Why doesn't it work when using Python to encrypt, then Ruby to decrypt?

Question

I'm trying to encrypt then base64_encode a string using Python, and base64_decode then decrypt the result with Ruby, but the string is messed up.

I don't see any difference in the two methods. I tried using the 128-bit AES-CFB algorithm, but had no success.

Here's my Python code:

from Crypto.Cipher import AES
from Crypto.Util.randpool import RandomPool
from base64 import standard_b64encode, standard_b64decode

key = "abcdefghijklmnop"
en = AES.new(key, AES.MODE_CFB, "0000000000000000")
cipher = en.encrypt("apple")
cipher64 = standard_b64encode(cipher)

cipher64 contains: WqF9Zj0=

My Ruby code is:

require "openssl"
require 'digest/sha2'
require 'base64'


de = OpenSSL::Cipher::Cipher.new("aes-128-cfb")
de.decrypt
de.key = "abcdefghijklmnop"
de.iv = "0000000000000000"
plain = de.update("WqF9Zj0=".unpack('m')[0])
de.final
puts plain

plain contains a different string than "apple". I get the same result if my string's length is 16, to avoid a padding problem.

I guess it's a parameter problem, but I can't figure out what. Does anyone have an idea?

falsetru · Accepted Answer · 2013-07-30 17:47:50Z

1

PyCrypto uses a different segment_size

When encoding, specify segment_size, and pad the plain text.

from base64 import standard_b64encode

from Crypto.Cipher import AES

def pad(x, n=16):
    p = n - (len(x) % n)
    return x + chr(p) * p

key = "abcdefghijklmnop"
en = AES.new(key=key, mode=AES.MODE_CFB, IV="0" * 16, segment_size=128)
cipher = en.encrypt(pad("apple"))
cipher64 = standard_b64encode(cipher)
print cipher64

Using the above code, you will get apple\x03\x03\x03. On the Ruby side, you should remove the padding.

Ruby part (decoding):

require "openssl"

de = OpenSSL::Cipher::Cipher.new("aes-128-cfb")
de.decrypt
de.key = "abcdefghijklmnop"
de.iv = "0000000000000000"
plain = de.update(ARGV[0].unpack('m')[0]) + de.final
plain = plain[0...-plain[-1].ord]
puts plain

Alternatively, you can use M2Crypto, which has no need to remove the padding on the Ruby side:

from base64 import standard_b64encode

import M2Crypto.EVP

key = "abcdefghijklmnop"
iv = "0000000000000000"
en = M2Crypto.EVP.Cipher('aes_128_cfb', key, iv, 1)
cipher = en.update('apple') + en.final()
cipher64 = standard_b64encode(cipher)
print cipher64

edited Jul 30, 2013 at 17:47

answered Jul 24, 2013 at 14:03

falsetru

371k69 gold badges769 silver badges659 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

user2194805 Over a year ago

Thanks, but i still have problem if the input apple string is longer than 8 characters. On python side i' ve:

from Crypto.Cipher import AES    en = AES.new(key, AES.MODE_CFB, "0000000000000000", segment_size = 64)   cipher = en.encrypt(pad(sys.argv[1]))

and on ruby side i' ve: require "openssl" decryptobj = OpenSSL::Cipher::Cipher.new("AES-256-CFB") . Unfortunately M2Crypto is not reachable in my case.

user2194805 Over a year ago

Ah, found the solution finally: i need to use segment_size = 128 by encoding and 16 byte long padding as well :- ).

falsetru Over a year ago

@user2194805, I fixed that. I also added ruby code that strip padding.

sscirrus Over a year ago

Note for future users - this answer almost tripped me up. The segment size is ONLY an issue in CFB mode.

Collectives™ on Stack Overflow

Why doesn't it work when using Python to encrypt, then Ruby to decrypt?

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related