[00:57:37] <^Mike> Is testwiki supposed to be reporting so much lag? http://test.wikipedia.org/w/api.php?action=query&meta=siteinfo&siprop=dbrepllag [00:57:52] <^Mike> its been doing so for quite a while now [01:03:35] It's because its DB backends are in Tampa and they're unloved [01:07:58] ^Mike: I've filed a ticket with our operations team [01:21:35] <^Mike> alright, thanks RoanKattouw [04:16:00] Hey folks, quick basic question. I'm looking at the guided tour code on gerrit, and I can see the recent change but I want to look at the entire file and I'm not sure where to find it, frankly. Last update is here: https://gerrit.wikimedia.org/r/#/c/67034/ [04:20:18] Ocaasi__: yeah, gerrit is a usability train wreck [04:20:54] Ocaasi__: Try https://git.wikimedia.org/tree/mediawiki%2Fextensions%2FGuidedTour.git/HEAD [04:21:03] bawolff: heh, thanks! [08:05:23] Why there are {{#expr: ( 36.222978 + 36.211321 + 36.247889 + 36.263799 + 36.21779 ) / 5 }} inside {{coord }} template ? [08:06:21] Is this a valid coordinate? [08:06:37] I get on my parser latitude = {{#expr: ( 42.585115 + 42.480296 ) / 2 }} longitude = {{#expr: ( 142.128639 + 142.074389 ) / 2 }} [08:06:58] Should I parse the contents of that and do the calculation myself? [08:08:04] i.e. ( 42.585115 + 42.480296 ) / 2 for latitude and 142.128639 + 142.074389 ) / 2 for longitude [08:09:56] BadDesign: How did you get that? By expanding the coord template somehow? [08:11:08] If you got that kind of output from your local MediaWiki, then maybe you need to install the ParserFunctions extension [08:11:30] RoanKattouw: from parsing the Wikipedia dumps [08:11:39] Ah right [08:11:53] I need to extract the main coordinate of the articles [08:11:56] Right [08:11:59] Wait [08:12:09] They put {{#expr}} stuff *in the article*?! [08:12:11] I think this is a corner case [08:12:43] RoanKattouw: some articles have that yes... inside {{Coord {{#expr }} }} [08:12:53] * RoanKattouw shakes fist at Wikpedians [08:12:56] OK, ignore what I said then [08:13:05] Wikipedians are much more evil than I had imagined [08:13:39] I also learned today that tags within tags (using the {{#tag}} parser function of course, because you can't do that otherwise) are apparently somewhat common [08:13:53] Which made me sad and angry that I now have to go and build support for that case in VisualEditor [08:14:15] I feel for you [08:16:16] Some areas of Wikipedia are really inconsistent [08:17:59] hmm I have an idea [08:18:43] I should start helping them to put the main article coordinate as a tag in the dumps instead of me having to parse it from text [08:19:23] [08:19:42] that'd be much easier [08:39:09] it's the edge cases that kill you.. every time [08:43:40] Yes. Also not trusting Nemo_bis when he says parsing the dumps is a VERY bad idea. [08:56:44] RoanKattouw_away: you're only discovering the horrors of wikitext now? :-D [09:12:06] Nemo_bis: I'll try to help too with the GeoData thing... [09:14:23] GeoData should have an API endpoint that returns all the articles on Wikipedia in language X that have a main coordinate attached [09:14:35] which pagination of course [09:14:42] s/which/with [09:15:23] I want to do something similar to Google Maps and their Wikipedia overlay [09:16:29] How does Google grab the main coordinates for each article and what criteria do they use to show a bigger marker for some articles and a smaller one for the others? https://en.wikipedia.org/wiki/Talk:Google_Maps#Wikipedia_markers_in_Google_maps [09:16:49] Someone asked that there but no one answered [09:41:53] valhallasw: to be fair, one thing is discovering and another taking a full-body full time dive in it [09:43:43] hmm, after I parsed the english wiki dump I only got 312180 articles with coordinates but the https://en.wikipedia.org/wiki/Template:Coord page claims it is used in 830,000+ pages. [09:44:06] that includes all the other languages not only the english dump? [09:44:36] looking [09:44:44] no, they would mean just en wp [09:45:05] did you just look at pages-articles or at all the meta pages too? [09:45:20] apergos: only page-articles [09:45:36] prolly used on a bunch of talk pages or whatever [09:46:00] I parsed coords usign {{Coord template as well as the infobox [09:46:30] if the {{Coord is not available [09:46:38] https://en.wikipedia.org/w/index.php?title=Special%3AWhatLinksHere&limit=500&target=Template%3ACoord&namespace=1 [09:46:41] see there's some [09:54:36] Inside the infobox https://en.wikipedia.org/wiki/Template:Infobox_settlement the coords can only be specified using the latd,latm,lats,latNS, longd, longm, longs, longEW right? [09:55:55] I first look if in the infobox there is a coordinates_display with the value of title in it (or inline,title) [09:57:36] if it is blank it is not present I assume that article doesn't use the infobox to specify its main coordiante [09:57:50] *or it is not present [10:16:40] I got 256423 articles with {{Coord template and 59553 within infobox that don't have {{Coord [10:17:13] Are there any other ways to specify the article's main coordinate? [10:25:04] Hi! Does anybody know how to configure the sidebar for the mobile view? We have MediaWiki:Sidebar for the desktop view for this purpose [10:30:02] there isn't a sidebar in the mobile view is there? [10:40:03] https://en.m.wikisource.org You can click a button in the left upper corner to see the side menu [11:57:06] What's up with the lat_d long_d way to specify the main coordinate of an article? [11:57:09] Theat's the old style? [11:57:32] BadDesign: what you're asking is likely en.wp-specific [11:57:56] BadDesign: various wikis use various styles, and only a few care about any sort of a consistency [11:58:10] (en.wp doesn't) [11:58:14] :(( [11:58:51] Why on earth the English Wikipedia which is the most important project doesn't have a consistent way to specify the main coordinate of an article? [12:00:34] Is there a wikipage where can I see the various ways the coordinates can be specified? [12:03:23] BadDesign: no idea, you might have more luck asking on #wikipedia or #wikipedia-en [12:03:43] BadDesign: if you're looking for infobox coordinates done mostly right, check out pl.wp :P [12:04:28] we have one standard way of specifying coords in infobox (apart from one or two outliers), and one single coordinates template [12:04:47] personally i just don't understand how did en.wp's templates get so unwieldy [12:22:53] the hell [12:23:02] in infobox there are at least 3 ways to specify coordinates [12:24:56] 1) |LATITUDE = val \n|LONGITUDE = val 2) |lat_d = val \nlong_d 3) lat_m lat_s lat_ns long_d long_m long_s long_ew 4) latm lats latns longd longm longs longew [12:25:55] Clearly, there should be only one way to specify the main coordinate of the article [12:26:02] otherwise this is BadDesign! [12:53:24] The VisualEditor is heavily broken: https://de.wikipedia.org/w/index.php?title=Oper_K%C3%B6ln&diff=119548868&oldid=118354900 some bug already opened for this [12:55:40] Raymond_: I'm not sure I would call something still in alpha "broken". [12:56:18] T13|mobile: it's beta since yesterday and until yesterday the alpha version had not make such a code mess [12:57:31] *sigh* beta =\= release [12:57:55] beta still == not finished [13:01:47] T13|mobile: I know very well! [13:02:09] thats not the point [13:03:07] Raymond_: there's a bug for that [13:03:26] MatmaRex: yes: https://bugzilla.wikimedia.org/show_bug.cgi?id=49577 [13:04:00] MatmaRex: But I guess it is some kind of wild running server currently [13:04:36] man, why does everything break on fridays [13:04:40] James_F|Away: *poke* [13:04:48] gwicke_away: *poke* [13:05:08] The beta worked last night better than today [13:06:17] Raymond_: it's a rapidly evolving piece of software [13:06:24] and sometimes stuff gets broken [13:06:36] and sometimes, all the people who could fix it are "away" [13:06:42] *especially* on fridays and weekends. :P [13:07:04] a variant of Murphy's law :P [13:25:54] Away in bed [13:25:54] Damn them for sleeping [13:27:14] you mean there are other spots people should irc from? [13:31:04] Reedy: WMF needs more people in or around UTC [13:31:46] There's a VE guy in the UK [13:31:47] @notify Ryan_Lane [13:31:47] I'll let you know when I see Ryan_Lane around here [13:32:34] I'm guessing you need a dev [13:33:51] I'm not sure [13:33:54] it's only VE ;) [13:34:13] :-D [13:36:41] * apergos thinks that wtp1008 looks odd [13:37:04] if its as broken as people say, could just temp turn it off >.> [13:37:14] Although I don't use it at all so it wouldn't bother me [13:37:53] well a few of the parosoid backends are down by the looks in -ops [13:38:11] they get restarted by puppet [13:38:22] that's been going on for a few days (at least) [13:39:20] well I've restarted parsoid on wpt1008 to see if it makes any difference (not even sure there's something wrong there but it won't hurt) [14:02:52] @notify petan [14:02:53] I doubt that anyone could have such a nick 'petan ' [14:02:59] @notify petan [14:02:59] This user is now online in #huggle. I'll let you know when they show some activity (talk, etc.) [14:03:06] hi [14:03:19] There you are [14:03:45] I'm curious, does wm-bot have a way to see what the current @config is for an item? [14:04:05] no... but it's a good idea to implement that [14:04:13] Need a ticket? [14:04:18] yes :) [14:04:23] Will do. [14:04:30] my memory usualy flushes often [14:05:23] @notify petan [14:05:23] I doubt that anyone could have such a nick 'petan ' [14:05:25] :p [14:05:32] that is already fixed [14:05:36] just not deployed [14:07:27] https://bugzilla.wikimedia.org/show_bug.cgi?id=49578 done [14:08:12] good that you did report my memory suck as well :P [14:08:57] I think I did... :p [14:09:09] I copied the conversation from here into the ticket.. [14:09:33] yep.. made the cut [14:58:59] hey, someone who is using visual editor or knows how to check [14:59:06] can they see f it is actually disabled now? [15:54:19] apergos: I know I am an hour late but it looks very disabled from where I am sat :) [16:12:46] excellent [16:42:24] marktraceur: heya, I put your fix for uploadwizard during the LD on Monday, cool? [16:42:39] Sounds great [17:24:34] apergos: wtp1008 is depooled currently, albeit not for a very good reason [17:24:56] well I understand chris is looking at a hardware issue [18:03:51] greg-g: Reedy: have you guys discussed this? https://bugzilla.wikimedia.org/show_bug.cgi?id=48506 [18:14:44] Is the syntax of a template on Wikimedia fixed? i.e. can I have |lat_deg = | lat_min = | lat_sec = |lon_deg = | lon_min = | lon_sec = AND |lon_deg = | lon_min = | lon_sec = |lat_deg = | lat_min = | lat_sec = ? [18:15:11] i.e. can the latitude and longitude be interchanged in the template? [18:15:49] robla: not explicitly since aude said it wasn't too urgent in the wikidata meeting yesterday (and since Reedy is working on the election email stuff today) [18:16:27] ok...we still have language links broken on 40,000 articles, which isn't a great place to be [18:18:27] ShiningThrough: yes they can [18:22:34] Nemo_bis: but they can also appear intermingled? [18:23:04] Nemo_bis: i.e. lon_deg = | lat_sec = | lon_min = ? [18:25:52] Nemo_bis: this http://www.regex101.com/r/cQ0rO4 seems to do the job :-) [18:31:22] apergos: We since found that the hardware issue also affects all of the other wtp* boxes, so it's not exactly a great reason to keep 1008 depooled :) [18:31:32] oh noes [18:31:37] what's the issue, do we know? [18:32:34] CPU power warnings [18:32:39] ok well [18:32:43] I dunno that this is hardware [18:32:43] OTOH there's a known kernel bug about those warnings being bogus in some cases [18:32:50] yeah I think I added that note about the bug [18:32:57] Right that was you [18:33:21] Well yeah the boxes are working well enough, so I'm keeping 1008 depooled to serve as Chris's guinea pig for when he gets around to poking at this [18:33:23] so it would be nice on one of those boxes to put a kernel on there with the fix and see how it behaves [18:33:33] he was moving cpus around earlier [18:33:41] and for now we'll just continue running things as usual on the other 23 backends [18:33:41] Oh good [18:38:14] if it's depooled I don't know how we are supposed to make it produce errors though [18:39:03] greg-g: Krenair was awesome and merged the change, so we'll have some sort of deployment to do in the nearish future, when would be good for that? [18:40:06] when it won't break things. [18:40:26] so... probably after lunch PDT, so there are people around [18:40:30] Mmkay. [18:40:32] * marktraceur orders food [18:40:39] where are you ordering from? [18:40:46] My house [18:40:54] apergos: True. Well there are 23 identical machines still pooled, and he can always ask for it to be repooled [18:41:07] ah well I see there are some syslog messages anyways after the reboot [18:41:11] marktraceur: well then [18:41:33] greg-g: It was a "stay at home and work on UploadWizard in pyjamas" kind of day. [18:41:37] it wasn't even running puppet at the time, wonder what it could have been doing [18:42:12] marktraceur: s'ok, it doesn't appear to be a real katurday today (unless it is only katurday on 6) [18:42:27] Ah [20:29:54] https://meta.wikimedia.org/w/index.php?diff=5570674 - don't the caching servers in esams do this for most requests from outside the US? [20:38:17] Reedy: Ping. People want to know where we're at with that election thing. [20:41:46] b [21:02:54] greg-g my darling, would ~15 minutes be an OK time to push this JavaScript change for UploadWizard? [21:03:44] marktraceur: my dear, 2:15 would be good [21:03:59] Wexcellent. [21:04:31] Hrm [21:04:36] There are some l10n updates in the middle [21:04:58] I guess I could cherry-pick or...just not scap? I'm not sure how terrible that is. [21:05:24] (third deploy, woo) [21:07:00] wee [21:07:18] greg-g: Opinions on whether cherry-picking things into wmf6 is a good idea, or if it would be better to include the l10n stuff and not sync it, or maybe the answer is I should just go ahead and scap because nobody else is around anyway? [21:09:52] nobody will be doing anything after you until Monday, so scap is fine if time is your concern [21:09:57] marktraceur: ^ [21:10:09] KK [21:10:13] right, everybody? no one will be doing anything until Monday, right? ;) [21:10:38] * marktraceur will do a final warning before running it [21:11:00] * greg-g nods [21:11:35] Effing git-review [21:13:27] yep [21:16:39] OK, merged...on to tin [21:20:07] OK, final notice: Nobody else is deploying today, because it's Friday and who else would be that insane. Right? Scapping in 2 minutes. [21:20:47] <^demon> I was thinking of upgrading gerrit on a friday at 5pm. [21:20:52] <^demon> Because that's always a good time. [21:21:36] ^demon: hush you [21:22:13] <^demon> Haven't even tested the build either. [21:22:18] <^demon> Let's just pick a build and pray :) [21:22:23] Right, eff you guys [21:24:11] "eff you, I'm scaping without you" [21:25:19] greg-g: Hey man, I warned y'all [21:25:28] And ^demon can deploy Gerrit without me, I think [21:26:16] no [21:30:41] No? [21:31:10] not today he can't [21:31:17] with or with you [21:31:25] * greg-g now has a song in his head [21:31:41] oh U2 [21:31:44] Ooh, I should get RATM on my laptop while I'm at home [21:31:52] indeed, not U2 [21:31:54] You keep bloody mentioning them, and I keep not having them on my laptop [21:32:39] <^demon> greg-g: You can't stop me! Bwahahahaa! [21:33:19] Just don't do it on Sunday, bloody Sunday. [21:33:39] sorry, these jokes are more appropriate in -staff [21:33:52] greg-g: These jokes are more appropriate on all the channels at once [21:33:53] brbs [21:45:05] greg-g: So I just deployed wmf6, is there an etiquette as to whether I should put the patch into wmf7 as well, and/or deploy wmf7? I just wanted to fix this on commons, so mission success, but I'd like to keep it fixed when we update commons...what, next week? [21:55:19] marktraceur: Normal behaviour is to backport, not prematurely bring forward, if it's core; if it's an extension... eh, Roan did wmf7->wmf6 backport for VE yesterday, but we had an excuse (A/B test ahead of wmf7 going out and we want consistent data). [21:56:01] marktraceur: I'd like it on test2 and beta commons. I'm still not sure why it didn't trigger a browser test failure. [21:57:03] Ah sure. [21:57:20] chrismcmahon: It only affects thumbnail loading, so maybe there's no test for that [21:57:44] marktraceur: probably. I'll revisit what we're actually checking in there RSN [21:57:53] checking for [21:58:20] James_F: I only wanted to fix it on commons, which is running wmf6, but it probably wouldn't be premature to bring forward a fix for Chrome users [21:58:46] marktraceur: You made the fix in wmf7 and wmf6? [21:59:11] Only in wmf6 so far, but I can quickly submit it for wm7 too [21:59:22] wmf7 is in the same place wmf6 was before I did this [22:01:26] And it's an extension. [22:01:36] (I guess that may not have been clear) [22:03:19] James_F: This would be why I'm asking, I did what I needed for the fix I was deploying but I'm not sure how far I should go for wmf7 [22:03:44] marktraceur: Put it into wmf7 but leave the deployment of wmf7 for Monday's automatic push. [22:03:55] Mmmkay. [22:04:26] (see, I wasn't sure if that would mess with someone else's deploy or what, now I know) [22:41:09] marktraceur: now you know, right? (ie: you all good? I just got back from my annual review 1:1 with rob) [22:41:29] also, I'm going to head out and be afk starting in about 25 minutes [22:41:33] Yup [22:41:45] greg-g: It seems to be working, I've merged into wmf7, etc. [22:43:22] rock [22:43:25] thanks [22:43:36] and you didn't break anything on a Friday, so I'm happy with you [23:57:08] gn8 folks