@Yardanico's answer is of course correct, but I want to try an example, which I think should be clearer, and also because I'd like to be corrected by someone who knows better if I'm wrong:
int a = 3 + 4;
const int b = dosomething();
int c = a + b;
b is also an l-value, though you can assign to it only once. The value returned by dosomething() is an r-value. You cannot compile dosomething() = 5.
In the third assignment, a, b, and c are all l-values, even though a and b appear on the right-hand side of an assignment, because you could assign to them (at least once) and (more importantly) their values persist beyond the assignment.
Now, suppose we have the following:
int *dosomething() {
int *d = (int *)malloc(sizeof(int));
return d;
}
const int e = *dosomething();
int f = *dosomething() = 5;
printf("%d %d\n", e, f);
@cantanima,
int f = *dosomething() = 5;
int f = *dosomething() ;
dosomething() = 5;
I can understand this. But this part is not digesting for me -
dosomething() = 5;
That assignment is invalid, and even a half-competent C/C++ compiler will report an error. The values that functions return are r-values; they disappear immediately after the statement. When I write dosomething() (without a star) I am referring to the function's return value, an r-value.
On the other hand, in the second version,
*dosomething() = 5;
Another way of saying this is that
f = *dosomething() = 5;
{
int *tmp = dosomething();
*tmp = 5;
f = *tmp;
}