Pure python3 implementation of SHA256 based on @thomdixon's pysha2

pysha2 sha2 sha256.py

#!/usr/bin/python3

__base__ = 'https://github.com/thomdixon/pysha2/blob/master/sha2/sha256.py'
__author__ = 'Lukas Prokop'
__license__ = 'MIT'

import copy
import struct
import binascii

F32 = 0xFFFFFFFF

_k = [0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5,
      0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5,
      0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3,
      0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174,
      0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc,
      0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da,
      0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7,
      0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967,
      0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13,
      0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85,
      0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3,
      0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070,
      0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5,
      0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3,
      0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208,
      0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2]

_h = [0x6a09e667, 0xbb67ae85, 0x3c6ef372, 0xa54ff53a,
      0x510e527f, 0x9b05688c, 0x1f83d9ab, 0x5be0cd19]


def _pad(msglen):
    mdi = msglen & 0x3F
    length = struct.pack('!Q', msglen << 3)

    if mdi < 56:
        padlen = 55 - mdi
    else:
        padlen = 119 - mdi

    return b'\x80' + (b'\x00' * padlen) + length


def _rotr(x, y):
    return ((x >> y) | (x << (32 - y))) & F32


def _maj(x, y, z):
    return (x & y) ^ (x & z) ^ (y & z)


def _ch(x, y, z):
    return (x & y) ^ ((~x) & z)


class SHA256:
    _output_size = 8
    blocksize = 1
    block_size = 64
    digest_size = 32

    def __init__(self, m=None):
        self._counter = 0
        self._cache = b''
        self._k = copy.deepcopy(_k)
        self._h = copy.deepcopy(_h)

        self.update(m)

    def _compress(self, c):
        w = [0] * 64
        w[0:16] = struct.unpack('!16L', c)

        for i in range(16, 64):
            s0 = _rotr(w[i-15], 7) ^ _rotr(w[i-15], 18) ^ (w[i-15] >> 3)
            s1 = _rotr(w[i-2], 17) ^ _rotr(w[i-2], 19) ^ (w[i-2] >> 10)
            w[i] = (w[i-16] + s0 + w[i-7] + s1) & F32

        a, b, c, d, e, f, g, h = self._h

        for i in range(64):
            s0 = _rotr(a, 2) ^ _rotr(a, 13) ^ _rotr(a, 22)
            t2 = s0 + _maj(a, b, c)
            s1 = _rotr(e, 6) ^ _rotr(e, 11) ^ _rotr(e, 25)
            t1 = h + s1 + _ch(e, f, g) + self._k[i] + w[i]

            h = g
            g = f
            f = e
            e = (d + t1) & F32
            d = c
            c = b
            b = a
            a = (t1 + t2) & F32

        for i, (x, y) in enumerate(zip(self._h, [a, b, c, d, e, f, g, h])):
            self._h[i] = (x + y) & F32

    def update(self, m):
        if not m:
            return

        self._cache += m
        self._counter += len(m)

        while len(self._cache) >= 64:
            self._compress(self._cache[:64])
            self._cache = self._cache[64:]

    def digest(self):
        r = copy.deepcopy(self)
        r.update(_pad(self._counter))
        data = [struct.pack('!L', i) for i in r._h[:self._output_size]]
        return b''.join(data)

    def hexdigest(self):
        return binascii.hexlify(self.digest()).decode('ascii')


if __name__ == '__main__':
    def check(msg, sig):
        m = SHA256()
        m.update(msg.encode('ascii'))
        print(m.hexdigest() == sig)

    tests = {
        "":
            'e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855',
        "a":
            'ca978112ca1bbdcafac231b39a23dc4da786eff8147c4e72b9807785afee48bb',
        "abc":
            'ba7816bf8f01cfea414140de5dae2223b00361a396177a9cb410ff61f20015ad',
        "message digest":
            'f7846f55cf23e14eebeab5b4e1550cad5b509e3348fbc4efa3a1413d393cb650',
        "abcdefghijklmnopqrstuvwxyz":
            '71c480df93d6ae2f1efad1447c66c9525e316218cf51fc8d9ed832f2daf18b73',
        "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789":
            'db4bfcbd4da0cd85a60c3c37d3fbd8805c77f15fc6b1fdfe614ee0a7c8fdb4c0',
        ("12345678901234567890123456789012345678901234567890123456789"
         "012345678901234567890"):
            'f371bc4a311f2b009eef952dd83ca80e2b60026c8e935592d0f9c308453c813e',
        "00baf6626abc2df808da36a518c69f09b0d2ed0a79421ccfde4f559d2e42128b":
            'b835e56173be2b5b7177d71bf02850dc578ac855ac60f91a108eec253bd5a543'
    }

    for inp, out in tests.items():
        check(inp, out)

I assume 00baf662… is considered as bytestring input to the hash algorithm. In this case, I store the source code of my gist in sha.py and run the following testcase:

>>> def f(content):
...     import hashlib
...     h = hashlib.sha256()
...     h.update(content)
...     return h.hexdigest()
... 
>>> def g(content):
...     import sha
...     h = sha.SHA256()
...     h.update(content)
...     return h.hexdigest()
... 
>>> inp = b"00baf6626abc2df808da36a518c69f09b0d2ed0a79421ccfde4f559d2e42128"
>>> f(inp)
'331f3d1a715ed0104af043d5705ab22bf3311d098814ffbc52a3029e448d8121'
>>> g(inp)
'331f3d1a715ed0104af043d5705ab22bf3311d098814ffbc52a3029e448d8121'
>>> f(inp) == g(inp)
True

As can be seen, the hexdigests match. As such, I cannot reproduce your problem. Did you purposefully assign the same value to a twice in your message?

Hey. Previous comment is correct. I am not yet sure whats the flaw here - but it is the same as with @thomdixon's pysha2 implementation (your gist is based on). If you add one more (64th) character or more to your inp byte string and run the same testcase again - it will unfortunately fail...

@daugela @prokls

This issue was brought up on my repository here: thomdixon/pysha2#6

The error is in this gist's update method. If len(m) == 64 then the update loop becomes:

for i in range(0, 64 // 64):
  self._compress(m[64 * i:64 * (i + 1)])
  self._cache = m[-64 % 64):]

which simplifies to:

self._compress(m[0:64])

# this will cause the finalization (digest() method) to include the previous message, 
# because we set m = self._cache + m in the update() method
self._cache = m[0:] 

To fix this, you can e.g., add an if before storing the _cache:

if len(m) % 64:
  self._cache = m[-(len(m) % 64):]

or you can keep the spirit of the original and process the data as a buffer with a white loop, so that it handles the boundary correctly:

def update(self, m):
    if not m:
        return

    self._cache += m
    self._counter += len(m)

    while len(self._cache) >= 64:
        self._compress(self._cache[:64])
        self._cache = self._cache[64:]

The annoying part of this code is the lack of comments.

@thomdixon Thanks for the details and the fix! I applied to last one to a new revision

@CesarCastellonE Thanks, now I was able to reproduce it!

@manuelinfosec Not sure what you mean. The algorithm and its subroutines themselves are well-documented → https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.180-4.pdf I am not sure what you are looking for.