Converting some c++ array code to c#

Question

I'm having some trouble converting the following code from c++ to c# because of pointers.

Basically I have a

STATE** State;
States = new STATE* [max_states];

for (int i=0; i < max_states; i++) {
  States[i] = new STATE(max_symbols);
}

If this was some double array I would say

STATE[][]  States;  
States = new STATE[max_states][];

for (int i = 0; i '<' max_states; i++) {
    States[i] = new STATE[max_symbols];
}

But the problem is the c++ code is not working "as" I expected it to.

States[i] = new STATE(max_symbols);

Has some strange behavior that for example allows

States[cur_state]->set_recur_out(k);

what exactly am I not seeing? This might be a beginner c++ question. Sorry if I don't make any sense at all =)

Answer 1

it is not a 2d-array, but a 1d-array containing pointers to single elements...
new STATE(max_symbols) constructs a single STATE object, calling the constructor which takes a single argument (in this case max_symbols).

i dont have that much clue of C#, but the following should be the correct representation of the C++ code in C#:

STATE[]  States;  
States = new STATE[max_states];

for (int i = 0; i '<' max_states; i++) {
    States[i] = new STATE(max_symbols);
}

Answer 2

This is simply an array of pointers. In C# that would be a one-dimensional array:

STATE[]  States = new STATE[max_states];

for (int i = 0; i < max_states; i++) {
    States[i] = new STATE(max_symbols);
}

The “strange behavior” that you are seeing in the C++ is simply the C++ way of accessing a member of a pointer to a type:

Type* x;
x->y();
// is the same as:
(*x).y();
// and corresponds to:
Type z;
z.y();