Electrolysis/Meetings/2016-09-08

« previous week | index | next week »

mconley

• bug 1298219 - about:sessionrestore shows a crashed tab after force-quitting the machine twice
  • Fix with a regression test landed on central, and should be in today's nightly. Will request uplift to Aurora soon.
• bug 1300411 - user reports excessive load spinning when switching tabs with e10s in 48.0.2 release
  • Several reports of this have caused some concern
  • We've landed a probe in bug 1301104 to figure out how bad the bad spinner cases (>1s) are, and hopefully how frequent
  • bug 1301131 has been filed with a tentative plan for disabling e10s if a user sees long spinners too often
• bug 1287178 - "you have an unsubmitted crash report" even though crash reporter is disabled
  • Adding a "submit always" button. Currently refactoring and adding automated tests.
  • This hooks into bug 1241459 - [e10s] A crashed background tab does not give option to submit, which will use the same notification mechanism
• bug 1255843 - Develop a tool to measure resource usage of each subprocess
  • Rutuja has finished this work, and the tool is available in about:performance
  • Some follow-up bugs have been filed, which she is interested in hacking on! \o/

gsvelto

• Nothing specific to report on the e10s front; I'm sill waiting for all the required machinery to land in order to get client-side stack traces for content process crashes but that's taking longer than expected

shell

• determining impact of file-system sandboxing on add-ons
  • MPC:true or add-ons from last experiment not tested against sandboxing at all
    • hoping for experiment in that area. are there telemetry probes way to separate
    • would hope for against all population - or even 50 add-ons experiment population
• 50 move to MPC:true and MPC:webextensions experiment(~1600 named and more unnamed)