[00:10:33] (03PS2) 10Nilesh: Added some code docs for the Python scripts [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77063 [00:11:24] (03CR) 10Nilesh: "No idea why this ends upon depending on the JSON change. :(" [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77063 (owner: 10Nilesh) [03:41:52] Join wikimedia-incuvator [03:42:34] hey [03:42:47] that is /j wikimedia-incubator [05:48:22] hello [08:26:00] (03PS1) 10Henning Snater: ValueFormatter for Time DataValue [extensions/DataValues] - 10https://gerrit.wikimedia.org/r/77089 [08:28:38] (03PS2) 10Henning Snater: Implemented MwTimeIsoFormatter in Wikibase [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76921 [08:29:31] (03CR) 10jenkins-bot: [V: 04-1] Implemented MwTimeIsoFormatter in Wikibase [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76921 (owner: 10Henning Snater) [08:33:27] (03Abandoned) 10Tobias Gritschacher: Include ChangeHandler tests on Jenkins again (DNM) [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/64266 (owner: 10Tobias Gritschacher) [08:52:30] (03PS3) 10Tobias Gritschacher: Added new ChangeOp for claim operations [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76710 [09:27:22] DanielK_WMDE: regarding the ClaimModificationHelper you suggested instead of https://gerrit.wikimedia.org/r/#/c/76725/2/repo/includes/api/ModifyClaim.php [09:27:49] just an ApiMain would not be enough, right? [09:28:39] I need access to e.g. attemptSaveEntity and addRevisionIdFromStatusToResult a.s.o. which are in ApiAwikibase [09:30:45] also the modules currently inheriting from ModifyClaim all have common permissions, possibleErrors, params and paramDescriptions [09:31:05] so I would have to define these things separately in each module again? [09:31:44] now they're just calling parent::getParamDescription, parent::getPossibleErrors, a.s.o. [09:33:01] Tobi_WMDE: first off: it's a suggestion, not a requirement. [09:33:16] but yes, ApiMain is not enough, you should provide any services ("collaborators") needed by the helper [09:33:29] if that turns out to be a lot, it should probably be multiple helper classes (one per concern) [09:33:45] DanielK_WMDE: you gave a -1 so I take it as an requirement.. ;) [09:33:57] the -1 was for something else iirc [09:34:02] ok [09:34:19] I guess for the non-handling of null values.. yes [09:34:29] I see [09:34:37] Tobi_WMDE: but note that jeroen reverted me when i did something like this a few months back, on the grounds of "abusing inheritance" [09:34:55] yes, indeed, the -1 was for the error handling [09:34:57] (03PS2) 10Henning Snater: Make use of Time data value formatter in EntityView [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/75084 (owner: 10Daniel Werner) [09:34:59] or lack thereof [09:35:05] (03CR) 10jenkins-bot: [V: 04-1] Make use of Time data value formatter in EntityView [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/75084 (owner: 10Daniel Werner) [09:35:55] Tobi_WMDE: basically, making a baseclass with a bunch of protected methods is bad: you can't use these in classes that have to inherit from something else for some reason, and you can't test these functions separately. so push them into a helper class instead and use compositioon [09:35:55] DanielK_WMDE: ok, so what's about keep a reduced ModifyClaim with the common functions like the getParamDescription, getPossibleErrors aso and have a ClaimModificationHelper for the rest? [09:36:36] Tobi_WMDE: i was just going to suggest that, yea [09:36:56] ok, then I'll do it that way for the moment.. [09:37:07] getParamDescription and getPossibleErrors are public, they are not "hidden" in the base class. [09:37:22] also, it makes sense to expect these to accummulate in a way following the inheritance structure [09:37:30] yes [09:38:09] DanielK_WMDE: ok, thx [09:38:51] Tobi_WMDE: hm... actually, at least for getPossibleErrors, you may want to return something like this: [09:39:20] return array_merge( parent::getPossibleErrors(), $this->helperThingy->getPossibleErrors(), array( /*some more errors*/ ) ); [09:39:36] any errors triggered inside the helper should be declared by the helper [09:39:49] DanielK_WMDE: ok [09:39:52] makes sense [10:22:55] DanielK_WMDE: one idea, if you have 5-8 minutes, no hurry [10:23:00] can even wait until wikimania [10:23:56] Denny_WMDE: what's up? [10:24:25] re keeping track which client article uses which wikidata item [10:25:00] if we wanted to actually track which properties are used, e.g. in order to warn an editor "if you change this you might change huwiki" [10:25:13] that would be totally non-trivial, right? [10:29:01] Denny_WMDE: i'm still trying to understand the actual use case. it touches on a question we havn't actually descussed explicitly, i think [10:30:14] Denny_WMDE: to re-iterate: we definitly track wich wiki page uses (data from) wich item. We might also track wich wiki page uses which property (by label or id), but that wouldn't include the info which property is used for data on which item. [10:30:21] that'S what you ar4e asking for, right? [10:30:34] item+property -> page? [10:30:57] in addition to item -> page / property -> page? [10:31:27] yes, right [10:31:56] i have a statement Q1 P2 Q4, if i change P2 to P4, which wikipedia aticles will I potentially break= [10:32:21] but you can't "change P2 to P4" [10:32:24] that [10:32:30] that's not an operation we support [10:32:36] well, you can decide that P2 is a bad property [10:32:38] and P4 is better [10:32:46] remove the P2 Q4 statement, and add P4 Q4 instead [10:32:55] a normal maintenance task [10:33:03] in general, or for that one item? [10:34:18] Denny_WMDE: pragmatically, we could just calculate the set of all pages using Q1 and then reduce that to the set that also uses P2. [10:34:50] some of them may be using P2 for some other item Q17 and not for Q1, giving false positive [10:34:57] i don't think that would be a huge problem [10:35:49] how do we know which pages us P2? [10:35:58] actually, for filtering out the false positives, we could have more detailed info in the ParserOutput, which is cached. [10:36:44] Denny_WMDE: the usage trackign tabel. it could/should cover all entities used on a page, including the properties [10:37:15] yes, but asking for the P2 of Q1 won't actually use P2 [10:37:26] we don't need P2's label for that [10:37:38] depends on whether we reference it by label or by id [10:37:56] right [10:38:06] wrong [10:38:20] referencing by label doesn't actually call the "get me the label of" thingy [10:38:26] it calls it the other way around [10:38:32] yes [10:38:32] so we would need to fix that too [10:38:39] but if the label changes, the page needs to be updated anyway [10:39:09] (03PS4) 10Tobias Gritschacher: Use ChangeOps for wbsetclaimvalue and wbcreateclaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 [10:39:28] good point [10:39:39] (03PS5) 10Tobias Gritschacher: Use ChangeOps for wbsetclaimvalue and wbcreateclaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 [10:40:07] Denny_WMDE: actually, that raises another, unrelated question: if some page references a property using the label "foo", but there is currently no such property (because the label was changed), we can't track that [10:40:18] so, when some property gets the label "foo", that page isn't found and purged [10:40:20] DanielK_WMDE: can you have a look at the changes in https://gerrit.wikimedia.org/r/#/c/76725/ again? and see if this is ok now? [10:40:43] yeah [10:40:44] bad luck [10:40:46] :) [10:41:11] Tobi_WMDE: sec [10:41:20] DanielK_WMDE: I'm going to implement tests now, but that could take a while. I just want you to look at it again before you're off.. ;) [10:41:42] Denny_WMDE: so we need to decide: a) track properties at all? b) track property/item pairs for each page? [10:42:41] if we assume that an item isn't used on a *huge* number of pages, we could just examine each one to find out whether it uses the property in a way we want to detect, even without an extra tracking table [10:42:58] hmm [10:43:14] the only issue i am raising this now is because it might involve changing a potentially big table later [10:43:25] i was just thinking about that [10:43:30] we could make the tracking table more generic [10:43:44] just add a blob? [10:43:49] instead of entity <-> page, we could define it to be tracking-key <-> page [10:44:04] a tracking key is a string which could be equal to an entity id, or a pair, or whatever [10:44:28] actually i don't think we need to index by id+property [10:44:34] Denny_WMDE: we could also add a blob that we can evaluate if we have a smallish set of pages to consider, yes. [10:44:45] need to think about it [10:44:58] ok, dont want to steal more time :) just had the thought as it is kinda current [10:45:02] we need that index if we have items that are used tousands of times [10:45:25] i kind of like the "tracking key" approach, it's lightweight, flexible, streight forward [10:46:19] it potentially increases the size of the table by an order or two of magnitude [10:46:28] and that's already a big table [10:46:40] the blob doesn't increase it at all [10:46:50] but we need to make sure all required operations still perform well [10:54:26] (03CR) 10Daniel Kinzler: [C: 04-1] "(8 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 (owner: 10Tobias Gritschacher) [10:54:28] (03PS6) 10Tobias Gritschacher: Use ChangeOps for wbsetclaimvalue and wbcreateclaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 [10:54:59] Tobi_WMDE: getting there, but please inject the service objects you need in the helper, instead of using global state. that's yet another advantage of moving the code to a helper class. [10:55:15] Tobi_WMDE: and please talk to addshore, he's also looking into refactoring api stuff [10:55:25] DanielK_WMDE: thx [10:56:30] Denny_WMDE: the blob doesn't increase the number of rows, that's correct. The blob works well enough if we never have to examine too many of them. if a million pages uses the one item and we need to examine all the blobs to find out which pages use a particular property of that page, that would suck [11:01:32] DanielK_WMDE: that is right, that's why i said, we need to figure out what we need the data for [11:01:51] if we only ever need to inspect one blob of one row, then everything is fine [11:01:55] and i think that is the case [11:04:10] i dont think there is a need for inspecting several blobs [11:04:15] in order to filter them further [11:04:24] but i am havent completely thought it through [11:06:49] Denny_WMDE: your use case was: change Q1 P2 Q3 ==> Q1 P4 Q3. That would require: find all pages that use Q1, and examine all the blobs to find out whether that page ises P2 of Q1. [11:07:08] or am i missing something? [11:08:43] uhm [11:08:45] right [11:08:52] yah, my fault [11:09:28] meh, stupid use case, it does make a huge table huger [11:16:17] (03CR) 10Daniel Kinzler: [C: 04-2] "I suppose this should be abandoned now that Ibacc8357d9a has been merged." [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/76640 (owner: 10Nilesh) [11:22:02] (03CR) 10Daniel Kinzler: "(1 comment)" [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77063 (owner: 10Nilesh) [11:22:39] (03CR) 10Daniel Kinzler: [C: 04-1] "(2 comments)" [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77060 (owner: 10Nilesh) [11:29:41] (03CR) 10Daniel Kinzler: [C: 04-1] "please get rid of global state in the helper class, see comments on PS 5" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 (owner: 10Tobias Gritschacher) [11:33:11] Tobi_WMDE, addshore: could you review two things for me please? [11:33:13] https://gerrit.wikimedia.org/r/#/c/75886/ [11:33:19] https://gerrit.wikimedia.org/r/#/c/74394/ [11:33:34] the first one is kind of blocking liangent wrt testability [11:33:42] the second one is blocking url values [11:55:30] (03CR) 10Tobias Gritschacher: "(6 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 (owner: 10Tobias Gritschacher) [11:55:32] (03CR) 10Tobias Gritschacher: "(2 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 (owner: 10Tobias Gritschacher) [11:55:33] (03PS8) 10Tobias Gritschacher: Use ChangeOps for wbsetclaimvalue and wbcreateclaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 [11:55:36] (03PS4) 10Tobias Gritschacher: Get rid of unused Autocomment code [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76727 [11:55:42] (03CR) 10Tobias Gritschacher: "(2 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76915 (owner: 10Tobias Gritschacher) [11:56:07] (03PS5) 10Tobias Gritschacher: Move claimGuidValidator into ModifyClaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76915 [12:04:34] (03PS1) 10Addshore: Fix incorrect test group [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77101 [12:04:37] Tobi_WMDE: ^^ [12:05:13] (03CR) 10Tobias Gritschacher: [C: 032] Fix incorrect test group [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77101 (owner: 10Addshore) [12:10:07] (03Merged) 10jenkins-bot: Fix incorrect test group [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77101 (owner: 10Addshore) [12:18:11] (03PS6) 10Tobias Gritschacher: Move claimGuidValidator into ClaimModificationHelper [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76915 [12:21:50] (03PS4) 10Tobias Gritschacher: Refactoring of wbremoveclaims API module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76929 [12:22:57] (03CR) 10jenkins-bot: [V: 04-1] Refactoring of wbremoveclaims API module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76929 (owner: 10Tobias Gritschacher) [12:44:55] (03PS5) 10Tobias Gritschacher: Refactoring of wbremoveclaims API module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76929 [12:45:43] how can one delete an object from wikidata? [12:53:38] (03PS1) 10Tobias Gritschacher: Move initialization of ClaimModificationHelper to ModifyClaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77112 [12:56:44] (03PS2) 10Tobias Gritschacher: Move initialization of ClaimModificationHelper to ModifyClaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77112 [12:58:51] addshore: DanielK_WMDE: so, I'm finished with all the comments on the summary stuff, I'm going to add tests for ClaimModificationHelper now, but probably you want already take a look: https://gerrit.wikimedia.org/r/77112 (and dependencies..) [13:07:41] Tobi_WMDE: will do in a few [13:12:06] Danwe_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=47435 this is the original bug [13:12:20] Danwe_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=47496 this is the idea with the hidden pref [13:29:04] Denny_WMDE: sounds good to me, except that this means you really *never* get to see the notice. [13:29:11] hm, i got an idea... i'll comment on the bug [13:30:59] DanielK_WMDE: that sounds pretty good :) once you agree, you never see it again [13:32:59] Denny_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=47496#c8 [13:33:57] meh, there is no need to explicitly store the version. but dont care [13:34:06] if you have a new version it could just be -v2 [13:34:09] it will be new then [13:34:15] Denny_WMDE: i'll make the original bug depend on the new one, since resolving that will resolve the original issue [13:34:28] but only if we do it like this [13:34:53] "it could just be -v2"? [13:34:59] it could be any string [13:35:01] message-v2 [13:35:07] yeah, right [13:35:16] no need to put in a -v1 in the first place [13:35:40] i don't quite understand what you mean [13:36:13] i'll set the prio from normal to high [13:36:18] this is a major annoyance [13:37:08] there is no need to use a message-v1 for the hidden pref [13:37:26] because if we ever should change the license, we can use message-v2 or newmessage or whatever [13:37:30] as long as it is different [13:38:20] even better would be to simply reset all hidden prefs for that message in the db [13:38:25] and keep the old message [13:39:14] (03CR) 10Nilesh: "(1 comment)" [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77063 (owner: 10Nilesh) [13:42:05] (03CR) 10Daniel Kinzler: [C: 031] "(3 comments)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76710 (owner: 10Tobias Gritschacher) [13:48:11] we have too many bugs [13:48:19] triaging them for hours…. meh [14:02:39] (03CR) 10Daniel Kinzler: [C: 032] "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76915 (owner: 10Tobias Gritschacher) [14:05:17] . [14:07:10] (03CR) 10Daniel Kinzler: [C: 031] "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76929 (owner: 10Tobias Gritschacher) [14:08:30] (03CR) 10Daniel Kinzler: [C: 032] Move initialization of ClaimModificationHelper to ModifyClaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77112 (owner: 10Tobias Gritschacher) [14:21:47] (03CR) 10Daniel Kinzler: "Looking at the output of" [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77063 (owner: 10Nilesh) [14:22:55] (03CR) 10Daniel Kinzler: "...oh, and: please add a README in this directory, very briefly explaining what to do with these scripts, and how." [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77063 (owner: 10Nilesh) [14:39:14] (03CR) 10Daniel Werner: "This change is not cool, it introduces Wikidata specific code into Wikibase. Why not simply overwrite the "wikibase-shortcopyrightwarning"" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/52406 (owner: 10Denny Vrandecic) [14:46:47] Lydia_WMDE: where are you hiding? [14:47:01] d [14:47:27] JeroenDeDauw: right now? in a hotel room in the friends hotel in taipei ;-) [14:47:30] why? [14:47:38] do you miss me? :P [14:48:41] (03CR) 10Nilesh: "Okay, I'll add the README." [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77063 (owner: 10Nilesh) [14:48:48] cheers [14:49:17] does anyone have an opinion about this? http://www.wikidata.org/wiki/Property_talk:P706#Use_for_events.3F [14:50:52] Why not just have a generic 'in' relation? then you can just work up the chain. ;) [14:51:49] My office is in London. London is in England. England is in Europe. There's no difference in the type of relation there. [14:52:39] But I expect the bureaucracy monkeys won't like it because 'in' hasn't been rigidly defined by the International Standard Organisation For Coming Up With Silly Standards That Make Librarians And Others Seem Like They Are Doing Something Important. [14:56:07] Lydia_WMDE: yes, I'm missing you terribly much right now. Raging over IRC does not work well :| [14:56:25] JeroenDeDauw: lol - what do you want to rage at me about? [14:56:39] or do you want me to rage at someone for you? [14:56:42] :P [14:57:09] Lydia_WMDE: I am not happy with having to take the swag stuff. I prefer only taking hand luggage with me, and now have to take a whole other bag just for the swag. Lame [14:57:25] Lydia_WMDE: enough people with an actual checked bag travveling to take these kinds of things [14:57:31] JeroenDeDauw: welcome to the club -.- i also had to take a huge load [14:57:49] tommorris: doesn't help with events [14:57:53] Lydia_WMDE: If I was taking an extra bag anyways, I would not mind [14:58:27] i understand - and if it were my decision we would have sent this stuff by mail [14:58:30] but alas [14:58:49] Lydia_WMDE: then whos decision is this? Abraham_WMDE told me you are the guilty one! [14:59:10] i am the guilty one because i asked abraham to distribute it yes [14:59:16] because we need to get it to HK somehow [15:00:16] just for comparison: i have a 22kg bag plus laptop plus handbag [15:00:20] ;-) [15:00:50] Lydia_WMDE: for comparison, I have to take 100% more bags with me [15:00:58] I think that beats anyone else actually [15:06:55] Schisma: a generic in property would be fine for that. ;) [15:07:42] tommorris: does it mean you support using P706 for events also? [15:08:07] it seems to be common sense anyway ^^ [15:08:14] Nope. I support creating a sane 'in' property that would do the job. ;) [15:08:32] it's the same problem we have with GND main type vs. generic type [15:08:39] ic [15:09:06] i'd agree [15:09:32] that's the thing that'll kill Wikidata: overly-specific bureaucratic prats who can't see that generic solutions are far better than bureaucratic ones [15:09:51] dumb properties are better than smart ones, because it'll be dumb people using the properties. [15:10:12] Abraham_WMDE: Actually tomorrow is not going to work. What about going to this place on Saturday? (I leave on Sunday, so this works) [15:10:25] sometimes clever is the opposite of intelligent ^^ [15:11:25] * aude gives all my stuff to JeroenDeDauw to take! [15:11:59] * aude is taking only hand luggage [15:40:48] DanielK_WMDE: The top N list for properties that are used in sources - it'll be of the same format as the global list right? pairs? [15:44:23] nileshc: yes [15:44:33] DanielK_WMDE: ok.. [15:48:12] (03PS9) 10Tobias Gritschacher: Use ChangeOps for wbsetclaimvalue and wbcreateclaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 [15:56:46] (03PS4) 10Tobias Gritschacher: Added new ChangeOp for claim operations [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76710 [15:56:47] (03PS10) 10Tobias Gritschacher: Use ChangeOps for wbsetclaimvalue and wbcreateclaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76725 [15:56:48] (03PS7) 10Tobias Gritschacher: Move claimGuidValidator into ClaimModificationHelper [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76915 [15:57:03] (03PS6) 10Tobias Gritschacher: Refactoring of wbremoveclaims API module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76929 [15:57:15] (03PS3) 10Tobias Gritschacher: Move initialization of ClaimModificationHelper to ModifyClaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77112 [16:05:40] i hate german wikipedia ^^ [16:11:54] aude: I'll take those dollar bills you have yes, do keep the rest [16:13:21] (03CR) 10Siebrand: [C: 04-1] "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76929 (owner: 10Tobias Gritschacher) [16:26:33] Danwe_WMDE: http://www.2ality.com/2013/07/es6-modules.html [16:45:53] (03PS1) 10Tobias Gritschacher: Adding tests for ClaimModificationHelper [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77143 [16:51:09] (03PS7) 10Tobias Gritschacher: Refactoring of wbremoveclaims API module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76929 [16:51:12] (03CR) 10Tobias Gritschacher: "(1 comment)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76929 (owner: 10Tobias Gritschacher) [16:51:13] (03PS4) 10Tobias Gritschacher: Move initialization of ClaimModificationHelper to ModifyClaim [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77112 [16:51:14] (03PS2) 10Tobias Gritschacher: Adding tests for ClaimModificationHelper [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77143 [17:02:44] (03PD2) 10Nilesh: Added MapReduce scripts to count global property frequencies [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77070 [17:03:28] (03PD1) 10Nilesh: Added code docs, now using docstrings. Fixed a few syntax errors. [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77146 [17:04:22] DanielK_WMDE: Tested the top-global-properties scripts on the full dataset with hadoop. Can confirm it works. Please review when you can. [17:06:44] DanielK_WMDE: If you want to test it for yourself I can give you quick Hadoop Streaming with Python introduction and you can fire 'em away on the labs instance where Hadoop is already setup. [17:12:10] nileshc: please fill out your midterm evaluation in the next hours [17:12:14] the deadline is soon [17:12:29] and you're making the admin team cry because of potentially missing the deadline ;-) [17:12:47] Lydia_WMDE: Yes, I'll fill it out before going to sleep tonight. :) [17:13:25] Lydia_WMDE: So that means within 4-5 hours. [17:14:00] nileshc: thank you! :) [17:14:11] Lydia_WMDE: :) [17:15:09] (03CR) 10Nilesh: "I would like to amend https://gerrit.wikimedia.org/r/77060 to simplify the nested code and add in-code docs too. I think it's best if I ad" [extensions/WikidataEntitySuggester] - 10https://gerrit.wikimedia.org/r/77063 (owner: 10Nilesh) [17:20:25] Lydia_WMDE: Sumanah sent me an email too about this. Should I reply in the thread to make everyone aware that I'm indeed filling it out in a few hours, or is it okay? [17:20:44] nileshc: i just talked to her about it - it's fine [17:20:57] Lydia_WMDE: okay, thanks. [17:52:14] nileshc_away: a quick hadoop intro would be great - can you write me an email? i'll find you on irc later with any questions [17:54:20] (03PS1) 10Daniel Werner: Remove "Wikidata" specific handling of copyright message introduced in Ie7dba11 [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 [17:57:30] (03CR) 10Jeroen De Dauw: [C: 031] Remove "Wikidata" specific handling of copyright message introduced in Ie7dba11 [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [17:58:24] (03PS2) 10Daniel Werner: Remove "Wikidata" specific handling of copyright message introduced in Ie7dba11 [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 [18:00:19] (03PS1) 10Daniel Werner: (bug 47435) Set expire time for cookies for dismissed copyright notice. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77155 [18:05:32] DanielK_WMDE: Alrighty, I'll write you an email. [18:07:52] DanielK_WMDE, Lydia_WMDE: I'm confirming that I've just submitted by midterm report in melange. [18:17:01] (03CR) 10Aude: [C: 04-1] "so how will it work for wikidata now? what do we need to do differently?" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [18:42:03] (03CR) 10Siebrand: [C: 031] "L10n reviewed." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/76929 (owner: 10Tobias Gritschacher) [18:58:44] (03CR) 10Denny Vrandecic: [C: 04-2] "Only submit if there is also a matching changeset in the configuration so that there is no change on Wikidata.org." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [18:59:44] (03CR) 10Denny Vrandecic: "And make sure that the work provided by the translators already is not lost." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [19:19:39] (03CR) 10Daniel Werner: "@Aude/Denny:" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [19:24:31] Denny_WMDE: you might want to check your inbox, we had a situation regarding sorting this afternoon which requires some more discussion tomorrow. Seems like we made some wrong assumptions yesterday :-/ [19:28:35] (03PS1) 10Daniel Werner: (bug 52412) dismissing copyright warning no longer bound to interface language [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77167 [19:31:28] (03CR) 10Denny Vrandecic: "It is worth a -2 because it must not be merged before it is ensured that it will not cause a disruption on the site and before it is ensur" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [19:40:31] (03CR) 10Daniel Werner: "I see. In my understanding, -2 means the whole concept of the change set is flawed. With a -1 the change set will already not be submitted" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [19:40:31] Danwe_WMDE: saw it, answered. [19:43:01] (03CR) 10Denny Vrandecic: "For me it says "-1 There's a problem with this change, please improve". And "-2 Do not submit". And given that the necessary conditions fo" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [19:45:10] Danwe_WMDE: thanks [19:48:10] Denny_WMDE: I know you discussed the actual handling of this in the UI with Henning. But it was just after you were gone that we noticed that we can not do this without backend changes anyhow. [19:50:29] ah, that's true [19:50:41] but yeah, as you say, the order needs to be encoded [19:55:39] (03CR) 10Aude: "how about make the message key to be used here customisable? a setting? Then for wikidata, we can change it and not have the wikidata-sp" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [19:58:44] (03CR) 10Daniel Werner: "I still don't see the problem. Making a message key configurable seems totally overkill. Simply move the messages on Wikidata. done." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [20:02:04] (03CR) 10Daniel Werner: "moving http://www.wikidata.org/wiki/MediaWiki:Wikidata-shortcopyrightwarning to http://www.wikidata.org/wiki/Wikibase:Wikidata-shortcopyri" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [20:04:01] (03CR) 10Daniel Werner: "Sorry, the 2nd link had to be http://www.wikidata.org/wiki/Wikibase:Wikibase-shortcopyrightwarning of course." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77154 (owner: 10Daniel Werner) [20:10:21] (03PS1) 10Daniel Kinzler: (bug 52325) validators for url schemes. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77183 [20:11:16] (03CR) 10jenkins-bot: [V: 04-1] (bug 52325) validators for url schemes. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77183 (owner: 10Daniel Kinzler) [20:12:19] wtf jenkins [20:20:36] (03PS2) 10Daniel Kinzler: (bug 52325) validators for url schemes. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77183 [20:21:17] Danwe_WMDE: the messages usually don't come from the wiki, but from translatewiki, where they are translated. [20:21:32] Danwe_WMDE: only /en in this case was locally overwritten [20:23:38] also I am unsure that wgRightsUrl is the right URL, because of the two licenses for wikidata [20:24:10] Denny_WMDE: so the actual reason why we have a "wikidata-" prefixed version is that otherwise translatewiki can't handle it because they are not able to use the same message key in the scope of different projects? [20:24:39] Danwe_WMDE: that might be a problem [20:24:49] i don't know [20:24:53] No, in case of wikidata wgRightsUrl is not the right url for wikidata I guess. But that's one of the reasons why the message is overwritten in Wikidata [20:25:08] wgRightsUrl is easy to change [20:25:58] errr, not sure because we have creative commons (for wikitext) and cc-0 / pd for data [20:26:08] aude: exactly [20:26:20] Danwe_WMDE: yes, there were some issues with that [20:26:28] i'd have to look at the code and test the setup [20:26:30] (03CR) 10jenkins-bot: [V: 04-1] (bug 52325) validators for url schemes. [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/77183 (owner: 10Daniel Kinzler) [20:26:56] once they are resolved, as said, it's fine to remove it. but if it would have been that trivial, we would have done it right back then [20:27:06] it was one of my evil hacks, and I admit it [20:28:00] but it will actually never surface in a non-wmf wiki, so I am not sure whether the benefit is sufficiently big to warrant the work in order to get this fixed [20:28:47] This is not the first time translatewiki.net restrictions cause trouble. [20:30:23] I wonder how it is done on Wikipedia.org, can't imagine we don't have overwritten system messages there. Or aren't they managed by translatewiki in the same fashion there perhaps? [20:45:08] wikimedia specific messages come from the WikimediaMessages extension [20:45:15] that's the same thing with wikibase [20:45:43] the wikidata-shortcopyrightnotice message is provided through Wikimedia-Messages, not through wikibase [20:46:04] that is why the if clause only ever gets fired if you have the WikimediaMessages extension installed as well [20:46:34] which is why this should never do anything for any other wikibase instance [20:46:43] Danwe_WMDE: ^^ [20:47:37] and what do you want to say with "this is not the first time translatewiki.net restrictions cause trouble"? are you suggesting to deal with translation in another way? [20:47:38] Denny_WMDE: we ought to avoid "if is wikidata.org" checks, that is really not good practice [20:48:04] Denny_WMDE: isn't defining messages in the "Mediawiki" namespace a core feature? [20:49:29] Danwe_WMDE: it is. That is what is used for EducationProgram for instance [20:49:31] JeroenDeDauw: I agree with that. and I said that once all issues are fixed, we can remove that if clause. but our primary goal is not the software, but the site. please do not forget that. [20:49:54] Denny_WMDE: translatewiki restrictions causing troubles in terms of translatewiki not being able to deal with the same massage keys appearing in different projects. [20:50:04] [travis-ci] wikimedia/mediawiki-extensions-WikibaseDatabase#36 (master - 94924d1 : Translation updater bot): The build has errored. [20:50:04] [travis-ci] Change view : https://github.com/wikimedia/mediawiki-extensions-WikibaseDatabase/compare/260e583ca12d...94924d13f28d [20:50:04] [travis-ci] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-WikibaseDatabase/builds/9751522 [20:50:05] Danwe_WMDE: but defining messages in the MediaWiki namespaces does not allow them to be translated through translatewiki. [20:50:12] lolwat [20:50:40] Denny_WMDE: Denny_WMDE that's why i suggest make it configurable [20:50:46] [travis-ci] wikimedia/mediawiki-extensions-WikibaseDataModel#48 (master - caf98a5 : Translation updater bot): The build passed. [20:50:46] [travis-ci] Change view : https://github.com/wikimedia/mediawiki-extensions-WikibaseDataModel/compare/59db3764309a...caf98a550382 [20:50:46] [travis-ci] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-WikibaseDataModel/builds/9751535 [20:50:48] not ideal but would work [20:51:08] then we don't need to move messages or anything [20:51:28] aude: i am fine with such a solution. but i require it before we merge Danwe_WMDE's patch :) [20:51:57] huh? that build didn't fail, did it? [20:51:59] Denny_WMDE: i don't care what solution, as long as it works and is not too much hassle for translators [20:52:13] aude: agreed. [20:52:41] I still don't get the actual problem. I thought it was all about translate wiki not being able to manage project scopes but perhaps I missunderstand? [20:53:45] Danwe_WMDE: where should the translated messages for wikidata-shortcopyrightnotice come from? [20:54:32] Denny_WMDE: from Translatewiki, "wikidata-shortcopyrightnotice" in the Wikidata.org project scope. [20:54:37] sorry [20:54:44] "wikibase-shortcopyrightnotice" [20:54:56] there is no wikidata.org project scope on translatewiki [20:55:04] which seems to be the problem [20:55:21] well, are you suggesting to make one for this one message? [20:55:25] assuming there is no two equal message keys in all the projects they support is rather optimistic [20:55:41] Danwe_WMDE: well, do you want to discuss that with me? [20:56:16] Danwe_WMDE: that's why our messages are prefixed with wikibase- [20:56:21] and other extensions do likewise [20:56:25] Denny_WMDE: no, but acknowledging that this is the actual problem would help understand [20:57:07] it is only a problem if you want to use the same message key in two different extensions [20:57:11] which would not work anyway [20:57:18] and same prefix [20:57:33] Denny_WMDE: what site? [20:57:43] aude: prefixed with "wikibase-" because of the MW messaging system (which should be another can and independent from mangagement of messages in translatewiki) [20:57:47] i don't know if it's possible for wikimediamessages to have the same message, but maybe [20:57:57] i can't say it's impossible... just don't know [20:58:37] JeroenDeDauw: wikidata.org and the other wikimedia project sites [20:59:11] aude: then it would be relevant to include them in the right order again [20:59:26] aude: because we must make sure that the right message overrides the wrong one at the right time [20:59:53] aude: and you told me dependency on inclusion order should be avoided [21:01:00] now I'm not sure if I should feel trolled or not [21:01:09] JeroenDeDauw: ? [21:01:26] JeroenDeDauw: maybe i misunderstood your question? [21:02:04] Denny_WMDE: I guess you missed the context, but nvm [21:02:12] * JeroenDeDauw goes off to delete some evil tags [21:02:41] http://translatewiki.net/wiki/MediaWiki:Wikibase-shortcopyrightwarning/en [21:02:53] looks like there can be just one [21:03:58] Denny_WMDE: and right, relying on inclusion order would be bad, though not quite the same problem [21:05:20] i dropped. so if anyone said something, i missed it [21:07:36] Denny_WMDE1: that i checked translatewiki and there can be only one wikibase-shortcopyrightwarning [21:08:01] and agree, relying on inclusion order should be avoided, though it's a different problem [21:08:02] ah, thanks. yeah, so this is one of the reasons [21:08:38] [travis-ci] wikimedia/mediawiki-extensions-Wikibase#514 (master - 41ef235 : Translation updater bot): The build has errored. [21:08:38] [travis-ci] Change view : https://github.com/wikimedia/mediawiki-extensions-Wikibase/compare/8477c23a82f9...41ef235bc680 [21:08:38] [travis-ci] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-Wikibase/builds/9751760 [21:09:11] what's up with travis? [21:09:20] hiccup on their side, I assume? [21:10:12] could be github is down? or access to github [21:10:18] * aude clicks to look [21:10:51] Could not open input file: maintenance/update.php [21:10:54] no idea [21:11:06] The command "bash .travis.sh $EXTENSION" failed and exited with 1 during before_script. [21:11:24] Danwe_WMDE: please to merge https://gerrit.wikimedia.org/r/#/c/76924/ ? [21:11:59] Denny_WMDE1: aude: looks like something went wrong with fetching some file [21:12:04] ok [21:12:16] Given that wmf git just had some serious issues, I guess it is probably not travis fault [21:12:32] I had some failing git commands locally [21:12:35] (03CR) 10Daniel Werner: [C: 032] Remove dead code [extensions/Diff] - 10https://gerrit.wikimedia.org/r/76924 (owner: 10Jeroen De Dauw) [21:12:38] ah, ok, but the actual tests are not failing [21:13:30] err, i mean, they are not failing because they are actually not run [21:14:44] Denny_WMDE1: right [21:16:58] Indeed, the build failed at cloning mw [21:17:06] Last SMW build ran into the same issue [21:17:22] Perhaps I should have it clone from the github mirror, much more reliable [21:18:06] could be, although github tends to have ddos and other issues sometimes, too [21:18:22] why people would ddos github, don't understand.... [21:18:47] aude: They don't like githubs unlimited diskspace for users? :P [21:20:17] aude: secretly it is you DDoS'ing GH, since you want everyone to switch to WMF git. Your evil plans are out now, I'm going to submit them to wikileaks [21:20:54]