win: Fix InitializationStateDcheck in Release build

util/misc/initialization_state_dcheck.h

 // Copyright 2014 The Crashpad Authors. All rights reserved.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
 // You may obtain a copy of the License at
 //
 //     http://www.apache.org/licenses/LICENSE-2.0
 //
 // Unless required by applicable law or agreed to in writing, software
 // distributed under the License is distributed on an "AS IS" BASIS,
 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
 #ifndef CRASHPAD_UTIL_MISC_INITIALIZATION_INITIALIZATION_STATE_DCHECK_H_
 #define CRASHPAD_UTIL_MISC_INITIALIZATION_INITIALIZATION_STATE_DCHECK_H_
 
 //! \file
 
 #include "base/basictypes.h"
+#include "base/compiler_specific.h"
 #include "base/logging.h"
 #include "util/misc/initialization_state.h"
 
 namespace crashpad {
 
 #if DCHECK_IS_ON || DOXYGEN
 
 //! \brief Tracks whether data are initialized, triggering a DCHECK assertion
 //!     on an invalid data access.
 //!
@@ skipping 117 matching lines @@
 //!     object.
 //!
 //! \sa crashpad::InitializationStateDcheck
 #define INITIALIZATION_STATE_DCHECK_VALID(initialization_state_dcheck) \
   DCHECK_EQ((initialization_state_dcheck).state(),                     \
             (initialization_state_dcheck).kStateValid)
 
 #else
 
 // Since this is to be used as a DCHECK (for debugging), it should be
-// non-intrusive in non-DCHECK (non-debug, release) builds. An empty struct
-// would still have a nonzero size (rationale:

[Mark Mentovai, 2015/02/11 20:47:12]

This comment explains why this wasn’t done before.

Since MSVC doesn’t accept the zero-length, can we at least make this conditional
on COMPILER_MSVC, to avoid wasting these bytes where it’s not necessary?

[scottmg, 2015/02/11 21:03:31]

Done.

It seemed unnecessarily tricky to me given the size and number of objects in
which it was used.

-// http://www.stroustrup.com/bs_faq2.html#sizeof-empty). Zero-length arrays are
-// technically invalid according to the standard, but clang and g++ accept them
-// without complaint even with warnings turned up. They take up no space at all,
-// and they can be “initialized” with the same () syntax used to initialize
-// objects of the DCHECK_IS_ON InitializationStateDcheck class above.
-using InitializationStateDcheck = bool[0];
+// non-intrusive in non-DCHECK (non-debug, release) builds.
+struct InitializationStateDcheck {
+};
 
 namespace internal {
 
 // This function exists to make use of the InitializationStateDcheck object so
 // that it appears to be used. It always returns true, so that it can be used
 // as the argument to a no-op DCHECK.
 inline bool EatInitializationState(const InitializationStateDcheck*) {
   return true;
 }
 
 }  // namepspace internal
 
-// The contents of these DCHECKs will never be evaluated, but they make use of
-// initialization_state_dcheck to avoid triggering -Wunused-private-field
-// warnings.
+// Avoid triggering warnings by making repurposing these macros when DCHECKs are
+// disabled.
 #define INITIALIZATION_STATE_SET_INITIALIZING(initialization_state_dcheck) \
-  DCHECK(::crashpad::internal::EatInitializationState( \
-      &(initialization_state_dcheck)))
+  ALLOW_UNUSED_LOCAL(initialization_state_dcheck)
 #define INITIALIZATION_STATE_SET_VALID(initialization_state_dcheck) \
-  DCHECK(::crashpad::internal::EatInitializationState( \
-      &(initialization_state_dcheck)))
+  ALLOW_UNUSED_LOCAL(initialization_state_dcheck)
 #define INITIALIZATION_STATE_DCHECK_VALID(initialization_state_dcheck) \
-  DCHECK(::crashpad::internal::EatInitializationState( \
-      &(initialization_state_dcheck)))
+  ALLOW_UNUSED_LOCAL(initialization_state_dcheck)
 
 #endif
 
 }  // namespace crashpad
 
 #endif  // CRASHPAD_UTIL_MISC_INITIALIZATION_INITIALIZATION_STATE_DCHECK_H_