[00:00:35] Hey. I just clicked a Wikidata link and was taken to http://imm.io/18VQ9 [00:00:44] legoktm: ^ [00:01:02] ugh [00:01:05] er [00:01:09] is that a screenshot? [00:01:27] legoktm: yes [00:01:30] of what I got [00:01:39] when I clicked the ru link at https://www.wikidata.org/wiki/Q4451759 [00:01:55] Sven_Manguard: yes, aude has a patchset waiting deployment [00:02:04] legoktm: well, it gets worse [00:02:21] I went to ru.wikipedia and put Тарасовка (Киево-Святошинский район) into the searchbar and got that too [00:02:37] Sven_Manguard: Yeah. [00:02:54] what is causing that? [00:03:31] idk, but https://gerrit.wikimedia.org/r/67900 is the hotfix [00:03:34] it's not happening when I click the uk.wiki link on the same page [00:03:57] maybe a template is loading Lua which parses the item? [00:04:00] idk [00:04:44] and it's taking out a page on Russian Wikipedia? [00:04:48] dammmmn [00:05:09] ¬_¬ [00:05:26] i bet if you remove the coordinate property itll fix it [00:05:41] oh, damn [00:05:43] that's lua [00:06:34] * aude fixing [00:06:52] :D [00:06:59] I don't even have Scribuntu installed on my test setup [00:07:12] * hoo heads off to do that [00:13:44] New patchset: Aude; "Handle failure to deserialize entity in lua" [mediawiki/extensions/Wikibase] (mw1.22-wmf5) - https://gerrit.wikimedia.org/r/67932 [00:13:59] hoo: does that look sane to you? [00:14:16] * aude has lua [00:14:24] I've just installed it [00:14:31] ok [00:17:24] Jenkins is crying about that change :/ [00:17:33] ignore [00:17:41] :D [00:17:42] it's testing with datavalues master and old branch of wikidata [00:17:48] :/ [00:17:51] wont' work [00:18:26] It looks sane and I guess you tested it?! [00:18:48] I can force the merge, if ou want [00:19:32] iam testing [00:22:26] it's good [00:22:46] sure the error handling could be more elegant but it works fine [00:23:28] or it's my lua script that's not perfect [00:23:30] aude: why is http://imm.io/18VYe happening? [00:23:48] Sven_Manguard: that i don't know [00:24:07] well it's an unnecessary precision loss [00:24:13] agree :/ [00:24:17] aude: Will merge... just looked a bit through the Scribunto code and this looks good [00:24:20] i can't do anythign tonight [00:24:23] hoo: thanks [00:24:30] i'll backport and poke someone [00:25:12] New review: Hoo man; "Looks sane and follows the Scribunto documentation." [mediawiki/extensions/Wikibase] (mw1.22-wmf5); V: 2 C: 2; - https://gerrit.wikimedia.org/r/67932 [00:25:13] Change merged: Hoo man; [mediawiki/extensions/Wikibase] (mw1.22-wmf5) - https://gerrit.wikimedia.org/r/67932 [00:25:37] aude: also it would be nice if Wikidata could take the curly apostrophes and quote marks from geohack (see upper left of https://tools.wmflabs.org/geohack/geohack.php?pagename=Kesteren_railway_station¶ms=51_55_34_N_5_38_13_E_type:railwaystation_region:NL ) and convert them into ones that Wikidata can use [00:26:12] if you straight cut and paste in from geohack, Wikidata won't accept it until you manually replace the curly apostrophes and quote marks with straight ones [00:27:47] Sven_Manguard: Are you suggesting that the Javascript should escape and replace the characters? [00:28:53] Sky2042: doesn't it already do that with other things? autoconvert? [00:29:18] Sven_Manguard: I think it does, I was just wondering if that's what you had in mind. [00:29:57] I mean, on the plus side, I never would have discovered that the decimal to minute/second conversion was broken if it weren't for the inability to just straight cut/paste in the geohack curly apostrophes [00:41:56] aude: in answer to my earlier question, the reason is that apparently the automatic setting was to round to the arcminute, and I wanted to express a value that had arcseconds [00:42:05] so I overwrote the default [00:42:14] hmmmm [00:42:17] but that's still an issue [00:42:39] default needs to be at least arcsecond, if not 1/10th arcsecond [00:42:51] and that's the only time I had that issue [00:43:36] seems reasonable [00:51:11] aude: well it's doing it again [00:51:37] whenever I cut paste in from geohack and then swap out the curly for the straight apostrophes [00:54:33] rschen7754: ping [00:54:38] hi [00:55:05] when you added the geocoords to https://www.wikidata.org/wiki/Q2504328 how did you do it? [00:55:12] it's a known issue [00:55:15] it was super annoying [00:55:16] the apostrophe [00:55:27] first i copy-pasted and it didn't work [00:55:28] rschen7754: did you cut and paste? [00:55:49] then i dropped all the degrees and added it manually [00:55:59] all the minutes and seconds that is [00:56:02] yeah, what I'm getting at is that the location you entered is 32°55'27"N, 96°45'49"E. The interchange is at 32°55'27"N, 96°45'49"W. [00:56:06] dms should certainly work [00:56:18] I wanted to know if it was the software or you that made that mistake... [00:56:24] but decimal degrees appear to work better at the moment [00:56:34] i think they might still get rounded though [00:56:51] unless you pick the right precision [00:57:05] https://bugzilla.wikimedia.org/show_bug.cgi?id=49386 about the apostrophe [00:59:30] Sven_Manguard: no idea [00:59:46] aude: yeah i ran into that issue [00:59:55] is this live on test.wikidata/ [01:01:40] rschen7754: should be [01:01:48] should be same version of everything now [01:02:02] hm, so if i could just get PC on test.wikidata... [01:02:06] go ahead [01:02:15] rschen7754: I'll give it [01:02:17] oh i can't give you that though [01:02:22] * aude does not have that right :o [01:02:53] {{done|color=pink}} [01:02:53] How efficient, color=pink! [01:02:55] oO [01:03:08] thx [01:03:09] aude: The devs. should get 'crat access there IMO [01:03:24] just creating properties is fine [01:03:40] should i go file a bug to un-restrict PC on test? [01:03:40] hoo: agreed, but I can't do anything about that [01:03:44] https://test.wikidata.org/wiki/Property:P10 [01:03:48] or can someone here just do it? [01:04:16] I don't think we'd need some kind of consensus, do we? :P [01:04:16] Reedy could [01:04:24] probably not tonight [01:04:25] But he's probably sleeping now [01:04:32] hoo: idk. the reason for restricting it doesn't really exist on test. [01:04:33] the stewards can grant access [01:04:33] not thaaaat important [01:04:43] indeed [01:04:57] aude: yeah, but it is kinda annoying that people keep on having to request it [01:05:05] * aude wants bot rights on test and will need to request [01:05:18] just so i can test everything works fine [01:05:27] I'm going to ping Reedy tomorrow... I guess he's fine with giving a couple of +crats out [01:07:18] now i can't add the property to the item :/ [01:07:45] oh, it doesn't work over https [01:08:04] ewk [01:09:22] *sigh* [01:09:28] what? [01:10:19] ok, i think i figured it out… one of my scripts was conflicting [01:10:26] rschen7754: Works for on test with https? Or did you mean wikidata.org [01:10:27] ah ok [01:10:54] i have my global js disabled on regular wikidata [01:11:28] rschen7754: Me too ... I was having problems with it IIRC [01:11:33] ah, ok [01:12:04] hoo: How nice :) I'm a ... "user" ;) [01:12:19] rschen7754: Yay for using my javascripts :D [01:12:24] Hazard-SJ: mh? [01:12:25] :D [01:13:10] hoo: I think the [save | cancel] things were replaces with ||| or something like that [01:13:33] Sven_Manguard: https://test.wikidata.org/wiki/Q14 [01:13:43] I'm not going to investigate that now... but rather my nice and warm bed :D [01:14:07] Jasper_Deng: rschen7754 what [01:14:16] err, how'd Jasper wind up there [01:14:24] requested [01:14:37] yeah, that's how it looked when i entered it [01:14:47] drops the seconds automatically though [01:15:26] okay, so go to options [01:15:31] err [01:15:32] aude: Good night... tomorrow I'll have time to hack on the client move thing :) [01:15:41] advanced adjustments [01:15:59] and change percision to arcsecond [01:16:24] hoo: thanks for your help :) [01:16:32] yeah, it works when you do that [01:16:45] but the seconds are lost forever if you don't do that right off the bad [01:16:47] *bat [01:16:53] rschen7754: yes, which is an issue [01:20:10] aude: I found another weird issue [01:20:24] go to https://www.wikidata.org/wiki/Q2504328 and try to edit the coordinate there [01:21:35] when I tried to change E for W it changed the value from 32°55'N, 97°14'W to 32°55'27"N, 96°45'49"E [01:21:46] err other way around, from 32°55'27"N, 96°45'49"E to 32°55'N, 97°14'W [01:22:33] or 32°55'27"N, 97°14'11"W after I force it into arcsecond [01:23:41] it absolutely can't handle itself dealing with the direction west [01:23:43] it's freaky [01:26:14] grrrr [01:27:04] alright, it's almost getting to be morning here ! [01:27:17] * aude hears the birds chirping [01:27:22] * aude sleep :) [01:28:41] aude: you're not in NY anymore? [01:28:55] no, in berlin [01:29:00] o.O [01:29:19] * aude invades NY sometimes [01:46:51] aude: the deeper I go, the more problems I find with the way that geocoords are being handled [01:52:53] aude: rschen7754: legoktm: https://bugzilla.wikimedia.org/show_bug.cgi?id=49415#c2 [01:58:36] so basically we cant do anything west? [02:01:57] shouldn't west be a negative number? [02:02:35] (from a basic OOP programmer's perspective, it sounds like not all the fields of the location object are getting migrated accordingly) [02:07:28] 02:07, 11 June 2013 Sven Manguard (talk | contribs | block) deleted page Property:P625 (This propery is being temprarily deleted (by admin consensus via emergency IRC chat) due to errors in the datatype causing false outputs) [02:07:47] legoktm: west or south, so yes, it's a negative number issue [02:07:54] possibly / likely [02:08:21] ok, you should probably update the bug summary then [02:11:01] https://bugzilla.wikimedia.org/show_bug.cgi?id=49415#c2 [02:16:12] !log Properties P625 and P626 have been temporarily deleted following a conversation in wikidata-admin and the escalation of the bug's complexity and scope [02:16:51] :O [02:31:18] Pink|TV: awesome of you to use my image for the testwikidata [02:31:39] well thanks for creating it :) [02:31:54] :D [02:32:08] I never thought I'd be used, considering that it's an inverse [02:32:49] need some way to signify it's a crazy alternate universe ;) [02:33:18] i had someone whip up a black logo like for test.wp, but Reedy didn't like it [02:33:58] but yeah, it's a really cool image :) [02:34:00] have you ever seen the image on my enWiki user page? [02:34:29] I have a consistent aesthetic when it comes to those kinds of things. I like parallel rectangles doing interesting things [03:24:01] I'm having fun with the Wikidata logo http://imm.io/18WOm http://imm.io/18WPN http://imm.io/18WQs http://imm.io/18WR4 http://imm.io/18WS6 http://imm.io/18WSY [03:24:10] if you want any, take em [03:24:18] All my work is CC-BY-SA [03:37:25] TCN7JM: http://imm.io/18WWg [03:37:51] That's a bit too blurry for me. [03:58:39] TCN7JM: http://imm.io/18X0N and http://imm.io/18X0S then [03:59:38] Second one is great. [04:01:28] TCN7JM: okay, here's a fixed version then http://imm.io/18X1o [04:01:38] Define fixed. [04:02:07] the reason it was blurry in the beginning is that the distortion left lines [04:02:23] I thought I got them all in 18X0S but I missed a few [04:02:55] Ah. Cool. [07:35:15] what's the official way to run tests in wikibase? [07:35:28] I only tried to create a suites.xml myself [07:54:35] hello :-D [07:54:39] oh men, hmm [07:54:41] Guten Tag [07:59:05] liangent: tobi should be here soon and able to help you with that [08:23:45] liangent: wikibase tests don't run standalone but as part of the mediawiki test suite [08:24:16] liangent: go to tests/phpunit in your installation root. run php phpunit.php --group Wikibase [08:58:32] DanielK_WMDE: hangout? [09:24:46] New patchset: Daniel Kinzler; "Fix string length check in StringtValidator." [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/67956 [09:24:52] Change abandoned: Daniel Werner; "seems to work again with current MW core" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/67574 [09:28:49] New review: Daniel Kinzler; "(1 comment)" [mediawiki/extensions/Wikibase] (mw1.22-wmf5) - https://gerrit.wikimedia.org/r/67932 [09:31:36] https://bugzilla.wikimedia.org/show_bug.cgi?id=49425 looks bad but i am not sure who'd need to take a closer look [09:33:38] New review: Daniel Kinzler; "(1 comment)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/67862 [09:33:45] New review: Daniel Werner; "(1 comment)" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/67404 [09:34:41] New review: Aude; "(1 comment)" [mediawiki/extensions/Wikibase] (mw1.22-wmf5) - https://gerrit.wikimedia.org/r/67932 [09:34:46] DanielK_WMDE: you haz comment [09:34:56] lua does not need to serialize the entity? [09:35:13] aude: i don't know - does it? what for? [09:35:20] i did not chang ethat [09:35:28] anyway, it's serializing but the error sais deserializing... c onfusing :) [09:35:38] you are right, but it was 2am [09:35:56] aude: sure, not a big deal. I'm just wondering why it is serializing at all. [09:36:04] * aude can correct that in patch for wmf6 and master [09:36:04] perhaps ask jens when you see him around [09:36:07] sure [09:36:18] would be good to have that fixed in master, i don't care about wmf6 [09:36:20] * aude needs to poke at lua more [09:36:22] it's a minor inconsistency [09:36:28] we need it in wmf6 [09:36:30] just in case [09:36:39] although wmf6 knows about globe [09:36:49] Henning_WMDE: Danwe_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=49425 <- please have a look - looks bad [09:37:01] * aude won't be around on thursday either, btw [09:37:19] we need the error handling, sure. i meant that it'S not too horrible to be mistakingly referring to "deserialzation" in an error message [09:37:30] DanielK_WMDE: you are right :) [09:37:36] Lydia_WMDE: known issue, patch on gerrit [09:37:40] will link the bugs [09:37:43] thx [09:37:46] i'll just call it serialize-error [09:37:48] which it is [09:37:55] right [09:38:59] you'll be around on thursday evening right [09:39:00] ? [09:39:25] there should be no submodule changes at that point, but just in case there are any problems (unlikely now) [09:40:40] aude: yes, i'll be in the office [09:40:58] Lydia_WMDE: https://gerrit.wikimedia.org/r/#/c/67816/ [09:41:14] aude: perhaps that should be backported ----^ [09:41:24] should be possible without the dependencies [09:41:45] DanielK_WMDE: can look [09:41:47] thanks [10:00:57] Tobi_WMDE: Henning_WMDE which patches? [10:01:22] https://gerrit.wikimedia.org/r/#/c/67257/ [10:01:26] k [10:02:52] New patchset: Aude; "Fix for globeCoordinate's toDegree()" [mediawiki/extensions/DataValues] (mw1.22-wmf6) - https://gerrit.wikimedia.org/r/67959 [10:06:10] aude, Henning_WMDE: why are we using degree/minute/seconds internally at all? [10:06:19] wouln't two floats be much simpler? [10:06:50] or use the system that was originally used by the user... [10:07:01] which would probably mean storing the original string [10:07:32] * aude would prefer a field for latitude and one for longitude [10:07:44] too much change though from status quo [10:07:46] New patchset: Nikerabbit; "Fix string length check in StringValidator." [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/67956 [10:08:17] https://gerrit.wikimedia.org/r/#/c/67816/ does not cherry pick cleanly [10:08:20] DanielK_WMDE: [10:08:23] * aude merging the other one [10:08:25] aude: not really, the DataValue just needs to accept the old and the new array representation [10:08:32] that's not pretty, but easy enough [10:08:51] DanielK_WMDE: sure [10:08:54] aude: just ignore the test [10:09:01] Change merged: Aude; [mediawiki/extensions/DataValues] (mw1.22-wmf6) - https://gerrit.wikimedia.org/r/67959 [10:09:18] the test depends on fixing the test infrastructure for data values, which was pretty broken [10:09:23] but we don't need to backport the test [10:09:24] would you make a patch for the time validation in wmf6 [10:09:36] sure, just patch the time value thing [10:09:54] since we already backporting in datavalues, this is not a problem [10:10:00] you mean git checkout wmf6 && git checkout wmf5 -- path/to/TimeValue.php? [10:10:01] sure :) [10:10:09] we don't need wmf6 [10:10:13] err, wmf5 [10:10:15] just wmf6 [10:10:32] right sorry, should be "master" there, not wmf5 [10:10:47] right [10:20:09] New patchset: Daniel Kinzler; "(bug 49264, bug 48965) validate time format." [mediawiki/extensions/DataValues] (mw1.22-wmf6) - https://gerrit.wikimedia.org/r/67962 [10:20:15] thanks [10:20:47] aude: i have only unit tested this, not tried to supply garbage via the api [10:21:15] noted [10:25:07] Lydia_WMDE, aude: i have a GSoC call in my calendar for 2pm... is that going to happen? Who is goping to attend? [10:25:25] DanielK_WMDE: yes it is going to happen [10:25:28] * aude be there [10:25:30] attendees: gsoc students, you, kaite and me [10:25:35] *katie [10:30:48] Lydia_WMDE: ah, ok. [10:33:42] DanielK_WMDE: your patch is no good :( [10:33:52] it for example does not let me remove values [10:34:03] aude: err? [10:34:27] i have no idea how that would be related [10:34:37] maybe i had badly formatted dates [10:35:06] and the validation prevents them from being removed? [10:35:10] I suppose that could happen [10:35:13] yes it did [10:35:23] * aude not convinced my dates are badly formatted either [10:35:36] * aude removes one added via the interface [10:35:45] is the ui putting in badly formatted dates? [10:35:48] basically, you cannot instantiate a TimeValue with an invalid string, so you can't instantiate a PropertyValueSnak for a broken date [10:35:50] and my bot? [10:35:58] then the UI is doing it wrong [10:36:01] i have no idea [10:36:03] * aude thinks [10:36:16] * aude does https://www.mediawiki.org/wiki/Extension:Wikibase/API#wbcreateclaim [10:36:19] perhaps change the error to include the actual value [10:36:22] that's what my bot does and it does not like it [10:36:54] i can't save dates [10:36:59] via the UI [10:37:08] wtf [10:37:12] ok, we can investigate but not backporting [10:37:14] * DanielK_WMDE is not going to get any work done today [10:37:19] :/ [10:37:25] better the way it is, as much as it sucks [10:37:59] New review: Aude; "i can't add dates via the UI or with my bot following the example https://www.mediawiki.org/wiki/Ext..." [mediawiki/extensions/DataValues] (mw1.22-wmf6) C: -1; - https://gerrit.wikimedia.org/r/67962 [10:38:15] if this is in master, i don't understand [10:38:36] does it really work in master? [10:39:11] dev repo seems fine [10:40:00] i have not tested extensively using the UI or the API [10:40:10] just seems incompatible with the branch [10:40:14] o_O [10:40:18] i'll check that again [10:40:19] * aude confused [10:40:23] the regex itself seems fine [10:40:33] it's in master so must be on dev repo [10:40:39] try print ( preg_match( '!^[-+]\d{1,16}-(0\d|1[012])-([012]\d|3[01])T([01]\d|2[0123]):[0-5]\d:([0-5]\d|6[012])Z$!', '+00000002010-01-02T00:00:00Z' ) ) ? "yes\n" : "no\n"; [10:40:40] El búfer 01 está vacío. [10:41:10] yes [10:41:22] the patch doesn't really do much else [10:41:39] * aude nods [10:41:41] i don't get it [10:42:03] http://www.wikidata.org/wiki/Wikidata:Project_chat#deleted_by_127.0.0.1_without_deletion_log.3F <- any idea what is going on there? [10:42:06] * aude imho would also validate the calendars [10:42:07] btw [10:42:33] easy for bots to put in the wrong thing like the /wiki/Qxxx page [10:42:40] instead of /entity/Qxxx [10:42:43] aude: there's DataValue valiudation, and there'S DataType validation [10:42:51] mmmm, ok [10:43:06] it is part fo the value [10:43:09] part of [10:43:10] DataValue is very basic, doesn't know about config or properties or high level types [10:43:33] yes, but as far as the DataValue is concerned, globe is a string. could be anything. [10:43:36] i suppose it gets combined at some point? [10:43:42] we could restrict the syntax a bit, but that wouldn't help [10:43:44] * aude talking atbout calendar [10:43:45] about [10:43:54] same thing [10:44:00] this might be at the api level or somewhere else then [10:44:01] globe and calendar [10:44:04] not for today [10:44:09] on the DataValue level, these are just arbitrary identifiers [10:44:15] but icky that the api accepts whatever [10:44:20] ok [10:44:24] validation of the URIs can only happen on a higher level [10:44:28] ok [10:44:39] i'm working on that, but it needs quite a bit of infrastructure and refactoring [10:44:44] fine [10:44:50] (that'S what i originally intended to do today...) [10:44:53] anyway :) [10:45:02] * aude working on https://bugzilla.wikimedia.org/show_bug.cgi?id=49417 [10:45:07] that needs another backport [10:45:50] * aude thinks code freeze and branching on wednesday is not enough time for deployment on monday [10:46:15] would be if we had actual copmmunity testing [10:47:45] anyway, i'll update test with henning's patch [10:47:52] and fix 49417 [10:49:33] aude: i can created, edit and remove dates via the UI with my patch applied to wmf6# [10:49:43] hmmmm [10:59:18] DanielK_WMDE: http://dpaste.com/1228627/ [10:59:24] now when just viewing an item [10:59:35] * aude wonders how i can have such different results on my wmf6 branch [11:14:50] DanielK_WMDE: that still fails.. [11:16:13] previously I had to put ../../extensions/Wikibase/lib/tests/phpunit in suites.xml to run it, so with --group=Wikibase, how does it know where tests are? [11:17:29] New patchset: Tobias Gritschacher; "Fix precision detection in globe coordinate parser" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/67967 [11:18:42] New patchset: Tobias Gritschacher; "Fix precision detection in globe coordinate parser" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/67967 [11:19:55] DanielK_WMDE: Hi! [11:19:57] I've just made a change to add support of unicode quotes that are used in Wikipedias. https://gerrit.wikimedia.org/r/#/c/67966/ [11:19:59] Is it possible to deploy it at the same time as the other bug fix for the globeCoordinate in order to avoid the bad experience to be unable to copy/paste from Wikipedia? [11:20:07] aude: this solves the precision issue (cutting off seconds): https://gerrit.wikimedia.org/r/#/c/67967/ [11:20:43] aude: can you please check with your examples and confirm it is solve. we can then backport this change as well. [11:21:15] at least we have two main issues solved then. [11:21:39] ok [11:22:05] Tpt_: Tobi_WMDE will have a look after lunch [11:22:07] thank you! [11:22:27] Tpt_: thx, will have a look at it! [11:22:40] Lydia_WMDE, Tobi_WMDE: Thanks! It was a very annoying bug. [11:23:24] [03:42:03 AM] http://www.wikidata.org/wiki/Wikidata:Project_chat#deleted_by_127.0.0.1_without_deletion_log.3F <- any idea what is going on there? <-- translate extension, known bug [11:23:55] sounds like translate [11:24:26] Lydia_WMDE: An off topic question: is Wikivoyage support is on scope of the next too weeks or will happen later? [11:25:59] New patchset: Henning Snater; "Improved GeoCoordinate parsers' coordinate detection" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/67968 [11:27:51] New patchset: Aude; "(bug 49417) Skip any change in client change handling that failed for some reason" [mediawiki/extensions/Wikibase] (mw1.22-wmf5) - https://gerrit.wikimedia.org/r/67969 [11:29:02] DanielK_WMDE: https://gerrit.wikimedia.org/r/#/c/67969/ [11:29:36] i'm sure this could be more elegant but think this is the place to be sure to catch anything that goes wrong [11:36:07] DanielK_WMDE: i am getting the Exception Caught: $time needs to be a valid ISO 8601 date when saving a geocoordinate :/ [11:36:18] * aude has evil stuff in my item [11:36:35] can't remove it [11:36:58] * aude has stuff with "The value does not comply with the property's definition." [11:37:03] can't remove those [11:37:11] * aude is on master [11:37:51] *gives coffee to aude* [11:38:01] :) [11:38:12] aude: just kill the item [11:38:23] yes, but we have items like that in wikidata [11:38:30] how do we kill those? [11:38:34] aude: "like that"... like what? [11:38:35] maybe via the api [11:38:38] sure we do [11:38:55] "The value does not comply with the property's definition." means that the property definition changed [11:38:56] sure there is garbage in wikidata [11:39:01] i don't think we have that on wikidata [11:39:14] no, it's a valid property with valid snaks [11:39:34] * aude looks for corrupt item [11:39:36] complaining about time when saving a geo-coord also points towards confusion about property types [11:40:02] well, it's saving the entire entity right? [11:40:06] trying to [11:40:07] aude: can you export the item in question, so i can look at it? [11:40:11] legoktm: thx [11:40:12] one sec [11:40:31] ok, you saw https://bugzilla.wikimedia.org/show_bug.cgi?id=49425 [11:40:35] yes, it's saving the entire entitiy... but i don't think it's instantiating all the snaks [11:40:37] it shouldn't [11:40:58] with http://www.wikidata.org/wiki/Q1825747 after we put your patch in / move to the next branch [11:41:06] when editing a claim, the other claims should stay in the json form and not be touched [11:41:18] that is unremovable, unless maybe by bot somehow [11:41:38] DanielK_WMDE: but the other claims became invalid [11:41:43] "ununserializable" :P [11:41:48] they were accepted before [11:41:48]