[07:58:42] petan: petan|lounge: You at the hotel? [08:05:49] session about the new EditGroups tool starting in 25 minutes in Q1/0003 :) [08:21:26] Can anyone verify if a t-10 ticket works on the buses around here? [08:34:35] o/ [08:35:10] o/ [08:35:28] good morning sjoerddebruin :) [08:35:39] Indeed, you too. [08:36:43] Isarra: The page at https://www.tmb.cat/en/barcelona/fares-metro-bus/travel-cards/t-10 seems to say that they work for both the train and the bus. [08:36:56] But I'm not a local :) [08:37:09] ask someone with a yellow bandana! [08:40:32] Nikerabbit: Thoughts on $wgAllUnicodeFixes? [08:42:22] bd808: Yeah, thanks. XD [08:42:24] Krinkle: out of context: It's something that maybe should be enabled by default but there are performance concerns? [08:42:29] ...except I keep missing them anyway. >.> [08:43:21] Nikerabbit: I'm just curious about the thoughts behind it. Like, whether the intention is to remove the per-language normalize and just do them all always. (Given we have multi-language content and things like usernames and translations etc.) [08:43:46] Or maybe the intention is that it should be per-content language, and is for rare cases, and $wgAllUnicodeFixes is paranoia that nobody uses? [08:43:47] What's the point of a bus every twenty minutes that saves me ten minutes when I keep just missing it and it only takes me twenty minutes to walk the entire way anyway? [08:44:19] Krinkle: IIRC we only have extra normalization for ml and ar, that are additional to some normalisations we do always [08:44:26] They usually let you on if you hail [08:44:41] Assuming they drove past. [08:44:49] The bus only makes sense for Holiday Inn tbh, the walk is pretty nice before sitting inside the whole day. [08:45:09] I would always walk if I stayed in the campus hotel. [08:45:17] Nikerabbit: Anyway, for the short-term - https://gerrit.wikimedia.org/r/#/c/434040/1/languages/Language.php [08:45:24] that might reduce overhead a bit and simplify things overall [08:46:21] I don't have data on current performance, but I suppose we could try and measure it, and one way to improve it would be to combine the normalise data sets in the maintenance script, and use it always in Language::normalize() in a single pass. Assuming that that works. [08:46:26] It currently does not appear to have tests. [08:46:37] When I was working on it, I misspelled the file name, and no tests failed. [08:47:59] Krinkle: I think the logic was: most wikis don't have malayalam or arabic text in significant amounts-> skip them by default [08:48:19] IIRC Tim introduced the setting so that I could easily enable it on translatewiki.net [08:48:22] Nikerabbit: Right. Could be a simple boolean that we disable by default, and enable in prod? [08:48:40] E.g. disable for small third party wikis. [08:48:46] But then again, if we perf is fine, we shoudl just enable it. [08:49:11] However, it did strike me that these files have been unchanged for a long time. [08:49:15] small third party wikis probably wouldn't notice anything in performance [08:49:46] I'm curious as to what it actually does for end-users and how it relates to our Utfnormal library. [08:49:58] Like, would it make sense to add to that instead? [08:50:13] i would've loved to walk from the holiday inn to the venue. but... you can't [08:50:25] Santhosh can probably explain the case for malayalam [08:51:02] aside, I love how the diff in https://gerrit.wikimedia.org/r/c/434040/1/languages/data/normalize-ar.php#603 displays for me with some arrows going => and some <= [08:52:50] mesmerizing [08:53:00] i didn't know PHP allowed right to left arrays! [09:00:37] WiFi dead for everyone? [09:05:31] next bus? [09:05:52] James_F: yes! [09:06:27] Spoken to the organisers, it’s known, they’re trying to fix it. [09:09:25] suddenly we have 13 wifi hotspots as everyone is tethering [09:14:58] Craptastic HHVM - https://github.com/composer/composer/issues/7361 [09:15:40] So apparently some of the 3.2x releases in HHVM either toggle php7.all by default, or otherwise make PHP_VERSION return 7.1.x instead of 5.6.99 [09:15:41] https://3v4l.org/Rn6BI [09:20:03] Which is a problem, because one of the more obscure changes in PHP 7, to support empty string as object property, is not supoorted by HHVM, and Composer makes use of that feature. [09:33:01] gilles: landed your blubber patch but `arc land` squashed the commits and put me as the author... because arcanist. once i figure out how to force push to diffusion i will fit it :) [09:34:41] *fix* it rather [09:39:01] gilles: fixed! thanks for the patch! [09:48:59] is stas on IRC? [09:53:27] harej: his bouncer at least is... nick SMalyshev [10:00:17] Spread the word! The location for the group photo at 15:00 will be outside the auditorium, around the back. [10:00:25] We'll direct people through the auditorium lobby to the location in the linked photo. @Mholloway, @Petr_Urbanec and I scouted the location this morning and we should all be able to fit. [10:00:30] If you want to help take photos or help wrangle folks to the location, please arrive a little early, say 14:45 or so. [10:02:05] https://usercontent.irccloud-cdn.com/file/YiaI2u7d/hackathon-group-photo-location.jpg [10:10:05] Quality Assurance SIG (special interest group) meetup started at QC0011 [10:24:05] zeljkof: is there a better version of say - `assert.equal( NewsletterPage.tableHasRowWithNewsletterDetails( newsletterName ), true );` - like I want an assert.isTrue() - they say it do not exist :-( [10:33:27] tonythomas: it's just assert() :) [10:33:35] I see :D [10:35:27] zeljkof: btw - https://gerrit.wikimedia.org/r/#/c/433737/ - kind of okeyish browser test to review [10:36:11] tonythomas: in a session, will take a look [10:36:16] nice. Thanks. [10:51:08] who broke notifications on wikitech? [10:51:29] https://usercontent.irccloud-cdn.com/file/yoIHkXG4/Wikitech%20Notifications.png [10:52:19] is there a way to detect the currently running ResourceLoader module from inside that module? (JavaScript) [10:52:26] (in case of code copy+paste) [10:52:40] I can iterate over mw.loader.moduleRegistry and check the state of each module, but that feels very fragile [10:52:51] s/is there a way/is there a supported way/ [10:56:28] addshore: You around ? :) [10:58:43] Lucas_WMDE: ? [10:59:07] Harmonia_Amanda: ? [10:59:22] did I miss an earlier message by you? [10:59:28] Ash_Kiwi is doing a new version of the deleting tool today [10:59:42] Lucas_WMDE: will that change something in your namescript? [11:00:11] I don’t think so [11:00:12] no, no new message, I just didn't want to disturb you if you were busy :) [11:00:18] ok, thank you! [11:00:22] unless it’s a change in functionality that you also want in namescript [11:00:34] Harmonia_Amanda: well I’m still busy with namescript, so it wouldn’t really be disturbing ;) [11:00:54] I don't think so, it's getting the browser interface better [11:01:03] ok [11:01:07] so no changes for what we need in namescript [11:01:42] Lucas_WMDE: don't hesitate to ping me if you need [11:01:46] ok [11:01:51] probably going for lunch soon though [11:01:55] :) [11:29:26] CindyCicaleseWMF: me and isarra was wondering if we could get a demo of commentstreams [11:30:18] Group photo is at 15:15 at the auditorium (where we had the opening) [11:30:26] Directly afterwards we have the showcase / closing [11:31:29] bawolff: sure! I'm at a picnic table by the cafeteria now, but I'm coming up soon. [11:32:43] We could head over there [11:38:40] bawolff: Great, I'll stay here. [11:38:52] If you want to coordinate your travel to the airport with others I have added a space for that in the social eitherpad [11:38:57] https://etherpad.wikimedia.org/p/WMhack18_Social [12:06:26] zeljkof: will you merge https://gerrit.wikimedia.org/r/#/c/346982/ ? [12:11:48] physikerwelt: sure, in session now, but later [12:25:54] bawolff: I'm now towards the far end of the main corridor [12:57:39] bd808: a person here in my table asks for some help with pywikibot [13:34:39] greg-g! [13:59:04] zeljkof, why there's no EU SWAT on Tuesday? Was there some change? [14:04:47] did siebrand say we can't use own laptops for showcase? [14:05:02] petan|lounge, yes, he did [14:05:17] ok in that case I have nothing to show :( [14:05:20] I guess that's because of the problem to connect their "provided laptop" :D [14:05:29] petan|lounge, make screenshots ;) [14:06:07] Screenshots, some online presentation and shortcut link and you can present Huggle :) [14:06:14] But of course, it's up to you [14:08:28] Urbanecm: not sure, copy/paste error? Cc greg-g [14:08:45] (no eu swat on Tuesday) [14:08:52] zeljkof, there's no error. There's just WMCS network maintenance instead of EU SWAT on Tuesday [14:09:39] Urbanecm: ah, ok [14:09:46] "WMCS network maintenance -- no SWAT " [14:09:49] :) [14:10:13] greg-g, exactly. When it will be back to normal? Or is this a normal state for now? [14:10:35] Urbanecm: the next day [14:10:36] (because EU SWAT and Morning SWAT are the only SWATs I can use without being awake 24/7) [14:10:52] ah, they aren't there the following week, that's a copy/paste error [14:11:04] greg-g, I meant...is it planned to have 2 SWATs again on Tuesdays? [14:11:12] yes, see last line :) [14:11:51] greg-g, so, there should be a EU SWAT on 05/22? [14:12:05] (just if I understood correctly) [14:14:51] Urbanecm: yes. (we need to automate this dang schedule sometime) [14:15:41] That's great. Thanks greg-g! [14:15:57] Ooooh, automatic schedule [14:16:05] ACTION edits [14:16:07] There should be a tool for that? :0 [14:16:21] Anyone in the room want a lemon Fanta? [14:16:36] Or a coke zero [14:16:50] Or even an orange Fanta.... [14:20:07] Urbanecm: check please :) [14:20:31] greg-g, it's fine now. Thanks a lot! [15:34:32] Urbanecm: I could demonstrate in the showcase how to piss off commons as something I accomplished this weekend ;) [15:39:11] bawolff, :D I don't think it is a good idea [15:39:45] I scheduled the patch for deployment for Tuesday EU SWAT, if you want to re-schedule it sooner, feel free to. [15:40:04] Urbanecm: If I'm free during the mid-day european swat (I'm not sure I will be) I will do so [15:41:10] yeah, I really screwed up handling the rate limit change stuff [15:41:54] I blame timing of it all, but that's not an excuse [16:00:26] Marshall_: I am back at the venue. Where are you? [16:11:21] I'm doing a session on Timeless (that horrible colourful mw skin) in five minutes or so in Flaix/Q3/0003. [16:11:26] Come and tell me how horrible it is. [16:17:15] Lucas_WMDE: hi again :p [16:17:27] hi! [16:17:33] back at the place with the working wifi :) [16:17:55] :) [16:33:40] Anyone up for a game of "BANG!"? I think it supports 7 players... :) [16:52:45] When do the busses leave [16:59:12] Or when is the next one? [17:19:28] bawolff: better ask in the help desk [17:20:22] I did. They did not know [17:28:11] !ping is pong [17:28:11] This key already exists - remove it, if you want to change it [17:28:14] yay [17:28:18] bot is working again [17:28:46] the bad news it was dead whole hackthon in this channel so there are no logs stored [17:57:51] @relay-on [17:57:52] Relay was enabled [17:57:54] goats are great [17:58:01] :D [17:58:05] @relay-off [17:58:05] Relay was disabled [17:59:14] darn it [17:59:16] My python is really rusty :( [17:59:19] I forgot how to generator. [17:59:33] petan|lounge: I don’t understand, are the logs at https://wm-bot.wmflabs.org/browser/index.php?display=%23wmhack incomplete? [17:59:47] I definitely saw something there for most of the weekend [18:05:33] Lucas_WMDE: I have a batch of 129 725 family names to clean using the off-browser namescript :D http://tinyurl.com/y8zm2ghg [18:06:05] oh dear lord [18:06:23] Lucas_WMDE: that's why I wanted an automatic tool in the first place [18:06:28] I guess I need to update the draft of the blog post I’m writing, the “2000” is outdated :D [18:06:45] because I knew I would never be able to clean that manually [18:06:56] but I’m very glad about that, because for just 2000 items I’m not sure if the time I spent on it was really less than just doing the 2000 items by hand :D [18:07:04] 100k is a lot more convincing :> [18:07:12] Lucas_WMDE: and that's just the first batch [18:07:20] I have other SPARQL queries waiting [18:07:37] do all of those have broken descriptions too? [18:07:38] you definitely didn't do it for nothing :) [18:07:42] yep [18:08:16] ok [18:09:10] I was very unhappy when GZWD created all these items without caring about descriptions [18:09:21] because I had no easy way to repair it [18:09:26] but now I can! [18:09:30] ACTION all happy [18:09:54] it might be worth it adding some extra features to the script, in that case [18:10:09] so you don’t have to copy all those item IDs into the command line [18:10:15] oooohhh [18:10:20] would that be possible? [18:10:21] perhaps the script could run the query for you [18:10:24] sure [18:10:38] Lucas_WMDE: ♥ [18:11:00] We can make a query with SELECT only the Qid [18:11:33] that’s probably also a lot faster, without the query service [18:11:46] yeah, I just wanted to verify the results [18:12:18] http://tinyurl.com/yd5ldcv4 [18:12:27] do you really need a CONTAINS for the reference URL? it looks like that should be the full URL [18:12:43] oh yeah, you are right [18:12:52] I started from another query [18:13:04] yeah, same number of results for http://tinyurl.com/yb9qn5j6 [18:13:56] sometimes when it work under the timeout limit, I'm not careful ^^ [18:14:04] :D [18:17:47] (:D, there are too many entries, LibreOffice Calc is hating me and refusing to work) [18:25:31] while people were talking about using two factor for mediawiki accounts. What would it take for a user (no admin, bureaucrat) to get this feature ? [18:25:48] */s/while/saw/d/' [18:26:29] tonythomas: mostly that there's a better user experience for when the user loses the second factor; if i remember correctly you have to get an engineer to manually reset the second factor at the moment and that won't scale to our whole user base [18:26:37] short of that, I think there are certain user rights you can get? [18:27:46] hmm. thats sad. probably this was extensively documented somewhere (I guess) ? would be great to look at it and what is blocking us in 2018 to enable some kind of fallback on losing 2A [18:27:52] *2FA [18:28:38] Lucas_WMDE: maybe not from a query (I think than other than this first batch, all other will need to be cleaned up), but could namescript use a csv with column A being all Qid? [18:28:50] Lucas_WMDE: that seems way more practical [18:28:54] sure [18:58:19] Lucas_WMDE: problem! [18:58:44] the browser version doesn't show anymore the error message when the script fail because of duplicate? [19:00:42] Lucas_WMDE: https://www.wikidata.org/wiki/Q37558027 and https://www.wikidata.org/wiki/Q15854124 [19:01:23] (I'm not correcting it so you can see for yourself) [19:36:20] Harmonia_Amanda: I’m ready to review it now but you just merged the items it seems [19:36:34] but I can take a guess at a fix… [19:36:39] ah [19:36:45] I'll find you another [19:39:32] Lucas_WMDE: https://www.wikidata.org/wiki/Q4115189 [19:39:44] https://www.wikidata.org/wiki/Q37096481 [19:40:01] kthx [19:40:29] ah, I see [19:40:34] I was confused by the sandbox for a sec :D [19:40:50] ^^ [19:41:09] the script was taking too long to find another duplicate [19:41:28] (I cleaned up too many duplicates these last few months it seems :p) [19:41:34] hm, I’m not seeing anything… [19:41:45] what should the script do if there wasn’t an error? replace all the labels? [19:41:50] Lucas_WMDE: that's the problem [19:42:37] Lucas_WMDE: it should gave you a warning (like the one which "success, updated labels and descriptions") but with the Qid of the duplicate and the language [19:43:02] Lucas_WMDE: basically the error message it gave us on shell should be visible in the browser too [19:43:02] but I’m looking at the API response from Wikidata and there’s no warning [19:43:09] ??? [19:43:15] I’m confused too [19:43:25] Lucas_WMDE: the warning was there in the shell [19:43:25] as far as I can tell the script isn’t even trying to set labels? only aliases? [19:43:38] that's why I opened the Wesemann items [19:43:57] because it failed off-browser and there was an error [19:44:01] the wbeditentity request that the item sends starts with {"descriptions":[],"labels":[],"aliases":[[{" [19:44:09] it’s not trying to adjust any labels or descriptions [19:44:17] what? [19:44:28] but it *did* adject labels [19:44:32] adjust* [19:45:02] on which item should I be running this on? sandbox or the real one? [19:45:10] Lucas_WMDE: the real one [19:45:12] ah [19:45:22] let’s try that :) [19:45:33] the sandbox should be the Qid given in the error message [19:45:40] with fr as language of duplicate [19:45:43] okay, yes, there’s an error in the console now [19:46:02] I can debug that [19:46:05] Lucas_WMDE: https://www.wikidata.org/w/index.php?title=Q15854124&type=revision&diff=682180647&oldid=682180568 I confirm the script work and add labels, descriptions and aliases [19:46:10] ok, great [19:46:33] might be offline for a bit during the bus ride to the hotel [19:47:58] Lucas_WMDE: duplicates: Q43610626 and Q37540754 [19:48:16] well no [19:48:19] urgh [19:48:25] I'll clean that one up [19:49:16] ACTION hates when names written in different writing systems get mixed up [19:53:41] great Hackathon everybody! [20:12:43] Harmonia_Amanda: so I shouldn’t test with Wesemann for now? [20:13:13] Lucas_WMDE_: you can test with https://www.wikidata.org/wiki/Q37096481 (with https://www.wikidata.org/wiki/Q4115189 acting as duplicate) [20:13:34] ok thanks [20:13:37] I found three new duplicates during that time but I cleaned them up [20:13:50] I will let you play with these two :) [20:19:57] okay :) [20:20:27] (I think I’m online again now?) [20:20:42] yes [20:41:53] urgh [20:42:07] mw.Api just gives you the error code on error, apparently [20:42:09] nothing else [20:43:34] aaaahhhh [20:43:37] it does give you more stuff [20:43:40] in extra parameters [20:45:18] not that it would document that, of course :/ [20:48:07] Lucas_WMDE: :s [20:48:52] and I keep forgetting that I need to run namescript-upload.js before reloading in the browser :D [20:49:03] ^^ [21:01:45] alright harmonia_amanda should be fixed now [21:01:55] do another `git pull` then `node namescript-upload.js` [21:02:27] Lucas_WMDE: I have an off-browser version working right now [21:02:32] ok [21:02:41] then it’s probably best to wait [21:03:07] Lucas_WMDE: but as there are already three duplicates in that batch, once it's finished i'll be able to test right after [21:03:20] looks like namescript edits at just under one edit per second: https://tools.wmflabs.org/editgroups/ [21:03:48] great