Question

I need to know how does this thing work? By this i mean ” all calls to an inline function must be recompiled”. I am reading a book which says that each time the inline function is used in our program, the compiler will recompile the short function definition and place a copy of this compiled short definition in your code.

I don’t understand this at all. An explanation with example showing the whole proces will be highly appreciated. Also, could you please explain how does it improve the efficiency.

Thanks a lot.

Answer 1

Suppose you have this code:

void foo() { bar(); }

int main()
{
   foo();
   foo();
   foo();
}

If foo is inlined, then the compiler will essentially rewrite the code as:

int main()
{
   bar();
   bar();
   bar();
}

Now if instead of bar(); you had a more complex piece of code, then that piece of code would appear (and be compiled) three times, rather than just once.

You trade the cost of a function call against increased and repetitive code.

The compiler may well refuse to actually inline a function. If you ever take the address of foo and pass it somewhere outside, you cannot even get around creating a stand-alone version. In practice, a happy mix of standalones and inlining will happen, depending on what fits best. (More important to what code ends up being generated is the effect of the inline keyword on the one-definition rule, though.)