[d8] cleanly force exit in d8 on windows.

src/d8-windows.cc

 // Copyright 2009 the V8 project authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "src/d8.h"
 
 
 namespace v8 {
 
 
 void Shell::AddOSMethods(Isolate* isolate, Local<ObjectTemplate> os_templ) {}
 
+void Shell::Exit(int exit_code) {
+  // Use _exit instead of exit to avoid races between isolate

[Michael Achenbach, 2016/11/04 10:40:44]

This comment doesn't apply anymore.

[Yang, 2016/11/04 11:54:17]

Done.

+  // threads and static destructors.
+  fflush(stdout);
+  fflush(stderr);
+  TerminateProcess(GetCurrentProcess(), exit_code);

[Michael Achenbach, 2016/11/04 10:40:44]

Do we have to wait for the process to exit? Similar to
https://cs.chromium.org/chromium/src/base/process/process_win.cc?q=TerminateP...

[Yang, 2016/11/04 11:54:17]

good point.

[brucedawson, 2016/11/04 15:03:48]

Actually, I'm pretty sure that other code is wrong and that the wait is not
needed. I can see where the idea came from because the documentation is
confusing. It says:

"This function stops execution of all threads within the process and requests
cancellation of all pending I/O. ...

TerminateProcess is asynchronous; it initiates termination and returns
immediately. If you need to be sure the process has terminated, call the
WaitForSingleObject function with a handle to the process."

The last paragraph suggests waiting. But the first paragraph says that it stops
execution of all threads, which makes waiting pointless.

I guess I don't know for sure, but I think that the waiting is for when you use
this on another process. I'm pretty sure it's not needed for suicide. You could
do a __debugbreak afterwards and I'm pretty sure it would never get hit.

Harmless either way.

+}
 
 }  // namespace v8