Significance of a .inl file in C++

.inl files are never mandatory and have no special significance to the compiler. It's just a way of structuring your code that provides a hint to the humans that might read it.

I use .inl files in two cases:

  • For definitions of inline functions.
  • For definitions of function templates.

In both cases, I put the declarations of the functions in a header file, which is included by other files, then I #include the .inl file at the bottom of the header file.

I like it because it separates the interface from the implementation and makes the header file a little easier to read. If you care about the implementation details, you can open the .inl file and read it. If you don't, you don't have to.


Nick Meyer is right: The compiler doesn't care about the extension of the file you're including, so things like ".h", ".hpp", ".hxx", ".hh", ".inl", ".inc", etc. are a simple convention, to make it clear what the files is supposed to contain.

The best example is the STL header files which have no extension whatsoever.

Usually, ".inl" files do contain inline code (hence the ".inl" extension).

Those files ".inl" files are a necessity when you have a dependency cycle between header code.

For example:

// A.hpp
struct A
{
    void doSomethingElse()
    {
       // Etc.
    }

    void doSomething(B & b)
    {
       b.doSomethingElse() ;
    }
} ;

And:

// B.hpp
struct B
{
    void doSomethingElse()
    {
       // Etc.
    }

    void doSomething(A & a)
    {
       a.doSomethingElse() ;
    }
} ;

There's no way you'll have it compile, including using forward declaration.

The solution is then to break down definition and implementation into two kind of header files:

  • hpp for header declaration/definition
  • inl for header implementation

Which breaks down into the following example:

// A.hpp

struct B ;

struct A
{
    void doSomethingElse() ;
    void doSomething(B & b) ;
} ;

And:

// A.inl
#include <A.hpp>
#include <B.hpp>

inline void A::doSomethingElse()
{
   // Etc.
}

inline void A::doSomething(B & b)
{
   b.doSomethingElse() ;
}

And:

// B.hpp

struct A ;

struct B
{
    void doSomethingElse() ;
    void doSomething(A & a) ;
} ;

And:

// B.INL
#include <B.hpp>
#include <A.hpp>

inline void B::doSomethingElse()
{
   // Etc.
}

inline void B::doSomething(A & a)
{
   a.doSomethingElse() ;
}

This way, you can include whatever ".inl" file you need in your own source, and it will work.

Again, the suffix names of included files are not really important, only their uses.


Since nobody else has mentioned it:

The use of .inl files to store your inline functions can be useful for speeding up compiles.

If you only include the declarations (.h) where you need declarations, and only include inline implementations (.inl) where you need them ( i.e. probably only in .cpp and other .inl files, not .h's ), it can have a beneficial effect on your header dependencies.

This can be a significant win on larger projects with many interacting classes.