asign to return value

[Lost User]

Yes, but Func() is not really necessary, since you have already declared S. You can just say S.x = 100; Creating a function to return a structure (or anything else) is usually necessary when you cannot declare the relevant object at compile time, but need to create it dynamically. Something like:

TestStruct* Funct()
{
TestStruct* pS = new TestStruct;
pS->x = 0;
ps->z = 0;
// other initialisation code

return pS;

}

//
TestStruct* pNewStruct = Funct();
pNewStruct->x = 100;

[Stephane Capo]

Hello, the code itself will work, but will not do what I suppose you expect. Here, Funct returns a copy of the TestStruct S. So Funct().x = 100 will affect 100 in a temporary TestStruct that you can not use after that. I don't know why you want to use a function to access your S structure, but the behavior you look for is probably :

// Funct returns a reference on S structure.
TestStruct& Funct()
{
return S;
}

[Calin Negru]

my actual code:

ChartNode Chart[100];
ChartNode NodeCoord(int x, int z)
{
return Chart[z * 10 + x];
}

if(NodeCoord(4,4).access)
//do something
NodeCoord(4,4).access = false;

The result I`m looking for is the same as the result achieved with this function:

void NodeCoord(int x, int z, bool writetobool)
{
Chart[z * 10 + x].access = writetobool;
}

[Stephane Capo]

So you probably need to return a reference :

ChartNode& NodeCoord(int x, int z)
{
return Chart[z * 10 + x];
}

else the NodeCoord(4,4).access = false will assign a temporary copy of your ChartNode struct and not the one in your array.

[Calin Negru]

ChartNode& NodeCoord() will work for both read and write?

[Lost User]

If you are referring to array cells with two dimension values then you should use a two-dimensional array. As it is you can pass any values in to NodeCoord but there is no way of telling if they are valid. So you could end up with lots of random memory corruption.

[Calin Negru]

Thanks Richard How do I declare a two (or more) dimensional array? The problem is I might need arrays with a great number of dimensions. I need a `one fits all` type of solution.

[Lost User]

Just like a one-dimensional, but you declare the two dimensions. Think of it as a block of items having some number of rows and columns. so you could have something like:

#define MAXROW 10
#define MAXCOL 5

ChartNode Chart[MAXROW][MAXCOL];
ChartNode NodeCoord(int row, int column)
{
if (row > 0 && row < MAXROW && // make sure row index
column > 0 && column < MAXCOL) // and column index are within range
{
return Chart[row, column]; // return the cell
}
else
{
return NULL; // error
}
}

All of these features are well documented in the C/C++ documentation and the various tutorials on the language.

[Stephane Capo]

Yes.

[Calin Negru]

the two dimensional array seems like something too good to be true :) The compiler is definitely doing something suspicious behind the courtain. Thanks.

[Lost User]

I think you could do with studying all the features of C (or C++ if that is what you are using). Trying to learn a language by posting questions here will take ten times as long, and you will still miss most of it.

[Calin Negru]

Often I edit my posts/replies shortly after making them, I can`t organise myself in a single swing.

[_Superman_]

Internally it's just continuous memory that is allocated for a 2 dimensional array (For any no. of dimensions for that matter). The compiler uses the specified dimensions to calculate the offset into the memory to fetch. For instance, offset of Chart[2][3] could be calculated as -

(sizeof(ChartNode) * MAXROW * 2) + (sizeof(ChartNode) * 3)

«_Superman_»  _I love work. It gives me something to do between weekends.

_Microsoft MVP (Visual C++) (October 2009 - September 2013)

Polymorphism in C