inline functions

[bpark]

It seems that inline functions are bound to the file it's defined in. Is there a possible way to overcome this?

[Dolda2000]

Where would you get such an idea. The difference is that inline functions defined in another file are not used inline in other files, for obvious reasons. In the kernel source code, they have overcome that by defining inline functions as static in the header files.

[bpark]

Well, I have two separate files. One is the file containing the main function with the appropriate function prototypes and the other is just the function definitions themself. When I compile these two, the compiler complains about the inline function.

[Dolda2000]

Can you give an example? This worked perfectly for me:
File 1:

#include <stdio.h>

inline void test&#40;int mast&#41;
&#123;
    printf&#40;"%i\n", mast&#41;;
&#125;

File 2:

int main(void)
{
    test(5);
}

Could it be that you declare the prototypes in the header files as inline? I don't think that will work, and it's possible that the compiler could complain about it.

[bpark]

file1

#include <stdio.h>

inline void test&#40;int mast&#41;
&#123;
    printf&#40;"%i\n", mast&#41;;
&#125;

file2

void test( int );

int main(void)
{
    test(5);
}

Result

undefined reference to `test(int)'

With or without the function prototype, I get errors. I'm using gcc2.95.

[Dolda2000]

What are you doing to compile this?

[bpark]

Well, I'm using a make file but the commands are really nothing but

g++ -Wall file.o file2.o

[Dolda2000]

Could it be something with C++? Try this:

gcc -c -o test.o test.c
gcc -c -o test2.o test2.c
gcc -o test test.o test2.o

[bpark]

Yea, this problem works fine in C. C++ however doesn't seem to like it.

[Dolda2000]

In that case I'm afraid I can't help you. Maybe it's a compiler bug? Yes, it's hard to believe, but it might not be impossible. You could try subscribing to the gcc mailing list and ask someone there.