Hi all, i am starting a new project, a bomberman fan game and i am gathering resources and testing codes...so i decided to make my code this way..

is this way bad praticing or is a good code??

in player and enemies classes i make this...

void cPlayer::logic()
{
  cAnimation::logic();
  // here more code of the logic of player
}

void cPlayer::render()
{
  cAnimation::render();
}

then i create a class cManager that will hold a list of all cEntity objects,
so i simplify the loop with this...

I just want to know if my code style is good, sugestions and so on

I agree with bamccaig that your entities should own an animation, not be an animation.

However, I don't agree that entities and animations shouldn't have their own logic and render methods. All animations share the same basic logic and rendering. I see how it might help if you really need flexibility in how you render an animation, but unless you need shaders/effects/something other than plain bitmap drawing, then I don't think an external object should render an animation. I definitely disagree that entities and animations should not have their own logic methods.

thanx to all for explanation...
so, after reading this i got new ideas...

class Entity
{
  public:
    virtual void logic();
    virtual void render();
};

so i think is this way good, right??

and one more thing to know your opinions...about tiles...
should i create a Level class that have a matrix tiles[][] to compare collision
or should i create a Level class that have a lot of tiles objects (class Tile, with information of x, y, size, is solid or not)??

Tiles: I'd recommend to separate two things:
- Available tiles: An array or list or vector where you enumerate the available tiles.
- Maps: each is (or contains) a 2D array or list<list> or vector<vector> of indices that refer to the distinct tiles above.

ie:

Firstly, your animation class doesn't seem to be an animation at all. It has a single bitmap. That is just a sprite. You do seem to have some of the workings for an animation (e.g., a frame) so perhaps you are just keeping the implementation simple for now and will extend it into an animation later (e.g., std::vector<ALLEGRO_BITMAP *> sprites)?

Secondly, I think being visible is more an attribute of an entity within the game, not the animation itself. You care if the entity is visible, not if its animation is visible. At least, that seems logical to me. I guess it might depend on the game, but I think that is true for any generic entity vs. animation. If they aren't generic classes then you might want to give them more specific names.

Audric's solution is good: tiles are usually reused in a level to make a bigger level from fewer tiles. To save memory you might want to load your tiles into memory once into a manager/container and then reference them somehow from your levels. In memory, your level can store a matrix of numeric indexes, string keys (probably a bad idea for performance reasons), or even pointers to these tiles. That way you don't waste too much memory, but you can still reference tiles relatively efficiently.

For example, you could use a single char to identify your tiles. Assuming you use mnemonics to assign characters to your tiles, this could make it relatively easy to create levels with them. Load them up into a map and look them up once at level load-time.

(Untested... There is some very questionable code regarding const/static initialization, but you get the idea...)

Well that all turned out pretty messy, but hopefully not too messy to follow.

The idea is that your levels can be like this:

# media/levels/the-alter
A=media/tiles/alter.bmp
g=media/tiles/grass.bmp
T=media/tiles/tree.bmp
-=media/tiles/wall_horz.bmp
|=media/tiles/wall_vert.bmp

|------------|
|gggggggggggg|
|ggTgg---gggg|
|ggggg|Aggggg|
|ggggg---ggTg|
|ggggTggggggg|
|gggggggggggg|
|------------|

<over-engineer carried-away="true" /></tangent></digress>

That's already exactly what happens..

Better yet, use an automatic variable instead, if possible.

first thanx to all for the atention....
responding the questions....

is just simple for this post. I use a bitmap with all animation sprite with specific size and clip them acordiling..

e.g. al_draw_bitmap_region(sprite, frame * w, stance * h, x, y, w, h, 0);

hmmm, dont know.. but i agree with you in this point

for tiles i will try to find someway first, then after ask!!

Not necessarily the stack. A class member can still be automatic, but the class instance itself can be on the heap.

*Something* like that anyway. Automatic variables are much easier to use and understand, which is why I suggested using them whenever possible.