Status

jmaher

• Identifying remaining tasks to get total automation failure rate <5%
  • bug 781162 - Resolve the twisted errors
  • bug 781159 - Mark devices as dead if they timeout

• Panda Board
  • updated bootargs for different resolution (720p). Not output on HDMI, but screencap works now
  • mochitests/talos/reftest all run, still working on errors

• bug 781179 - 2 pass build for optimized mobile coming up on my radar
• ATM Beagle boards are P2

• bug 737961 - Blocked on [reftest resolution]

mcote

• Finished first version of new SUTAgent design.
• Started implementation--have a hello-world binary building and running on my Galaxy SII.

jgriffin

• Working on adding Gaia unit and Gaia integration tests to the B2G CI
• Working on expanding the set of mochitests we're currently running
• Need to write some sheriffing docs for B2G so I can share the joy

mdas

• Got a version of mouse support for the B2G pandaboard, so I can now get the rest of the eideticker tests going
• Submitted a patch to get current eideticker+b2g pandaboard work pushed to main repository. This adds support for running a basic html/js test.

ahal

• Reftest desktop running reftests against b2g nightly builds reported to autolog: http://brasstacks.mozilla.com/autolog/?tree=b2g&source=autolog
• Patch for reftest b2g infrastructure up for review, bug 770490
• Starting work on getting reftests running on pandaboards (still in the setup/flashing phase)
• Wrote getb2g to download latest b2g nightlies (pip install getb2g, https://github.com/ahal/getb2g)

TODO

• Modify reftests to use iframe mozbrowser instead of browser remote, bug 778072
• Figure out "program error managing snapshot" problem, bug 778725
• Figure out why remote xul recently got disabled and how to enable it
• Figure out reftest resolution problem

mihnea

ctalbert

Make all your tracking bugs block this uber tracking bug:

• bug mobile-automation (bug 781041)
• Using this to generate a scrum burndown using this scrumbugs tool

wlach

• Working on a pandaboard burnin script based on Autophone
  • Seems to work ok on phone, but long panda reboots seem to be causing problems
  • Wondering what the best way of getting this standing up against a network of pandas - Autophone's architecture seems to assume it's a persistent daemon, but we really just want it to come up, run 500 or so tests, then shut down.
• Some work on getting ARMv6 device working with Eideticker (in non-HDMI capture mode). Will finish work and post preliminary results after this burn in test finished.
• Next priority is getting better Fennec startup tests going with Eideticker, to replace S1S2. Have some preliminary startup measurement code but nothing finished yet.

Callek

• (todo add all bug #'s, including filing bugs)
• Worked with ateam and got "tegras rebooting" bug fixed

Linux Foopy

• Got the issue I was hitting with mochitests in xpcshell failing fixed with glandium's help. (Was missing $HOME and causes crash)
• Finish writing all patches for buildbot/puppet for Linux foopy and get them reviewed

More Tegras

• Most tegras filled
• Need to fill in releng system lists for the tegras
• Need to test PDU entries for all new tegras
• Blocked on Linux Foopy for moving these to production
• Likely need a new buildbot master for the lot.

bc

• working on method of using standard updater and su binaries in a recovery mode updater to create a more universal minimal rooting technique.
• will send out request for data on various phones to everyone.

Issues of The Week

Round Table

• [jmaher] - Mozharness support? foopie-less tegra support?
• [jmaher] - What work needs to be shared between fennec automation and b2g for pandas?
• [jmaher] - b2g emulator?
• [jmaher] - how do we reduce our differences between autophone, eideticker, b2g, buildbot? is mozharness the solution? The more we reuse code, the less documenting of stacks and debugging random things.