[00:52:04] jdlrobson, do you know why some tests are failing here? https://integration.wikimedia.org/ci/job/mwext-mw-selenium/435/console [00:52:32] bmansurov: check the phabricator board before asking any question like this. :) [00:52:42] main board? [00:53:48] sprint board [00:53:55] it's the highest priority task in code review [00:58:58] jdlrobson, hmm the phab task doesn't tell much about why tests started failing all of a sudden. Didn't they use to work earlier? [00:59:19] i have no idea. i've spent the entire day trying to work that out with Dan [00:59:22] it makes no sense [01:00:04] ok [01:04:27] bmansurov hoping the latest patchset does the job [01:11:19] bmansurov: finallllyy.. https://gerrit.wikimedia.org/r/232827 [01:11:34] good job, dude [01:27:36] ergg just a pain [01:27:39] anyway i'm calling it a day [09:22:08] hi! [11:02:08] hey joakino [11:02:20] felt pretty rotten at 10 am so i had a nap [11:02:25] feeling a wee bit better now [11:33:09] phuedx, thanks for the review. I didn't find in the requirements that the survey should be hidden when viewed. [11:33:21] phuedx, I thought the user may still want to read the privacy policy. [11:33:42] okie poke -- then we've got two survey implementations that do different things [11:33:48] i'll make a note on the ticket [11:33:56] phuedx, ok thanks [14:09:29] bearND|afk niedzielski-afk coreyfloyd ping :) [15:04:32] hi FlorianSW [15:04:44] hi bearND [15:06:26] venue change, brb [15:08:49] bearND: ahh, sorry, I pinged you :P I had a problem with building a production release for an apk based on the Wikipedia App. It always tried to load content without showing it, while a prodDebug built works fine. The debugger said, that marshaller.getPayload isn't a function. [15:09:13] I already found the problem, I forgot to adjust the proguard rules to the new package name :) [15:09:30] changing and rebuilding the app and all works fine for a prodRelease built now. [15:09:42] FlorianSW: ah, cool. Good! [15:14:10] FlorianSW: what's up? [15:14:33] oh, just seeing this message to bernd [15:15:00] niedzielski: the same as for bearND :) [15:15:24] FlorianSW: ok cool :) [15:15:34] niedzielski: thanks for replying :) [15:15:42] np! [15:15:51] and thanks to bearND, too! [15:31:08] trying to join the call [15:37:20] bmansurov: can you look at my comment on https://gerrit.wikimedia.org/r/#/c/232670/ ? I'm happy to merge after it :) [15:44:25] FlorianSW, replied [15:44:41] looking [16:03:13] jdlrobson: yt? [16:07:45] mdholloway: want to rejoin the mtg? [16:08:16] yep, sorry, one sec, forgot [16:14:42] joakino: do we have a ticket for disabling the lead image? [16:15:05] phuedx: not that I know of, i know there's one for pushing it to stable [16:15:06] an epic [16:15:40] joakino: cool -- i know i've mentioned it once or twice [16:15:56] but it's a beta feature that we're not really doing anything with and has /a lot/ of open bugs [16:16:29] phuedx: what kind of bugs? [16:16:39] minors around reflows [16:16:42] choosing the right images [16:16:53] then there's focal area stuff to figure out [16:16:53] bleh [16:16:54] w/e [16:17:04] all dangling off the "promote to stable" epic [16:17:22] i think it's a worthy change -- but as jdlrobson says, it's mostly done by this new fangled extension [16:17:23] phuedx: we're safe to ignore them then hehe [16:17:39] and using other people's hard work and taking the credit for it is my job [16:17:47] lol [16:18:12] phuedx: I'll talk to jdl and see how we can present this to prod & design and explain what's up [16:18:28] it may be worthy work for next [16:18:29] Q [16:18:39] phuedx: would you mind adding that to the brainstorming etherpad? [16:20:47] joakino: link [16:20:58] don't have it listed anywhere :/ [16:26:28] done [16:27:33] gonna go spend some time with the family and get the boys to bed [16:27:48] ^ professional [16:39:13] phuedx|afk: joakino it's also impacting our first paint since beta/stable are both mobile ;-) [16:39:50] what is? [16:43:08] banners :) [16:43:15] joakino: btw your design changes are epic [16:43:30] :D thanks jdlrobson [16:43:31] do you mind removing nearby quickly [16:43:39] since i'm not sure i'm gonna have that done [16:43:41] hit a few road blocks [16:43:52] then i'll merge away [16:44:06] today i'm gonna upstream some stuff to jscs-dev [16:44:09] and fix this up [16:45:23] also dunno if you saw my infinite random patch for Gather [16:45:29] since i was there i couldnt help myself... [16:47:35] lol no i havent [16:47:36] niedzielski: for the TWN sync patch: what's the manual formatting do? [16:48:33] bearND: i just pushed a revision to fix the space count. the diffs should make more sense now. the manual formatting just pretty prints so it's not all on one line [16:49:19] niedzielski: ah, cool. Would be nice to add to the commit message since the command is a bit cryptic [16:49:32] bearND: will do [16:58:04] bearND: for app_store_description, do you know why this eng translation was used: https://gerrit.wikimedia.org/r/#/c/232952/4/app/src/main/res/values-zh-rTW/strings.xml [16:58:53] niedzielski: no, that looks like a mistake [16:59:05] niedzielski: let's check the history on TWN [17:03:56] https://translatewiki.net/w/i.php?title=Wikimedia:Wikipedia-android-strings-app_store_description/zh-hant&action=history [17:05:04] niedzielski: since that is not actually used inside the app (only for the store description) it doesn't matter for this release. But we should still ask this translator why this was changed to English [17:06:19] bearND: sounds good. i'm almost done walking the diffs [17:07:02] niedzielski: good job on the formatting fix, btw. [17:07:32] bearND: thanks. diffs look sane to me with that one exception [17:07:56] bearND: the tests all pass with the exception noted in the commit message. any objections to merging? [17:10:05] bearND: i think the reason why the translation may have changed to english is because the last one was quite old [17:11:29] niedzielski: probably. I think in that case it would have been better to just delete the translation. [17:11:38] bearND: agreed but maybe that wasn't clear [17:11:53] i guess this is a case where i wish it had been... lost in translation [17:12:02] * niedzielski shows himself out [17:12:22] lol [17:14:04] jhobs: which config on the portal is for adding personal IPs. TEST1, right? [17:14:32] dr0ptp4kt: any of the tests ones are fine, but I believe TEST1 is what most people have been using, yes [17:14:38] jhobs: thanks! [17:18:02] dbrant: we're standing [17:27:03] mbinder: hey! can you hit me up with that meeting invite? [17:29:21] niedzielski: sent you the invite [17:30:12] bearND: thanks! [17:31:24] dr0ptp4kt: jhobs : yes, I was using TEST1 [17:35:39] bearND: is the meeting hangout link still valid? [17:36:28] niedzielski: hmm, maybe I should let mbinder let send it to you [17:37:13] niedzielski: https://plus.google.com/hangouts/_/wikimedia.org/sprint-planning?authuser=0&hceid=bWJpbmRlckB3aWtpbWVkaWEub3Jn.rngl8lb2dl3voo04j1h6ig8hic [17:37:18] does that work for you? [17:37:42] vibha: kaity anyone from design joining mobile apps sprint prioritization? [17:38:27] bearND: it's hanging on trying to join the call. maybe i have to be invited to the meeting (not just forwarded email)? [17:38:35] mbinder: ^ [17:38:43] I'll try to bump you and invite again [17:39:11] niedzielski: i've added you to the meeting. Maybe restart your browser [17:42:10] mbinder: vibha and kaity are trying to connect. connection issues abound [17:42:21] thx dr0ptp4kt [17:42:30] crossed connection signals [17:42:43] (i guess kaity was already on) [18:04:07] dbrant: bearND: niedzielski: ok, so dbrant's link preview tweaks are in, i looked at bearND's load strategy variant version patch, is there anything else that needs reviewing for the beta release? [18:04:46] mdholloway: i've also merge dbrant's other patch for the search button highlight [18:05:21] all set from my end [18:05:30] bearND: dbrant: awesome! [18:07:53] maybe I'll look at bearND's Retrofit and performance test patches, in that case [18:08:20] (is adding Retrofit a hygiene patch?) [18:10:00] mdholloway: The idea was that once it's complete (it's still WIP) it shouldn't add new functionality. Using the RESTBase endpoint will happen in a later patch. But I have no problems dropping that from the commit title [18:10:29] yeah, it's probably good to drop it [18:11:13] bearND, caching is under way as well: https://phabricator.wikimedia.org/T109742 [18:12:03] gwicke: awesome. Thank you! [18:12:08] dr0ptp4kt: ^ [18:12:42] gwicke: bearND: awesome! [18:26:13] mhurd: coreyfloyd bgerstle see you for estimation shortly [18:26:50] k [18:32:14] hey dbrant [18:32:25] I downloaded the latest [18:33:18] the line height looks still too small https://www.dropbox.com/s/pjobkuve0s22xkd/Screenshot_2015-08-21-11-21-57.png?dl=0 [18:34:31] kaity: what's the version number on that? ("More" -> "About") [18:35:17] alpha 2015-08-21 [18:35:58] dr0ptp4kt: Is Toby around today or did he just forget to nack my 1:1 with him? [18:36:18] found him! [18:36:42] kaity: weird! that's totally not the line height that I see. [18:37:31] dbrant: hmm should i uninstall and reinstall? [18:37:48] kaity: what version of Android is that? [18:38:17] dbrant: oh its old [18:38:31] 4.2.2 [18:41:25] kaity: huh! it seems to be specific to 4.2.2... [18:41:34] kaity: is there anything else odd? [19:00:14] bd808: he is around, not sure if he's gotten into his inbox very much [19:19:17] bearND|afk: gwicke thx for heads up about caching. is the determination about when to cache purge delegated to the implemented service? or are there some standardized patterns (e.g., if we know article x changed, then by definition the entire composite response is stale)? [19:20:56] joakino: phuedx|afk this does sum up my thoughts on the experiment https://twitter.com/bintuwrong/status/634806469603430400/photo/1 [19:21:18] whoops [19:21:25] i'm here but set as afk [19:21:29] yeah inorite [19:22:06] dr0ptp4kt: we are tracking several page related events, including edits and template updates [19:22:17] phuedx|afk: jdlrobson the approach used on the apps is to collapse that first infobox. there's also collapsing for subsequent infoboxes/tables (remember in the apps the articles are fully expanded) [19:22:50] dr0ptp4kt: that would be irresponsible on web since we still serve it and it's brittle [19:22:59] we are using those to pre-generate & actively update parsoid content currently, but can piggy-back purges and possibly pre-generation for the app service as well [19:23:01] gwicke: i forget is there a quick lookup for the service to look at the current revision versus its last served revision? [19:23:10] any sort of content changes need parsoid [19:23:43] i think if we want to collapse infoboxes we really need to work with gwicke to make that possible rather than resorting to js hackery [19:24:01] dr0ptp4kt: not sure I understand your question re current revision [19:24:51] jdlrobson: yeah, i think either doing a transform at the restbase tier or something deeper in the stack is more sustainable [19:25:07] the parsoid folks are also really keen to help you, btw [19:25:59] currently those purges are dumb, as in the entire page is re-rendered [19:26:03] gwicke: sorry for confusion. i can't remember, how would the mobile content service be able to know when it should purge? suppose it's if the article content has changed, for instance, that should be the signal for the mobile content service to know it needs to serve something fresh. [19:26:40] oh, RB / Varnish will take care of that [19:26:51] basically, you'll see a new request when needed [19:27:00] jdlrobson: lolol [19:27:07] gwicke: oh, so it's keyed off of domain name and article or something like that out of band? [19:27:15] gwicke: that is... [19:27:22] the api path has the title [19:27:26] varnish is told that an article on a wiki is stale [19:27:43] yeah [19:27:46] that's the purge [19:28:08] in the first phase, we might only do that, which means that the next access will be a cache miss [19:28:10] and thus the mobile content service only gets its code invoked if the last served response was stale. right? [19:28:34] in other words, the mobile content service would be relieved of the need to write code to figure out if content has become stale? [19:28:34] in the second phase, we might pre-generate the new content in RB and purge Varnish, so that the first access will just get the content from RB storage [19:28:48] correct [19:28:52] nice! [19:29:07] jdlrobson: phuedx that works fine in my app lol http://chimeces.com/webkipedia/#/wiki/Al_Franken [19:29:10] dr0ptp4kt: https://phabricator.wikimedia.org/T109742 [19:29:16] Who needs infoboxes [19:29:49] https://en.m.wikipedia.org/wiki/User:Jdlrobson/minerva.js < just bring back wikidata infoboxes :) [19:30:21] dr0ptp4kt: there's another possible third phase, which is replicating the pre-generated content to the caching datacenters, so that accesses to that data will always be served directly from the cache DC [19:31:00] maybe in Q4, if we find the time [19:31:49] gwicke: right, saw that task. yeah, replication could be really nice. [19:32:50] jdlrobson: i don't remember, did you write a phab task to remove the lead image? [19:33:12] phuedx: mmayybee lemme see [19:33:15] jdlrobson: that's awesome! [19:33:32] gwicke: needs a bit more work - need to cover more instanceOfs [19:33:48] but I think it's a lot better for mobile [19:34:06] i'm planning to get it working on desktop just to show parity between wikidata infoboxes and manually made ones [19:34:08] yeah, a lot more control over selection & layout [19:34:55] phuedx: nah there's just https://phabricator.wikimedia.org/T98145 [19:36:58] okie poke [19:37:34] jdlrobson: I do wonder if we can get away with doing wikidata infoboxes only, or if we need to support per-instance parameters too [19:38:30] gwicke: the wikidata inboxes use the wikidata instanceOf field [19:38:35] this gives them a template [19:39:10] Check out mw.config.values.wgMFInfoboxConfig [19:39:36] So the key '5' corresponds to https://www.wikidata.org/wiki/Q5 [19:39:52] and it defines rows that should be included - the idea being an admin can configure it to their needs [19:40:07] brb food time [19:40:11] that's pretty powerful [19:40:18] kk, ttyl! [19:44:41] joakino, jdlrobson: https://phabricator.wikimedia.org/T109889 [19:48:05] should that be linked to the other one phuedx ? [19:48:59] linked 'e [19:49:00] 'e [19:49:02] *em [19:52:16] okie poke [19:52:24] outty 5000 [19:52:28] g'night y'all [20:04:19] @dbrant|brb: i haven't been using it very long but i think i like the page animation removal on link previews [20:05:09] dbrant: letting just the preview paper slide over makes it feel less detached and "new tabby" [20:27:53] @dbrant: i think this is likely a data issue, but i have seen a couple cases where the link preview text had some wikitext-ish looking stuff in it. one i can't repro any longer but one can be seen on the main page by tapping oberdorf am hochegg [20:28:36] niedzielski: it's an issue in the TextExtracts extension :( not much we can do. [20:34:22] vibha: kaity: would one (or both) of you be able to jump on a hangout? [20:52:10] niedzielski: I think I've seen those, too, esp. on disambig pages [20:53:24] bearND dbrant: hm, i've only seen a couple since TNG. i'm not sure how worried we should be, especially if we can't do anything aobut it [21:05:19] niedzielski: probably not a big deal. If we find patterns of this behavior we might be able to do something about it [21:06:38] bearND: makes sense [21:07:26] bearND: by the bye, are we planning on this guy making beta? https://gerrit.wikimedia.org/r/#/c/231786/ i put some comments in there but it's looks pretty good to me [21:10:19] niedzielski: I thought we were but it's not a big deal. We could also skip it for now. I'd prefer to address your patch comments in a separate patch since this one is just a revert + manual rebase [21:10:46] bearND: ok, would you have any qualms about me merging that? [21:11:16] none [21:11:54] gr8! [21:12:04] dbrant: would you mind taking another pass on this guy? https://gerrit.wikimedia.org/r/#/c/233038/ [21:13:06] bearND: niedzielski: mdholloway|afk: can we have another huddle for a minute? [21:13:49] dbrant: i'm down. i think mdholloway|afk may have stepped out. it's getting late over there and he has a big trip planned [21:13:59] sure [21:31:48] niedzielski: thanks for th emerge [21:32:05] np, thanks for the review :) [21:35:36] niedzielski: hang on... what was the reason for removing layout_gravity="center_vertical"? [21:35:51] it's height is wrap_content [21:36:00] dbrant: ^ [21:36:34] oh nuts, that's layout_gravity not gravity [21:36:45] right [21:36:49] dbrant: niedzielski : opps, just merged it [21:37:06] niedzielski: bearND: i'll put it back in my patch ;) [21:37:14] dbrant bearND: sorry! thanks, dbrant [21:43:26] dbrant: hey, i don't mean to harp on this but the short link previews in the new patch are invisible when opened in portrait and rotated to landscape. i thought it was worth mentioning but not necessarily a blocker [21:45:03] niedzielski: hmm, yeah :( we'll have to live with it for this round. it's high-priority for sure, though. [21:47:05] i'm "watching" your use of link previews via eventlogging ;) [21:52:02] dbrant: :) [21:52:15] dbrant: hey, it looks like there's a small gap between the preview dialog and the screen edge on gb :/ [21:52:44] dbrant: and if you see the value of the link preview type change, that's because i used the dev settings [21:53:03] dbrant: let me get a screencap [21:55:40] dbrant: https://img.bi/#/tuIDZBA!fjdqTbjkabnGvdchaihtxbdcxMirercGfayBqdve [21:56:45] dbrant: it's a little worse in landscape it seems [21:57:21] niedzielski: that's a little weird... it must be a consequence of how API 10 does ListViews... [22:00:53] dbrant: that was on a note i. note ii which has the same form factor but kitkat looks fine [22:01:17] niedzielski: yeah, i seeit on my API 10 device, too [22:01:27] more GB woes [22:17:04] dbrant: still online? [22:17:14] kaity: yep [22:17:27] dbrant: i'm playing with link preview a lot [22:17:45] dbrant: I think an arrow by the title would be really useful [22:17:59] dbrant: what do you think? [22:19:52] kaity: even with the whole title being clickable? perhaps... [22:20:09] dbrant: it doesn't feel that clickable [22:20:18] dbrant: or we could try a blue link color [22:21:05] kaity: sure, we can do additional usability testing with that. [22:21:15] kaity: (the current Beta is locked in, though) [22:22:12] dbrant: ok cool [22:27:31] dbrant: is it possible to move the overflow button to the bottom? [22:27:38] dbrant: or is that too weird? [22:27:57] kaity: mmm.....pretty weird [22:28:19] dbrant: ok [22:28:40] kaity: i'd have to see a mockup :) [22:30:54] dbrant: https://www.dropbox.com/s/r5hsqi21lt3k9mr/link%20preview%20test-02.png?dl=0 [22:32:58] kaity: hmm, it's a bit too close to the system buttons... and I would want the overflow options to be accessible without having to scroll it fully into view. [22:33:34] dbrant: true [22:33:49] dbrant: but I think a clear way to the article is the most important [22:34:03] dbrant: just got to find a good place for everything [22:35:19] kaity: yes indeed... [22:36:03] @dbrant @bearND: any objections to my merging dbrant's final patch? [22:36:36] niedzielski: no [22:37:53] none! [22:38:46] bearND dbrant: sweet! merge in progress... [22:40:02] dbrant: you mentioned something like this https://www.dropbox.com/s/r5hsqi21lt3k9mr/link%20preview%20test-02.png?dl=0 [22:40:56] bearND dbrant: ok. merge is done. i'm going to start cutting [22:41:10] nice [22:41:52] 8< [23:05:14] dbrant: hey! whenever you get a moment, would you mind making the final wording on the release notes? [23:05:35] or the "rellie" notes, as i've come to know them [23:10:19] bearND dbrant: beta is looking pretty good on everything i have over here. any objections on moving forward with the release? [23:13:35] niedzielski: no objections! [23:13:48] niedzielski: all systems go [23:14:07] no objections you, honor [23:14:11] your honor [23:15:10] lol [23:24:14] dbrant bearND: is everyone happy with the release notes? it looks like they saw a few last minute updates [23:24:56] happy, pappy. [23:25:57] lgtm, ship it. \o/ [23:29:02] dbrant bearND : apk is in alpha now if you want to send a heads up to our feedback friends [23:29:20] niedzielski: thx! will send. [23:39:42] niedzielski: sent! [23:39:56] \o/ [23:39:57] bearND: niedzielski: awesome work today, gents [23:40:02] agreed! [23:40:12] hey, i'm working on the release mail but other than that we're done afaik