Electrolysis/Meetings/2014-11-13

From MozillaWiki
Jump to: navigation, search


« previous week | index | next week »

handyman

  • bug 1044182 - Plugin focus is broken in e10s on Mac
    • working on tests
    • will probably land without (these are very edge-casey tests)
  • bug 1075670 - [e10s] event.screenX and event.screenY is wrong
    • patch posted - uses mac plugin work
    • issue is that, in content proc, screen (0,0) is assumed tab origin. (So far) plugins and user-js need correct coordinates (but the engine wants (0,0)-relative). This patch simply reports actual screen coordinates for events, which could still be wrong (if the engine somehow uses it). The setup is overly precarious.
  • bug 1094328 - [e10s] Detaching XUL tabs (about:newtab, about:preferences, about:addons) leaves a ghost tab in the old window, and opens a blank tab in the new window
    • remote vs non-remote tab issue
    • re-triaged yesterday. m5.
  • bug 1076820 - Redraw on resize performance for content is pretty bad with e10s
    • dunn

cpeterson

jimm

  • PTO today due to bad internet. On and off irc today. Tech was out yesterday, made things worse. next tech will be out Saturday.
  • catching up on review requests / need infos. Sorry for the delays this week.
  • bug 669200 (M4, windowed plugins for win/linux) - landed
  • mac update - https://wiki.mozilla.org/Electrolysis/plugins#mac_update
    • follow ups:
      • bug 1095754 - sync window updates over the compositor (wip)
      • bug 1098055 - enable windows deferred message protection for PContent (patch posted)
      • bug 1098057 - potential deadlock risk, need to investigate.
  • bug 1089008 - enable e10s more broadly on windows (patch posted)

allison

  • bug 1080212 - [e10s] Add Telemetry probe for js exceptions that occur in the addon js compartment
    • worked with gfreitzche to expose keyed histogram to cpp
    • f+ from mccr8 for the js-gecko callback code
    • f? billm for the most recent changes to jsexn.cpp (original code had f+)
    • gfretziche is ok with the telemetry touchpoint in the patch(no formal review)
    • Rountable Question: Do we want the addonId as part of the keyed histogram?
  • last week tl;dr closed 1010737, did 6 reviews, and logged errors thrown in the js addon compartment
    • see 2014-11-06's update for more information

mrbkap

  • bug 1088457 - fixed tracking protection with e10s
  • Speaking to people around the office, trying to figure out what's causing pain.

mconley

  • M4 bug 1090444 - The Printing Progress dialog does not show progress or close when opened from the child process
    • Fixed and landed
    • Causing ~3% Dromaeo DOM regression... ? bug 1098244... really strange. Feeling stabby.
  • M4 bug 1008435 - [e10s] Port the built-in Gecko profiler to e10s
    • I've got the infrastructure in place - the content process is now gathering profiles and sending them up to the parent! \o/
    • The parent process is receiving the profiles, but it needs to know how to deal with them. Right now, the end result is empty samples / markers for both the GeckoMain and Content process profiles, which is thoroughly anti-climactic.
  • bug 984821 - Intermittent browser_CTP_iframe.js | Test 1, Waited too long for the overlay to become invisible.
    • Dealing with this between builds

billm

  • Fixed some plugin hangs
  • Fixing LastPass performance by pre-fetching CPOWs
  • Log warnings about using CPOWs outside event handlers
  • Investigating CPOW and plugin deadlocks

Roundtable

  • bug 1080212 - [e10s] Add Telemetry probe for js exceptions that occur in the addon js compartment
    • Roundtable Question: Do we want the addonId as part of the keyed histogram?
      • agree error::addonid, followup bug to get filename&line number, privacy-safe