mozilla :: #mobile

18 May 2017
04:11finkeytron3000bah, why noez readingList :/
15:16* ahunt grumbles about having no test devices compatible with the O beta
15:24mkaplyHow do you view crash reports on mobile? about:crashes doesn't seem to work?
15:25mkaplyAlthough I see it being sent in the log
15:25mkaplyhttps://crash-reports.mozilla.com/submit?id={aa3c5121-dab2-40e2-81ca-7ea25febc110}&version=53.0.2&buildid=20170504210220
15:25mkaplyBut searching on that ID doesn't show the crash
15:25mkaplyThe actual crash seems to start with java.lang.UnsatisfiedLinkError: dlopen failed: library "/data/data/org.mozilla.firefox/lib/libmozglue.so" not found in the log
15:25mkaplywhich seems really bad
15:41ahuntabout:crashes should work, unless theres a bug and/or its your own build
15:43ahuntUploads arent immediate, but it its not in about:crashes then something went wrong...
16:02snorpum
16:02snorprbarker, jchen, droeh: I can't use vidyo because I have to reset my ldap :(
16:02snorpso....
16:02snorpI'm gonna go do that.
16:02droehalright
16:04snorpkbrosnan: can you grant triage privs to my folks
16:04snorpor did we already try that
16:05snorpoh we're supposed to be setting the firefox tracking flags anyway
16:05snorpso maybe you guys already have those privs?
16:11jchensnorp: can you use vidyo yet? i'm in another vidyo room watching the profiler talk
16:11snorpI cannot
16:31mkaplyI was testing about:crashes on nightly. I've asked the partner to go to about:crashes and give us the ID.
16:35kbrosnansnorp: pm/email me a list
16:57ahuntDoes the project dawn/aurora transition also include disabling the mozilla updater, or is that being kept?
16:58ahunt*for nightly updates
17:18nalexanderahunt: it's definitely being kept and will be tested on Android to some extent.
17:18nalexanderahunt: but no user population of any size will be using it, so expect rot :(
17:19ahuntnalexander: good to know, I was looking at ways of making it a bit more robust in face of startup crashes, and wasnt sure whether it would be a complete waste of time to do so.
17:19ahuntIf its still being used I guess thats at least slightly useful.
17:19nalexanderahunt: mostly a waste of time, since nobody will be really depending on it; but not entirely a lost cause.
17:20ahunt(Bug 1359730 for more, it shouldnt need huge changes to survive a fennec startup crash)
17:20firebothttps://bugzil.la/1359730 NEW, ahunt@mozilla.com Safe Updater for nightly
18:30rfeeleysleroux: i see this from sync fest "Make sure there is a bug on file to disconnect iOS device via push when disconnected through device manager"
18:30rfeeleysleroux: is it on file somewhere or should i file it?
18:30slerouxrfeeley: not yet
18:39mkaplyahunt: MozillaOnline uses the updater.
18:43droehdaleharvey: ping
18:55daleharveydroeh: pong
18:56droehdaleharvey: I'm working on converting WebAppActivity to use GeckoView, but I can't seem to get anything to actually install as a web app -- it looks like the manifest url never gets set on the tab
18:56droehdaleharvey: (testing with https://www.pokedex.org/ )
18:56daleharveydroeh: anything in pwa.rocks should work
18:57droehdaleharvey: And I should just need to hit "Add to home screen" to install one?
18:57daleharveydroeh: hrm sounds like something else is up, pokedex definitely works as a pwa
18:58daleharveyyup, with manifest.install.enabled pref set to true, add to homescreen
18:59bcsnorp: do you have a bug already on file to support extensions in geckoview? would these by legacy extensions or webextensions do you think?
18:59snorpbc: webextensions
18:59snorpbc: don't think there is a bug
18:59snorpoh
18:59snorpno there is
18:59snorphttps://bugzilla.mozilla.org/show_bug.cgi?id=1322580
18:59firebotBug 1322580 NEW, nobody@mozilla.org [geckoview] Consider using Web Extensions to allow content script access
19:00bcthanks
19:00snorpbc: I'm working on adding content script access directly right now
19:00droehdaleharvey: ah! the pref was what I was missing, thanks!
19:00snorpcause I need it for something
19:00snorpit's almost working
19:00nalexandersnorp: bc: a lot of that doesn't really make sense; much of the Web Extension API is intimately tied to browser chrome and UI (and specifically Chrome's UI).
19:00bcsnorp: is this something that will allow me to quit the app?
19:00snorpnalexander: right
19:00snorpnalexander: it would be some kind of subset
19:00snorpnalexander: but there is value in having, e.g. uBlock Just Work
19:01snorpbc: hmm, not currently
19:01snorpbc: why do you want to quit, I forget
19:01snorpbc: can't just kill it?
19:01bcI wonder what the future of quitter is in the tree.
19:01snorpI can't remember why you want the quitter
19:02snorplike, android doesn't "quit" us
19:02snorpit just kills us like a savage
19:02snorp(sometimes)
19:02bcsnorp: Isn't all kinds of bad things related to the app screwed up when you kill the app rather than just shut it down?
19:02snorpbc: generally not
19:02snorplike I said, normal
19:02snorpbut you may need to trigger onPause first
19:02snorpby going to the home screen
19:02snorpwhich I think you can do via adb
19:02bcI thought we had this discussion ages ago when discussing creating the profile etc.
19:03snorpyeah it does seem familiar
19:03bcmmm. didn't know that.
19:03snorpis that what it is?
19:03snorpcustom profile creation causes problems?
19:03bcnot sure if it causes real problems apart from having to burn cycles with an arbitrary wait before killing.
19:04snorphmm
19:04bcwe'll come to it for sure when we kill legacy extensions i guess.
19:04bcthe in tree quitter may have a solution. i don't think they want to lose that.
19:05snorpok
19:08bcIt's only used in pgo optimization I guess.
19:23SeburoHi. Can I have some help with > https://support.mozilla.org/en-US/questions/1160485 (Not help finding drink and drugs, just help with answering the question). :-)
19:43ahuntI though we had a quit button in the main menu? (Which only appears if you flip some prefs in the browser settings)
19:45Seburoor if you install the Quitnow add-on.
19:46ahuntSettings->Privacy->"Clear private data on exit seems to be whats needed for quit to appear
19:48SeburoYup
19:48SeburoBut that leaves you with a privacy pref changed.
19:56ahuntSeburo: tbh Im not a fan of the English play store name ("Firefox Browser fast & private), and the google translation of the russian name doesnt sound much better ("Fast Browser Firefox).
19:57Seburoahunt: Is it Google that translates the text from what we supply?
19:58ahuntSeburo: I doubt it, I just google translated from russian back to english ( https://play.google.com/store/apps/details?id=org.mozilla.firefox&hl=ru )
19:58Seburoahunt: So Mozilla supplies the text?
20:16PikeCC delphine
20:17delphineah I remember this
20:17delphinelet me dig into my bugmail
20:18delphinehere we go: https://bugzilla.mozilla.org/show_bug.cgi?id=1347318#c3
20:18firebotBug 1347318 FIXED, nobody@mozilla.org Rework app title of Firefox product in the Google Play Store listing
20:19delphineSeburo: our localizers translate from the English string we supply
20:19delphineCheck out bug above, may give you more info
20:25Seburodelphine: Okay, fab, thank you. :-)
20:38delphinesure :)
20:40Seburodelphine: I have asked the user to supply details of the incorrect translation. Assuming I get a response, I will file a bug to get this fixed. Thanks again, appreciate the steer.
20:41delphine(oh sorry, seems like I totally missed that sumo link above and answered thinking you were talking about something else. Hope this helped, nevertheless)
20:42delphineIt may be because of the way the English string sounds once translated (and is why some localizers have chosen to translate it differently, as explained in the bug I sent)
20:45SeburoTrue. I do not know what the exact inaccuracy is, so hopefully the user can provide more info. Thanks again for your help and time with this. :-)
21:12TitanNanohttps://bugzilla.mozilla.org/show_bug.cgi?id=1366066
21:12firebotBug 1366066 UNCONFIRMED, nobody@mozilla.org Can't add Google IO to home screen
22:30mkaplyWow. Can't believe the fragmentation here - https://developer.android.com/about/dashboards/index.html
22:32mkaplyTitanNano: It's working for me on 53 and nightly.
23:11glassercHi! I have a couple questions about the versions of fennec that makes it to archive.mozilla.org. I see that we upload android-api-15 and android-x86 versions. Are both used by users or is the x86 flavor special somehow?
23:11glasserc(I'm looking at https://archive.mozilla.org/pub/mobile/releases/53.0/)
23:11glassercI also see that sometimes we've simultaneously released versions for android-api-X and android-api-Y, like in https://archive.mozilla.org/pub/mobile/releases/45.0/
23:12glassercWhat determines that? What differences would there be in the two packages?
23:18ahuntglasserc: both are used (although IIRC there arent that many x86 devices in the wild)
23:19ahuntI cant remember what the reason was for having separate api-9 package (they were for Android 2.X devices)
23:21glassercSo generally for a given release, I should expect it to be only on one API version?
23:25glassercBesides x86, do we cut versions for different processor architectures (e.g. ARMv6 and ARMv7)?
23:26ahuntBack when api-9 and api-15 existed: api-9 builds were for devices running Android 2.X, api-15 for Android 4 and newer. (Before that we had api-9 for Android 2.X, api-11 for Android 3.X and newer). Google play automatically distributes the right version to your device.
23:26ahuntAnd x86 was only supported on api-11+ (maybe api-15+) hence there was only one x86 apk
23:27ahuntglasserc: nope, theres just one arm build (I dont know which specific arch it targets though)
23:28ahuntIIRC nalexander might know more of the details, I only saw this from the sidelines.
23:28glassercOK. The reason I'm asking is because I'm trying to figure out how important it is to track the different Android builds for a given release, or whether it's "good enough" to say "Oh that's Fennec 54.0"
23:28glassercIt's for https://github.com/mozilla-services/buildhub/
23:28glassercWe're figuring to track "platform" and the question is whether Android is one platform or several
23:29nalexanderglasserc: I expect you'll want OS + platform, so {Android} x {ARM, x86, aarch64}.
23:29glassercOne reason it might be several is if crashes on one architecture might not affect other ones, or if the versions distributed in different API builds include different code
23:29nalexanderglasserc: correct.
23:30nalexanderglasserc: snorp will have an opinion.
23:30glassercAnd the first {Android} means Android API version? i.e. {Android API 9, Android API 11, Android API 15} x ...?
23:31glassercOops, nalexander ^^
23:31nalexanderglasserc: yeah, probably worth having the API version or version range in there, since at some point we're likely to split in some way.
23:32nalexanderglasserc: Android generally supports very fine-grained package splits: SDK, device density, screen size, language, processor architecture, etc. We're just not set up to do that.
23:32glassercYikes, OK, thanks
23:39kbrosnanfacebook pushes something like a 100 differnt splits of their messenger and primary fb app
19 May 2017
No messages
   
Last message: 126 days and 9 hours ago