I'm walking through the examples and exercises in "Effective Perl Programming" by Joseph N. Hall and Randal Schwartz (2 asides: 1) yeah, Merlyn and, 2) reading on my Kindle, where the formatting is inconsistent and some fairly odd typos mar the text; looks like it was translated to Kindle format by someone unfamiliar with Perl).

This script, pure ASCII -- or so I think, is adapted from an example on sorting:

#!/usr/bin/perl
use Modern::Perl;

say "\n\n\t sort on multiple keys";

my @first = qw(John Joe John Jon Fred Rich);
my @last = qw(Smith Smith Jones Jackson Jones Adler);

my @index = sort {
    $last[$a] cmp $last[$b]
    or
    $first[$a] cmp $last[$b]
} 0 .. $#first;

for (@index) {
    say "$last[$_], $first[$_]";
}

say "\n \t But, testing, single sort (cmp):";


my @sorted = sort { $a cmp $b  } qw(John Joe John Jon Fred Rich);
for (@sorted) {
    say "\t $_";
}


=head execution

         sort on multiple keys
Adler, Rich
Jackson, Jon
Jones, John
Jones, Fred
Smith, John
Smith, Joe

         But, testing, single sort (cmp):
         Fred
         Joe
         John
         John
         Jon
         Rich

## sort puts "John" before "Joe" in the first instance but not in the 
+second: Why?

=cut
[download]

Can someone point me to the document I failed to read or the section I failed to read carefully enough that explains the disparity in the output?