[07:30:01] Hello [07:30:16] MJ94! [07:30:22] JD|cloud! [07:30:24] http://tree.opentreeoflife.org/ :D [07:30:31] Bsadowski1! [07:30:52] That would be a nice thing to have on... [07:30:54] well [07:30:55] Wait [07:32:45] Heh [07:32:50] Wikidata or Species? :P [07:32:56] http://opentreeoflife.org/ [08:19:11] hi [08:19:27] i submitted a patch for review for this task [08:19:46] https://phabricator.wikimedia.org/T112795 [08:19:55] can anyone check it out? [08:24:50] d3r1ck: link? [08:25:12] also, see https://www.mediawiki.org/wiki/Gerrit/Commit_message_guidelines#Phabricator [08:25:39] addshore: did you comment on the patch i sent? [08:25:48] i don't understand what youu mean by recheck [08:25:53] what is wrong [08:27:26] d3r1ck: "recheck" causes jenkins-bot to run the tests on the change! [08:28:23] When you submit a patch only "checks" are performed, not tests as you are not on the whitelist (see https://www.mediawiki.org/wiki/Continuous_integration/Zuul#Split_between_check_and_test) [08:29:55] ok [08:32:42] addshore: is it ok now? [08:34:02] addshore: you can +2 the correct one [08:34:15] +2ed now, it will be merged shortly [08:34:31] hah, that got a bit confusing, apparently we all went to edit the commit message at the same time..! :) [08:35:20] hahaha.... [08:35:25] addshore: it is fine now right? [08:35:30] yup [08:35:48] nice [08:35:57] addshore: this is my second bug i am fixing [08:35:58] :) [08:38:08] addshore: it has been merged [08:41:02] Woo! [08:49:45] addshore: :) [08:56:34] aude: https://gerrit.wikimedia.org/r/#/c/239783/ ;) [08:59:44] addshore! [08:59:44] :D [09:00:35] Bsadowski1! [09:52:45] addshore: \o/ [10:09:40] addshore: would you at all be interested in helping with https://phabricator.wikimedia.org/T113033 ? [10:16:47] Yes aude, I can probably take a look! [10:18:33] yay [10:18:47] i am trying to setup the extension and understand how it's supposed to work [11:21:30] aude: addshore Am I understanding correctly that we have not deployed this extension at all so far? [11:23:37] JeroenDeDauw: not yet [14:10:54] addshore: are you taking over maintenance of mw/core RecentChange stuff ? :-} [14:11:30] hah! no :) just writing some tests ;) (as I am touching it a little bit with the category watching sutff)! [14:12:21] addshore: beware that the code is being used for the broadcasting of RC changes to IRC [14:12:29] (well maybe not anymore to be honest) [14:12:42] and a ton of bots are hardcoded to parse a very specific format [14:12:49] but I think it is covered by tests [14:12:51] hehe, well, I'm trying to avoid touching anything like that ;) [14:12:55] yeah [14:12:59] don't put your fingers in those gears [14:13:00] :D [14:13:03] yeh, there are tests for some irc stuff [14:13:09] ideally we should drop support for irc.wikimedia.org [14:13:31] and migrate everyone to https://wikitech.wikimedia.org/wiki/RCStream [14:17:49] hashar: https://phabricator.wikimedia.org/T87780 ;) [14:21:07] addshore: yet another project which is bit rotting :-/ [14:22:25] addshore: got a straightforward patch for WikidataPagebanner for you :} https://gerrit.wikimedia.org/r/#/c/239841/ [14:22:40] marks some property as protected to please PHP CodeSniffer [14:24:21] hashar: you need to switch the visibility and the static keyword! [14:24:27] otherwise I would merge ;) [14:24:45] why doesn't code sniffer catch it ! [14:25:06] I dont know! :O that might not be in the MW cs stuff yet actually :? [14:25:09] but it should be! [14:25:17] It is in our ones ;) [14:25:28] yeah [14:25:40] addshore: better? https://gerrit.wikimedia.org/r/#/c/239841/3/includes/WikidataPageBanner.functions.php,unified [14:26:01] ya :) [14:50:54] aude: could you re +2 https://gerrit.wikimedia.org/r/#/c/239791/ ? :) [14:51:09] oh..... hashar just did it :D [14:51:41] :p [17:06:47] jzerebecki: can you check the teahouse thing again? now that i hassled krinkle about it, it seems to just work o_O [17:07:05] you can check for mw.util.teahouse.board in the js console. that should indicate whether the gadget got loaded [17:11:34] DanielK_WMDE_: Remember RL modules take <= 5min to propagate [17:12:01] and of course individual pages may be cached for <= 30 days and the load() call is in the html [17:16:46] Krinkle: the change was made more than 24 hours ago. More testing here indicates that the behavior is inconsistent: after a reload, sometimes the gadget is there, sometimes it isn't. [17:17:19] Krinkle: maybe something about LocalStorage? Different versions in different caches? [17:17:36] No [17:17:44] the object that should exist if the gadget is loaded is mw.util.teahouse.board [17:17:52] DanielK_WMDE_: OK. one more time, are you talking about the presence of the gadget or the version [17:18:00] E.g. ext.gadget.teahouse being loaded [17:18:14] Those live in completely different pages. [17:18:18] i'm talking about it being loaded for anons. [17:18:20] It being default means it will be added to the page. [17:18:27] at save or purge time [17:18:40] (for anons) [17:18:50] logged-in users don't have skin cache so it's always present [17:19:07] for anons it rolls over as article cache rolls over (either from edits/purges or from 30-day expiry) [17:19:11] ah ok - so if i'm viewing a page that wasn't purged since the gadget was enabled, i will not see the gadget? [17:19:21] Yeah, it's an html change. [17:19:25] effectively [17:19:55] If you want to bypass that, add mw.loader.load('ext.gadget.teahouse') to dewiki Common.js [17:20:00] And remove it after a few days [17:20:06] Krinkle: ok, now it makes sense. sorry for bugging you. is this documented somewhere? [17:20:20] Not particularly, it's part of how MediaWiki works in general. [17:20:28] i assume there's no way to fix this other than 2touching all pages, or waiting? [17:20:32] RL doesn't magically inject itself out-of-cache into every page [17:20:49] Workaround is to use a secondary resoruces to load it instead of inside the page [17:20:54] ^^ common.js [17:21:00] mw.loader.load("ext.gadget.teahouse") [17:21:47] keep |default of course, so that it rolls into pages [17:22:30] Krinkle: that line would cause it to load all associated resources? I don't know how gadgets integrate with RL [17:22:46] Define associated resources [17:22:56] Gadgets register as proper modules [17:23:27] Krinkle: the .js and .css "messages" [17:23:36] Gadget-teahouse/main.js, Gadget-teahouse/config.js, Gadget-teahouse/teahouse.css [17:23:49] Yeah, it creates a script/styles bundle like any other module. The origin or source doesn't matter. [17:24:00] ok cool [17:24:01] https://de.wikipedia.org/w/load.php?modules=ext.gadget.teahouse [17:24:16] Krinkle: how about generic code in common.js to load all default gadgets always? [17:24:25] that would cause too many requests, right? [17:24:26] [17:24:26] mw.loader.load(["mediawiki.action.view.postEdit","site","mediawiki.user","mediawiki.hidpi","mediawiki.page.ready","mediawiki.searchSuggest","mediawiki.page.watch.ajax","ext.cirrusSearch.loggingSchema","mmv.bootstrap.autostart","ext.echo.overlay.init","ext.eventLogging.subscriber","ext.wikimediaEvents","ext.wikimediaEvents.geoFeatures","ext.wikimediaEvents.lo [17:24:26] ggedin","ext.navigationTiming","ext.gadget.toolserver-integration","ext.gadget.WikiMiniAtlas","ext.gadget.OpenStreetMap","ext.gadget.old-movepage","ext.gadget.CommonsDirekt","ext.gadget.teahouse","ext.visualEditor.targetLoader","schema.UniversalLanguageSelector","ext.uls.eventlogger","ext.uls.interlanguage"]); [17:25:09] DanielK_WMDE_: Not really. You can think of load as an async require_once [17:25:20] It won't start loading the same thing twice [17:26:06] However I'd discourage this from Common.js in general as it isn't capable of doing conditionals that the server also does. [17:26:19] But for unconditional things during cache transition, seems sensible [17:26:35] how about changing the machnism so it's not bound to html generation at all? [17:26:49] have one js module that loads all the default gadgets? [17:26:57] That would delay things considerably [17:27:11] yea, too bad... [17:27:18] Anyway, thanks for clearing this up. I wasn't aware this was tied to HTML generation (yea, yea, *someone* implemented it like this at some point...). I assumed the gadgets now got loaded along with common.js via a JS module [17:27:22] And also, it's reliant on user and context. [17:27:31] I can disable the gadget [17:27:31] default gadgets arn't [17:27:35] Yes they are [17:27:42] or, true, i forget :D [17:27:57] that would work only for forced always-on gadgets [17:28:09] So you'll want to do in common.js a "if mw.user.options.get('' ... ') to make sure not to annoy users who choose to disable it [17:28:18] or just wrap it in mw.user.isAnon() and be safe :) [17:28:28] logged-in users don't need it anyway since they getit properly from the html diretly [17:30:03] DanielK_WMDE_: There are no forced gadgets [17:30:05] but there are forced modules [17:30:10] i know, i just ,ade them up [17:30:14] :) [17:30:17] anyway, thanks for the help :) [17:30:20] yw [17:30:36] enwiki uses this trick quite often when changing defaults [17:30:52] and we also used it during the VE roll-out initially. Although I fixed that for VE since then with a work-around I'll explain some other time. [17:38:40] sjoerddebruin: more context? [17:39:08] Your bot updated the new properties in the status update, right? It's a long time ago. [17:39:15] hrm [17:40:07] Traceback (most recent call last): [17:40:07] File "/data/project/legobot/wikidata/new_properties.py", line 11, in [17:40:07] import datetime [17:40:07] ImportError: No module named datetime [17:40:08] WTF [17:42:44] sjoerddebruin: something is messed up...I'm asking on -labs [17:52:55] ok [18:06:06] sjoerddebruin: someone changed the format of the page, so my bot couldn't find where to add the new entries in [18:06:40] sjoerddebruin: https://www.wikidata.org/w/index.php?title=Wikidata:Status_updates/Next&diff=next&oldid=240830968 [18:06:55] :'( [18:07:01] sigh. [18:07:30] https://github.com/legoktm/wikidata/blob/master/new_properties.py#L50 [18:07:54] I might as well fix this properly [18:17:05] sjoerddebruin: https://www.wikidata.org/w/index.php?title=Wikidata%3AStatus_updates%2FNext&type=revision&diff=253079698&oldid=253079649 [18:17:14] <3 [18:17:24] sjoerddebruin: as long as no one touches those comment markers, it should continue to work [18:17:29] Lydia_WMDE: ^^ fyi :) [18:29:42] SMalyshev: i have 2 patches signed up for tomorrows puppet swat. if you want I could also take care of yours, feel free to move the patch on the deployment page to me. [18:31:40] jzerebecki: sure, that would be excellent. I can still be present for at least first half hour but if we group it that would be great [18:32:58] jzerebecki: moved the patch to you [18:34:05] SMalyshev: ok. no need for you to also be there, its mostly waiting anyway. easy to test stuff. [18:34:24] jzerebecki: ok, thanks [18:34:30] yw [19:47:13] JeroenDeDauw: your phpmd things were merged! [20:14:24] addshore: yeah I just saw [20:14:29] Only took half a year [20:14:31] djeeez [20:27:29] * benestar waves [20:27:37] * JeroenDeDauw quickly hides