[02:50:18] Basic question: is it currently possible to geotag an item about an historical event with coordinates on Wikidata? [02:50:35] If so, where might I find a list of such items? [02:50:48] I don't think we have that datatype yet [02:52:32] Jasper_Deng: so we can't make http://www.ragtag.info/2011/feb/2/history-world-100-seconds/ with information from Wikidata? :P [02:52:42] PiRSquared: we're still waiting on the geocoordinate datatype [02:52:44] https://www.wikidata.org/wiki/Template:Autoarchive_resolved_section [02:52:47] wtf happened there? [02:53:01] Autoarchive_resolved_section? [02:53:04] PiRSquared: not yet! [02:53:14] the doc is all messed up when i look at it [02:54:32] legoktm: fixed. Now can you tell me when this will be possible? [02:54:41] he can't [02:54:48] not sure, its when the devs get it done [02:54:51] ^^ [02:57:50] legoktm: Before 2014? [02:58:48] hopefully :P [02:59:12] and thanks for the fix PiRSquared :D [03:03:57] np [07:25:01] New patchset: Daniel Kinzler; "Add overview of dispatch states to DispatchStats." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58464 [07:38:56] Reedy_: several users are reporting that their preferences were set back - was there any update or anything you think could have caused this? [07:39:10] they say it happened yesterday or the day before [07:39:25] New patchset: Daniel Kinzler; "Made Diff fully standalone from MediaWiki" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/57408 [07:46:59] can we access a running version of wikidata somewhere in a testing env, where we can actually submit changes safely? [07:51:55] Lydia_WMDE_: around? [07:51:59] tobyS: http://wikidata-test-repo.wikimedia.de/ [07:52:06] rschen7754: yes for a few more mins [07:52:08] wasup? [07:52:10] PM [07:52:12] k [07:52:20] lbenedix: thx :) [07:56:46] Change merged: Daniel Kinzler; [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/57408 [07:57:10] New patchset: Daniel Kinzler; "Improved docs and code style" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/57409 [08:00:11] Change merged: Daniel Kinzler; [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/57409 [08:00:22] DanielK_WMDE: what do you have in mind as an api module for dispatch stats? [08:02:48] New review: Daniel Kinzler; "As per comments on PS 14" [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/52796 [08:02:48] tobyS: thx for what? [08:03:14] lbenedix: sorry, wrong nick completed. thx was meant for Lydia_WMDE_. [08:03:23] ok [08:05:11] btw. setting up an own wikidata-repo is not very complicated [08:14:15] aude: something simple... action=wbdispatchstats would return structures equivalent to what you see on Special:DisplayStats [08:14:36] aude: could have options to get info for a specific client wiki too, and perhaps a mode for returning details for all client wikis [08:14:57] (though that would need paging, so it's non-trivial and probably won't be in the original version) [08:15:20] err, Special:DispatchStats [08:15:25] {{like}} [08:15:25] 04[1] 04https://www.wikidata.org/wiki/Template:like [08:15:30] * legoktm facepalms [08:15:32] (second time today, i should take a nap) [08:15:33] hmmm, ok [08:15:41] Thanks for working on it DanielK_WMDE and aude :) [08:15:57] aude: why, what do you have in mind? [08:16:10] * aude working on stats to log the edit rate, etc. [08:16:15] per wiki and on wikidata [08:16:17] legoktm: dealing with the dispatch lag is pretty much top prio right now [08:16:26] having the lag per wiki would be important [08:16:44] yes, but not necessarily all at once [08:16:51] edit rate... would be nice too# [08:16:54] glad to hear it :) [08:16:57] ? [08:17:07] edsu has an api you can pull to get the EPM for wikidata [08:17:12] aude: do you think we should rather have a more generic wbstats module, so it can be used for other stats too? [08:17:32] e.g. edit rate, number of items, etc? [08:17:45] DanielK_WMDE: no, it alreay provides edit rate, in some fashion [08:17:51] legoktm: yea, but scraping sucks :) [08:18:01] i want the lag per wiki [08:18:03] per client [08:18:24] true :P [08:18:28] aude: for a given client, no problem. for all... problem [08:18:42] or get a list of them all? [08:18:52] aude: that list needs paging [08:18:52] query list [08:18:58] doable, sure, but not a quick one-off [08:18:59] use the query module [08:19:09] uhhh [08:19:12] also, the results will not use the same notion of "now", so they may be inconsistent [08:19:25] the query modules lists pages [08:19:42] ...it doesn't really work for listing anything else, iirc [08:19:46] http://en.wikipedia.org/w/api.php?action=query&meta=siteinfo&siprop=statistics [08:19:51] it can list any sor tof thing [08:20:15] abuselog [08:20:18] hm... ok, i'll hgave to look at that again [08:20:20] whatever [08:20:32] * aude shall work on something else today [08:20:37] but still, way more complicated than the simple stats. [08:20:43] it would definitly be nice to have [08:20:43] job queue [08:20:49] but i'd call that a separate feature [08:21:02] aude: what doi you plan to do with the jq? [08:21:05] it would be useful to diagnose per wiki [08:21:16] see how to make the dispatcher work with it? [08:21:25] * aude could do something else entirely [08:21:51] don't want to overlap [08:21:55] aude: poke at it, but i think any solution needs diuscussion with aaron first (and i'd like to add my 2c too) [08:22:03] yes [08:22:35] there are several questions there... when an edit happens, does it create one job? or one per effected target wiki? [08:23:00] for dispatch stats, i'd incorporate it with query site info module [08:23:03] please do that [08:23:25] what if that job gets executed while the target is still "too fresh"? then the update would be lost [08:23:40] yep [08:23:48] unless we re-post the job, until the target is either touched by another process, or "stale enough". [08:24:20] if we could tell the jq to delay execution, that would avoid churn. maybe you can find out how to do that with the redis based jq [08:24:50] sure [08:25:10] "lost" updates would not happen on a wiki with many edits - the next job would just pick it up. [08:25:26] if there are not many edits, an external dispatcher process may be needed to avoid this [08:25:34] ...or the re-posting mechanism [08:25:38] anyway. [08:25:42] shall look at it [08:25:56] see what i can propose and what might work [08:26:10] for stats, can it be part of the site info queyr module? [08:26:32] i was wondering... posting 300 jobs to the queue for every edit seems a bit much. But posting 1 job means that when that job gets executed, it would have to dispatch to 300 wikis, making for an extremely slow job [08:26:46] hmm, wondering what a "snak" is? [08:26:47] or it needs to re-post jobs (chaining) [08:27:06] or additional meta option [08:28:05] tobyS: snak is tasty :) [08:28:19] it's a collection of property and data value [08:28:23] wouldn't that be a snack? [08:28:25] (in some cases value = no value) [08:28:30] tobyS: a word we made up to refer to a property-value-assignment (that may not actually assign a value) [08:28:37] or unknown value [08:28:39] tobyS: https://www.wikidata.org/wiki/Help:Glossary [08:28:56] tobyS: https://meta.wikimedia.org/wiki/Wikidata/Data_model#Snaks [08:29:17] thx :) [08:29:30] tobyS: a snak may say that a certain property has a certain value (for a given entity), or no value, or the value isn't known, or there's a range of values, etc [08:31:29] ok, so in terms of berlin, the attribute "founding date" could have a PropertyValueSnak like 1234, a PropertySomeValueSnak if nobody knows when it was founded? [08:32:21] * aude back in 30 minutes [08:33:12] tobyS: exactly. it may also have both, if one source claims the date, and anotherr source claims the date is unknown. [08:33:36] k, thx [08:33:55] (in that case, there would be two statements, one haveing hte PropertyValueSnak as it's main snak, the other having the PropertySomeValueSnak as its main snak) [08:43:11] hi [08:43:31] is there a way to get all Items that use a vertain property? [08:43:42] certain [08:44:48] yes [08:44:52] special:whatlinkshere [08:46:27] is it possible to export an item with all related properties? [08:47:50] Change merged: Tobias Gritschacher; [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/57912 [08:51:52] lbenedix: do you mean something like that: http://wikidata-test-repo.wikimedia.de/wiki/Special:EntityData/Q3.jsonfm [08:52:56] legoktm: thanks ^^ [08:53:01] np :) [08:53:52] Schisma_: you can use the whatlinkshere page on the property [08:54:10] err. way too late, sorry legoktm [08:54:18] :P [08:55:19] Tobi_WMDE: if this is importable into my testwiki: yes [08:56:07] Tobi_WMDE: https://wikidata.org/wiki/Special:EntityData/Q3.jsonfm doesn't work on the live site? [08:56:09] lbenedix: no, you need Special:Export for that. But we don't have a way to include all the property definitions there [08:56:30] lbenedix: Special:EntityData is still experimental/incomplete, so it's not live [08:56:33] hmpf [08:56:37] silly www, https://www.wikidata.org/wiki/Special:EntityData/Q3.jsonfm still doesnt work though [08:56:39] ah [08:56:40] ok [08:57:52] legoktm: importing entities is tricky, btw... i'm not sure it will work correctly. [08:58:07] you mean lbenedix ;) [08:58:07] entity IDs are a problem [08:58:14] yeah [08:58:25] legoktm: indeed [08:59:03] the test data (the chemical elements with some properties) are also not very helpful [08:59:17] New patchset: Jeroen De Dauw; "Work on DataType interface [DO NOT MERGE]" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/58472 [08:59:26] the data type of all properties is 'item' [09:00:19] make some new properties! [09:00:24] yay [09:00:36] creating test data by hand sucks [09:02:01] lbenedix: guess you can use wbeditentity and pass the json from Special:EntityData there. but haven't tried.. [09:02:10] are the {{#property}} syntax and relevant lua modules enabled on wikidata.org? [09:04:54] * lbenedix started the Wikidatameter : http://www.youtube.com/embed/9-UAURdr0kI [09:06:36] * lbenedix needs to add a clock for later evalutation [09:06:49] DanielK_WMDE: good to know ^^ [09:15:45] DanielK_WMDE: can you explain me why the class map is wrapped into that construct in DataModel/DataModel.classes.php? [09:17:01] tobyS: probably question for JeroenDeDauw [09:18:25] DanielK_WMDE: Hi! Is it possible to allow the use of the mw.wikibase library on wikidata.org? That will help us for the i18n of https://www.wikidata.org/wiki/Wikidata:List_of_properties [09:19:58] how many edits/min can wikidata.org handle? [09:22:54] New review: Daniel Kinzler; "(10 comments)" [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/51164 [09:24:33] tobyS: i don't have the slightest idea, it seems redundant to me. [09:24:48] yeah, same impression here :) [09:25:34] Tpt: running client and repo code on the same wiki is currently not supported... it *nearly* is, but there are a few nasty assumptions burried deep in the code. [09:25:45] Tpt: please file a bug about it, i don't think we have one yet :) [09:26:03] DanielK_WMDE: Thanks :-) [09:27:03] tobyS: we also have closures as hook handlers in some places... which makes the static context less obvious, but doesn't avoid it, as far as i can see. [09:27:46] tobyS: hm... one thing this does is it protecteds the $classes variable, by giving it a local scope to live in. otherwise, it may interfere with other stuff in the global scope. [09:27:49] tobyS: are you talking about the array being wrapped in a function? [09:29:23] tobyS: DanielK_WMDE: danielks last line sums up the reason why [09:29:36] This is a pattern very commonly used in modern JS, and seems to make sense here as well [09:31:44] DanielK_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=47071 [09:32:21] Tpt: thanks for that :) [09:32:55] legoktm: It's the easy part of the work ;-) [09:33:05] indeed :P [09:40:44] can you tell me how to add the new search, that is deployed on wikidata.org, to my testwiki? [09:41:58] and how to resolve the things on the help-pages I imported now? [09:46:28] translate extension for the latter [09:47:10] DanielK_WMDE, JeroenDeDauw: well, "return array( … );" would do the same trick, i think [09:48:05] dont see install instructions on mw.org/Extension:Translate [09:51:06] tobyS: it would yes, since there you would not (could not) create any extra vars - however here we "need" some extra vars [09:51:50] Don't think PHP has a higher level function that takes in an array and allows mapping it in a way where you can set the keys [09:52:04] *higher order function [09:53:08] that is true, yes [09:53:09] tobyS: oh oops, sorry, I was looking at the wrong file [09:53:31] tobyS: I was looking at Database/Database.classes.php [09:53:45] hehe :) [09:53:52] tobyS: it is true that right now the array could just be returned yes [09:54:09] so its a general pattern for the named issues, that is only used there for consistency, i guess [09:54:14] Though idially we would be able to get the stuff to look more like in the other file [09:54:39] Not really a patytern or convention [09:54:43] I just copied it [09:54:52] And then ended up not having any local vars in there any more [09:55:03] Could remove the wrapping, though I do not see how it hurts [09:57:54] tobyS: esp for the standalone code remarks on how the loading stuff is done are very welcome - it'd be good to follow generally accepted ways of doing things from the PHP community there [09:58:13] For instance it'd be good if the standalone libraries could easily be included somewhere using composer or whatever [09:58:48] That also includes loading of unit tests [09:59:27] Diff and Ask have a phpunit.php file to run their tests with, though I'm not sure if the approach used there is particularly good [10:00:59] well, then PSR-0 is the way to go [10:05:19] New patchset: Henning Snater; "Introducing toolbarbase widget" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58474 [10:05:19] New patchset: Henning Snater; "Renamed/Moved toolbar jQuery widgets" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58475 [10:05:35] New patchset: Tobias Gritschacher; "Fixing missing JS dependency" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58477 [10:07:26] Change merged: Henning Snater; [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58477 [10:24:29] can someone help me getting a working export of http://www.wikidata.org/wiki/Help:Contents ? [10:32:26] I think changing the help-link in the sidebar to http://www.wikidata.org/wiki/Help:Contents is enough [10:33:14] how do I get the working searchbar for my testwiki? [10:42:05] lbenedix: you might want http://www.wikidata.org/wiki/MediaWiki:Sidebar ? [10:43:34] nope [10:43:47] i want the search in the upper right corner of wikidata.org [10:44:17] the broken search on wikidata was replaced by something or got reparied [10:45:57] btw. can you file a bug regarding this language changing thing in the sidebar for me? [10:49:37] it should not be required to click the select button after selecting the language in the dropdown menu [10:50:17] and I found just another strange thing on wikidata.org the name of a property is different for two different items [10:52:19] http://lbenedix.monoceres.uberspace.de/screenshots/kemhl7ijxm_(2013-04-10_12.49.06).png [10:53:46] in german the Property has the same name for both items [10:57:38] tobyS: on the topic of PSR-0... If you look at the stuff in Wikibase\Query, it is not fully compliant as the actual classes are in Wikibase/Query/Query since there is registration code in Wikibase/Query. And the tests are certainly not compliant either. [10:57:48] If you have suggestions on how to improve that .. [10:59:26] New patchset: Jeroen De Dauw; "Wikibase/Query => Wikibase/QueryEngine" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58484 [10:59:49] New patchset: Tobias Gritschacher; "Fixed missing JS dependencies" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/58486 [11:00:08] New patchset: Jeroen De Dauw; "Wikibase/Query => Wikibase/QueryEngine" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58484 [11:00:12] JeroenDeDauw: improve what? [11:00:18] the different names? [11:00:26] or the select-language-thingy? [11:00:37] Change merged: Tobias Gritschacher; [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58484 [11:00:45] The directory and namespace structure of the Query component [11:00:53] New patchset: Henning Snater; "Added missing snakview dependency on snaklistview" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58487 [11:01:11] ahh.. you were not talking to me [11:01:33] Lydia_WMDE_: Ping [11:02:38] Change merged: Tobias Gritschacher; [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58487 [11:07:53] New patchset: Henning Snater; "Fixed missing JS dependencies" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/58486 [11:10:59] Change merged: Henning Snater; [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/58486 [11:15:15] am i correct that properties can arbitrarily be nested? like here the "testproperty"? http://wikidata-test-repo.wikimedia.de/wiki/Q3 [11:16:57] is there a numeric data type and if not, why? [11:32:14] DanielK_WMDE: saw my question? [11:47:18] lbenedix: pong [11:47:54] lbenedix: there is no numberic datatype yet [11:47:59] because it's not coded yet [11:50:32] okay [11:51:22] do you have defined test-cases for ui testing or when you demonstrate wikidata to people who never heared about it ? [11:52:53] no they're not written down [11:53:02] it's mostly ad-hoc [11:55:07] do you have recordings that I can use for creating that test cases? [11:55:55] I've seen some talks on youtube but they were mostly kind of meta or about technical things [11:56:09] i don't think we've recorded any of them yet [11:56:35] sorry - need to run to a meeting [11:56:52] Okidoki [12:02:52] tobyS: I'm not completely sure what you mean, though I suspect the answer is no [12:03:43] tobyS: perhaps reading this answers your question: https://meta.wikimedia.org/wiki/Wikidata/Notes/Data_model_primer [12:04:11] JeroenDeDauw: well, there is a list of values displayed in this testproperty. i was wonderinh how that could be. because i was only aware of the 3 value types we discussed yesterday [12:04:46] tobyS: these are qualifiers and references [12:05:04] qualifiers and references consists out of snaks [12:05:24] the main value for a claim is called its main snak [12:05:41] we have 3 types of snaks [12:05:55] some value snak, no value snak and property value snak [12:05:56] lbenedix: I can't reproduce your issue with different labels of the same property in the same language [12:06:10] for me it says "continent" on both items [12:06:17] Only the later has an actual value, which is an instance of DataValue [12:06:25] Tobi_WMDE: try in german [12:07:12] tobyS: the types you got introduced to yesterday apply to the last value [12:07:48] tobyS: I recommend you read the linked document which explains this stuff [12:07:51] lbenedix: "Kontinent" for both properties [12:08:00] JeroenDeDauw: k, i'm on it [12:08:17] not sure if i actually need that for then review, seems to be somewhat complex [12:08:18] lbenedix: sorry, for both items [12:08:18] no... in german its fine [12:08:38] it's fine for me in german AND in english [12:08:51] I dont't see the "is on continent" anywhere [12:08:59] funny [12:09:08] when i add ?uselang=en It is "Continent" [12:09:31] btw. I'm not logged in [12:09:38] lbenedix: me neither [12:09:55] JeroenDeDauw: could we do a quick skype call around this topic? need to get an insight quickly so i can work more on the code. [12:10:10] lbenedix: so, when exactly are you seeing "is on continent"? [12:10:19] for property p30 [12:10:45] tobyS: sure [12:10:45] yepp [12:10:52] http://www.wikidata.org/wiki/Q365 [12:11:24] I see "continent" [12:12:11] hmmmm [12:12:38] what should I do to convice you that I see "is on continent"? [12:12:51] lbenedix: you don't have to [12:13:09] only thing I could imagine is that it's some kind of caching issue [12:13:27] I don't know where else "is on continent" should come from [12:13:32] * lbenedix cleares the cache [12:13:52] seems not to be the browser cache [12:14:42] next statement related issue is the different order for items of the same "type" like Q365 and Q1726 [12:16:41] lbenedix: addition to issue 1: I just checked the list of labels of P30: there is no "is on continent" label in any language [12:16:56] ad issue2: what do you mean with order? [12:17:00] oder of statements? [12:17:18] if you mean the order of stetements: there is no such "order" of statements yet [12:17:53] they are ordered in the way the got added (I guess) at the moment [12:17:54] there is an order (obviously not a defined one) and its confusing [12:18:10] lbenedix: : no there iis not [12:18:24] it's just how they got added [12:18:52] "maintain order" is something we have on our todo list, but can't promise when it will be available [12:19:03] okay [12:20:09] issue 1: I dont see any de-alias in the edit history of P30 [12:21:48] lbenedix: so what? [12:21:58] sop nobody added a de alias [12:22:00] yet [12:22:22] sorry... [12:23:38] http://www.wikidata.org/w/index.php?title=Property%3AP30&diff=22685633&oldid=10212388 [12:24:32] one month ago there was tha label "is on the continent" [12:27:30] JeroenDeDauw: a claim is a statement that does not have any sources assigned? [12:28:09] tobyS: a Statement is a Claim that can have Reference objects attached yes [12:28:32] Statement extends Claim [12:28:37] http://www.wikidata.org/wiki/File:Wikidata_statement.svg [12:28:56] lbenedix: yes and it got changed to "continent" on march, 27: http://www.wikidata.org/w/index.php?title=Property%3AP30&diff=15887883&oldid=15033702 [12:28:57] a Propertz with value has no name [12:29:16] Tobi_WMDE: isnt it strange that the cache is soooo old? [12:29:38] lbenedix: yes, I agree [12:29:50] but nothing I can do about this, as it is not reproducable [12:29:52] JeroenDeDauw: property with value but without qualifiers [12:31:29] Tobi_WMDE: I'll test it from another computer [12:33:57] is "continent" [12:47:08] New patchset: Jeroen De Dauw; "Added base for new Wikibase Query extension" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58498 [12:47:09] New patchset: Jeroen De Dauw; "Added wikibase extension type" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58499 [12:54:34] JeroenDeDauw: is daniel w. also on the channel? [12:54:42] Danwe: [12:54:44] ^ [12:55:27] He is obsessing over his UML diagram ATM though, not noticing IRC [12:55:34] tobyS: I am here [12:55:44] Danwe: ah, nice :) [12:56:08] Danwe: how can we determine what is part of the old and new ui (js wise)? [12:56:19] do specific files belong to one or the other? [12:57:40] tobyS: mostly, everything in wikibase.ui is old UI. Also, this is a (little outdated) description of how the OLD Ui works: http://meta.wikimedia.org/wiki/Wikidata/Notes/JavaScript_ui_implementation including a diagram [12:59:06] Danwe: ok, so all files that start with wikibase.ui are out of scope? [13:00:09] New patchset: Jeroen De Dauw; "Rename some renaming things from Query to QueryEngine in QueryEngine component" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58503 [13:00:18] tobyS: yes. But you should have short look at http://meta.wikimedia.org/wiki/Wikidata/Notes/JavaScript_ui_implementation perhaps to get a basic idea [13:07:00] tobyS: i was away for a bit. about properties being nested... i'd say no, but it depends on what exactly you mean [13:07:08] but generally, our data model is not recursive anywhere [13:07:23] DanielK_WMDE: thx, already clarified with JeroenDeDauw [13:07:29] ok [13:10:17] tobyS: can you do me a favor? there's a relatively new and isolated class i'd like some feedback on, if you can squeeze it in. It's Wikibase/repo/includes/store/sql/DispatchStats.php [13:11:43] DanielK_WMDE: can you keep that in mind for later? need to first dig through the code at all. [13:13:08] i'll put it into my queue, so i don't forget [13:18:15] is there documentation somewhere, how the hooks provided by the repo are expected to work? e.g. "WikibaseAddConstraintChecksForLabel" [13:21:09] can someone enlighten me on the mysthery of MultiLangConstraintDetector? [13:21:16] ^ DanielK_WMDE, JeroenDeDauw [13:22:33] mhhh [13:22:41] looks like that class has a bunch of dead code in it [13:23:05] I suspect the thing is about restricting length of labels? [13:23:52] JeroenDeDauw, tobyS: yes, it was introduces to have a generic way to impose restrictions on labels, aliasses and descriptions. [13:23:52] Written by John and merged by Danwe [13:24:03] and was/is used to enforce length limits. [13:24:16] i'm not sure we need such an elaborate system for that, though [13:24:47] ok, so how is such a label-constraint represented? [13:25:30] tobyS: we don't have documentation for our hooks, sadly. it belongs into docs/hooks.txt, just like core does it. but it's missing [13:25:33] another todo. [13:26:07] tobyS: i don't remember how it's represented exactly, sorry [13:26:18] hmm, no good. then i can only remark there that the code is quite hard to understand and should be refactored [13:26:48] some of the hooks are obsolete, i believe [13:26:54] related to polling for changes [13:27:08] others, don't know [13:28:16] tobyS: used in EditEntity, line 718. ON a first glance, it seems like MultiLangConstraintDetector simply implementes the constraints directly [13:28:21] that is, they are hardcoded [13:28:49] (so an appropriate name would be MultiLangConstraints or even TermConstraints) [13:29:27] tobyS: ah... you should look into the Status object. It's used a lot in the editing/saving process. [13:29:35] WikibaseAddConstraintChecksForLabel for example [13:29:36] i disliked it at first, but now i find it quite useful [13:29:37] don't see it used [13:29:39] anywhere [13:30:23] DanielK_WMDE: ok, thx [13:30:23] WikibaseEntityDeletionUpdate looks not used anywehere [13:30:27] anywhere [13:30:51] don't know which are used but probably not need any (i hope) [13:32:29] EditFilterMergedContent is used i think to interact with the abuse filter extension [13:32:35] DanielK_WMDE: right? [13:32:41] there might be some use cases like that [13:42:37] JeroenDeDauw: regarding your code (mail), inserting the different kinds of snaks, there is a nice pattern for such stuff, see https://gist.github.com/tobyS/5354730 [13:44:40] aude: yes, right. WE trigger EditFilterMergedContent so extensions that usually react on the editpage also get notified when we edit an entity. [13:45:06] DanielK_WMDE: ok [13:45:09] i think that's the only core hook we emulate, the rest are our own. [13:45:14] but i may be forgetting something [13:45:18] and not sure which are actually still used? [13:45:36] no, not sure. [13:45:43] most look unused or i don't understand them all [13:45:48] but unused hooks are not a problem, if they don't clutter the code and actually make sense [13:45:53] right [13:45:58] other extensions may indeed want to hook in [13:49:31] New patchset: Jeroen De Dauw; "Move Query out of lib and QueryContent out of repo, both to Wikibase Query" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58510 [13:50:37] Danwe: are you guys using jslint/jshint? [13:51:14] tobyS: yes, with some concious exceptions. [13:51:42] Danwe: in its default config? or do you have the configs stores somewhere externally? [13:51:47] Danwe: hint or lint? [13:52:49] tobyS: we're just using it in our IDEs, so its kind of up to the developers. We don't automatically run it over new commits. [13:53:19] err we do [13:53:29] JeroenDeDauw: since when? [13:53:33] JeroenDeDauw: core does [13:53:48] https://gerrit.wikimedia.org/r/#/c/57449/ [13:53:54] Look at the comment [13:53:58] Danwe: ok, which tool you actually use? jshint or jslint? [13:54:15] This also used to be done for WIkibase [13:54:28] tobyS: jslint [13:54:51] Looks like jenkins-bot is dead ATM, at least for Wikibase [13:56:10] Tobi_WMDE: if you could haz a look at https://gerrit.wikimedia.org/r/#/c/58498/ and commits depending on that - no new logic, just moving stuff about [13:59:10] * lbenedix restarted the wikidatameter-livestream http://www.youtube.com/embed/zBz95tDEZa4 [13:59:18] New patchset: Jeroen De Dauw; "Ran optimize imports" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/58511 [14:01:18] there is something wrong with the irc-monitoring [14:06:46] what is a "usekeys state"? [14:13:26] ohno, that evil thing [14:13:41] tobyS: that is related to the serialization format of the API [14:13:51] Needs a different structure if it is XML or JSON [14:14:07] XML needs extra keys, else the formatter blows up [14:14:24] ah, ok [14:14:27] Think John and DanielK_WMDE can explain that further [14:14:41] would be nice, because i smell some abstraction issue … [14:20:59] tobyS: what code are you looking at? [14:21:47] Tobi_WMDE: you want haz belgian chocolate! [14:22:13] JeroenDeDauw: sorry, no time currently [14:22:31] !nyan | Tobi_WMDE [14:22:31] Tobi_WMDE: ~=[,,_,,]:3 [14:24:24] tobyS: basically: from our data model, we build a nested array structure that is then serialized, e.g. into JSON or XML. The XML serialization is pretty generic, but does make sume assumptions [14:25:05] tobyS: for JSON, it's nice to use the language code as the key for the respective Label structures, for convenient lookup [14:25:06] DanielK_WMDE: i'm currently looking at ApiWikibase [14:25:26] tobyS: haha, have fun :) [14:25:39] :-P [14:26:05] anyway, about the usekeys thingy: if we used such keys in the structure for the XML output, it would generate tags from these codes: , , etc. Not what we want. [14:26:38] so, the code that generates the array structure from our data model needs some hint as to whether we want plain indexed arrays, or preferr associative arrays. [14:27:07] DanielK_WMDE: ok, understood the issue [14:27:09] thx :) [14:27:22] tobyS: the abstraction issue here is of course that the "arrayalizer" is to some degree aware of the implementation details of the XML serializer [14:28:08] well, that basically means there should be 2 implementation of an interface, one that produces a json-suited array and the other one for xml :) [14:34:31] tobyS: yes, but the nested structure is using nested arrayalizers. So you'll need the respective knowledge some levels down in the recursion. Passing in a set of "hints" that are fairly generic is much simpler to realize. [14:41:08] DanielK_WMDE: don't get what you mean. can you point me to some code? [14:43:00] legoktm: yeah, like is aid i should probably move it from heroku (free) to wmflabs, or start paying heroku [14:43:12] legoktm: it should be running again fwiw [14:43:54] New review: Jeroen De Dauw; "(2 comments)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58464 [14:46:06] tobyS: think that daniel is referring to lib/includes/serializers with "arrayilizers" [14:46:44] There is an EntitySerializationOptions there which also has some of this keys stuff in it [14:47:02] tobyS: a lot of serialization code was initially written in API modules [14:47:10] Or rather some base class of the modules [14:47:25] This got split off later since it made no sense in the API [14:47:46] EntitySerializationOptions and the stuff that uses it is still very similar to what was in the API [14:47:57] ie, it is cleaned up a bit, though not all the way [14:48:47] The only serialization code left in the API now (AFAIK) is in the giant method in Api\EditEntity [14:53:11] hmm, not sure if i get that [14:55:04] tobyS: well, if you have any specific question.. [14:59:36] tobyS: the point is that it's not one class that would need to be duplicated, but several, each with fectory methods that generate the correct kind of serializer/arrayalizer for some sub-structure. [14:59:50] and if we have more options, we'd need to duplicate again... [15:00:33] that could be solved by decoration, imo [15:00:51] sounds like a typical visitor pattern case to me … [15:00:58] will think something out [15:01:35] yes, it's possible, but doesn't seem necessary to me. [15:02:34] "use associative arrays where possible" is generic and minor enough to consider it an option to the process; other options could include a list of languages to include/exclude for multilingual parts of the structure, etc [15:02:58] but if you find a nicer way, sure :) [15:05:36] DanielK_WMDE: does not really have something to do with "nicer". such options put regular if … else … cases all over the place. each of these creates a new bunch of test cases to be added. its easy to miss out one of the cases, where the option must be respected. separating the specialities into 2 classes solves both problems. [15:08:25] tobyS: true, but... well, perhaps there's a similar issue with the TermIndex interface and friends. I'm curious what you make of the $options parameter of getMatchingTerms() there. [15:08:41] will look into it [15:10:00] thanks! [15:23:22] any idea how long to wait until the number data type is implemented? [15:39:49] New patchset: Jeroen De Dauw; "Added QueryStoreUpdater tests" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58524 [15:39:49] New patchset: Jeroen De Dauw; "Some implementation for in the SQLStore updater" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58525 [15:52:12] New patchset: Daniel Kinzler; "Generalized access to Claims by property ID." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56272 [15:52:25] New review: Daniel Kinzler; "rebased" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56272 [16:04:42] New review: Daniel Kinzler; "(5 comments)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56272 [16:07:07] New patchset: Jeroen De Dauw; "Use PHPUnit mocking support" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58531 [16:07:07] New patchset: Jeroen De Dauw; "Removed ObservableQueryInterface" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58532 [16:19:56] New review: Jeroen De Dauw; "(1 comment)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56272 [16:22:12] New review: Jeroen De Dauw; "Can you please not mark stuff as done when it is in fact not changed in the latest patchset?" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56272 [16:30:45] New patchset: Daniel Werner; "minor documentation update on wikibase.entity, added some TODOs" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/58536 [17:51:55] the dispatch lag is rising... [18:07:53] legoktm: https://www.stathat.com/stats/D6QK [18:08:31] http://tools.wmflabs.org/legobot/dispatchstats.log has EPM too :) [18:08:58] xD [18:09:42] legoktm: i found a wikidata php framework, one problem, any idea what i should put in /extensions/DataValues/DataValues.php ? :P [18:10:02] errr [18:10:06] idk [18:11:42] maybe this :P https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/DataValues.git;a=blob;f=DataValues.php;h=68786108916daf2e65f5604ddc13b5862aa98b22;hb=HEAD [18:11:48] oh [18:11:49] yeah [18:37:16] addshore: Hi! [18:38:13] "include_once __DIR__ . '/extensions/DataValues/DataValues.php is to include the Mediawiki extension DataValue that provides classes to manage values. [18:38:34] There is a bug with composer so you have to clone it by hand. [18:42:51] New patchset: Daniel Kinzler; "Generalized access to Claims by property ID." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56272 [18:45:00] New review: Daniel Kinzler; "Addressed concerned with the original patch set, and provided additional testing after rebase." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56272 [18:45:21] New patchset: Daniel Kinzler; "Rename PropertySQLLookup to PropertyEntityLookup." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56319 [19:02:56] Do we have code review for gadgets yet? [19:03:32] New patchset: Daniel Kinzler; "Implement PropertyLookup on top of TermIndex." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56330 [19:06:30] hoo: not really…should we? [19:07:08] Of course... people seem to enable gadgets w/o consensus again [19:07:16] eg. https://www.wikidata.org/w/index.php?title=MediaWiki:Gadget-CommonsMedia.js [19:08:10] That one has been enabled by default, by it's author with hardly any consensus and ZERO code review [19:09:28] hoo: You could ask the author why they did it, Bene*. (benecloud for a ping) [19:09:37] wasnt that already a gadget in something else? [19:13:15] New patchset: Daniel Kinzler; "Implement PropertyLookup on top of TermIndex." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56330 [19:13:34] New patchset: Daniel Kinzler; "Use the term based property lookup on the client." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56717 [19:16:44] New review: Daniel Kinzler; "(1 comment)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/56272 [19:17:01] legoktm: Dunno... but the code got significant flaws, not security related this time, tough [19:17:55] {{sofixit}} ;) [19:17:55] 04[1] 04https://www.wikidata.org/wiki/Template:sofixit [19:18:06] legoktm: (partly) on it [19:18:11] :D [19:21:22] legoktm: Did it... [19:21:27] * legoktm looks [19:22:22] :) [19:24:22] I guess people don't realize that, then they default a gadget it can cause major disruption... [19:25:22] The one who set it at default should know that. When you see where they currently are. [19:26:21] Yes, would be nice if someone could let him know... we should establish a process for this... [19:26:45] I just can't review all gadgets alone... that eats a lot of time... [19:26:52] Off for eating ;) [19:31:53] what is wrong with the gadget? [19:32:28] Nothing as I can see, its just people have a problem with gadgets being made default out of the blue. [19:32:41] me too [19:33:10] because I dont see any difference between Extensions that uses a lot of JS and a gadget [19:36:21] the only difference seems to be that there are no rules for gadgets and no review process [19:38:30] o/ Swob [19:38:35] o/ [19:38:48] hi [19:38:55] are Wikidata item descriptions always in English? [19:39:05] what if there's an entity that has no English page? [19:39:48] Swob: you only see the english one because your interface is in english [19:39:51] ok [19:40:01] if you install the label lister gadget you can see all the others easily [19:40:04] well Im trying to fix this page [19:40:05] https://www.wikidata.org/wiki/Q1454125 [19:40:08] a bot keeps reverting me [19:40:24] and Im wondering if it's because the bot thinks English is the "source" wiki [19:40:29] * Jasper_Deng blocks Swob for edit warring [19:40:29] and the others are dependent on it somehow [19:40:31] jk [19:40:36] but it's the English link that doesn't belong [19:40:59] Swob is fighting with Legoktm's bot :D [19:41:17] heh [19:41:32] ok [19:41:40] Swob: can you create a new item for the english link [19:41:43] thatll stop my bot [19:41:56] I dont think there are any wikis that have a page for "gullibility" [19:42:11] right [19:42:15] are you just trying to make us look? [19:42:16] :< [19:42:28] lol. no [19:42:33] https://www.wikidata.org/wiki/Q10841408 [19:42:35] Haha. [19:43:17] ok, thanks [19:43:19] how did you do that? [19:43:30] https://www.wikidata.org/wiki/Special:NewItem [19:43:30] do you have to be an admin? when I tried it told me the page already exists [19:43:37] no [19:43:39] umm [19:43:53] if that happens, i just normally hit the button a few times and itll go through [19:44:00] presumably because there cant be two items with the same English name [19:44:17] no, there cannot be two items with the same links [19:44:26] but there can be with the same name [19:44:26] well this wasnt going to have the same links [19:44:34] just the same name [19:44:48] anyway, it happened also to me, just click another time [19:44:54] I guess it's just that they had the same name in English [19:45:17] but OK it seems to be fixed now? [19:45:46] yeah [19:45:51] it was the "label" that was the problem [19:46:00] I tried to use the same label as the previous page [19:46:14] and an empty description [19:46:21] that must be the problem [19:46:33] two pages can have the same label, but not the same label and the same description [19:46:37] and presumably empty = empty [19:47:37] right [19:47:50] thats the uniqueness thing [19:48:26] ok [19:49:13] i think the Hebrew one might be appropriate [19:51:35] is there a way to leave edit summaries? [19:51:50] I think if I just keep removing things with no explanation people will wonder what Im doing [19:51:53] no [19:51:57] well [19:51:59] last time I tried leaving a message on the talk page, but nobody saw it [19:52:01] if you're a bot :P [19:52:16] dont worry about it for now [19:52:22] usually when people revert theyll ask why [19:53:03] ok [19:54:24] also it seems GOogle Translate doesnt do Wikipedia anymore [19:54:43] I remember someone telling me that there was a browser extension that would run the text of a page through Google to translate it [19:54:49] o.O [19:56:39] ah i found this [19:56:39] http://translate.google.com/translate_buttons [19:56:43] they still work if you use that [19:57:34] still doesnt work on Russian though [19:57:43] it must be something to do with the characters [19:57:50] what are you trying to do? [19:57:53] oh well [19:58:08] i was trying to see if the Russian and Hebrew pages should also be removed [19:58:49] I guess it cant handle characters like ö [20:00:29] oh well, not worth it [20:00:39] I know for sure that the German, Dutch, and Swedish pages are the same though [20:17:18] edsu: ping [23:03:47] hi [23:05:37] Hi, Paintman :) [23:08:07] Why Wikidata Phase 2 was posponed on some wikis? [23:08:39] there are a few technical issues with dispatch lag [23:09:06] thanks for the great work! [23:09:21] Wikidata is awesome! [23:09:43] :) [23:09:59] \o/ [23:11:16] :D