This is purely theory based, but I have this code:

int i = 3, k[] = {2, 4, 6, 8, 10, 12}, *x = &i, *y = k;
double d = 1.5;
struct point_tag {
 int x, y;
 char *name;
} pt[] = {{200, 40, "begin"}, {300, 100, "end"}}, *pp = pt;

and these two expressions:

pt[i--].y 50 which causes undefined behavior at run time

*(*pp.name 2) which does not compile

I would just like to know why the top one can not run and why the bottom one does not compile, even though *((*pp).name 2) does.

CodePudding user response：

pt[i--].y 50 which causes undefined behavior at run time

The array pt has the size of 2, because you added two initializer items. This results in valid index values as 0 and 1. i has an initial value of 3. You just access the array out of bounds.

*(*pp.name 2) which does not compile

The operator preceedence favors . before *.

CodePudding user response：

You declared the array pt as having 2 elements. The variable i is initialized by the value 3

int i = 3,...;

So this expression with the subscript operator

pt[i--].y 50 

accesses memory beyond the array because the valid range of indices for the array is [0, 2).

As for this expression

*(*pp.name 2)

then it is the same as the expression

*(*( pp.name )   2 )

As the variable pp is a pointer you may not apply the dot operator.

You need to write at least

*( ( *pp ).name 2)