How does this searching work in Perl? - perl

Question

my ($len, $longest) =0;
length > $len and ($longest, $len)=($_, length) for #matches;
#matches stores some substrings. This code catches the longest substring from #matches, then stores it in $longest.
Source code:
#!usr/bin/perl
use strict;
use Data::Dumper;
my $needle = "axibidm";
my $haystack = "axididm";
my #matches;
for my $start (0..length $needle) {
for my $len (1 .. ((length $needle)-$start)) {
my $substr = substr($needle, $start, $len);
push #matches, $haystack =~ m[($substr)]g;
print "$substr\t";
print "#matches\t\n";
}
}
my ($len, $longest) = 0;
length > $len and ($longest, $len) = ($_, length) for #matches;
print "The longest common substring between\n", $needle, "\nand\n", $haystack, "\nis '$longest'\n";

Answer 1

Someone was being far too clever, or they made a typo. Or both. Probably both.
There's some things going on with that code which don't do what it appears to be doing. This does not initialize both variables to zero.
my ($len, $longest) = 0;
Its a deceptive way to write this.
my $len = 0;
my $longest;
for $matches is silly, there's only one thing to iterate over so why use a loop? This idiom is very occasionally used to put the value into $_ and use it in various default constructs, but that's not much use here.
Next, the following is a very tortured way to write do this if that. It's written that way to make it a single expression which will work within the for loop statement modifier.
length > $len and ($longest, $len)=($_, length)
It is much better written like this.
if( length > $len ) {
$longest = $_;
$len = length;
}
Expanding it out, and removing the useless for loop, we get...
my $len = 0;
my $longest;
if( length $match > $len ) {
$longest = $match;
$len = length $match;
}
The other option is that $matches is an array references and they meant for #$matches. for $matches would still "work" but it would always return a length of 21 because an array reference stringifies as something like ARRAY(0x7fc07c800468).

Answer 2

That looks almost deliberately obfuscated. Here's a more verbose expression of the same logic.
my $len = 0;
my $longest;
foreach my $match (#matches) {
if (length($match) > $len) {
$longest = $match;
$len = length($match);
}
}
So let's compare.
my ($len, $longest) = 0;
This declares the two lexical (my) variables $len and $longest, and sets the first one ($len) to 0, leaving $longest at its default value of undef.
This structure:
(code goes here) for #matches;
is the same as this:
for (#matches) {
(code goes here)
}
So we're iterating over the #matches array and running the code once per element. Within the body of the code, the special variable $_ will hold the current element.
length > $len and ($longest, $len) = ($_, length);
First, (expression) and (code) is a shorthand way of writing if ( (expression) ) { (code) }. It works because in Perl, and is evaluated left-to-right in a short-circuiting fashion. That is, if the left side expression is false, Perl doesn't bother evaluating the right side, since its value doesn't matter; false and anything is false.
When length is called without an argument, it means length($_), so that's the length of the current element of #matches that is being examined.
And ($var1, $var2) = ($val1, $val2) is parallel assignment that sets $var1 to $val1 and $var2 to $val2.

Answer 3

EXPR for LIST;
is roughly the same as
for (LIST) { EXPR; }
EXPR1 and EXPR2;
is roughly the same as
if (EXPR1) { EXPR2; }
(This is not a generally accepted practice, except when EXPR2 is a flow control expression (next, die, etc).)
length defaults to using $_ as its argument (length($_)).
( $x, $y ) = ( EXPR1, EXPR2 )
is roughly the same as
$x = EXPR1;
$y = EXPR2;
(One notable difference is that you can do ($x,$y)=($y,$x) to swap values, but that's used here.)
(Using a list assignment when two scalar assignments would do is also not a generally accepted practice.)
A more conventional way of writing the code would be:
my $len = 0;
my $longest;
for (#matches) {
if (length($_) > $len) {
$longest = $_;
$len = length($_);
}
}