Electrolysis/Meetings/2016-03-17: Difference between revisions

← Older edit

Electrolysis/Meetings/2016-03-17 (view source)

Revision as of 17:31, 17 March 2016

5,636 bytes added , 17 March 2016

http://www.unbiased.name/ff/e10scrashes/

Jmathies

Confirmed users

1,982

edits

@@ Line 14: / Line 14: @@
 == mconley ==
-* TBD
+* Back in TO after travels
+* '''Perf P1''' {{bug|1226564}} - [e10s] CACHE V2 telemetry probe regressions
+** [https://bugzilla.mozilla.org/show_bug.cgi?id=1226564#c20 Posted my analysis here.]
+* '''Perf P2''' {{bug|1174770}} - tpaint regressions (3%-33%) in e10s mode compared to non-e10s mode
+** I'm currently trying to confirm that we're measuring the right thing with MozAfterPaint, because I'm not convinced. Coordinating with mstange who is doing similar work in {{bug|1254628}}
+* {{bug|1256468}} - [e10s] Add-on ReloadEvery doesn't refresh (unsafe CPOW usage in RemoteAddonsParent.jsm and webNavigation.sessionHistory is undefined)
+** One of our shims was slightly broken, so I fixed it and added a test. Will request uplift soon.
+* {{bug|1255511}} - Onbeforeunload dialogs appear twice for each tab when I exit/restart Firefox
+** Landed on central. Will request uplift soon.
+* {{bug|1255578}} - "winData is undefined(TypeError)" error on getWindowState called during window unload event
+** Working with an add-on author to try to find a workaround for breakage due to the CPOW-less async window flushing stuff
 == felipe ==
-* todo
+* Running first beta46 experiment
+* Planning for the next phase of tests on beta46
+** Improvements to the staged rollout code (tracked at {{bug|1257265}})
+** Test-plan written
+* Many test reviews
+== gabor ==
+* {{bug|1247420}} - (m9 1247420 - Hovering over a select popup inside ul elements breaks hover css selector on list elements) - regression from Bug 1083365, I spent a lot of time reading hover state management code and debugged this for a couple of days, starting to understand slowly. Discussing the results with Smaug. I need some more time...
+* {{bug|1254081}} - (p1 perf - tabpaint regression on XP and linux) - trying to figure out why is profiling not working on this test...
+* {{bug|1225160}} - (crash in JSAutoCompartment::JSAutoCompartment | mozJSComponentLoader::ImportInto) - this is unrelated to e10s but... is the browser toolbox debugger not function with e10s really??
+== mrbkap ==
+* {{bug|1254856}} - sending cookies for flash-initiated requests when 3rd party cookies disabled broke - fixed and wrote a mochitest.
+* {{bug|1253307}} - opening web protocol handler in a new tab doesn't update the UI correctly - fixed and enabled Florian's mochitest for e10s.
+* {{bug|1140115}} - topcrash in ChildImpl::OpenProtocolOnMainThread - this seems to have gone completely away, the only crashes I can find are on beta and there are only 18 of them.
+* More test shuffling/reviewing/investigating (Ehsan has been knocking over my review queue with test patches, he rocks!).
+== RyanVM ==
+* Ash is up and running, chasing down some permafails still
+* Draft test plan is under review
+* Discussed addons testing w/ Kev, Jorge, and Andy
+==elan==
+*Clarity on HB Survey situation: {{Bug|1223895}}
+*(WIP)Trackback schedule to organize all the details for Beta 47: https://docs.google.com/spreadsheets/d/101ra3HjqyufjP79JkGa4e0RyHPLVlJXS43rXtKuoJw4/edit#gid=0. This week's focus needs to be getting the second phase of current experiment ready for next week (Felipe).
+== gw280 ==
+* {{bug|1114647}} - content process rename. feedback is in, Jeff managed to get a BD contact for Panda AV and they're going to whitelist the new executable names. I have some minor review comments to address then we can probably land this on nightly. Will coordinate with Jeff to minimise disruption when it lands.
+* {{bug|1176019}} - retain layers for background tabs. have a WIP patch.
+* {{bug|1255841}} - hidden select elements get selected. I'm pretty sure I know what the issue is here, and I will have a patch today.
+* {{bug|1213432}} - D3D11 WARP perf. Spoke to jrmuizel about this. Basic layers is much faster than WARP for scrolling, and Milan's concern was that we need WARP to have decent video performance. {{bug|1254877}} was filed to investigate WARP video performance, and it looks like Basic is faster than WARP for video as well. I think we can make a very strong case to not support WARP.
+==aklotz==
+* {{bug|1256943}} - Fixed BLOCKED_ON_PLUGIN_MODULE_INIT_MS in e10s chrome, awaiting uplift
+* a11y+e10s - Ongoing exploratory work on how well COM/RPC security plays with sandboxed processes.
+* a11y+e10s standing meeting set up for Wednesdays at 10AM Pacific, 1PM Eastern, biweekly, starting next Wednesday, 23 March
+* {{bug|1256541}} - Exception-time crash annotations using incorrect paths when MOZ_CONTENT_SANDBOX is not defined - preventing OOM annotation uplift, should be up for review today if my try tests finally ran
+== bsmedberg ==
+* e10s experiment cohort usage: https://sql.telemetry.mozilla.org:443/embed/query/69/visualization/109?api_key=7adc64d1a4ad7d24ba95ba226e2a0fb0987770c5
+** surprised by the size of the "first-session" cohorts
+* e10s experiment crash rates: https://sql.telemetry.mozilla.org:443/embed/query/69/visualization/110?api_key=7adc64d1a4ad7d24ba95ba226e2a0fb0987770c5
+** look only at "control-no-addons" and "experiment-no-addons"
+** initial results: e10s is at 108, non-e10s is at 91 (main+content)
+** plugins crashes are much higher but I don't believe it yet -- expect some data error
+** topcrash lists are showing memory issues even more than usual
+== cpeterson ==
+* [https://wiki.mozilla.org/Electrolysis/Release_Criteria#Jank Jank], [https://wiki.mozilla.org/Electrolysis/Release_Criteria#Page_Load Page Load], and [https://wiki.mozilla.org/Electrolysis/Release_Criteria#Startup.2FShutdown_Time Startup/Shutdown Time] release criteria look green.
+== Roundtable ==
+* https://wiki.mozilla.org/Electrolysis/Release_Criteria#UI_Smoothness
+* http://www.unbiased.name/ff/e10scrashes/