joshua-davis-rose / slowaes Goto Github PK

def strip_PKCS7_padding(s):
    """return s stripped of PKCS7 padding"""
    if len(s)%16 or not s:
        raise ValueError("String of len %d can't be PCKS7-padded" % len(s))
    numpads = ord(s[-1])
    if numpads > 16:
        raise ValueError("String ending with %r can't be PCKS7-padded" % s[-1])
    return s[:-numpads]


slowaes-0.1a1

Instead of:

   if(c[3] >= 0 && c[2] < 64){

should be:

   if(c[3] >= 0 && c[3] < 64){

Best regards,
Igor.

missed an '$' there on line #326

See attachment for a modified version of decrypt that doesn't require 
originalsize as a parameter.

You need to add some example code. I was straggling with the PHP implementation 
and had lot of problems. In any case the data to be send is not an array, so 
please put down some examples about strings being encrypted and decrypted. I 
personally need a working solution for JavaScript and PHP communicating with 
each-other, but any code explaining the steps to make it work is OK. I can 
contribute some examples myself if I can figure it out.

"i" should be a variable and have to be written as "$i"

encryptData should take IV as an optional argument, and only generate one 
randomly if None is passed in.

What are the requirement to complete this task?

Implement PKCS7 padding as described on
http://code.google.com/p/slowaes/wiki/OpenSSLStandards

Please use labels and text to provide additional information.

Hello-

I have a few reference implementations of AES-256-CBC in C#, Ruby, PHP, and 
Flash.

I am know trying to get it to talk to JS via slow aes using the same settings.

I've been toiling for hours trying to get it to work, so I thought I'd ask to 
see if anyone had any ideas.

On a working system, using the following settings:

IV = 0123456789012345
key = 67a4f45f0d1d9bc606486fc42dc49416
plaintext = yay!

Across my reference implementations, I get the following Base64 encoded output:

q/xJqqN6qbiZMXYmiQC1Fw==

I am having trouble setting up slowaes to give me the same. Can anyone give me 
some pointers?

tia

In
http://slowaes.googlecode.com/svn/trunk/js/cryptoHelpers.js Revision 39.

There is a couple of minor syntax issues (found when running it through JSLint).

The issues don't occur when using it in browser, as the browser auto-fixes them 
for you.

But if you run it in a stricter environment (e.g. Windows Script Component), 
then it complains.

I've attached a patch file. 

btw. Awesome project guys, keep up the good work.

What steps will reproduce the problem?
1. run code in "2." (i am running a gs script in google spreadsheets)
2.
function aesTester() {
  var password = "mypw";
  // generate the private key based on user's input password
  var privateKey = cryptoHelpers.generatePrivateKey(password, 128);

  var myplaintext = "hello plain text";
  // using OPENSSL_256 key size but could use
  var keysize = slowAES.aes.keySize.SIZE_256;
  var mode = slowAES.modeOfOperation.CBC;
  var key = cryptoHelpers.toNumbers("5e884898da28047151d0e56f8dc6292773603d0d6aabbdd62a11ef721d1542d8");
  // initialization vector, make this unique and RANDOM!!
  var iv = cryptoHelpers.toNumbers("6bbda7892ad344e06c31e64564a69a9a");
  var plaintext = cryptoHelpers.convertStringToByteArray(myplaintext);
  // keysize is no longer a parameter because it auto finds size
  var encrypted = slowAES.encrypt(plaintext, mode, privateKey, iv);
  var base64 = cryptoHelpers.base64.encode(encrypted);
  // this is the plain text
  Logger.log(myplaintext);
  // this is the encrypted text
  Logger.log(base64);
  // decryption in reverse order of encryption
  var decodedText = cryptoHelpers.base64.decode(base64);
  // keysize is no longer a parameter because it auto finds size
  var decrypted = slowAES.decrypt(decodedText, mode, privateKey, iv);
  var decryptedText = cryptoHelpers.convertByteArrayToString(decrypted);
  // this is the decrypted text
  Logger.log(decryptedText );
}

What is the expected output? What do you see instead?
expected output is the input but you will see an extra character at the end. I 
have tracked the bug down to the "unpadBytesOut" function. it does not remove 
the "=" from the string. should the unpadding be done AFTER the 
decryption??!!?? or should it be removed before. (removing a decrypted "=" 
sounds difficult.)


What version of the product are you using? On what operating system?
not sure, i just copied the code today. running on windows 7. google chrome. 
google docs as a .gs (google script). js code runs as-is.

Please provide any additional information below.
i can try to figure out the fix, but hopefully someone beats me to it.

What is the expected output? What do you see instead?
A License File checked into the repository


For packaging it would be nice to have a LICENSE file stating a copy of the 
license for the package.

What are the requirement to complete this task?

The encrypt and decrypt methods should accept an array of bytes instead of
a string. Any conversion code within the methods should be removed under
the assumption that the conversion will occur before the method is called.

Please use labels and text to provide additional information.

What are the requirement to complete this task?

The encrypt and decrypt methods should accept an array of bytes instead of
a string. Any conversion code within the methods should be removed under
the assumption that the conversion will occur before the method is called.

Please use labels and text to provide additional information.

Hi guys.

I made an version for python 3 using bytes and bytearrays.
Attach here. Tested with some utf-8 string with åäö letters :-)

What steps will reproduce the problem?

test cases, these should both return the array [97,97]

cryptoHelpers.base64.decode(cryptoHelpers.base64.encode([97,97]))
97,97,64

cryptoHelpers.base64.decode("YWE=")
97,97,64

What do you see instead?
The array: [97,97,64]


What version of the product are you using?
revision 39 from svn, running under Mozilla Rhino

My proposed fix is attached.  After this patch is applied:

js> cryptoHelpers.base64.decode(cryptoHelpers.base64.encode([97,97]))
97,97

js> cryptoHelpers.base64.decode("YWE=")
97,97

What are the requirement to complete this task?

Implement PKCS7 padding as described on
http://code.google.com/p/slowaes/wiki/OpenSSLStandards

Please use labels and text to provide additional information.

If you take the encrypted cipher resulting from any CBC operation and
decrypt the message using the same IV, key, key size, etc, the decrypted
string is longer than expected and results in gremlins being appended to
the end of the decrypted text.

SlowAES seems to only work in a round-trip scenario where the unencrypted
text is known at the time.

Using Josh Davis' implementation, you can see the issue first hand:

http://josh-davis.org/ecmaScrypt

Password: Test001!
256 Bit
Hex Key: e8cfe0b4e39dcc886ca83116b21db375ca231eeebf1df98f0b38163221385fcd
CBC
Encrypted String:
88ff1654ef2badfc22bc7dfa6c0d5d145cbfa547bce4d87c663473b09ce8d9c3

The result you will get is 'Hello, World! This is a test' but at the end
there will be extra characters at the end.

I tried using the fix submitted here:
http://code.google.com/p/slowaes/issues/detail?id=6

which works correctly unless I pass in a 16 character hex string as the
plain text.

What steps will reproduce the problem?
1. Create an encrypted data with static KEY and IV
2. Once you try to decrypt the data with the same KEY and different IV the 
message is displayed and only the first line is scrambled

What is the expected output? What do you see instead?
I was expecting that once the IV is changed the whole message will be scrambled 
but instead it is only the first line

What version of the product are you using? On what operating system?
The last version available for download.

When encrypting 16 bytes long blocks (and multiples of 16 bytes) of
plaintext with the Javascript version of SlowAES/CBC the padding block is
not appended.

PKCS7 needs a full 16-bytes padding block made of 0x10 added to the
plaintext before encryption in such case.

Thanks for the library!
stefano

Small point.  Your test code has a length of 15 characters, fitting in a single 
block, so your example does not test the block sequencing methods at all.

How about a text string:

"This is a test with several blocks!"