With a bit of research (here and here primarily), I sorted out a pure perl way to calculate the etag for an object on S3.

My motivation was that I'm having to sync data from an S3 bucket for $work so a user can do work on the files that are put there by a third party. I wanted to ensure that the files I had were complete and hadn't changed.

The only external requirement to actually use the calculator is to know the chunk size of the multi-part upload. I was able to guess this by running some rough math, but the person who uploaded the file (or you, if you're uploading the file) can provide this value to you.

use strict;
use warnings;
use Digest::MD5 qw(md5 md5_hex);

# needs a file name and a value for the multi-part chunk size in MB.
print "etag for $ARGV[0] = " . calculate_etag(@ARGV) . "\n";
sub calculate_etag {
    my ($file, $chunk_size) = @_;
    print "Calculating etag of $file...\n";
    my $string;
    my $count = 0;
    open(my $FILE, '<', $file);
    binmode $FILE;
    while (read($FILE, my $data, 1048576 * $chunk_size) {
        my $chunk_md5 = md5($data);
        $string .= $chunk_md5;
        $count++;
    }
    close($FILE);
    return(md5_hex($string) . "-$count");
}
[download]