Could you please explain to me what $x[(scalar @x) - y] is doing?

(scalar @x) is the length of @x here 3

(scalar @x) - 1 == 2

In other words

$x[-y] is a shortcut for $x[(scalar @x) - y]

=>

$x[-1] eq $x[2] eq "light"

$x[-2] eq $x[1] eq "bar"

$x[-3] eq $x[0] eq "foo"

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}

As an historical aside, note that using the result of evaluating the expression  (scalar @ra) - 1 is not quite the same as using an index of -1 or $#array if you have set the array base $[ (which defaults to 0) to be 1 — but don't do that! See $[ in Deprecated and removed variables in perlvar. Assignment to $[ was first deprecated in Perl version 5.12 and has been increasingly constrained in subsequent versions.

c:\@Work\Perl\monks>perl -wMstrict -le
"$[ = 1;
 ;;
 my @ra = qw(zero one two three);
 ;;
 print scalar @ra;
 print 0 + @ra;
 print $ra[ (scalar @ra) - 1 ];
 print $ra[         @ra  - 1 ];
 print $ra[         @ra      ];
 print $ra[  -1  ];
 print $ra[ $#ra ];
"
Use of assignment to $[ is deprecated at -e line 1.
4
4
two
two
three
three
three
[download]

• evaluating an array in any scalar context yields the number of elements (or length) of the array;
• $#array yields the maximum index of the array;
• negative indices are handled consistently.

Thanks for the clarification.