Clean Up Windows Thread Join Debug Code

base/threading/platform_thread_win.cc

Index: base/threading/platform_thread_win.cc
diff --git a/base/threading/platform_thread_win.cc b/base/threading/platform_thread_win.cc
index 3571c522f7907d9cf1740564f4b6d6587064a5db..80ce760dde29b15751859a14b6cb20e8d4e6ea67 100644
--- a/base/threading/platform_thread_win.cc
+++ b/base/threading/platform_thread_win.cc
@@ -217,15 +217,8 @@ void PlatformThread::Join(PlatformThreadHandle thread_handle) {
 
   // Wait for the thread to exit.  It should already have terminated but make
   // sure this assumption is valid.
-  DWORD result = WaitForSingleObject(thread_handle.platform_handle(), INFINITE);
-  if (result != WAIT_OBJECT_0) {
-    // Debug info for bug 127931.
-    DWORD error = GetLastError();
-    debug::Alias(&error);
-    debug::Alias(&result);
-    CHECK(false);
-  }
-
+  CHECK_EQ(WAIT_OBJECT_0,

[danakj, 2016/06/10 17:38:25]

Can you make it a DCHECK also?

[danakj, 2016/06/10 17:41:06]

I see you made a comment about this is bad so we should CHECK. It was a CHECK
because it was tracking down a bug, which makes sense. Once a bug is resolved, I
think we should make it back into a DCHECK.

[robliao, 2016/06/10 18:06:42]

We CHECK on the equivalent call in POSIX.
https://cs.chromium.org/chromium/src/base/threading/platform_thread_posix.cc?...

So it seemed reasonable to keep the CHECK here. A failure to join could mean a
thread is now running around unexpectedly, which could cause a crash later down
the line.

[danakj, 2016/06/10 22:03:32]

Oh, okay thanks for the pointer. LGTM

+           WaitForSingleObject(thread_handle.platform_handle(), INFINITE));
   CloseHandle(thread_handle.platform_handle());
 }