From MozillaWiki
Jump to: navigation, search

The Highlights

  • [camd] pushed MozTrap to prod
  • [henrik] - Mozmill: We have our hack work week from today to Friday!
  • [mdas] Documented perf test approach at
  • [dburns] Documented how to get marionette on a build that wasnt enabled with Marionette
  • [griffin] Transition Pulse to new multi-node hardware complete!

The Overview


  • [jhammel] added support in mozrunner to attach a debugger
  • [wlach] added a command-line client for SUTAgent enabled devices to mozdevice. handy for debugging


  • [henrik] We have our hack work week from today to Friday!
  • [jhammel] added support for BROWSER_PATH in mutt so the python tests now work
  • [davehunt] mozmill-ci now sends notification emails to public mailing list Examples
  • [davehunt] mozmill-ci now pre-fetches the remote hg repositories to circumvent downtimes of



  • [glob] work on returning the qa field back to qa (good to go now, trial run prep work)
  • [glob] investigated webservice error management and arecibo integration
  • [glob] investigated sentry and browserid updates
  • [glob] updates to the lost device form
  • [glob] fixed xss issue with guided bug entry
  • [glob+dkl] much fun investigating a "duplicate dependencies" issue
  • [glob+dkl] release tracking work
  • [glob+dkl] general administration


  • [wlach] Dashboard for LG G2X results back up and running. Was down due to power issues, fixed by moving device communication to tcp/ip and plugging devices into wall.
  • [wlach] Support for the Galaxy Nexus trucking along. Working locally, just need a few tweaks in Mountain View and we'll have Galaxy Nexus results coming in.
  • [wlach] Moved Eideticker tests to a private third party repository, so they can be revision controlled without copyright issues contaminating the main Eideticker repo


  • No updates, aside from decision to drop histogram and evangelism goals in favour of mobile-blockers dashboard and further AutoPhone work.


  • [edmorley] Dealing with tree closure for buildbot DB issues bug 754813.
  • [edmorley] Helping Gerv interpret several try result failures for the tree-wide MPL2 change + planning of its landing.
  • [edmorley] Dealing with tree closure due to tests not being scheduled bug 754885.
  • [edmorley] Tracking down cause of 5 min zero change windows builds. Filed bug 755684 and some dependants.
  • [edmorley] Finally got a local TBPL dev environment working after windows-specific python library issues + VirtualBox networking bugs :-(
  • [edmorley] Fixed bug 755845 to improve the TBPL Vagrant/puppet config so a data import is run as part of setup.
  • [edmorley] Fixed bug 756209 for protecting the local php/config.php since it contains TBPL passwords.
  • [edmorley] Started work on improving TBPL dev environment docs, particularly for Windows.
  • [edmorley] Have now been added as a superuser on the new sheriff duty site, so can start testing.
  • [edmorley] Issues with clobberer not working if you try to clobber everything at once, filed bug 756532.

War on Orange

  • [mcote] - Designing Orange Seed; found appropriate ES query for the necessary data.
  • [edmorley] Disabled the test in bug 668716 on OS X for too many intermittent passes.

Signal From Noise

  • [jmaher] - identified 26 pages that fit a statistical model, working with them to see what we can do for moving forward
  • [jmaher] - chmanchester is leading the effort on cleaning up tp5 pages that access the internets. Step 1 is to identify exactly what is references and where it is referenced from, Step2 is to programatically clean it up (which might be next to impossible)
  • [jmaher] - we can use a*team staging server for posting to datazilla now, time for queueing up the runs


  • [jhammel] added ability to do per-test filters (also for Signal from Noise)

Mobile Automation

  • [ted] Finished up Fennec build time improvements project


  • [dburns] Documented how to get marionette on a build that wasnt enabled with Marionette
  • [dburns] Trying to get a workflow going for Open Web Apps testing, raising Marionette bugs as I hit them
  • [dburns] Bug fixes for Marionette client
  • [jgriffin] Trying to get buildserver functional; running into lots of weird build problems
  • [mdas & jgriffin] Misc bug fixes to support WebAPI and Apps testing
  • [mdas] Documented perf test approach at

Web QA


New Graphserver (Datazilla)

  • [carljm] Added test runner and test coverage measurement
  • [carljm] Worked on logging of JSON-parsing errors; in-progress
  • [carljm] Vendored dependencies for IT and wrote script for easy updating
  • [carljm] Converted config to use IT's preferred style of Django settings, no required env vars
  • [carljm] Converted from FastCGI to WSGI, with sample Apache/mod_wsgi config

autophone/Noah's Ark

  • [mcote] - A bunch of fixes and improvements, as usual mostly to stability, error detection/recovery, and diagnostics. Started setting up a staging server.


  • [mcote] - No real work but thinking about how to package it for easy setup in a dev environment.



  • Transition to new multi-node hardware complete!
  • Other goodies like SSL on the back burner for now until I have more time.


  • [carljm] - Fixed environment cascade bug.
  • [camd] - pushed to prod!



Upcoming Events

Round Table

Take notes here and copy and paste them to this section afterwards.

  • [henrik] Final decision about the Mozbase check-in policy?
    • Rebasing vs. 'merge --squash' for patches (no magic github merge button, but a local testrun)
    • Who checks code in? Reviewers or patch author himself?
    • Commit message like "Bug 123456: the title of the bug (if applicable);r=whoever"


  • [henrik]
    • Checkin policy "merge --squash"
      • dburns: against github button. don't like rebasing, prefer use squash.
      • clint: always only has 1 commit on feature branch to keep things straight.
      • prefer squash over rebase.
    • Who should check in code?
      • patch authors
    • commit messages
      • Stick to what we have in the firefox repos

Holidays and Trips

US is out next Monday, Memorial day 5/28. Germany, too. Britain out week after. Canadians decide if they want a meeting next Monday 5/28

Next week MC

  • TBD - Meeting may not happen since 3/4 of team is out next Monday.