[00:52:59] If anyone intends on posting a comment at https://www.wikidata.org/wiki/Wikidata:Requests_for_comment/Bureaucrats%27_role_in_removal_of_permissions please do it now. I will be closing it "soon" [02:27:26] GWAH HA HA [02:27:32] GWAH HA HA [02:27:40] GWAH HA HA HA HA HA HA [02:27:43] GWAH HA HA HA HA HA HA [02:27:50] ....right [02:27:59] So IRCClould works now, I think [02:29:01] (gwah ha ha) [02:30:46] >.< [06:12:25] who can tell me the difference between a 'church' and a 'local church'? [06:12:35] someone needs a good smack on the bottom [06:13:13] the current moving base and the lack of place to read about it is becoming quite problematic [06:14:54] then changes happening to useful.js are perplexing and need communicating [06:17:20] well, useful.js is a userscript [06:25:22] so? [06:25:31] it is linked to be used from your pages [06:25:42] and it doesn't change the nature of comment [06:28:18] true [06:32:19] Zolo has been hassling there, so there is an underlying reason why [06:32:35] and that isn't communicated [06:33:10] a changelog would be a good idea [06:33:19] you can suggest it :) [06:39:44] Published patchset: Aude; "(bug 48473) validate claim guid in api and return error" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63787 [06:44:39] Published patchset: Aude; "(bug 48497) Check for offsetExists in ByPropertyIdArray" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63710 [06:50:59] Change merged: Tobias Gritschacher; [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/63695 [08:04:53] Change merged: Tobias Gritschacher; [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63656 [08:27:03] Change merged: Tobias Gritschacher; [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63664 [08:50:11] New review: Nikerabbit; "Hmm? Fixing a typo in the commit message does not warrant a mention in the commit message." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63653 [08:54:58] legoktm: did you make any progress getting a postgresql server? [08:56:41] Abraham_WMDE, Denny_WMDE1: i'm back. i'll wait in the hangout :) [08:56:52] oooh, a hangout? [08:58:22] DanielK_WMDE: give me a sec [08:58:31] aude: it's in the calendar [08:59:52] * aude working on https://bugzilla.wikimedia.org/48503 probably [09:00:07] then poke at failing tests in the client [09:00:14] and work on enhanced changes [09:00:33] can wait until your back to discuss and work further on dispatch changes and tests for it [09:01:03] though could do just a patch to move stuff from the dispatcher script to the separate class and then follow up after [09:06:46] Mineo: no sorry not yet. i've been rather busy lately so I haven't had the time to poke Coren (he would be the one who's able to do it) [09:10:42] no problem, I saw you talking about having exams :) [09:11:19] aude: for today's deployment, we should not forget to include the easyrdf submodule, right? or will that happen automatically? [09:11:31] not automatic, afaik [09:11:41] * aude not forget :) [09:11:41] ok, let's keep it in mind [09:24:05] aude: with the hackathon coming up, i'm not sure when i'll be in berlin again. perhaps we can do a hangout? [09:24:21] true..... [09:24:22] aude: but i do think small patches are good, so yea, start with factoring out *one* thing. [09:24:28] agree [09:24:46] it's just that if i move stuff, then the stuff still does not have tests [09:24:53] * aude does not want anyone to rage about it :) [09:26:11] New patchset: Tobias Gritschacher; "Make the RDF tests not fail when the RDF stuff is not installed" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63660 [09:29:22] aude: I will rage [09:29:29] heh :) [09:30:13] suppose selenium tests count, though [09:30:22] cucumber tests :D [09:35:05] DanielK_WMDE: let me know when you are available for 10 minutes [09:52:01] Denny_WMDE1: having an early lunch, will ping you wehn i'm done [09:59:18] ok [09:59:18] Tobi_WMDE: is https://gerrit.wikimedia.org/r/#/c/63660/ ready for review? [10:00:21] aude: yes, please check if tests pass when the submodule is there and that they are skipped correctly when it is not there [10:00:51] ok, doing once my maintenance script is done running [10:03:08] aude: running only the specific test behaved diffrent from running the whole group. that should be fixed now. please also check this.. [10:03:41] ok [10:18:41] New patchset: Denny Vrandecic; "trim incoming string values" [mediawiki/extensions/DataValues] (mw1.22-wmf4) - https://gerrit.wikimedia.org/r/63834 [10:19:10] aude: is it possible that https://bugzilla.wikimedia.org/show_bug.cgi?id=48112 is related to https://bugzilla.wikimedia.org/show_bug.cgi?id=48506 which you just filed? [10:19:21] legoktm: looking [10:19:58] legoktm: i don't think so [10:20:24] but hmmm.... how many entries in items per site are like that? [10:20:24] hmm, ok [10:20:35] i dont know, havent had the time to check [10:20:40] could be problematic [10:20:58] and wonder why [10:21:20] * aude theory for entity per page is these were missed when we added the table and ran the script [10:21:34] e.g. between time the script started and finished, maybe [10:21:57] this was in december and same time roughly as when the missing items were created [10:22:46] New patchset: Henning Snater; "(bug 48145) Input extender: Triggering "animationstep" event" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/63835 [10:23:04] apparently that table doesnt have a view on the toolserver :/ [10:23:10] it does not [10:24:30] New patchset: Henning Snater; "(bug 48145) Handling "animationstep" event in die value variation" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63836 [10:24:37] is there any private info in it? [10:24:41] legoktm: no [10:25:02] ok, ill send a mail to ts-admins [10:25:09] i think i opened a jira request for that a while ago. [10:25:18] it has a list of page ids (int), entity type ('item', 'property'), entity id (int) [10:25:19] well, i did for some tables. not sure wich ones exactly [10:25:21] New patchset: Denny Vrandecic; "trim incoming string values" [mediawiki/extensions/DataValues] (mw1.22-wmf4) - https://gerrit.wikimedia.org/r/63834 [10:25:23] Denny_WMDE1: now? [10:25:34] DanielK_WMDE: great, lemme try and find it [10:25:35] would also be nice to get a dump of these, but that's another bug ticket [10:26:38] DanielK_WMDE: yep [10:26:52] i'd like to quickly brainstorm on the following idea: [10:27:16] how hard would it be to write a php script that goes through the Wikidata dump [10:27:28] grabs the entity content from it [10:27:35] New review: Aude; "I get some failures when running the Wikibase group, with easyRdf missing." [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/63660 [10:27:36] pipes it through the RDF serializer [10:27:44] and makes an output file [10:27:52] the first two steps i understand [10:28:07] what i am curious about is the "pipe it trough the serializer" part [10:28:10] not hard at all. just very slow. [10:28:33] slow is fine. i assume 3-4 hours per run [10:29:04] do you have an idea on how to make this much faster? [10:29:37] or what would your idea be how to do it? [10:29:46] i'd assume much more than a few hours, sadly. [10:30:02] oh why? [10:30:06] one reason for this is value types. i think we made a mistake there. [10:30:29] data value types are stored inline. property types are not. [10:30:51] so, in order to generate the rdf, we have to look up the property's type for every property in every item [10:31:07] ok, caching helps with that, we'll just löoad all the properties into memory. but still [10:31:29] ok, got this [10:31:36] other than that, parsing xml with php is just sluggish, but not horrible [10:31:52] yep, that's why I am assuming a few hours :) [10:31:54] actually... maybe it's doable in a few hours. we'd have to give it a try. [10:32:05] i hope php can parse bzipped xml [10:32:11] yes, it can [10:32:27] and the infrastructure for generating the appropriate Content objects is already in core [10:32:57] Denny_WMDE1: there are also some more details to sort out. the "pipe to serializer" part is essentially: $serializer = new RdfSerializer(); $rdf = $serializer->serializeEntity(); [10:33:18] but at the moment, that generates a full document; in the XML case, including the XML header and all. [10:33:18] where is the entity handed over? [10:33:37] you just register a callback to the dump importer [10:33:46] basically, you install a handler that is called for every revision [10:34:00] ok, so what would be the better approach? [10:34:15] i give it a try and you improve it afterwards? [10:34:24] or you do it? :) [10:34:33] oh, the handler/callback approach is fine. EasyRdf just needs a "snippet" mode [10:34:55] currently, there'S just no way to tell it to output a graph as a *part* of a document. [10:35:09] that's probably the most tricky bit at the moment. [10:35:38] so, shall I have a first cut at it, or do you prefer to do it yourself? [10:35:53] not in a hurry for that, btw [10:36:05] although it would be really nice to have that by the end of the month [10:36:11] wrt the property types... if we take them from the database (or even memcached - it's already there, for the clients), that will probably be fine... but it's ugly. it should really use *only* the dump, so it would need to read the dump twice - once to load the property types, and once "for real" [10:36:32] we wouldn't be running that script in production [10:36:47] i don't think we have the db available [10:36:56] then we have to read the dump twice [10:37:03] yeah [10:37:03] or two dumps, one just for the properties [10:37:09] (faster that way) [10:37:42] nah, i can create a script that gets us al properties and their types in about 15 minutes [10:37:52] during the first run, you'd just feed the type info intop an InMemoryDataTypeLookup [10:37:53] i'll start with that [10:38:17] ...and then use that lookup in the serializer (it requires one) [10:38:27] yep [10:38:28] so... [10:38:37] how do we get the additional functionality into EasyRdf? [10:38:46] dunno yet [10:38:48] Ask upstream? Hack it in and submit upstream? [10:39:03] nick said it can be used for streaming already [10:39:03] did they react to the security patch at all? [10:39:07] oh yes, [10:39:14] they integrated it within a day [10:39:19] awesome [10:39:21] and are also working on the other things [10:39:26] we should rebase our fork, then [10:39:38] i'd suggest not today [10:39:43] as we want to deploy :) [10:39:44] hm, if he sais it can be used for streaming, i'd be curious to know how [10:39:45] but at some point [10:39:57] yeah, didnt check it out yet [10:40:00] New review: Aude; "not blaming these failures on having easyRdf or not. I still get them when I git init the submodule..." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63660 [10:40:08] perhaps not using the output-one-graph-per-entity approach we are currently using [10:40:22] need to check dunno [10:40:40] might need refactoring - a serializer callback registered in the rdf builder instead of a graph builder used by an "outer" serializer [10:41:15] i might give it an hour today [10:41:18] let's see :) [10:41:38] anyone willing to +2 this one? https://gerrit.wikimedia.org/r/#/c/63834/ [10:41:43] the change is rather… trivial :) [10:41:48] and in the wrong place [10:42:03] oh, i bet you want a comment there first explaining that [10:42:04] ok, ok [10:44:26] New patchset: Denny Vrandecic; "trim incoming string values" [mediawiki/extensions/DataValues] (mw1.22-wmf4) - https://gerrit.wikimedia.org/r/63834 [10:45:29] the normalization does seem to be in the wrong place, yes [10:45:42] and, i don't quite understand the change to the test case [10:45:58] but then, i don't understand how the test case works exactly anyway [10:46:40] * DanielK_WMDE thinks abusing data providers to generate test subjects is evil. [10:47:00] the test could check ' foo ' = false [10:47:09] instead of trimmign it and then making it true [10:48:02] aude: the test does not trim at all [10:48:14] denny trimmed the test case [10:48:20] yes [10:48:37] it could also check ' foo ' (input) = 'foo' (expected) [10:49:17] maybe just nitpicking for a evil hack backport though :) [10:49:24] unnecessary nitpicking [10:50:40] if no one else merges it, i'll run it with *all* the tests and then approve if all is okay [10:51:10] it's not a nitpick, it's an important thing to test [10:51:35] ...and if i understand correctly, the way the test is currently set up, it's not easy to even create a test case for this [10:52:05] which is why i think it'S evil to use providers to create test subjects, instead of sets of parameters, along with the expected result. [10:52:20] Denny_WMDE1: the test should actually check that the trimming is applied, and applied correctly. [10:53:50] Tobi_WMDE: yes, I would actually like the selenium tests to be run [10:53:59] there *might* be something depending on the not-trimming [10:54:06] the string value is very deep in the system [10:54:25] so if we could run all the tests, i would be very happy [10:54:30] and only then merge this [10:55:03] and it is just one in my series of "evil hacks by denny" [10:55:38] and this should NOT be at this place, but after discussing where it should be, i figured out if we want to have trimming before the month is over, this is how this will be done as a stopgap [11:04:48] Denny_WMDE1: trying to test your patch, but encoutering problems with JS, and finding that xdebug doesn't quite work on the new box... [11:04:58] taking longer than expected :) [11:05:17] generally, i'll approve it for the branch. this shouldn't go into master, though [11:09:35] Danwe_WMDE, Tobi_WMDE: Error: Unknown dependency: jquery.ui.toggler [11:09:39] what ma i missing? [11:14:36] DanielK_WMDE: sure you pulled ALL extensions? [11:14:46] yea. [11:14:52] oh! but that might be the problem! [11:15:04] i'm testing a patch for wmf4 - so i need to siwtch all extensions to that [11:15:06] will try... [11:18:37] Danwe_WMDE: ok, that worked :) [11:20:32] New review: Daniel Kinzler; "This should NOT go into master." [mediawiki/extensions/DataValues] (mw1.22-wmf4); V: 2 C: 2; - https://gerrit.wikimedia.org/r/63834 [11:20:32] Change merged: Daniel Kinzler; [mediawiki/extensions/DataValues] (mw1.22-wmf4) - https://gerrit.wikimedia.org/r/63834 [11:20:46] oh, right... [11:21:09] now that jenkins is running, are we still supposed to click "submit", or will jenkins merge after CR+2 if the tests pass? [11:26:13] DanielK_WMDE: CR+2 will merge if jenkins verified the changeset [11:39:42] Tobi_WMDE: ok, thanks [11:39:47] i'll stop merging myself, then [12:01:06] Abraham_WMDE: which hangout to use for story time? [12:02:37] sec [12:03:20] DanielK_WMDE: check the calendar, there is the link [12:04:42] DanielK_WMDE: I invite you [12:05:51] there's no link in my calendar entry. i'm using the one from your personal calendart [12:06:08] I sent you via PM [12:06:23] and invited you through g+ [12:21:54] Abraham_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=48453 [12:37:02] New review: Daniel Kinzler; "Works, but I'm worried about data providers returning null. It seems to skip the tests silently, ins..." [mediawiki/extensions/Wikibase] (master) C: 1; - https://gerrit.wikimedia.org/r/63660 [12:43:42] New review: Tobias Gritschacher; "(1 comment)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63660 [13:13:17] New patchset: Daniel Kinzler; "Revert "(bug 47610) Keep default settings out of globals."" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63845 [13:13:38] New patchset: Daniel Kinzler; "Revert "(bug 47610) Keep default settings out of globals."" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63845 [13:18:29] New review: Jeroen De Dauw; "Should throw an OutOfBoundsException, not return null." [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/63710 [13:42:09] aude: Ask is missing in https://gerrit.wikimedia.org/r/#/c/63833/ [13:42:21] Tobi_WMDE: ask has not been reviewed for deployment [13:42:25] ok [13:42:26] it's experimental [13:42:44] * aude hopes someone requests review for it soon [13:43:07] don't like it missing, however it's not required yet [13:47:36] Change merged: Jeroen De Dauw; [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/63796 [13:54:00] New patchset: Daniel Kinzler; "(bug 47610) Use $wgWBXXXSettings instead of $wgWBSettings" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63849 [13:55:36] Change abandoned: Daniel Kinzler; "as per aude" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/62453 [14:06:43] New patchset: Tobias Gritschacher; "Selenium tests for false edit-conflicts in claim editing" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63854 [14:29:12] New patchset: Aude; "(bug 48497) Check for property id exists in ByPropertyIdArray" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63710 [14:34:26] Denny_WMDE: tests for wmf4/wmf4 for statements ok [14:34:41] do we need to test any other combination again? [14:34:43] great, thank you Tobi_WMDE [14:34:50] no [14:34:53] ok [14:35:11] thank you [14:53:30] New patchset: Tobias Gritschacher; "Use composition instead of inheritance" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63857 [14:53:46] New patchset: Tobias Gritschacher; "Use composition instead of inheritance" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63857 [14:54:19] New patchset: Tobias Gritschacher; "Selenium: use composition instead of inheritance" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63857 [15:00:21] aude: deployment will happen around 8 again? [15:01:09] if true, I'll be around [15:13:18] New patchset: Henning Snater; "(bug 48145) Adding calendar hint message" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/63859 [15:21:28] New patchset: Henning Snater; "(bug 48145) Implemented input extender jQuery widget" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/62151 [15:21:28] New patchset: Henning Snater; "(bug 48145) Input extender: Triggering "animationstep" event" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/63835 [15:21:29] New patchset: Henning Snater; "(bug 48145) Adding calendar hint message" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/63859 [15:21:29] New patchset: Henning Snater; "(bug 48145) Introducing listrotator widget to set time precision" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/62585 [15:21:29] New patchset: Henning Snater; "(bug 48145) Simplified input extender widget" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/62586 [15:21:30] New patchset: Henning Snater; "(bug 48145) TimeInput expert: Applied toggler widget" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/63641 [15:21:30] New patchset: Henning Snater; "(bug 48145) Applying input expander widget to TimeInput expert [DO NOT MERGE]" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/62152 [15:21:30] New patchset: Henning Snater; "(bug 48145) TimeInput expert: Added listrotator to select calendar" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/63640 [15:39:57] New review: Jeroen De Dauw; "You are not testing if the exception is thrown correctly yet." [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/63710 [15:47:00] New patchset: Aude; "(bug 48497) Check for property id exists in ByPropertyIdArray" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63710 [16:03:49] New review: Jeroen De Dauw; "(1 comment)" [mediawiki/extensions/Wikibase] (master) C: 1; - https://gerrit.wikimedia.org/r/63710 [16:09:32] New patchset: Aude; "(bug 48497) Check for property id exists in ByPropertyIdArray" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63710 [16:09:56] New review: Aude; "(1 comment)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63710 [16:16:00] New patchset: Henning Snater; "(bug 48145) Assigning absolute position to snak type selector menu" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63865 [16:40:04] New patchset: Henning Snater; "Using Proper event namespaces in the snak type selector widget" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63869 [16:53:32] New review: Jeroen De Dauw; "-1 for the new behaviour not being tested at all" [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/63787 [18:00:40] dong [18:01:02] * aude waves [18:29:18] so, is the action going on somewhere else? [18:29:33] i think we might try localisation update first [18:29:50] since we have a new special page (entitydata), albeit not with much text [18:30:07] We need gerrit to work first :p [18:30:13] k [18:30:18] I'm still waiting for the wmf4 changes to push [18:30:26] * aude has learned from past mistakes [18:30:44] gerrit is crazy slow [18:31:15] Yes... it's a pain [19:03:53] fucks sake [19:04:03] tin doesn't have internet access [19:04:12] no internet access means no pulling from remote repos [19:04:15] :( [19:04:55] if there's no workaround, maybe we can put easyRdf into gerrit? [19:05:05] i was just going to suggest that [19:05:10] it'll replicate back to github anyway [19:05:13] it'd need to be a separate repo [19:05:21] That shouldn't be much of an issue [19:05:25] k [19:05:34] The workaround would be doing it on another server, and then copying across. every time it needed updating :/ [19:05:41] hmm [19:05:50] unless.. can git use a proxy for https? [19:05:56] not so nice [19:06:38] git config --global http.proxy $http_proxy [19:07:18] http://xmodulo.com/2012/04/how-to-use-git-behind-proxy-on-ubuntu.html [19:26:18] Sooooooo [19:27:09] proxy or put easyRdf into gerrit? [19:30:49] Proxy doesn't seem to want to work for HTTPS [19:32:58] then we can put it in gerrit [19:34:09] enwiki RFC closed (with the expected sensible answer) [19:34:26] https://en.wikipedia.org/wiki/Wikipedia:Requests_for_comment/Wikidata_Phase_2 [19:42:03] ^demon|lunch: Are you away? [19:42:19] <^demon|lunch> nom nom. [19:42:21] <^demon|lunch> sup? [19:42:59] tin hasn't got internet access, so can't clone easyRdf from the github repo wikimedia de have created [19:43:41] https://github.com/Wikidata/easyrdf_lite.git [19:43:47] <^demon|lunch> Dangit. Could make a repo on gerrit and pull from that. [19:44:04] That's what we were thinking was probably the easiest [19:44:33] agree [19:44:37] not sure where it'd go as it's not a MW extension itself [19:44:52] Create a wikidata top level folder? [19:45:04] <^demon|lunch> Easiest for permissions would just be a child of Wikibase. [19:45:32] <^demon|lunch> Wikibase/easyrdf [19:45:40] Guess that doesn't really matter as long as it's a seperate repo [19:45:53] that wouldn't interfere with the extensions since they are in mediawiki/extensions, right? [19:46:32] <^demon|lunch> It'd be fine, we do it for....VisualEditor I think? [19:46:38] Yeah [19:46:49] ok [20:12:39] hello there [20:12:46] Wikidata is broken [20:12:58] eptalon: Hm? Howso? [20:13:05] hi eptalon [20:13:10] What exactly? [20:13:32] logged in? logged out? which browser? [20:14:18] Go to http://simple.wikipedia.org/wiki/Plastic_explosive, klick on "add links", and try adding the english page. [20:14:41] Logged in,Firefox) [20:14:45] Oh, that's my widget.... [20:14:51] eptalon: Same name? [20:14:59] (the enwiki article, I mean) [20:15:09] hoo: should map to en:plastic explosive, yes [20:16:02] eptalon: https://www.wikidata.org/w/index.php?title=Q898359&curid=849197&diff=42665271&oldid=1488208 [20:16:14] I did that using the widget with Firefox 20 [20:16:46] Did you get an error message or smth.? [20:16:56] hoo: I get a message that the item is already linked to that other item. [20:17:14] Which other item? [20:17:57] Q898359 [20:18:06] ("plastic exlosive) [20:18:20] Isn't that the right item? [20:18:29] https://www.wikidata.org/wiki/Q898359 [20:18:31] yes. [20:18:56] error message is frustrating, because if its linked, I expect to see iw links [20:19:05] It wasn't linked [20:19:12] which i didn't until very recently [20:19:32] I just linked it [20:21:28] eptalon: So, you were on "http://simple.wikipedia.org/wiki/Plastic_explosive" clicked "Add links" and then it told you that it already is linked to an item and can't be linked? [20:21:55] FF 20.0.1, same game, (to Composite Compound 4) [20:21:59] Did it show you the table with other language links [20:22:02] ? [20:22:21] hoo: I could select en,typed the page, and got the error. [20:22:32] no table withother language links [20:23:02] eptalon: "Site link [[enwiki:C-4 (explosive)]] already used by item [[Q220775]]." that one? [20:23:03] [1] 04https://www.wikidata.org/wiki/enwiki:C%2D4_%28explosive%2913 => [20:23:05] [2] 10https://www.wikidata.org/wiki/Q220775 [20:24:18] The external client site did not provide page information [20:25:40] eptalon: I see the problem [20:25:42] logged in via https? [20:25:48] aude: Not the issue [20:25:52] k [20:26:02] eptalon: Try it with the enwiki article "C-4 (explosive)" [20:26:37] hoo: in what respect does it matter? [20:26:47] eptalon: Redirects... please try it [20:26:54] aude: wbgetentities isn't as smart as the special page with the same purpose as it seems [20:27:09] hmmmmm [20:27:10] That's why the widget tried to create new items (which of course failed) [20:28:13] hoo: that worked [20:28:22] Ah, great :) [20:28:23] but thats completely intransparent [20:28:29] Yeah [20:28:58] I'm just on implementing a SiteTitle class in lib, but am a bit stuck on architecture... [20:29:01] can't you silently follow the redirect, and link to that? [20:29:08] If we switch to that this should be resolved [20:29:24] hoo: in what way: stuck on arch? [20:29:49] if the thing the user selects is a redirect, you directly link to the redirected page. [20:29:53] eptalon: Of course we can... and I thought we already did... sadly this code isn't centralized yet, so that every module that takes user input has to do this at it's own... and wbgetenteties seems to don't [20:30:15] eptalon: Yes... I'm just a bit stuck on function names and specific functionality ;) [20:30:54] bang it on the wall until it does what you like it to do. otherwise, starving it also works wonders [20:31:00] aude: This is like https://gerrit.wikimedia.org/r/60760 ... shall we open a bug? [20:32:07] the user expects to get all pages, and does not care if it is a polka-dotted squirrel [20:32:44] neverending issues of normalizing foreign wiki page titles... [20:33:07] aude: Yes... that's why I'm on building a SiteTitle class for that to centralize the logic [20:33:12] good [20:33:18] so that it's only broken once, then :D [20:33:29] consistency would be very nice [20:33:38] +1 [20:34:08] hoo: and yes a bug for it would be good [20:34:26] don't think we have one yet [20:34:57] eptalon: I'm already working on that (you just showed me that it's more than just a code refactoring). I/ we hope to get this resolved soon ;) [20:35:11] aude: I don't think so as well... repo is weird here [20:35:21] Will you open one or shall I? [20:35:33] hoo: please do [20:35:41] or no, lemme do it [20:36:02] related ones might be https://bugzilla.wikimedia.org/show_bug.cgi?id=45282 [20:36:05] and https://bugzilla.wikimedia.org/show_bug.cgi?id=45131 [20:36:16] hoo: if you want to open, ok with me [20:36:18] eptalon: Well, I guess it's better if a dev. does this [20:36:28] hoo: ok. [20:36:28] as the cause lies deeper than in the client widget [20:37:10] aude: That's basically it [20:37:16] my basic way of doing it would be: if it is a redirect, tell the client, perhaps confirm, and link to the page redirected to [20:37:27] My patch should resolve them both [20:38:04] eptalon: We already have a function which asks the client to normalize the title (resolve redirects etc.)... it's just not used consistently [20:39:35] good :) [20:42:32] aude: i'll revert that revision for now [20:43:12] ok [20:43:39] aude: "normalizeItemByTitlePageNames" weird Setting name [20:43:49] heh [20:45:10] But I guess I can keep using that, no? [20:45:50] i think so [21:29:58] New patchset: Hoo man; "[WIP] Introduce a SiteTitle class for normalizing page titles" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63967 [22:02:02] New review: Jeroen De Dauw; "Code not registered. Test file not ending on Test." [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/63967 [22:06:43] New review: Jeroen De Dauw; "I'm not entirely sure what you are trying to do here, but the current class really does not strike m..." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63967 [22:07:34] Change merged: Jeroen De Dauw; [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63710 [22:08:09] New patchset: Jeroen De Dauw; "Removed ValueParsers Error interface" [mediawiki/extensions/DataValues] (master) - https://gerrit.wikimedia.org/r/58879 [22:23:35] "// If the code is under test, don't call out to other sites, just normalize locally." Superb... there's no way to test this functionality in Wikibase [22:24:25] * hoo cries and cripples his unit test [22:28:16] JeroenDeDauw: You wrote MediaWikiSite and crippled it for unit tests? [22:29:00] VERY NICE HIGH FIVE [22:30:51] :D [22:41:20] New patchset: Hoo man; "[WIP] Introduce a SiteTitle class for normalizing page titles" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63967 [23:02:26] New patchset: Hoo man; "[WIP] Introduce a SiteTitle class for normalizing page titles" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63967 [23:04:25]