I never really liked the way pointers are declared in C/C++:
int *a, *b, *c; // a, b and c are pointers to int
The reason is that I am used to reading variable declarations as MyType myVar1, myVar2, myVar3;
and I always read “int*
” as the type “integer pointer”�. I therefore wanted the following
int* a, b, c; // a is a pointer to int, b and c are ints
to mean that a
, b
and c
all were of type int*
, i.e. pointers to int
. and I therefore found it slightly annoying to repeat the asterisk for every variable. This also meant that the symbol *
had two slightly different meanings to me: (1) It declares a pointer or (2) it dereferences a pointer. I usually don't declare a whole lot of pointers in one line, but still, this is a (minor) annoyance I have briefly discussed with few fellow programmers over the years. Today I started reading C Traps and Pitfalls by Andrew Koenig and after reading one sentence, in chapter two, the pointer declaration syntax suddenly makes – at least some – sense:
[…] Analogously,
float *pf;
means that *pf is a float and therefore that pf is a pointer to a float.
Of course! If we instead of looking at it as a variable a
of type int*
, read it as *a
– i.e. “a
dereferenced” –� it makes sense. That is indeed an int
, and that also means that *
always means “dereference”�.
My problem with attaching it to the variable is twofold: It blends in with pointer dereferences and it doesn't make sense in the function return value scenario:
The pointer is not to the function, it is to the return value. My preference is to not take sides, I just put spaces on both sides: