int (*a)[10] = NULL; // initilaized to NULL pointer
printf("a :: %p\n", a); // print nil
printf("*a :: %p\n", *a); // seg fault??
return 0;
I executed this code on my x86 code using gcc, and it compiles without any error.
I expected something like
a :: (nil)
Segmentation Fault
but it actually prints out
a :: (nil)
*a :: (nil)
Why here referencing null pointer does not cause seg fault?
The reason in general is that dereferencing a pointer to array does not read any memory, the result is an array having same address - the result is just a different type.
The behaviour is undefined nevertheless.
Note that you used %p but you could have used %s for the second one (with char (*a)[10] = NULL;) which would have even more undefined behaviour, and yet still GCC/Glibc printf would often print (nil) instead of crashing.
That's because this is undefined behaviour, anything can happen on any compiler and even different versions of a compiler.
You are using gcc and the code that gcc generates is merely that it moves NULL (a zero) to rax which does not segfault.
mov QWORD PTR [rbp-8], 0
mov rax, QWORD PTR [rbp-8]
mov QWORD PTR [rbp-16], rax
You can see assembler output for C programs with different versions of gcc at
