java 和 Python Blowfish 的实施产生不同的结果

java and Python implementation of Blowfish produce different results

我在 java 中有一个遗留的河豚实现,我正试图将其移植到 Python。

Java:

import blowfishj.*;
import org.apache.commons.codec.binary.Hex;

private static byte[] EncryptBlowFish(byte[] sStr, String sSecret) {        
    byte[] key = sSecret.getBytes();
    byte[] cipher = new byte[sStr.length];
    BlowfishECB blowfish = new BlowfishECB(key, 0, key.length);
    blowfish.encrypt(sStr, 0, cipher, 0, sStr.length);

    return (new String(Hex.encodeHex(cipher)));

}

Python:

from Crypto.Cipher import Blowfish
import binascii

def encrypt(encr_str, key_str):
    cipher = Blowfish.new(key_str, Blowfish.MODE_ECB)
    return binascii.hexlify(cipher.encrypt(encr_str)).decode('utf-8')

如果要加密的字符串是“12345678”,密钥是“1234567890123456”,java代码输出"e00723bbb58234aa",python代码输出“61d2570dc6e09632”。

由于 java 代码是遗留的,我不能碰它。 This indicates there's a problem in pycrypto's implementation of blowfish. However, I can confirm the accepted answer here works. Not sure why though. I tried both pycrypto as well as this blowfish module 结果相同。

关于如何在 Python 中复制与遗留 java 代码相同的河豚输出的任何想法?

感谢@Kai Iskratsch 指出了正确的方向。

参考:

  1. What's the difference between Blowfish and Blowfish-compat?

  2. https://gist.github.com/adamb70/1f140573b37939e78eb5%22

这是对我有用的代码。

Python:

from Crypto.Cipher import Blowfish
from binascii import hexlify

def encrypt(key, string):
    """
    Encrypts input string using BlowFish-Compat ECB algorithm.
    :param key: secret key
    :param string: string to encrypt
    :return: encrypted string
    """
    cipher = Blowfish.new(key, Blowfish.MODE_ECB)
    return hexlify(_reverse_bytes(cipher.encrypt(_reverse_bytes(string)))).decode('utf-8')

@staticmethod
def _reverse_bytes(data):
    """
    Takes data and reverses byte order to fit blowfish-compat format. For example, using _reverse_bytes('12345678')
    will return 43218765.
    :param data as bytes
    :return: reversed bytes
    """
    data_size = 0
    for n in data:
        data_size += 1

    reversed_bytes = bytearray()
    i = 0
    for x in range(0, data_size // 4):
        a = (data[i:i + 4])
        i += 4
        z = 0

        n0 = a[z]
        n1 = a[z + 1]
        n2 = a[z + 2]
        n3 = a[z + 3]
        reversed_bytes.append(n3)
        reversed_bytes.append(n2)
        reversed_bytes.append(n1)
        reversed_bytes.append(n0)

    return bytes(reversed_bytes)