[00:05:46] Need help. Anyone around ? [00:07:47] More or less [00:07:59] hai WikiJunkie [00:08:15] hei [00:08:49] wikidata first phase started running on the Heb Wikivoyage recently [00:09:01] everything seems to be working fine so far [00:09:09] except for one annoying thing [00:09:15] which is? [00:09:41] all the links to the parrallel articles in the English Wikivoyage used to be shown on the top of the list [00:09:59] now the link to the German Wikivoyage article appears first [00:10:35] because the order is now alphabetically [00:10:43] ok [00:10:43] WikiJunkie: That can be solved :) Would you open a bug? [00:11:01] there is no easy way to solve that ? [00:11:50] Requires a change to Wikibase and a config change :/ [00:12:04] hmm [00:12:07] ok [00:15:01] thanks for the help [00:17:37] WikiJunkie: i think we already have a patch for that [00:17:45] just waiting to deploy it [00:17:51] oh [00:18:06] so this would be fixed soon ? [00:18:11] yes [00:18:29] in the next 24 hours, i'd say ... hopefully much sooner [00:18:47] * aude checks the status [00:19:19] https://gerrit.wikimedia.org/r/#/c/75617/ [00:19:42] thanks [00:20:01] sure [00:20:07] thanks for reporting the issue [07:32:13] aude: not exactly what I saw but similar, mine included an enwikivoyage link appearing with the wikipedia links ;p [08:58:02] (CR) Addshore: "(3 comments)" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74171 (owner: Tobias Gritschacher) [09:00:37] DanielK_WMDE: hangout? [09:08:35] Denny_WMDE: http://tools.wmflabs.org/addbot/addbot/iwlinks/index.html?lang=en&site=wiki (it is still being generated but shows some interesting gaps the bots have missed :) ) [09:11:18] addshore: i found a bunch of stuff missed on enwiki [09:11:29] there is lots everywhere :) [09:11:30] like 500 places out of 10000 [09:11:36] for my bounding box [09:12:00] http://tools.wmflabs.org/audetools/coords/index.php/browse/1000 (see all the pink ones) [09:12:08] though i made a bunch last night [09:12:14] so it's less pink [09:12:33] http://tools.wmflabs.org/audetools/coords/index.php/browse/2000 [09:12:39] :) [09:12:58] hehe :) [09:13:14] at least i can find them now [09:13:27] hmm, think I need to refine my list a bit as it includes things such as Wikipedia:WikiProject Spam/COIReports/2007, May 1 [09:13:44] will have to wait a few hours for it to finish generating first [09:48:49] (PS6) Daniel Kinzler: Make ReferencedEntityFinder use Entity::getAllSnaks [extensions/Wikibase] - https://gerrit.wikimedia.org/r/72962 [09:48:55] (PS2) Daniel Kinzler: Improve tests for EntityView. [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75103 [09:51:20] (PS4) Daniel Kinzler: (bug 49306) Introducing ReferencedUrlFinder [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74394 [09:51:35] (PS5) Daniel Kinzler: Adding tests for claims to EntityTest. [extensions/WikibaseDataModel] - https://gerrit.wikimedia.org/r/72974 [09:51:57] (PS11) Daniel Kinzler: (bug 51363) Add more tests to for bad diffs. [extensions/Wikibase] - https://gerrit.wikimedia.org/r/73772 [09:52:18] (CR) jenkins-bot: [V: -1] Adding tests for claims to EntityTest. [extensions/WikibaseDataModel] - https://gerrit.wikimedia.org/r/72974 (owner: Daniel Kinzler) [09:52:22] (PS5) Daniel Kinzler: (bug 46352) Implement EntityHandler::getUndoContent [extensions/Wikibase] - https://gerrit.wikimedia.org/r/73942 [09:53:05] (PS10) Tobias Gritschacher: Use ChangeOps to generate edit summaries in API modules [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74171 [09:53:11] (CR) Tobias Gritschacher: "(3 comments)" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74171 (owner: Tobias Gritschacher) [09:55:46] (PS11) Tobias Gritschacher: Use ChangeOps to generate edit summaries in API modules [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74171 [10:21:21] (PS4) Tobias Gritschacher: Record multiple summaries in ChangeOps [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75338 [10:22:15] https://dl.dropboxusercontent.com/u/172199972/map/index.html [10:30:24] (PS12) Tobias Gritschacher: Use ChangeOps to generate edit summaries in API modules [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74171 [10:37:24] addshore: updated https://gerrit.wikimedia.org/r/#/c/74171/ [10:37:27] :) [10:37:34] looking now :) [10:38:41] * aude eager to have the summaries all consistent, etc :) [10:38:50] hehe :P [10:39:00] (PS5) Tobias Gritschacher: Record multiple summaries in ChangeOps [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75338 [10:39:07] but then we won't be able to tell which module was used for which edit :/ [10:41:07] (PS6) Tobias Gritschacher: Record multiple summaries in ChangeOps [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75338 [10:45:22] (CR) Addshore: [C: 2] Use ChangeOps to generate edit summaries in API modules [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74171 (owner: Tobias Gritschacher) [10:45:58] (Abandoned) Tobias Gritschacher: Record multiple summaries in ChangeOps [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75338 (owner: Tobias Gritschacher) [10:46:49] (Merged) jenkins-bot: Use ChangeOps to generate edit summaries in API modules [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74171 (owner: Tobias Gritschacher) [10:49:28] mhhm Tobi_WMDE I have just found one you have missed! [10:49:32] https://bugzilla.wikimedia.org/show_bug.cgi?id=44813 [10:50:06] still always appears as ‎Added link to [xxwiki]: [10:50:26] :/ [10:50:56] addshore: will see [10:50:57] and just "Modified site link [xxwiki]:" would be nicer [10:51:00] :) [10:51:06] site-specific is strange wording [10:51:22] aude: I think the "site-specific" is gone already.. will have a look [10:51:28] Tobi_WMDE: ok [10:52:06] still appears as Added link to [enwiki]: User:Addshore/Sandbox :) [10:56:08] (CR) Addshore: [C: -1] "(1 comment)" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74954 (owner: Liangent) [11:12:52] properties cannot be accesed by name yet, no? [11:13:31] (CR) Daniel Kinzler: [C: 2] Do not store empty aliases in the datamodel [extensions/WikibaseDataModel] - https://gerrit.wikimedia.org/r/75321 (owner: Addshore) [11:13:39] (Merged) jenkins-bot: Do not store empty aliases in the datamodel [extensions/WikibaseDataModel] - https://gerrit.wikimedia.org/r/75321 (owner: Addshore) [11:13:44] :> [11:13:53] cheers DanielK_WMDE :) [11:18:28] (CR) Daniel Kinzler: [C: -1] "If I understand correctly, the purpose of this patch is to disambiguate when looking up Site objects based on local langlink prefixes." [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75477 (owner: Hoo man) [11:19:20] (CR) Daniel Kinzler: [C: -2] "see above" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75477 (owner: Hoo man) [11:19:22] (PS1) Tobias Gritschacher: Slightly more meaningful autocomment for wbsetsitelink [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75848 [11:23:51] (PS1) Tobias Gritschacher: Treat not set parameter 'linktitle' as request to remove sitelink [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75849 [11:24:15] (CR) Daniel Kinzler: [C: -1] "(2 comments)" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75533 (owner: Hoo man) [11:24:38] addshore: fixed! https://gerrit.wikimedia.org/r/#/c/75849/ [11:24:58] Tobi_WMDE: will check ina second :0 [11:25:36] addshore: and the improvement to the autocomment for setting sitelinks: https://gerrit.wikimedia.org/r/#/c/75848/ [11:25:47] (CR) Daniel Kinzler: [C: 2] add render function to Template class [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75082 (owner: Aude) [11:26:10] (PS15) Daniel Kinzler: Label and description serialization now accepts LanguageFallbackChain as the language option [extensions/Wikibase] - https://gerrit.wikimedia.org/r/72225 (owner: Liangent) [11:26:31] (PS4) Daniel Kinzler: Allow mw.wikibase.label( ) to find labels written in variants [extensions/Wikibase] - https://gerrit.wikimedia.org/r/72751 (owner: Liangent) [11:26:40] (PS28) Daniel Kinzler: Include preferred labels / descriptions for entities in JS on EntityView [extensions/Wikibase] - https://gerrit.wikimedia.org/r/72226 (owner: Liangent) [11:27:21] (Merged) jenkins-bot: add render function to Template class [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75082 (owner: Aude) [11:29:04] (CR) Daniel Kinzler: "Please provide basic tests for the new behavior." [extensions/Wikibase] - https://gerrit.wikimedia.org/r/72226 (owner: Liangent) [11:29:14] (CR) Daniel Kinzler: [C: -1] "Please provide basic tests for the new behavior." [extensions/Wikibase] - https://gerrit.wikimedia.org/r/72226 (owner: Liangent) [11:30:44] (CR) Daniel Kinzler: [C: 1] Allow mw.wikibase.label( ) to find labels written in variants [extensions/Wikibase] - https://gerrit.wikimedia.org/r/72751 (owner: Liangent) [11:31:16] (CR) Liangent: "How .. ? by grepping in generated HTML?" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/72226 (owner: Liangent) [11:37:12] [travis-ci] wikimedia/mediawiki-extensions-Wikibase#472 (master - 39d2e06 : jenkins-bot): The build was broken. [11:37:12] [travis-ci] Change view : https://github.com/wikimedia/mediawiki-extensions-Wikibase/compare/586e9284350a...39d2e0607fd9 [11:37:12] [travis-ci] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-Wikibase/builds/9469647 [11:38:28] (CR) Addshore: [C: -1] "(1 comment)" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75849 (owner: Tobias Gritschacher) [11:41:06] (PS2) Tobias Gritschacher: Slightly more meaningful autocomment for wbsetsitelink [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75848 [11:42:56] Change on 12mediawiki a page Extension:WikibaseLib was modified, changed by Kghbln link https://www.mediawiki.org/w/index.php?diff=746648 edit summary: [-11] - update [11:47:33] thanks DanielK_WMDE :) [11:47:55] aude: hm? [11:48:00] we're now a little bit closer to being able to test stuff with Templates [11:48:01] (PS1) Jeroen De Dauw: Add getDataValueFactory to top level factory [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75851 [11:48:10] aude: yay :) [11:48:21] like pragun's stuff :) [11:48:30] (PS2) Tobias Gritschacher: Treat not set parameter 'linktitle' as request to remove sitelink [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75849 [11:51:01] . [11:51:16] Tobi_WMDE: (→‎wbsetsitelink-add:1|enwiki: User:Addshore) [11:53:13] why??? [11:54:51] Tobi_WMDE, JeroenDeDauw, aude: plz review! https://gerrit.wikimedia.org/r/#/q/owner:%22Daniel+Kinzler%22+status:open,n,z [11:55:13] DanielK_WMDE: ok! [11:56:49] oh, and please move the two "bad value" tickets from "doing" to "review" [11:57:11] ...and make danwe review that - basically, everything leading up to https://gerrit.wikimedia.org/r/#/c/69659/ [11:57:46] (CR) Addshore: [C: 2] Treat not set parameter 'linktitle' as request to remove sitelink [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75849 (owner: Tobias Gritschacher) [11:57:51] DanielK_WMDE: ok [11:57:59] oh... [11:58:19] ? [11:58:43] Tobi_WMDE: can you have a look at the mail that ju8st came in on wikidata-l, regarding API changes? [11:59:06] sounds like something was lost when the serialization code got refactored [11:59:07] (Merged) jenkins-bot: Treat not set parameter 'linktitle' as request to remove sitelink [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75849 (owner: Tobias Gritschacher) [11:59:11] get entities? [11:59:29] yea - sitelinks/urls doesn't seem to work any more... [11:59:46] hmmm [11:59:52] sounds like a bug. should be easy to fix. [11:59:58] sure [12:00:36] DanielK_WMDE: I've put that on my todo list [12:01:51] we should file a bug if you can verify the issue... [12:02:28] I'll do that now :) [12:03:00] DanielK_WMDE: what do you think? should a custom user summary replace the whole autocomment, or only the autosummary? [12:03:35] DanielK_WMDE: I think the user summary should follow a automatically generated comment [12:04:21] addshore : for BugMaster! [12:04:24] https://bugzilla.wikimedia.org/show_bug.cgi?id=52020 [12:04:42] (PS1) Jeroen De Dauw: Some refactoring in ExtendedMySQLAbstraction [extensions/WikibaseDatabase] - https://gerrit.wikimedia.org/r/75852 [12:04:50] I also think the summary should follow the automatically generated comment! [12:05:04] but In some cases these could start getting quite long :/ [12:05:36] Tobi_WMDE: huh? shouldn't the user supplied summary be first? [12:05:49] and then the autosummary be appended if and only if it fits? [12:06:21] * aude htinks no appending [12:06:24] thinks [12:06:32] just use the user supplied summary [12:06:32] no appending at all? [12:06:35] yes [12:06:43] (PS1) Jeroen De Dauw: Remove ClaimsTable [extensions/WikibaseQueryEngine] - https://gerrit.wikimedia.org/r/75854 [12:06:43] trust the user :) [12:06:48] hah! [12:06:56] oh, it is not about trust [12:06:59] Denny_WMDE: we now have somehing like "Added link to [enwiki]: Hamburg" where "Added link to [enwiki]:" is the autocomment part and "Hamburg" is the autosummary. [12:07:10] the auto comemtns are great [12:07:18] * aude thinks human or bot supplied summary will be better and probably duplicative / loo long to append [12:07:19] Denny_WMDE: https://gerrit.wikimedia.org/r/#/c/75854/ [12:07:19] the only time we would need to totally replace them currently is in editentity [12:07:26] addshore: no! [12:07:29] we then will have something like "Added link to [enwiki]: " [12:07:42] where the user or bot would probably be able to identify what they are doing better than the code can [12:07:48] they don't say why the user / bot changed [12:08:03] right, that is why i am saying - i never even thought about it differently - append the auto to the user provided [12:08:11] and do not replace neither [12:08:21] only when there is not enough space, cut it off [12:08:24] hmmmmm [12:08:26] so [12:08:30] don't cut off the user summary [12:08:37] then eliminate the auto part [12:08:43] This is nonsense! Removed link to [enwiki]: Hamburg. [12:08:46] right [12:08:47] but the auto part is very usefull! [12:08:49] agree [12:08:54] don't cut off the user summary [12:08:59] cut off the autopart [12:09:04] addshore: useful but redundant [12:09:20] and we should never loose the user's comments [12:09:28] it could be "blah blah blah (Removed link to [enwiki]: Hamburg)" [12:09:31] what is the limit on length of summaries? [12:09:33] maybe, if possible [12:09:39] put the user summary first? [12:09:43] yes [12:09:46] 264 chars I think [12:09:51] put the user summary first [12:09:57] if we have to have both :) [12:10:02] then the auto parts [12:10:05] (Added link to [enwiki] User:Addshore) [12:10:11] addshore: yes [12:10:19] might be technically more complicated but right thing to do [12:10:22] and don't add auto if there is no space [12:10:33] yep [12:10:37] not so sure about the brackets :P [12:10:50] nah, leave the brackets [12:10:58] leve them out [12:10:59] per denny [12:11:11] i don't want to make assumptions about brackes in an i18n project :) [12:11:15] would be lovely if there was not enough space for the full auto summary with the list of things changed to instead just include (Added link to [enwiki]) [12:11:17] needs to be distinguised visualy though [12:11:22] like change tags [12:11:46] well, the auto summary already appears as a light gray color [12:12:00] addshore: but appending it... [12:12:02] although the list of things changed following do not [12:12:02] addshore: no, that's the autocomment [12:12:19] "Added link to [enwiki]: Hamburg" where "Added link to [enwiki]:" is the autocomment part and "Hamburg" is the autosummary. [12:12:34] so simple :) [12:12:43] * aude loves the autocomment code in core [12:13:15] so ( ) :P [12:13:23] Tobi_WMDE: re replacing the automatic summary... it's a tough call, really. We could allow bgoth, using different parameters, but that seems confusing. [12:13:32] yeah, which is not quite as simple / out of the box to do that way [12:13:51] Tobi_WMDE: maybe it should depend on the action/module... for editentity, the user supplied summary should probably replace the automatic one. [12:14:02] for other api modules... not so sure [12:14:13] I share that opinion! [12:14:25] most simple solution, and already finished is, have the autocomment followed by the user-summary [12:14:55] Tobi_WMDE: can we just turn that around? [12:15:01] * DanielK_WMDE is still reading the backlog [12:15:02] not easily [12:15:06] user, autocomment, autosummary [12:15:07] maybe [12:15:15] so, e.g. for editentity: "Updated item. " [12:15:17] usercomment, autocomment, autosummary [12:15:18] i'd have to look at the core code [12:15:31] it it's possible or easy to change the rendering around like that [12:15:32] oh, that's a core issue [12:15:36] ? [12:15:41] well, it's the recent changes page [12:15:50] and other places it gets displayed [12:16:07] has a prescribed way of doing things, which hopefully can be customised [12:16:14] aude: but we have a hook funktion for the formating stuff [12:16:17] but don't count on it being easy [12:16:26] Tobi_WMDE: for formatting links? [12:16:29] or everything? [12:16:46] the formatting links only happens on recent changes and certain places, by the way [12:16:51] not on history pages [12:16:57] aude: look at Autocomment::onFormat [12:16:59] Denny_WMDE, addshore: please consider that the "autocomment" part is special: it must always be first, and is enclosed in /* .... */. That'S the bit we can translate. [12:17:05] Tobi_WMDE: ok, maybe [12:17:19] but that might just be the autocomment part [12:17:23] DanielK_WMDE: indeed, [12:17:34] the sumamry should show you what has happened in an edit eithout looking at the edit [12:17:42] Denny_WMDE, addshore: we also automatically add stuff after that (the "autosummary"),m which is usually just a list of codes or some such, which is marginally informative and generally ugly [12:17:44] and the auto summaries in all modules except for editentity currently do that [12:17:57] so, perhaps, always keep the autocomment, but replace the autosummary with the user-supplied one? [12:17:59] addshore: that's why I would always prepend the autocommenrt [12:18:11] DanielK_WMDE: that's what I say [12:18:31] Tobi_WMDE: yea, just wanted to make sure everyone's on the same poage on that: [12:18:32] the user comment will only be usefull for people in a single language [12:18:38] there is THREE parts to consider, not just two [12:18:52] it's autocomment/autosummary/user-summary [12:18:59] ick! [12:19:12] ...and the autocomment must always be first. [12:19:14] DanielK_WMDE: yes, and I would just replace autosummary with usersummary [12:19:17] * aude not happy about the order but understand the limitations of the software [12:19:23] () ? [12:19:24] ok, i didnÄt know about the autocomment-first constraint [12:19:34] so i would suggest: [12:19:46] Tobi_WMDE: yes, i agree, except perhaps for cases where the autocomment is barely infomrative (that is, editentity and maybe setclaim) [12:20:30] "Updated item" !:) [12:20:34] DanielK_WMDE: I think especially in those cases the autocomment would be helpful [12:21:14] hmmm, for create item (however it was done) did or does support bot summaries [12:21:21] Created a new item: Bot: adding interwikilink: de [12:21:25] http://www.wikidata.org/w/index.php?title=Q760780&action=history [12:21:30] so you can see how it looks [12:21:53] aude: we had partly support for bot summaries already [12:22:02] ok [12:22:15] just inconsistent [12:22:21] forgotten about [12:22:22] https://bugzilla.wikimedia.org/show_bug.cgi?id=46537 implies that it got removed [12:22:28] yeah [12:22:32] but it was never consistant though [12:22:34] Tobi_WMDE: really? the autocomment for editentity is what, "updated"? [12:22:36] yep [12:22:45] DanielK_WMDE: "Updated item" [12:22:55] although think it's being improved, per change ops [12:23:09] if it's one change op, then the summary can be specific for that [12:23:19] DanielK_WMDE: if theres no autocomment at all a user which does not speak the language the usersummary is in dows not have a clue what happend in that edt at all [12:23:21] can only be improved if there was only one changeop [12:23:21] if it's a bunch of change ops, then the summary can't list all the cahnges [12:23:45] Tobi_WMDE: true, but "updated item" is not helpful in any language :) [12:23:58] Tobi_WMDE: that's the way it is on commons and other places [12:24:01] we live with that [12:24:11] autocomment autosummary usercomment [12:24:12] aude: and nither can the autocomment - our translation magic only works for a single message (iirc) [12:24:19] if not enough space, remove autosummary [12:24:21] DanielK_WMDE: I've changed the message already to "Applied changes" [12:24:25] if not enough space remove autocomment [12:24:37] never cut off the usercomment [12:24:51] Denny_WMDE: the autosummary is usually something like "enwiki|dewiki" or something cryptic like that... [12:24:52] Denny_WMDE: ok [12:25:10] yes, but it goes with the autocomment. [12:25:16] if there is enough space, show it. [12:25:19] if not, drop it. [12:25:26] Denny_WMDE: i don't like surprises like that. it basically means the autosummary is "sometimes" there. [12:25:28] but Denny_WMDE if you started seeing summaries that said [12:25:28] '选举的两名参议员组成”,因此联邦参议员原本是由州议会选举而非公民直接选举。修正案还改变了参议员席位出现空缺时的补选程序,规定各州州长“在人民依该州议会指示举行选举填补缺额以前”“任命临时参议员”,而原本的规定这一缺额仍然是“在州议会下次集 [12:25:28] 会”来进行“填补”' is that usefull ? :> [12:25:42] * aude thinks the bot committee would have somethign to say about that [12:26:01] DanielK_WMDE: yes, and it makes sense for it to be always there, but user comments are more important [12:26:02] that doesnt inform us at all about what has happened [12:26:05] let the community manage that [12:26:20] addshore: if that is what the user provided, yes, this is what we will take [12:26:29] you can always click to see the diff [12:26:42] if curious, but still think the bot committee might object [12:26:50] mhhm, aude, but the point of summaries is to not need to click the diff :P [12:26:53] per aude - you can always click and see [12:26:56] addshore: yeah [12:27:09] addshore: your comment was a whole discussion [12:27:09] Added link to [enwiki]: 选举的两名参议员组成”,因此联邦参议员原本是由州议会选举而非公民直接选举。 [12:27:17] if they are shorter, then everything is fine [12:27:17] is a lot better :) [12:27:31] but we are not cutting the user comment short [12:27:31] unicode ones will be long [12:27:35] that is the most valuable thing [12:27:40] more likely to cut off the auto part [12:27:45] anything else can be re-created later automatically [12:27:54] people adding summaries should just expect there to always be the 15 char or however long the auto is to be there [12:28:14] we could make the usercomment be max. 200 chars or sth [12:28:24] * DanielK_WMDE thinks the autosummary is usually not very important can can be dropped [12:28:25] so that it usually wont cut off the autocomment [12:28:52] DanielK_WMDE: the auto summary is usefull if there is no user provided summary (ie. use through the api) [12:28:53] Denny_WMDE: 200 *byte* [12:28:58] *use from the ui [12:29:06] yes, byte [12:29:23] i strongly disagree with the autosummary not being important or useful [12:29:24] http://www.wikidata.org/wiki/Q2717980http://www.wikidata.org/wiki/Q2717980‎; 12:28 . . (+28)‎ . . ‎http://www.wikidata.org/w/index.php?title=User:Shadster&action=edit&redlink=1 (http://www.wikidata.org/wiki/User_talk:Shadster | http://www.wikidata.org/wiki/Special:Contributions/Shadster)‎ (‎Changed Czech label: Sativa Rose) [12:29:32] that's a great autosummary [12:29:44] yes [12:29:45] (‎Created claim: http://www.wikidata.org/wiki/Property:P150http://www.wikidata.org/wiki/Property:P150: http://www.wikidata.org/wiki/Q488693http://www.wikidata.org/wiki/Q488693) [12:29:50] that's a great autosummary [12:29:58] and that's just recent changes [12:30:03] if there is space, there should be there [12:30:09] but the user comments are even more precious [12:30:15] autosummaries can be recreated [12:30:16] Denny_WMDE: these are autocomments, not autosummaries. [12:30:34] though i would have to look at the code to see whether the translation magic makes use of thge sum,mary part too [12:30:35] thats autocomment + autosummary, no? [12:30:45] the autocomment is the Created claim: part [12:30:46] not 1000% sure, that code is... interesting :P [12:30:55] autosummary is the part behind the : [12:31:01] i *think* the translation can't handle the summary part, but i could be wrong [12:31:02] love the confusion about terminology :) [12:31:12] couldn't be more confusing [12:31:34] you are wrong. [12:31:50] just checked, both autocomment and autosummary are translated [12:31:52] just checked [12:31:56] could be a JS thingy [12:32:07] but that's what i see on wikidata.org recent changes [12:32:20] so i stick with: [12:32:29] autocomment + autosummary + usercomment [12:33:10] Denny_WMDE: i think the stuff after the : is just appended, not translated... [12:33:11] if not enough space, remove autosummary, if still not enough space remove autocomment [12:33:26] but anyway. yes, keep all three, or only the user comment [12:33:26] DanielK_WMDE: can you just try on Wikidata.org? [12:33:28] Denny_WMDE: DanielK_WMDE if it has a message key, then it's translated [12:33:34] i dont think we should allow there to be not enough space for the autocomment [12:33:35] DanielK_WMDE: it does translate for me [12:33:36] we might be doing that [12:33:46] a user summary will not have a message key [12:34:20] Denny_WMDE: i don't see anything after the : that even needs translation (that's intentional, i guess). Or, and link labels are of course translated - that's an independent step [12:34:36] (‎Erstellte Aussage: http://www.wikidata.org/wiki/Property:P21http://www.wikidata.org/wiki/Property:P21: http://www.wikidata.org/wiki/Q6581097http://www.wikidata.org/wiki/Q6581097) [12:34:50] hmm, copy pasting didnt work [12:34:50] Erstellte Aussage is the autocomment? [12:34:54] for me it says: [12:34:59] addshore: yes [12:35:10] Erstellte Aussage: Geschlecht (P21): männlich (Q65…)) [12:35:11] * aude understands what it says [12:35:20] when I switch to english it says [12:35:21] Denny_WMDE: yea, link label translation is independent of the other stuff. that's done in the linker. [12:35:30] that's the label being turned from QXXXX to a label [12:35:37] it'll also work if you put [[q12345]] into the user summary. at least, it should. [12:35:38] [3] 10https://www.wikidata.org/wiki/q12345 [12:35:42] we do that with a hook and hte hook only works in certain places [12:35:51] like recent changes, but not history [12:35:52] other auto summaries are not translated / translatable [12:35:53] (wikibase-item-summary-wbsetsitelink-set: 1, enwiki)(colon-separator) User:Addshore/Sandbox) [12:36:04] same for all language specific changes [12:36:10] wooot? [12:36:11] aude: o_O why doesn't it work in the history? gah :/ [12:36:15] :D good gosh [12:36:22] the linker thing is not used there [12:36:25] apparently [12:36:26] i Would never have expected that to work in recent changes but not in the history [12:36:30] it is a linker hook [12:36:33] or somethign like that [12:36:46] ah well. i still stand to it. [12:36:49] usercomments rule. [12:36:51] core issue [12:36:57] not always :p [12:37:03] addshore: trust the user [12:37:23] :) [12:37:26] but how can something i have no chance of being able to read be usefull to me at all? ;p [12:37:37] trust the community [12:37:40] addshore: because this won't happen that often [12:37:50] and if it does, check that one diff [12:38:10] if it's a bot, i think the bot committee would expect reasonable summaries understandable to the community [12:38:14] don't you think [12:38:14] * addshore decides to make his bot start adding 250 char edits summaries in chinese to make a point [12:38:18] :> [12:38:29] * aude bans the bot! [12:38:36] community in action :) [12:38:45] * DanielK_WMDE wants support for {{int:xyz}} in summaries [12:38:46] [4] 04https://www.wikidata.org/wiki/Template:int:xyz [12:38:52] :) [12:39:01] thank you AsimovBot [12:39:25] so, shall we restrict usercomments to 200 bytes? [12:39:30] yes! [12:39:33] sure [12:40:12] ok, so do we have agreement on : autocomment + autosummary + usercomment, drop as if not enough space, drop ac if not enough space, and restrict uc to 200 bytes? [12:41:03] addshore: what's with https://gerrit.wikimedia.org/r/#/c/75848/ ? [12:41:20] I am taking this as agreement? :) [12:41:28] if we restrict to 200 bytes it is almost impossible to loose the ac :) *is happy* [12:41:42] Denny_WMDE: i still don't like the magic aspect much, but we can go with that [12:42:02] ok. Tobi_WMDE, I think we have an agreement [12:42:08] i hope it is not hard to implement :P [12:43:15] (CR) Addshore: [C: 2] Slightly more meaningful autocomment for wbsetsitelink [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75848 (owner: Tobias Gritschacher) [12:44:39] (Merged) jenkins-bot: Slightly more meaningful autocomment for wbsetsitelink [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75848 (owner: Tobias Gritschacher) [12:45:29] * aude silently disagrees :) [12:45:35] don't care enough [12:55:05] [travis-ci] wikimedia/mediawiki-extensions-Wikibase#474 (master - eb49e51 : jenkins-bot): The build is still failing. [12:55:05] [travis-ci] Change view : https://github.com/wikimedia/mediawiki-extensions-Wikibase/compare/1d21834fbbce...eb49e5163020 [12:55:05] [travis-ci] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-Wikibase/builds/9472235 [13:00:40] DanielK_WMDE: question: the /* and */ arould the autocomment, is this a weird mediawiki thing? [13:02:22] yes [13:02:44] thx [13:03:18] so, does this get stored in the database together with the rest? [13:03:31] the autocomments code dates from the days when erik did coding on mediawiki :) [13:04:02] * aude checks [13:04:48] . [13:04:53] aude: time for a reqrite, no? [13:04:54] /* wbsetlabel-set:1|de */ Afrika [13:04:59] *rewrite [13:05:00] Tobi_WMDE: yes! [13:05:21] * aude also loves the create page summaries [13:05:35] * aude creating q16 (canada) says "Created page with "country" " [13:05:37] Tobi_WMDE: 250 was the limit? [13:05:39] country being the description [13:05:51] yes [13:05:54] wbsetaliases-set: 7,en */ qwertyuiopasdfghjklzxcvbn, qwsrcyfyfytfkftfhfhfhfhfhfhfghfhfhfghgjhkjhkjfshuwy8th98wihuhi, hdh32ithiu389tghigweijgiow3ot2ngwmfpojoh9giu, iohfio2h3hgijwiogjiwng32ihgowh0fhiasnfiuhqiuhuhuhi, hjhiudhwkjfjfn32jfbwiigwd79qgwa... [13:05:55] and deleting a page, is deleted page with "country" or such [13:05:59] thats 249 :) [13:06:00] hah [13:06:10] addshore: check WikibaseLib line 130 [13:11:53] (PS1) Aude: (bug 52023) don't display multiple anon bubbles at once [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75862 [13:13:52] (CR) Denny Vrandecic: [C: 2] "(1 comment)" [extensions/WikibaseQueryEngine] - https://gerrit.wikimedia.org/r/75854 (owner: Jeroen De Dauw) [13:14:08] (Merged) jenkins-bot: Remove ClaimsTable [extensions/WikibaseQueryEngine] - https://gerrit.wikimedia.org/r/75854 (owner: Jeroen De Dauw) [13:27:19] JeroenDeDauw: addshore: 19:45 is one and another one at 22:55 - need to order tickets now - decisions please :D [13:27:27] Talking of limits, is the maximum length for item descriptions also something like 255? [13:27:31] And will this stay? [13:27:40] * Granjow needs moar [13:30:53] heh Tobi_WMDE commented on https://gerrit.wikimedia.org/r/#/c/75847/ (didnt notice it was a draft) [13:30:56] (PS6) Tobias Gritschacher: Support for custom edit summaries [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75845 [13:33:11] Granjow: item descriptions are used to differentiate two items that have the same label [13:35:54] Denny_WMDE addshore aude DanielK_WMDE: here's a first version of custom edit summaries: https://gerrit.wikimedia.org/r/#/c/75845/ [13:35:59] Denny_WMDE: I know, I need them for the descriptions here: http://ec2-54-234-76-44.compute-1.amazonaws.com/filter.php [13:37:10] it is not exactly what we agreed on but that is a bit more tricky to implement and could take longer.. [13:37:24] Granjow: so you are using your own wikibase instance to power the data for this? [13:37:39] Denny_WMDE: If I can put it into the description field, I can use one single wiki. [13:38:07] you are using wikidata.org for this? [13:38:35] Denny_WMDE: This version is plain mysql without Wikibase, but the Wikibase extension now works. No, it is not wikidata.org since some points would not be regarded as relevant data, I guess. [13:39:07] Granjow: then i am confused [13:39:17] Lydia_WMDE: 19:45 [13:39:29] JeroenDeDauw: ok [13:39:32] addshore: how about you? [13:39:32] Denny_WMDE1: Why? :) [13:40:19] Granjow: are you planning to put long descriptions into wikidata.org to support the species identifier tool? [13:41:05] aude: https://bugzilla.wikimedia.org/show_bug.cgi?id=51870 https://bugzilla.wikimedia.org/show_bug.cgi?id=51871 [13:41:16] Denny_WMDE1: Yes, that would be easiest. Like the one for «Immergrün». [13:42:39] yes Lydia_WMDE :) [13:42:55] addshore: perfect - will oder tickets then for you, JeroenDeDauw and me [13:42:59] :) [13:43:00] for 19:45 [13:43:19] Lydia_WMDE: Where are you going? :P [13:43:25] overcurios /me [13:43:27] hoo: seeing wolverine :D [13:44:01] heh, have fun :) [13:44:47] * aude stay home and work on my bot :) [13:45:20] it's run out of tasks to do [13:45:35] Granjow: this should not happen, but up to the community [13:46:49] Granjow: it could happen with the text properties [13:46:55] but not with what we currently call descriptions [13:47:10] text properties are planned for later this year [13:52:42] Granjow: btw cool site [13:56:19] aude: http://de.wikivoyage.org/wiki/G%C4%81ra_(Neues_Tal) [13:59:12] heh, Denny_WMDE1 one of the screenshots showing why the bot made such an impact on statistics :P http://grab.by/oMUw [14:00:19] 900 changes in one minute, addshore? [14:00:24] full throttle [14:00:28] yes, :> [14:00:32] but thats only on a single wiki ;p [14:00:49] :o [14:01:04] not exactly sure how fast it was going once phase 1 was deployed to the majority of sites [14:01:18] not really sure the best way to find out such a figure either :O [14:02:11] heh, and that was even with maxlag set to 0 (so if anything was lagged at all it would slow)! [14:04:37] * Reedy bans addshore [14:04:48] :] [14:11:29] (PS29) Liangent: Include preferred labels / descriptions for entities in JS on EntityView [extensions/Wikibase] - https://gerrit.wikimedia.org/r/72226 [14:15:19] Error: Unknown dependency: jquery.autocompletestring [14:19:13] (PS1) Tobias Gritschacher: Fix: Edit summaries have a double space [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75869 [14:20:28] (CR) Addshore: [C: 1] "It works and does what you would expect :)" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75845 (owner: Tobias Gritschacher) [14:20:39] (CR) jenkins-bot: [V: -1] Fix: Edit summaries have a double space [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75869 (owner: Tobias Gritschacher) [14:21:12] Denny_WMDE1: Thanks! :) More to come. Text properties are like Wikitext? [14:21:30] Have to go now, later :) [14:21:53] (PS2) Tobias Gritschacher: Fix: Edit summaries have a double space [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75869 [14:22:16] Granjow: no, never Wikitext [14:22:20] just plain text [14:22:43] let's see about their restriction in size [14:23:24] (CR) jenkins-bot: [V: -1] Fix: Edit summaries have a double space [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75869 (owner: Tobias Gritschacher) [14:23:28] Text is fine :) [14:23:35] (PS8) Liangent: api.php?action=wbgetentities gets a new parameter 'languagefallback' [extensions/Wikibase] - https://gerrit.wikimedia.org/r/74954 [14:25:17] (PS7) Tobias Gritschacher: Support for custom edit summaries [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75845 [14:26:11] Only about 1% of all global users have an account on wikidatawiki :/ [14:27:18] hoo :< [14:27:22] (PS1) Jeroen De Dauw: Fix multiple design issues in this component [DNM] [extensions/WikibaseDatabase] - https://gerrit.wikimedia.org/r/75873 [14:27:27] (CR) jenkins-bot: [V: -1] Fix multiple design issues in this component [DNM] [extensions/WikibaseDatabase] - https://gerrit.wikimedia.org/r/75873 (owner: Jeroen De Dauw) [14:29:11] (CR) Addshore: [C: 1] "It works and does what you would expect :)" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75845 (owner: Tobias Gritschacher) [14:29:19] dejavu [14:30:46] Denny_WMDE1: https://gerrit.wikimedia.org/r/#/c/75876/1/repo/tests/phpunit/includes/EntityViewTest.php [14:32:13] DanielK_WMDE: The page you requested was not found, or you do not have permission to view this page. ? [14:32:37] liangent: oh, right, drafts are only visible to reviewers. just a sec [14:32:41] hoo: what is the percentage of active users? [14:32:44] (silly policy) [14:32:59] liangent: try now [14:33:22] Denny_WMDE1: mh [14:33:51] We don't have a global user touched atm, so that's hard to tell [14:34:10] data from the account audit extension is per wiki either :/ [14:35:11] hoo: or who at least made one edit [14:35:46] would also require to scan all wiki databases as the centralauth database isn't caching that [14:36:10] meh [14:36:21] that of course can be done, but it's a quite huge pile of work that script that (sadly not all databases are on the same server) [14:37:21] DanielK_WMDE: put context in arg list of the test function [14:37:43] liangent: possibly. or just the bits that need to be set in the context [14:37:49] it'S up to you, this is just a mockup [14:38:26] liangent: i'll work on the FIXME though: isolating entityview from the database (at least more than before) [14:40:44] DanielK_WMDE: which fixme? [14:41:28] liangent: the on in the example code [14:44:07] DanielK_WMDE: and it'll create a merge conflict with my code [14:47:09] liangent: i do not intend to merge this. it's just a mockup for you and denny to look at. [14:47:20] it's the answer to your question how to test this [14:47:52] or do you mean the stuff i want to do with EntityView? Yea, there will be merge conflicts [14:47:57] not too horrible, though [15:13:12] (PS1) Daniel Kinzler: Introducing EntityTitleLookup [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75885 [15:13:13] (PS1) Daniel Kinzler: Inject access to entities into EntityView. [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75886 [15:14:29] liangent: ---^ [15:17:51] (CR) Daniel Kinzler: [C: 2] Support for custom edit summaries [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75845 (owner: Tobias Gritschacher) [15:19:16] (Merged) jenkins-bot: Support for custom edit summaries [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75845 (owner: Tobias Gritschacher) [15:19:37] Hello. I believe that there is a need to make embedding of wikidata infos in wikipedia HUMAN readable, so that people can easily work with wikidata within wikipedia. On this page the fact was pointed out today: https://www.wikidata.org/wiki/Wikidata_talk:Infoboxes_task_force#Make_embedding_data_in_wikipedia_human_readable [15:20:45] testtest_: the right way to do this i think would be a nice lua module [15:20:51] e.g. instead of {{#property:p35}} one should use {{#property:NAME OF PROPERTY}} e.g. {{#property:head of state}} [15:20:54] parser functions arn't up to the job. [15:21:18] testtest_: that already works. has been working for months. (unless it's broken for some reason?) [15:21:32] you can refer to properties by therir label in the local wiki's content language [15:21:51] i tried it out yesterday and it didn't work. but i will try it again [15:22:26] testtest_: please check with aude, but i think it's supposed to work [15:22:36] (i vaguely recall some discussion about/against it though) [15:25:53] for me it seems to be broken [15:26:29] i tried it in this article: https://de.wikipedia.org/wiki/Vereinigte_Staaten have a look at the last two changes [15:28:02] testtest_: looking into it [15:30:18] [travis-ci] wikimedia/mediawiki-extensions-Wikibase#475 (master - c5eb128 : jenkins-bot): The build is still failing. [15:30:18] [travis-ci] Change view : https://github.com/wikimedia/mediawiki-extensions-Wikibase/compare/eb49e5163020...c5eb128a1254 [15:30:18] [travis-ci] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-Wikibase/builds/9478506 [15:31:20] maybe further discussion can be performed on https://www.wikidata.org/wiki/Wikidata_talk:Infoboxes_task_force#Make_embedding_data_in_wikipedia_human_readable [15:31:39] i need to quit the chat. have a nice day :) [15:33:20] Updating ExtensionMessages-1.22wmf12.php...Unexpected non-MediaWiki exception encountered, of type "Exception" [15:33:20] exception 'Exception' with message 'Bad initialization order: When running the Wikibase repository extension and the WikibaseClient extension on the same wiki, WikibaseClient has to be included AFTER the repository.' in /a/common/php-1.22wmf11/extensions/Wikibase/repo/Wikibase.php:54 [15:33:22] grrr, again [15:33:25] aude: ^^ :( [15:33:32] wtf [15:33:43] wow [15:33:54] ok, patch coming [15:34:01] give me a couple minutes [15:34:05] thanks [15:34:11] * aude has an idea of how to make this better [15:38:00] Lydia_WMDE: when are we leaving? [15:38:15] Lydia_WMDE: on foot btw, dont want to be roasted inside of the ubahn atm [15:38:17] JeroenDeDauw: good question [15:38:20] addshore: ^ [15:38:32] i am not going to run just so you know ;-) [15:38:34] addshore: vote for walking, unless you want to die out of heat stroke [15:38:44] * addshore has no idea how far away it is [15:38:46] Lydia_WMDE: 6kmph is not running [15:38:58] You need to go 12.7kmph before it is running [15:39:10] Or was it 11.7? [15:40:29] Denny_WMDE1: so you saw the video? [15:40:35] Did you listen to it all the way? [15:40:44] Reedy: https://gerrit.wikimedia.org/r/#/c/75892/ [15:40:51] let me know if you think that will work or not [15:41:02] * aude thinks it will avoid the problems and still work fine [15:41:17] JeroenDeDauw: yes, i was frozen in shock and could not stop it [15:41:44] aude: DanielK_WMDE: is there anything we can do to get this one kicking? https://gerrit.wikimedia.org/r/#/c/65443/ [15:41:54] oh, that.... [15:43:25] Denny_WMDE1: mutante and mark will be at wikimania [15:43:49] i might be a bit busy to sit and focus, but maybe DanielK_WMDE can try to make progress [15:43:58] might involve varnish in some way, afaik [15:43:58] aude: great, thanks [15:44:04] which might be more possible now [15:44:06] Reedy: anything we can do to get https://gerrit.wikimedia.org/r/#/c/65443/ out? [15:44:23] probably above reedy to do anything :) [15:44:30] i wouldn't touch if i was reedy [15:44:43] tim might also have ideas, though not sure if he is at wikimania [15:45:45] I think he is now [15:45:49] Though he wasn't supposed to be [15:45:50] I think [15:46:00] Reedy: that's what i thought [15:46:07] i just don't see him on the attendee list [15:46:37] tim + mark are probably the people to figure this out about apache / varnish [15:47:14] it would be nice to get this done. i am baffled about its complexity [15:47:18] but whatever: https://gerrit.wikimedia.org/r/#/c/65443/ ?? [15:47:26] wah [15:47:27] sorry [15:47:28] wrong one [15:47:29] well, seems it won't work with http + https [15:47:44] with entity data, we always redirect to https [15:47:52] with the sub domains, that is not a good idea [15:47:54] i meant the one with the hevoyage [15:47:58] oh [15:48:13] i think reedy can do that along with my patch (if that's acceptable) [15:48:19] https://gerrit.wikimedia.org/r/#/c/75617/ this one [15:48:29] ah, what's your patch aude? [15:48:35] Denny_WMDE1: to fix the localisation update [15:48:40] great [15:48:42] it's choking again [15:48:59] if my patch works, then it should also work for data types [15:51:16] DanielK_WMDE: you joining the SF call? [15:59:35] (CR) Hoo man: "(2 comments)" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75533 (owner: Hoo man) [15:59:36] (PS4) Hoo man: Normalize item ids before trying to remove doubles [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75533 [15:59:37] (CR) Hoo man: "Addressed Daniel's comments and added a test case." [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75533 (owner: Hoo man) [15:59:49] addshore: do you want to join for food before the movie? [16:00:36] yus :) [16:00:38] DanielK_WMDE: Denny_WMDE1 Lydia_WMDE call? [16:00:40] addshore: ok, then lets leave at 18:30 [16:01:08] yus :) [16:23:06] !admin [16:23:16] wikiuser: Yes? [16:23:18] yus ? :O [16:23:32] addshore: Go back to fixing Wikivoyage. I have this :P [16:23:46] haha! [16:24:02] is there any problem with wikidata table heading [16:24:51] Oh was there is. [16:24:53] *wow [16:25:02] addshore: Get back here actually. [16:25:03] What is happening? [16:25:08] Everything is messed up [16:25:16] I just created a thread at the PC lol [16:25:28] is there any problem with the interface pages? [16:25:33] wikiuser: what exactly is happening? [16:25:52] on user contribution it shows " " [16:25:53] [16:25:56] addshore: [16:26:00] Interface pages are missing [16:26:06] https://www.wikidata.org/wiki/MediaWiki:Wikibase-terms [16:26:10] https://www.wikidata.org/wiki/MediaWiki:Wikibase-itemlink [16:26:27] And thus, mediawiki is not showing them [16:26:33] this is everywhere [16:26:59] heh, im going to say this is due to the small issue we are currently having :> [16:27:05] oh lol [16:27:16] they should all come back once the issue is fixed ;p [16:28:30] wtf all the wikibase messages are missing lol [16:28:37] Addshore: Okay :) [16:28:42] :) [16:29:02] (CR) Jeroen De Dauw: "(3 comments)" [extensions/DataValues] - https://gerrit.wikimedia.org/r/70433 (owner: Daniel Kinzler) [16:29:15] (CR) Jeroen De Dauw: [C: 2] (bug 49264) Handle bad values using UnDeserializableValue. [extensions/DataValues] - https://gerrit.wikimedia.org/r/70433 (owner: Daniel Kinzler) [16:31:14] is something wrong on wikidata? why do I only see "wikibase-itemlink" now instead of the name: https://www.wikidata.org/wiki/Special:Contributions/Trijnstel? [16:33:52] (CR) Hoo man: "So... we don't need this at all? :)" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75477 (owner: Hoo man) [16:34:15] Trijnstel: yes, all wmf wikis besides wikipedia are broken right now [16:34:21] really? :O [16:34:58] wait [16:35:00] all of them? [16:35:07] try voyage [16:35:10] * Hahc21 runs to Wikinews [16:35:34] Wikinews looks goog [16:35:39] Let me try WV [16:35:55] probably just a different amount of brokenness. they currently seem to have the wrong version deployed [16:35:58] it is being fixed [16:35:59] woah I can't access WV [16:36:25] (PS1) Aude: remove bad initialization order exception for repo (branch only!) [extensions/Wikibase] (mw1.22-wmf11) - https://gerrit.wikimedia.org/r/75899 [16:36:33] it says eswikinews has 1.22wmf11 (1c12100) [16:36:37] https://gerrit.wikimedia.org/r/#/c/75899/ [16:36:47] i will self approve in 20 seconds if no one else does [16:37:01] also doing this for wmf9 branch [16:38:10] (CR) Aude: [C: 2 V: 2] "must do this" [extensions/Wikibase] (mw1.22-wmf11) - https://gerrit.wikimedia.org/r/75899 (owner: Aude) [16:38:50] It still passes the unit tests, so shouldn't be to bad [16:40:10] yeah [16:40:20] it just requires humans to provide the correct load order [16:40:36] or things might not work perfect [16:42:47] second branch only fix already :/ [16:42:59] aude: Shall I abandon https://gerrit.wikimedia.org/r/75477 ? [16:43:57] hoo|away: we're goign to rebuild the site identifiers table [16:44:00] k [16:44:11] it will have for example wikivoyage only + external interwiki prefixes [16:44:23] But I don't get how they can be unique if en: leads you to enwiki on WP but to enwikivoyage on wikivoyage [16:45:43] aude: Ok, so you do what I did in application logic now on the DB level? Sounds snae [16:46:25] yes [16:46:34] they will be unique [16:46:36] one en only [16:46:46] the en will be different per site group [16:46:48] (Abandoned) Hoo man: Attribute iw-links set via wikitext to the right site group [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75477 (owner: Hoo man) [16:46:55] depending which site we are on [16:46:57] Yes, that makes sense :) [16:47:16] Just have to make sure that this gets set consistent when creating new wikis [16:48:13] yes [16:48:43] the rebuild interwiki script used by wmf is especially complex but we can integrate with it for wmf-specific stuff [16:48:53] and provide a generic populate sites table script for core [16:49:50] That sounds great... afterwards we can also move parts of hte LangLinkHandler logic into core maybe [16:50:27] aude: I wont be around for the deployment tonight, but as it seems the move propagation is working anyway... just in case, you know how to set the setting to disable it and you know my email to get me to fix it ;) [16:52:01] we already working on stuff [16:52:11] busy in ops :) [16:52:35] :) [16:52:47] I guess (and hope) the Wikipedians will love that feature [17:11:58] (CR) Matmarex: "For future reference, if you have a blank line between "Bug: NNNNN" and "Change-Id: abcdef", the references to the bugs won't be recognize" [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75845 (owner: Tobias Gritschacher) [17:20:36] Denny_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=52039 [17:20:45] please assign DanielK_WMDE to that [17:21:33] meanwhile i can try to come up with a better way in deployment to handle loading and localisation update [18:06:09] I'm sure this is a frequently asked question, but I could not find the answer anywhere ... [18:06:28] 42 [18:07:12] This page lists several authority identifiers I am interested in: https://www.wikidata.org/wiki/Q75812 [18:07:45] But when I try to retrieve them via API, I cannot find them. They are not here: https://www.wikidata.org/w/api.php?action=wbgetentities&sites=dewiki&titles=Norbert_Elias&format=json [18:09:21] WikiData's API documentation seems to be unchanged from the generic one that comes with MediaWiki -- it does not tell me how I can get these properties directly. [18:10:29] Should not it be easy to get a JSON or XML representation of any page such as https://www.wikidata.org/wiki/Q75812 ? [18:11:21] The wikidata api help page has a lot more entries on it [18:11:41] Been there. [18:12:59] bielewelt: you then have to load the item that is referenced there [18:13:05] errr [18:13:26] that's if the property is referring to another item [18:13:41] wikibase-entityid [18:13:43] type [18:14:14] p269 has a string 027288579 [18:14:36] you might also be interested in http://meta.wikimedia.org/wiki/Wikidata/Development/Queries which is not available yet [18:15:01] bielewelt: But when I try to retrieve them via API, I cannot find them. They are not here: https://www.wikidata.org/w/api.php?action=wbgetentities&sites=dewiki&titles=Norbert_Elias&format=json [18:15:03] you said [18:15:06] but they are there? [18:15:28] search for 027288579 [18:15:37] and you will find it [18:16:18] I am looking for properties like GND (P227) or VIAF (P214). [18:16:52] And I would appreciate a URL schema for retrieving them. [18:17:53] it's in the JSON [18:18:03] both P227 and P214 [18:18:14] Gosh. Semantic MediaWiki was trivial to use in comparison. [18:19:08] OK, thanks, Denny. [18:19:11] (PS3) Daniel Kinzler: (bug 49264, bug 48965) validate time format. [extensions/DataValues] - https://gerrit.wikimedia.org/r/68152 [18:19:24] did you find it? [18:19:31] I did not realize I had to search for p269 instead of SUDOC. [18:19:38] (CR) Daniel Kinzler: [C: 2] "restoring CR+2 from danwe after rebase" [extensions/DataValues] - https://gerrit.wikimedia.org/r/68152 (owner: Daniel Kinzler) [18:19:44] (Merged) jenkins-bot: (bug 49264, bug 48965) validate time format. [extensions/DataValues] - https://gerrit.wikimedia.org/r/68152 (owner: Daniel Kinzler) [18:20:27] when i load the JSON into an object [18:20:37] say result [18:20:52] result.entities.q75812.claims.p214[0].mainsnak.datavalue.value [18:21:00] returns me "27064019" [18:21:26] Thanks everyone. This will do. :) [18:21:33] :) [19:22:53] (PS10) Daniel Kinzler: (bug 49264) Handle UnDeserializableValue gracefully. [extensions/WikibaseDataModel] - https://gerrit.wikimedia.org/r/70443 [19:24:55] With properties can you use different separators for lists? [19:25:02] Lika a, b and c. [19:25:11] *Like [19:25:25] Two separators: "," "and" [19:33:52] 2013-07-25 19:16:28 mw1132 wikidatawiki: [0cea464b] /w/api.php?action=translationaids&format=json&title=Translations%3AWikidata%3ATools%2FArray+properties+gadget%2F16%2Fpl Exception from line 179 of /usr/local/apache/common-local/php-1.22wmf11/extensions/Wikibase/lib/includes/EntityFactory.php: failed to deserialize [19:35:56] Paintman: you'd have to code that in lua [19:36:01] generally, the answer is "yes" [19:36:11] but we don't provide this functionality. [19:36:12] thaaaaaaaaaks! [19:36:18] +n [19:36:22] would be nice to have a nice and handy lua module that offered stuff lkike this [19:41:06] Reedy: sounds evil [19:43:27] DanielK_WMDE: Reedy if a magic word is not found, like noexternallanglinks, would be better just no have the magic word not do anything? [19:43:33] e.g. show all the wikidata links? [19:43:44] or throw the exception and have the site down? [19:44:20] i see that we can handle it in the hook and just ignore / throw warning but wonder if that's best to do [19:44:51] or it means {{#property:capital}} appears in wikitext or something instead of getting resolved [19:45:35] Reedy: and how easily / well do we track the warnings? [19:48:47] [travis-ci] wikimedia/mediawiki-extensions-Wikibase#477 (master - ce1a577 : Translation updater bot): The build was fixed. [19:48:47] [travis-ci] Change view : https://github.com/wikimedia/mediawiki-extensions-Wikibase/compare/c5eb128a1254...ce1a5774441a [19:48:47] [travis-ci] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-Wikibase/builds/9488616 [20:18:30] (CR) jenkins-bot: [V: -1] (bug 52038) catch exception when magic word not found and warn [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75997 (owner: Aude) [20:19:27] (PS2) Aude: (bug 52038) catch exception when magic word not found and warn [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75997 [20:20:57] (CR) jenkins-bot: [V: -1] (bug 52038) catch exception when magic word not found and warn [extensions/Wikibase] - https://gerrit.wikimedia.org/r/75997 (owner: Aude) [20:24:16] DanielK_WMDE: around? [20:24:23] jenkins says "no" [20:42:31] (PS6) Zfilipin: Use cucumber for selenium tests [extensions/Wikibase] - https://gerrit.wikimedia.org/r/62598 (owner: Tobias Gritschacher) [20:43:06] (PS4) Zfilipin: Selenium: use composition instead of inheritance [extensions/Wikibase] - https://gerrit.wikimedia.org/r/63857 (owner: Tobias Gritschacher) [20:45:21] (CR) jenkins-bot: [V: -1] Use cucumber for selenium tests [extensions/Wikibase] - https://gerrit.wikimedia.org/r/62598 (owner: Tobias Gritschacher) [20:46:41] (CR) jenkins-bot: [V: -1] Selenium: use composition instead of inheritance [extensions/Wikibase] - https://gerrit.wikimedia.org/r/63857 (owner: Tobias Gritschacher) [21:09:36] DanielK_WMDE: I've finished the property suggestions, some unit tests, code docs etc. Added a servlet that can be used to provide property suggestions for pre-existing items. [21:10:43] Tested everything on the labs instance too, works. Now it's just properties and items. I've removed the value suggestion code, it's backed up on my own PCs. [21:11:41] DanielK_WMDE: I'm pushing a few changes. Will you approve/review them? [21:12:41] hey nileshc [21:12:48] i'll review in the morning [21:12:51] about to go to bed now [21:13:34] DanielK_WMDE: Oh, okay. Question - shall I add a service (servlet) to add Item-property pairs? [21:14:15] That would enable training with single data points from the php side...this may come of good use if [21:14:41] later the team decides to keep the suggester updated to the second, know what I mean? [21:16:00] That is, whenever a new property is added, the suggester is updated, when some properties/items are deleted, the suggester is updated, so that it doesn't lag behind until the next wiki dump is generated and parsed. [21:19:59] (CR) Zfilipin: [C: -1] "Tests fail at wikidata.beta.wmflabs.org." [extensions/Wikibase] - https://gerrit.wikimedia.org/r/62598 (owner: Tobias Gritschacher) [21:20:44] DanielK_WMDE: Just so that I'm clear on this - I'll checkout a new branch -> git add blah blah -> git commit -> git review -R .....then again checkout a new branch and so on...after the code is reviewed, it'll be merged with master on Gerrit right? Won't there be unnecessary branches left dangling on my local machine or on remote? Sorry if this is too obvious. [21:24:08] I think git branch --merged and git review -f are what I need in cases like these...hmmm... [21:26:52] (CR) Zfilipin: [C: -1] "Failed tests at wikidata.beta.wmflabs.org." [extensions/Wikibase] - https://gerrit.wikimedia.org/r/63857 (owner: Tobias Gritschacher) [21:36:20] aude: I'm creating a new branch for a separate commit+change and doing git review -R. And yet I'm seeing that one ends up "depending" on another in Gerrit. After doing git review it gives me the multiple commits warning. What am I missing? [21:37:03] try git rebase -i master and see what it says [21:37:07] you can exit that [21:38:03] Cannot rebase: you have unstaged changes. Please commit or stash them [21:38:13] hmmm [21:39:05] you can do git stash, i suppose [21:39:18] http://git-scm.com/book/en/Git-Tools-Stashing [21:39:32] Oh! That should work. Let me try it out. [21:40:17] (PS1) Reedy: Adding easyrdf to .gitignore to prevent noise from checkout [extensions/Wikibase] - https://gerrit.wikimedia.org/r/76015 [21:41:00] ^ I'm going to do a few commits like that to numerous extensions to remove the noise [21:41:38] Reedy: jenkins will say no [21:41:45] because we have a failing test [21:41:47] lol [21:41:49] (CR) jenkins-bot: [V: -1] Adding easyrdf to .gitignore to prevent noise from checkout [extensions/Wikibase] - https://gerrit.wikimedia.org/r/76015 (owner: Reedy) [21:41:53] can this wait until tomorrow to be merged? [21:41:56] Yeah [21:41:59] ok [21:42:00] awesome. Thanks a lot aude :) What do I do about the already submitted change? Just unapprove it, do a git fetch --all and move on? [21:42:12] There's no rush, just trying to stop submodules being marked as dirty [21:42:23] nileshc: do you still want that change? [21:42:39] is the new thing dependant on that? [21:42:46] or should it work against master? [21:43:13] Yes I do. But it isn't dependent. These are separate changes. [21:43:24] ok, then git rebase -i master would work [21:43:39] you can then remove the unrelated dependency when you do that [21:43:48] and then your new commit will be against master [21:43:57] you can git review -d to get the other one [21:44:32] So I'll remove this change from gerrit, then make a new commit, then rebase -i master and so on? [21:44:40] no not from gerrit [21:44:46] one sec.... [21:44:53] okay. [21:45:00] see if i have an example [21:47:38] so http://dpaste.com/1317538/ [21:47:51] alright. [21:47:55] if i didn't mean to base my commit on reedy's then i would do git rebase -i master [21:47:58] (on command line) [21:48:12] just remove the easyrdf line [21:48:12] then save [21:48:18] and then it won't have the dependency [21:48:27] assuming it's a clean rebase [21:48:57] sure daniel has another trick [21:49:41] I see...this doesn't involve creating multiple branches it seems. Sounds neat. [21:49:50] right [21:50:07] you will want the two different patches on different branches (ideally) [21:50:14] so you can switch back and forth [21:50:18] and have both based on master [21:50:22] not on each other [21:50:51] Ah. Yes. I get the flow now. [21:50:55] k [21:51:20] sometimes we do end up fighting with git and gerrit [21:51:26] But I have two changes already submitted to Gerrit and they have garbled dependencies. Shouldn't I just abandon them, and start from where I began? [21:51:38] you can git review -d 00000 [21:51:45] then do the git rebase -i master [21:51:53] then git review -R to resubmit [21:52:06] Oh, that's what you were telling me a while back. [21:52:22] Gotchya. Indeed. It's somewhat confusing to get hold of first... [21:52:29] git review -d will make a new local branch for you [21:52:30] like review/daniel_kinzler/fix-ByPropertyListUnserializer [21:53:08] Okay...let me try it out. [21:53:11] ok [22:03:47] sadly, doing the git review -R again gives remote rejected....no new changes...failed to push [22:04:04] did you rebase? [22:04:38] if it's https://gerrit.wikimedia.org/r/#/c/76012/ [22:04:50] then i see it depends on somethign already merged [22:05:06] Yup. I did git stash -> git review -d 76013 -> git rebase -i master -> git review -R [22:05:09] weird [22:05:28] when you did rebase, you removed the dependency line [22:05:29] ? [22:05:30] why would we have wikibase unittests that rely on datavalues [22:05:44] daniel's changeset to datavalues broke wikibase [22:05:49] Denny_WMDE: wikibase depends on datavalues [22:06:01] it's all tested together [22:06:08] i know, but the unit tests for that stuff should be in datavalues not in wikibase [22:06:20] a changeset submitted to datavalues does not get tested with wikibase [22:06:29] aude: removed the dependency line? elaborate please? [22:06:31] that's why jenkins does not catch that error [22:06:31] right and it breaks something in wikibase [22:06:43] probably the test has an invalid value [22:06:48] yes, i guess [22:06:53] i will see if i can fix it [22:06:54] i'm not fixing it tonight [22:06:58] ok :) [22:07:08] aude: that is fine :) you did more than your share tonight :) [22:07:21] although i'm being entertained by an event in the building behind mine [22:07:29] might not sleep anytime soon :) [22:07:32] :P [22:07:41] seems like it's just starting [22:07:41] a good german would call the police [22:07:48] no [22:07:58] it's a ballroom [22:08:21] * aude shall go over there and take a beer :D [22:08:22] :D good for me. ballroom event = aude staying up late = me learning some gerrit-fu possibly. [22:08:29] * nileshc chuckles. [22:10:01] aude: where should I look for the dependency line? How will I remove it? [22:11:37] you don't see anything like http://dpaste.com/1317538 ? [22:13:21] I see only one line. I'm in the tests branch (for change 76013) and doing git rebase -i master. [22:14:14] pick 35bc5ca Added unit tests for ......... Recommender [22:14:46] When I try to git review -R it tells me no new changes made so it'll not push. [22:14:50] http://dpaste.com/1317554/ is what i see [22:14:50] what? [22:14:55] why did I make it a draft? [22:14:56] https://gerrit.wikimedia.org/r/#/c/76022/ [22:15:13] i cloned the entity suggester [22:15:24] then git review -d 76013 [22:15:27] git rebase -i master [22:15:39] maybe you need to checkout master again and [22:15:44] git pull --rebase origin master [22:15:54] weird. Then I'll make a fresh clone. Must've done something wrong. [22:15:54] it rebases against your local master [22:15:57] oh, okay.. [22:16:01] no try updating your master branch first [22:16:05] ok [22:16:22] can a draft still be merged, or do i have to do something? [22:16:42] Denny_WMDE: Possibly.. Or is there a publish button? [22:16:44] Denny_WMDE: looking [22:16:49] what reedy says [22:17:00] or do git review again without the -D [22:17:03] yeah there is a publish button [22:17:07] use that [22:17:15] wohoo [22:17:18] thank you [22:17:21] jenkins passes [22:17:54] yay! [22:17:57] the respective test even had a comment [22:18:05] saying "this needs to be removed once..." [22:18:07] :P [22:18:19] aude: nope. same thing.. [22:18:33] but there is a part of the comment saying also to remove a validator, so i left it in and just commented it out, and will ask tomorrow what it means [22:18:47] but for now, if you want jenkins to be ok, you can merge that i guess [22:19:01] same one line...while rebasing master it says "current branch master is up to date"...git review -d 76013 reuses the old existing branch [22:19:05] Denny_WMDE: i want to make sure it's tested for illegal values somewhere [22:19:12] that it has proper behavior [22:19:19] so i'll let daniel review it [22:20:02] suppose data values is testing it though [22:20:41] i very much hope so [22:20:52] but i agree [22:21:07] no one is going to do a lot of coding tonight that requires merging [22:21:14] * aude doubts it [22:21:42] https://git.wikimedia.org/blob/mediawiki%2Fextensions%2FDataValues/de632da601bc9c079470ccd4e2be3afb5f408dd9/DataValues%2Ftests%2Fphpunit%2Fincludes%2Fvalues%2FTimeValueTest.php [22:21:45] or if so, they can merge the patch [22:21:48] ok [22:21:58] here are the tests for illegal values [22:22:05] right where they should be [22:22:24] i am slightly worried now that we might have a number of illegal values in wikidata.org [22:23:33] we do [22:23:41] (PS2) Denny Vrandecic: Removed failing test for illegal time value [extensions/Wikibase] - https://gerrit.wikimedia.org/r/76022 [22:24:02] the idea is that a single illegal value won't corrupt the entire item [22:24:12] ok [22:24:13] we need to verify that [22:24:18] yeah [22:24:33] ah well [22:24:35] before, the time value validation corrupted the items [22:24:35] i need to sleep [22:24:38] ok [22:24:47] gn8 [22:24:54] night! [22:24:55] and thanks again for the firefighting [22:24:59] sure [22:25:07] i hope this won't become a habit :P [22:25:26] yeah [22:25:59] we just have to stop moving things around eventually and be stable [22:26:28] and not do weird things with insisting on certain loading order, if possible [22:30:16] i wonder why we had that [22:30:24] Denny_WMDE: it's the settings [22:30:31] trying to merge them correctly [22:30:34] it's not like the client and repo are actually running on the same machine anywhere [22:30:43] well the idea is that they can [22:31:01] yes, but obviously we are not smart enough to figure out how to use settings [22:31:03] shouldn't be a problem if localisation update loads all the extensions [22:31:18] we can get rid of the backwards compatiblity now, for one thing [22:31:24] no $wgWBSettings [22:31:27] only client or repo [22:31:34] and then we know which is which and no merging [22:31:40] sounds good [22:31:48] that should be oh so done until the 26th [22:31:49] i don't understand what other reasons daniel had [22:31:53] yes! [22:32:11] for data types, we can do the think we tried tonight [22:32:34] i think but will have to test the best way to add that into production [22:34:00] did eventually everything get deployed tonight? [22:34:14] yeah [22:34:18] wikipedias are fine [22:34:18] i missed that through the whole voyagedown thing etc. [22:34:20] ok thanks [22:34:27] should've been uneventful [22:34:34] and voyages and data are back up and humming [22:34:49] the problem was with the new wmf12 branch and localisaiton [22:35:05] and somehow localisation also got run for wmf11 with a problem [22:35:35] wikipedia never affected, fortunately :) [22:35:47] well, that is good [22:35:50] yeah [22:36:10] alright, time to work on my bot [22:36:17] canada bot? [22:36:20] yes [22:36:25] did prince edward island [22:36:27] awesome [22:36:28] (the smallest) [22:36:33] more data to the people! [22:36:37] :D [22:37:03] ok, see you tomorrow [22:37:33] ok [23:00:49] aude: are you still there? [23:04:05] ? [23:08:54] nileshc: ^ [23:13:20] aude: Sorry, was away. Never mind. It's 4:45 here and I'm fed up with gerrit. I don't know when I'll be able to do this without those needless dependencies. For now I'm just letting it be, get the changes to gerrit, let Daniel review them. I guess the dependencies won't hurt now much would they? [23:21:07] Daniel probably has ideas about gerrit [23:23:24] Hmm...I'll ask him in the morning. Thanks again aude. :) [23:27:01] ok, good night :) [23:42:47] aude: good night :)