[00:03:44] niedzielski: got it. okay, i did the w0 stuff, i can re-run the non-w0 traffic tomorrow. but i gotta get going. have a good night! [00:04:09] dr0ptp4kt: thanks adam! have a good night! [00:14:54] rmoen: so what is wrong with your browser tests now and why didn't you ask? [00:15:00] let's pair first thing tomorrow and sort that out [00:15:14] jdlrobson: the tests pass [00:15:17] they don't [00:15:18] but they don't run locally [00:15:23] there's some issues with them locally [00:16:05] they are passing only by a false positive - always check the log https://integration.wikimedia.org/ci/job/mwext-mw-selenium/617/console [00:16:29] all your tests don't fail because they don't run :) [00:17:03] jdlrobson: weird. I see the issue [08:38:24] good morning! [08:38:43] hi phuedx|afk [08:48:35] hey joakino [09:52:43] phuedx, joakino howdy [09:56:27] hey bmansurov [10:31:40] going for a super early lunch [10:37:01] bb! [11:34:43] back [11:37:39] joakino: fancy doing some code review on that fix ;) :P [11:38:27] what fix [11:39:08] https://gerrit.wikimedia.org/r/#/c/234507/ [11:40:23] done [11:40:24] hehe [11:42:29] phuedx: more often than not i get a ton of module errors [11:42:32] https://www.irccloud.com/pastebin/gLYE0gsP/ [11:42:36] is it only me? [11:42:53] hrrrm [11:43:02] everything up to date? [11:43:41] i think so [11:43:52] phuedx: i can't see the survey even though the token is fine now [11:44:08] ext-quicksurvey-external example survey [11:44:26] it should be properly bucketing me with 100 in coverage, right? [11:45:12] joakino: that should be fine [11:45:16] test url? [11:45:48] wat? [11:45:52] http://localhost:8080/wiki/Cat [11:45:57] phuedx: ^ xD [11:46:13] https://2bbccf2a.ngrok.com [11:46:17] phuedx: ^ [11:47:09] joakino: you need quicksurvey=true in the query string -- everything's disabled without it [11:47:24] * phuedx notes down that that should be gotten rid of [11:50:30] lol [11:50:33] phuedx: nice to know [11:53:51] alright signed off [11:53:57] off to family meeting [11:56:38] joakino: hf! [15:18:00] dbrant: hey! i saw your morning mail from our friends. i loled hard when i saw that line about trashing history and even harder at your apples mail. i personally think a trash bin on the search page is a poor design choice, but i respect if you have a differing opinion on the matter [15:23:07] niedzielski: lol well, i do agree that we should follow the actual guidelines that advise against "destructive" actions in the FAB. [16:47:37] bearND: would you mind reviewing this guy? it's the rename of the language strings we had to undo because the twn diffs were big. https://gerrit.wikimedia.org/r/#/c/232215/ [17:01:40] dbrant: What on sprint 64 can be counted as done, and what will make it to 65? [17:02:31] dbrant & bearND: batcave? [17:03:11] k [17:07:41] phuedx|afk: let me know when you are back from keyboard [17:07:50] *to [17:14:55] mbinder_away: i think only the stuff that's currently in the Done column should be counted as done. [17:16:14] ok [17:19:52] phuedx|afk: (https://gerrit.wikimedia.org/r/#/c/233083/4/ needs to be merged asap in some form - but need to work out the caching problems) would appreciate your input [17:23:29] dbrant: are link previews on android on tap or long-press? [17:24:25] mhurd: on tap! [17:24:44] dbrant: how's beta feedback been? [17:32:21] dbrant: just curious - wondering if we get push-back from users if giving them an option to switch the default behavior to link-previews on long-press would make everyone happy [17:33:19] mhurd: it's not really conclusive yet... there isn't much feedback in our reviews that explicitly calls out link previews in a positive or negative way. [17:33:26] mhurd: but we've actually had users complain when we took away link previews (when we started the a/b test of 50% of people getting them). [17:33:47] dbrant: ah. interesting! [17:34:52] dbrant: ready to estimate? [17:39:10] jdlrobson: i shall give it a look when the boys are asleep [17:51:12] I want to work on bug suggested by jdlrobson Bug:https://phabricator.wikimedia.org/T98829 but I am confused with what exactly is rquired to do? [17:55:52] can anyone help? [18:00:55] edsanders, hi, i am taking off for the next week (burning man), can we chat briefly re VE maps? [18:11:03] thanks phuedx|afk ping me when you want to chat it over [18:24:54] mbinder dbrant: in case it comes up, i am fine with keeping with keeping the atomic number theme for sprint names. i think that would make next sprint Ter­bium [18:25:11] TURBOium [18:25:28] correct! [18:26:10] one of the least-creatively-named elements. [18:26:42] jdlrobson: so instead of loading the gallery styles for /everything/ you're only injecting 'em when a base module (page.gallery.something.or.other) is request? [18:27:00] :) [18:27:06] phuedx: hey so yeh it's complicated.. [18:27:29] so the core change that got merged adds a new module that only gets loaded on pages which use the gallery tag [18:27:49] this loads on mobile and desktop unlike the older legacy styles module [18:28:01] the change i pointed to gives skin specific styles for the gallries [18:28:16] to override the desktop ones [18:28:42] dr0ptp4kt: thanks for the wiki zero analysis! [18:28:42] 'course [18:28:50] jdlrobson: mind writing that in the commit message? ;) [18:28:52] niedzielski: sure thing [18:29:07] k let me update the commit message now. [18:30:15] phuedx: updated [18:30:23] bgerstle: coreyfloyd dr0ptp4kt vibha kaity etonkovidova iOS retroing now [18:30:27] jdlrobson: minerva.content.styles is loaded on all page views, right? [18:31:06] yup but since the gallery styles is removed from that no gallery styles will be removed on any old cached html [18:31:17] i can change the patch to double load though instead if we think that's better [18:31:43] (e.g. just don't remove the file from minerva.content.styles and add a fixme ) [18:32:45] jdlrobson: feels like we need a cache fixme! :D [18:33:08] phuedx: yeh so the only possible cache fix me is to double load [18:33:19] mhurd: https://etherpad.wikimedia.org/p/MobileAppRetrospective [18:33:21] so shall i go with that? [18:33:32] jdlrobson: yarrrp [18:33:36] add a fixme [18:34:01] i've also been thinking about tagging commit messages [18:34:12] with something like [CACHE] [18:34:18] so that we can grep the git logs easily [18:34:33] having said that, grepping files is easy too [18:34:35] might be overkill [18:34:58] phuedx: i'll create a bug [18:35:02] phuedx: resubmitted [18:36:28] jdlrobson: you clobbered your commit message update [18:36:36] phuedx: arrrhgh [18:38:32] phuedx: amended [18:47:29] joakino: you around? [18:47:38] for a bit yep [18:47:43] i'm looking at your renaming email [18:47:46] can you jump in tracy island quickly? [18:48:06] jdlrobson: they're named as quarter sprints to reuse the columns every quarter, but we can definitely add cues to the column names [18:48:25] jdlrobson: unfortunately we're doing numbers now hehe [18:48:32] yeh heh [18:48:49] but yeh would be great to quickly run through next sprint goals with you [18:51:11] joakino: that a no? [18:51:18] apparently you're competing against pizza now :) [18:51:32] ok i choose pizza :P [18:52:25] jdlrobson: i've renamed the sprint columns https://phabricator.wikimedia.org/tag/reading_web_planning/ [18:52:58] jdlrobson: i'm super packed with the strategy meetings next week :( unless we do super early (for you) on monday it's going to be complicated [18:53:04] i'm leaving in a bit now [18:56:04] joakino: no worries. can wing it [18:56:21] jdlrobson: wat [18:56:53] bgerstle: good morning. I found out Facebook wrote a wrapper around Xcode to simplify IOS apps testing ( https://github.com/facebook/xctool ) [18:56:58] wanna meet monday 8am or 8:30 to 9am? [18:57:13] jdlrobson: ^ [18:57:22] bgerstle: comes with some CI related section https://github.com/facebook/xctool#continuous-integration ! [18:57:25] hashar: sorry in retro then another meeting [18:57:37] bgerstle: just copy paste for later reading :} [18:57:38] joakino mayyyybeee. we'll see if i can mak eit [18:57:38] hashar: oh, we're already using xctool [18:57:51] hashar: not sure how it effects CI though, maybe they added a new feature [18:57:58] oki [18:58:18] jdlrobson: i'm pinging you in a few epics, i'd love to get some help putting some more work on board 55 [18:58:23] hashar: yeah, we're using it to run tests, through another (awesome) tool called fastlane: https://github.com/KrauseFx/fastlane [18:58:28] free( hashar_brain_xctool ); [18:58:51] hashar.brain.push(fastlane) [18:58:57] jdlrobson: is that ok? (sorryy) [18:59:13] bgerstle: seems there is enough tooling around Xcode! rather nice [18:59:33] hashar: some, the built-in tools are a bit rough though [18:59:47] and still requires OSX [19:01:16] (y) [19:01:28] cheers phuedx :) [19:04:51] mhurd: https://phabricator.wikimedia.org/tag/wikipedia-ios-app-product-backlog/ [19:05:04] mhurd: http://hatjitsu.wmflabs.org/05743 [19:06:58] mbinder: https://usercontent.irccloud-cdn.com/file/FWP99pEA/ [19:07:43] mhurd: http://hatjitsu.wmflabs.org/20273 [19:19:32] https://en.m.wikivoyage.org/wiki/United_States_of_America !! [19:19:33] banners are live! [19:20:47] bearND: would you mind double checking my twn sync? it's one message and my understanding is that i should mark it as ignored. https://translatewiki.net/wiki/Special:ManageMessageGroups [19:23:33] jdlrobson: fancy ! [19:23:48] thedj: just working on sending an announcement to wikitech [19:24:08] thedj: has image centering for mobile - so now for USA on mobile view you can see the statue of liberty :) [19:26:56] hmm, the text looks a bit meh on narrow screen [19:27:45] niedzielski: yes, ignore that one [19:27:57] bearND: thanks! [19:32:53] jdlrobson: and with working section headings :D Great, thanks for SWAT deploying! :) [19:39:43] gonna head off for the weekend folks [19:39:49] have a lovely weekend y'all [19:40:16] dbrant: have you created a task for the analytics stuff earlier? [19:46:29] bearND: yep - https://phabricator.wikimedia.org/T110702 [19:46:55] dbrant: thanks [19:49:28] jerkin are you really there? [19:49:34] joakino ^ [19:55:22] FlorianSW: yuppp thanks so much for the help [19:55:25] it's looking great [20:00:38] rmoen|lunch: take a look at my comments on https://gerrit.wikimedia.org/r/#/c/233763/ and let's wrap this up when you get back from lunch [20:27:22] rmoen|lunch: back yet? [20:27:40] jdlrobson: yes [20:27:43] working on it [20:27:48] awesome. trying to plan next sprint so want to work out where we are [20:28:01] shame baha's not around would love to finish up his patch. [20:28:11] i'm reworking the availabeSurveys bit [20:28:13] so its more readable [20:29:35] jdlrobson: hey :) [20:29:46] \o benestar long time no see [20:29:52] rmoen: you can do that in a follow up patch [20:29:54] when do you have time so that we can talk about the wikidata search thing? [20:29:59] if you just fix the sampling i'll +2 as is [20:30:06] jdlrobson: ok [20:30:10] benestar: now works if you want :) [20:30:19] sure [20:30:46] I've talked to the team today and got the following thoughts [20:31:14] In future, we want to use Cirrus search for Wikidata because our current search api doesn't work very well anymore [20:31:42] this will make it possible to search for more than just the id using the normal mediawiki search api [20:31:52] which is great, isn't it? [20:32:39] benestar: yeh that sounds great [20:32:48] so you would use the same api query on wikidata? [20:32:58] at least partly [20:33:06] moving to cirrus contains a few problems [20:33:31] first we cannot search by language but will go through all languages instead afaik [20:33:41] I'm not sure if that is an issue [20:34:09] i can imagine. But usually the hard things are good things to be working on :) [20:34:14] the bigger problem however is that currently the api as well as the MobileFrontend code knows only about pages [20:34:32] they don't know about descriptions/aliases and the fact that the label is not the page title [20:34:49] so what we also have to resolve is how search results get displayed then [20:35:26] we don't want to show the plain id for sure but rather the label and the description [20:35:28] jdlrobson: and yay, the tests pass. Going to do a followup nwo [20:35:30] now [20:35:38] benestar: could you use displayTitle in some way? [20:35:41] Perhaps also the aliases if they matched the search term [20:35:49] mobile search results use displayTitle rather than actual page title [20:36:05] how is displayTitle implemented in wikidata? [20:36:24] hmm, afaik mobile uses the title [20:36:55] kristenlans: can you meet? [20:37:00] https://github.com/wikimedia/mediawiki-extensions-MobileFrontend/blob/master/resources/mobile.search/SearchApi.js#L89 [20:37:33] I'm not sure if the display title functionality mediawiki provides works well for multilingual labels... [20:38:30] jdlrobson: the displayTitle property can only be changed in javascript so we'd still need custom MobileFrontend code in Wikibase. See my link above [20:40:28] benestar: make core mediawiki support that well :) [20:41:20] we do too much working around its inadequacies. It would be worth exploring having a way to have a stored multilingual title with a wikidata entry [20:41:39] otherwise this vicious cycle of working around problems.. stuff getting broken... fixing them happens [20:42:15] jdlrobson: but labels arent titles ;) [20:43:11] benestar: why not? :) [20:43:18] they aren't unique [20:43:42] mediawiki thinks of titles like of identifiers [20:44:04] what may work better is showing what actually matched the search term in the search results [20:44:31] laity sorry I'm here [20:44:37] kaity ^ [20:44:51] That way the user know why they found the item and they get the expression they look for [20:45:01] kristenlans: im in the hangout [20:45:10] so if a label matched, show the label, if an alias matched, show the alias etc. [20:45:56] benestar: i guess i'm just wondering why that cannot be revisited to make integration with other wiki extensions easier [20:46:11] right now it looks like wikidata is going to have to reinvent everything for mobile :( [20:46:30] what about flow? [20:46:44] it will also need lots of rework in mobile I guess [20:46:47] or needed [20:46:52] rmoen: can you update the commit message about why the i18n message is needed? [20:47:01] benestar: yeh flow suffered from the same problem [20:47:21] the issues begin when you start changing the content model [20:47:21] it's not as bad now - in that it works in mobile but it surfaces problems such as history view link etc.. [20:47:52] yup. but they should all share a common interface in an ideal world [20:48:21] mediawiki is far away from an ideal world though ;) [20:48:41] anyways, do you think it would be possible to provide an option in MobileFrontend's search to show a custom title [20:48:51] jdlrobson: done [20:48:54] rmoen: also you made it a string... [20:49:13] jdlrobson: i'm not sure what you are talking about [20:49:21] what did i make a string? [20:49:35] I already tried to just set the title to the label but then the search results also link to that title instead of the id [20:51:01] jdlrobson: could you please be specific ? [20:51:57] rmoen i don't know what this is for -> https://gerrit.wikimedia.org/r/#/c/233763/25/i18n/en.json, [20:52:11] https://gerrit.wikimedia.org/r/#/c/233763/25/tests/browser/LocalSettings.php < also that should be a number [20:52:54] jdlrobson: oh the number, is a string. oops. I thought a survey description i18n key would be fairly obvious why it is needed. but i updated the commit message like you asked [20:53:34] rmoen: i'm just not sure how it's related as you didn't add a new survey to LocalSettings.php [20:53:36] was it previoulsy undefined? [20:54:06] is it not needed because of line 52 on init.js ? [20:54:33] the survey view is trying to pass survey description to mw.msg. it previously did not exist [20:54:41] so it was a bug in master? [20:55:32] jdlrobson: yes seemed like a trivial fix, would you like it in a separate commit ? [20:55:53] The LocalSettings.php has "description" => "yo", [20:55:57] should that change to be that key? [20:58:11] jdlrobson: done [20:58:17] on it [20:59:09] rmoen: thanks all much clearer now what's going on [21:01:20] jdlrobson: do you think it would be possible to provide an option in MobileFrontend's search to show a custom title? [21:02:19] rmoen: great. i'm just moving the @integration tag to the top and dropping the @wip before merging as an fyi https://gerrit.wikimedia.org/r/#/c/233763/28..29/tests/browser/features/internal_survey.feature [21:02:19] the tags at the top apply to all the tests [21:02:20] ok [21:03:30] jdlrobson: yay thanks [21:03:59] benestar: if you can create an api that takes the same parameters and responds with a format similar to prefixsearch you have no problem [21:04:18] that would be what I'd do (and just change the query against it via a config variable) [21:05:09] jdlrobson: yes, I created that ;) [21:05:18] see https://gerrit.wikimedia.org/r/#/c/234560/ [21:05:32] so can't you respond with the displayTitle you want? [21:05:45] jdlrobson: nope, displaytitle is only set in javascript [21:05:49] the api only returns the title [21:06:08] we could change it to displayTitle: info.displatTitle || this._highlightSearchTerm( info.title, query ), [21:06:11] so the displayTitle used there is only for adding the html to underline the prefix [21:06:24] yes, that would work [21:06:42] that's what i'd suggest then. There's no reason to use js for the highlighting [21:06:58] but people would wonder where the info.displaytitle should come from because it's set by mediawiki [21:07:00] in fact core api should probably respond with that anyway [21:07:15] yes, we will have to add something like that when using cirrus [21:07:23] for now it should just in the way you described [21:07:24] thanks :) [21:07:33] perfect. yeh that sounds great to me [21:07:48] will you create a patch to do the above? [21:07:59] create a task if you need us to do anything for you [21:09:05] will provide a patch, yes [21:09:24] I hope we don't need to do more changes on MobileFrontend for now :) [21:52:43] jdlrobson: hey, you still in the office? [21:53:26] woah! kaldari! [21:53:33] :-) [21:53:45] heh, still lurking :) [22:03:12] kaldari: yup [22:03:16] kaldari: 3rd floor [22:03:30] too late, I already figured out what I was going to ask you about :) [22:30:49] kristenlans: replied to your offsite comment [22:35:29] thanks jdlrobson I'm trying to keep things in sync between the email thread w/travel and phab - the travel gang aren't on Phab :P I'll float those alternate space options by them, [22:35:52] i have a feeling for some unknown reason we avoid airbnbs [22:35:56] but i'd feel more comfortable [22:36:31] jdlrobson: yeah I think there's some weirdness about airbnb but I'll get the facts and see what's possible