It will make the crash go away. However it won't fix the problem of running the wrong ABI, all this does is catch the crash and handle it by alerting the user that they downloaded the wrong build [1]. There is at least one related bug [2] that would need to be included as well. Sebastian did most if not all the work on the feature.
[1] https://www.youtube.com/watch?v=99TifhtYgew [2] https://bugzilla.mozilla.org/show_bug.cgi?id=1222925 Kevin On Thu, Dec 31, 2015 at 10:17 AM, Ritu Kothari <[email protected]> wrote: > Hi Kevin, Margaret, > > I am also in support of uplifting the patch from bug 1119915 to Beta44 > even though it includes a string change. This has been a long standing > crash across multiple releases and if we can alert users in some if not all > cases, I think we are in a better place than not doing anything about it. > > Bug 1119915 was fixed in Aurora45 when it was still on the Nightly > channel. Kevin, *does the crash-stats data show any improvement on this > specific issue on those channels* (I was unable to tell)? If yes, that is > a pretty strong argument towards uplifting this to Beta44. > > Thanks, > Ritu > > > On Thu, Dec 31, 2015 at 8:54 AM, Kevin Brosnan <[email protected]> > wrote: > >> We don't understand this issue well. Current theorys are a 3rd party >> store that only provides an ARM binary. Some sort of forum post that is >> popular or some sort of direct link on the internet to download Firefox. I >> have seen a different crash that is an emulator that allows users to run >> Android apps on desktops [1]. Since it is a a crash on startup it is simple >> for a user to crash many times inflating the crash rate. >> >> [1] http://www.bluestacks.com/comparison-chart.html >> >> Kevin >> >> On Thu, Dec 31, 2015 at 8:48 AM, Margaret Leibovic <[email protected] >> > wrote: >> >>> On Wed, Dec 30, 2015 at 10:52 PM, Kevin Brosnan <[email protected]> >>> wrote: >>> >>>> >>>> The second issue that is skewing the crash data is that the number 2 >>>> crash is the ABI mismatch [5]. Running Firefox for Android for ARM devices >>>> on x86 or vice versa. We have some code changes in v45 to try to alert >>>> users that they need to use a different build. They are rudimentary but at >>>> least give some feedback other than crashing [6]. >>>> >>> >>> Do we understand how beta users are even encountering this issue? Is >>> Google Play giving them the wrong version? Is there any way for us to know >>> the source of the builds running into this problem? >>> >>> Uplifting bug 1119915 would reduce the crash rate, but I'm worried about >>> how users are even getting these incorrect builds in the first place. Also, >>> if we wanted to uplift that, we'd need to uplift a string. >>> >>> Margaret >>> >> >> >> _______________________________________________ >> release-drivers mailing list >> [email protected] >> https://mail.mozilla.org/listinfo/release-drivers >> >> >
_______________________________________________ mobile-firefox-dev mailing list [email protected] https://mail.mozilla.org/listinfo/mobile-firefox-dev
