In C, a function pointer (evidently) can be referred to with or without the &
operator. Is there a stylistic reason to choose one over the other?
The following code compiles and runs in gcc x86-64 12.2 -Wall without any warnings:
#include <stdio.h>
int cb1(void) { return 1; }
int cb2(void) { return 2; }
void do_cb(int (*cb)(void)) { printf("%d\n", cb()); }
int main(void) {
do_cb(cb1);
do_cb(&cb2);
}
CodePudding user response:
Is there a reason to choose one over the other?
&
is demanded in front of function identifiers by MISRA 2004. I like it, I think it looks nice and explicitly conveys the intent.
MISRA C Rule 16.9 (required): A function identifier shall only be used with either a preceding '&', or with a parenthesised parameter list, which may be empty.
Rationale
A function identifier can implicitly convert to a pointer to a function. In certain contexts this may result in a well-formed program, but which is contrary to developer expectations. For example, if the developer writes
if ( f )
then it is not clear whether the intent is to test if the address of the function is NULL or if a call to the function 'f()' should be made and the brackets have been unintentionally omitted. The use of the '& (address-of)' operator will resolve this ambiguity.
From https://analyst.phyzdev.net/documentation/help/reference/misra.func.addr.htm , https://rules.sonarsource.com/cpp/RSPEC-936 .
(I think the rule was removed in MISRA 2012. I wonder what is the rationale.)