ubuntu-phone team mailing list archive

[Didier Roche <didrocks@xxxxxxxxxx>]

Hey,

Two images today, one revert and then a fix, some jenkins outage, busy 
Friday!

# 265:
- faster phone boot time thanks to some tweaks to force all CPU cores to 
be used during boot
- new unity8 to get better stacktrace on crashers (timeout increased), 
some string enabled for translations, compatibility with latest powerd
- some click scope enhancement and removal of old vala scopes
- fixes for powerd
- some various HUD bugfixes
- new accounts plugin for a better scope/account integration  in google 
accounts with evolution backend, and a lot of optimization in the loading
- other webbrowser app changes to prepare for the oxide switch

-> we couldn't get test results on that one as powerd did prevent the 
screen unlocking to happen. We timely reverted (once identified) and got 
a powerd fix that entered the next image. The screen unlock only impact 
the tests (as it's unlocking from a terminal, which was what the 
policykit profile denied). We asked to add that script into upstream's 
test plan for delivering next powerd.

Dogfooding showed similar results than previously, meaning that a lot of 
crashers and nothing else apart from the known blockers.


#266:
- got the powerd fix
- system-settings got the click update integration into their update UI
- indicator-power change to be compatible with this new powerd
- other changes from the platform (glib, dconf, theme…)

-> tests are currently running. We can already confirm that the screen 
is unlocking during the tests. :)


* On the promotion blocker list, here is where we currently stands at:
** gallery shows blank dates and photos (Bill)
https://bugs.launchpad.net/ubuntu/+source/gallery-app/+bug/1296634
-> the debian package is landing, we need a click package release then.

** Music/Calendar bug (Qt event look stopping when the device is 
suspend) (Saviq)
Discussion is happening and work is ongoing while discussing with 
upstream. Was set as a blocker while whitelisted on the previous 
promotion blocker.

** Numerous new telephony-service-indicator crashes (Bill)
http://ci.ubuntu.com/smokeng/trusty/touch/mako/262:20140326:20140304/7386/dialer_app/ 
(and on other tests)
-> Some fixes are landing as we speak.

** Multiple crashes on images since the scope transition (Saviq/Thomas)
https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1297240 (Saviq)
-> Upstream is able to reproduce some now and some fix is under way.
It seems as well that any application crashing can make Unity8 and other 
services crashing. For this chain of issues, Dave opened:
https://bugs.launchpad.net/ubuntu/+source/mediascanner2/+bug/1298931 
(Thomas)
https://bugs.launchpad.net/ubuntu/+source/mediascanner/+bug/1298950 (Thomas)
https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1298952 (Saviq)
-> seems as well the mediaplayer crash is back… but Dave will expand.


* Issues not impacting on the user experience or whitelisted by QA as 
not a promotion blocker:

QA is going to track the previously whitelisted ones (removed from the 
list).

** qtubuntu crash causing the dialer app to crash (only during tests) 
(Kevin)
https://bugs.launchpad.net/ubuntu/+source/qtubuntu/+bug/1297900
-> seems to only impact on tests, when the application isn't started by 
upstart.


On a finale note on the CI status:
Some people would have noticed that the CI Train jenkins got killed 
multiple times by the OOM. The machine went high on memory as jobs were 
piling up (the archive had a lot of builds and powerpc was taking time 
to catch up, as more jobs running there) and 1 GB wasn't enough. As we 
can't add more RAM dynamically on this machine, after the 4th outage 
within 2 hours, we decided to add as a band-aid some swap file to it. It 
was a coincidence that the discussion on finishing the move to prodstack 
(with more memory) happened some hours before the outage… (and the move 
should hopefully happen next week).

The outage was mostly transparent to most users: we rerun the jobs and 
reconciliate the status ourselves. Hower, it chained to an out of disk 
space issue as a consequence. As most of the jobs were just monitoring 
launchpad PPA builders, no time was lost apart from few jobs which 
needed to be built with new content which were delayed in their build or 
landing.

Due to the builders catching up and the high number of requests (and 
some being in a silo without movement for days), we had to play with the 
silo assignment strategy, getting no buffer at some point and flushing 
duplicates. Thanks to that, all requests are now fulfilled!

Cheers,
Didier