Hi guys,

I'm programming a game with various game states. I've created an Entity Manager class, and in that class I add all my objects. However, I also create a special, static pointer to my Player. This is for transferring the Player data between game states.

/*pseudocode*/

c_Entity_manager{
private:
    static c_Player *player;

public:
    static void addPlayer( c_Player *p ){ player = p; };
    static c_Player* getPlayer(){return player;};
}; 

c_State1{
private:
    c_Player*  state1_player;

public:
    state1_player = c_Entity_manager::getPlayer();
};

c_State2{
private:
    c_Player*  state2_player;

public:
    state2_player = c_Entity_manager::getPlayer();
};

/*etc...*/

I'm almost certain this is considered a "singleton" pattern, and although I know that some argue that I should focus on whatever solves my problem and not worry too much about design, I know this is a common problem and I would like to break what could be a bad habit.

So, my questions are:
Is this considered a singleton pattern? If not, is it an effective way of passing data between states?

If it is, what is an alternative way to transfer a persistent objects' (for example, a player) data between states?

This isn't a singleton pattern.
A singleton is a class of which one and only one instance is allowed to exist. This is normally achieved by making the constructor, and the copy constructor private, and having a static function (usually named Instance() ) that returns the instance of that class (creating it if it doesn't exist), one version of which looks like this:

SomeClass* SomeClass::Instance(){  static SomeClass instance;  return &instance;}

You can just have the requirement that each of your states is initialized with a pointer to the player, which is what I do. Just create the player in whatever class owns/creates/manages each state, since it looks like you want the player to be persistent outside those states anyway. You don't even have to dynamically allocate the player, just make it a non-pointer class member of the class that manages your states.

If your game states inherit from a common base class, Player can belong to the base class as a static pointer.

Mine does that - the base class has a static pointer to GameStructure which is only pre-declared and not used in the base class at all.

The actual GameStructure is declared separately and has no internal knowledge of game states, only having a pointer to the root namestate.

Derived GameStates are defined elsewhere, and #include the GameStructure in their .cpp file, so they can actually use GameStructure.

GameStructure contains the Area, Player, Camera, the root GameState, and receives and forwards the player input and messages. It used to include the Window as well, but for ease-of-porting I preemptively separated the windowing system from GameStructure.

This seemed the most straightforward and easiest route for my code, while maintaining clear divisions between classes.