[06:05:10] good morning [08:43:46] (03PS11) 10Addshore: Refactor GetEntities Test [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81003 [08:44:36] (03CR) 10Addshore: "As far as I can tell all comments have been addressed" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81003 (owner: 10Addshore) [08:47:32] Isn't this rather a rewrite than a refactor? :P [08:50:55] (03PS5) 10Tobias Gritschacher: toolbarcontroller: Detach event handlers before re-attaching. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/79776 (owner: 10Henning Snater) [08:59:27] (03CR) 10Hoo man: [C: 04-1] "(3 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81003 (owner: 10Addshore) [08:59:36] yay for nitpicking [08:59:42] :D [09:08:05] (03CR) 10Addshore: "(3 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81003 (owner: 10Addshore) [09:08:19] hoo: ^^ [09:08:56] (03CR) 10Tobias Gritschacher: [C: 032] Fix coverage tags for Entity [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82058 (owner: 10Jeroen De Dauw) [09:09:08] (03Merged) 10jenkins-bot: Fix coverage tags for Entity [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82058 (owner: 10Jeroen De Dauw) [09:10:41] (03CR) 10Hoo man: "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81003 (owner: 10Addshore) [09:11:18] hah! :P [09:11:21] the only thing you *could* do would be to compile an array of combined props and lang or something like that and loop over that... although that's not necessary nicer [09:11:24] (03CR) 10Addshore: "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81003 (owner: 10Addshore) [09:14:33] hoo ^^ :) [09:14:35] (03PS12) 10Addshore: Refactor GetEntities Test [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81003 [09:14:37] oh.. ^^ [09:17:23] :) [09:18:45] You could have linked https://gerrit.wikimedia.org/r/75533 but I doubt that matters much [09:27:26] !admin [09:27:37] yes [09:27:43] ! [09:27:46] DangSunM_alone: Urm... [09:27:47] testing sorry. [09:28:09] I think !admin is working like tagged user [09:28:20] DangSunM_alone: I was about to say. Why is an admin asking for an admin :p [09:28:20] >.< [09:28:38] just test [09:29:07] !admin is get attention (users....,) [09:29:07] You are not authorized to perform this, sorry [09:29:21] It is result I think [09:29:33] DangSunM_alone: !_admin is to get attention of admins :p [09:29:54] testing again [09:29:56] !_admin [09:30:04] works? [09:30:20] no... thats why we use it in examples :P [09:30:22] DangSunM_alone: !_admin does nothing. The _ is there to prevent it pinging users. [09:31:15] thanks [09:31:19] Tobi_WMDE: in the past have the entity ids in guids been stored as lowercase? [09:31:50] guids used to be whatever the api user supplied without validation [09:31:59] sure some supplied it with lowercase [09:32:04] * aude thinks likely [09:32:09] I mean when they are generated ;p [09:32:16] not sure [09:32:18] i think the ui generated them [09:32:23] and there was some inconsistency [09:32:48] * aude thinks we should assume yes [09:32:50] I guess it used getPrefixedId to get the entityid [09:32:57] How can I create a link to MyPage?debug=true with Wikitext? Is this possible? [09:33:03] so whatever that returned was used to construct the guid [09:33:54] addshore: Yes, they were lowercase q42, now they are uppercase Q42. [09:34:00] okay [09:34:33] * aude remembers the ui generated guids failing validation when i added a validator [09:34:55] then we made it more flexible for deserializing and stricter for new guids [09:39:05] hmm, guidgenerator is in lib but used in datamodel, seems a bit silly... [09:39:17] hmm [09:43:32] Change on 12mediawiki a page Extension:WikibaseLib was modified, changed by 86.176.186.77 link https://www.mediawiki.org/w/index.php?diff=775038 edit summary: [-29] [09:45:13] Change on 12mediawiki a page Extension:Wikibase Client was modified, changed by 86.176.186.77 link https://www.mediawiki.org/w/index.php?diff=775039 edit summary: [+10] [09:46:10] Change on 12mediawiki a page Extension:WikibaseLib was modified, changed by 86.176.186.77 link https://www.mediawiki.org/w/index.php?diff=775041 edit summary: [+6] [10:21:08] The current inability to import data from Wikipedia is the best argument for importing data from DBpedia .. http://ultimategerardm.blogspot.nl/2013/09/the-president-of-latvia.html [10:24:21] GerardM: dbpedia also doesn't have the information that he was a communist until 1990. I don't follow your point? [10:24:23] http://dbpedia.org/page/Andris_B%C4%93rzi%C5%86%C5%A1_(Latvian_President) [10:31:01] (03CR) 10Tobias Gritschacher: [V: 04-1] "This fails qualifiers & references selenium tests." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/79776 (owner: 10Henning Snater) [10:35:23] Tobi_WMDE: https://gerrit.wikimedia.org/r/#/c/81910/ [10:36:13] Denny_WMDE: got five minutes? [10:39:13] DanielK_WMDE: i do [10:40:23] Denny_WMDE: i have trouble recalling what exactly we agreed on for SMWCon. Who is going to do what? [10:40:38] smwcon or semantic mediaweek? [10:41:00] smwcon [10:41:15] let's ask lydia tomorrow [10:41:20] ok [10:41:21] there should be no deadline today [10:41:43] not for smwcon, right. i got confused about that, that deadline if for the social media week [10:42:01] are you doing somethign there? so far i was only aware of mathias [10:42:29] Tobi_WMDE: the other way I see of doing the same would be to leave the generation to the changeop, and make getIdFromClaimGuid public [10:42:42] then the changeop could see if the id looked correct and fix if needed [10:43:00] the the datamodel stuff could just validate and throw and exception if it is wrong [10:43:28] Denny_WMDE: ? [10:43:36] DanielK_WMDE: checking [10:44:13] hm, can't find anything relevant in the scehdule http://socialmediaweek.org/berlin/schedule/ [10:44:16] addshore: no, not everybody uses changeops. [10:44:28] changeops are mainly for the api [10:44:34] well, we chould have the guid stuff everywhere in the repo anyway.. [10:44:41] so datamodel doesnt depends on lib [10:44:42] :> [10:44:52] *depend [10:44:53] let's see what jeroen says [10:45:05] Tobi_WMDE, addshore: would be nice if our special pages (like NewPoperty) would use ChangeOps [10:45:29] move everything that uses guid out of data model :D [10:45:55] DanielK_WMDE: mail from felix sasaki, 16. July [10:46:05] aude: then the DataModel needs a guid provider interface [10:46:18] heh :) [10:46:36] DanielK_WMDE: there was another mail on july 18 [10:46:41] i have not followed up [10:46:55] DanielK_WMDE: indeed [10:47:14] DanielK_WMDE: and then 12th of august [10:47:20] DanielK_WMDE: I think Special:SetLable etc do already [10:47:21] best to ask felix i guess :) [10:47:36] Denny_WMDE: ah, right. uh, me nither. yes, the august one is the one i should reply to today. [10:47:37] not sure about NewItem [10:47:48] Tobi_WMDE: cool [10:47:52] GerardM: are audio files connected to omegawiki entries? do you have an example? [10:48:44] haha Tobi_WMDE I found it https://bugzilla.wikimedia.org/show_bug.cgi?id=53204 [10:49:18] :) [10:49:21] that's from you [10:49:26] yes :P [10:49:28] Denny_WMDE: DBpedia has correct current affiliation ... THAT is relevant [10:50:00] addshore: guid generator is trivial piece of code [10:50:29] meaning? :P [10:50:46] what about guid generator depends on lib? [10:51:00] the fact thats where the guidgenerator is :P [10:51:11] mediawiki\extensions\Wikibase\lib\includes\GuidGenerator.php [10:51:20] yeah [10:51:39] ie. you cant use entity->newClaim in the datamodel without having the whole of lib present [10:51:59] or I guess making your own guid generator class O_o [10:52:55] it could just take $guid as a param [10:52:59] not guidgenerator [10:53:20] would be nice to check it's valid, but maybe that doesn't need to be at data model level [10:53:27] well, it need not exist at all :P addClaim exists [10:53:37] true :) [10:53:41] with addclaim you simply make the claim object and give it a guid then add it xD [10:53:52] GerardM: indeed interesting [10:53:57] GerardM: anything on the audio? [10:54:02] newclaim just seems to put that last tiny bit of logic in the datamodel [10:54:41] newClaim appears unused except in tests and ChangeOpMainSnak [10:54:43] heh, and newClaim only has 18 uses in our code base... [10:54:47] yup [10:54:50] can those be made to not use newClaim? [10:55:05] yes, just make the tests give the claim a guid and use addclaim instead :> [10:55:11] maybe thats what I will do today [10:55:12] yes [10:55:19] ok [10:55:19] Denny_WMDE: hm, I guess we should have provided a summary for the Social Media Week event a while ago... I guess it'S fine if we do that tomorrow. [10:55:23] Denny_WMDE: I think OmegaWiki does do audio but Christoph knows better [10:55:24] seems it would get us one step further to getting rid of this from the datammodel [10:55:30] yep [10:55:58] I'm not going to leave this chair until I have done them all ;p (of course im getting another dirnk first) ;p [10:56:03] heh [11:27:05] hmm aude should I mark it as deprecated? [11:27:15] yes [11:27:25] * aude tempted to just remove it :) [11:29:25] (03PS2) 10Addshore: Get rid of the use of Entity->newClaim() in tests [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82233 [11:29:26] heh [11:29:32] :) [11:29:33] I have got rid of it in all tests [11:29:36] ok [11:29:40] just got to get rid of it in 1 changeop now [11:29:44] ok [11:58:57] (03PS1) 10Aude: add backwards compatiblity for change dispatcher with badges [extensions/Wikibase] (mw1.22-wmf14) - 10https://gerrit.wikimedia.org/r/82235 [11:59:14] (03CR) 10jenkins-bot: [V: 04-1] add backwards compatiblity for change dispatcher with badges [extensions/Wikibase] (mw1.22-wmf14) - 10https://gerrit.wikimedia.org/r/82235 (owner: 10Aude) [12:01:41] Tobi_WMDE: for wbCreateClaim I guess rank should only be returned if the claim that is created is a statement...? [12:02:17] sure [12:02:34] (03CR) 10Aude: [V: 031] "tests pass for me and tested manually with mw1.22-wmf14 client + repo and with mw1.22-wmf14 client and repo master" [extensions/Wikibase] (mw1.22-wmf14) - 10https://gerrit.wikimedia.org/r/82235 (owner: 10Aude) [12:03:01] https://gerrit.wikimedia.org/r/#/c/82235/ gets a red dot [12:05:31] (03PS3) 10Addshore: remove Entity->newClaim from MainSnakChangeOp [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82234 [12:07:21] (03PS1) 10Addshore: Deprecate Entity->newClaim [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82237 [12:08:17] (03PS2) 10Aude: add backwards compatiblity for change dispatcher with badges [extensions/Wikibase] (mw1.22-wmf14) - 10https://gerrit.wikimedia.org/r/82235 [12:12:00] DanielK_WMDE: Tobi_WMDE review of https://gerrit.wikimedia.org/r/#/c/82235/ [12:12:03] woudl be helpful [12:12:07] would [12:12:08] (03CR) 10jenkins-bot: [V: 04-1] (bug 53678) add backwards compatiblity for change dispatcher with badges [extensions/Wikibase] (mw1.22-wmf14) - 10https://gerrit.wikimedia.org/r/82235 (owner: 10Aude) [12:12:33] jenkins will say "no" because entity id changes in data model are incompatible with the branch [12:13:03] * aude has corresponding patch almost ready for master [12:13:54] (03CR) 10Tobias Gritschacher: [C: 032 V: 032] Updated test configuration to the latest version [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82039 (owner: 10Zfilipin) [12:22:52] fff [12:22:55] Hellow [12:22:56] (03PS11) 10Addshore: Implement merge api module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/79988 [12:22:57] (03PS9) 10Addshore: Add editentity claims support [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81671 [12:24:23] DanielK_WMDE: I guess there shouldn't really be a strict format for the guid in datamodel? [12:24:36] (03CR) 10jenkins-bot: [V: 04-1] Implement merge api module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/79988 (owner: 10Addshore) [12:25:33] heh, useful , Build step 'Execute shell' marked build as failure 12:24:31 Recording test results 12:24:31 Finished: FAILURE [12:32:39] (03CR) 10Zfilipin: [C: 031] Cucumber: simplify css selectors for labels and descriptions [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81926 (owner: 10Tobias Gritschacher) [12:50:29] (03PS3) 10Addshore: ChangeOpClaim now uses the claim guids [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82223 [12:50:51] (03CR) 10Tobias Gritschacher: [C: 031] (bug 53678) add backwards compatiblity for change dispatcher with badges [extensions/Wikibase] (mw1.22-wmf14) - 10https://gerrit.wikimedia.org/r/82235 (owner: 10Aude) [12:53:57] (03CR) 10Tobias Gritschacher: [C: 032] Added extra tests for EntityIdValue [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82061 (owner: 10Jeroen De Dauw) [12:54:14] (03Merged) 10jenkins-bot: Added extra tests for EntityIdValue [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82061 (owner: 10Jeroen De Dauw) [12:54:23] (03CR) 10Tobias Gritschacher: [C: 032] Add extra tests for EntityId [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82062 (owner: 10Jeroen De Dauw) [12:54:32] (03Merged) 10jenkins-bot: Add extra tests for EntityId [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82062 (owner: 10Jeroen De Dauw) [12:56:29] (03PS12) 10Addshore: Implement merge api module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/79988 [12:56:30] (03PS10) 10Addshore: Add editentity claims support [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81671 [12:58:10] (03CR) 10jenkins-bot: [V: 04-1] Implement merge api module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/79988 (owner: 10Addshore) [13:03:04] Tobi_WMDE: any idea about the above failure O_o [13:03:20] the tests, yet again, giving me helpful messages [13:04:48] addshore: the data model probably shouldn't care. [13:05:11] does it use "composite" ids anywhere ("entityId$guid") [13:08:33] aude: do you know if simply linking different stylesheets like in https://git.wikimedia.org/blob/mediawiki%2Fskins%2Ferudite/0e271e351fc9208f462e97c8cac925e283462540/erudite.php will make the skin responsive? [13:08:46] lines 30-36 [13:09:19] http://bots.wmflabs.org/~wm-bot/dump/%23wikimedia-wikidata.htm [13:09:51] (03PS13) 10Addshore: Implement merge api module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/79988 [13:10:11] pragunbhutani: she is out for lunch :) [13:10:42] DanielK_WMDE: indeed, think I have got it all sorted now :) [13:11:11] addshore: okay, I'll contact here in a bit then, thanks :) [13:11:27] pragunbhutani: Ill ping you when she walks in (if I remember) :) [13:12:33] addshore: okay cool [13:23:45] mhhm Tobi_WMDE did you say there was a way to go from claim guid to entity it belongs to? [13:30:43] addshore: well, the claim guid has the entity id in front of it [13:30:59] i can't imagine it being more complicated than that [13:31:28] pragunbhutani: i think you need to ask jon [13:32:32] Jon's away till the 5th [13:32:41] hmmmm [13:32:46] aude: I'll get in touch with somebody else from the mobile team [13:33:08] Jon told me I could reach out to a few people, but I think they're all away right now [13:33:20] I'll check in again at night [13:33:32] i'm not sure who else there to ask, but maybe yuvipanda or brion or kaldari [13:33:44] Question about the tools account. Working on the replica means what exactly? Do I change the “live” data of the wikis? I'd like to have a copy of wikidata.org to mess around with without breaking anything serious. [13:33:46] he said I could talk to MaxSem, awjr or jgonera [13:33:56] jgonera and awjr are good [13:34:02] max also, though he does more backend stuff [13:34:11] I'll ping them later [13:34:14] ok [13:34:26] * aude doesn't know better than you do about css :) [13:34:34] aude: haha okay :) [13:34:45] i don't know what is acceptable [13:34:50] otherwise, how do you feel I should plan out my timeline for the project over the coming month? [13:35:39] well, we need to know how the different "views" (e.g. when you click on a statement, then what) tie together [13:35:40] (03CR) 10Daniel Kinzler: "(2 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/72751 (owner: 10Liangent) [13:35:59] (03CR) 10Daniel Kinzler: [C: 031] "seems fine, want to test it though." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/72751 (owner: 10Liangent) [13:36:05] i can try to help with getting the data available to the skin [13:36:07] aude: indeed :) [13:36:45] aude: hmm, that's true [13:37:01] that's what's on my plate as issue number one [13:37:07] ok [13:37:13] I think we'll have a clearer picture once I have the scripts there as well [13:37:35] making the data available is not necessarily difficult [13:37:41] but touches entity view which is ugly [13:37:54] so makes it more difficult than it ought ot be [13:37:56] to be [13:39:57] addshore: Entity::getIdFromClaimGuid [13:40:47] explode( '$', $claimKey ); [13:40:53] * aude knew it was that simple :D [13:41:16] it is marked as deprecated though [13:41:47] * aude not sure what the "right" function ot use is now [13:44:34] ill make it up as I go along for now ;p [13:44:41] ok [13:44:48] should probably amke a guid parser soon;p [13:44:53] or something [13:48:14] aude: Do you perhaps know how the replica DBs work? Can I play around on them for testing purposes or will I breakk stuff? [13:48:18] aude: so once we know how the views tie together [13:48:25] Granjow: on labs? [13:48:34] aude: yes. [13:48:35] Granjow: they are read only ;p [13:48:42] and then the 'getting the data' bit, we should have the first working iteration! [13:48:48] those are read only, but if you do a super mega query then it might have limits [13:49:01] pragunbhutani: yay! [13:49:09] aude: :D [13:49:41] Okay. Well, I do have a huge query, meaning that it scans all data. And I just saw that a select count(id) takes 5 seconds already. [13:50:14] Granjow: i don't know how much query is too big or how to handle that [13:50:23] you can ask in #wikimedia-labs [13:50:26] What I need to do is e.g. find all items that have “instance of“ set to e.g. “tree” [13:50:29] okay, thanks [13:50:37] hmmm, you might want to use a dump [13:51:05] the replicas do not have the full text [13:51:16] of the items (e.g. the json blobs) [13:51:40] but you can use tables like page links to see something is connected with another item [13:52:01] and then connected to "tree" perhaps, but seems hacky [13:52:18] Ok. Connections alone are not sufficient I think, yes ... [13:52:45] i would use a dump [13:53:05] but you can try with the replicas [13:53:36] Will try the dump. I'd also like to write some test data. [14:00:01] Granjow: you can try exporting pages from wikidata [14:00:07] do we have a special page for all properties (I mean an automatic one)? [14:00:09] to get a subset for testing [14:00:30] Sannita: no, but probably something we should have [14:00:37] in my opinion [14:01:01] we introduced a new "property info" table so think it would be more feasible [14:01:09] aude: thanks, because I just got aware that https://www.wikidata.org/wiki/Property:P691 is not listed in WD:P/Generic (at least I can't find it) [14:01:15] oh, ok [14:01:22] so I was wondering how many properties we're missing [14:01:38] there's special:allpages which can be filtered by namespace [14:01:43] but it's not terribly useful [14:01:51] well, for a start... [14:01:53] thanks [14:01:56] ok [14:02:14] would be nice to get a list of all properties of a particular type [14:02:44] * aude has to finish my open todo items before starting anything new, though [14:02:51] np, thanks ;) [14:02:57] I'll put it as a paper cut [14:02:58] sure [14:03:12] i was just thinking of that last night when trying to find all time properties :) [14:03:32] :d [14:03:34] :D [14:03:56] aude: A dump would be more useful since the extension works, but performance is a big issue I think [14:04:06] Granjow: hmmmm [14:04:39] if you don't care about page history, then you can have a dump of current revisions [14:04:50] it's still a bit large amount of data [14:05:05] but a lot more manageable [14:05:08] Problem is, I need those sets {instance of Tree}, {instance of Question} etc. updated live, so as soon as you change an item, I need the updated data. [14:05:20] aude: I don't need page history. [14:05:56] hmm, how do we simply get from an entityId to the entity itself now with the entityid changes? >.< [14:05:58] when the wikibase query stuff is ready, that might help [14:06:07] no idea of a timeline for that [14:06:32] and those queries will get cached (not sure how long) [14:06:54] Granjow: you can use the daily incremental dumps to keep the data updated [14:07:12] more frequent than that, i am not sure what is feasible [14:09:36] aude: Ah -- no, old wikidata data is not the problem. Here I just need lots of data. I was talking about my extension which has to keep track of changes. [14:10:00] hmmmm [14:10:08] And initially go through all data, which I wonder if it is possible because of the execution time limit of PHP [14:11:02] php can probably work though i would use something else like python or java to parse the dump [14:11:39] i think denny parses the dumps with python [14:11:55] byrial uses c (http://toolserver.org/~byrial/wikidata-programs/) [14:12:10] https://github.com/vrandezo/wikidata-analytics [14:12:52] those are probably quicker, but php probably would also work [14:13:54] aude: But then I cannot do life updates. Like you add a new claim that item Asdf is now an instance of “Tree” too, then this has to be visible instantly in my extension when it is online. [14:14:21] aude: But then I cannot do life updates. Like you add a new claim that item Asdf is now an instance of “Tree” too, then this has to be visible instantly in my extension when it is online. [14:14:29] i don't know the best way to do that [14:14:38] (was offline for 2 min somehow, not sure if it arrived before?) [14:15:16] JeroenDeDauw: Can Ask do that? [14:15:41] not sure he's really online now, but might be [14:15:56] ok, otherwise tomorrow :) [14:16:00] ok [14:16:21] hi all :) [14:16:29] $entity = WikibaseRepo::getDefaultInstance()->getEntityLookup()->getEntity( $entityId ); ? O_o [14:16:59] addshore: ... [14:17:14] just a brain check, but thats right? :P [14:19:21] how can I merge two items? [14:20:28] fale: https://www.wikidata.org/wiki/Help:Merge [14:20:29] (03CR) 10Tobias Gritschacher: "(1 comment)" [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82237 (owner: 10Addshore) [14:20:56] JeroenDeDauw: In case you read this: I will have wikibase behind this editor: http://ec2-54-234-76-44.compute-1.amazonaws.com/editor.php If I add a new Taxon, the editor will create a new Item with “instance-of” set to “Woody Plant” (or whatever topic is chosen). Then, I need an updated list of all items that have “instance of“ set to “Woody Plant”. Does Ask do such kinds of queries? [14:21:24] lazowik: thanks :) [14:21:49] np [14:21:53] aude: Do you happen to know how I can load a DB dump? [14:22:43] https://www.mediawiki.org/wiki/Manual:MWDumper perehaps [14:22:46] * Granjow cannot really imagine that this is ascomplicated as https://meta.wikimedia.org/wiki/Data_dumps/Import_examples [14:22:50] perhaps [14:23:01] or there may be other tools or you can create a script [14:23:13] (03CR) 10Siebrand: [C: 032] GENDER support in MediaWiki:wikibase-mylanguagefallbackchain-babel [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82118 (owner: 10Wizardist) [14:23:24] the page dump is the same structure as what you get when exporting pages [14:23:38] only difference is that a dump has a lot of pages [14:24:31] aude: thanks, will try that. [14:25:09] (03CR) 10Tobias Gritschacher: [C: 032] Remove unused imports [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/81952 (owner: 10Jeroen De Dauw) [14:25:50] (03Merged) 10jenkins-bot: Remove unused imports [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/81952 (owner: 10Jeroen De Dauw) [14:26:11] (03CR) 10Tobias Gritschacher: [C: 032] Improvements to StatementTest [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/81948 (owner: 10Jeroen De Dauw) [14:26:37] (03Merged) 10jenkins-bot: Improvements to StatementTest [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/81948 (owner: 10Jeroen De Dauw) [14:27:00] (03CR) 10Tobias Gritschacher: [C: 032] Improvements to EntityPerPageBuilderTest. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81557 (owner: 10Jeroen De Dauw) [14:27:20] (03Merged) 10jenkins-bot: GENDER support in MediaWiki:wikibase-mylanguagefallbackchain-babel [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82118 (owner: 10Wizardist) [14:28:22] (03PS4) 10Addshore: ChangeOpClaim now uses the claim guids [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82223 [14:28:47] (03PS6) 10Aude: Split code for handling site link comments in client (DRAFT) [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 [14:28:56] (03PS14) 10Addshore: Implement merge api module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/79988 [14:28:57] (03PS11) 10Addshore: Add editentity claims support [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81671 [14:29:08] (03CR) 10jenkins-bot: [V: 04-1] Improvements to EntityPerPageBuilderTest. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81557 (owner: 10Jeroen De Dauw) [14:29:19] So one should build mwdumper from source. Would be useful if they also told how to do that. [14:29:33] lazowik: where was your changeset for getentites and badges? [14:30:01] addshore: https://gerrit.wikimedia.org/r/#/c/81934/ [14:30:07] (03CR) 10Aude: "enough refactoring here for one changeset. shall do further work subsequent to this." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 (owner: 10Aude) [14:30:45] addshore: I marked as draft when decided to make editing in the same [14:30:57] but that can be marged by itself also [14:30:57] Granjow: i've not used mwdumper but looks like they have builds available [14:31:00] https://integration.wikimedia.org/ci/job/MWDumper-package/org.wikimedia$mwdumper/ [14:31:01] (only getting) [14:31:48] (03CR) 10Aude: "this also fixes failures in the ChangeHandler tests and incompatibilities with site link badges" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 (owner: 10Aude) [14:33:07] (03CR) 10Tobias Gritschacher: [C: 032] Update PropertyValueSnak to use PropertyId [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/81549 (owner: 10Jeroen De Dauw) [14:33:08] addshore: of course after "enabling" your tests [14:33:28] (03PS13) 10Addshore: Refactor GetEntities Test [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81003 [14:33:29] (03PS3) 10Addshore: Extend api to support badges [DO NOT MERGE, DRAFT] [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81934 (owner: 10Michał Łazowik) [14:33:37] stupid gerrit! [14:33:51] (03PS7) 10Aude: Split code for handling site link comments in client (DRAFT) [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 [14:33:53] (03Merged) 10jenkins-bot: Update PropertyValueSnak to use PropertyId [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/81549 (owner: 10Jeroen De Dauw) [14:35:30] addshore: yes! let's send patches via traditional mail and merge by hand! [14:35:38] :> [14:35:41] (03PS8) 10Tobias Gritschacher: Improvements to EntityPerPageBuilderTest. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81557 (owner: 10Jeroen De Dauw) [14:36:03] (03CR) 10Daniel Kinzler: [C: 04-1] "Two issues:" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/72751 (owner: 10Liangent) [14:38:18] (03PS1) 10Aude: add missing wikibase-comment-add message [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82250 [14:38:46] jenkins has some issue sorry [14:38:51] :/ [14:39:19] I just loaded zuul and see a large pile of stuff :P [14:39:55] addshore: devs write code faster than jenkins is able to prcess ^^ [14:40:42] * lazowik reboots, off for a while [14:40:48] so lazowik I rebased your commit ontop of the getentities tests and added a followup testing the badges (just waiting for the tests to run now) [14:40:51] pah! [14:42:01] addshore: lol [14:42:17] I see that a lot of data are now in Wikidata, but they are often unused by local wiki. Is planned a massive change of this? [14:44:38] fale: local wikis have to decide which data they will take and when. [14:45:02] Sky2042: so it may be that some wiki will not use wikidata for the next 10 years? [14:45:33] fale: Yup. I think most though are waiting for people to develop the Lua necessary, or possibly some are waiting for the data to be sourced, or some simply don't know about it. [14:45:35] * Sky2042 shrugs. [14:48:01] I hope there will be a different future, maybe with something like what has be done with interwiki [14:49:40] Be patient. :) [14:54:11] Sky2042: I'm :) [15:01:47] (03PD2) 10Addshore: Add badges check to getentities test [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82246 [15:02:34] lazowik: ^^ [15:02:44] yup [15:04:50] feel free to expand on the stuff it actually checks for badges ;p [15:05:20] sure [15:05:41] the only thing to check is that every badge is a proper itmeid [15:06:06] um, more [15:06:14] can compare provided with expected [15:06:38] addshore: so I'll change https://gerrit.wikimedia.org/r/#/c/81934/ back to only getting [15:06:58] (which == change commit message...) [15:07:00] that sounds like the best plan :) [15:07:04] ahha xD [15:07:25] well, youd need to do that at some point anyway, as it currently has DNM;p [15:07:29] just fix it when you do that :) [15:08:09] (03PS4) 10Michał Łazowik: Extend api to support badges [DO NOT MERGE, DRAFT] [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81934 [15:08:19] addshore: ^^ ah, the power of gerrit! [15:08:42] xD [15:09:12] addshore: so, if now someone +2 this it will automagically wait for the dependency to be merged? [15:09:19] yup [15:09:23] nice [15:09:31] clever gerrit ;p [15:09:45] Don't! You'll spoil him! [15:09:55] but you can +2 [15:10:00] who says gerrit is male :O [15:11:02] addshore: it's a male dutch name :) [15:11:08] says wikipedia [15:11:11] pah! [15:11:21] I say we fork it and rename it ;p [15:11:25] win! [15:11:26] we even had a collegue here named gerrit [15:11:30] :D [15:11:40] now, that must have been amusing :P [15:11:42] aude: seem like a lot of fun :D [15:12:05] addshore: There is (was?) a WMDE employee called Gerrit.. [15:12:10] yep [15:12:41] " stupid gerrit!" [15:12:46] :> [15:13:01] lucky there isnt an employee here called jenkins... or travis... [15:13:08] or php... [15:13:11] they annoy me more than gerrit sometimes :/ [15:13:13] hah! [15:13:18] :) [15:13:28] aude: Gerrit Holz left WMDE? [15:13:36] hoo: yes [15:13:52] :P Didn't know... but yes, that was a weird coincidence [15:14:11] alright, time to go home [15:14:19] * aude on later to do more coding [15:14:29] * lazowik is again off for a while [15:14:35] :O [15:14:37] see ya! [15:14:38] lazowik: if you want to take a look at the change handler patches, that would be helpful [15:14:47] aude: sure [15:15:10] since you are most familiar with the site link structure [15:15:26] master handles the "old format" fine [15:15:57] i had to make the currently deployed branch able to handle the new format [15:16:11] and make sure master handles the new format okay [15:16:14] mhm [15:16:16] great [15:16:41] but in a while, I'm afk for some time [15:16:43] this all does mean backwards compatibility breaking..... [15:16:57] (some ~= an hour) [15:17:03] e.g. there would be no good way to switch back to the old branch on wikidata, should there be some other issue [15:17:09] we've never had to do that, though [15:17:27] any items edited with the new format become undeserializable in the old code [15:17:40] * aude back in an hour or so [15:17:51] hmm [15:17:58] that's a thing to cosider [15:18:16] but is there any way to change that? [15:19:45] lazowik: by using a different key for the new link structure (e.g. "newlinks" instead of "links"), and for a while, always writing both [15:19:53] that's a pain though, and i'm not sure it's worth it. [15:20:06] lazowik: i don't think it's necessary this time [15:20:09] we should fine out what exactly happens when feeding the new structure to old code... [15:20:23] DanielK_WMDE: i get an exception [15:21:37] if there was a problem with core, they could always switch core back but then everything on the old core branch would have to get wikibase updated [15:21:57] it's rare that this happens but good to be aware there would be such issue [15:22:07] aude, lazowik: consider that a client running old code may load entities containing the new structure [15:22:09] that would be bad [15:22:25] DanielK_WMDE: hence my patch for the branch [15:22:33] ah, i missed that [15:22:38] it is addressed, at least for change handler [15:22:49] if you think of any other place i forgot, please say so :) [15:23:03] aude: the lua lib, and the parser function [15:23:13] it also means importing a page would not work [15:23:21] DanielK_WMDE: ok, yes need to test it [15:23:21] yea, never mind that [15:23:34] especially showing labels of referenced entities [15:23:55] we should test, and since i am away the next few days, don't count on me testing [15:24:04] although i can perhaps this evening [15:25:00] maybe some more backwards compatibility would be a good thing in this case... [15:25:07] although it's indeed a pain [15:25:18] Just FYI, today is a US holiday, so no deployments [15:25:25] Reedy: that's fine [15:25:35] * aude shouldn't be working today :) [15:25:40] (even though last monday was a holiday here but I did it anyway ;)) [15:25:45] heh [15:27:08] alright, really going home.... [15:27:32] (03PS14) 10Liangent: api.php?action=wbgetentities gets a new parameter 'languagefallback' [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/74954 [15:29:18] (03CR) 10Daniel Kinzler: [C: 04-1] "The fallback doesn't seem to work correctly." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/72225 (owner: 10Liangent) [15:30:20] DanielK_WMDE: can you post the content on Special:MyLanguageFallbackChain for every similar report? [15:30:30] claim guids are just annoying :D [15:33:36] probably the most annoying thing is there is no guid class :P [15:37:11] (03CR) 10Daniel Kinzler: "...another issue:" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/72225 (owner: 10Liangent) [15:38:38] (03PS1) 10Addshore: Remove verbose headers [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82253 [15:39:26] (03CR) 10Daniel Kinzler: "...for the previous two reports, I used this fallack chain:" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/72225 (owner: 10Liangent) [15:41:48] liangent: done for the one patch - sadly, this page doesn't exist on the client. [15:43:11] night. [15:44:26] DanielK_WMDE: "In this case, I would expect the en label to be returned, because en-x is a variant of en. But in fact, the de label is returned." by design de is returned... [15:45:08] because with given information you can read both en-x and de [15:45:27] however en-x may have a poorer quality [15:45:38] as it's machine converted [15:45:44] so it decide to provide de [15:48:45] DanielK_WMDE: any thoughts on what I should call an object that assists in working with guids? :P a formatter or parser don't seem to quite fit.... [15:49:08] liangent: munger? [15:49:51] liangent: yea, but if I *explicitly* requested en-x, and not de, to me it makes more sense to get the en variant. [15:50:16] basically, i would expect the fallback to depend on what it's falling back *from*. [15:50:45] liangent: that's not a an absolute blocker, but i find it somewhat surprising. I guess it would be nice to have some control over which kind of fallback is applied. [15:50:51] but that doesn't have to happen in this change [15:51:13] that's some LanguageFallbackChainFactory's job [15:51:50] but from its view, it can't well distinguish language requested from uselang or from userprefs [15:52:20] for userprefs info, we can't weight it so much [15:52:40] as all users must set one, even if they can speak two languages, both natively [15:52:42] i'd like to be able to just ignore user prefs for api calls [15:52:45] as an option [15:52:56] maybe languagefallback shouldn't be a switch but could take values? [15:53:36] but anyway - if this is "working as designed", it can go in; i just think that we should keep thinking of ways to give more control to api clients [15:54:30] liangent: about the second report (missing for-language)... did I misunderstand something? or should that indeed be there? [15:54:52] hm the languagefallback was intended to be some "helper" function for entityview ui [15:55:26] so I didn't try to make it so flexible [15:55:26] yea, but people *will* use it for other stuff [15:55:26] they always do [15:55:27] but as I said: larter is fine. just reply to my comment to that effect [15:55:39] liangent: i'm about to go and fix dinner. i may or may not be online later. i'll be on all day tomorrow, though [15:58:14] DanielK_WMDE: can you please ping me when you're back? [16:00:41] sure [16:01:20] DanielK_WMDE: Even when you try to make it so people do other stuff with it, we do even more other stuff with it than you thought. :D [16:09:32] DanielK_WMDE: aude: so now what? need to add compability with old code? [16:09:59] * lazowik hopes that the badges structure hasn't been deployed yet [16:15:25] ah, only for client if I understand correctly [16:15:43] s/code/formaet [16:16:00] s/formaet/format :p [16:41:56] (03PD1) 10Addshore: Uppercase Api Examples [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82262 [16:45:13] tata for now all [17:08:55] lazowik: if it's easy enough to do, make site link deserialization understand the old structure? [17:09:02] * aude doesn't know if that's easy [17:09:13] that's done [17:09:18] and needed [17:09:19] ok [17:09:32] otherwise it wouldn't work at all :) [17:09:38] hmmmm [17:10:30] give me a minute and i can provide a stack trace of what happens when i switched my repo back to the branch [17:10:47] it's in the branch, of course, so maybe nothing we can do [17:10:51] that won't work [17:10:55] yeah [17:11:19] it would require to change the code that is on computers already :) [17:11:26] * aude might be overly concerned and we've never had to switch wikidata back [17:11:41] the only thing we can do is to ensure that client works with both versions [17:11:46] yeah [17:11:50] s/can/need [17:12:00] s/need/must :) [17:12:08] and test test test master and test.wikidata on thursday :) [17:12:22] ok, so that's priority [17:12:33] make sure there's nothing in the new branch that would be a problem on monday [17:12:36] I'll leave the api until this is resolved [17:12:59] there's no going back... :) [17:13:06] ye [17:13:08] p [17:13:10] :D [17:13:12] and badges are to be deployed? [17:13:21] on monday? [17:13:27] they will be in the data model [17:13:33] mhm [17:13:53] actually I'm going to finish all of this this week [17:13:57] i'm not sure if someone could enter them in the api [17:14:09] I'll need to leave anyways [17:14:10] if they are there but can't be used, that's fine [17:14:15] schools time [17:14:19] ok [17:14:49] how can I get a patch from gerrit with all dependencies? [17:15:27] checkout the last one in the chain [17:16:54] magic! [17:17:07] and where do I find propagation errors? [17:18:12] aude: and what's the current status of client? everything should work? [17:18:19] (with your patches ofc.) [17:18:28] do you have a client setup? [17:18:32] yep [17:18:44] and change propagation working? [17:18:53] my patches should make everything work fine [17:18:53] not now :) [17:18:58] hmmm [17:19:03] did it work before? [17:19:05] (without patches) [17:19:09] ok [17:19:16] setting up now [17:19:26] so, 1) make sure change propagation works with master repo and master client [17:19:38] per the phpunit failures, you can see that it's broken without my patch [17:19:57] 2) put client on mw1.22-wmf14 branch and keep repo on master [17:20:10] try change propagation without my patch to see errors [17:20:26] aude: but your patches are merged on master? [17:20:33] are they? [17:20:53] noo [17:21:12] so master is currently broken? [17:21:12] ok, you can try without my patches to see the problems [17:21:26] and check that my patches fix the problems [17:21:34] yes, master is broken [17:21:51] we set it to skip change handler test (e.g. not voting) so we don't block the other developers [17:21:52] (03CR) 10Umherirrender: "(6 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81671 (owner: 10Addshore) [17:21:59] but shouldn't skip it for very long [17:22:55] tobi can also help verify my patches and run selenium tests [17:23:38] wmf14 is for monday? [17:24:16] wmf14 is deployed currently [17:24:54] we normally deploy to wikidata and wikivoyage on mondays and wait until thursdays for wikipedia [17:25:02] hmm [17:25:15] and Wikidata and WikidataDataModel are deployed at once? [17:25:25] * aude thinks we could actually consider switching everything at the same time, although there still can be a time period where they are different [17:25:28] lazowik: yes [17:25:34] maybe I'll just ask what bothers me [17:25:43] sometimes reed.y deploys wikivoyage first, for example [17:25:57] why the change for wmf14 if badges are not deployed [17:26:15] lazowik: they are still in the data model [17:26:32] and why https://gerrit.wikimedia.org/r/#/c/81924/ is not checking for legacy diffs? [17:26:33] wikibase does not understand the new site link structure [17:27:24] ah, the deployment on monday is still for wmf14? [17:28:06] on other words: the branch patch is in case https://gerrit.wikimedia.org/r/#/c/81924/ is not merged until monday? [17:28:30] hmmm, i should try the legacy diffs [17:28:54] branch patch is for monday - thursday next week when wikipedia still has wmf14 [17:29:01] wikidata will have wmf16 [17:29:20] ah of coure [17:29:28] forgot client/repo is not at the same time :) [17:29:39] yeah, complicated [17:31:07] so, is there a possibility that after https://gerrit.wikimedia.org/r/#/c/81924/ gets deployed the queue will have some diffs in old format? [17:31:20] I think that this should be checked still [17:31:31] someone might update client but not repo [17:32:23] aude: I think that the master patch should also contain the checks from the branch patch [17:32:38] lazowik: i can add the checks [17:33:15] if the client gets updated before repo, it worked fine as far as i tested [17:33:47] em [17:33:50] how could it be [17:34:36] the updated client will assume that $siteLinkDiff[ $siteGlobalId ] is _not_ an atomic diff [17:34:47] and it'll search for $siteLinkDiff[ $siteGlobalId ]['name'] [17:37:10] aude: and wmf is not the whole world :) [17:37:37] even if someone updates at once, there might be still old diffs in the queue [17:38:15] what? wmf is not the whole world? [17:38:22] ofc it is not, its the whole of reality! [17:38:24] :) [17:38:31] :D [17:39:51] aude: but.. if someone updates only the repo things will break anyways :( [17:40:11] (03CR) 10Jeroen De Dauw: [C: 032] Uppercase Api Examples [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82262 (owner: 10Addshore) [17:41:02] (03CR) 10Jeroen De Dauw: [C: 04-2] "Will conflict with https://gerrit.wikimedia.org/r/#/c/77344/" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82253 (owner: 10Addshore) [17:41:17] lazowik: i can do more testing [17:41:20] (03PS4) 10Jeroen De Dauw: Imrpovements to GUID generation code [do not submit] [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77344 [17:41:52] aude: of what? [17:41:59] anyways, it's not stable :) [17:42:05] things _can_ break :) [17:42:07] lazowik: of the different combinations of repo and client [17:42:48] aude: there's no way to fix new repo and old client configuration [17:42:59] ? [17:43:16] (03CR) 10jenkins-bot: [V: 04-1] Imrpovements to GUID generation code [do not submit] [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77344 (owner: 10Jeroen De Dauw) [17:43:21] well.. the repo will send new diffs [17:43:34] that's what my patches are for [17:43:45] and you have no control over old client [17:43:57] it's already on some computer [17:44:07] that we can't control [17:44:13] yep [17:44:16] we can do backports for wikipedia [17:44:22] exaclty [17:44:38] outside of wikidata, people should update period. [17:44:42] tep [17:44:44] *yep [17:45:05] not good to have different versions of repo and client [17:45:05] but I'd still add the checks [17:45:12] agree [17:45:37] and get this merged as soon as possible [17:45:41] yes [17:45:49] so we can make the test voting again [17:45:57] I feel anxius [17:46:01] *anxious [17:46:04] yes [17:46:29] test test test stuff :) [17:47:09] so can you add this checks now? [17:47:15] so I can test [17:47:39] (03PS3) 10Jeroen De Dauw: Add badges check to getentities test [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82246 (owner: 10Addshore) [17:48:35] (03CR) 10Jeroen De Dauw: "(2 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/82246 (owner: 10Addshore) [17:52:42] JeroenDeDauw: https://gerrit.wikimedia.org/r/#/c/81934 easy one [17:59:43] aude: ping [17:59:56] lazowik: i'm updating my patch [18:00:28] ah, ok [18:01:40] (03PS5) 10Jeroen De Dauw: Imrpovements to GUID generation code [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77344 [18:05:34] (03CR) 10Aude: "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77344 (owner: 10Jeroen De Dauw) [18:11:55] aude: ok, I got my local changes working [18:12:27] (03PS8) 10Aude: (bug 53678) make change handler compatible with site link badges [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 [18:12:30] lazowik: ok [18:12:31] is there actually any other thing to test than adding, removing and changing the link? [18:12:39] deleting and restoring the item [18:13:03] i updated my patch [18:13:44] the entire thing could be made a lot nicer but i only want to do so much per patch [18:13:51] incremental changes [18:13:52] mhm [18:15:29] aude: maybe also reverting and restoring old rev? [18:17:57] aude: {adding, deleting, modifying} sitelinks and {deleting, restoring} item and undoing change and retoring old rev works \o/ [18:18:05] yay! [18:23:01] (03CR) 10Jeroen De Dauw: "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81934 (owner: 10Michał Łazowik) [18:23:14] lazowik: you can haz opportunity to clean some stuff up ;p [18:23:44] yey yay [18:25:42] (03CR) 10Michał Łazowik: [C: 031] "Tested:" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 (owner: 10Aude) [18:32:16] thanks lazowik [18:32:27] hehe [18:32:36] I have broken that actually :) [18:32:39] we also need to check that there are no issues with lua and parser function access [18:34:21] (03PS1) 10Jeroen De Dauw: Cleanup in SimpleSiteLink [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82268 [18:34:29] aude: lua should be ok [18:34:46] lazowik: i think so but someone needs to check that [18:34:52] ok, I will now [18:34:59] with new repo / old client and new for both [18:35:00] lazowik: https://gerrit.wikimedia.org/r/82268 [18:35:39] JeroenDeDauw: "an ItemId" [18:35:41] hehe [18:35:54] * lazowik was recenlty pointing that to Jeroen [18:36:09] * lazowik and has written incorrectly also [18:36:44] lazowik: I'm not thrilled with the toArray and newFromArray in this class... though lets leave it like that for now [18:37:04] lazowik: what is really odd though is that the new method newFromArray has a legacy clause [18:37:16] Why would it have support for legacy stuff if it is new? [18:37:34] deseriazlization uses that [18:37:36] Better keep the legacy stuff where it came from, and not pollute this class with it [18:38:02] JeroenDeDauw: so that should go to Item> [18:38:05] * ? [18:38:14] Yeah [18:38:37] not too much too change [18:38:46] can do when will be waiting for some review [18:40:05] (03CR) 10Michał Łazowik: [C: 031] Cleanup in SimpleSiteLink [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82268 (owner: 10Jeroen De Dauw) [18:40:47] * lazowik wonders how could he have missed spaces after/between parenthesis [18:41:07] *parentheses [18:46:30] https://www.mediawiki.org/wiki/Extension:Scribunto/Lua_reference_manual somewhat large... [18:49:18] (03CR) 10Aude: [C: 032] Cleanup in SimpleSiteLink [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82268 (owner: 10Jeroen De Dauw) [18:49:29] (03Merged) 10jenkins-bot: Cleanup in SimpleSiteLink [extensions/WikibaseDataModel] - 10https://gerrit.wikimedia.org/r/82268 (owner: 10Jeroen De Dauw) [18:49:47] lazowik: I'm wondefring more how the reviewer could have missed that ;p [18:59:54] hehe [19:00:01] aude: lua works [19:03:03] aude: and parser is only for propeties, right? [19:03:10] *properties [19:04:45] (03CR) 10Liangent: ""for-language" is "de", which exists in output explicitly only when the output format doesn't have such a place to put it - eg. XML." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/72225 (owner: 10Liangent) [19:05:13] liangent: i'm back now [19:06:11] (03CR) 10Liangent: "Also, < lazowik: yes [19:06:22] DanielK_WMDE: ^ ^ [19:06:27] aude: checked anyways, works \o/ [19:06:33] it currently only works for properties of item type or strings [19:06:42] lazowik: good! [19:06:57] of course lua need adding badges support [19:07:02] but nothing is broken [19:07:08] lazowik: that's fine [19:07:56] soooo [19:08:00] https://gerrit.wikimedia.org/r/#/c/81924 meeeeeeeeergeeeeeee [19:08:19] :) [19:08:31] * lazowik does not know to whom he was shouting [19:08:33] lets find a typo :D [19:08:44] yay typos! [19:09:03] JeroenDeDauw: on the quest: Typos! [19:09:07] ah, missing space found! [19:09:19] subheader: things might get sloppy [19:09:40] liangent: ah, i see. yes, that works. i see a problem with consistency, though [19:10:07] SiteLinkCommentSerializer??? [19:10:16] liangent: if a client script asks for a label in language X, it would expect to get back a result with X as the key (or none). [19:10:16] JeroenDeDauw: what would you call it? [19:10:45] liangent: in the XML case, the client would look at one well known attribute (e.g. language) to check if there was a result. [19:10:55] * aude hates naming stuff [19:11:03] now, the script would have to check for-language first, then check for language. that's somewhat odd. [19:11:18] DanielK_WMDE: and the issue described in https://gerrit.wikimedia.org/r/#/c/72751/ : [19:11:20] patch https://gerrit.wikimedia.org/r/#/c/71996/ first [19:11:26] (03CR) 10Jeroen De Dauw: [C: 04-1] "(4 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 (owner: 10Aude) [19:11:29] in the json case, the originally request language is always the key, right? [19:12:08] DanielK_WMDE: in json, right [19:12:42] hah! [19:12:49] but in xml, it looks more consistent to have language stick to what the label is really written in [19:13:36] liangent: then maybe the for-language attribute should always be there, for consistency, even if it's redundant [19:13:43] then the client can just always check for that [19:14:33] DanielK_WMDE: denny had some different idea iirc [19:15:07] bah :) [19:17:04] liangent: anyway, this all looks really close to finished now. i'll play with it some more tomorrow, and probably merge it. [19:17:34] (03PS9) 10Aude: (bug 53678) make change handler compatible with site link badges [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 [19:18:00] (03CR) 10Jeroen De Dauw: "(3 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 (owner: 10Aude) [19:18:15] DanielK_WMDE: and dependency in my patches looks already obscure [19:18:46] even when I tried to push you to review, instead of stacking more patches [19:20:04] (03CR) 10Jeroen De Dauw: "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 (owner: 10Aude) [19:20:11] (03CR) 10Jeroen De Dauw: [C: 04-1] (bug 53678) make change handler compatible with site link badges [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 (owner: 10Aude) [19:20:15] liangent: yea, i know, sorry. i was off track for three weeks (wikimania, vacation), still trying to get back on track. [19:20:52] liangent: also, before, it was really hard for me to test this stuff. with piglatin, this is much better. thanks for that :) [19:21:15] aude: try to describe what the class does in one short line. If you can do that, a good name should not be to hard to find [19:21:16] maybe i'll adopt your piglatin patch to add support for de-ch [19:21:23] all it does is s/ß/ss/g [19:21:33] hey people, i am trying to add commonscat to https://www.wikidata.org/wiki/Q2876421 but it won't let me [19:22:08] JeroenDeDauw: SiteLinkCommentCreator ? [19:23:44] lazowik_: that is certainly better then the current name [19:23:44] DanielK_WMDE: is it the same in the other direction? [19:23:46] Builder, Generator, ... [19:23:52] ie. s/ss/ß/g [19:24:01] liangent: no. it's not reversible. [19:24:12] de-ch doesn't make a distinction between ß and ss. [19:24:13] Generator seems the best [19:24:13] DanielK_WMDE: Creator is actually better, as it does not imply a pattern [19:24:19] hehe [19:24:21] It is not a generator [19:24:27] ah, yes [19:24:30] JeroenDeDauw: ok [19:24:36] JeroenDeDauw: true [19:24:50] hm. [19:25:05] https://en.wikipedia.org/wiki/Voltage-controlled_oscillator <- that's a generator :) [19:25:14] i expect another iteration to either use the summary class or some other object [19:25:19] anyway before piglatin you can always use sr-ec and sr-el, which are western alphabet based too [19:25:24] russavia: no problem for me. Be careful to not have spaces at the beginning [19:25:37] only so much change for one patch [19:25:43] but for me, piglatin now looks like a good tool to watch unexpected double conversion [19:26:02] pyb -- check it -- what is Hydro Graphics? [19:26:03] lazowik_: i was rather thinking of https://en.wikipedia.org/wiki/Generator_%28computer_programming%29#2001:_Python [19:26:36] liangent: indeed [19:26:44] https://www.wikidata.org/w/index.php?title=Property:P373&action=history looks like someone changed the label name [19:27:03] :) [19:27:11] DanielK_WMDE: heh [19:27:21] and it won't let me undo those edits [19:28:41] russavia: for me too [19:28:42] russavia: AbuseFilter is blocking the revert [19:28:56] but I tried the restore [19:29:02] hoo|away: is AbuseFilter supposed to apply to undo edits? Perhaps it shouldn't. [19:29:12] in any case, why isn't this high profile property protected? [19:29:25] highly used properties probably should be [19:29:35] addshore: is there a policy about this? [19:30:48] lazowik_: revert will probably work even when undo doesn't. [19:30:59] undo is an actual edit, i guess that's why abusefilter is triggering [19:31:12] revert isn't [19:31:46] DanielK_WMDE: I tried restore, and got some error about language code length [19:32:18] lazowik_: yea, restor is also an "edit". revert is special, iirc. [19:32:26] mhm [19:32:51] lazowik_: in any case, abosefilter should probably be silenced for that kind of edit. like to file a bug? [19:33:10] yep [19:33:11] *abusefilter [19:33:24] but anyway... [19:33:42] if there's any admisn around, please protect high profile property pages [19:34:20] lazowik_: i *think* it'S abusefilter [19:34:34] it might be something else [19:34:51] mhm [19:35:26] (03PS10) 10Aude: (bug 53678) make change handler compatible with site link badges [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 [19:35:46] (03PS11) 10Aude: (bug 53678) make change handler compatible with site link badges [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 [19:36:37] (03PS12) 10Aude: (bug 53678) make change handler compatible with site link badges [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 [19:36:55] alright, removing the change handler test skipping [19:38:01] \o/ [19:39:22] http://www.youtube.com/watch?v=NiTcmahEjiY [19:40:18] lazowik_: https://www.youtube.com/watch?v=oHg5SJYRHA0 [19:40:29] JeroenDeDauw: you loose [19:40:42] my irc client shows thumbnails of photos and videos [19:40:46] * lazowik_ wins! [19:41:02] lazowik_: you lose for not hearing the best song ever then! [19:41:09] hehe [19:41:26] lazowik_: but don't worry, i'm not going to give you up [19:41:36] * lazowik_ feels relieved [19:41:51] :D [19:46:05] aude: why is the test still deriving from MediaWikiTestCase? Not used as far as I can see... [19:50:31] DanielK_WMDE: just to let you know: badges do not break lua or parser <- don't know if you got that, got disconnected around that time [19:56:25] lazowik_: ok, good [20:11:08] (03PS13) 10Aude: (bug 53678) make change handler compatible with site link badges [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81924 [20:11:09] JeroenDeDauw: good point! [21:27:04] JeroenDeDauw: alive? [21:27:21] this $props thing is very strange [21:27:44] they are used in handleEntity... [21:28:07] ...only when $params['props'] !== array() [21:28:08] ... [21:46:49] (03PS5) 10Michał Łazowik: Extend api to support getting badges [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81934 [21:51:03] (03CR) 10Michał Łazowik: "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81934 (owner: 10Michał Łazowik) [21:57:13] (03PS6) 10Michał Łazowik: Extend api to support getting badges [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81934 [21:58:30] (03CR) 10Michał Łazowik: "Comment for PS 5 still applies" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/81934 (owner: 10Michał Łazowik)