Nesting Subroutines (Perl Cookbook, 2nd Edition)

10.16. Nesting Subroutines

10.16.1. Problem

You want subroutines to nest, such that one subroutine is visible and callable only from another. When you try the obvious approach of nesting sub FOO { sub BAR { } ... }, Perl gives warnings about variables that will not stay shared.

10.16.2. Solution

Instead of making the inner functions normal subroutines, make them closures and temporarily assign their references to the typeglob of the right name to create a localized function.

10.16.3. Discussion

If you use nested subroutines in other programming languages with their own private variables, you'll have to work at it a bit in Perl. The intuitive coding of this kind of thing gives the warning "will not stay shared." For example, this won't work:

sub outer {
    my $x = $_[0] + 35;
    sub inner { return $x * 19 }   # WRONG
    return $x + inner( );
}

The following is a workaround:

sub outer {
    my $x = $_[0] + 35;
    local *inner = sub { return $x * 19 };
    return $x + inner( );
}

Now inner( ) can be called only from within outer( ) because of the temporary assignments of the closure. Once called, it has normal access to the lexical variable $x from the scope of outer( ).

This essentially creates a function local to another function, something not directly supported in Perl; however, the programming isn't always clear.

10.16.4. See Also

The sections on "Symbol Tables" in Chapter 10 in Programming Perl and in perlmod(1); the sections on "Closures" and "Symbol Table References" in Chapter 8 of Programming Perl and the discussion of closures in perlref(1); Recipe 10.13; Recipe 11.4