using c++ in c code

Printable View

10-26-2005
hannibar

using c++ in c code

I'm writing a program in c, and I try to use mmgr with it. Mmgr is a c++ file, and a h file. To use mmgr, I have to include the .h file in all my .c files. When I try to do this I get several errors :

Does anyone knwo how I can solve this ?

Code:

F:\Cvs\Gui\mmgr\mmgr.h(57) : error C2061: syntax error : identifier 'bool' F:\Cvs\Gui\mmgr\mmgr.h(58) : error C2061: syntax error : identifier 'breakOnRealloc' F:\Cvs\Gui\mmgr\mmgr.h(58) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(62) : error C2059: syntax error : '}' F:\Cvs\Gui\mmgr\mmgr.h(101) : error C2143: syntax error : missing '{' before '&' F:\Cvs\Gui\mmgr\mmgr.h(101) : error C2059: syntax error : '&' F:\Cvs\Gui\mmgr\mmgr.h(102) : error C2143: syntax error : missing '{' before '&' F:\Cvs\Gui\mmgr\mmgr.h(102) : error C2059: syntax error : '&' F:\Cvs\Gui\mmgr\mmgr.h(119) : error C2061: syntax error : identifier 'm_validateAddress' F:\Cvs\Gui\mmgr\mmgr.h(119) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(119) : error C2059: syntax error : 'type' F:\Cvs\Gui\mmgr\mmgr.h(120) : error C2061: syntax error : identifier 'm_validateAllocUnit' F:\Cvs\Gui\mmgr\mmgr.h(120) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(120) : error C2059: syntax error : 'type' F:\Cvs\Gui\mmgr\mmgr.h(121) : error C2061: syntax error : identifier 'm_validateAllAllocUnits' F:\Cvs\Gui\mmgr\mmgr.h(121) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(121) : error C2059: syntax error : ')' F:\Cvs\Gui\mmgr\mmgr.h(127) : error C2143: syntax error : missing ')' before '*' F:\Cvs\Gui\mmgr\mmgr.h(127) : error C2143: syntax error : missing '{' before '*' F:\Cvs\Gui\mmgr\mmgr.h(127) : error C2059: syntax error : ')' F:\Cvs\Gui\mmgr\mmgr.h(134) : error C2143: syntax error : missing ')' before '*' F:\Cvs\Gui\mmgr\mmgr.h(134) : error C2143: syntax error : missing '{' before '*' F:\Cvs\Gui\mmgr\mmgr.h(134) : error C2059: syntax error : 'type' F:\Cvs\Gui\mmgr\mmgr.h(134) : error C2059: syntax error : ')' F:\Cvs\Gui\mmgr\mmgr.h(135) : error C2143: syntax error : missing ')' before '=' F:\Cvs\Gui\mmgr\mmgr.h(135) : error C2072: 'm_dumpMemoryReport' : initialization of a function F:\Cvs\Gui\mmgr\mmgr.h(135) : error C2059: syntax error : 'type' F:\Cvs\Gui\mmgr\mmgr.h(135) : error C2059: syntax error : ')' F:\Cvs\Gui\mmgr\mmgr.h(142) : error C2061: syntax error : identifier 'new' F:\Cvs\Gui\mmgr\mmgr.h(142) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(142) : error C2059: syntax error : 'type' F:\Cvs\Gui\mmgr\mmgr.h(143) : error C2061: syntax error : identifier 'new' F:\Cvs\Gui\mmgr\mmgr.h(143) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(143) : error C2059: syntax error : '[' F:\Cvs\Gui\mmgr\mmgr.h(144) : error C2061: syntax error : identifier 'new' F:\Cvs\Gui\mmgr\mmgr.h(144) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(144) : error C2059: syntax error : 'type' F:\Cvs\Gui\mmgr\mmgr.h(145) : error C2061: syntax error : identifier 'new' F:\Cvs\Gui\mmgr\mmgr.h(145) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(145) : error C2059: syntax error : '[' F:\Cvs\Gui\mmgr\mmgr.h(146) : error C2061: syntax error : identifier 'delete' F:\Cvs\Gui\mmgr\mmgr.h(146) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(146) : error C2059: syntax error : 'type' F:\Cvs\Gui\mmgr\mmgr.h(147) : error C2061: syntax error : identifier 'delete' F:\Cvs\Gui\mmgr\mmgr.h(147) : error C2059: syntax error : ';' F:\Cvs\Gui\mmgr\mmgr.h(147) : error C2059: syntax error : '[' Error executing cl.exe.

Thanks in advance.
10-26-2005
lemnisca

Yes.

Use a C++ compiler.
10-26-2005
hannibar

And how can I do it without using a c++ compiler ?
10-26-2005
kryptkat

Quote:

using c++ in c code

<Hisssssss!>

You can use c in c++ but not c++ in c code. You can .cpp your c code and compile it fine. If you use c++ you have to .cpp your .exe compile.

You ask how you can solve your errors. Just .cpp your file. That should take care of most of your errors. Some are syntax errors with out seeing the code i can only guess what you did.
10-26-2005
hannibar

I've seen something like :

Code:

#ifdef __cplusplus extern "C" { #endif #include "cppfile.h" //c code #ifdef __cplusplus } #endif

I tried it, but it doesn't work, but maybe it is a key to the solution ? I don't want to make the c files into cpp files. I just hope there is a 'cleaner' solution.
10-26-2005
durban

all that code is doing is checking whether you are using a c++ compiler, if you are, then it launches that code. That line proves that kryptkat is right - you CANNOT use C++ code in a C compiler, otherwise you wouldn't check for it right? To make that code work with C++ you would have to #define __cplusplus.
10-26-2005
hannibar

Ok this is hard to explain... I'll try again.

I am writing a program in c using MSVC++6. Because I want to test this program on memory leaks, I decided to use a library called mmgr. (can be found here : http://www.fluidstudios.com/publications.html). This small library is written in c++. I have read that it can be used for c++ AND for c. Now the only thing I want to know is how I can do it. I've searched quite a bit, but I haven't found any info on it.
10-26-2005
durban

Check the documentation. If they promote that it can be used for C and C++ then they will have a reference on how to do it.
10-26-2005
hannibar

The documentation doesn't say anything about it.
10-26-2005
Ancient Dragon

yes you can use c++ code in C programs, but there are some restructions.
1. C programs can't access any c++ classes, or overloaded functions. If you want a function in a c++ file to be used from C, then you have to declare it as "extern C"

Code:

// this is in the implementation file something.cpp extern "C" void foo() { // do something }

now, function foo() above can be called from either C or C++. And function foo() can access all normal c++ classes and other language features.

Do the same in the header file

Code:

// myheader.h #ifdef __cplusplus extern "C" { #endif void foo(); // other functions here #ifdef _cplusplus } #endif

I don't think the extern "C" will work the way you posted earlier -- I couldn't get VC++ 6.0 compiler to like that either. I had to put the extern "C" stuff in the .h file itself.

Note: __cplusplus is not ansi standard macro (or it wasn't a few years ago), so different compiles may define something else.
10-26-2005
durban

wouldn't extern "C" be linking to the "C" libraries? Saying a C++ program can compile on a C compiler is ridiculous. You are actually using C in C++, not the other way around.
10-26-2005
Ancient Dragon

I didn't say the C++ modules are compiled with a C compiler. You have to use a c++ compiler to do that, such as g++, Dev-C++, VC++ 6.0/7.0, and many other c++ compilers. c++ compilers can compile both c and c++ modules, and it knows how to compile by the file extension (*.c for C code and either *.cpp or *.cc for c++ code). Many c++ compilers have additional options to recognize other file extensions.

If you want to link both C and C++ into the same executable, then yes, the C runtime libraries will also be linked into the program.

[edit]The above assumes that the source code for all *.c and *.cpp files is available and can be modified as needed. In OP case, I don't think he does -- it appears he may be trying to call c++ library functions from c, which as you said before, cannot be done.
10-27-2005
hannibar

Thanks for the help everyone. It seems that my problem is not trivial to solve, so I'll try to code my own memory manager library.

btw : as a side-question :
in c++, when you declare a function like this :

int &function(...);

hen what does the & mean?
10-27-2005
Ancient Dragon

it equivalent to this in C
int* function();
10-27-2005
grumpy

Quote:

Originally Posted by Ancient Dragon

it equivalent to this in C
int* function();

No, it is not. If you wanted an equivalent functionality in C, you would emulate it using a function that returns a pointer. But the way you use that function is syntactically different.

In

Code:

int &function()

means that a function returns a reference to an int. A sample usage is;

Code:

int a = 0, b = 0; int &function(int t) { if (t) return a; else return b; } int main() { function(1) = 24; // sets the global a to 24 function(0) = 42; // sets the global b to 42 }

will print 24 42.

The (approximate) equivalent in C is this;

Code:

int a = 0, b = 0; int *function(int t) { if (t) return &a; else return &b; } int main() { *(function(1)) = 24; /* sets the global a to 24 */ *(function(0)) = 42; /* sets the global b to 42 */ }

While it achieves the same net effect, both the implementation of the function and the way it is invoked to achieve that effect are different.
10-27-2005
Prelude

>No, it is not.
It depends on how you define equivalent. If you see it as having similar or identical effects but potentially different ways of obtaining those effects (a common interpretation) then using a pointer to simulate a reference is equivalent. Now, if Ancient Dragon had said that the two methods were equal, then you would be right and he would be wrong. But as it is, you're just being overly pedantic. I don't imagine that there was any confusion about what he meant.
10-28-2005
grumpy

Quote:

Originally Posted by Prelude

>No, it is not.
It depends on how you define equivalent. If you see it as having similar or identical effects but potentially different ways of obtaining those effects (a common interpretation) then using a pointer to simulate a reference is equivalent. Now, if Ancient Dragon had said that the two methods were equal, then you would be right and he would be wrong. But as it is, you're just being overly pedantic. I don't imagine that there was any confusion about what he meant.

Sigh.

You're suggesting that I'm pedantic in using a "common interpretation", and then using a very pedantic but (in your view) obvious view of why my statement is incorrect.

Anyone who applies "a common interpretation" to Dragon's words could have been easily misled. I agree he was not incorrect, but there was enough potential for people to misunderstand that I considered it worthwhile pointing that out.

Your view of what is "obvious" is based on some fairly specialised knowledge. People who lack that knowledge will be fairly easily misled.

In a court of law, the litmus test of "reasonable" is something like "how would the majority of sane people interpret this statement". People who interpret statements in a highly specialised (or not "common") way tend to be the ones who are found to be misleading others --- and therefore the ones who lose in legal proceedings.
10-28-2005
Rashakil Fol

Saying that 'int *function()' and 'int &function()' are equivalent definitely implies that you can use function() the same way, which is untrue. Sure, it can be taken mean something else, but that requires ignoring the words that were written.