已具有返回类型的函数声明中的模板声明？

Question

So I'm following a C++/SDL2 programming tutorial series for 2D gaming. I was following along on this video and he wrote this bit of code.

using ComponentID = std::size_t;

inline ComponentID getComponentID()
{
    static ComponentID lastID = 0;
    return lastID++;
}

template <typename T> inline ComponentID getComponentID() noexcept
{
    static ComponentID typeID = getComponentID();
    //typeID below is being used like a function???
    return typeID(); //<- error and won't compile, I believe this was a typo
}

I don't understand why template<typename T> is put on that line for inline ComponentID getComponentID() noexcept . Why is that there? Obviously, there's no semicolon after a template declaration so I suppose it can be put on the same line but why? Is there something I'm missing? I haven't finished the video yet as this perplexed me and didn't wanna just go through the whole video copying code and not understanding it. Any hints?

EDIT: I know how templates work, I don't need a lesson on that. But there's already a return type in the declaration of that function which is what I was asking about. Not for strangers to teach me about the C++ language.

template <typename T> inline ComponentID getComponentID() noexcept

ComponentID aka std::size_t already specifies the return type does it not? T 's definition is put on the same line and is not used in that function. This is what I'm asking about. Did I miss something? Because I didn't think you could have multiple return types for a function. So since it isn't the return type and it isn't used in the function itself or as a type for a parameter what is its use in this context?

Answer 1

The goal is to have unique incremental ID by type.

So with

ComponentID getComponentID()
{
    static ComponentID lastID = 0;
    return lastID++;
}

template <typename T>
ComponentID getComponentID() noexcept
{
    static ComponentID typeID = getComponentID();

    return typeID; // () was a typo indeed
}

We will have

struct Component1 {};
struct Component2 {};


getComponent<Component1>(); // Will return 0
getComponent<Component2>(); // will return 1
getComponent<Component1>(); // Will return 0
getComponent<Component2>(); // will return 1

Answer 2

Templates are used in order to adapt the function or a class to support more than one build in or custom type. by doing so you can have different types used in the same code, ultimately reducing the duplicate code.

imagine that your getComponentID() had to return an int and also in another scenario need it to return a double. if getcomponentid() wasn't a template function, you had to overload the function to handle both int and double return types. Since you are using templates you don't need to deal with that. the compiler will handle it for you.

one thing to note though, templates can slow down your compilation time.

http://www.cplusplus.com/doc/oldtutorial/templates/

Following up on your edit:

he is not returning a template type. instead of doing;

template<typename T>
inline ComponentID getComponentID()
{
...
...
}

he is putting everything in one line.as follows.

template <typename T> inline ComponentID getComponentID()

yours truly, some stranger from the net -.-