ubuntu-phone team mailing list archive

[Loïc Minier <loic.minier@xxxxxxxxxx>]

       Hi there,

Friday (and Saturday!) have been very eventful in terms of landings;
hence a pretty long email!

The original plan to finish some relatively safe landings -- bug fixes,
low hanging fruits etc. -- to clear the runway for Mir landing end of
Friday.

The prerequisites for Mir on Friday were:
* fixing input focus bug when trying to pick up a call when not within
 dialer-app
* fixing the webbrowser input issues (seemed to relate to volume up/down
 input events not reaching Unity 8)
* getting Mir along the SurfaceFlinger results results on the dashboard
 (as to compare test results and as to keep the option open to roll
 back to SF)
* decent autopilot pass rate for Mir
* fixing an AutoPilot input issue on maguro

Good progress has been made debugging the input issues over Friday, but
eventually time ran out to fix these.  The good news is that the image
testing dashboard now shows Mir (touch_mir) results along the
SurfaceFlinger ones (touch_ro):
   http://reports.qa.ubuntu.com/smokeng/saucy/
and the other good news is that results are pretty good!  Lastly, we
also landed a fix for the AutoPilot input issue on maguro (there are
testers with only a Galaxy Nexus for testing).
 The Mir input issues will likely be fixed over the course of Monday,
allowing us to aim for a switch end of Monday.  Sergio has prepared a
livecd-rootfs patch to turn it on by default [1].

Over Friday and Saturday, the landing team tested and landed:
* unity 8 with tons of bug fixes + unity-notifications +
 indicator-network with cool new snap decisions for wifi
* indicators stack
* settings stack
* webcred stack
* phone stack
* ubuntu-keyboard
* fix for AutoPilot input issue on maguro
* boot related speed up and bug fixes (dbus job output,
 upstart-file-bridge, lxc-android-config fix for shutdown/reboot,
 ureadahead fix)
* initial langpacks
* and even an unity 7 update (desktop)

There was an unfortunate regression in unity8 that caused it to use 50%
of the CPU all the time (screen on or screen off), that only got caught
after image #81 had been rolled and that translated to dramatic image
test results, so the unity8 update was reverted in #82 and we got a
fixed unity8 in image #83.



What's next?

Mir is our top priority for Monday, before end of day we will likely
have some fixes for the input issues mentioned above.

In the mean time, Monday morning we ought to:
* consider promoting image #83
* retry landing the UI toolkit update; this update wasn't actually that
 buggy in itself, but it caused many app testsuites to fail because
 they hardcoded a database pathname; it's also blocking a lot of other
 landings right now
* try landing fixes for unity8 autopilot tests under Mir
* land other non-risky things (e.g. thumbnailer which is unseeded)


   Cheers,


PS: sorry for the delay in sending this

[1] https://code.launchpad.net/~sergiusens/livecd-rootfs/mir_default/+merge/189429
--
Loïc Minier