Closures

From JmPm

Contents 1 Closure And its Usage 2 Anonymous Function 3 Closure - basic 4 Example - counter 5 Counter - use 6 Counter – more interesting 7 Interval 8 Interval - use 9 Regexes - Task 10 Regex with Closure

Closure And its Usage

sub xx {
    my $n;
    $n = 5
}

1. $n inaccessible from outside function
2. here $n always disappears
3. Or does it always???

Anonymous Function

sub {
    ...
    ...
    ...
}

What is it good for?

Closure - basic

Lets combine my + anonymous function

sub xx {
    my $n;
    return sub {
        $n++;
    }
}

• $n not accessible outside xx
• $n does not disappear, since anonymous function needs it.
• For good explanation see M.J. Dominus Higher Order Perl pp. 71-79

Example - counter

sub create_cnt {
    my $start = 1;
    return sub {
        $start++;
    }
}

Counter - use

$c1 = create_cnt(); # cnt1
$c2 = create_cnt(); # cnt2
$c1->();  
$c1->();
$c1->();
$v1 = $c1->();
 
$c2->();
$v2 = $c2->();
print "v1 = $v1, v2 = $v2\n";

sub create_cnt {
    my $start = 1;
    return sub {
        $start++;
    }
}

Counter – more interesting

$c1 = create_cnt(1, 1);
$c2 = create_cnt(7, 3);
$c1->();
$c1->();
$c1->();
$v1 = $c1->();
$c2->();
$c2->();
$v2 = $c2->();
print "v1 = $v1, v2 = $v2\n";

Different start and delta:

sub create_cnt {
    my ($start, $inc) = @_;
    my $current;
    return sub  {
        $current =  $current ? $current += $inc: $start;
    }
}

Interval

sub time_int () {
    my $start = time();
    return sub (){
        my $current = time();
        my $diff = $current - $start;
        $start = $current;
        return $diff;
    }
}

Interval - use

my $t1 = time_int();
my $t2 = time_int();
for (1..4) {
    sleep(3);
    my $d1 = $t1->();
    print "$d1\n";
}
my $d2 = $t2->();
print "$d2\n";

sub time_int () {
    my $start = time();
    return sub (){
        my $current = time();
        my $diff = $current - $start;
        $start = $current;
        return $diff;
    }
}

Regexes - Task

• For ‘12345678’
• For any substring, e.g.34
• Print char before, then substring, then char after, e.g 1234
• Program to implement:

Regexes – implementation 1

my @x = @ARGV;
$str = '12345678';
foreach $cur  (@x) {
    my ($res) = $str =~ /(.?$cur.?)/;
    print "cur = $cur, res = $res\n"
}

Problem:

• Regex is compiled for each substring
• We cannot anticipate all substrings
• Closure to the rescue!

Regex with Closure

my @x = @ARGV;
$str = '12345678';
foreach $cur  (@x) {
    unless (defined $regex{$cur}) {
        $regex{$cur} = sdd_regex($cur); # key substr, val closure
        print "create for $cur\n";
    }
    else {
        print "already exists for $cur\n";
    }
    my $res = $regex{$cur}->(\$str); # use closure
    print "cur = $cur, res = $res\n";
}

sub sdd_regex () {
    my $s = shift;
    my $reg = qr/(.?$s.?)/; ## compile once for substr
    sub () {
        my $str = shift;
        #my ($r) =  $$str =~ /(.?$s.?)/;
        my ($r) =  $$str =~ /$reg/;
        return $r;
    }
}