SDL UpdateWindowSurface() returns -1 if called from a class (in separate file)

Question

Today I started a C++/SDL2 Snake clone, and I've been looking for ways to make my code neater, -particularly with classes. I tried to put all the SDL code used for the window/display in a class. When I run the code, the window closes instantly. From the error tests I set up in display.cpp, it also tells me through the console that SDL_UpdateWindowSurface() (in display.update()) is always returning -1. Why does this happen when I rearrange my code like this?

With this code I load an image in main() and display it through my class' function applySurface(). The idea is to have classes/objects for the game's grid/board, the snake, etc., -each calling applySurface() for their own images. Feel free to tell me if this is a bad idea altogether.

main.cpp:

#include <SDL.h>
#include "display.h"

SDL_Event event;
SDL_Surface* image = nullptr;

int main(int argc, char* args[])
{
    Display display;
    display.loadImage("image.bmp");

    if (SDL_Init(SDL_INIT_EVERYTHING) == -1)
        return false;

    bool quit = false;

    while (!quit)
    {
        while (SDL_PollEvent(&event))
        {
            if (event.type == SDL_QUIT)
                quit = true;
        }

        display.applySurface(0, 0, image, display.windowSurface);
        display.update();
    }

    SDL_FreeSurface(image);
    SDL_Quit();
    return 0;
}

display.h:

#pragma once
#include <SDL.h>
#include <string>
#include <iostream>

class Display
{
public:
    SDL_Window* window;
    SDL_Surface* windowSurface;

    Display();
    SDL_Surface *loadImage(std::string fileName);
    void applySurface(int x, int y, SDL_Surface *source, SDL_Surface *destination, SDL_Rect *clip = nullptr);
    void update();
    ~Display();
private:
    const int WINDOW_WIDTH = 612;
    const int WINDOW_HEIGHT = 632;
    const int SCREEN_BPP = 2;
};

display.cpp:

#pragma once
#include "display.h"

Display::Display()
{
    window = SDL_CreateWindow("Snake", SDL_WINDOWPOS_CENTERED, SDL_WINDOWPOS_CENTERED, WINDOW_WIDTH, WINDOW_HEIGHT, SDL_WINDOW_SHOWN);

    if (window == NULL)
        std::cout << "Error: SDL_CreateWindow failed." << std::endl;

    windowSurface = SDL_GetWindowSurface(window);
}

SDL_Surface* Display::loadImage(std::string fileName)
{
    SDL_Surface* loadedImage = NULL;
    SDL_Surface* optimizedImage = NULL;

    loadedImage = SDL_LoadBMP(fileName.c_str());

    if (loadedImage != NULL)
    {
        optimizedImage = SDL_ConvertSurface(loadedImage, windowSurface->format, 0);

        SDL_FreeSurface(loadedImage);

        if (optimizedImage != NULL)
            SDL_SetColorKey(optimizedImage, SDL_TRUE, SDL_MapRGB(optimizedImage->format, 255, 255, 255));
    }

    return optimizedImage;
}


void Display::applySurface(int x, int y, SDL_Surface *source, SDL_Surface *destination, SDL_Rect *clip)
{
    SDL_Rect offset;

    offset.x = x;
    offset.y = y;

    SDL_BlitSurface(source, clip, destination, &offset);
}

void Display::update()
{
    if (SDL_UpdateWindowSurface(window) == -1)
        std::cout << "Error: SDL_UpdateWindowSurface() failed." << std::endl;
}

Display::~Display()
{
    SDL_FreeSurface(windowSurface);
    windowSurface = NULL;

    SDL_DestroyWindow(window);
    window = NULL;
}

Answer 1

This is a valid use of classes to structure your code. SDL_Init needs to come before any other SDL functions, which means you're best off moving SDL_Init to the top of main or adding it to the display constructor. If you add it to the beginning of the display constructor this means that you can only have one display class object running at a time, which would likely be fine in this case.