[00:23:29] New patchset: Jeroen De Dauw; "Add docs to README file" [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73718 [00:25:51] Change merged: Jeroen De Dauw; [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73718 [01:20:28] Hey all. Is there a plan for the ability to reorder the order that properties appear on a page? [01:20:51] Yes [01:20:57] They're going to be called "ranks" [01:21:12] Right now everything is set to a rank of "normal" [01:21:21] There was a post about this somewhere... [01:24:07] Sven_Manguard: https://meta.wikimedia.org/wiki/Wikidata/Notes/Data_model_primer#Ranks [01:26:14] legoktm: no, no, I mean the order that the properties appear [01:26:16] erm... [01:26:32] the way i understand it, ranks will control that [01:26:45] so in http://www.wikidata.org/wiki/Q7747#sitelinks-wikipedia the order is place of birth, then office held, then alma matter [01:26:51] that's what I meant [01:27:22] oh [01:27:25] why does order matter? [01:28:01] aesthetics? [01:28:08] really that's the only reason I can think of [01:28:33] it would make sense to have the same order for all of the items on a specific subject, say for example, video games. That way I could look at one item and scroll down and say "oh, wait, there's no composer property here" without having to dig through what could become a very long page [01:29:03] ctrl+f :) [01:29:04] the data itself won't care, but the people that are looking though it for completeness will [01:29:28] You could probably write some JS to re-arrange it [01:29:42] no, *I* can't write a js... [01:30:12] Someone* ;) [01:30:21] legoktm: I don't want to search for every term. If there are seven properties that every video game article should have, I don't want to search for all seven, I want to be able to do a quick check, see that all seven are there, and then move to the next idem on the list [01:30:38] yeah, it's not a big deal if there are seven properties, but what about 35? [01:30:55] having a set order would allow people monitoring the data to work with it easier [01:31:05] I understand your usecase, but I don't think this should be implemented within the software itself. [01:31:56] why not? At the very least, have it auto-reorder properties by numerical order or something [01:33:23] Because for the sake of storing data it doesn't matter. It's a viewing/editing issue, so imo a user-script makes sense [01:35:10] legoktm: I cannot be the only one that wants this... [03:57:21] Sven_Manguard: you're not, I asked about it a few weeks ago too [03:57:58] * Sven_Manguard sticks his tongue out at legoktm [03:58:09] Heh. [06:59:27] * Lydia_WMDE gives Sven_Manguard's tounge a piercing [06:59:28] ;-) [06:59:33] also: https://bugzilla.wikimedia.org/show_bug.cgi?id=44678 [06:59:41] nikki: legoktm Sven_Manguard ^ [07:00:03] Ah. thanks [07:00:05] thanks, reading it now [07:00:24] Lydia_WMDE: by the way, who takes over as team lead once we lose the Mr. Q [07:00:39] Mr. Secret Q* [07:00:54] Denny* [07:01:15] we'll have to figure that out still [07:01:23] Denny "Secret Q" Vrandecic [07:01:27] heh [07:01:36] also i am totally on vacation [07:01:38] * Lydia_WMDE slaps self [07:46:30] :O [07:47:50] ohai addshore [07:47:56] wanna fix a trivial bug for me? [07:47:57] ohai [07:48:01] what is it? :P [07:48:07] i have some things to catch up on ;p [07:48:35] https://en.wikipedia.org/wiki/Barack_Obama?action=info <-- where it says "q76", the "q" should be uppercase. [07:48:50] mhhhm, but why? ;p [07:49:21] Because it shouldn't be lowercase? [07:49:33] but the id is lowercase ;p [07:49:39] technically ;p [07:49:58] all human facing interfaces will use the capitalized form [07:50:22] mhhm, silly human facing interfaces :p [07:51:36] legoktm: where are they used on ui? [07:51:52] here, {{#property: }} missing label [07:52:01] and what more? [07:53:03] hmm page titles but they're "MediaWiki page names" not essentially the original "entity id" [07:53:06] DanielK_WMDE_: hi [07:53:12] thats it i think [07:53:31] liangent: hi. [07:53:33] if you load a item page w/o javascript they show up in lowercase….weird. [07:53:35] what are you looking for? [07:53:47] legoktm: they should always be lowercase in my opinion :P [07:53:55] lowercase looks ugly >.> [07:53:59] true :P [07:54:06] thats what the js is there for ;p [07:54:59] addshore: we switched IDs from lower- to uppoer case at sopme point, it's not really consistent in the code [07:55:06] the EntityId class is case insensitive [07:55:27] everything should be upper now? :P [07:55:37] and prefixed IDs should not be used anywhere internally (they are in some places, though) [07:55:45] addshore: you said all pages were imported to wikidata some days ago. Why are there so many pages without interwiki links which have no item yet? [07:55:51] liangent: does EntityPerPage help you? [07:55:52] yup, the api is a big culprit [07:56:03] DanielK_WMDE_: what's that? [07:56:08] Pyfisch: when did I say that? :O [07:56:14] liangent: a mapping between mediawiki pages and entity ids [07:56:31] it's a DAO for wb_entity_per_page [07:56:37] DanielK_WMDE_: well I'm not working on it. I was just answering legoktm's question [07:56:39] personally i dont see the point in creating an item with a single site link on it, so my bot has never done that [07:56:44] addshore: good thing that this channel is logged. I will look [07:56:48] hehe :) [07:57:11] liangent: ah, well, then perhaps it helps legoktm. sorry for jumping in in the middle of the conversation :) [07:57:18] I am aware of a list of articles that still have multiple interwiki links on, so if I really said that I must have been asleep :> [07:58:07] DanielK_WMDE_: if you try https://gerrit.wikimedia.org/r/#/c/72982/ where fallback values simply replace original values, you'll see everything gets fallback automagically, though without any visible language labels [07:58:54] DanielK_WMDE_: but it prevents me from implementing it progressively [08:01:49] addshore: 20130712 but maybe I misunderstood you. You mean with importing everything again scanning all wikipedias again for articles without item? [08:02:17] what exactly did I say? :P [08:03:15] This is what I know, the majority of the import bots have stopped (as far as I am aware), my bot scans over every page with interwikis on every day trying to remove them but it doesnt currently try to add them to an item (that should be back next week) [08:03:33] Pyfisch: take a look at http://tools.wmflabs.org/addshore/addbot/status/ [08:03:42] [14:58:51] addshore: which wikipedias will you import :-O [08:03:44] [14:58:54] ? [08:03:45] [14:58:58] everything, again xD [08:03:47] [14:59:02] thats the plan [08:04:30] yus, thanks to legoktm we have just scanned all of the most recent dumps and compiled a fresh list of articles that still have links on them [08:04:45] once I finish coding it it will run over everything again :) [08:04:59] DanielK_WMDE_: also I still doubt whether it'll break any external scripts depending on the fact that wbEntity and wbUsedEntities are taken literally from database ie without fallback [08:05:40] Pyfisch: including wikivoyage i think there are 1,470,975 articles with interwikilinks on to check (if the list is correct) [08:06:02] if not, I can simply serialize a fallback-resolved version for wbEntity, and eliminate all usages of qqc [08:07:05] addshore: much work for your bot ;-) how many edits is it doing per minute? [08:07:56] (on wikidata) [08:09:16] anyway the whole approach which includes labels with resolved fallback can't be deployed as is (without js change) [08:14:23] Pyfisch: none currently [08:14:38] the wikidata end broke ;p [08:14:42] and I havnt had time to fix it :) [08:16:39] it is now only doing a couple of edits/m globally [08:36:45] Morning! Why would I get a «Fatal error: Call to undefined method Language::getParentLanguage() in /var/www/mediawiki/mediawiki-1.21.1/extensions/Wikibase/lib/includes/LanguageFallbackChainFactory.php on line 140»? [08:36:59] backtrace http://codepad.org/O2PQ3WLx [08:37:56] (git master, f080b62 works) [08:40:47] Granjow: upgrade your mediawiki to master [08:58:12] liangent: It is on master. [08:59:00] 6f31d3e [09:02:02] Granjow: upgrade your mediawiki core to master [09:02:06] not just wikibase [09:15:00] aude: when showing the diff of the change which seemed to break dispatch i get the same errors we were seeing the other day [09:15:15] Warning: strtr() expects parameter 1 to be string, array given in mediawiki\includes\Html.php on line 154
[09:16:41] oh really? [09:18:09] addshore: on wikidata? [09:18:34] I imported it to my local test instal [09:18:38] taking a look at the json now [09:18:41] oh, ok [09:19:55] hi [09:20:36] aude: I almost think that item is broken... [09:20:39] I am a little bit confused. If I click on languages I would like to add interwiki, but instead comes some confusing language settings: https://commons.wikimedia.org/wiki/File:Jazykov%C3%A1_nastaven%C3%AD.jpg [09:21:00] my import shows the current revision only has 2 sitelinks in the json... [09:21:02] addshore: not surprised [09:21:04] even after reading that, I havent figured out how it works [09:21:15] and where I can add iw than? [09:22:26] aude: if thats right I have a feeling another edit to it would break dispatch again O_o [09:22:35] oh no [09:22:48] if this has imported correctly.. [09:23:04] in the db it has 2 sitelinks, on the page it actually shows 4 O_o [09:23:40] Juandev: seems the "add links" link is hidden :/ [09:23:53] click languages to expand [09:24:08] and you can use the widget [09:24:47] or we now have a "data item" link in the toolbox -- if the page is connected already to wikidata [09:25:14] and you can edit directly in the item [09:25:59] heh, actually i was looking at the wrong row! :D [09:26:07] ? [09:27:07] aude: oh [09:27:26] on http://www.wikidata.org/wiki/Special:ItemByTitle/czwiki/Eternit [09:27:29] aude: its a new page [09:27:29] hmm, so what happend in the db was just [09:27:29] "links":{"cswiki":"Projekt 661","dewiki":"Papa-Klasse","frwiki":"Classe Papa","svwiki":"Projekt 661 Antjar"} [09:27:31] i see the page is not in wikidata [09:27:35] to "links":{"svwiki":"Projekt 661 Antjar","dewiki":"Papa-Klasse","frwiki":"Classe Papa","cswiki":"Projekt 661"} [09:27:51] i wonder what made it switch around O_o [09:28:25] * aude wonders where the invitation to create item is in http://www.wikidata.org/wiki/Special:ItemByTitle/czwiki/Eternit [09:28:55] Juandev: then i would go create it in http://www.wikidata.org/wiki/Special:NewItem and then add site links [09:29:14] that would be nice to have that "add links" there as default for everybody [09:29:26] Juandev: i don't know why i can't see it [09:29:47] aude: well I have no problem with that, I just add in [[en:eternit]] [09:29:47] [2] 10https://en.wikipedia.org/wiki/eternit [09:29:48] could be you have [[:en]] in the wikitext [09:29:49] [3] 04https://www.wikidata.org/wiki/:en [09:29:51] ah, yes [09:30:18] should I fill that as a bug? [09:30:19] if you take that out then "add links" should be there [09:30:38] Juandev: sure [09:30:58] * Juandev things its the second time he is filling such bug [09:31:01] that the add links should be there if the page has no links from wikidata (even if there are wiki text ones) [09:31:12] Juandev: ok [09:31:25] maybe the problem is, the page of at the state nonexistent? [09:31:35] not sure [09:38:06] aude: but it was there. Lets wait if it come up [09:38:16] Juandev: ok [09:38:17] aude: do you have a copy of a change record that failed? the info blob, most importantly [09:38:31] re the dispatcher failure [09:38:54] also, are there any changes up regarding that that i should look at? [09:39:02] DanielK_WMDE_: check my mail [09:39:20] * aude digs up the dpaste [09:39:30] http://dpaste.com/1300944/ [09:39:43] http://www.wikidata.org/w/index.php?title=Q13360583&diff=57098298&oldid=54370606 [09:40:08] should I file a bug request for already added coordinates on the moon, while it is not displayed that the coordinate is on the moon and not on earth? [09:40:09] as far as dispatcher and change code, i don't think anything has been modified recently [09:40:15] in the diff code, don't know [09:40:34] [[Q607451]] [09:40:36] [4] 10https://www.wikidata.org/wiki/Q607451 [09:41:02] DiffChange->getDiff() and getSiteLinkDiff returned a DiffChangeObject while it should be an ItemDiff [09:41:08] DiffOpCHange [09:41:12] gah :) [09:41:47] only difference appears to be the key order of site links between old and new value [09:44:14] aude: thanks [09:44:59] looks like someone deleted links and labels [09:45:00] http://www.wikidata.org/w/index.php?title=Q13360583&diff=54057292&oldid=49269981 [09:45:08] then someone else added them again [09:45:40] and then the first person restored to http://www.wikidata.org/w/index.php?title=Q13360583&oldid=49269981 [09:45:48] which is before all of that [09:46:47] aude: DiffChangeObject vs. DiffOpChange? gah. [09:46:54] DiffOpChange [09:47:12] vs. ItemDiff which it should be [09:47:28] getSiteLinkDiff should return an ItemDiff? Really? That sounds wrong [09:47:43] idk [09:47:46] that's what it does [09:47:56] where is the itemdiff expected? [09:48:10] in the dispatcher / item change [09:49:11] There's no SiteLinkDiff or anything like that [09:49:15] ItemDiff is ok [09:49:57] o_O [09:50:30] lazowik: ItemDiff is a diff between items. it defines stuff like getSiteLinkDiff... [09:50:53] yep [09:52:01] so the SiteLink diff isn't the/a ItemDiff [09:52:02] hmmmmm [09:52:20] now i get Wikibase\ItemChange for the change (fine) [09:52:37] and Diff\Diff for the site link diff (for the ones around the bad one) [09:52:57] instead of diff, i get diffopchange for the site links [09:52:58] that's what it should be [09:53:01] yeah [09:53:23] * aude must've been tired last night  [09:54:45] aude: which tag(s) were deployed when that happened? [09:54:53] diffopchange wmf9 [09:55:03] for diff and wikibase both? [09:55:07] yes [09:55:17] ok, i'll try to reproduce [09:55:21] alright [09:55:47] * DanielK_WMDE_ suspects that the diff is empty because only the order changed, and thus is hitting some edge case [09:55:59] me too [09:56:15] so it's like no diff yet it's a change [09:56:31] if the code is checking difference in array keys, tehn no diff [09:56:33] then [09:56:35] such a change can be skipped [09:56:46] (we could even make it a true null edit) [09:56:47] agree [09:57:04] although label and desc changed [09:57:11] in that case [09:57:15] right [09:57:18] ic [09:57:25] can't skip it alltogether, then [09:57:28] just the site link part is null [09:57:43] null? [09:57:56] no diff, unless you consider the order of the array keys [09:58:47] yea, but i'd expect it to be an empty change object, not null [09:58:56] anyway, i'll try to reproduce. [09:59:26] k [09:59:42] aude: same problem we had with the alias diff being null! [09:59:47] addshore: yes! [10:07:35] New patchset: Aude; "(bug 43238) Add very simple weighting for entity search" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73405 [10:07:35] New patchset: Aude; "(bug 51227) add term_weight column to wb_terms table" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73420 [10:07:48] New review: Aude; "rebased" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73420 [10:13:14] New review: Addshore; "I presume this can be abandoned? :)" [mediawiki/extensions/Wikibase] (mw1.22-wmf9) - https://gerrit.wikimedia.org/r/73175 [10:14:19] aude: ok, I can reproduce the issue. will dig in now. [10:14:21] Henning_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=51347 [10:14:53] Abraham_WMDE: new bug for the board, https://bugzilla.wikimedia.org/show_bug.cgi?id=51347 [10:16:48] DanielK_WMDE_: ok [10:17:02] with an undo? [10:17:08] Denny_WMDE: done, who's working on it? [10:17:25] aude: undo? [10:17:37] DanielK_WMDE_: errr, restore [10:17:42] how did you reproduce? [10:18:12] i inserted the broken change into my db and ran dispatch stats (had to truncated wb_changes_dispatch first). [10:18:49] DanielK_WMDE_: ok [10:18:54] aude: http://dpaste.com/1302234/ [10:19:38] Abraham_WMDE: put my head on the dispatcher breakage, please [10:19:51] New review: Aude; "(1 comment)" [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/73420 [10:19:55] and the undo bug too [10:21:34] DanielK_WMDE_: do you have the bug numbers? [10:21:48] New patchset: Henning Snater; "Added site link group name to section heading id" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73739 [10:25:05] Abraham_WMDE: Henning_WMDE for now [10:26:10] Henning_WMDE: i suspect some selenium or other test needs to be updated to use the new section heading [10:26:14] * aude doesn't know what [10:26:34] No, I just looked, Selenium is using css classes only. [10:26:37] ok [10:27:23] Abraham_WMDE: DanielK_WMDE_ : i think it is this one [10:27:31] if yes, i will set it to assigned [10:28:16] * aude can actually review some javascript :) [10:28:46] Denny_WMDE: which one? [10:28:54] https://bugzilla.wikimedia.org/show_bug.cgi?id=49338 [10:28:59] sorry forgot to actually paste [10:29:52] Denny_WMDE: no, that's not it [10:29:58] oh, sorry [10:30:20] that's *another* critical bug? [10:30:22] ugh. [10:30:31] Denny_WMDE: sounds nasty though [10:31:00] ick [10:33:27] New review: Aude; "when entity view gets refactored, we might find a better way. I don't see a nicer way to handle thi..." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/72090 [10:33:53] Denny_WMDE, Abraham_WMDE: https://bugzilla.wikimedia.org/show_bug.cgi?id=46352 is the undo thing [10:33:54] DanielK_WMDE_: i was hoping that's the one you were talking about [10:34:16] Denny_WMDE: nope. can look at the too, though [10:34:25] New patchset: Addshore; "Removing redundant parameter check" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73741 [10:34:26] thanks [10:34:27] sorry [10:34:48] i will assign both then. Abraham_WMDE will you put up both on the wall? thank you [10:35:08] i have a fix ready for the dispatcher thing [10:35:21] DanielK_WMDE_: yay! [10:35:23] it's not a nice fix, though. will have a go at the underlying issue [10:35:50] DanielK_WMDE_: well, a not nice fix i also had :) [10:35:52] aude: needs backport and deployment asap. any edit that only changest he order of links will cause this error to appear again [10:35:57] ...and break dispatching [10:36:05] DanielK_WMDE_: i'll take a look at the patch [10:36:13] will be up in a minute [10:36:30] hm... ten minutes, want to add tests [10:36:54] * aude might also backport henning's patch if we're doing backport [10:41:20] New patchset: Addshore; "Move wbgetclaims validation to validateParameters" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73742 [10:44:30] Change merged: jenkins-bot; [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73739 [10:47:19] New patchset: Addshore; "Move wblinktitles validation to validateParameters" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73743 [10:52:49] JeroenDeDauw: got a while? [10:53:20] lazowik: he's not here yet [10:53:36] New patchset: Addshore; "use EntityIdParser in api/getclaims" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/72532 [10:53:43] mhm [10:54:08] lazowik: btw, i'm setting to have unit tests run for your patches [10:54:27] hmm? [10:54:29] which? [10:54:40] lazowik: any, at least any wikibase ones [10:55:32] thanks :) [10:55:36] sure [10:55:42] makes it easier to review your patches [10:55:43] DanielK_WMDE_: around and available for some talk? [10:56:35] New review: Addshore; "I will have to do some digging," [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/72316 [11:02:06] test time = guitar time [11:10:29] addshore: that's a great idea :) [11:10:36] :D [11:10:52] suddenly long running tests became your firend! [11:10:53] New review: Aude; "is the hack to put a Site object in the siteGlobalID setting the best we can do to facilitate the te..." [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/72939 [11:11:47] New review: Aude; "(1 comment)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/72939 [11:12:42] New review: Aude; "i'd say so" [mediawiki/extensions/Wikibase] (mw1.22-wmf9) - https://gerrit.wikimedia.org/r/73175 [11:23:38] liangent: thanks, upgrading mediawiki to master worked. [11:24:11] :> [11:28:37] aude: tests are more tricky than i thought, had lunch first. will hack at it now [11:29:36] ok [11:29:42] liangent: i'm a bit tied up today looking at some critical bugs. can you write me an email? [11:29:48] What about https://gerrit.wikimedia.org/r/71543 ? Still awaiting CR... [11:30:21] hoo: on my list, but see above [11:30:55] ok [11:31:36] DanielK_WMDE_: ok [11:33:40] hey [11:34:17] maybe some along the way a diff for sitelinks is created as if for non associative array? [11:34:33] that is with second argument set to false [11:35:10] lazowik: no... it's simmilar, but not quite that. i'll hack at the root cause later. [11:35:45] lazowik: basically, if the contents of the map doesn't change, but the order does, it's reported as an atpmic change (of the order) instead of a (Map)Diff [11:36:02] mhm [11:37:18] I'm still digging into Diffs [11:51:07] I cannot add claims anymore though. Error message: http://codepad.org/DTk2kqRp [11:51:43] aude: whats the most up to date test wikibase client we have? [11:52:29] Hm, after running maintenance/update.php it seems to work [11:52:34] :P [11:52:41] addshore: wikidata-dev-repo? [11:52:49] err dev-client [11:52:51] your database was likely missing someting! [11:52:58] aude: where is it? :> [11:53:04] http://wikidata-dev-client.wikimedia.de/wiki/Helium [11:53:09] and how old is it? :D [11:53:20] it's updated every 20 minutes [11:53:26] yay! :D [11:53:32] just want I need :) [11:58:33] hmm, are they linked to ech other? [11:59:44] dev-repo? [11:59:50] should be as far as I know [12:00:42] mhhm, whats should the siteid of the client be? [12:04:51] New review: Denny Vrandecic; "(1 comment)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73420 [12:05:27] addshore: 'enwiki', probably [12:05:36] :> *checks* [12:05:53] looks like devclient [12:06:51] that's the wikiid [12:07:01] I don't know whether this is used [12:07:05] I cant seem to add a sitelink to it O_o [12:07:51] New patchset: Denny Vrandecic; "(bug 51227) add term_weight column to wb_terms table" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73420 [12:12:13] DanielK_WMDE_: thanks for your 2 patches that make getting the entity id possible from http://wikidata-dev-client.wikimedia.de/w/api.php?action=query&prop=pageprops&meta=wikibase&titles=Hydrogen ;p [12:12:18] its perfect :> [12:20:45] addshore: glad it's helpful :) [12:22:03] Denny_WMDE, aude, Abraham_WMDE: what's the bug id for the dispatcher issue? [12:22:33] make one! [12:22:36] * aude has no idea [12:22:46] i dont think we made one [12:22:48] i am making one now [12:22:50] aaarrrggg [12:22:54] please [12:22:57] and put it on the borad :) [12:23:02] aye [12:25:09] here we go https://bugzilla.wikimedia.org/show_bug.cgi?id=51363 [12:25:17] thanks [12:33:54] grrr, [12:34:26] aude: this is much bigger than I thought. the problem exists with diffs of all associative structures. [12:34:38] ugh! [12:34:49] my fix only adresses sitelinks, that's no good. i'll have to actually get to the root of this [12:34:53] right [12:34:58] i think this happens with aliases [12:35:04] or can happen [12:35:06] and labels. and descriptions. and claims. [12:35:11] yes [12:35:18] right. [12:35:58] * aude could've been able to patch the site link part of the code but thought it's a deeper problem [12:36:55] * DanielK_WMDE_ just wasted two hours [12:37:00] grrrr1 [12:37:00] ! [12:37:45] btw, did you see https://bugzilla.wikimedia.org/show_bug.cgi?id=51192 ? [12:38:35] there's more stuff going on with the data structure inconsistent [12:41:33] fuck, there's too much stuff coming apart. can't follow. [12:41:41] bugzilla doesn't load [12:41:52] :( [12:42:24] in combing through the api code, etc. we're finding these issues [12:42:31] addshore: https://gerrit.wikimedia.org/r/#/c/73316/ [12:42:42] for this to work only test modification is needed [12:43:01] because tests expect different keys [12:43:17] JeroenDeDauw: can you help out with the diff stuff? when the order of things in an assoc array changes, we are seeing a DiffOpChange instead of an empty Diff [12:43:28] he's not here yet..... [12:43:41] JeroenDeDauw: this happens in ItemDiff for sitelinks, but also in EntityDiff for at least aliasesw [12:43:45] aude: grr. [12:43:54] tell him to read the log when he comes in :) [12:43:56] expect he's here soonih [12:43:58] soonish [12:44:32] i though the issue must be with MapDiffer, but it doesn't seem like it is... [12:44:55] i really don't know [12:45:07] it would take me a while to dig through that part of the code [12:45:44] don't know if it's a new issue or what [12:45:58] * aude expected the dispatcher to crash before, if it's an older issue [12:46:18] maybe it's with the patching or something [12:46:49] no, i don't see a connection to patching [12:46:53] ok [12:46:58] * aude has no idea [12:49:36] * DanielK_WMDE_ needs a wtf-o-meter that goes to 11 [12:49:49] :( [13:12:05] New review: Zfilipin; "Tobias, apologies for the late reply. Are you still working on this, or did you gave up on Cucumber?" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/63857 [13:12:18] New review: Zfilipin; "Tobias, apologies for the late reply. Are you still working on this, or did you gave up on Cucumber?" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/62598 [13:14:26] New patchset: Pragunbhutani; "Bug 50578 - Factor out getHtmlForClaim from EntityView" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/72909 [13:15:35] aude: hello! [13:22:59] New patchset: Daniel Kinzler; "(bug 51363) Fix MapDiffer for equivalent substructs." [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73752 [13:23:56] New patchset: Daniel Kinzler; "(bug 51363) Fix MapDiffer for equivalent substructs." [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73752 [13:27:18] aude: hi :) I have deployed your Zuul change to whitelist some more Wikibase people ( https://gerrit.wikimedia.org/r/#/c/73744/ ) [13:28:42] lazowik: indeed :) [13:28:53] aude, Denny_WMDE: the problem existed at least since May 25, see 4a8c67d71b0216b2 in the Diff extension [13:29:15] not sure whether that change introduced it, but it touched the relevant code [13:29:27] anyway, fix is up [13:29:54] i'll follow up with patches for improved robustness. [13:35:14] DanielK_WMDE_: Jeroen just walked in ;p [13:39:58] addshore: make him look at my patch :) [13:57:23] Danny_B: Hi. I've uploaded the svg version of Persian Wikivoyage for incubator, would you please update http://incubator.wikimedia.org/wiki/MediaWiki:Gadget-ProjectSpecificLogos.js with http://commons.wikimedia.org/wiki/File:Wikivoyage-logo-fa.svg thanks [14:19:35] DanielK_WMDE_: How should I get the lastrevid (via JavaScript) for adding a claim? [14:23:11] Granjow: it's probably in wbEntity, but ask Henning_WMDE [14:23:16] i'm not into the JS stuff [14:24:17] New patchset: Liangent; "LanguageWithConversion translation change" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73754 [14:26:01] DanielK_WMDE_: Ok. How would I do it in PHP? [14:26:35] Granjow: depends on what you have. $title->getRevision(), if you have the title. [14:27:01] $title->getLatestRevID(), sorry [14:27:37] DanielK_WMDE_: I have the EntityID. [14:29:52] Granjow: use EntityContentFactory::getTitleForId [14:30:10] it's not the best place for that functionality, but that's where it is currently located [14:30:31] DanielK_WMDE_: Okay, thanks! [14:37:23] hey there. I got a change by Jeroen to make mwext-Wikibase-testextensions-master job voting (aka block whenever it fails) [14:39:29] hashar: sorry for the stupid question is there a link to the job? [14:39:51] https://integration.wikimedia.org/ci/job/mwext-Wikibase-testextensions-master/? [14:40:01] some patches are passing, some others dont [14:40:11] I guess the failing ones need to be rebased [14:41:50] sounds good. I will just quickly check back with him to understand [14:42:06] I guess jeroen fixed most the non passing tests [14:45:38] hashar: i understood now, yes, it is good [14:45:43] hashar: it can be made voting [14:45:48] hashar: thank you! [14:45:56] awesome [14:46:42] deployed it [14:46:49] thanks! [14:47:33] New patchset: Hashar; "Jenkins job validation (DO NOT SUBMIT)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/70395 [14:47:39] trying out :-] [14:48:02] which is https://integration.wikimedia.org/ci/job/mwext-Wikibase-testextensions-master/3055/ [14:49:48] green! [15:00:49] Jeroen is awesome [15:01:17] Denny_WMDE1: if he is around, maybe he could post some message on your internal list to let other people knows about it :-] [15:01:22] DanielK_WMDE_: your patch doesn't seem to work [15:01:25] am i missing something? [15:01:58] hashar: thank you - i will let him know and also the mail [15:02:13] \O/ [15:02:29] :> [15:04:21] New review: Aude; "I still get a fatal error when looping through the changes (in the same way dispatcher does). Am I ..." [mediawiki/extensions/Diff] (master) C: -1; - https://gerrit.wikimedia.org/r/73752 [15:06:41] aude: it does not prevent failure when hitting bad changes. i'm working on that. [15:06:47] New review: Aude; "it looks to me that isEmpty() counts the number of array elements. In case of the site links, old a..." [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73752 [15:06:49] it just prevents new bad diffs from being generated [15:07:00] hmmmmm [15:07:26] would isEmpty() be sufficient? [15:08:23] aude: no. but that's one part that needs fixing. [15:08:31] DanielK_WMDE_: ok, then i wait [15:09:18] nice to set wikidatawiki_p to be my changes database :) [15:09:42] * aude hopes these don't expire soon or else i'll have to load them into my test wiki [15:10:49] pragunbhutani: how's it going? [15:19:07] aude: hello! it's going well [15:19:13] I uploaded another patch set todau [15:19:14] yay! [15:19:20] can you please take a look at it whenever you can? [15:19:26] doing [15:19:35] I made the changes that you had suggested [15:19:53] is see the phpunit tests are reporting some problems [15:19:56] i see [15:20:18] yes they are [15:20:31] it does seem to work just fine on my local install, though [15:20:33] * aude trying to see why [15:20:52] I'll be back in about 20 mins, dinner [15:20:56] ok [15:23:16] heh, 50+ changes to be reviewed [15:24:43] New review: Aude; "some issues (see comments)" [mediawiki/extensions/Wikibase] (master) C: -1; - https://gerrit.wikimedia.org/r/72909 [15:32:40] New patchset: Addshore; "validation in wbgetentities to validateParameters" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73759 [15:39:51] heh, well this makes no sense... [15:39:51] if ( !( strlen( $params['fromtitle'] ) > 0 && strlen( $params['totitle'] ) > 0 ) ) { [15:39:56] $this->dieUsage( 'The from title can not match the to title' , 'param-illegal' ); [15:40:39] what??? [15:40:43] hah! [15:41:20] im not quite sure what it is meant to be checking [15:42:41] the description of the error is deifnatly wrong [15:43:00] is that for wblinkpages ? [15:43:06] yup [15:43:40] that is weird [15:44:06] it should be if strlen(from) == 0 : die (empty from) [15:44:19] if streln(to) == 0 : die (empty to) [15:44:36] indeed what I was thinking, 'The from title can not match the to title' just makes no sense [15:44:40] but again, that seems something that the API can deal with anyway in parameter checking [15:44:47] New patchset: Daniel Kinzler; "(bug 51363) Work around broken diffs." [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73763 [15:44:48] this is also plain wrong [15:44:52] aude: EntityIDFormatter is already there in the Use section [15:45:00] "from=Berlin&to=Berlin" is peerfectly valid [15:45:15] pragunbhutani: then you can use it in the class simply as EntityIdFormatter [15:45:20] without the namespace part [15:46:26] okay, fixed [15:46:37] that's just a style issue [15:46:53] New patchset: Daniel Kinzler; "(bug 51363) Work around broken diffs." [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73763 [15:47:12] aude: this one should do it (with or without the diff fix) [15:47:20] also, why is getFormattedValue an undefined function? [15:47:23] DanielK_WMDE_: ok [15:47:55] aude: the data model fix contains the workaround, the diff fix keeps bad diffs from being created [15:47:55] pragunbhutani: not sure [15:48:35] aude: I've defined it right there! [15:49:00] pragunbhutani: are you seeing the same errors i do? [15:50:21] aude: I just see Undefined variable : entitiesPrefixMap [15:50:41] pragunbhutani: ok [15:50:49] but that's from the test logs posted by jenkins-bot [15:50:59] on my local installation, things are running just fine [15:51:11] do you see any errors in the apache logs? [15:51:33] pragunbhutani: in the jenkins output, can you see which test case causes the error? [15:51:38] can you run that test case locally? [15:51:57] also, make sure you have your error reporting cranked up all the way. [15:52:14] DanielK_WMDE_: let me check and confirm [15:52:32] pragunbhutani: in local settings you want error_reporting( E_ALL ); at the top or somewhere [15:52:40] $wgShowExceptionDetails = true; [15:53:01] also ini_set( 'display_errors', 1 ); [15:53:04] there are other debug settings [15:53:15] yes [15:53:19] $wgShowSQLErrors = true; [15:53:25] $wgDevelopmentWarnings = true; [15:53:47] $wgEnableJavaScriptTest = true; [15:53:51] okay let me get those included [15:54:00] i think these are the most important ones [15:55:33] aude: DanielK_WMDE_ : yeah, now I see the same erros [15:55:35] *errors [15:55:43] pragunbhutani: ok :) [15:55:44] \o/ [15:56:57] valueFormatterOptions is shown as an undefined variable [15:58:13] I heard yours storks have psychological problems [15:58:16] http://www.nydailynews.com/news/world/german-stork-attacks-reflective-objects-article-1.1398595 [15:58:27] New patchset: Jeroen De Dauw; "Fix strict warning" [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73767 [15:59:04] pragunbhutani: let me guess: you forgot $this-> [15:59:30] hello Denny_WMDE1 [15:59:40] what time did you say yesterday? [16:01:13] DanielK_WMDE_: yes he did [16:01:35] lazowik: whut? [16:02:03] pragunbhutani forgot $this-> [16:05:22] New patchset: Daniel Kinzler; "Remove usage of dumpcate methods in ItemChange." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73768 [16:05:30] lazowik: DanielK_WMDE_ : right you both were! :) [16:06:17] may I make a note, master Yoda, that DanielK_WMDE_ guessed, I checked [16:06:23] New patchset: Pragunbhutani; "Bug 50578 - Factor out getHtmlForClaim from EntityView" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/72909 [16:06:53] JeroenDeDauw: got a while? [16:07:17] sorry for bothering you all the time... [16:09:13] New patchset: Daniel Kinzler; "(bug 51363) Work around broken diffs." [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73763 [16:10:14] New patchset: Daniel Kinzler; "Remove redundant methods from ItemDiff." [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73769 [16:10:50] New patchset: Daniel Kinzler; "Remove usage of dumpcate methods in ItemDiff." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73768 [16:11:13] New patchset: Daniel Kinzler; "Remove redundant methods from ItemDiff." [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73769 [16:12:11] liangent: it's 3pm Berlin time [16:12:37] still fine with you? [16:12:56] that's 9pm in China, I think [16:13:28] hmm. maybe we should go for earlier? 11:30 Berlin / 17:30 Beijing? [16:14:40] Denny_WMDE1: nope I may be even impossible to get back in time [16:14:49] :) [16:14:56] but given it's already so late I don't think it'll have a big chance to happen [16:15:05] everyone works so late [16:15:15] liangent: that's tomorrow, no? [16:15:35] today in my timezone :) [16:15:43] New review: Aude; "this allows the dispatcher and other places that diffs are used to be more robust. I verify this ad..." [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73763 [16:15:50] oh, right :) [16:15:53] DanielK_WMDE_: you have comments https://gerrit.wikimedia.org/r/#/c/73763/ [16:16:01] i don't want to mark it +1 or -1 yet [16:16:10] ok, so we stay at 3pm / 9pm ? [16:16:49] * aude confused... today for liangent = tomorrow for us? [16:16:56] aude: yes [16:17:01] ah, ok [16:17:09] since 17 minutes, though [16:19:06] liangent: we stick with 3pm/9pm? [16:19:11] (i need to go … ) [16:20:27] ok, i assume we stick with 3pm Berlin tomorrow/9pm Beijing today … i am off for now! see you liangent [16:21:14] New patchset: Daniel Kinzler; "(bug 51363) Add more tests to for bad diffs." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73772 [16:22:40] hm, i see test failures on master locally :/ [16:22:51] DanielK_WMDE_: like what? [16:23:03] 43) Wikibase\Test\ChangeHandlerTest::testHandleChange with data set #17 [16:23:20] 4) Wikibase\Lib\Test\Serializers\ByPropertyListUnserializerTest::testNewFromSerializationInvalid with data set #3 (array(42)) [16:23:21] oh, i think jenkins is skipping them [16:23:27] doesn't mean that is okay though [16:23:32] o_O [16:23:34] gah [16:23:41] this is new [16:23:52] :( [16:24:23] aude: the tests work when run by themselves [16:24:36] this is nasty interactions again [16:24:48] we can't have such interactions [16:24:56] can't have tests poisoning other tests [16:25:09] well, easier said than fixed [16:25:16] obviously [16:25:27] Denny_WMDE1: well I left for a while... [16:26:49] liangent: he's gone but perhaps back later or can catch him via email [16:27:34] New patchset: Daniel Kinzler; "(bug 51363) Add more tests to for bad diffs." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73772 [16:28:26] damn, i'm losing concentration [16:28:39] time to go outside [16:28:44] play with the kids [16:28:48] New patchset: Daniel Kinzler; "(bug 51363) Add more tests to for bad diffs." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73772 [16:28:58] aude: if it do that, my work day is over [16:29:12] i wouldn't mind JeroenDeDauw to also review the stuff, since he knows the code better [16:29:12] no chance getting my head back into gear later [16:29:16] DanielK_WMDE_: ok [16:29:28] i can verify the patches "work" and look okay to me [16:29:30] but yea, it's just a matter of wrapping up [16:29:34] for that ONE damn bug [16:29:42] yeah, but we've had it for a while [16:30:05] if the dispatcher breaks, i know how to find the bad diff / we can skip it [16:30:06] hm, my changes seem to increase the test breakage [16:30:17] can't see why. but that patch isn't critical anyway [16:30:38] yeah, it's important to get it right and not introduce new bugs :) [16:30:46] even if it takes a little time [16:31:27] amazing to be able to see a copy of the changes table :) [16:31:39] hm, tests on a clean master run fine. I415a8b5c breaks tests, if you run --group Wikibase [16:31:45] hmmmm [16:31:48] if you run the respective tests alone, they pass [16:32:01] ain't it fun [16:32:02] huh [16:34:39] ok, off for now. [16:35:09] ok [16:36:42] New patchset: Jeroen De Dauw; "Added top level factory for Ask" [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73774 [17:22:17] New patchset: Jeroen De Dauw; "Added documentation on serialization code usage" [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73781 [17:26:40] aude: what? [17:28:04] JeroenDeDauw: to review daniel's patches for handling diffs [17:30:03] hi folks [17:30:10] exception log on the wikimedia cluster is full of these: [17:30:12] 2013-07-15 17:01:20 mw1200 wikidatawiki: [XXX] /w/api.php?action=wbeditentity&format=xml Exception from line 1699 of /usr/local/apache/common-local/php-1.22wmf9/includes/GlobalFunctions.php: Internal error in ApiFormatXml::recXmlPrint: (zh, ...) has integer keys without _element value. Use ApiResult::setIndexedTagName(). [17:30:58] looks like it's wikibase, and that the error rate picked up sunday [17:31:39] huh, edit entity and xml? [17:31:43] * aude has no idea [17:33:16] JeroenDeDauw: ping me when you'll have time and willingness [17:35:32] zh? is it related to me? [17:35:51] liangent: are you doing stuff with xml? [17:36:33] aude: I touched serializer at least. not sure whether they've got merged or not [17:37:24] aude: https://gerrit.wikimedia.org/r/#/c/73322/1/repo/includes/api/EditEntity.php possibly? [17:37:39] ori-l: is it only for zh? [17:37:56] liangent: no [17:38:30] Jasper_Deng: I saw that you just blocked Kotakodaga for me it looks like it is a sock of user Tolikara, if this is important [17:38:44] I was the blocker of Kotakodaga. [17:38:47] Pyfisch: I blocked? isn't that TCN7JM's block? [17:39:55] aude: that change looks suspicious... [17:40:01] Jasper_Deng: ok, yes it was not you, but you was the last who editet the page ;-) sorry [17:40:39] ori-l: possibly [17:41:13] TCN7JM: well that ends the story. account was locked by MF-W. [17:41:30] * TCN7JM nods. [17:43:45] aude: have to run, but filed https://bugzilla.wikimedia.org/show_bug.cgi?id=51376 [17:43:51] ori-l: thanks [17:44:05] np, see ya. [17:50:09] New patchset: Addshore; "Ask DeserializerFactory and Tests" [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73786 [17:50:15] JeroenDeDauw: ^^ [17:51:13] aude: link? [17:51:19] lazowik: on what>? [17:51:22] addshore: yay [17:51:28] :> [17:51:33] they even all pass ;p [17:51:49] hacking ofcourse! [17:53:54] addshore: http://pastebin.com/4PEi95vs [17:55:26] New patchset: Jeroen De Dauw; "Also register the integration tests with MW" [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73788 [17:55:38] JeroenDeDauw: https://gerrit.wikimedia.org/r/#/q/owner:daniel.kinzler%2540wikimedia.de+status:open,n,z [17:55:47] addshore: this is why Jenkins did not fail: https://gerrit.wikimedia.org/r/#/c/73788/ [17:56:09] aude: can you be more specific>? [17:56:09] please REEEEEEEEEVIEW the top 5 things in the list (can skip the one that jenkins failed, though) [17:56:17] REEEEEEEEEEEEEVIEW :) [17:56:28] New patchset: Addshore; "Ask DeserializerFactory and Tests" [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73786 [17:56:50] New review: Addshore; "yes please :)" [mediawiki/extensions/Ask] (master) C: 2; - https://gerrit.wikimedia.org/r/73788 [17:56:50] Change merged: Addshore; [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73788 [18:01:36] aude: regarding that recently discovered format inconsistencies in aliases and others [18:02:05] could you check if that doesn't also happen on restore? [18:02:28] (if restore is done by patches) [18:02:56] lazowik: i think daniel might be looking at that [18:03:17] I think that this could also break in the same way [18:03:23] probably does [18:04:15] I need to wait with badges [18:04:43] because badges are also affected [18:04:54] they are non associative array [18:05:24] yep [18:05:50] but I'm pretty sure I've got one-line fix [18:05:53] to Diff [18:06:17] will try [18:09:55] New review: Jeroen De Dauw; "I recommend you update the QueryRoundtrip integration test to use these new factories. That will unc..." [mediawiki/extensions/Ask] (master) C: -1; - https://gerrit.wikimedia.org/r/73786 [18:20:42] I'm not the only person that sees "Tag:" twice, right? http://www.wikidata.org/?diff=58266799&oldid=58266778&rcid=58327650 [18:24:53] TCN7JM: yes [18:24:56] thats a new core change [18:25:05] Erm...why? [18:25:18] I have a script to fix it though [18:25:32] because tags now show up on diff pages [18:26:13] hrmm [18:27:14] ill run it in a few minutes [18:54:19] addnight: my patch to patches (^^) solves one of the failed tests on yours https://gerrit.wikimedia.org/r/#/c/73316 [18:55:17] and here comes the funny part: one of the patches needs to be first [18:55:55] and if it is to pass tests, it needs to "correct" the test to wrong [18:56:03] and then the second patch needs to do the opoosite [18:56:21] correct from "corrected" to correct... [18:56:41] precisely it's this: Wikibase\Test\ItemDiffTest::testApply with data set #7 [18:58:48] wait [18:59:02] it cannot be "corrected" to wrong even [19:10:48] New patchset: Michał Łazowik; "(51192) Resolve data inconsistency in patcher." [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73797 [19:12:09] New patchset: Michał Łazowik; "(bug 51192) Resolve data inconsistency in patcher." [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73797 [19:27:17] lol [19:27:21] got stuck on gerrit bug [19:27:45] and couldn't see a patch on gerrit web [19:28:21] so after checking out I saw different code than on gerrit [19:32:30] what's the preferred way to get a string representation of an EntityId for internal use [19:32:57] more specifically, to use as keys of a flag array to avoid duplication [19:34:51] liangent: getPrefixedId() ? [19:35:04] lazowik: it's tagged as deprecated [19:35:13] ah [19:35:15] yep [19:35:55] liangent: then just do what it does :) [19:36:07] $this->getId()->getPrefixedId() [19:36:31] nope [19:36:38] also deprecated [19:37:30] there's __toString() in EntityId [19:38:18] is it generally fine to use a double underscore function literally? [19:38:37] I'm not an expert... [19:40:21] lazowik: liangent use the entity id formatter [19:41:41] aude: is it too expensive? [19:41:53] I don't need to format it to any readable format [19:41:54] liangent: and ifyou have no better way to get the id formatter, it is registered in the "global" registry for repo [19:42:26] i suppose you won't instantiate a new one each time [19:42:36] it might even be persistent [19:43:15] yes it is, if obtained from the registry thing [19:43:41] aude: ok [19:43:53] i wouldn't use a deprecated function [19:47:49] aude: and it would be better if the person who tags a function as deprecated also note the alternative function to use [19:51:50] liangent: agree [19:53:23] liangent: aude: the registry should not be directly accessed from inside classes specifying runtime logic in new code [19:54:06] actually now that data model is split from wikibase..... [19:54:40] the part that asks for Settings entity prefix stuff is depending on wikibase and i think that will be eliminated soon [19:55:05] liangent: if you want to get an unique key, you can serialize the id [19:55:15] i'm not sure where the prefix stuff should go [19:55:29] JeroenDeDauw: well it's just a bit longer [19:55:50] liangent: and that is a problem? [19:56:39] JeroenDeDauw: no [19:58:48] New patchset: Jeroen De Dauw; "Fix strict warning" [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73853 [20:01:03] JeroenDeDauw: the change is totally wrong? [20:01:14] lazowik: what change? [20:01:22] https://gerrit.wikimedia.org/r/#/c/73797/ [20:01:44] I mean "totally" [20:02:03] is it a bigger issue than that? [20:04:30] New patchset: Liangent; "LanguageWithConversion has a new reverseTranslate() function" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73754 [20:04:51] New patchset: Liangent; "LanguageWithConversion has a new reverseTranslate() function." [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73754 [20:06:20] New patchset: Liangent; "LanguageWithConversion fix" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73854 [20:07:17] lazowik: Just considered this for a few minutes and bounced it off Danwe_WMDE: conclusion: this is not the right place to put the array_values [20:07:23] This patchcer can create gaps in your array [20:07:35] Removing gaps at the end will make it remove all gaps though [20:07:47] Which leads to unexpected behvaiour when you input an array with gaps [20:08:01] Removing gaps is not part of this things responsibility [20:08:09] mhm [20:08:12] If you don't want gaps, remove them after getting the patches thing [20:08:16] *patched [20:08:17] New patchset: Bene; "Special page for entities without description" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/67634 [20:08:32] this also affects badges [20:08:33] New review: Jeroen De Dauw; "(22:07:16) JeroenDeDauw: lazowik: Just considered this for a few minutes and bounced it off Danwe_WM..." [mediawiki/extensions/Diff] (master) C: -2; - https://gerrit.wikimedia.org/r/73797 [20:08:47] they are also an array [20:08:58] lazowik: yeah sure [20:09:06] so now I'm waiting for this to settle down [20:09:08] Sounds like putting in an array_value si the correct thing to do [20:09:12] Just not where you did it now [20:09:33] lazowik: this however indicates the docs for this patcher where inadequate, I will fix this [20:09:59] but probably I'll need to write a function to Diff sitelinks [20:10:13] lazowik: what code is running into this issue? [20:10:32] because now when sitelink is added Diff treats this as modification [20:10:50] I can get you print_r of diff if you want [20:11:01] patching is ok [20:11:22] when I do DiffOpAdd( whole sitelink structure ) [20:12:27] but EntityDiff outputs Diff( DiffOpAdd( pageName ), DiffOpAdd( badges ) ) [20:12:38] or something along the lines [20:12:50] New patchset: Jeroen De Dauw; "Improved ListPatcher docs" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73856 [20:12:57] New patchset: Jeroen De Dauw; "Improved ListPatcher docs" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73856 [20:15:34] lazowik: the structure needs to be normalized before being handled to the diff code [20:19:46] New review: Nikerabbit; "(1 comment)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73768 [20:40:19] New review: Michał Łazowik; "OK, but maybe it would be good to write why is this a desired behaviour." [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73856 [21:33:39] Change merged: Jeroen De Dauw; [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73768 [21:33:55] Change merged: Jeroen De Dauw; [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73769 [21:34:35] New review: Jeroen De Dauw; "This code does not have tests?" [mediawiki/extensions/Wikibase] (master) C: 1; - https://gerrit.wikimedia.org/r/73854 [21:36:52] New review: Jeroen De Dauw; "(1 comment)" [mediawiki/extensions/Wikibase] (master) - https://gerrit.wikimedia.org/r/73772 [21:43:31] Change merged: Daniel Werner; [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73853 [21:44:36] New review: Jeroen De Dauw; "violating method contract" [mediawiki/extensions/Diff] (master) C: -1; - https://gerrit.wikimedia.org/r/73752 [21:45:02] Change merged: Daniel Werner; [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73856 [21:48:14] New review: Daniel Werner; "(1 comment)" [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/72960 [21:51:15] Change merged: Daniel Werner; [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73774 [21:52:48] Change merged: Daniel Werner; [mediawiki/extensions/Ask] (master) - https://gerrit.wikimedia.org/r/73781 [21:55:25] New review: Jeroen De Dauw; "On second look, the contract is not actually violated." [mediawiki/extensions/Diff] (master) C: 2; - https://gerrit.wikimedia.org/r/73752 [21:55:28] Change merged: Jeroen De Dauw; [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73752 [21:59:43] New patchset: Jeroen De Dauw; "Update release notes" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73872 [21:59:43] New patchset: Jeroen De Dauw; "Remove incomplete Merger code from master so it does not block release" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73873 [22:02:21] New review: Jeroen De Dauw; "(1 comment)" [mediawiki/extensions/WikibaseDataModel] (master) C: 2; - https://gerrit.wikimedia.org/r/73763 [22:02:24] Change merged: Jeroen De Dauw; [mediawiki/extensions/WikibaseDataModel] (master) - https://gerrit.wikimedia.org/r/73763 [22:02:43] New patchset: Jeroen De Dauw; "Remove incomplete Merger code from master so it does not block release" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73873 [22:02:51] Change merged: Jeroen De Dauw; [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73872 [22:03:28] Danwe_WMDE: can has merge https://gerrit.wikimedia.org/r/#/c/73873/ ? [22:04:27] New patchset: Jeroen De Dauw; "Remove incomplete Merger code from master so it does not block release" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73873 [22:06:42] Change merged: Daniel Werner; [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73873 [22:08:23] New patchset: Jeroen De Dauw; "Re-add incomplete Merger code which was pulled from master so to not block a release." [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73875 [22:08:56] New patchset: Jeroen De Dauw; "Re-add incomplete Merger code which was pulled from master so to not block a release." [mediawiki/extensions/Diff] (merger) - https://gerrit.wikimedia.org/r/73876 [22:09:39] Change merged: Jeroen De Dauw; [mediawiki/extensions/Diff] (merger) - https://gerrit.wikimedia.org/r/73876 [22:11:17] New patchset: Jeroen De Dauw; "0.7 release" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73878 [22:11:28] Change merged: Jeroen De Dauw; [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73878 [22:13:29] [travis-ci] wikimedia/mediawiki-extensions-Diff#33 (merger - c389299 : Jeroen De Dauw): The build passed. [22:13:29] [travis-ci] Change view : https://github.com/wikimedia/mediawiki-extensions-Diff/compare/1cb4159fdb66...c389299a193c [22:13:29] [travis-ci] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-Diff/builds/9115626 [22:17:43] New patchset: Jeroen De Dauw; "Improve doc" [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73881 [22:18:55] Change merged: Jeroen De Dauw; [mediawiki/extensions/Diff] (master) - https://gerrit.wikimedia.org/r/73881 [22:21:27] https://github.com/mame/quine-relay [22:21:29] winwinwin [22:25:14] looool [22:25:53] on gentoo might do similiar but with compilers/interpreters [22:26:17] of course if one has several lifes to spent [22:27:09] JeroenDeDauw: isn't the time when one starts making others angry with written code getting dangerously closer?