Added block padding to be able to work with leading zeroes, breaks all kind of stuff

4 files changed, 70 insertions, 11 deletions

diff --git a/rsa/__init__.py b/rsa/__init__.py
index d9999ae..7fda4e1 100755
--- a/rsa/__init__.py
+++ b/rsa/__init__.py
@@ -42,6 +42,19 @@ def decode64chops(string):
 
     return chops
 
+def block_size(n):
+    '''Returns the block size in bytes, given the public key.
+
+    The block size is determined by the 'n=p*q' component of the key.
+    '''
+
+    # Set aside 2 bits so setting of safebit won't overflow modulo n.
+    nbits = rsa.common.bit_size(n) - 2
+    nbytes = nbits / 8
+
+    return nbytes
+
+
 def chopstring(message, key, n, int_op):
     """Chops the 'message' into integers that fit into n.
     
@@ -56,13 +69,10 @@ def chopstring(message, key, n, int_op):
     Used by 'encrypt' and 'sign'.
     """
 
-    msglen = len(message)
-    mbits = msglen * 8
 
-    # Set aside 2 bits so setting of safebit won't overflow modulo n.
-    nbits = rsa.common.bit_size(n) - 2
+    nbytes = block_size(n)
 
-    nbytes = nbits / 8
+    msglen = len(message)
     blocks = msglen / nbytes
 
     if msglen % nbytes > 0:
@@ -90,14 +100,25 @@ def gluechops(string, key, n, funcref):
 
     messageparts = []
     chops = decode64chops(string)  #Decode base64 strings into integer chops
+
+    nbytes = block_size(n)
     
     for chop in chops:
         value = funcref(chop, key, n) #Decrypt each chop
         block = rsa.transform.int2bytes(value)
+
+        # Pad block with 0-bytes until we have reached the block size
+        blocksize = len(block)
+        padsize = nbytes - blocksize
+        if padsize < 0:
+            raise ValueError('Block larger than block size (%i > %i)!' %
+                    (blocksize, nbytes))
+        elif padsize > 0:
+            block = '\x00' * padsize + block
+
         messageparts.append(block)
 
     # Combine decrypted strings into a msg
-    
     return ''.join(messageparts)
 
 def encrypt(message, key):
diff --git a/rsa/transform.py b/rsa/transform.py
index 9cd1ef3..6719dcb 100644
--- a/rsa/transform.py
+++ b/rsa/transform.py
@@ -12,7 +12,9 @@ def bit_size(number):
     return int(math.ceil(math.log(number,2)))
 
 def bytes2int(bytes):
-    """Converts a list of bytes or a string to an integer
+    """Converts a list of bytes or an 8-bit string to an integer.
+
+    When using unicode strings, encode it to some encoding like UTF8 first.
 
     >>> (((128 * 256) + 64) * 256) + 15
     8405007
diff --git a/tests/test_binary.py b/tests/test_binary.py
new file mode 100644
index 0000000..f770b72
--- /dev/null
+++ b/tests/test_binary.py
@@ -0,0 +1,37 @@
+'''Tests string operations.'''
+
+import struct
+import unittest
+
+import rsa
+
+class BinaryTest(unittest.TestCase):
+
+    def setUp(self):
+        (self.pub, self.priv) = rsa.newkeys(64)
+
+    def test_enc_dec(self):
+
+        message = struct.pack('>IIII', 0, 0, 0, 1) + 20 * '\x00'
+        print "\tMessage:   %r" % message
+
+        encrypted = rsa.encrypt(message, self.pub)
+        print "\tEncrypted: %r" % encrypted
+
+        decrypted = rsa.decrypt(encrypted, self.priv)
+        print "\tDecrypted: %r" % decrypted
+
+        self.assertEqual(message, decrypted)
+
+    def test_sign_verify(self):
+
+        message = struct.pack('>IIII', 0, 0, 0, 1) + 20 * '\x00'
+        print "\tMessage:   %r" % message
+
+        signed = rsa.sign(message, self.priv)
+        print "\tSigned:    %r" % signed
+
+        verified = rsa.verify(signed, self.pub)
+        print "\tVerified:  %r" % verified
+
+        self.assertEqual(message, verified)
diff --git a/tests/test_strings.py b/tests/test_strings.py
index c5803e4..8baa63d 100644
--- a/tests/test_strings.py
+++ b/tests/test_strings.py
@@ -11,8 +11,7 @@ class StringTest(unittest.TestCase):
 
     def test_enc_dec(self):
 
-        # TODO: test with unicode strings and non-ascii chars
-        message = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+        message = u"Euro=\u20ac ABCDEFGHIJKLMNOPQRSTUVWXYZ".encode('utf-8')
         print "\tMessage:   %s" % message
 
         encrypted = rsa.encrypt(message, self.pub)
@@ -25,8 +24,8 @@ class StringTest(unittest.TestCase):
 
     def test_sign_verify(self):
 
-        # TODO: test with unicode strings and non-ascii chars
-        message = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+        message = u"Euro=\u20ac ABCDEFGHIJKLMNOPQRSTUVWXYZ".encode('utf-8')
+        print "\tMessage:   %s" % message
 
         signed = rsa.sign(message, self.priv)
         print "\tSigned:    %s" % signed