[Bug tree-optimization/61502] == comparison on "one-past" pointer gives wrong result

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61502

--- Comment #49 from post+gcc at ralfj dot de ---
While checking whether some LLVM bugs also occur in GCC, I stumbled upon
another variant of this bug:

#include <stdio.h>
#include <stdint.h>

int main()
{
  // also try swapping these
  int x = 7777; 
  int y = 6666;

  int* p = &y+1;
  int* q = &x;

  if (q == p) {
    printf("pointers are equal!\n");
  } else {
    printf("pointers are NOT equal!\n");
  }

  uintptr_t dist = (((uintptr_t)p) - ((uintptr_t)q));
  if (dist == 0) {
    printf("Distance is 0 (i.e., poiners are equal)\n");
  }

  return 0;
}

(https://c.godbolt.org/z/TdGq4q5bK)

This prints

pointers are NOT equal!
Distance is 0 (i.e., poiners are equal)

which is logically inconsistent, i.e., a miscompilation. The pointers may be
equal or not, but they can't be sometimes equal and sometimes not just because
you look at them from a different direction.