#!/usr/bin/perl
#

use Benchmark qw/ cmpthese /;

open F, "conf/English/stops" or die "Pb : $!\n";
chomp(@stops = <F>);

open G, "test/texte" or die "Pb : $!\n";
chomp(@text = <G>);

for $line (@text) {
    push @data, split /\s+/, $line 
}

%seen = map { $_ => 1 } @stops;

my @regexens = map { qr/\Q$_\E/ } @stops;
{
    local $" = '|';
    $r = qr/@regexens/;
}

cmpthese(100, {
    'string' => sub {
        WORD:
        foreach my $word (@data) {
            foreach my $stop (@stops) {
                next WORD if ($stop eq $word);
            }

            push @words, $word;
        }
    },
    'regex' => sub {
        WORD2:
        foreach my $word (@data) {
            next WORD2 if ($word =~ m/^$r$/);

            push @words, $word;
        }
    },
    'hash' => sub {
        WORD:
        foreach my $word (@data) {
            next WORD if $seen{$word};

            push @words, $word;
        }
    }
});
__DATA__
            (warning: too few iterations for a reliable count)
         Rate string  regex   hash
string 30.0/s     --   -70%   -99%
regex  99.9/s   233%     --   -97%
hash   3704/s 12230%  3607%     --