EngineeringProductivity/Meetings/2012-02-13
From MozillaWiki
The Overview
- bugzilla
- [dkl] Implemented AutoLand BMO extension for Lukas Blakk for integrating with Auto Try system (under review)
- [glob] more push work (https://github.com/globau/bugzilla-push)
- [glob] tweaked bmo's memory usage, no longer recompiling perl with each request
- [glob] minor bmo updates (added 'cc count' column to bug lists, added sort-by-bug to user-activity-report)
- [glob+dkl] administrative work, as always :)
- bughunter
- [bc] Presented to the security and qa teams on friday.
- [jeads] Fixed some more bugs
- [jeads] Gave two presentations on bughunter on Mountain View (QA) and Santa Cruz (security). Good feedback coming in.
- marionette (https://developer.mozilla.org/en/Marionette)
- [mdas] Heavily updated documentation and comments
- [mdas] Updated it so it loads as a component only if it is enabled via a build-flag and a pref is set
- [mdas] added a 'make test' command to make it easier for the b2g guys to start testing with it.
- [mdas] Added logging, so test users can add their own log messages (https://developer.mozilla.org/en/Marionette/Marionette#Debugging)
- [jgriffon] Many small bug fixes and improvements to make writing tests and running them easier and more reliable.
- [jgriffon] Worked with QA to write some mozTelephony tests.
- [jgriffon] Security review next Monday.
- [jgriffon] Have a local copy of Jenkins running B2G builds and Marionette tests; will set up a production copy on EC2 next week, then work on re-purposing the existing gitbuilder machine as a Jenkins node that will perform automatic merges from m-c into B2G.
- pulse
- [jgriffon] Work on setting up Nagios monitoring for pulse.
- [jgriffon] Work on developing scripts to automatically identify mis-behaving pulse queues, which historically seem to have been the biggest source of problems.
- [jgriffon] Trying to identify a better way of publishing buildbot messages to pulse that won't break consumers by unpredictable changes.
- fennec ARMv6 (bug 697205)
- [ted] got an armv6 build successfully up and running with some local patches, just need to get things landed and then it's in RelEng's hands.
- signal from noise
- [jmaher] starting to get clarity on limitations of how to run tests
- [jmaher] first side by side staging is done, will reconfig early in the week
- [jmaher+jhammel] lots of churn on cleaning up talos configuration and result parsing
- [jhammel] Added user configurable data filters to talos (bug 723569). Interim step to moving all data processing out of Talos.
- [jeads] Some headway in converting some bughunter javascript to work with s4n
- [jeads] Got 4 hosted VM's from IT for use with s4n
- graphserver
- [jhammel] talked with armen, jeads, and jmaher about plans for graphserver-next. Agreed on some fundamental architecture decisions:
- We should be sending all data to graphsever
- There should be a configurable way to apply filters and drill in to the data (jeads is working on the UI)
- The schema needs to be reworked to both make sense and be more flexible
- [jhammel] talked with armen, jeads, and jmaher about plans for graphserver-next. Agreed on some fundamental architecture decisions:
- mobile automation
- [jmaher] robocop is landed and live on tbpl
- [jmaher] working on updating the foopies (mobile device masters) with updated talos and toolchains
- [jmaher] will turn on talos robocop tests hopefully next week
- [jmaher+wlach] evaluating panda boards, working on getting builds and installation steps defined (http://pandaboard.org)
- [mcote] made some much nicer looking graphs (http://10.250.2.223:8100/#/rawfennecstartup)
- [mcote] updated the raw-fennec-start web UI. After some failed attempts at rooting phones, got the real-phone automation system ("autophone") running and making improvements on it.
- Addon compatible by default automation (https://wiki.mozilla.org/Auto-tools/Projects/Compatibile_By_Default_Automation):
- [wlach] no change
- Eideticker (https://wiki.mozilla.org/Project_Eideticker):
- [wlach] Some more work on Eideticker dashboard but project temporarily on hold while we sort out the issues with the card in mountain view.
- [wlach] Spent time working to make MozRunner work on Android (https://github.com/wlach/mozbase/tree/mozrunner-remote/mozrunner/mozrunner)
- Ateam Services
- [bc] Waiting on network configuration changes in Phoenix.
- jetperf harness (http://k0s.org/mozilla/hg/jetperf)
- [jhammel] In beta state but good enough for production.
- [jhammel] next steps: adding to buildbot and graphserver (though no concrete work there yet).
- peptest
- [mcote] Prerequisite mozbase changes landed. Submitting a patch for peptest itself shortly.
Upcoming Events
All times Pacific Time, click on link to find your timezone
- Platform Meeting (all platform engineers) Tuesday at 11AM, Conf # 8605.
- Mobile Test Automation Wednesday at 10:30AM #304
- Marionette/Eideticker Meeting Thursday at 10AM Conf # 304
- Signal from Noise Thursday at 11AM #343
Round Table
- Next week's MC?
- Next week is a holiday. So jeads will catch us up at the next meeting, Feb 27.
- We're going to try to pin Jgriffin down after that for the one following.