It seems that I cannot run any mochitest with patch in bug 1335987 applied. Running "./mach mochitest --disable-e10s layout/style/test" would crash with the following stack very quick: Assertion failure: recreateContext, at c:/mozilla-source/stylo/layout/base/ServoRestyleManager.cpp:189 #01: mozilla::ServoRestyleManager::RecreateStyleContexts (c:\mozilla-source\stylo\layout\base\servorestylemanager.cpp:256) #02: mozilla::ServoRestyleManager::RecreateStyleContexts (c:\mozilla-source\stylo\layout\base\servorestylemanager.cpp:256) #03: mozilla::ServoRestyleManager::RecreateStyleContexts (c:\mozilla-source\stylo\layout\base\servorestylemanager.cpp:256) #04: mozilla::ServoRestyleManager::RecreateStyleContexts (c:\mozilla-source\stylo\layout\base\servorestylemanager.cpp:256) #05: mozilla::ServoRestyleManager::RecreateStyleContexts (c:\mozilla-source\stylo\layout\base\servorestylemanager.cpp:256) #06: mozilla::ServoRestyleManager::RecreateStyleContexts (c:\mozilla-source\stylo\layout\base\servorestylemanager.cpp:256) #07: mozilla::ServoRestyleManager::RecreateStyleContexts (c:\mozilla-source\stylo\layout\base\servorestylemanager.cpp:256) #08: mozilla::ServoRestyleManager::ProcessPendingRestyles (c:\mozilla-source\stylo\layout\base\servorestylemanager.cpp:361) #09: mozilla::PresShell::FlushPendingNotifications (c:\mozilla-source\stylo\layout\base\presshell.cpp:4173) #10: nsRefreshDriver::Tick (c:\mozilla-source\stylo\layout\base\nsrefreshdriver.cpp:1837) #11: mozilla::RefreshDriverTimer::TickDriver (c:\mozilla-source\stylo\layout\base\nsrefreshdriver.cpp:326) #12: mozilla::RefreshDriverTimer::TickRefreshDrivers (c:\mozilla-source\stylo\layout\base\nsrefreshdriver.cpp:297) #13: mozilla::RefreshDriverTimer::Tick (c:\mozilla-source\stylo\layout\base\nsrefreshdriver.cpp:317) #14: mozilla::VsyncRefreshDriverTimer::RunRefreshDrivers (c:\mozilla-source\stylo\layout\base\nsrefreshdriver.cpp:669) #15: mozilla::VsyncRefreshDriverTimer::RefreshDriverVsyncObserver::TickRefreshDriver (c:\mozilla-source\stylo\layout\base\nsrefreshdriver.cpp:590) #16: mozilla::detail::RunnableMethodImpl<mozilla::VsyncRefreshDriverTimer::RefreshDriverVsyncObserver * __ptr64 const,void (__cdecl mozilla::VsyncRefreshDriverTimer::RefreshDriverVsyncObserver::*)(mozilla::TimeStamp) __ptr64,1,0,mozilla::TimeStamp>::Run (c:\mozilla-source\stylo\obj-firefox-stylo\dist\include\nsthreadutils.h:869) #17: nsThread::ProcessNextEvent (c:\mozilla-source\stylo\xpcom\threads\nsthread.cpp:1262) #18: NS_ProcessNextEvent (c:\mozilla-source\stylo\xpcom\threads\nsthreadutils.cpp:389) #19: mozilla::ipc::MessagePump::Run (c:\mozilla-source\stylo\ipc\glue\messagepump.cpp:96) #20: MessageLoop::RunHandler (c:\mozilla-source\stylo\ipc\chromium\src\base\message_loop.cc:232) #21: MessageLoop::Run (c:\mozilla-source\stylo\ipc\chromium\src\base\message_loop.cc:212) #22: nsBaseAppShell::Run (c:\mozilla-source\stylo\widget\nsbaseappshell.cpp:158) #23: nsAppShell::Run (c:\mozilla-source\stylo\widget\windows\nsappshell.cpp:264) #24: nsAppStartup::Run (c:\mozilla-source\stylo\toolkit\components\startup\nsappstartup.cpp:283) #25: XREMain::XRE_mainRun (c:\mozilla-source\stylo\toolkit\xre\nsapprunner.cpp:4461) #26: XREMain::XRE_main (c:\mozilla-source\stylo\toolkit\xre\nsapprunner.cpp:4638) #27: XRE_main (c:\mozilla-source\stylo\toolkit\xre\nsapprunner.cpp:4729) #28: do_main (c:\mozilla-source\stylo\browser\app\nsbrowserapp.cpp:235) #29: NS_internal_main (c:\mozilla-source\stylo\browser\app\nsbrowserapp.cpp:307) #30: wmain (c:\mozilla-source\stylo\toolkit\xre\nswindowswmain.cpp:118) #31: __scrt_common_main_seh (f:\dd\vctools\crt\vcstartup\src\startup\exe_common.inl:253) #32: BaseThreadInitThunk[C:\WINDOWS\System32\KERNEL32.DLL +0x8364] #33: RtlUserThreadStart[C:\WINDOWS\SYSTEM32\ntdll.dll +0x670d1]

Emilio, did you do a try push for bug 1335987?

Since mochitest is not green yet, having a try push may not help a lot...

(In reply to Bobby Holley (:bholley) (busy with Stylo) from comment #1) > Emilio, did you do a try push for bug 1335987? Whoops, I did, but I might have added the assertion after that (thought that the assertion was trivially correct). That being said, I can't repro at all apparently with the branch I have recently? This may be because of the recent animation changes? Will rebase and see.

I can't reproduce this at all with origin/bookmarks/stylo + my patch. Xidorn, do you have any other patch locally applied?

I manual-typed a simplified version of the command... which doesn't crash... I actually use a slightly more complicated command which crashes: > ./mach mochitest --disable-e10s -f plain --keep-open true layout/style/test/ Sorry for the confusion.

Since "--keep-open true" is necessary to trigger this assertion, I guess this is not necessarily a big blocker for now then...

This patch just removes the assertion, which doesn't hold in the case of explicit change hints going through the restyle manager.

Comment on attachment 8834349 [details] [diff] [review] patch Review of attachment 8834349 [details] [diff] [review]: ----------------------------------------------------------------- Not sure what does it mean... redirect to bholley.

Comment on attachment 8834349 [details] [diff] [review] patch Review of attachment 8834349 [details] [diff] [review]: ----------------------------------------------------------------- Ah yes, good point!

Pushed by ecoal95@gmail.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/ffd74513a734 remove bogus assertion in presence of explicit change hints. r=bholley