CEF3: Crash disposing browser instances asynchronously

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Open any page (a browser instance needs to be allocated)
2. Call browser deconstructors from other thread than the one calling the 
constructors

What is the expected output? What do you see instead?
Should not crash. See call stack below.

What version of the product are you using? On what operating system?
CEF3 r593 @ Windows 7 64bit

Please provide any additional information below.

[0414/165442:FATAL:non_thread_safe_impl.cc(12)] Check failed: 
CalledOnValidThread(). 
Backtrace:
    base::debug::StackTrace::StackTrace [0x056E8501+33] (c:\users\alexander\development\google\chromium@131752\src\base\debug\stack_trace_win.cc:149)
    logging::LogMessage::~LogMessage [0x056D4B1F+63] (c:\users\alexander\development\google\chromium@131752\src\base\logging.cc:564)
    base::NonThreadSafeImpl::~NonThreadSafeImpl [0x06F1D700+160] (c:\users\alexander\development\google\chromium@131752\src\base\threading\non_thread_safe_impl.cc:13)
    base::NonThreadSafe::~NonThreadSafe [0x05776706+22]
    CefResponseManager::~CefResponseManager [0x057766DC+44]
    CefBrowserImpl::~CefBrowserImpl [0x057AD337+103] (c:\users\alexander\development\google\chromium@131752\src\cef\libcef\renderer\browser_impl.cc:267)
    CefBrowserImpl::`vbase destructor' [0x057B4506+22]
    CefBrowserImpl::`scalar deleting destructor' [0x057B44C6+22]
    CefBrowserImpl::Release [0x057B43ED+93] (c:\users\alexander\development\google\chromium@131752\src\Crystalbyte.Chocolate.Application.Windows.exe has triggered a breakpoint
cef\libcef\renderer\browser_impl.h:121)
    CefCppToC<CefBrowserCppToC,CefBrowser,_cef_browser_t>::UnderlyingRelease [0x0569B16A+74] (c:\users\alexander\development\google\chromium@131752\src\cef\libcef_dll\cpptoc\cpptoc.h:118)
    CefCppToC<CefBrowserCppToC,CefBrowser,_cef_browser_t>::Release [0x0569A916+38] (c:\users\alexander\development\google\chromium@131752\src\cef\libcef_dll\cpptoc\cpptoc.h:109)
    CefCppToC<CefBrowserCppToC,CefBrowser,_cef_browser_t>::struct_release [0x0569B36D+189] (c:\users\alexander\development\google\chromium@131752\src\cef\libcef_dll\cpptoc\cpptoc.h:142)
    (No symbol) [0x61213917]

Original issue reported on code.google.com by krasshir...@googlemail.com on 14 Apr 2012 at 3:01

[GoogleCodeExporter]

Thanks, fixed in revision 632.

Original comment by magreenb...@gmail.com on 18 May 2012 at 9:18

• Changed state: Fixed

[GoogleCodeExporter]

This crash also occurs in single-process mode when the context is shut down 
before the render thread has stopped. Fixed in trunk revision 978.

Original comment by magreenb...@gmail.com on 6 Jan 2013 at 4:39

• Added labels: NeedsMerge

[GoogleCodeExporter]

Some people are reporting that IsRenderThreadShutdownComplete() introduced by 
revision 978 never returns true: 
http://magpcss.org/ceforum/viewtopic.php?f=6&t=10417&p=15625#p15616

Original comment by magreenb...@gmail.com on 1 Feb 2013 at 7:26

• Changed state: Accepted

[GoogleCodeExporter]

Original comment by magreenb...@gmail.com on 1 Feb 2013 at 7:26

• Added labels: NextRelease
• Removed labels: NeedsMerge

[GoogleCodeExporter]

@comment#3: The change in behavior was caused by Chromium revision 170644. The 
render "process" is now never shut down in single-process mode. For this reason 
the approach taken in revision 978 is no longer valid. Revision 978 changes 
have been reverted in trunk revision 1076 and 1364 branch revision 1077.

Original comment by magreenb...@gmail.com on 1 Feb 2013 at 8:04

[GoogleCodeExporter]

@comment#3: Fixed in trunk revision 1078 and 1364 branch revision 1079 by 
causing the RenderProcessHost to be deleted during shutdown.

Original comment by magreenb...@gmail.com on 2 Feb 2013 at 12:23

• Changed state: Fixed

[GoogleCodeExporter]

Issue 874 has been merged into this issue.

Original comment by magreenb...@gmail.com on 2 Feb 2013 at 12:27