When to roll your own game engine?
I've been interested in writting games for a while, and now that I have acquired a bit of experience in programming, I feel like I could step up to making a little more complicated things. I've had some fun with SFML but since now I only had the basic classes I needed for my game objects, and a big run() method of my Game object to "run them all".
I've only written gameplay tests so far, but I feel like a complete game in the same style of programming would be a gigantic headhache source. I am now facing a problem/question : should I be considering creating my own little "game engine", learning through it, adding features as I need them ? I think it could be a really good experience, but it would certainly take time as I'm no professional, and I don't have a clear view of "how-to" yet.
I am using SFML with C++.