Translating from C to Haxe

gcardozo

Hello, guys!
I'm translating some C code and I just faced this function:

cvgQERef *cvgQERot(cvgQERef *edge)
{
   return (edge->r < 3) ? edge+1 : edge-3;
}

As you can see, he's returning a pointer to the memory address where that edge is. That edge is part of an array, of course.
My question is: can I do that with HaxeFlixel or should I workaround it (like passing the array as a parameter)?

Thanks in advance!

starry-abyss

You pass objects (array elements), and Haxe cares about reference counting and GC. But for optimization you can see how HaxeFlixel does pools with FlxPoint get() and put().

gcardozo

I'm not sure if I understood your answer correctly, but in case you're suggesting me to use a pool, I can't, because that function needs to return a specific edge. Thanks for your answer anyway.
(:

starry-abyss

You don't need a pool. Basically, just have an Array<Edge> a = new Array<Edge>() (where Edge is your type) and return a[i]. Or do you ask about linked lists?

gcardozo

I was trying to achieve is something similar to the original C function.
Which would be something like this:

public function rot(edge:Edge):Edge
{
   return (edge.r < 3) ? edge + 1 : edge - 3;
}
//where edge+1 is equivalent to arr[edgeIndex + 1]

//somewhere else:
var arr = new Array<Edge>();
arr.push(specificEdge1);
arr.push(specificEdge2);
arr.push(specificEdge3);
arr.push(specificEdge4);

//in case arr[2].r == 2, then specificEdge would be arr[3]
var specificEdge = rot(arr[2]);

but apparently I can't do that in Haxe. So a workaround would be passing the array "arr" to the function, instead of just one edge.

starry-abyss

Hmm, you could cook your own wrapper that includes array and index and acts similar to pointer.

gcardozo

Ok, that might work. Thanks @starry-abyss!