Implementation of SHA-1 Hash in pure ruby

SHA1.rb

#http://stackoverflow.com/questions/5940316/left-rotate-through-carry-in-ruby
class Integer
  def lotate(n=1)
    self << n | self >> (32 - n)
  end
end

# FIPS 180-2 -- relevant section #'s below
# Pulls parts from Wiki pseudocode and http://ruby.janlelis.de/17-sha-256
class SHA1

  def self.digest(input)
    # 5.3.1 - Initial hash value
    z = 0x67452301, 0xEFCDAB89, 0x98BADCFE, 0x10325476, 0xC3D2E1F0
    
    # 5.1.1
    length = input.length*8
    input << 0x80
    input << 0 while input.size%64 != 56
    input += [length].pack('Q').reverse

    #6.1.2
    input.unpack('C*').each_slice(64){|chunk|
      
      #6.1.2 - 1. Prepare the message schedule
      w = []
      chunk.each_slice(4){|a,b,c,d| w << (((a<<8|b)<<8|c)<<8|d) }

      #Expand from sixteen to eighty -- 6.1.2 - 1. Prepare the message schedule
      (16..79).map{|i|
        w[i] = (w[i-3] ^ w[i-8] ^ w[i-14] ^ w[i-16]).lotate & 0xffffffff
      }

      #6.1.2 - 2. Initialize the five working variables.
      a,b,c,d,e = z

      (0..79).each{|i|  # 6.1.2 - 3. & 4.1.1 - SHA-1 Functions
        case i
        when 0..19
          f = ((b & c) | (~b & d))
          k = 0x5A827999
        when 20..39
          f = (b ^ c ^ d)
          k = 0x6ED9EBA1
        when 40..59
          f = ((b & c) | (b & d) | (c & d))
          k = 0x8F1BBCDC
        when 60..79
          f = (b ^ c ^ d)
          k = 0xCA62C1D6
        end

        temp = (a.lotate(5) & 0xffffffff) + f + e + k + w[i] & 0xffffffff

        e = d
        d = c
        c = b.lotate(30) & 0xffffffff
        b = a
        a = temp
      }

      # 6.1.2 - 4. Compute the ith intermediate hash value
      z[0] = z[0] + a & 0xffffffff
      z[1] = z[1] + b & 0xffffffff
      z[2] = z[2] + c & 0xffffffff
      z[3] = z[3] + d & 0xffffffff
      z[4] = z[4] + e & 0xffffffff
    }

    # Produce the final hash value (big-endian)
    return '%.8x'*5 % z
  end
end

if $0 == __FILE__
  input = gets(nil) || ''
  if RUBY_VERSION >= '1.9'
    input = input.force_encoding('US-ASCII')
  end

  puts SHA1.digest(input)
end

Github doesn't notify you about gist comments apparently... I wrote this for a college network security/crypto class so I didn't expect it to be perfect. @strags did you correct the padding issue you described? If you could share the fix that would be awesome! I would like to keep this implementation as accurate as possible for anyone else who stumbles across it.

@tstevens I wrote a fix for your padding algorithm; check out my fork of this gist. I verified that it gives the same output as the Ruby library's SHA-1 for messages whose length % 64 bytes (or 512 bits) fall into that weird range (56..63) where the padding has to be adjusted.

There's still a bug, though! When I try to hash a message that's >= 64 bytes, I get a NoMethodError in your leftrotate() function. Can you verify this?

Thanks for making your implementation free for all to use!

Over a year behind but I finally fixed this up. Both the NoMethodError and padding issues have been resolved with this update!

Hey, can you please explain this expression?
'%.8x'*5 % z
Thanks!