[16:22:23] joakino: I'm living in a 25 foot trailer, which is parked on my property. I'm renting my house for the summer to make some ca$h for fall road trip. [16:22:46] trailer = caravan for my euro pals [16:23:06] * joakino googles 25 feet [16:23:11] haha kristenlans, awesome [16:24:05] joakino: I ahould blog about it, it's pretty crazy [16:24:06] kristenlans: do we have a team map with the members? [16:24:20] kristenlans: do it! [16:24:38] joakino: we should have a team map [16:25:13] kristenlans: how would we do that [16:25:30] joakino: I have an idea... [16:26:30] the land of liberty is so cool some times https://www.youtube.com/watch?v=kDMAmMy_OqU Outdoor Channel Dock Dog National Championship Series Eastern Regional Championship. [16:26:43] (appeared googling 25feet haha) [16:26:51] lol [16:26:58] kristenlans: i shall destroy you [16:27:12] "euro pals" [16:27:28] joakino: team map would be neato [16:28:57] joakino phuedxx my friend makes adventure maps. Maybe he would make us one: http://z6z.co/london2london [16:29:13] haha this one's funny http://z6z.co/runwithsean [16:29:57] kristenlans: you should road trip to see each team member [16:31:31] lol those maps [16:37:01] hmmm phuedx joakino maybe this? https://www.mapbox.com [16:43:43] bgerstle: coreyfloyd jdlrobson dbrant kaldari would you please join the meeting? [17:01:01] jdlrobson: we stand and deliver! [17:15:19] standup fer apps :) [17:28:19] kristenlans: mbinder: no sprint 60 board yet? [17:28:53] bearND: kristenlans held up by Phab access. kristenlans can you make it? [17:41:19] team map would be fun [17:41:49] gonna head out for the day [17:42:18] jdlrobson: i've fixed up the pagination patch, i'd understand if you want to wait for tests, i'll address them on monday, but review the changes if you can [17:42:45] and thanks for the browser tests work, looks like we're going to get some focused time to nail it down, that's going to be awesome [17:45:18] ciao joakino [17:46:59] joakino: no problemo [17:47:06] seeya joakino [17:47:14] i'll probably write pagination patches myself as a folllow up [17:49:34] bmansurov: wanna chat about the overlay stuff? [17:49:42] yes [17:49:50] So if you look at resources/mobile.startup/OverlayManager.js [17:50:03] you'll see there is an _om_hide event being emitted on every hide [17:50:08] (hide event) [17:50:18] and _hideOverlay called hide [17:50:37] so the reason the tests are failing is if you call overlay.hide() on something using the OverlayManager it calls hide twice [17:50:55] It's messy but this is why the browser tests are failing i suspect :) [17:51:13] so we probably want some option like usingOverlayManager: true [17:51:24] i'm not sure what the best architecture is there [17:51:31] jdlrobson: i see. but does removing window.history.back() make sense? i don't see a reason to keep it. [17:51:34] i try to avoid overlay hiding/overlay manager as much as possible [17:51:47] bmansurov: not really because you don't know where they are coming back from [17:51:58] for instance say there was a link in the talk overlay to the editor [17:52:19] personally i wouldn't touch the default behaviour - you'll cause yourself more trouble than it's worth. [17:52:36] ok, make sense. I think i'll do what ImageOverlay does [17:52:46] i must confess i haven't even looked at the bug you are trying to fix [17:52:56] it would be a good idea to get a sense of urgency since i don't think it went triaged [17:53:01] it's been reported before but only fixed with the image overlay [17:53:18] it doesn't seem as important to fix as the templating stuff for instance [17:53:39] when I picked it up it was at the top of the pile [17:54:24] yeh i can imagine.. moving columns adds a card to top. [17:54:42] but Barry is not giving false alarms [17:54:50] he passes without your patch but doesn't with it. [17:57:46] tgr|away: there? https://en.m.wikipedia.org/wiki/Special:Gather/all/active is showing lots of min items as 0. Do we need to run a script there or is this related to the bug? [18:04:07] mhurd: http://hatjitsu.wmflabs.org/25443 [18:06:43] mhurd: https://phabricator.wikimedia.org/tag/wikipedia-ios-app/board/ [18:10:37] coreyfloyd: http://hatjitsu.wmflabs.org/25443 [18:12:46] jdlrobson: there? [18:12:53] hey codezee :D [18:14:36] the test build here https://gerrit.wikimedia.org/r/#/c/216735/ failed and the log says "Cannot open file..." [18:15:13] jdlrobson: mysql> select gl_id, tracked, actual from (select gl_id, gl_item_count tracked, (select count(*) from gather_list_item where gli_gl_id = gl_id) actual from gather_list)t where actual != tracked; [18:15:17] Empty set (0.01 sec) [18:15:28] seems to be correct [18:16:00] also I only see lists with 4+ items when I open that link [18:16:06] codezee: yeh this happens sometimes [18:16:09] do you have a specific example? [18:16:15] codezee: if you reply with "recheck" it will force it to run again. [18:16:25] looking at the report i think the problem is just jenkins going down - not you [18:16:44] codezee: i tried again [18:17:25] tgr: interestinglly it seems i get different results when anon [18:17:31] tgr: looks like it only happens when you are logged in [18:17:43] tgr: oh no wait [18:17:44] not true [18:17:50] 'Game' shows up [18:17:52] that has 1 item [18:17:55] tgr: can you see it? [18:18:03] (on https://en.m.wikipedia.org/wiki/Special:Gather/all/active) [18:18:42] same result... [18:20:39] jdlrobson: uh, sorry, I was looking at beta [18:20:56] the transaction patch is not deployed in production yet, right? [18:22:17] on beta, the bug seems fixed - the selenium tests used to trigger it and looks like they don't any more [18:23:02] at a wild guess, when you add and remove items at the same time, the frontend sends two simultaneous API requests, and that caused a race condition somehow [18:23:48] anyhow, I can run updateCounts.php to fix this temporarily, but it will have to be re-run after the patch has been deployed [18:25:51] jdlrobson: oh, actually it is deployed - https://git.wikimedia.org/history/mediawiki%2Fextensions%2FGather.git/66ca5ddc95db61365ae1c9e14630fc76a5a9f680 [18:26:40] I'll just re-run the script then and hopefully that's the end of the story [18:31:36] sounds good :D [18:31:43] tgr: let me know when i need to check [18:31:47] codezee: mmm [18:32:14] codezee: taking a closer look [18:32:24] jdlrobson: you can check now [18:32:26] codezee: looks like ebernhardson beat me to it :) [18:32:28] jdlrobson: I've got the feedback, I'll update the patch [18:32:50] codezee: awesome. [18:33:01] the more interesting question is, will it be still correct when you check a few weeks from now? [18:33:20] although based on the selenium user on beta, I would say the answer is yes [18:33:36] jdlrobson: do you have shell access to production? [18:33:40] tgr: nope [18:33:46] probably a good thing ;-) [18:34:59] just remind me in a week or two then to re-check the counts to be sure they are correct [18:35:07] for now, I am assuming this fixed [18:38:29] tgr: looks good on enwiki :) [18:39:19] jdlrobson: well, updateCounts.php recalculates everything so that's to be expected [18:39:27] we'll see if it stays good [18:45:04] brb going to have some lunch. [18:45:24] Krinkle: it looks like the canonical url stuff is preventing users landing on mobile site in their mobile browser [18:45:37] either that or google are doing bad things [18:45:56] but a search on mobile is linking to http://en.wikipedia.org/?title=San Francisco [18:46:16] jdlrobson: But for articles indexed as /wiki/ it links to .m ? [18:47:11] it might just be google doing bad things [18:47:18] Fascinating. https://en.wikipedia.org/?title=San%20Francisco redirects to /wiki [18:47:21] to normalise the encoding [18:47:35] but title=San_Francisco doesn't redirect [18:47:48] if you have %20 + any query parameters it does not redirect (as it shouldn't) [18:47:59] this means it may be easier to fix then we think. There's existing logic for it! [18:48:22] Krinkle: on google searching for san francisco takes me to desktop [18:48:27] searching for santa barbara mobile [18:48:36] Yeah [18:48:48] Something is feeding google those urls [18:48:54] I see them more and more often lately [18:49:03] view-source:https://en.wikipedia.org/?title=San_Francisco [18:49:11] canonical url is wrong [18:49:14] Yeah, I know. I filed many bugs about that [18:49:20] seems without the _ it doesn't redirect [18:49:27] Yeah [18:49:29] * with the _ [18:49:39] Canonical url is obsolete problem if there is a 301 redirect, naturally [18:50:11] So we should look into the possibility of changing the redirect logic to look not only at the encoding but also the incoming url itself. [18:50:15] MaxSem for reference https://phabricator.wikimedia.org/T67402 [18:50:27] We can preserve the logic for not redirecting if there are query parameters (just for back-compat) [18:50:37] but it'd be good to normalise this whole class of urls [18:51:24] the regex for redirecting to mobile is horrible.. :-/ but you are talking about the redirect behaviour on wikipedia? [18:51:34] is there a task open for this Krinkle or do you think it's something we can hot fix today? [18:51:34] Indeed [18:51:50] jdlrobson: Wait wait, there is redirect logic in Mobile and that is dependant on the url? [18:51:57] In what case would it not redirect? And why? [18:53:39] Krinkle: MaxSem can point you at the link but https://en.wikipedia.org/?title=San_Francisco doesn't redirect to mobile on a mobile browser [18:54:29] OK. [18:54:36] I'm curious why that is. [18:54:58] Are there any pages inside index.php that we don't want to redirect? [18:55:29] MaxSem says it's because of feature parity. He's gonna grab us a link to the puppet code that handles the redirect [18:56:14] moving the discussion to #-ops [18:58:43] kaldari: around? [18:59:29] dr0ptp4kt: kaldari may still be on vacation [19:05:56] mdholloway|afk bearND|afk: sorry, it wasn't clear from the comments but is a formatversion=2 change needed in FullSearchArticlesTask? [19:06:04] comments here: https://gerrit.wikimedia.org/r/#/c/218983/4/wikipedia/src/main/java/org/wikipedia/search/FullSearchArticlesTask.java [19:12:11] hey coreyfloyd [19:12:13] do you have time to chat? [19:20:11] FlorianSW: I'm sad to hear that development for Wiktionary was ceased. [19:21:32] SuppliedRelic: yeah :( But the limited resources have to priorized in some way :( [19:22:54] FlorianSW: I can understand that. It would have been great if the foundation could have supported both the encyclopedia and dictionary since they complement each other. [19:23:05] well, it wasn't a WMF project at all [19:23:33] SuppliedRelic: MaxSem is right, and you're welcome to take over the project and moving it forward :) [19:23:33] just a few GSoC students whipped up an app [19:25:41] That is a good point. It wasn't actually developed by the foundation but if they had the resources then they could create their own application. I'm not a developer myself but I'll look into finding somebody who could possibly take over. [19:25:43] right -- i'm going for the weekend [19:25:57] i'll see y'all on tuesday [19:26:06] phuedx: have fun! [19:26:37] jdlrobson: I responded about gerrit stream [19:27:30] \o YuviPanda :) [19:27:42] YuviPanda: on the bug? [19:27:48] have't got to that email yet :) [19:27:49] jdlrobson: Ah, I see you want jsduck inside 'npm test' [19:27:53] I'll add it back in there [19:27:53] Yes [19:28:00] Krinkle: yup you mind reader you :) [19:28:01] Most projects have a separate Jenkins job for that [19:28:08] Ok [19:28:20] trying to maintain documentation quality [19:28:23] jscs does some of it [19:31:22] * jdlrobson breaks for lunch [19:34:01] dbrant: T100173 Articles with no Wikidata description show an erroneous description upon refreshing / going back. - How I can find articles that have no Wikidata description? [19:42:36] back [20:09:35] @mdholloway|afk pinged you on this ticket https://phabricator.wikimedia.org/T101436 [20:09:56] Kindly get back to me when you have a chance [20:13:03] Krinkle: all merged [20:13:12] Krinkle: does this mean we can publish docs now?! [20:15:11] jdlrobson: Yep :) [20:16:53] I'm writing the patch [20:17:04] I'll templatetise it for future use in other extensions [20:18:05] * jdlrobson hugs Krinkle :) [20:23:05] :) [20:23:52] (no job-template actually, as whether a repo has jsduck, doxygen or both depends). It's only a few lines boilerplate. Perhaps we can make a macro at some point, but will defer to later. For now just copying what VisualEditor has [20:28:25] niedzielski: Yes, we need more tasks with formatversion=2 since PageProperties(JSONObject) constructor is used all over the place. /cc:mdholloway|afk [20:29:08] bearND: can / should this be in a new patch? is morelike dependent on formatversion=2? [20:31:15] niedzielski: i don't know the answer to your latter question for sure. I doubt it is, though. To answer your first question: no, this should be in the same patch that mdholloway|afk is writing, so we have a consistent usage of formatversion=2 [20:31:30] https://www.irccloud.com/pastebin/2ZhKYKyQ/ [20:31:35] tgr: ^ [20:32:27] bearND: ok thanks [20:37:11] niedzielski: yeah, i meant to write this for the formatversion=2 patch mdholloway|afk is working. [20:37:37] bearND: sounds good, thanks! [20:40:26] ok brb, getting on the irc cloud for the first time [20:43:21] bearND: dbrant - I am using apk provided in the email - loading/saving long pages(2015 9n sport) takes a really long time...and crashes [20:43:50] YuviPanda: look like we've got an alpha build again. Did you do anything? [20:44:02] mostly die from exhaustion :) [20:44:05] it's been a long 48h [20:44:12] and yes they should have been back a few hours ago [20:44:15] sorry didn't notify you [20:44:36] bearND: btw, iOS is fine with loading/saving long articles [20:44:50] YuviPanda: np. Thanks you for bringing it back online. :) [20:45:44] etonkovidova: you can now use the regular http://android-builds.wmflabs.org/. But that other apk shouldn't have the issues you mentioned [20:46:00] bearND: understood. [20:49:49] bmansurov: cannot replicate "Not sure if it's part of this patch, but while logged out I visited http://127.0.0.1:8080/wiki/Special:Gather/all/recent and clicked on 'My collections'. I was taken to the login page. Once logged in I was taken to the 'Change password' page. I should have been taken back to the gather page." [20:49:54] kaity: I'm around now if you want to talk early [20:49:55] can you give me more information? [20:50:12] jdlrobson: It was the desktop site [20:50:25] I have a 2-2:30 [20:50:39] np [20:50:54] we can do the time you scheduled, just figured id ask [20:50:57] kaity: ^ [20:51:06] bmansurov: works for me in desktop and mobile.. [20:51:08] coreyfloyd: thanks! [20:51:19] jdlrobson: ok let me try again [20:51:22] the clear issue is a problem in MobileFrontend - the content header class isn't compatible on desktop [20:51:26] not quite sure what to do about that [20:51:50] (mobile watchlist would have same issue if surfaced on desktop) [20:53:20] jdlrobson: the redirect happened again. I followed the steps that I gave. Do you want me to create a screencast? [20:54:23] bmansurov: no i'm clear on the steps.. i just can't replicate it [20:54:34] what url is the login page? [20:54:39] I'm using Chrome Version 43.0.2357.124 (64-bit) if that helps [20:54:47] (that the tab sends you to) [20:54:53] Hello all - query. On the mobile apps and the beta version of the mobile site, there are subtitle legends visible under the article title. How does one edit these? [20:55:05] jdlrobson: http://127.0.0.1:8080/w/index.php?title=Special:UserLogin&returnto=Special%3AGather&warning=gather-anon-view-lists [20:55:05] For example: [[en:Emergency medicine]] has this mobile subtitle: "Medical specialty in which physicians care for patients who require immediate medical attention". No such text to be found within the article itself [20:55:46] jdlrobson: oh i see, this happens when your password is not secure [20:55:49] jdlrobson: http://127.0.0.1:8080/w/index.php?title=Special:UserLogin&action=submitlogin&type=login&returnto=Special:Gather [20:55:58] Your password is not valid: Passwords must be at least 8 characters. [20:56:01] jdlrobson: ^ [20:56:06] jdlrobson: just ignore it then [20:56:13] oh yeh that's a new thing for core [20:56:17] fix for tabs on way [20:57:52] jdlrobson: whenever you get time before leaving, have a look at the icons patch https://gerrit.wikimedia.org/r/#/c/215029/7 so that I can add the remaining options after it [21:00:30] codezee: will do! [21:19:30] dbrant good news on link preview--group 0 starts well before the other groups, so when I limit group 0 to just this past week and compare it to group 2 for this past week, group 2 outperforms! [21:20:27] dbrant this is an important reminder of: A. the importance of RCT over simply measuring before/after, B. how much data can fluctuate from week to week. [21:22:56] jdlrobson: do you know if that index existed before? [21:23:25] it does exist in production, and it's part of gather_list.sql [21:23:44] but maybe the update script tries to delete it twice [21:24:04] I can test on a fresh box if you think it's not a local problem [21:28:56] So in regards to CollectionsContentOverlay are we tied to invoking the CollectionEditOverlay via a route? [21:29:11] jdlrobson: ^ [21:29:28] hey sorry - lemme catch up [21:29:49] tgr: i think it's a fresh box.. [21:30:00] rmoen: yes i think so [21:30:11] sorry i've been wrapping up some old patches [21:30:31] rmoen: assume this is https://phabricator.wikimedia.org/T99109 ? [21:30:41] indeed [21:30:41] JonKatz: yay, excellent! very true. [21:31:10] i realize we load the editor on demand which i like because you wont need it when toggling a collection. [21:31:37] but i'm finding it difficult to construct the editor overlay with a tutorial flag when its created by the route [21:31:48] i can't remember the inner details of this patch too much but i'd suggest you CollectionsContentOverlay and CollectionEditOverlay would be invoking some kind of CollectionsPointerOverlay [21:32:04] the tutorial flag could be passed in on creation [21:32:13] sorry let me rephrase [21:32:22] you could change the route to have an extra parameter [21:32:26] right the route is the thing blocking me from passing [21:32:28] ah i see [21:32:28] e.g. #/collection-editor/1/tutorial [21:32:33] kristenlans: Android Board is clear. [21:32:39] as long as it's backwards compatible that's not a big deal [21:32:40] i was thinking that but wanted to see what you thought first [21:32:43] or even collection-editor-onboarding/1 [21:32:54] whatever you think makes most sense [21:32:57] cool [21:33:07] coreyfloyd: hey you there? [21:33:51] @bgerstle_afk @kaity Can we see a screenshot with the updated icon? https://phabricator.wikimedia.org/T97780 [21:36:19] Sorry, out for the day vibha. You can follow up with coreyfloyd or we can sync on Monday [21:36:50] ok, thanks bgerstle_afk [21:43:25] hashar: hey there! hey i'm still seeing that "sniedzielski is missing the Job/ExtendedRead permission" error. i have other things i can work on as part of this card but thought i'd ping you on irc in case it was an easy fix [21:47:58] niedzielski: :( [21:48:47] hashar: i added a note on the card so don't drop everything to get it sorted. thanks for the support! [21:49:05] niedzielski: i have no idea what that permission is :( [21:49:12] niedzielski: is that when reading the configuration? [21:50:07] hashar: that's what i get following this link in the thread: https://integration.wikimedia.org/ci/job/test-T62720-android-emulator/configure [21:50:32] hashar: which i believe is just the configure link (i can't see it when i go to the job since i don't have the permission) [21:51:24] niedzielski: is that your labs account ? https://integration.wikimedia.org/ci/user/Sniedzielski/ Sniedzielski [21:51:49] and you are a WMF staff [21:52:14] so your labs account should be in the ldap groups 'wmf' and 'nda' , the 'wmf' group would grant you full admin rights on jenkins :} [21:53:45] vibha: I'm happy to upload a screenshot but it's the same progress bar animation as while waiting for a page to load [21:53:52] hashar: i think so. it's what i use to login to wikitech.wikimedia.org [21:54:05] niedzielski: so you want to be in the wmf ldap group, I left a note [21:54:30] niedzielski: I also edited/saved your Jenkins account, that might have populated some missing field and unlocked it (unlikely) [21:54:51] niedzielski: so in short, fill a task for ops and you will have admin rights. I have left some details at https://phabricator.wikimedia.org/T62720#1384115 [21:55:45] hashar: nuts, ok. these account permissions are tricky to get right. thanks for the help! [21:56:23] If you can figure out what Chad Horohoe is called on IRC today, he's usually the easiest way to get into the 'wmf' group [21:56:39] he's the only active LDAP admin outside of ops, iirc [21:57:18] Looks like he's still "ostriches" [21:59:47] niedzielski: we should also give you some shell access on jenkins slaves. That is the integration project. Fill a task for #continuous-integration-infra as well :D [22:02:22] kristenlans: Could you kindly move this to Sprint 60? https://phabricator.wikimedia.org/T101782 [22:04:16] Krenair: thanks! i'll cc him on the ticket [22:04:27] vibha: https://usercontent.irccloud-cdn.com/file/PuVaHS23/Screen%20Shot%202015-06-19%20at%206.04.04%20PM.png [22:04:29] hashar: will do, thanks again! [22:04:59] niedzielski: feel free to hangout in #wikimedia-releng , rest of team should be able to assist :-} [22:05:12] vibha: hmm, i think we actually wanted to reorder the icons didn't we? [22:05:20] hashar: ok, will join. thanks [22:06:21] niedzielski: also you might want to update your Phabricator profile https://phabricator.wikimedia.org/p/Niedzielski/ :D [22:06:48] https://phabricator.wikimedia.org/p/hashar/ !! [22:07:37] hashar: ha! i like the "so you like thumb ups and clicked on my profile... here's DOUBLE the thumbs up!" approach you took [22:08:35] hashar: actually i was hoping to keep my profiles consolidated so folks can click on my mediawiki username (linked) and get there [22:09:06] coreyfloyd: oh i re-ordered them on the 5.0 ticket: < > heart lang share [22:09:23] mhurd: ha! [22:09:38] @Fjalapeno Design comments from the card say: Evenly space them out [22:09:44] coreyfloyd: and they're all equally spaced cause we get that for free [22:09:48] vibha: k - we will fix [22:09:49] lol [22:10:03] vibha: ^ [22:10:07] :) [22:10:17] vibha: is done [22:10:47] The order for the icons is: Left to right: Back, Forward, Language, Save, Share [22:11:03] Language goes to the left of Heart/Save [22:11:19] https://usercontent.irccloud-cdn.com/file/WxZC9Mi0/Screen%20Shot%202015-06-19%20at%203.11.02%20PM.png [22:11:29] vibha: can we have the heart in the middle? ^ [22:11:50] vibha: it's symmetrical, nice to center [22:11:58] *heard is symmetrical* [22:12:02] *heart* lol [22:12:22] vibha: no worry either way [22:13:01] vibha: did you have new svg for the lang icon? [22:13:23] It does feel like the center and the corners are easier to get to and should be reserved for frequent actions. Likely save is more frequent than language? [22:13:41] Let me discuss with kaity|away and get back to you guys on the ordering [22:13:58] vibha: super easy to flip it however you want :) [22:14:35] yes, we have new svg for language icon. attached to phab - https://phabricator.wikimedia.org/T97780 [22:14:47] bearND: niedzielski: about formatversion=2, the thing is, unless i'm missing something the meaningful booleans are only implemented for mobileview [22:14:53] Corey - does this mock have the updated icon? [22:15:03] Corey/ Monte ^ [22:16:02] bearND: niedzielski: if you make the query for read more pages in your browser and get a result that's a mainpage or disambiguation page, you'll see that it's still using an empty string for "true" [22:16:21] bearND: niedzielski: even if you tack on the "formatversion=2" parameter [22:16:58] bearND: niedzielski: i have an example in one of these tabs... [22:17:07] mdholloway: oh is this about the comment bearND left in code review? [22:17:46] niedzielski: i saw some discussion about it come through here earlier but i'll look there too [22:17:57] vibha: gotcha. note that that ticket dispensed (for some reason) w the canonical lang names (useful) and added titles (redundant as you're already on that page)... not sure why we'd swap out something useful for something redundant... [22:18:06] mdholloway: can you provide an example URL? [22:18:22] niedzielski: i guess for purposes of your patch the answer was that it should be handled in mine, right? [22:19:10] bearND: sure, one sec [22:19:25] mdholloway: yes, all format version changes should be in your patch [22:20:40] bearND: https://en.wikipedia.org/w/api.php?action=query&prop=pageterms|pageprops&ppprop=mainpage|disambiguation&wbptterms=description&generator=search&gsrsearch=Main_Page&gsrnamespace=0&gsrwhat=text&gsrinfo=&gsrprop=redirecttitle&gsrlimit=5&formatversion=2 [22:21:55] bearND: and that's not surprising since the patch to implement the meaningful boolean values seems only to handle it in mobileview: https://gerrit.wikimedia.org/r/#/c/214636/3/includes/api/ApiMobileView.php [22:22:21] (and not to handle disambiguation pages...) [22:25:01] mdholloway: hmmm, that almost sounds like a blocker for your patch. If some requests that have PageProperties behave one way and other behave differently then this doesn't seem like a step for us to make yet. [22:26:10] bearND: i don't think merging would hurt anything but i was thinking it does seem a little premature. [22:28:48] mdholloway: Once you tap share, do you see a card right away and license fades in or do you wait to see the card (on slow connections for share a fact) [22:29:29] vibha: you wait to see the card. want me to send you an apk to try it out? [22:30:13] vibha: more precisely, you wait to see the whole share dialog [22:30:25] bearND: hey do you remember how to trigger a captcha warning on edit? [22:30:33] mdholloway: I actually think it would make it worse. With your patch, you create PageProperties from mobileview parse the two values as booleans; whereas coming from FullTextSearchTask you'd still have the old behavior and needed to parse those same two values with json.has("mainpage") and json.has("disambiguation"). [22:30:41] bearND: testing some editing code... [22:31:04] bearND: yeah, you're right. [22:31:45] bearND: i mean, they're handled in different spots but it would get messy. [22:32:23] mhurd: sorry, i don't remember [22:32:59] bearND: hmm... [22:33:53] mdholloway: yeah, my understanding of this card is to make the Android code simpler. If it can't accomplish that then we should postpone it until we can do this consistently. [22:34:24] mdholloway: It would be good to raise this issue with the API team, and also add to the respective Phab task [22:34:53] dr0ptp4kt: would you recall how we can trigger a captcha warning on edit? [22:34:57] bearND: will do. [22:35:30] dr0ptp4kt: bearND: editing test wiki i thought it would trigger if we tried to enter a url, but it doesn't seem to anymore... [22:39:17] bearND: whoa plot twist! action edit is now saying it doesn't recognize the "captchaid" and "captchaword" parameters!!!! we may have an issue... [22:40:03] bearND: or maybe it's a logic bug on iOS... investigating further... [22:41:24] mdholloway: thank you. I've added some of that discussion on the Phab task. Feel free to add what's missing here, too. [22:43:44] Deskana|Away: did capchas on edit go away? i don't see "captchaid" and "captchaword" on action edit... [22:45:23] bearND: cool, thanks, will do. [22:46:06] bearND: ah check it out: https://www.mediawiki.org/wiki/API:Edit#CAPTCHAs_and_extension_errors < seems to say captchaid and captchaword are gone after MW 1.18 [22:46:27] bearND: i think that means both apps can safely remove the captcha logic from editing... [22:46:31] coreyfloyd can you paste that image in gchat [22:46:34] bearND: double checking... [22:46:40] I don't know what Irc did [22:46:44] We can't see it anymore [22:49:32] mhurd: I was able to add a URL to a testwiki page from the Android app. No captcha needed. [22:49:58] bearND: quick bat cave? [22:50:41] mhurd: sorry, can't right now. Maybe later or Monday? [22:51:54] bearND: k. the reason you are able to save a link is that the logic for knowing if a captcha was required has moved - formerly the app *would* show the captcha interface. need to make sure this is deliberate [22:52:24] bearND: here's the commit that removed those 2 parameters from the edit api response: https://phabricator.wikimedia.org/rSVN104064 [22:53:30] bearND: if it's ok we're good and we can remove the edit captcha handling, if we should be checking some other param now we'll need to switch over [22:54:34] mhurd: not sure how to make it captcha [22:55:20] dr0ptp4kt: no worry - the api actually changed. just need to find out now if action = edit *ever* requires captcha now that it no longer has those captcha params [22:56:16] mhurd: maybe email the api list and someone will have a quick idea there [22:56:37] mhurd: at least on monday... [22:56:44] mhurd: ...they would probably respond then [22:57:00] dr0ptp4kt: ya doing that now [23:01:58] vibha: emailed you an APK for the image license/share-a-fact patch [23:20:03] thank you. [23:51:35] marxarelli: care to look at https://gerrit.wikimedia.org/r/219501 ? [23:51:44] smoke tests failing :-/ [23:51:51] trust my patch to break them! [23:51:56] sure thing [23:52:45] jdlrobson: you want me to verify it against beta? [23:53:42] marxarelli: if you want. you can blind merge it if you want [23:53:46] i've got Barry testing it [23:54:05] he's +1ed it :) [23:54:41] haha, nice [23:54:44] go Barry [23:55:56] durn Barry, taking our jarbs