[00:03:05] niedzielski: Good work today. I'm going to head out. Will be back later. [00:06:04] bearND|afk: night. i'm still looking into a checked state corner case improvement. should have something soon [09:10:56] hullo [09:20:05] hi phuedx [09:24:21] hey joakino [09:24:23] how're you? [09:25:01] phuedx: good, almost recovered from the jetlag [10:13:32] joakino: phuedx: hello! never get to chat w/ you guys anymore [10:15:46] hullo bgerstle [10:15:51] i haven't been around for a while [10:16:01] (which might also explain things) [10:16:11] glad you're back, then [10:18:57] sup bgerstle [10:19:15] o y'know, just rendering HTML natively NBD [10:23:36] bgerstle: we need to meet so that you guys tell me wat you're doing so that I can do it too here chimeces.com/webkipedia [10:24:22] joakino: yeah, once we get ourselves together on iOS (i.e. ship this to the app store) we plan to join Android on the "content service" [10:24:35] and perhaps that would be a good time to try consolidating web stuffs [10:24:58] so we can all try to share a mobile web view, or at least small variations on it [10:26:13] bgerstle: so you're not using the content service now? [10:26:39] joakino: not yet [10:26:57] joakino: android guys themselves are still hooking it up, last i checked [10:29:16] oki [10:29:27] bgerstle: so how are you rendering native html then? :p [10:30:12] https://github.com/Cocoanetics/DTCoreText [10:30:25] same thing wordpress ios is using (for now) [10:30:44] android has similar components, but afaik they're not using them [10:30:47] HtmlSpanner, IIRC [10:37:01] interesting [10:37:07] does it work well? [10:37:17] the article contents have some weeeeird html [10:45:40] joakino: that's why we have xpaths [10:45:50] joakino: will show a demo soon, for now, dogs clamoring for a walk [10:45:50] bbiab [10:46:03] bb! [10:48:35] bb [11:08:50] I'm out for lunch [14:41:59] back [14:59:09] dbrant: hey! may i have a little time this morning to talk to your singleTask thread? [14:59:21] niedzielski: of course [15:00:01] dbrant: thanks i wanted to collect my thoughts on it a little bit so i can elucidate in more detail [17:12:56] joakino: my stuff on those phab tasks about the surveys making sense? lemme know if you want to hop on video and discuss anything... [17:14:27] coreyfloyd: mdholloway mhurd bearND dbrant|brb standUP [17:14:49] dr0ptp4kt: want to chat about it? [17:17:36] joakino: sure. want to send me the url for your batcave? [17:17:54] (or was it haven of antman or spiderman?) [17:17:56] dr0ptp4kt: https://plus.google.com/hangouts/_/wikimedia.org/tracyisland [17:18:04] we're doing that one for now hehe [17:19:12] i'm in [17:19:18] ^ joakino [17:58:20] dr0ptp4kt: any chance we could move our 1:1 any earlier in the afternoon (to anytime after 11:30 a.m. Pacific)? [18:31:22] dbrant|brb: do you need anything from me before the beta release? [18:32:02] coreyfloyd: retro [18:32:03] bearND: could you weigh in on this: https://gerrit.wikimedia.org/r/228191 [18:32:09] ru roh [18:33:40] https://etherpad.wikimedia.org/p/MobileAppRetrospective [18:40:32] dbrant: ok, added a comment. [18:40:49] mdholloway: this one time or always? [18:41:30] dbrant: Any thoughts on what stories we can count towards completed velocity for sprint 62? [18:42:52] mbinder: at least https://phabricator.wikimedia.org/T104714 [18:43:59] dbrant: sorry, lemme rephrase. I meant what is done or expected to be done in sprint 62? And what is gonna carry over? Jsut wanna calculate points [18:46:51] dr0ptp4kt: just today (sent PM) [18:47:42] mdholloway: oh, cool. thx [18:47:45] mbinder: ^ that's the big point-carrying story that, if QA'd today, would contribute to velocity for this sprint. [18:48:10] mbinder: the things in Design signoff will probably all carry over. [18:48:15] dbrant: gotcha, thanks [18:48:36] dbrant: we can hold off calc until monday I think [18:48:54] mbinder: yep, cool [18:49:58] etonkovidova: mind tackling some Android QA when you have a chance? (especially https://phabricator.wikimedia.org/T104714 ) [18:51:46] bearND|afk: dbrant niedzielski mdholloway is the android 2.3 s2 something on which you need frequent testing from etonkovidova ? or is it more spot checking bugs and things of that nature? bearND|afk the device list reflects devices carved out specifically for etonkovidova per your list. we have several s3, and i'll have to ask someone to give theirs up. in [18:51:46] theory the s2 is in the cabinet somewhere. are there any devices etonkovidova should *not* really be using so much anymore? [18:54:50] dr0ptp4kt bearND|afk mdholloway dbrant: since s2 is #4 device, i think it would be good for qa [18:57:20] dr0ptp4kt: i think generally 2.3 is more for spot-checking... but it would be a nice bonus if all stories were tested on a 2.3 device, especially stories with visual changes. [18:58:52] dbrant: agreed. i would say that if gb is tested, the s2 would be a great choice [19:00:08] niedzielski: dbrant bearND|afk mdholloway what would be the order preferred for OSes for etonkovidova when she's doing testing? 5.x, then 4.x, then 2.3? something else? [19:00:37] niedzielski: dbrant mhurd bearND|afk trade offs, trade offs [19:01:10] dbrant bearND|afk mdholloway dr0ptp4kt: i vote 4.x, 5.x, 2.3.x, 3.x. most of my dev work is on 5.x [19:01:14] i think regular testing on 2.3 would be valuable. (google emphasizes the need to take care to support lower-end devices, which make up a significant share of the market; see, for example, here: https://medium.com/google-developers/developing-for-android-i-understanding-the-mobile-context-fd2351b131f8) [19:01:41] i spend most of my time in 4.4+ but probably shouldn't... [19:01:56] dr0ptp4kt: niedzielski: agreed with that order [19:02:11] ^+1 [19:02:27] mhurd: http://hatjitsu.wmflabs.org/23550 [19:06:52] dbrant: thx! Looking into "some Android QA" and "especially https://phabricator.wikimedia.org/T104714 " :) [20:21:32] dbrant: I am unable to access deployment-eventlogging02.eqiad.wmflabs even after having ops update with my new SSH key. Any ideas why this would be? [20:31:51] hey guys, who's maintaining the Barry and Frankie bots? they're spamming Phabricator with hundreds of blank pastes: BarryTheBrowserTestBot (and relatives) spamming Phabricator Paste application with blank pastes [20:32:00] er, wrong paste: https://phabricator.wikimedia.org/T107549 [20:32:40] jdlrobson: joakino ^ [20:37:39] bmansurov: rmoen ^^ [20:50:42] hey coreyfloyd [20:50:54] dr0ptp4kt, greg-g, jdlrobson , Blanks pastes are no bueno. Let me try to shut them down [20:51:05] rmoen: thanks! [20:51:33] rmoen: danke [20:52:35] dr0ptp4kt: I was kinda fumbling deciding/finding the right projects to put that in to get the reading team's attention. Is #reading-web the right "first entry point"? [20:53:20] greg-g: reading-web is correct [20:53:40] coolio [20:55:43] greg-g: hmm looks like i don't have access to the instances jdlrobson is running the bots on. ;( [20:56:05] not sure why they are pasting blank, i know some changes were made to the bot recently to make them more realtime [20:58:09] rmoen: this looks wrong: [20:58:11] if args.paste: [20:58:11] if not is_good: [20:58:11] print 'Pasting commit %s with (is good = %s)..' %(commit, is_good) [20:58:14] paste_url = get_paste_url(output) [20:58:17] print "Result pasted to %s"%paste_url [20:58:19] else: [20:59:24] well, not sure [20:59:37] * greg-g doesn't have a box set up with the dependencies [21:01:04] heh, an hour ago, jdlrobson did the same thing: https://phabricator.wikimedia.org/p/Jdlrobson/ [21:01:09] see the last entry in his activity [21:01:21] a blank paste, maybe he was testing the code locally/with his account [21:02:11] dr0ptp4kt: rmoen can I assume you two will take this to completion in a timely maner? [21:02:13] greg-g: looking at the code, get_paste_url runs a shell command to arc to paste with the text.. [21:02:15] greg-g: why does the above code look wrong? [21:02:40] bmansurov: because I forgot which function did which thing :) [21:02:44] ie: I was wrong [21:02:51] greg-g: ok [21:03:18] greg-g: is it pasting against all repositories? there are not many open MS or Gather patches [21:03:24] MF [21:03:46] Grabbing late lunch [21:03:54] no way to tell since the pastes are blank, but I assume only responding to patches in correct repos, since it's not at 20,000 yet :) [21:04:25] bmansurov: rmoen who else has access to the box that runs barry? [21:04:34] greg-g: I will do the best I can, keep in mind, I don't have access to the instances or know where they are running. Also I didn't write this bot, I created an install script, so my knowledge of its internals is minimum [21:04:41] get someone from -labs to add you to it if you need to [21:04:55] yuvi or andrew or coren can add you to it [21:05:16] it's a 240 line python script :P [21:05:17] greg-g: Again, Not sure what project or instance i need to be added to. [21:05:21] ok [21:05:44] greg-g: I'll see if I still have access [21:06:01] rmoen: you also should have access already [21:06:07] actionables: get the rest of the team access to the box [21:06:15] bmansurov: do you know what box / [21:06:15] ? [21:06:37] rmoen: looking [21:08:39] greg-g, bmansurov for good measure, I've killed the instances in the project reading-smoketest, which is where these should have been running [21:08:45] not sure where jdlrobson is hosting the bots otherwise [21:08:58] * greg-g nods [21:09:14] afaik the two instances running in reading-smoketest were not running anything, but its possible i suppose [21:09:38] rmoen, greg-g: here I think: https://wikitech.wikimedia.org/wiki/Nova_Resource:Mobile-smoketests [21:09:56] other than that i have no idea, Doesn't seem like they are running in reading-staging. as they shouldn't be. I've looked into editor-engagement but don't see anything there. so im assuming the instance is invisible to me [21:10:23] bmansurov: those look right [21:10:25] rmoen: you're an admin there ^ [21:10:42] ok cool, dont know why they are invisible but cool [21:10:54] he doesn't look like an admin on mobile-smoketests.... [21:11:06] i will login if i can [21:11:08] unless his name is "novaadmin" :P [21:12:12] correct, i'm not a member of the project " mobile-smoketests " [21:12:14] which is why i cannot see it [21:12:30] bmansurov: you can add him [21:12:35] rmoen: i'm sorry, I was looking at reading-smoketest admins [21:12:36] er, no you can't [21:12:41] only jon and yuvi [21:13:12] greg-g: yeah, i'm not an admin [21:13:15] pinged him [21:13:19] in ops chan [21:13:34] ah, I pinged him in -labs :) [21:13:39] should i kill instance of frankie and barry? [21:13:43] s* [21:13:47] yeah [21:13:50] ok [21:13:52] just kill them [21:14:19] 21:13 < greg-g> YuviPanda: please add rmoen and bmansurov to https://wikitech.wikimedia.org/wiki/Nova_Resource:Mobile-smoketests as admin [21:14:22] 21:13 < greg-g> that project has an instance with a script that is spamming phabricator empty pastes [21:14:25] 21:14 < andrewbog> greg-g: I can do it, one moment... [21:14:31] oh ok maybe not then [21:14:33] bmansurov: [21:14:46] just letting you know andrew is adding you two as admins [21:14:51] perfecto [21:14:56] i will login asap [21:14:59] oops i killed a couple of frankies [21:14:59] and kill the screens [21:15:25] bmansurov: its pretty painless to set these up so no worries [21:15:46] thanks a ton, rmoen and bmansurov [21:16:00] np, sorry couldn't act faster [21:16:39] well, it just sat in the #phabricator project over night, you all responded after I pinged you, so thanks :) [21:17:00] (ie: it's obvious you didn't see it until I pinged you) [21:18:01] rmoen: bmansurov can one of you update that task, plz, when your done? kthx [21:18:32] greg-g: ok [21:18:44] rmoen: are you on to this? or should I go and kill the remaining bots? [21:18:52] bmansurov: i'm on it [21:18:59] thanks [21:19:48] bmansurov: actually, i've logged into the three instances and don't see any processes running for barrybot.py [21:21:20] try frankie? [21:22:04] bmansurov: ps aux | grep barrybot [21:22:07] returns nothing [21:22:22] rmoen: hmm, grep barry maybe? [21:23:06] bmansurov: which ones did you kill ? [21:23:11] maybe they are not running anymore [21:23:28] rmoen: frankies on Mobile-browser-tests.mobile-smoketests.eqiad.wmflabs [21:23:48] ok i still havent found barry yet [21:23:50] only seen frankies [21:23:56] bmansurov ^ [21:24:22] arent they the same guy, but different names? [21:25:17] do a test patch and see what happens? [21:25:24] to maybe trigger him or not? [21:25:58] last frankie paste was at 19:57, last barry was at 20:09 [21:27:42] https://gerrit.wikimedia.org/r/#/c/223158/ for brarry [21:28:13] i also tested crontabs, no jobs are running [21:28:54] ok so i've checked all three instances running under :Mobile-smoketests [21:28:59] no barrybot running. [21:29:39] bmansurov: i would imagine barry is running under jdlrobson's user under Gather-browser-tests as the script is in his home directory [21:29:49] but i see no process [21:29:52] even when logged in as him [21:31:05] rmoen: this is the latest by jon https://phabricator.wikimedia.org/P1807 [21:31:26] rmoen: i think it's a good idea to link back to the patch in the paste [21:34:51] niedzielski: hey! i've been in meetings most of the day :( got a few minutes for a hangout? [21:35:12] dbrant: yeppers! omw [21:40:24] hey greg-g sorry i was in an interview. Yeh i was debugging some issues earlier with the empty pastes from the bots [21:40:35] i thought i'd fixed it wed but there was a rogue process going on [21:40:44] heh [21:41:05] well, we may have killed all of your bots, sorry! :) [21:42:17] np, phabricator probably shouldn't accept empty pastes imo not sure if we can configure it not to [21:42:28] i'm double checking the script logic because it's not suppose to paste an empty paste :) [21:42:47] https://phabricator.wikimedia.org/P1807 < result of echo '' | arc paste [22:06:00] niedzielski: sorry to bother... can you take a look at https://phabricator.wikimedia.org/T107631 -- I know that "no reliable reproducible steps" sounds lame :( [22:06:29] niedzielski: does it make *any* sense? [22:06:48] etonkovidova: hmm [22:07:03] etonkovidova: this might be a network issue we had seen very rarely. let me see if i can repro [22:07:57] niedzielski: just for the first load of the article - try to be aggressive and hit ToC several times :) [22:11:56] etonkovidova: the scenario for reproing makes sense to me but i haven't been able to repro yet. how often does this happen? [22:12:25] niedzielski: sigh... often-ish [22:12:38] niedzielski: ok, ok, let me think about it [22:13:03] niedzielski: but that "flickering" action on Nexus 5 was spectacular :) [22:13:09] etonkovidova: i'll try some more devices [22:13:10] etonkovidova: this is on wifi? [22:13:18] niedzielski: yup [22:13:56] etonkovidova: is this a regression or an issue in the previous release? [22:14:13] niedzielski: never seen it before [22:14:37] etonkovidova: right but the nexus 5 is a new test device, right? [22:15:23] niedzielski: right. But I also saw the issue - error unknown on my favorite Samsung tablet 4.0.4 [22:21:09] etonkovidova: no luck so far :| i have another device or to i want to try. i will have to show you how to capture the debug logs next time so we can include those in bug report [22:28:59] etonkovidova: i can't repro :| [22:31:54] hi niedzielski. I'm back. What did I miss? [22:34:14] bearND: dbrant and i discussed singleTask launch mode further. we'd like to do some experiments and also talked briefly about an alternate singleton approach. etonkovidova is seeing a weird regression with TOC. i've not been able to repro yet but she has it repros frequently on here devices. https://phabricator.wikimedia.org/T107631 [22:35:02] bearND: i don't have a nexus 5 but it would be good to test on that [22:35:09] niedzielski: Thanks. Yeah, I commented on that task. Maybe there is some info in the logcat re:the ToC issue [22:35:39] niedzielski: I've got a Nexus 5. I'm going to try it. Basically just latest master build, right? [22:35:51] bearND: that's right [22:36:48] bearND: it might be worth working with etonkovidova on the logcats sooner rather than later. [22:36:49] etonkovidova, do you have time to do a brief walkthrough using android studio to collect debug logs? [22:37:11] etonkovidova: it should be pretty straightforward [22:37:37] niedzielski: sure google hangout? [22:38:36] etonkovidova: yep [22:38:42] etonkovidova: i'll call you in just a moment [22:38:54] niedzielski: ok [22:53:37] etonkovidova: niedzielski: hmm, no luck reproducing this one so far. I've tried the '2015 in sports' article and tap on a section in ToC that's not loaded. Nothing bad happens. [22:54:09] bearND: working with niedzielski [22:54:35] good. Yeah, was just FYI [23:06:31] etonkovidova: if you have trouble with logcat in android studio, you can install this app on phone: https://play.google.com/store/apps/details?id=com.nolanlawson.logcat [23:06:58] etonkovidova: oh nvm, it says this requires root access :( [23:07:24] niedzielski: no trouble :) I just added logcat for flickering cause I saw it again [23:08:04] etonkovidova: hm, nothing telling in this unfortunately :/ is there any more logs showing? [23:08:19] niedzielski: exactly... [23:09:26] niedzielski: what about: Background partial concurrent mark sweep GC freed 212(7KB) AllocSpace objects, 3(18MB) LOS objects, 15% free, 88MB/104MB, paused 7.257ms total 59.266ms or [23:09:29] niedzielski: file:///android_asset/bundle.js:602 - Uncaught TypeError: Cannot read property 'offsetTop' of null [23:10:08] etonkovidova: the "sweep GC" line is normal. i have seen this though... "i: file:///android_asset/bundle.js:602 - Uncaught TypeError: Cannot read property 'offsetTop' of null" [23:10:11] that might be telling [23:10:18] haven't* [23:11:15] etonkovidova: yeah the 'offsetTop' looks promising [23:11:22] etonkovidova: is there any other output near that? [23:11:37] niedzielski: oh my - I think it's simpler - when a page is still loading, scroll ToC several time trying to get to the bottom [23:12:49] bearND: ^^ [23:12:55] niedzielski: if ToC is still open and the page starts "flicker" reflecting the fact that more sections of the article are loaded [23:13:23] bearND: i think we intend to hide TOC + FAB until page load is complete [23:14:31] niedzielski: I thought we did but I guess this has changed since we moved to the ToC FAB button [23:14:38] niedzielski: even with "Using page from cache: 2015 in sports" and stuff like "Background partial concurrent mark sweep GC freed 25372(2MB) AllocSpace objects" [23:17:57] etonkovidova: I scrolled the ToC all the way to the bottom several times, and scrolled in there some more but did not notice any flickering [23:18:11] etonkovidova: are you still in video call? [23:18:24] I wouldn't mind seeing it myself [23:18:24] bearND: the article is cached quite effeciently [23:18:32] bearND: i think the issue is calling setToCButtonFadedIn(true) in displayNewPage. this should move to onPageLoadComplete [23:19:11] bearND: I saw it twice - after that no more flickering [23:20:24] bearND: delete it from History - and I reproduce it very well now [23:20:50] etonkovidova: I had cleared data even [23:21:44] bearND: it could be only for nexus 5 with 5x ... [23:24:10] bearND: another tip - scroll ToC to near bottom and click on a section - ToC closes and re-open it again. It opens and and starts "flickering" [23:24:23] etonkovidova: I've got 5.1.1 [23:25:33] howdy all. just got back and tried on my samsung siii (4.4.4); no luck reproducing. [23:26:12] bearND: yes, try actually clicking on the ToC item and then re-open ToC again - in benign case the ToC will just scroll up [23:26:18] etonkovidova: oh, i see, what you call flickering I call scrolling [23:27:14] bearND: somehow it look much-much worse on nexus 5 5.0.1 [23:27:29] yeah, the ToC entries scroll to the place where you currently are on the page. Since you had previously scrolled to the bottom of the ToC but on the page you are still at the top it'll take a long time to scroll all the way up on such a long page [23:27:36] the scrolling i am seeing. [23:28:22] bearND: yup, the benign scrolling :) I see it on 4. 0.4 [23:29:05] bearND: still think that ToC should not be availble until page is loaded - like it is on iOS [23:30:44] etonkovidova: You could shoot a video from Android Studio. Go the the Android tab (press Cmd-6). There you'll see two vertical columns of small buttons with little icons. The 2nd on from the top on the left hand side is for recording videos from the device. [23:31:04] If you hover over that it'll say Screen Record. [23:31:48] bearND: thanks - I was looking for it! :) will do [23:33:05] bearND: although I think the issue is clear - even with less extreme case with 4.0.4 there a weird blank page displayed with a ToC keeps scrolling up... [23:36:48] etonkovidova: would you show it to me, either on a video or in a hangout (e.g. in the batcave)? [23:37:23] bearND: have to go in 20min :( [23:50:32] bearND: I got the video - after some creative editing(just kidding!) I'll send it to you