I work in Visual Studio but my project is for a POSIX-based environment (marmalade sdk). In this project, the release
build is compiled with gcc for ARM but the debug version works on windows and is compiled by MS compiler. Also this environmet has its own implementation of STL and other standard libraries.
Many of these c++ librares have code like this:
#if defined( _MSC_VER )
#include <Windows.h>
#else
#include <pthread.h>
#endif
Is it possible to undefine the _MSC_VER macro? - So that the C++ libraries will detect a POSIX system here.
_MSC_VER is (and always should be) defined when compiling with the Microsoft compiler so that it "evaluates to the major and minor number components of the compiler's version number". Therefore, the code is using the wrong macro test, since it will always be defined to some value for your compiler regardless of the Windows environment differences.
Rather than destroy the definition of _MSC_VER (which could lead to other problems if any code really does want to know the compiler version), what you really should do instead is to correct the condition so that a more appropriate macro test is used that distinguishes between the kinds of Windows environments that you might encounter.
See the more complete list of predefined macros you could consider here
You could either replace the condition ...
#if someOtherConditionGoesHere
... or extend it with additional conditions, e.g.
#if defined(_MSC_VER) && someOtherConditionGoesHere
Of course:
#undef _MSC_VER
#if defined( _MSC_VER )
#include <Windows.h>
#else
#include <pthread.h>
#endif
Or, #undef it before you include the file where _MSC_VER is used.
Related
I have a VS (C++) project that relies on OpenCV and TBB, so I created property sheets for each library and included them in the project. Everything worked fine and the code compiled.
Yesterday, I have started using vcpkg package manager. I installed OpenCV and TBB via vcpkg and everything seemed to work. I created an empty project, included the headers of both and tested if the new compiled libraries work. After verifying that, I went back to my main project and removed the property sheets, so I can use the libraries from vcpkg. I did not change the code in any way since the last successful compilation.
But when I try to compile the code now I get this error two times (in main.cpp and in a submodule)
tbb\critical_section.h(53): error C3861: 'InitializeCriticalSectionEx': identifier not found
Does anybody know what is going on here or why this error occurs?
Update
I found the error myself. I'm adding the poco-libraries tag, because it's actually a conflict between TBB and Poco.
I found the source of the problem and it has actually nothing to do with TBB but with the Poco library.
Consider the minimum example:
#include <Poco/Poco.h>
#include <tbb/tbb.h>
void main()
{
}
This will throw an compiler error.
Tracing down the path
When including tbb.h, critical_section.h is included in line 51 of tbb.h. However, ciritcal_section.hpp includes machine/winwdows_api.h which looks like this (unnecessary stuff is cut out):
tbb/machine/winwdows_api.h:
#if _WIN32 || _WIN64
#include <windows.h>
#if _WIN32_WINNT < 0x0600
#define InitializeCriticalSectionEx inlineInitializeCriticalSectionEx
inline BOOL WINAPI inlineInitializeCriticalSectionEx( LPCRITICAL_SECTION lpCriticalSection, DWORD dwSpinCount, DWORD )
{
return InitializeCriticalSectionAndSpinCount( lpCriticalSection, dwSpinCount );
}
#endif
As you can see, windows.h is included before the check of the _WIN32_WINNT macro. This macro is defined in sdkddkver.h (which is included in windows.h), iff it's not already defined (in my case it's set to Win10):
sdkddkver.h:
#if !defined(_WIN32_WINNT) && !defined(_CHICAGO_)
#define _WIN32_WINNT 0x0A00
#endif
In windows.h, the _WIN32_WINNT macro controls which version of the windows header files are actually included. If _WIN32_WINNT is set to an earlier version than Windows Vista, the function InitializeCriticalSectionEx is not defined.
This issue is catched by machine/winwdows_api.h (as you can see in the code block of that file) by simply defining a macro InitializeCriticalSectionEx that calls an appropriate alternative function.
So far so good.
The problem
The root of all evil lies in Poco/UnWindows.h of the Poco library. When including a poco header, at some point UnWindows.h will be included.
Poco/UnWindows.h (shortened):
#if defined(_WIN32_WINNT)
#if (_WIN32_WINNT < 0x0501)
#error Unsupported Windows version.
#endif
#elif defined(NTDDI_VERSION)
#if (NTDDI_VERSION < 0x05010100)
#error Unsupported Windows version.
#endif
#elif !defined(_WIN32_WINNT)
#define _WIN32_WINNT 0x0501
#define NTDDI_VERSION 0x05010100
#endif
#endif
#include <windows.h>
The preprocessor checks, if _WIN32_WINNT is already defined, and if not, sets it to 0x0501 which is Windows XP. After that, windows.h is included. In the previous chapter I mentioned that _WIN32_WINNT controls which version of the windows header files are actually included.
Now imagine, the very first include in our project is a header from Poco. This means, that _WIN32_WINNT will be set to Windows XP and windows.h will include the windows headers of Windows XP (which imo is already a bad sign).
But don't worry, it gets worse.
If we trace the include hierarchy one level up, we reach Poco/Platform_WIN32.h.
Poco/Platform_WIN32.h (shortened):
#include "Poco/UnWindows.h"
...
#if defined (_WIN32_WINNT_WINBLUE)
#ifdef _WIN32_WINNT
#undef _WIN32_WINNT
#endif
#define _WIN32_WINNT _WIN32_WINNT_WINBLUE
...
Funny, isn't it? First, it includes UnWindows.h, which sets _WIN32_WINNT and causes Windows XP headers to be included, and next it redefines _WIN32_WINNT to be Windows 8.1. I have no clue why it does that, maybe there is a good reason, idk.
If we now look at the minimum example at the very top we see that Poco is included before TBB. What now happens is:
Include Poco headers
Set _WIN32_WINNT to Windows XP
Include windows headers (Windows XP version, because of 2)
Reset _WIN32_WINNT to Windows 8.1
Include TBB headers (windows headers are already included, so TBB doesn't need to include them again in tbb/windows_api.h)
TBB checks the windows version via _WIN32_WINNT and recognizes Windows 8.1 (as set by Poco)
TBB thinks InitializeCriticalSectionEx is defined, because the Windows version is 8.1 (or is it? Poco says: get rekt) and InitializeCriticalSectionEx is defined since Windows Vista.
Unfortunately Poco ensured that the Windows XP headers are loaded, so compiler says: no.
The solution
Either include windows.h yourself beforehand, or set _WIN32_WINNT yourself beforehand:
#define _WIN32_WINNT 0x0A00 // either this
#include <Windows.h> // or this
#include <Poco/Poco.h>
#include <tbb/tbb.h>
void main()
{
}
Maybe someone of the Poco contributors can clarify some things here. The Poco version is 1.8.1-1 built with x64 (via vcpkg).
Update
Poco is on the issue. Updates can be found here.
When testing for features in Visual Studio, I found this weird behaviour:
#ifndef __cpp_constexpr
#error Opposite day! //Compiler error
#endif
#define test_macro
#ifndef test_macro
#error But only for feature macros? //No compiler error
#endif
int main() {}
__cpp_constexpr is definitely defined, I used it in the actual program.
From my testing, it seems that with feature macros #ifndef behaves like #ifdef
and vice versa.
Is this a Visual Studio bug, or am I missing something?
Compiled in VS 2017, Visual C++14 or later, with C++17 standard enabled.
P.S. Intellisense is working as intended, it's only the compiler.
You are using incorrect pre-defined macro and pre-proc code. Next code can help:
#if defined(__GNUG__) && (__cplusplus > 201103L)
# define HAS_CONSTEXPR
#elif defined(_MSC_VER) && (_MSC_FULL_VER >= 190024210)
# define HAS_CONSTEXPR
#endif // __GNUG__
#ifndef HAS_CONSTEXPR
# error Opposite day! //Compiler error
#else
# define test_macro
# ifndef test_macro // useless by why not ?
# error But only for feature macros? //No compiler error
# endif
#endif
Suggesting to use boost config library. It is cross compiler, an you can solve it like this:
#ifdef BOOST_NO_CXX11_CONSTEXPR
# error Opposite day! //Compiler error
#endif
And it is going to work for all compilers supported by boost (according to the boost license, you can exam the source code and replicate is in you project)
I see this being done all the time for example in the Linux Kernel. What is the purpose of using the preprocessor commands vs just normal C++ if else block? Is there a speed advantage or something?
A preprocessor changes the C/C++ code before it gets compiled (hence pre processor).
Preprocessor ifs are evaluated at compile-time.
C/C++ ifs are evaluated at run-time.
You can do things that can't be done at run-time.
Adjust code for different platforms or different compilers:
#ifdef __unix__ /* __unix__ is usually defined by compilers targeting Unix systems */
#include <unistd.h>
#elif defined _WIN32 /* _Win32 is usually defined by compilers targeting 32 or 64 bit Windows systems */
#include <windows.h>
#endif
Ensure header file definitions are included only once (equivalent of #pragma once, but more portable):
#ifndef EXAMPLE_H
#define EXAMPLE_H
class Example { ... };
#endif
You can make things faster than at run-time.
void some_debug_function() {
#ifdef DEBUG
printf("Debug!\n");
#endif
}
Now, when compiling with DEBUG not defined (likely a command line parameter to your compiler), any calls to some_debug_function can be optimized away by the compiler.
Preprocessor is run before the compilation pass, so the compiler won't even see anything that was in the not-taken #if branch.
#if DEBUG
int a;
#else
double b;
#endif
gcc -c -DDEBUG=1 file.c will see "int a"
gcc -c file.c will see "double b"
Preprocessor allows you to actually cut out or paste in to your source file, code to be compiled. If its cut out, its gone, its like a comment, does nothing, is not compiled, produces no code in the binary. Devs will often use this technique to add code only in debug build for debugging purposes or for adding or excluding code for specific operating systems.
I am trying to migrate existing c++ 32 code to 64 code on windows7 with visual studio 2010.i never did 64bit compilation before. with the help of internet references i did the setup for 64 bit compilation. like VS2010 with 64 bit compiler etc and other configuration changes.
In the preprocessor i removed WIN32 and added WIN64. i have some other pre-processors like OS_WIN_32 and some other which are specific in my code.
In the code wherever WIN32 was getting used i added extra condition as || WIN64 this is just to ensure that application should get compiled with win32 as well as win64.
When i am trying to compile the code i am getting the compilation error saying
fatal error C1189: #error : Only one of the WIN32 and WIN64 symbols should be defined
this error is coming from the local code where we have a check whether both WIN32 and WIN64 are defined. that code is as shown below.
#if defined WIN32 && defined WIN64
# error Only one of the WIN32 and WIN64 symbols should be defined
#endif
in VS2010 if macros are not enabled then the code inside the macro gets greyed out. in my code also the above error is greyed out. but still i am getting that error.
The code where i added WIN64 is including windows.h. for reference givine it below.
#if defined WIN32 || defined WIN64
#include <windows.h>
#include <process.h>
#endif
So my question is why i am getting this error? shouldnt we add windows.h for 64bit compilation.? i tried by commenting this inclusion but i am getting other errors wrt HANDLE which are used in the code.
If i go to WIN32 definition VS2010 is pointing to a definition in windef.h file. This file is present in Microsoft SDKs\windows\v7.0A\include folder i.e. not my local code.
for referance that definition is given below.
#ifndef WIN32
#define WIN32
#endif
So i want to know why compiler is getting both pre-processors WIN32 and WIN64.
Thanks in advance for your help.
You shouldn't define either yourself. The macro's that should be used to check this are
_WIN32 // always defined for Windows apps
_WIN64 // only defined for x64 compilation
These are defined by the compiler (see here).
Often, the IDE adds the unprefixed macros to the commandline to not let legacy projects which use the non-documented unprefixed versions fail to build. The fact that they work is not a reason to use them, when documented alternatives are present.
It boils down to this:
#ifdef _WIN32
// We're on Windows, yay!
#ifdef _WIN64
// We're on x64! Yay!
#else // _WIN64
// We're on x86 (or perhaps IA64, but that one doesn't matter anymore). Yay!
#endif // _WIN64
#else // _WIN32
// We're not on Windows, maybe WindowsCE or WindowsPhone stuff, otherwise some other platform
#endif
Please look at the following file: (it is a complete file)
#ifndef TEES_ALGORITHM_LIBRARY_WRAPPER_H
#define TEES_ALGORITHM_LIBRARY_WRAPPER_H
#ifdef _TEES_COMPILE_AS_LIB
#include <dfa\Includes\DFC_algorithms.hpp>
#include <DFA\FuzzyClassifier\FuzzyAlgorithmIntialization\InitFuzzyAlgorithm.hpp>
typedef teesalgorithm::tees_fuzzy_algorithms algorithms_switchyard_class;
#else
#include <DFA\Includes\commercial_algorithms.hpp>
//An incomplete class to hide implementation
class algorithms_switchyard_class;
#endif
class AlgorithmLibraryWrapper {
algorithms_switchyard_class * algorithmPtr_;
typedef teesalgorithm::tees_paramObj paramObj_type;
typedef teesalgorithm::tees_errorObj errorObj_type;
typedef teesalgorithm::tees_statusObj statusObj_type;
typedef teesalgorithm::tees_dataObj dataObj_type;
typedef teesalgorithm::tees_outputObj outputObj_type;
public:
AlgorithmLibraryWrapper(const std::string& sAlgName, paramObj_type& paramObj, errorObj_type& errObj, statusObj_type& statusObj, const char* sFilePath);
static bool dataReader(const std::string& sFileName, dataObj_type& dataObj, errorObj_type& errObj, statusObj_type& statusObj);
bool runalgorithm(const dataObj_type& dataObj, outputObj_type& outObj, errorObj_type& errObj, statusObj_type& statusObj);
~AlgorithmLibraryWrapper();
};
#ifdef _TEES_USE_COMPILED_ALGORITHM_LIB
# ifdef _MSC_VER
#if _MSC_VER < 1400 // If VC 2003
#ifdef _DEBUG
#error No AlgorithmLibWrapper libraries compiled for this version of VC
#else
#error No AlgorithmLibWrapper libraries compiled for this version of VC
#endif
#elif defined(UNDER_CE) // Win CE
#ifdef _DEBUG
#pragma comment( lib, "AlgorithmLibWrapperCEd" )
#else
#pragma comment( lib, "AlgorithmLibWrapperCE" )
#endif
#else // If VC 2005
#ifdef _DEBUG
#pragma comment( lib, "AlgorithmLibWrapperd" )
#else
#pragma comment( lib, "AlgorithmLibWrapper" )
#endif
#endif
#endif
#endif
#endif //TEES_ALGORITHM_LIBRARY_WRAPPER_H
I am getting the following errors; I don't know why. I manually counted the preprocessor directives also.
AlgorithmLibraryWrapper.hpp:10:1: unterminated #ifdef
AlgorithmLibraryWrapper.hpp:7:1: unterminated #ifndef
I am using the poor vxWorks gcc compiler. Please let me know if the fault is mine or the compiler's.
It could be that the problem is in the included files (if there actually are unbalaced #if/#endifs.
I would try preprocessing with another compiler. You can use gcc for that, doesn't matter it wouldn't compile. Just get gcc (or MinGW if you're on Windows) and run
cpp -Iinclude_direcories your_file
Or, if you don't like gcc, get MSVC Express edition. Again, you can preprocess code that even doesn't compile, so no problem with nonworking library etc.
Most compilers have an option that will give you the output from the preprocessor so you can check what it's doing. For example,
gcc -E file.c >file.preproc
will give you the pre-processed source so you can check the balancing of #if against #endif.
At a guess, one of the files you are #including from this one has a mismatched #ifdef/#endif pair. You need to look at all the files (as the preprocesor does), not just this one.
As others have noted, this is most likely due to mismatched include guards.
If the files you are including are under your control (i.e. not part of a 3rd party closed source library), you could consider replacing the #ifndef et. al. guards (which are used to prevent multiple inclusion) with #pragma once. This will eliminate the possibility of having mismatched preprocessor directives.
One caveat of this is that pragma once is non-standard, so it will only work if your compiler supports it.
I have tried to compile your source using vs 6.0 but did not get the error you have mentioned. As others said may be the error is coming from the included header file . For me to get your code compiled i need to comment the above header.
Please check the above header once.
I'd debug this by commenting out sections one by one and trying to identify which section is causing the error.
It could be your compiler does not like the nested #ifdefs or does not interpret the syntax correctly. Maybe it doesn't like the #pragmas.
This is a long shot, but in your source file you have the following line:
# ifdef _MSC_VER
where there is whitespace between the '#' character and the directive name (ifdef). This is valid in C/C++; however, it's not too commonly seen so I wouldn't be very surprised if the odd compiler choked on it.