perl unisanity

test.pl

#!/usr/bin/perl -CSAD
# You probably want to run as perl -CSAD /tmp/test.pl
# This is a little exploration of http://perldoc.perl.org/perlunicode.html#The-"Unicode-Bug"
# See also http://stackoverflow.com/questions/6162484/why-does-modern-perl-avoid-utf-8-by-default
use utf8;
use warnings;
use strict;

my $u = 'ü';

print "'$u' is utf8: " . utf8::is_utf8($u) . "\n";
print "$u =~ /\\w/: " . ($u =~ /\w/) . "\n";
my $ordu = ord($u);
printf("ord('$u') = 0x%x\n", $ordu);
my $xfc = "\xfc";
tests($xfc, $u);
print "utf::upgrade('\\xfc')\n";
utf8::upgrade($xfc);
tests($xfc, $u);

sub tests {
    my ($raw, $chr) = @_;

    print "\nTesting '$chr':\n";
    print "'\\xfc' is utf8: " . (utf8::is_utf8($raw) || 0) . "\n";
    print "'\\xfc' eq '$chr'? " . ($raw eq $chr || 0) . "\n";
    print "'\\xfc' =~ /\\w/: " . ($raw =~ /\w/ || 0) . "\n";
    print "uc('\\xfc') eq uc('$chr')? " . (uc($raw) eq uc($chr) || 0) . "\n";

    return;
}