What's the difference between a Perl sub declaration with and without parentheses?

Question

What is exactly the difference (if there is any) between:

sub foobar() {
    # doing some stuff
}

and

sub foobar {
    # doing some stuff
}

I see some of each, and the first syntax sometimes fail to compile.

Answer 1

By putting the () on the end of the subroutine name, you are giving it a prototype. A prototype gives Perl hints about the number and types of the arguments that you will be passing to the subroutine. See this section in perlsub for details.

Specifically, () is the empty prototype which tells Perl that this subroutine takes no arguments and Perl will throw a compilation error if you then call this subroutine with arguments. Here's an example:

#!/usr/bin/perl

use strict;
use warnings;
use 5.010;

sub foo {
  say 'In foo';
}

sub bar() {
  say 'In bar';
}

foo();
bar();
foo(1);
bar(1);

The output from this is:

Too many arguments for main::bar at ./foobar line 18, near "1)"
Execution of ./foobar aborted due to compilation errors.

It's that last call to bar() (the one with the argument of 1) which causes this error.

It's worth noting that Perl's implementation of prototypes is generally not as useful as people often think that they are and that outside of a few specialised cases, most Perl experts will avoid them. I recommend that you do the same. As of Perl v5.22, the experimental "signatures" feature is in testing that will hopefully accomplish many of the goals programmers from other languages would have expected from prototypes.