#!/usr/bin/perl

use Encode qw(encode decode from_to);
use Data::Dumper;
use warnings;
use strict;

my $str = 'раз два';
print "len: ".length($str)."\n";
print hexdump($str),"\n";

my $arr  = utf8ToBytes($str);
my $str2 = join('',@$arr);
print "len: ".length($str2)."\n";
print hexdump($str2),"\n";
 
my $str3 = encode('UTF-8', $str);
print "len: ".length($str3)."\n";
print hexdump($str3),"\n";


exit;

#========================================================================
sub hexdump {
    my $stuff = shift;

    my $retbuff = '';
    my @stuff   = ();

    return '' unless defined($stuff);

    for (my $i = 0; $i < length($stuff); $i++) {
        push @stuff, substr($stuff, $i, 1);
    }

    while (@stuff)
    {
        my $i = 0;
        $retbuff .= "\n" if($retbuff);
        my @currstuff = splice(@stuff, 0, 16);

        for my $currstuff (@currstuff) {
            $retbuff .= ' ' unless $i % 4;
            $retbuff .= ' ' unless $i % 8;
            $retbuff .= sprintf("%02X ", ord($currstuff));
            $i++;
        }
        for(; $i < 16; $i++) {
            $retbuff .= ' ' unless $i % 4;
            $retbuff .= ' ' unless $i % 8;
            $retbuff .= '   ';
        }

        $retbuff .= '  ';
        
        $i = 0;
        for my $currstuff (@currstuff) {
            $retbuff .= ' ' unless $i % 4;
            $retbuff .= ' ' unless $i % 8;
            if ( $currstuff ge chr(0x20) and $currstuff le chr(0x7E) ) {
               $retbuff .= $currstuff;
            } else {
               $retbuff .= '.';
            }
            $i++;
        }
    }

    return $retbuff;
}

sub utf8ToBytes {
    my $str = shift;

    my $p    = 0;
    my @arr  = (); #split // => $str;
    my @data = ();
    my $len  = length($str);
    
    for (my $i = 0; $i < $len; $i++) {
        push @arr, substr($str, $i, 1);
    }    

    for (my $i = 0; $i < $len; $i++)
    {
        my $c = ord($arr[$i]);

        if ($c < 128) {
            push @data, $c;
        }
        elsif ($c < 2048) {
            push @data, (($c >> 6) | 192);
            push @data, (($c & 63) | 128);
        }
        elsif ( (($c & 0xFC00) == 0xD800) && ($i + 1) < $len &&
                ((ord($arr[$i + 1]) & 0xFC00) == 0xDC00) ) {
            # Surrogate Pair
            $c = 0x10000 + (($c & 0x03FF) << 10) + (ord($arr[++$i]) & 0x03FF);

            push @data, (($c >> 18) | 240);
            push @data, ((($c >> 12) & 63) | 128);
            push @data, ((($c >> 6) & 63) | 128);
            push @data, (($c & 63) | 128);
        }
        else {
            push @data, (($c >> 12) | 224);
            push @data, ((($c >> 6) & 63) | 128);
            push @data, (($c & 63) | 128);
        }
    }
    
    return \@data;
}