Linux Archive

Linux Archive (
-   Ubuntu Mobile and Embedded (
-   -   Mobile Team Weekly Summary, 2010-01-12 (

Jamie Bennett 01-13-2010 07:48 PM

Mobile Team Weekly Summary, 2010-01-12

The weekly activity reports and meeting minutes from the mobile team are
now available. They can be viewed online at the following links and are
reproduced in full below.

Activity Reports: Reports

Meeting Minutes:

= Weekly Reports =
== Jamie Bennett (JamieBennett) ==
=== This week ===
* More work on 2D launcher dependencies. Embryo needs a commitment
from the Mobile team if it is to be moved to main. As Embryo is only
used by edje it was suggested that maybe we combine the two packages?
Worked with EFL upstream to discuss code issues. Some progress.
* More MIR's produced, this time for [[Bug: 502912|efreet]] and [[Bug:
502909|edbus]], as of writing they are yet to be picked up.
* Discounted Moblin's media player as a possible candidate for the
alternate media player spec. It
clutter]] which isn't appropriate for us.
* [[|Analysed
bootcharts]] from the ARM iMX51 image. Seems initial investigations were
right and that the slowness is all in Casper. Combined it with the
Casper information under the umbrella page
* Made a start on the implementation side of the Casper spec. Talked
with interested parties to discuss a solution that would not only work
but would be the correct solution. Committed
code]] which isn't working yet but shows the idea behind the fix. Left
this to finish off Alpha-2 targets specs.
* Chased a GDM issue on daily builds. Something crept in but was fixed
just as quickly.
* Bit of seed work for 2D launcher, hit a launchpad data corruption
bug. Worked with the launchpad team to track it down but to no conclusion.

=== Future ===
* Continue on getting a working implementation to fix the
debconf-communicate issues in Casper.
* Get 2D launcher dependencies finalised one way or another.
* Investigate kernel patches that may be worth backporting.
* Compare bootcharts from i386 and iMX51 to see if there is any arch
specific slowness.

== Tobin Davis ==
* Reloaded Babbage 3 image from Freescale to use as a benchmark and to
see if certain hardware issues were working. Discovered a number of
issues with current images; audio issues, video refresh issues, etc.
Most already have bugs filed against them.
* Finished porting the T2C test suite. Biggest hangup was in
packaging, but that has been passed off to Emmet.
* Purchased another SATA drive so that I can do installation testing
on dove without disrupting the development image.
* Switched focus to image testing. Dove images broken to the point of
not booting.
* As part of swith to full image testing, I have been asked to find
the obscure issues. First issue I filed affects all Ubuntu
installations; Bug 504116 apport-collect returns incorrect name with -h
option. Simple bugs like this can be easily fixed and released, making
Ubuntu look more like a highly refined distribution, not a throw
together distro.

=== Future ===
* More deep testing for issues not normally detected or brushed off
with work-arounds.
* Finish any remaining tasks with the mobile-lucid-arm-lib-tests spec.

== Emmet Hikory ==
* Packaging of LSB tests
* tet-harness packaged, pushed to lp:~persia/lsb-arm-port/tet-harness
* t2c-framework packaged, tarball unsuitable for redistribution,
packaging at lp:~persia/t2c-framework
* initial work done on t2c-desktop

== Steve Kowalik ==
* Test and re-test the ubuntu-netbook image, in preperation for Alpha 2.
* Investigate why netbook-launcher is constantly restarting itself,
have asac upload a patch.
* A bunch of NBS work.

== Paul Larson ==
=== This week ===
* Pairwise install testing blueprint
* documented test cases on the qa wiki
* Discussed next steps with Ara
* Requested access to iso tracker DB so that I can insert tests
* This blueprint is done, except for adding the virtual milestones
to begin testing 'after' A2 is released. Since it doesn't help
anything to do that now, that bit has been deferred until after A2 is
actually out
* Arm Bringup testing blueprint
* New script to use ldtp to time application startup times
* New tests added
* All complete except for one deferred item (script to gather
additional arm-specific information). The remaining item has been done
to the extent that it can be, but leaving it open because there could be
additional things that make it easier to get some more information in
the A3 timeframe.
* Arm Suspend/Resume Blueprint
* reorganized the table on the wiki page to align with work items
* Discussed the tests, priorities, and implementation ideas with
others on #ubuntu-arm
* assigned priorities, all A2 targeted work items complete
* Updated bug workflow to include new bug tag "daily" for bugs found
during the testing of daily images
* Tested suspend on imx51, still seems to be broken but there are some
upcoming patches that should help with that
=== Future ===
* Look into an issue that is preventing LTP from building properly in
* Look into some new tests
* Testing A2 candidate images
* Start running bringup testing suite, document what is needed to run
it, start posting results
* Create virtual milestones for pairwise testing, put tests into
tracker database
* start on suspend/resume tests

== Alexander Sack ==
=== Blueprints/Planning/Team ===
* ensure progress on alpha-2 items; get alpha-2 status down to trendline
* release team meeting
* work with plars to give suspend-resume spec a boost:
* reviewed prioritization
* discussed ways to test
* discuss lib-test spec with GrueMaster, persia and davidm
* outcome: we will share the work load better on this and assign
specialists for special tasks like packaging etc.
* assigned first batch of ftbfs to team members; peers for ftbfs tasks
are NCommander, dyfet and JamieBennett

=== Chromium ===
* resolve chromium-browser runtime issues by tracking down gcc bug
* two days worth of work on getting chromium license situation in
shape; only a few files left; extended to produce a
compressed dep-5 presentation; current copyright file:
* review and sponsored gyp - buld dependency of chromium.

=== Porting ===
* fix postgresql-8.4 ftbfs - LP: #497331
* uploaded fix for clutk to lucid - LP:#495066
* sponsored d-shlibs, libtool, ocaml for ncommander
* fixed apex build on armel/lucid
* fixed db4.2 ftbfs
* gave back a few packages that failed early in cycle
* firefox/xulrunner upload that includes patch to fix thumb crashes

=== Booting ===
* u-boot experiments and partition issues for images; also, broke my
bbg3 board

=== other team contributions ===
* pushed emergency #fontconfig update to unbreak buildds after defoma
update in lucid
* prepare/QA and upload firefox 3.0.17/3.5.7 and xulrunner security/stability updates for hardy to lucid
* mentor micah on firefox/xulrunner release process
* move xpi location for mozilla langpack production to a place where
ArneGoetje can maintain them.

== OliverGrawert ==
=== DONE ===
* tested new imx51 kernel from latest freescale code drop RC
* updated uboot to latest freescale code drops
* added MIR for uboot
* tried to create image creation script for uboot from proof of
concept code
* found several issues with partitioning and vfat handling of uboot
that prevent to load a kernel image uncorrupted. researched these with
asac with no result yet

=== TODO ===
* start working on the rootstock specs
* prepare and test A2 images for imx51
* find more info about the uboot issues that prevent proper image builds

== Dave Martin (dmart) ==
=== DONE ===
* Initial browser benchmark runs:
* Firefox-3.5 (firefox-3.5 3.5.7+nobinonly-0ubuntu0.9.10.1;
xulrunner1.9; karmic)
* Chromium (chromium-browser; lucid)
* Chromium looks good at the moment: up to ~ 80-90% faster on general
page rendering.
* Caution: firefox-3.5 used for comparison, was old, ARM (not
Thumb-2) and running on Karmic.
* Still need a firefox build to test on lucid: we don't have a
working Thumb-2 build yet (asac working on it).

* List of packages of packages to review for v7/Thumb-2 compatibility
* Published initial lists on the wiki, based on the results of the
source archive grep:

=== TODO ===
* List of packages of packages to review for v7/Thumb-2 compatibility
* Agree with the asac how we work through this list together.

== David Sugar (dyfet) ==
=== Arm ===
* Initial chromium testing on lucid arm
* Porting effort for canola to build on arm
* Initial pass at fails to build for arm
=== Server ===
* Handoff of asterisk blueprint to jmdault
=== Motu ===
* Rebuild of ccscript for package revu
=== Next Week ===
* Focus on failed to build priority bugs and Motu

== Michael Casadevall (NCommander) ==
=== ARM ===
* FTBFS work on armel, resolved ocaml, libtool, d-shlibs, and fontmatrix
* Retried and rescored several failed packages that went bust due
libtool being broken
* Forwarded patches to ocaml upstream
* Investigated non-PIC shared libraries on ARMum
* Dove instability investigation continues
* Assisted JamieBennett with some bash shell scripting in changing
debconf for speeding up casper

=== Next Week ===
* FTBFS work continues
* Thumb2 compatibility work continue
* Dove instability research continues

== Minutes ==
* NCommander was asked to invite the relevant kernel people to the
meeting in the meeting reminder email. He took up the action to do that
next week.
* asac, ogra, and persia has the action item to produce a wiki page
detailing .32 iMX51 kernel backporting information. The page is not yet
complete so they extended the action item to next week.
* There was an action item for ogra, asac and NCommander to blog on
planet about ARM related work. It was discussed that this will happen
when there is some news to report, possibly on the uboot progress ogra
and asac are making.
* NCommander asked about the 2D launcher dependency situation,
JamieBennett informed the team that a resolution had been found and that
the Mobile team are to take ownership of embryo for future bugs. This
unblocks the other MIR dependencies that have been filed. JamieBennett
also mentioned that he had filed bugs to upstream for all concerns that
the MIR process had highlighted. JamieBennett also noted that 2 MIR's
are yet to be reviewed.
* NCommander commented that the burn down chart for Alpha-2
( was looking
good. The team talked about the current short-comings of the script that
produces the burndown chart and asac mentioned that he will go through
the blueprints this week for alpha-3 planning and to clean them up.
* persia invited dmart (Dave Martin from ARM) and others to add their
weekly activities to the report. dmart and ian_brasil indicated that
they would.
* asac asked NCommander what the problem was with alternate images.
NCommander replied that he hasn't managed to install one due to changes
in the archive which tends to happen at this point in the cycle. asac
asked to be kept informed about the situation.
* Ubuntu Liquid Remix was discussed. rbelem indicated that seed work
had started: and
that they are now working on customizing kdm and kwin for mobile
environments. Riddell offered to help with the project and NCommander
added a standing item to discuss Liquid at subsequent meetings.
* The team had a lengthy discussion on how to process the
ARMv7/Thumb-2 compatibility review dmart had produced: dmart was thanked
for his effort so far and asac decided that it would be best if they
talked about the details after the meeting.
* Talk moved to kernel status. cooloney posted that " 2.6.31 Freescale
BSP patches based kernel source was uploaded on Jan 11. Kernel packages
cross compiled was tested on ogra and cooloney hardware. Freescale
updated final BSP is just released. I will try to review the new kernel
patches and try to apply them for testing". ericm posted that "Marvell's
patchset based on v2.6.32.2 has been rebased again on top of lucid,
config files sorted out, and building is OK. There are several issues
though (not specifically related to this patchset) including: unexpected
segfaults and X desktop freezing, still investigating the root cause
with Marvell, and this is holding the patches from being uploaded atm".
ogra asked ericm to subscribe all ARM related kernel bugs to ubuntu-armel.
* Maybe related to kernel issues, NCommander informed the team that
the dove image frequently hangs after boot. ericm noted that this didn't
happen with Lucid Alpha-1. The team thought the crashes were due to X.
NCommander and GrueMaster took the action item to investigate this
further with ericm's help. dmart took up the action to investigate if
ARM had any specific tests for Thumb2 that could be run on dove.
* QA (plars, Gruemaster) reported that the bringup testing blueprint
is done and plars plan to start running through that and posting results
* ogra mentioned that there were still quite a few packages on the
FTBFS list. NCommander indicated that he had fixed a few of them and was
continuing on with the effort along with dyfet.
* ogra informed the team that images hadn't been built due to a
hardware failure on the image builder. IS have been informed and an
image build was in progress again.
* asac said that had failed to build. ogra took the
action to remove it from the seeds if it was posing a problem to Alpha-2
image builds.

Jamie Bennett
Ubuntu Mobile Developer

Ubuntu-mobile mailing list
Modify settings or unsubscribe at:

All times are GMT. The time now is 08:48 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.