[00:07:19] lzia, atgomez1: mobile survey's up [00:09:45] yay!! [00:09:52] thank you!! [00:10:01] jhobs: also i just made this, fyi: https://phabricator.wikimedia.org/T119152 [00:11:18] atgomez1: yeah, I just saw that. I'll bring it up during sprint kickoff next week [00:12:22] thank you! it would be great to know at least if it's intentional or not, because leila may try to run a multiple choice on mon-wed next week if we can do it. if it's broken, we should just pause the hustle :) [00:12:30] to give time to fix it [00:13:25] I would imagine it's not intentional [00:13:40] but we'll find out during this survey I suppose [00:13:50] errr nevermind it's still an external one ;D [00:13:54] yeah leila was thinking it might be because she didn't need it the first survey and maybe jdlrobson didn't enable (but we're not so sure about how this should work) [00:14:10] haha yeah… we can also use the multiple choice version next week to test if that helps debug [00:55:51] jhobs: one question. did you put the sampling rate 1 out of 2000? :d [00:56:05] I'm getting a lot of responses, something is weird. [00:56:15] I made it what's on the task [00:56:50] .05 IIRC [00:57:06] 0.05 times 0.01 right? that number is a percentage. [00:57:40] Yeah, should be [00:58:02] is there a way to fix this? [00:58:07] cuz it's too high [00:58:45] There are no more deployment windows until Monday :/ [00:59:18] mmm. :D [00:59:32] I'm sure Ops can make some exceptions for us, right? [00:59:32] :D [00:59:55] Let me see what I can do [01:00:07] jhobs we can ask greg-g over pie, too ;) [01:01:02] gonna check with the guy who did my deployment earlier [01:12:06] jhobs: I talked to greg-g. He said you should stop the test. Can you do that? [01:12:22] (He said if you have deploy rights, you can do it, and I'm assuming you have.) [01:12:22] lzia: bryan davis just reduced the rate for us [01:12:27] oh! [01:12:29] i don't have deploy rights [01:12:29] perfect! [01:12:57] So are we at 5 out of 10000? (just to confirm) jhobs? [01:13:15] yep [01:13:34] perfect. thank you! it's much slower now. :-) [01:54:15] jgirault: you patch looks good. I don't have +2 so I added bmansurov as a reviewer. He's out tomorrow, but hopefully he'll get a chance to check it out on Monday (or hopefully someone else can by then) [01:54:22] s/you/your [02:03:57] jhobs: cool thanks [08:50:42] morning [09:13:53] o/ joakino [09:51:01] back in 10 [14:33:27] phuedx: do you think I should leave the second survey card in "doing" until the campaign is over, or move it to signoff because it's already been deployed (and is running)? [14:33:49] ^ or joakino [14:33:56] jhobs: only if there's another card that tracks turning it off [14:34:03] there is not [14:34:24] and that was my thinking, hence why it's still in doing atm [14:34:41] just didn't know if that would mess with our sprint (although it'll be taken down before kickoff on monday) [14:37:00] oh also, i've been experimenting with a "vim everywhere" extension for chrome :D [14:37:13] quite convenient to use j and k to scroll [14:38:02] * phuedx searches for said extension [14:38:09] lemme find it [14:38:15] Vimium [14:38:23] https://chrome.google.com/webstore/detail/vimium/dbepggeogbaibhgnhhndojpepiihcmeb [14:39:14] there's also one (with much more overhead) that lets you use any text editor for any textarea or contentEditable [14:39:22] So now I can write email in sublime, for instance :D [14:42:40] (with VIntage Mode, of course) [17:10:07] mbinder: I'll miss stand ups estimations and retros today, in a research workshop [17:13:42] phuedx: https://gerrit.wikimedia.org/r/#/c/253036/ nice :) and it seems to work great in zuul https://integration.wikimedia.org/zuul/ *yay* [17:34:16] niedzielski: Regarding T114524 Add button to play recorded audio pronunciation of the article. -- it says:"Add article title pronunciations" [17:34:41] etonkovidova: that's right [17:35:11] niedzielski: and I am supposed to see it? I do not see play bottom - there is the pronounciation icon for sure that shows a sort of pop-up [17:35:37] niedzielski: but not like in the ticket specs ... hmm [17:35:55] etonkovidova: it uses the content service. do you have that turned on? [17:36:32] niedzielski: hmm... it's in the dev options or something? [17:37:13] etonkovidova: yeah, let me double check the changes have been deployed so it's easy to test first. just a moment [17:38:29] etonkovidova: ok it looks like it should work. i'm testing on the neptune article. here's the setup: go to developer settings and turn on both useRestbase_setManually and useRestbase [17:38:45] etonkovidova: then tap "Crash" at the bottom of the app and press start over [17:38:50] niedzielski: ok [17:39:17] etonkovidova: note: title pronunciations only work for articles with recorded content using the IPA template [17:39:31] niedzielski: ok again [17:39:55] niedzielski: Are you in Design Research Workshop meeting :) ? I am sitting here - will go back to my desk at 10:00 [17:40:53] etonkovidova: i am not [17:41:26] niedzielski: ok then - I like listening to designers [17:41:44] :) [17:54:37] kaity_: ok, we'll catch you up if needed [18:08:02] niedzielski: to just enable or disable RB usage you don't need to force close/crash the app. (That is only needed if you change the RB URI format) [18:08:51] bearND: oh, cool [18:24:55] getting the kids to bed [18:37:40] hi halfak [18:37:49] hey leila [18:37:50] :) [18:37:56] * halfak sends message in -research [18:38:05] but -mobile is cool too :D [18:51:15] bearND: would you like to join us for Android retro? [18:51:34] mbinder: omw [18:53:03] mbinder: are you starting now or at the top of the hour? [18:53:16] i don't see anyone else in the h-o [18:53:46] bearND: sorry, didn't mean to alarm, just wanted to check since we all said our goodbyes :) [18:53:55] we'll be there in 5-10 [18:54:48] mbinder: ok. Yeah. I think I had my calendar showing the wrong week earlier [18:55:11] didn't think there was a retro but i do see it now [19:25:56] bearND: niedzielski have TWN updates always been done manually? [19:27:24] bgerstle: no. Previously l10n-bot did that for us, and it (or should I say he) does it still for other projects. [19:28:00] bearND: right, but IIRC our recent changes should have had the bot push directly to our repo's master branch [19:28:03] so they should be automated for us [19:28:18] we haven't gotten an update since 10-22, but there have been edits according to https://translatewiki.net/w/i.php?title=Special:RecentChanges&limit=500&namespace=1206&days=30&translations=only [19:29:16] is there a mailing list for 18n? [19:31:51] bgerstle bearND: i've been doing them manually lately. we recently had a couple patches from l10n-bot but there were some issues with waiting for tests to pass before merging [20:30:26] mdholloway: https://gerrit.wikimedia.org/r/#/c/254366/ should be ready to be merged [20:30:55] bearND: cool, was just looking at it, will merge now [20:31:18] mdholloway: i just amended a minute ago to fix a typo [20:31:23] bearND: ok [20:31:40] bearND: out of curiousity, do we have even a guess of what % of users have images turned off? [20:33:33] we could make a game out of it. I'm guessing <1%. [20:53:47] niedzielski: so, researching the opengl issue a bit further, it looks like ES 2.0 is supported in Android 2.2+, so the current check that I added is meaningless... [20:54:28] dbrant: so the emulators are just an incomplete implementation? [20:55:07] niedzielski: Right, this means that the crash only happens on the AOSP API15 emulator. Do we want to exert a lot of effort to prevent that specific crash on the emulator? [20:55:34] dbrant: i guess we'll find out in the beta if there's any issues. sorry for the red herring [20:55:54] niedzielski: ok! i'll take out the check, then [20:56:06] dbrant:sounds good. thanks for checking [21:25:29] niedzielski: well, shoot. how are you repro'ing? [21:26:07] mdholloway: oh, i just installed the apk with your patch. launched the app. then i reinstalled the apk from master and launched the app again [21:26:36] mdholloway: i mentioned it because it's interesting that it fails one way but not the other. i don't know if that's expected or not [21:34:56] niedzielski: yup, same here. that's pretty weird. [21:39:25] niedzielski: it seems like this might be related to the upgrade to gson v2.4 -- that change would have been in the beta where we started seeing this, i think. [21:40:03] mdholloway: i thought we saw this issue from day one with bernd being the first to report it? [21:40:26] yep, that's why we're suppressing it in production [21:43:18] niedzielski: dbrant: i have no recollection of that, but i believe you. [22:33:26] mdholloway: hey you still around? [22:33:40] kaity_: hey, yep, i'm here [22:34:00] mdholloway: do you know how wiktionary window will be invoked? [22:34:19] mdholloway: when you highlight 1 word and it returns a wiktionary match? [22:35:25] dbrant and niedzielski and i just discussed that earlier today. it sounds like what we're thinking is a nonmodal popup that appears at bottom of screen somewhat like a link preview [22:35:37] mdholloway: right [22:35:40] but persists so that you can highlight a different word and it will remain open [22:35:50] until closed by the user [22:36:00] mdholloway: what if you highlight multiple words or there's no matching definition? [22:37:53] kaity_: hmm, i hadn't thought about the case of multiple highlighted words. for no existing definition (which i imagine would occur frequently) i imagine we'd just display a "no definition found" type message [22:37:56] kaity: i was kind of wondering, assuming wiktionary goes well, if we might want to remove the system "contextual action bar" at the top and move those buttons into the new popup. my reasoning is that the user is highlighting text in each case and we should have a more focused experience [22:38:25] in the multi word case, we might assume, for example, that the user wants to share a fact [22:38:47] niedzielski: right thats a good point [22:39:04] er, but probably not a v1 type thing [22:39:08] :) [22:41:10] kaity_: i suppose there are a couple of different ways we could handle multiple-word selections: display the first word (if any) with a match, display all matches, display all matches up to a certain max... [22:41:34] niedzielski: kindle has them in context to the highlighted word, but then there are buttons all over the plce [22:42:20] kaity_: i will have to check that out. i think kindle is the perfect starting place for this [22:42:25] mdholloway: thats a good idea [22:42:40] * mdholloway needs to spend some time with the kindle [22:43:03] mdholloway: tho if multiple words are highlighted maybe the intention is to copy/share [22:43:08] mdholloway: we should think about how that might look for the api too. do we want a "jumble o words" endpoint that returns very terse output for each or just not handle that case? [22:43:49] kaity_: niedzielski: that is possible. probably true for multiple words, though i don't think it would be valid to assume the inverse: that a single highlighted word isn't intended for sharing [22:44:47] mdholloway: niedzielski ibooks does not display definition window at first but has it as one of the options: define, highlight, share [22:47:40] kaity_: that sounds like a good interface. in my head, at least, immediately showing a definition when a word is highlighted seems kind of over-eager [22:48:15] but again, i'll have to spend some time with both the kindle and the iphone i recently received to compare and contrast approaches [22:50:58] kaity_: i _think_ highlighting a word for dictionary lookup would be primary use case of highlighting a word. i only use share a fact for fun and rarely copy text from a wiki. offhandedly, i think i personally would prefer to see the definition first and then choose a different option if i wanted a rarer use case [22:51:26] kaity_: of course, i'm sure eventlogging will reveal that users actually highlight text for another option i've never even thought of :) [22:52:11] niedzielski: ping [22:52:24] dr0ptp4kt: hey! [22:52:47] niedzielski and mdholloway : kaity is having a connection issue [22:53:08] dr0ptp4kt: oh ok, thanks for the heads up [22:54:20] dr0ptp4kt: word [22:55:01] i actually have to run now. have a good weekend, all. [22:55:17] peace mdholloway [22:55:50] mdholloway: night! [23:30:34] hey jhobs! just to be 100% clear, how do i define percentage for the survey? [23:32:02] atgomez: put .01*the percentage as the value of the coverage field (so 1.0 is 100%) and then list the percentage as a comment behind it [23:32:35] for extra clarity :) thanks [23:33:14] Yeah, the comments are all just for clarity and double-checking that the wiki pages have the correct values [23:54:11] jhobs: i'm working on writing the messages for a multiple choice version (phab task almost exists) and i understand i need to have staff rights to make the messages? [23:56:03] atgomez: on enwiki, yes. Or get someone with staff rights to make them for you. I, unfortunately, don't have staff rights either and don't know the process for getting them :/ [23:57:28] hm. [23:57:32] It's because the pages are in the MediaWiki namespace I believe [23:57:35] yeah… [23:57:41] * atgomez thinks a UI would be really rad [23:57:52] JonKatz: do you have staff rights on enwiki? [23:57:55] * jhobs agrees wholeheartedly [23:58:02] don think so [23:58:09] how do i check? [23:58:33] leila doesn't either, she's had Guillaume create them for her in the past [23:58:42] JonKatz: go here, can you edit? https://en.wikipedia.org/wiki/MediaWiki:Reader-segmentation-1-message [23:59:04] hm [23:59:17] JonKatz: (If you can, make sure you don't) [23:59:27] thanks jhobs [23:59:30] :P [23:59:53] Haha yeah wouldn't want to mess with the live survey