r/100375 (owner: 10Zfilipin) [17:57:46] (03CR) 10Jhall: [C: 031] update for new Actions button and elements [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100808 (owner: 10Cmcmahon) [17:58:01] ori-l: woot. [17:58:14] ori-l: well, his examples don't work for me, but this *was* interminnent [17:58:19] everything else seems to match [17:58:36] hm, or maybe not [17:58:44] this is about "\nfunction" instead of "function" [17:58:53] we get "\n" instead of "function" [17:58:54] ori-l: ^ [18:01:51] hmm [18:01:58] gaaaah [18:02:08] i just wrote an api module only to learn it already exists under a different name [18:02:32] anomie|meeting: ApiSetNotificationTimestamp, why did you not document it *anywhere* :( [18:03:30] * bawolff bets it was documented on the api help page :P [18:04:01] it's not [18:04:04] its docs link to a 404 [18:06:29] dammit [18:07:06] MatmaRex: https://en.wikipedia.org/w/api.php?action=help&modules=setnotificationtimestamp is documentation ;) [18:07:19] MatmaRex: Mainly though, I never got around to it [18:07:39] anomie: https://www.mediawiki.org/wiki/API:SetNotificationTimestamp is a 404 :( [18:07:53] eh, that module is kinda weird, though [18:08:08] it doesn't call User#clearNotification or anything else sane, working on raw db instead :( [18:08:59] MatmaRex: User::clearNotification only does one at a time, meaning up to 5000 calls with 5000 individual DB writes. [18:09:51] anomie: tbh i don't see the use case for clearing more than one and less than all notifications, but there's no reason why it couldn't be extended [18:10:07] * anomie hopes MatmaRex will volunteer to fix this particular example of bug 1 [18:10:40] i can pastebin you the module i wrote so you can copy over the relevant keywords :P [18:10:48] 'mark as visited' is one [18:10:59] (03CR) 10CSteipp: [C: 031] Validate redirect destination on login [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100750 (owner: 10BryanDavis) [18:11:06] (03CR) 10Cmcmahon: [C: 032] "merging for the +1s and so the builds will pass" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100808 (owner: 10Cmcmahon) [18:11:26] anomie: i actually recently extended clearNotification to accept an $oldid parameter, by the way [18:11:28] (03CR) 10CSteipp: [C: 031] Do not store IP address [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100748 (owner: 10BryanDavis) [18:11:38] (03PS1) 10Yurik: Updated zero [core] (wmf/1.23wmf5) - 10https://gerrit.wikimedia.org/r/100813 [18:11:49] duplicating lots of your code i had no idea was there [18:11:55] (03CR) 10CSteipp: [C: 031] Validate application id [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100746 (owner: 10BryanDavis) [18:12:42] (03PS1) 10Yurik: Updated zero [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100814 [18:12:54] MatmaRex, hi [18:13:17] hi mayankmadan_ [18:13:19] (03CR) 10CSteipp: [C: 031] Sanitize csv export data [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100745 (owner: 10BryanDavis) [18:13:31] MatmaRex, any new tasks? [18:14:20] none right now [18:14:21] (03CR) 10Yurik: [C: 032 V: 032] Updated zero [core] (wmf/1.23wmf5) - 10https://gerrit.wikimedia.org/r/100813 (owner: 10Yurik) [18:15:00] hi I have a some problems with 1.23wmf6 and "@import url" in css. has anything changed? [18:15:27] a lot of things changed [18:16:11] anomie: i'll file a doco bug [18:16:41] in this page http://it.wikivoyage.org/wiki/MediaWiki:Common.css there are a lot of "import url" but now only the first one works [18:16:55] MatmaRex: Don't file it, just fix it! ;) [18:17:13] i've spent enough wasted time on this crap today [18:18:20] rotpunkt: wow [18:18:33] thanks hoo [18:19:15] rotpunkt: I mean: Those are a lot of imports... not that I really have time to look at the issue [18:19:43] rotpunkt: oh wow [18:19:46] why would you do that [18:19:57] it's not a script of mine [18:20:07] however until yesterday it worked [18:20:26] rotpunkt: just so you know, this destroys the loading speed benefits MediaWiki's ResourceLoader brings to scripts and gadgets [18:20:59] what can I suggest to the author for fixing the problem? [18:21:45] make these imported pages gadgets and make them default [18:21:49] (or just make them one bug gadget) [18:21:53] one big* [18:22:06] ok thanks! [18:22:13] rotpunkt: also, hmm [18:22:55] even though this is a bad idea, it should be working [18:23:50] the imports after the first one become "@import url(//it.wikivoyage.org/w/index.php);" [18:24:01] this is what the browser console logs [18:24:02] wtf really? [18:24:04] sure [18:24:04] We mangle them? [18:24:19] (03PS1) 10Krinkle: Enable mwext-TemplateData-jslint [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/100815 [18:24:21] hmmmmmmmm [18:24:34] Yup [18:24:35] maybe that's because of my cssmin changes [18:24:36] That's so weird [18:24:44] try to load https://it.wikivoyage.org/wiki/Pagina_principale?debug=true [18:25:02] i improved url() data uri @embed-ding recently [18:25:04] Not even CSSMin [18:25:10] It looks like it's in the .... yeah that [18:25:14] https://bits.wikimedia.org/it.wikivoyage.org/load.php?debug=true&lang=it&modules=site&only=styles&skin=vector&* [18:25:19] :( [18:25:20] Even that has mangled URLs [18:25:34] still [18:25:43] this should not happen [18:25:59] I would be nice :) [18:26:08] *it [18:26:20] !g If9082f553fa920c606f12093f39f4a163ebacc32 [18:26:20] https://gerrit.wikimedia.org/r/#q,If9082f553fa920c606f12093f39f4a163ebacc32,n,z [18:26:39] (03PS1) 10UltrasonicNXT: Add 'text' data for Special page links [core] - 10https://gerrit.wikimedia.org/r/100816 [18:26:53] actually, the question should be, why did it work before [18:27:36] or, why do i `return $file` at the end of CSSMin::remapOne() [18:27:49] (03CR) 10Catrope: "FIXME: CSS stylesheets with multiple successive @import url()s with query strings are getting mangled (the query string is removed from al" [core] - 10https://gerrit.wikimedia.org/r/94511 (owner: 10Bartosz Dziewoński) [18:28:19] RoanKattouw: file a bug, you silly you, not a fixme :( [18:28:25] Fine [18:28:31] (or just fix it) [18:28:41] :D [18:28:42] Probably better, yeah, since bugs can be tagged in commits [18:28:53] Well I don't know what causes it, and I have wmf6 bugs in VE to worry about [18:29:12] i was pretty sure i didn't change the behavior for external links there [18:30:55] (03CR) 10Catrope: "Filed as bug 58338" [core] - 10https://gerrit.wikimedia.org/r/94511 (owner: 10Bartosz Dziewoński) [18:32:53] ok thanks! I have to monitor bug 58338 is it correct? [18:34:16] rotpunkt: If you create an account on bugzilla, and all yourself to cc list for that bug, you'll get an email anytime someone updates the bug [18:35:55] ok thanks again [18:36:22] anomie: filed https://bugzilla.wikimedia.org/show_bug.cgi?id=58339 [18:41:16] (03CR) 10Ori.livneh: "Ping!" [extensions/TimedMediaHandler] - 10https://gerrit.wikimedia.org/r/99597 (owner: 10Ori.livneh) [18:46:21] (03CR) 10Umherirrender: [C: 031] Fix QueryPage transaction plan [core] - 10https://gerrit.wikimedia.org/r/100728 (owner: 10Tim Starling) [18:47:02] DarTar: Would it be possible to request node.js and npm on stat1? :D [18:47:26] (03PS1) 10Cmcmahon: add test for Topic actions to previous Post actions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100818 [18:47:45] * marktraceur is ever optimistic [18:50:02] (03CR) 10Brion VIBBER: "If there ever a need to customize the classes versus just removing them? I don't really like those as primary constructor parameters... I'" [core] - 10https://gerrit.wikimedia.org/r/100723 (owner: 10MaxSem) [18:51:10] (03CR) 10Yurik: [C: 032 V: 032] Updated zero [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100814 (owner: 10Yurik) [18:54:35] (03CR) 10Chad: [C: 032] "Actually, we'll just leave it off by default for now, I can see how it'd be confusing." [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/100392 (owner: 10Manybubbles) [18:54:40] (03PS2) 10Cmcmahon: add test for Topic actions to previous Post actions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100818 [18:54:47] (03Merged) 10jenkins-bot: Let Cirrus return the score [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/100392 (owner: 10Manybubbles) [19:01:00] (03PS1) 10Chad: Rename $wgCirrusSearchContentReplicaCount to $wgCirrusSearchReplicaCount [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/100821 [19:03:28] (03PS3) 10Aaron Schulz: Changed TIME_ONLY to NAME_AND_TIME_ONLY in FileRepo [core] - 10https://gerrit.wikimedia.org/r/100478 [19:04:13] (03CR) 10jenkins-bot: [V: 04-1] Changed TIME_ONLY to NAME_AND_TIME_ONLY in FileRepo [core] - 10https://gerrit.wikimedia.org/r/100478 (owner: 10Aaron Schulz) [19:07:01] (03CR) 10Mdale: "Is the idea that the outputHook in TimedMediaHandler_body.php will continue to output this module only where its needed?" [extensions/TimedMediaHandler] - 10https://gerrit.wikimedia.org/r/99597 (owner: 10Ori.livneh) [19:08:40] (03PS2) 10MaxSem: Allow customizing classes inside of diffs [core] - 10https://gerrit.wikimedia.org/r/100723 [19:08:41] brion, is ^^^ better?:) [19:08:45] RoanKattouw: debugged https://bugzilla.wikimedia.org/show_bug.cgi?id=58338 and commented. tl;dr not my fault :P (but i'm writing a patch) [19:08:49] looking [19:09:06] heh [19:09:10] MaxSem: that'll probably do :D [19:09:14] public properties ftw [19:10:10] (03CR) 10Brion VIBBER: [C: 032] "Still feels a little hacky to me, but the code using it will be clear and maintainable and that's what matters! :D" [core] - 10https://gerrit.wikimedia.org/r/100723 (owner: 10MaxSem) [19:10:20] lol nice [19:10:31] thanks brion [19:11:01] MatmaRex: Now I'm idly wondering if @import url(data:text/css,.....) works :D [19:13:10] (03PS1) 10Brion VIBBER: Blacklist Google Glass web browser from JS [core] - 10https://gerrit.wikimedia.org/r/100823 [19:13:55] (03Merged) 10jenkins-bot: Allow customizing classes inside of diffs [core] - 10https://gerrit.wikimedia.org/r/100723 (owner: 10MaxSem) [19:14:25] RoanKattouw: i see no reason why it shouldn't [19:15:03] i have the patch already, i'm coming up with creative tests now [19:15:23] https://gerrit.wikimedia.org/r/#/c/100824/ (draft) [19:15:55] (03CR) 10MaxSem: "Needs tests at tests/qunit/suites/resources/startup.test.js updated." [core] - 10https://gerrit.wikimedia.org/r/100823 (owner: 10Brion VIBBER) [19:16:15] :D [19:16:58] (03CR) 10Siebrand: (bug 54951) Add Malayalam aliases for 828 and 829 (031 comment) [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/95637 (owner: 10Odder) [19:17:05] MatmaRex: Maybe also add a test case for an @import url with a query string? It's probably a duplicate of one of the other cases in terms of functionality, but I'm a bit wary of fixing a bug then not putting the bug itself in the tests [19:17:26] okay, sounds reasonable [19:17:38] where's that stupid jenkins [19:19:06] (03PS1) 10Aaron Schulz: Avoid calling protected File field [core] - 10https://gerrit.wikimedia.org/r/100825 [19:25:17] (03PS2) 10Brion VIBBER: Blacklist Google Glass web browser from JS [core] - 10https://gerrit.wikimedia.org/r/100823 [19:31:18] (03PS4) 10Aaron Schulz: Changed TIME_ONLY to NAME_AND_TIME_ONLY in FileRepo [core] - 10https://gerrit.wikimedia.org/r/100478 [19:32:13] (03PS2) 10Matthias Mullie: Refactor header JavaScript [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100788 [19:32:15] (03PS6) 10Matthias Mullie: [WIP] Edit conflicts [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98836 [19:36:44] (03PS1) 10Matthias Mullie: Add callback when form & editor have completed loading [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100830 [19:38:26] (03PS1) 10Matthias Mullie: When initialized, editor should immediately take text size [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100831 [19:42:26] (03CR) 10EBernhardson: "In general i like this, but i was thinking perhaps an Error class:" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100387 (owner: 10Matthias Mullie) [19:44:53] (03CR) 10Anomie: [C: 04-1] "Found one bug in testing: RepoGroup::findFiles needs to pass the flag on to the subrepos' findFiles." [core] - 10https://gerrit.wikimedia.org/r/100478 (owner: 10Aaron Schulz) [19:49:44] (03PS1) 10Cmcmahon: update test for more IE robustfulness [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100832 [19:53:48] (03PS1) 10Yaron Koren: New version: 1.6.3 [extensions/ExternalData] - 10https://gerrit.wikimedia.org/r/100833 [19:55:20] (03CR) 10Anomie: "Both code paths appear to produce the same results. However, I note a change from the old version of the code: both the image redirect and" [extensions/GlobalUsage] - 10https://gerrit.wikimedia.org/r/99527 (owner: 10Aaron Schulz) [19:56:09] (03PS5) 10EBernhardson: [WIP] Add Flow events to Special:Contributions. [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99038 (owner: 10Werdna) [19:56:29] (03CR) 10EBernhardson: "PS5: cherry-picked onto current master" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99038 (owner: 10Werdna) [19:56:31] (03CR) 10Siebrand: "Thanks for the positive review, people. Is there a way to get this merged, too?" [extensions/EducationProgram] - 10https://gerrit.wikimedia.org/r/98567 (owner: 10Siebrand) [19:57:23] (03PS5) 10Aaron Schulz: Changed TIME_ONLY to NAME_AND_TIME_ONLY in FileRepo [core] - 10https://gerrit.wikimedia.org/r/100478 [19:58:01] (03CR) 10Siebrand: "Ping?" [core] - 10https://gerrit.wikimedia.org/r/96244 (owner: 10Siebrand) [19:58:33] (03PS7) 10Legoktm: Merge AssertEdit extension into core [core] - 10https://gerrit.wikimedia.org/r/90263 [19:59:15] (03CR) 10Legoktm: "PS7 was a rebase for the release notes using MatmaRex's rebase tool." [core] - 10https://gerrit.wikimedia.org/r/90263 (owner: 10Legoktm) [19:59:24] wheee [19:59:33] (03CR) 10Siebrand: "recheck" [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100785 (owner: 10Siebrand) [19:59:35] :D [19:59:38] it was so easy [20:01:31] it would be even easier if jenkins could do that as part of gate-and-submit [20:01:44] but we need to harness hashar to do that [20:01:50] (03CR) 10Anomie: [C: 032] Changed TIME_ONLY to NAME_AND_TIME_ONLY in FileRepo [core] - 10https://gerrit.wikimedia.org/r/100478 (owner: 10Aaron Schulz) [20:02:00] unless somebody else wants to poke at it [20:02:26] (03CR) 10Siebrand: [C: 04-1] Add message for 1.22 now that it is in Special:ExtensionDistributor (031 comment) [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/100803 (owner: 10Raimond Spekking) [20:05:13] (03CR) 10Siebrand: [C: 04-1] "Doesn't merge anymore, and needs release notes moved to 1.23." [core] - 10https://gerrit.wikimedia.org/r/87288 (owner: 10Liangent) [20:06:02] (03CR) 10Krinkle: [C: 032] "Still non-voting for now, but this should at least run." [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/100815 (owner: 10Krinkle) [20:06:26] (03Draft5) 10Bartosz Dziewoński: CSSMin: Fix remapOne() for URLs that are proto-relative or have query part [core] - 10https://gerrit.wikimedia.org/r/100824 [20:07:06] (03Merged) 10jenkins-bot: Enable mwext-TemplateData-jslint [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/100815 (owner: 10Krinkle) [20:07:48] (03Merged) 10jenkins-bot: Changed TIME_ONLY to NAME_AND_TIME_ONLY in FileRepo [core] - 10https://gerrit.wikimedia.org/r/100478 (owner: 10Aaron Schulz) [20:08:58] (03CR) 10Bartosz Dziewoński: CSSMin: Fix remapOne() for URLs that are proto-relative or have query part (033 comments) [core] - 10https://gerrit.wikimedia.org/r/100824 (owner: 10Bartosz Dziewoński) [20:09:14] James_F: https://gerrit.wikimedia.org/r/#/c/100824/ [20:11:50] (03PS3) 10Aaron Schulz: Use FileRepo::findFiles() to avoid a bunch of queries [extensions/GlobalUsage] - 10https://gerrit.wikimedia.org/r/99527 [20:12:20] MatmaRex: Thanks. [20:12:25] (03CR) 10Amire80: [C: 032] Add GENDER support for 'ep-articletable-remarticle-text' [extensions/EducationProgram] - 10https://gerrit.wikimedia.org/r/98567 (owner: 10Siebrand) [20:12:41] (03PS1) 10Robmoen: Configure ext.wikimediaEvents.ve to be compatible with mobile [extensions/WikimediaEvents] - 10https://gerrit.wikimedia.org/r/100835 [20:13:08] James_F: i assume "highest" priority means you're committing to finding someone to review that [20:13:20] and hopefully two more changes to that area i am preparing right now that make it suck less [20:13:28] MatmaRex: Yeah. [20:13:52] (03CR) 10jenkins-bot: [V: 04-1] Add GENDER support for 'ep-articletable-remarticle-text' [extensions/EducationProgram] - 10https://gerrit.wikimedia.org/r/98567 (owner: 10Siebrand) [20:14:03] (03CR) 10Anomie: [C: 032] Use FileRepo::findFiles() to avoid a bunch of queries [extensions/GlobalUsage] - 10https://gerrit.wikimedia.org/r/99527 (owner: 10Aaron Schulz) [20:14:05] (03Merged) 10jenkins-bot: Use FileRepo::findFiles() to avoid a bunch of queries [extensions/GlobalUsage] - 10https://gerrit.wikimedia.org/r/99527 (owner: 10Aaron Schulz) [20:16:07] (03CR) 10Jhall: [C: 032] update test for more IE robustfulness [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100832 (owner: 10Cmcmahon) [20:17:26] (03Merged) 10jenkins-bot: update test for more IE robustfulness [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100832 (owner: 10Cmcmahon) [20:19:20] (03PS2) 10Amire80: Add GENDER support for 'ep-articletable-remarticle-text' [extensions/EducationProgram] - 10https://gerrit.wikimedia.org/r/98567 (owner: 10Siebrand) [20:19:29] (03CR) 10Amire80: [C: 032] Add GENDER support for 'ep-articletable-remarticle-text' [extensions/EducationProgram] - 10https://gerrit.wikimedia.org/r/98567 (owner: 10Siebrand) [20:19:31] (03CR) 10jenkins-bot: [V: 04-1] Add GENDER support for 'ep-articletable-remarticle-text' [extensions/EducationProgram] - 10https://gerrit.wikimedia.org/r/98567 (owner: 10Siebrand) [20:19:43] (03CR) 10jenkins-bot: [V: 04-1] Add GENDER support for 'ep-articletable-remarticle-text' [extensions/EducationProgram] - 10https://gerrit.wikimedia.org/r/98567 (owner: 10Siebrand) [20:21:10] (03CR) 10Chad: Allow applicants to specify 3 wikis (031 comment) [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100517 (owner: 10BryanDavis) [20:21:31] (03CR) 10Jhall: [C: 031] add test for Topic actions to previous Post actions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100818 (owner: 10Cmcmahon) [20:22:35] (03CR) 10Chad: [C: 032] Validate redirect destination on login [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100750 (owner: 10BryanDavis) [20:23:18] (03Merged) 10jenkins-bot: Validate redirect destination on login [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100750 (owner: 10BryanDavis) [20:23:48] (03CR) 10Chad: [C: 032] Validate application id [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100746 (owner: 10BryanDavis) [20:24:05] (03Merged) 10jenkins-bot: Validate application id [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100746 (owner: 10BryanDavis) [20:24:14] (03CR) 10CSteipp: "This should use libxml_disable_entity_loader instead" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99164 (owner: 10EBernhardson) [20:24:22] (03CR) 10Yaron Koren: [C: 032 V: 031] New version: 1.6.3 [extensions/ExternalData] - 10https://gerrit.wikimedia.org/r/100833 (owner: 10Yaron Koren) [20:24:23] (03CR) 10Chad: [C: 032] Do not store IP address [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100748 (owner: 10BryanDavis) [20:24:39] (03CR) 10Yaron Koren: [V: 032] New version: 1.6.3 [extensions/ExternalData] - 10https://gerrit.wikimedia.org/r/100833 (owner: 10Yaron Koren) [20:25:28] (03Merged) 10jenkins-bot: Do not store IP address [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100748 (owner: 10BryanDavis) [20:25:47] (03Draft1) 10Bartosz Dziewoński: CSSMin: Change behavior for missing files [core] - 10https://gerrit.wikimedia.org/r/100837 [20:25:56] (03CR) 10Bartosz Dziewoński: CSSMin: Fix remapOne() for URLs that are proto-relative or have query part (031 comment) [core] - 10https://gerrit.wikimedia.org/r/100824 (owner: 10Bartosz Dziewoński) [20:28:25] (03CR) 10Chad: [C: 032] Sanitize csv export data [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100745 (owner: 10BryanDavis) [20:29:39] (03Merged) 10jenkins-bot: Sanitize csv export data [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100745 (owner: 10BryanDavis) [20:32:35] ori-l: re: https://saucelabs.com/tests/bc2c08bf827946f299f3476e319ae3f6 - I noticed the "Errors" tab count keeps increasing up to 950. Did you check those? Could be false positives about css or native deprecation warnings, but could be a clue. [20:32:53] Just making sure, since Opera 12 console looks like Chrome's but is different (errors don't always go the console tab) [20:35:17] also interesting: http://my.opera.com/community/forums/topic.dml?id=162114&t=1386794103&page=1#comment1907319 [20:35:28] normalisation is everwhere, I didn't expect to see it in function stringification though [20:35:46] (03PS1) 10Cmcmahon: check that login succeeds before navigating to targt page this should help with false failures in IEs [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100902 [20:36:48] (03PS2) 10Cmcmahon: check that login succeeds before navigating to target page this should help with false failures in IEs [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100902 [20:37:03] (03PS3) 10Cmcmahon: check that login succeeds before navigating to target page this should help with false failures in IEs [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100902 [20:40:49] I am trying to change an extension to not use $wgRequest etc. but I'm running into problems in the ArticleFromTitle hook. I need a Request instance (formely $wgRequest) but the hook does not provide Request or a context source ($article is not normally set). Is there any other way to get the Request object, or is there maybe another hook that can also change which Article is used? [20:45:06] !hook ArticleFromTitle | bawolff [20:45:06] bawolff: https://www.mediawiki.org/wiki/Manual:Hooks/ArticleFromTitle [20:45:39] (03PS2) 10Aude: Update Wikibase and DataValues [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100701 [20:46:18] (03CR) 10BryanDavis: Allow applicants to specify 3 wikis (031 comment) [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100517 (owner: 10BryanDavis) [20:46:28] (03PS3) 10Aude: Update Wikibase and DataValues [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100701 [20:46:53] leaves: Hmm, maybe that hook should be extended to have a $context object passed. For now I'd just leave it to use $wgRequest [20:47:26] Reedy: greg-g https://gerrit.wikimedia.org/r/#/c/100701/ [20:48:59] leaves: Although mostly that hook is meant for using a custom child class instead of Article [20:49:08] bawolff: passing a $context would be very useful, indeed. I do not see any reason not to (it's set to the Article right after the hook is called anyway). But I'm afraid there's not much choice right now to use $wgRequest. [20:49:48] (03CR) 10Jhall: [C: 032] check that login succeeds before navigating to target page this should help with false failures in IEs [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100902 (owner: 10Cmcmahon) [20:50:43] I think the idea is that that hook is supposed to inject a custom child class, and that child class of Article deals with anything needed from $wgRequest [20:51:24] (03Merged) 10jenkins-bot: check that login succeeds before navigating to target page this should help with false failures in IEs [qa/browsertests] - 10https://gerrit.wikimedia.org/r/100902 (owner: 10Cmcmahon) [20:52:01] bawolff: okay. I think I might take a look at FlaggedRevs (although my gut feeling tells me I shouldn't) how they do it there. [20:52:16] lol, yeah, flaggedrevs is scary [20:53:25] leaves: CategoryTree extension changes its behaviour based on a request variable, but does it later and not in that hook [20:55:40] bawolff: will also take a look there. Thank you :) [20:56:40] Hmm, maybe i should just add that variable to the hook, let people have maximun flexibility [20:57:23] (03CR) 10Raimond Spekking: Add message for 1.22 now that it is in Special:ExtensionDistributor (031 comment) [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/100803 (owner: 10Raimond Spekking) [21:00:01] (03CR) 10Hashar: job-delay (031 comment) [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/100787 (owner: 10Dan-nl) [21:00:22] (03PS1) 10EBernhardson: Disable external entity loading in libxml [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100904 [21:00:39] (03CR) 10EBernhardson: "Switching to libxml_disable_entity_loader in https://gerrit.wikimedia.org/r/100904" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99164 (owner: 10EBernhardson) [21:00:43] (03PS1) 10Brian Wolff: Add $context to the ArticleFromTitle hook [core] - 10https://gerrit.wikimedia.org/r/100906 [21:00:48] leaves --^ [21:02:22] James_F: i think wmf7 is getting branched today, so it'd be nice if you could make that priority really highest so that we have one less version to backport it to [21:04:03] bawolff: ? [21:04:30] leaves: Not that it helps you right now, but eventually https://gerrit.wikimedia.org/r/100906 would give you a $context from that hook [21:06:19] bawolff: Nice! [21:08:29] MatmaRex: It gets branched in about 20 hours' time. [21:10:49] James_F: alright, thanks [21:11:11] MatmaRex: (And yeah, this is probably worthy of a back-port anyway.) [21:12:49] (03CR) 10CSteipp: [C: 04-1] "You'll need to store and restore the old value. You can follow the example code on https://www.mediawiki.org/wiki/XML_External_Entity_Proc" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100904 (owner: 10EBernhardson) [21:15:41] (03CR) 10EBernhardson: [C: 04-1] [WIP] Add Flow events to Special:Contributions. (036 comments) [extensions/Flow] - 10https://gerrit.wikimedia.org/r/99038 (owner: 10Werdna) [21:15:48] (03PS2) 10Aaron Schulz: Set root job information for purge jobs to avoid excess work [extensions/GlobalUsage] - 10https://gerrit.wikimedia.org/r/100687 [21:17:33] (03CR) 10Siebrand: Add message for 1.22 now that it is in Special:ExtensionDistributor (031 comment) [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/100803 (owner: 10Raimond Spekking) [21:18:39] (03PS2) 10EBernhardson: Disable external entity loading in libxml [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100904 [21:19:05] (03CR) 10MaxSem: [C: 032] Adding convenience function isDisambiguationPage [extensions/Disambiguator] - 10https://gerrit.wikimedia.org/r/100413 (owner: 10Kaldari) [21:19:07] (03Merged) 10jenkins-bot: Adding convenience function isDisambiguationPage [extensions/Disambiguator] - 10https://gerrit.wikimedia.org/r/100413 (owner: 10Kaldari) [21:22:21] (03CR) 10Siebrand: [V: 032] "Ignoring unrelated Ci fail." [extensions/EducationProgram] - 10https://gerrit.wikimedia.org/r/98567 (owner: 10Siebrand) [21:22:37] (03CR) 10Siebrand: "Thanks, Amir." [extensions/EducationProgram] - 10https://gerrit.wikimedia.org/r/98567 (owner: 10Siebrand) [21:25:11] (03CR) 10Bartosz Dziewoński: CSSMin: Fix remapOne() for URLs that are proto-relative or have query part (031 comment) [core] - 10https://gerrit.wikimedia.org/r/100824 (owner: 10Bartosz Dziewoński) [21:26:21] (03CR) 10Anomie: [C: 032] Set root job information for purge jobs to avoid excess work [extensions/GlobalUsage] - 10https://gerrit.wikimedia.org/r/100687 (owner: 10Aaron Schulz) [21:26:23] (03Merged) 10jenkins-bot: Set root job information for purge jobs to avoid excess work [extensions/GlobalUsage] - 10https://gerrit.wikimedia.org/r/100687 (owner: 10Aaron Schulz) [21:27:02] (03Draft3) 10Bartosz Dziewoński: CSSMin: Correctly format 'url()' values with parentheses etc. [core] - 10https://gerrit.wikimedia.org/r/100907 [21:27:50] (03PS11) 10EBernhardson: [SCHEMA CHANGE] Shorten uuids to 88 bits. [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98995 [21:28:37] (03CR) 10jenkins-bot: [V: 04-1] [SCHEMA CHANGE] Shorten uuids to 88 bits. [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98995 (owner: 10EBernhardson) [21:29:22] (03CR) 10EBernhardson: "Updated patch removing the angry red lines and adding a BC hack for url anchor's with 128bit uuids." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98995 (owner: 10EBernhardson) [21:30:15] (03PS12) 10EBernhardson: [SCHEMA CHANGE] Shorten uuids to 88 bits. [extensions/Flow] - 10https://gerrit.wikimedia.org/r/98995 [21:41:33] (03CR) 10EBernhardson: [C: 04-1] When initialized, editor should immediately take text size (031 comment) [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100831 (owner: 10Matthias Mullie) [21:41:38] (03PS2) 10EBernhardson: When initialized, editor should immediately take text size [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100831 (owner: 10Matthias Mullie) [21:41:48] (03CR) 10EBernhardson: [C: 032] When initialized, editor should immediately take text size [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100831 (owner: 10Matthias Mullie) [21:42:10] (03Merged) 10jenkins-bot: When initialized, editor should immediately take text size [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100831 (owner: 10Matthias Mullie) [21:42:26] (03PS1) 10Spage: Added --wikidb param to sql.php (backport to wmf6) [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100909 [21:43:58] (03PS4) 10Hashar: (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 [21:44:33] (03CR) 10jenkins-bot: [V: 04-1] (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 (owner: 10Hashar) [21:51:36] (03CR) 10Leaves in Motion: [C: 04-1] "The .orig file seems to have been added by Gerrit Patch Uploader. I certainly did not include it in the diff." [extensions/CategoryTree] - 10https://gerrit.wikimedia.org/r/99649 (owner: 10Gerrit Patch Uploader) [22:03:35] hey Aaron|home, i've been running a 10169 and trying out that delayedJobsEnabled() method, but it doesn't seem to be working [22:04:16] Aaron|home: i'm wondering if the reason why is that i'm not properly adding the jobReleaseTimestamp … when you can please look at http://git.wikimedia.org/blob/mediawiki%2Fextensions%2FGWToolset.git/172ba5fb6308b989674e0c0a88edec06fbbcf3cc/includes%2FJobs%2FUploadMetadataJob.php#L101 [22:05:38] (03CR) 10Catrope: [C: 032] CSSMin: Fix remapOne() for URLs that are proto-relative or have query part [core] - 10https://gerrit.wikimedia.org/r/100824 (owner: 10Bartosz Dziewoński) [22:06:25] MatmaRex: I will backport that ---^^ to wmf6 today [22:07:17] * marktraceur cues ominous music [22:07:20] (03PS1) 10Brian Wolff: Always give sane output for Title::getUserCaseDBkey() [core] - 10https://gerrit.wikimedia.org/r/100917 [22:07:30] (03PS1) 10Jforrester: CSSMin: Fix remapOne() for URLs that are proto-relative or have query part [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100918 [22:07:36] ---^ That was 3 hours of debugging time for such a small issue [22:07:40] bd808, do you happen to know anything about the delayedJobsEnabled method and adding the jobReleaseTimestamp? [22:08:00] RoanKattouw: thanks. there's a couple followups, too, if you care / have time [22:08:15] MatmaRex: Yeah, reviewing those too, but they're cosmetic enough that I won't be backporting them [22:08:32] yeah, true [22:08:55] Reedy: bah left by mistake [22:08:57] (I did actually tell James to just put them all on his list, then changed my mind after reading the commit summaries) [22:08:58] superm401: ^^ :-D [22:09:55] Reedy, do you want a look at https://gerrit.wikimedia.org/r/#/c/100573/1 (routing fix for Beta)? [22:10:19] (03CR) 10Catrope: [C: 032] CSSMin: Change behavior for missing files [core] - 10https://gerrit.wikimedia.org/r/100837 (owner: 10Bartosz Dziewoński) [22:10:25] I am heading bed in a couple minutes so can't follow up [22:10:53] superm401: we have unit tests for that part of the code and the the preg_match() is using 'beta', so I don't think it is going to impact production [22:11:03] but definitely worth some other pair of eyes [22:11:19] (03CR) 10Bsitu: [C: 032] Added --wikidb param to sql.php (backport to wmf6) [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100909 (owner: 10Spage) [22:11:21] If not, I'll standardize the spacing, review (mainly checking the explicit list), then merge. [22:11:32] (03Merged) 10jenkins-bot: CSSMin: Fix remapOne() for URLs that are proto-relative or have query part [core] - 10https://gerrit.wikimedia.org/r/100824 (owner: 10Bartosz Dziewoński) [22:12:57] (03CR) 10Catrope: [C: 032] CSSMin: Correctly format 'url()' values with parentheses etc. [core] - 10https://gerrit.wikimedia.org/r/100907 (owner: 10Bartosz Dziewoński) [22:15:28] (03CR) 10Spage: [C: 032] "recheck" [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100909 (owner: 10Spage) [22:16:53] greg-g: LD today OK? MW core regression in wmf6, and some VE copy-and-paste things also? [22:18:12] (03Merged) 10jenkins-bot: CSSMin: Change behavior for missing files [core] - 10https://gerrit.wikimedia.org/r/100837 (owner: 10Bartosz Dziewoński) [22:18:16] (03CR) 10Chad: [C: 032] Allow applicants to specify 3 wikis [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100517 (owner: 10BryanDavis) [22:20:13] (03Merged) 10jenkins-bot: Added --wikidb param to sql.php (backport to wmf6) [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100909 (owner: 10Spage) [22:21:29] (03CR) 10Faidon Liambotis: [C: 031] "+1 for the intention and site impact; no opinion on the PHP/Javascript code itself." [extensions/WikimediaShopLink] - 10https://gerrit.wikimedia.org/r/98885 (owner: 10Ori.livneh) [22:23:06] (03PS1) 10Jdlrobson: WIP: Allow Skins to define other HTML tag attributes [core] - 10https://gerrit.wikimedia.org/r/100920 [22:23:53] (03Merged) 10jenkins-bot: CSSMin: Correctly format 'url()' values with parentheses etc. [core] - 10https://gerrit.wikimedia.org/r/100907 (owner: 10Bartosz Dziewoński) [22:25:10] (03Merged) 10jenkins-bot: Allow applicants to specify 3 wikis [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100517 (owner: 10BryanDavis) [22:26:16] superm401: heading bed myself sorry :/ [22:26:48] hashar, no problem. [22:27:03] thx! [22:31:14] dan-nl: Sorry, I was in a meeting. What are you trying to figure out? [22:31:44] np [22:32:08] bd808: http://git.wikimedia.org/blob/mediawiki%2Fextensions%2FGWToolset.git/172ba5fb6308b989674e0c0a88edec06fbbcf3cc/includes%2FJobs%2FUploadMetadataJob.php#L101 [22:32:29] i'm trying to sort out how to add the jobReleaseTimestamp [22:32:36] (03CR) 10Bartosz Dziewoński: "Fixed in https://gerrit.wikimedia.org/r/#/c/100824/" [core] - 10https://gerrit.wikimedia.org/r/94511 (owner: 10Bartosz Dziewoński) [22:33:16] i was hoping this piece of code would take care of it, but i can see that it's not. the last exception in beta cluster shows me that there is no jobReleaseTimestamp added as a param [22:33:47] so i'm guessing that means the delayedJobsEnabled test isn't doing what i expected it would [22:34:20] (03CR) 10Chad: Use select to chose wikis (031 comment) [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100730 (owner: 10BryanDavis) [22:34:33] Aaron|home: and i discussed the possibility you see on that code page, but also a test to see what the default job class is and use that as the test instead [22:35:42] it doesn't seem that there's a straightforward way to test if the job queue supports delayed jobs … at least not that i can feel atm [22:35:51] dan-nl: Ok. Let me read some code and see if I can figure anything out. [22:35:52] James_F: crazy multitasking and I'll forget, can you send an email or just edit wiki page? thanks [22:36:05] Maybe we can use eval on the beta cluster to find out? [22:36:23] that would be good, i'm not sure how to use that properly yet [22:36:27] it throws an exception if not support and there is the supportsDelayedJobs() method [22:37:06] Aaron|home: But beta should support the operation shouldn't it? It is using the redis job queue [22:37:11] Aaron|home: right, i'm trying that method, but it doesn't look like it's working [22:37:21] bd808: yes it is [22:37:30] greg-g: is there time available for https://gerrit.wikimedia.org/r/#/c/100701/ to be deployed? :) [22:37:35] but i may be adding the jobReleaseTimeStamp incorrectly [22:37:40] (03CR) 10Jdlrobson: [C: 031] Blacklist Google Glass web browser from JS [core] - 10https://gerrit.wikimedia.org/r/100823 (owner: 10Brion VIBBER) [22:37:41] whenever available.... [22:38:04] Aaron|home: here's that code link again …. http://git.wikimedia.org/blob/mediawiki%2Fextensions%2FGWToolset.git/172ba5fb6308b989674e0c0a88edec06fbbcf3cc/includes%2FJobs%2FUploadMetadataJob.php#L101 [22:38:16] or at least before deployments tomorrow [22:38:28] maybe i'm not adding the jobReleaseTimestamp properly? [22:38:44] what is the problem? [22:38:50] some error message? [22:38:54] greg-g: Yes. [22:38:59] (03CR) 10CSteipp: [C: 031] Disable external entity loading in libxml [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100904 (owner: 10EBernhardson) [22:39:29] there is no jobReleaseTimestamp as a param so i'm guessing that means the delayedJobsEnabled() method is not true [22:39:31] csteipp: ;) [22:40:05] dan-nl: did anyone enable delayed jobs in the job queue config for that queue type? [22:40:19] (03CR) 10Platonides: "I was probably thinking in the strange historical contents it had. Although HTML5 description seems saner http://www.whatwg.org/specs/web-" (031 comment) [core] - 10https://gerrit.wikimedia.org/r/53529 (owner: 10devunt) [22:40:22] have no idea … it's not automatic with redis? [22:41:02] it adds some slight overhead to the runners, so it is not enabled by default, only for queues that make use of it, like the Echo notification queue [22:41:16] Ah ha [22:41:26] * Aaron|home looks [22:41:42] $wgJobTypeConf['MWEchoNotificationEmailBundleJob'] = array( 'checkDelay' => true ) + $wgJobTypeConf['default']; [22:41:59] a similar line in CommonSettings will do the trick for this job type [22:42:14] thanks! will work on that now [22:46:05] we probably could just enable it for all queues...it's just that every so often one job runner gets appointed to making sure all stale and delayed jobs are brought into the main queue in one big serial loop (950 wikis * ~10 queues). The queues that tend to be active are handled in a much more distributed manor...this loop is only for wikis with low activity. [22:46:41] that could be better coded, though it still only takes a matter of seconds though [22:46:43] (03CR) 10Bartosz Dziewoński: [C: 031] "Looks good at a glance, but I'll let someone else double-check first :)" [core] - 10https://gerrit.wikimedia.org/r/100724 (owner: 10MegaAlex) [22:47:04] * Aaron|home really needs to rewrite most of the job runner code... [22:47:34] (03CR) 10BryanDavis: Use select to chose wikis (031 comment) [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100730 (owner: 10BryanDavis) [22:48:05] (03PS1) 10Brian Wolff: [WIP] Make API follow file redirects like normal redirects [core] - 10https://gerrit.wikimedia.org/r/100921 [22:49:30] Aaron|home: i don't mind adding the config, just needed to know about it. if it saves on performance then i'd say keep it this way … but i don't know enough about it yet [22:50:17] $wgJobTypeConf['gwtoolsetUploadMetadataJob'] = array( 'checkDelay' => true ) + $wgJobTypeConf['default']; [22:51:04] (03CR) 10Jforrester: "Scheduled to be deployed in ~ 70 minutes' time in a lightning deploy." [core] (wmf/1.23wmf6) - 10https://gerrit.wikimedia.org/r/100918 (owner: 10Jforrester) [22:51:18] thanks that's perfect [22:51:32] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Make API follow file redirects like normal redirects [core] - 10https://gerrit.wikimedia.org/r/100921 (owner: 10Brian Wolff) [22:51:48] of course... [22:53:00] (03CR) 10Chad: Use select to chose wikis (031 comment) [wikimedia/wikimania-scholarships] - 10https://gerrit.wikimedia.org/r/100730 (owner: 10BryanDavis) [22:54:57] hey bd808 just added you to that commit https://gerrit.wikimedia.org/r/#/c/100922/1 [22:55:33] and Aaron|home if you also want to approve it [22:56:13] thanks bd808 [22:56:16] dan-nl: Approved [22:56:47] dan-nl: Has greg-g talked to you about the prod deploy? [22:57:08] briefly … he wants to make sure we've cleared any issues .... [22:57:17] two remaining things i can think of atm [22:57:29] one has to do with https://bugzilla.wikimedia.org/show_bug.cgi?id=58299 [22:58:09] hasher and i followed the /tmp dir on the job runner instance. we discovered that sometimes orphaned files are left there but we don't know why [22:58:32] my guess is that a php script stops in mid process for some reason [22:58:56] That's possible. A segfault would keep temps from being cleaned up [22:59:04] he saw that the MimeMagic guessMimeType creates temp files in there for some tiff reader [22:59:15] bd808: hey, temps are people too [22:59:19] you can't just clean them up [22:59:19] and i uploadfromurl also uses it [22:59:24] segfault or no segfault [22:59:31] werdna: :) [22:59:53] we're just going to move you desk back into this corner [22:59:53] hasher and i were thinking of running a cron job that would clear out all URLxxxx files that are older than x hours, maybe 24? [23:00:29] bd808 would you be able to add that cronjob? [23:00:34] Having a clean up cron would be a good safety net. [23:01:07] I could submit the patch, but so could you. I don't have +2 on puppet things [23:01:16] hashar pointed to a config for it, but i can't find it atm [23:01:21] I do know people ot poke who do however [23:01:35] i don't know how to write good corncobs yet ;) [23:02:09] i'm sure if you submitted it and added hashar he'd correct it or merge it tomorrow [23:02:10] I think there was a patch to cleanup the tiff mess recently. I'll look for it [23:02:21] my god this spellchecker is a bit lunny [23:02:31] I figured. :) [23:03:54] (03PS1) 10Dr0ptp4kt: Dynamically generate HTML5 webapp manifests and cache manifests. [core] - 10https://gerrit.wikimedia.org/r/100924 [23:04:40] (03PS2) 10Robmoen: Configure ext.wikimediaEvents.ve to be compatible with mobile [extensions/WikimediaEvents] - 10https://gerrit.wikimedia.org/r/100835 [23:04:54] bd808: i think this it https://gerrit.wikimedia.org/r/#/c/100784/1 [23:05:15] (03PS2) 10Brian Wolff: [WIP] Make API follow file redirects like normal redirects [core] - 10https://gerrit.wikimedia.org/r/100921 [23:06:30] dan-nl: Sure. We'd want something like the cleantmpphp command but for URL* [23:07:11] right [23:07:29] shall i ask hasher to create the job tomorrow? [23:07:40] the same would need to be added to production i imagine? [23:07:57] who approves those cronjobs? [23:08:25] I'll put in the patch. It will apply to labs and production both [23:09:02] cool, i guess if you put in bug 58299 it will be documented as well? [23:09:39] dan-nl: Yes. Then if we find and squash the bug we can get rid of the cleanup job [23:12:27] who's working on Flow on beta? [23:12:54] dan-nl: "The Flow team"? Not sure. [23:13:09] Exception from line 188 of /data/project/apache/common-local/php-master/includes/db/LBFactory_Multi.php: LBFactory_Multi::newExternalLB: Unknown cluster "extension1" [23:13:44] #wikimedia-corefeatures [23:13:56] dan-nl: Be careful. You'll get hired as a QA engineer [23:14:03] uh oh [23:14:05] (03PS2) 10Dr0ptp4kt: WIP: Allow Skins to define other HTML tag attributes [core] - 10https://gerrit.wikimedia.org/r/100920 (owner: 10Jdlrobson) [23:14:46] https://www.mediawiki.org/wiki/Flow_Portal/Team and the channel that legoktm mentioned [23:15:01] k, just added the comments there … thanks [23:15:16] I already copied it into there :P [23:15:26] :) [23:18:42] (03CR) 10Daniel Friesen: WIP: Allow Skins to define other HTML tag attributes (031 comment) [core] - 10https://gerrit.wikimedia.org/r/100920 (owner: 10Jdlrobson) [23:18:46] bd808 the other issue i can think of is with hitting a glam server … what throttle would make sense so their server is not dos'd by the extension? [23:19:24] dan-nl: I suppose that depends on the GLAM and what they can support. [23:19:32] (03PS3) 10Dr0ptp4kt: WIP: Allow Skins to define other HTML tag attributes [core] - 10https://gerrit.wikimedia.org/r/100920 (owner: 10Jdlrobson) [23:19:58] That's not in your requirements document? :) [23:20:17] no it's not [23:21:17] You could do something like N seconds / MB I suppose [23:22:07] I'm not sure what throttles the job queue has available. [23:22:33] for now i'm relying on number of total jobs and then adding the delay … [23:23:03] Delay is between batches rihgt? [23:23:04] but it's arbitrary atm … the total could be 100 on one server and 900 on another [23:23:07] *right [23:23:29] yes [23:23:52] atm it checks for 3000 total media file jobs and delays 10 minutes before trying again [23:24:05] And the jobs in a batch will fire as fast as the workers can get to them [23:24:28] yes … that's what i'm concerned about atm [23:24:39] Soul you could end up downloading N files in parallel where N is the number of runners [23:25:21] so i like the idea of adding up the total mb instead or in addition [23:25:51] and then delay the adding of additional media file jobs if a mb threshold is reached [23:26:11] does that make sense [23:26:34] That gets tricky the way things are rigged I think. You'd need global storage that the jobs could read and then decide to requeue themselves with a delay or something. [23:27:01] i was thinking of passing it in the jobparams [23:27:22] that head curl that's done would be able to add the bytes from the header [23:27:46] and then the metadata batch job would keep track of the total [23:28:06] if would re-create itself with a delay when the threshold was reached [23:29:00] So you would fire N jobs in a batch and then delay the next batch based on how much you'd pulled in the prior batch? [23:29:11] i think that would work but i'd have to explore further ... [23:29:46] yes, something like that ... [23:29:48] The only thing you can't control for there is how many M of the N run in parallel [23:29:59] i see [23:30:29] Which would ideally be what you would want to throttle, but our queue may not be structured for that sort of thing at the moment [23:30:43] hmm ... [23:31:37] maybe i could just push N jobs out, and automatically delay the next N jobs by 5 minutes ... [23:33:22] Sustained load shouldn't really be the problem (one would hope). Burst load is waht kills [23:34:16] So the question is if all 16 (or whatever the real prod number is) downloads all hitting the same GLAM at the same time is too much or not [23:35:49] hmm ... [23:36:36] is there a way the job could detect an issue on their end? [23:36:52] The HTTP status code [23:37:13] you mean looking for anything other than 200 ? [23:37:39] Yes. And potentially specific codes that mean "slow down jerk" [23:37:57] ha! [23:38:05] which code is that? [23:38:20] 429 http://tools.ietf.org/html/rfc6585#page-3 [23:38:29] noticed that the rijksmusuem has been turing up with here was a problem during the HTTP request: 502 Proxy Error every once in awhile on a valid url [23:39:33] 502 usually means that a proxy can't reach the backend host. Which could indicate overloading [23:39:39] original URL: http://www.rijksmuseum.nl/media/assets/RP-T-1908-6 [23:39:39] evaluated URL: https://www.rijksmuseum.nl/mediabin.jsp?id=RP-T-1908-6 [23:39:55] yes, if i browse directly there's no issue ... [23:41:05] started the 10169 batch at 2013-12-11 20:18:16 utc, it's now 0:41 utc [23:41:22] and it's at about 9000 records [23:42:30] dan-nl: https://gerrit.wikimedia.org/r/#/c/100928 [23:42:31] the /tmp folder definitely has a few URLxxx orphans in there now [23:43:04] cool thanks! [23:43:57] i imagine that those 502's may be part of the issue [23:44:20] (03PS1) 10Bartosz Dziewoński: mw.loader.store: More fault tolerance [core] - 10https://gerrit.wikimedia.org/r/100930 [23:44:40] MatmaRex: wow, <3! [23:45:38] ori-l: o hi [23:45:54] ori-l: you know, logging how often these eval errors happen could provide for some interesting graphs [23:46:08] (but hopefully they will be some very boring ones) [23:46:59] who was https://www.mediawiki.org/wiki/Requests_for_comment/Services_and_narrow_interfaces ? [23:47:15] (03CR) 10Bartosz Dziewoński: "(I did not actually test this.)" [core] - 10https://gerrit.wikimedia.org/r/100930 (owner: 10Bartosz Dziewoński) [23:48:30] MatmaRex: https://gerrit.wikimedia.org/r/#/c/99547/ is on the road to that [23:50:18] MaxSem: Rewrite MW to be a SaaS stack? Good luck to that RFC [23:51:04] I guess SOA would be a better description than SaaS [23:52:33] Amazon did that change at one point. It made them better but took a huge investment [23:55:17] (03PS1) 10Tim Starling: Plural rules: updates for UTS #35 Rev 33 [core] - 10https://gerrit.wikimedia.org/r/100931 [23:59:53] (03PS1) 10Cmcmahon: Flow is on test2wiki now [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100932