Electrolysis/Meetings/2015-02-26

gw280

Have a working patch for bug 1066531 (tab switching sluggish), need to run it through try and going to go through it with mconley later to try and get it landed.
Investigated issues regarding OMTC, and concluded that we are now covering all cases with OMTC and e10s interaction (ie - if OMTC is not available we don't allow e10s).
Going to be looking at the imglib security bug today and tomorrow.

== ejpbruel == (devtools)

 - This is purely a performance issue, and would take a long time to fix.
 - Propose moving this to M7 or higher

 - This bug was newly reported, but not confirmed yet.

 - Haven't really had time to look into this. Can either gabor or mcmanus take a look at this?

read a lot of interposition code
Bug 1072350 - pdfjs events - cleaned up patch, created a pull request on github, still waiting for review
Bug 1126014 - domainPolicy - invetigated, suggested a different approach in the bug
Bug 1134006 - unexpected CPOW traffic - tried helping out Dave but got stuck with it
addon exceptions: https://etherpad.mozilla.org/AqBjvKDSjU

- cathegories for bugs
- requested new data
- made some progress with some of them already

(+ small web components related fixup)

(I might be late for the meeting today)

Bug 1074971 m5 - don't send some shortcut keys to content - work in progress, incoming patch
Bug 1079665 m6 - findbar focuscontent - need to update patch with Neil's feedback
Bug 903016 m6 - paused this one to focus on 1074971 which is a m5. Will resume it shortly

M5 - bug 1077168 - Installing an app on marketplace fails with e10s
- Fixed and landed
M5 - bug 1066812 - e10s: child process continues running if parent process crashes
- This has been a real pain in the ass. I finally got IPC logging out of the child process - it turns out that, at least on Windows XP, the sandbox prevents logging to stdout, so that's a thing.
- The IPC logging has me chasing down 2 hypotheses on one of the oranges that got me backed out. Details: https://bugzilla.mozilla.org/show_bug.cgi?id=1066812#c24
- This is slow going. Most of the time is spent waiting on try builds. I've got a loaner machine to use now, so that'll hopefully (?) speed things up some.
M5 bug 1090448 - Make e10s printing work on Linux
- Have more review feedback from karlt that I'm currently addressing.
M5 bug 1090439 - PPrinting calls from child to parent via ShowProgress and ShowPrintDialog should not be sync
- jimm and I have agreed to punt on trying to find a way around spinning the nested event loop in the child for now.
- I have a tentative patch up that I need to test more.
bug 1136855 - Print settings are not saved from print job to print job
- I caused this regression with my e10s work for printing, and we just shipped this regression in 36. :(
- I've landed a fix in Nightly, and requested uplift. Also seeing if we're going to spin a point release for 36, which I can pack it into.
- Writing a regression test.

bug 1132874 (content process aborts in PPluginWidget::Msg_ParentShutdown - fix landed which didn't solve the problem. reopened and a new patch is currently up for review.
bug 1082510 (Select drop down issue with click-drag) - under review.
bug 1127794 (Windowed plugin origin doesn't always match rendered content) - wip
bug 1105891 (closing tab hangs firefox) - blocked on issues with the profiles the reporter has generated.

Landed nullprincipal cookie crash bug 1087646.
Came up with a plan of action for showModalDialog bug 1137025 bug 1077002 and bug 981796.
Up next: Landing showModalDialog telemetry, trying to finish my m5 list.

bug 1075670 - [e10s] event.screenX and event.screenY is wrong
- waiting patch approval
- previously backed out for bc1 test and B2G issues
bug 1124408 - [e10s] Crash in [ChildView keyDown:] when using password manager
- landed
bug 1125325 - Web content is not rescaled when OS X moves browser window from HiDPI laptop display to external display
- Looking for a reliable repro.
- cpeterson sees it a lot with the same hardware
bug 1081691 - [e10s] 'mouseover' and 'mouseout' events don't work correctly for <browser> element or remote <browser> element
- started
bug 1130435 - [e10s] Mouse cursor doesn't disappear over Flash object when it should. Works in non-e10s.
- patch awaiting approval