I'm reading a code that has stuff like that.
*(int *)ptrArray = intN;
*(float *)ptrArray= floatN;
It is clear that different types are being assigned to the same vector of type void* . The question is, what is the meaning of *(int *) or *(float *) ?
Let's say I have a pointer to an integer
int* iPtr;
Given that it already points to a valid memory address, if I want to change the value of it I need to first derrefer the pointer
*iPtr = 100; //atribuo o valor 100
So this * is what indicates that I'm derreferencing this pointer and assigning the value to the memory it points to, if I did not have * there I would be trying to change the address this pointer points to
But this works because the compiler knows the type, when you use void* you have a generic pointer, since you do not have a defined type you can not simply derrefer this pointer, so you first need to cast it to a another type of pointer to be able to derrefer it, considering that we have a void* voidPtr variable, in the example
*(int*)voidPtr = 100;
First the (int*) cast void* to int* , then the first * derreferences the pointer to be able to assign the value
The expressions should be read as follows:
1) The value of ptrArray , which I am treating as a pointer to int , gets intN (which is int )
1) The value of ptrArray , which I am treating as a pointer to float , gets floatN (which is float )
This happens because you want to put a value (not an address) in the place indicated by ptrArray , which is a pointer to a data stream of undefined type ( void* ). So we need the forced conversion before the variable name, or the compiler will point to an error.
GCC 4.9.2 would say: error: invalid use of void expression