Electrolysis/Meetings/2016-04-14

« previous week | index | next week »

mconley

• Perf P2 bug 1174770 - tpaint regressions (3%-8%) in e10s mode compared to non-e10s mode
  • bug 1262946 - Don't focus the initial browser in a new window until it has been painted
    • This had a really great impact on tpaint across the board - ignore the Windows XP stuff, since that's kinda been tainted by the disabling of e10s stuff that happened recently.
  • Really close to the 5% threshold now. cpeterson has given his opinion that we should move on.
    • Hoping to drive towards sign-off of all UI responsiveness borderline regressions, leaving only tps
• Perf P1 bug 1186585 - 19.66% - 44.84% tps talos regression with e10s enabled and APZ disabled
  • I'm reasonably certain the test is flawed in its measurement, and is favouring non-e10s. See my arguments in bug 1261152
  • Currently in the process of re-writing part of tps to be more accurate. Initial results look promising, hoping to have solid data from a try push tomorrow.
• bug 1263887 - [e10s] The dropdown selector is not working properly.
  • We triaged this last Tuesday, and I was asked to do investigation. The problem is that script on the page is mutating the <select> dropdown after it has opened.
• Fixed regression bug 1260461 - Firefox fails to exit on system shutdown
  • This has now been uplifted to aurora / beta. Waiting to hear back about ESR.
• Fixed regression bug 1255773 - Page Info - Media is incomplete

billm

• Worked on fallout from bug 1235633 (perf problems and jemalloc4 regression).
• Continuing to work on OOM issues. May move to bug 1262671. Will talk to mccr8 first.

felipe

• A lot of reviews and feedback
• Maintaining the tests spreadsheet
• m9 bug 1262332 - Drop-down list does not appear in the right place - In progress
• bug 1264345 - Do not disqualify users of other experiments - Landed and uplifted
• bug 1264437 - Improve telemetry for disqualified users
• A number of uplifts

gabor

• bug 1247420 - (m9 1255138 - Resizing windows from JS is not supported in e10s mode) - Mostly completed, it needs some tuning/testing. There is an offset issue with sizeToContent which I might file a followup for...
• bug 1263953 - (Reduce the growth rate of Pickle) - Filed a quick patch for feedback, read some related bugs.
• bug 1264383 - (Add-on multiprocess compatible shims stops TabSelect event listeners being removed from the tab container) - shims bug, filed a bug for it, probably should be uplifted...
• bug 1099406 - (regression in SIMPLE_MEASURES_FIRSTLOADURI) - probe was completly broken, I would not bother fixing it unless there is a good reason for it.
• bug 1261299 - (Mac OS X Services is not available for e10s tab content) - looked into it briefly...
• lots of reviews, mostly addons

chutten

• bug 1261362 PLUGIN_JANK analysis is complete, :aklotz has signed off.
• Having conversations with :cpeterson, :erahm, :rvitillo about MEMORY_TOTAL and how ipynb analyses can look different from t.m.o dashboards.

RyanVM

• The draft e10s test plan is ready for review.
• Recent improvements in platform-level automated test coverage.

haik

• bug 1258925 ([e10s] Browser window is resized when click) - Have new patch out based on Gabor's patch, going through code review

kanru

• Perf P1 Investigating bug 1259641 [e10s] MediaFeatureValuesChangedAllDocuments takes tons of time when resizing e10s window
• Investigating bug 1257387 crash in OOM | unknown | js::AutoEnterOOMUnsafeRegion::crash |js::TenuringTracer::moveToTenured
• Investigating bug 1262661 Many large SessionStore:update messages

ting

• bug 1259512 significantly higher rates of OOM crashes in the content process of Firefox with e10s than in the main process of non-e10s
  • checked why minidump-memorylist does not work
• bug 1260736 3.71 - 5.47% damp / tp5o (windows7-32) regression on push 4de1094b41b2 for interruptible reflow (Tue Mar 29 2016)
  • updated PeekMessages() and confirmed it helped with the regression, recevied r+, about landing
• Went throguh bug 1005844, bug 1005849, bug 1219914 and got the idea, ni'd glandium to confirm whether getting jemalloc4 re-enabled is the first step

cervantes

• Filed bug 1264244 to make it easier to analyze crashes of different SystemError values
• Investigating bug 1258317 crash in mozilla::ipc::TransferHandleToProcess (PBackground related)
  • Trying reproducing on windows with 3rd-party security software installed
• Investigating bug 1259480 [e10s] FatalError crash in PContentChild::Read from PContentChild::SendReadPrefsArray

wcpan

• Perf P1 Investigating bug 1259641 [e10s] MediaFeatureValuesChangedAllDocuments takes tons of time when resizing e10s window