On Jan 15, 2008, at 4:54 PM, "Ken Kixmoeller -- reply to ken@xxxxxxxxxxxxxx
" <Kixjaguar@xxxxxxxxxxx> wrote:
Hey --- - -
I am in the process of upgrading the encryption technology I am
using from (64 bit) blowfish to (256 bit) rijndael.
The code (and some explanations) is below, but the results are, um,
unusual, and I can't see what I am doing wrong. For testing, I have
a program that generates a random 16-character string, encrypts it
to a variable, and decrypts it. Running it in 500 iteration loops,
it fails roughly 4% of the time. By "fails" I mean that the original
string and the eventual decrypted one don't match.
Anybody able to spot why?
Ken
--------------------------------------
function jagencdecr($text,$EorD,$encpass='') {
// parameters:
// - $text = string to be en/decrypted,
// - $EorD = Encrypt or Decrypt
// - $encpass = key phrase
if (empty($text)) {return "";}
$text = trim($text);
$cypher = mcrypt_module_open('rijndael-256', '', 'ecb', '');
// "ecb" mode produces the above results.
// "ofb" mode produces 100% errors
$size = mcrypt_enc_get_iv_size($cypher);
$phprand = rand(1000,9999);
$iv = mcrypt_create_iv($size,$phprand); // produces the same
results as below, platform independent
//$iv = mcrypt_create_iv($size,MCRYPT_RAND); // for Windows
//$iv = mcrypt_create_iv($size,MCRYPT_DEV_RAND); // for 'NIX
$ks = mcrypt_enc_get_key_size($cypher);
/* Create key */
$key = substr(md5($encpass), 0, $ks);
mcrypt_generic_init($cypher,$key,$iv);
if ($EorD == "D") {
$text_out = mdecrypt_generic($cypher,$text);
} else {
$text_out = mcrypt_generic($cypher,$text);
} // endif ($EorD == "D")
mcrypt_generic_deinit($cypher);
mcrypt_module_close($cypher);
return trim($text_out);
} // endfunc jagencdecr Jaguar Ecnrypt/Decrypt
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
Maybe you could echo the results of the failed ones and compare.
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
