MD5 在 Ruby 中的实现

发布于 2025-01-13 16:33:31 字数 4014 浏览 5 评论 0 原文

我正在尝试在 Ruby 中实现 MD5,遵循 wiki 中编写的伪代码

这是代码,运行效果不佳:

# : All variables are unsigned 32 bit and wrap modulo 2^32 when calculating
# s specifies the per-round shift amounts
s = []
s.concat([7, 12, 17, 22,  7, 12, 17, 22,  7, 12, 17, 22,  7, 12, 17, 22])
s.concat([5,  9, 14, 20,  5,  9, 14, 20,  5,  9, 14, 20,  5,  9, 14, 20])
s.concat([4, 11, 16, 23,  4, 11, 16, 23,  4, 11, 16, 23,  4, 11, 16, 23])
s.concat([6, 10, 15, 21,  6, 10, 15, 21,  6, 10, 15, 21,  6, 10, 15, 21])

# Use binary integer part of the sines of integers (Radians) as constants:
k = 0.upto(63).map do |i|
  (Math.sin(i+1).abs * 2 ** 32).floor
end

# Initialize variables:
a0 = 0x67452301 # A
b0 = 0xefcdab89 # B
c0 = 0x98badcfe # C
d0 = 0x10325476 # D

message = File.read(ARGV[0])

# Pre-processing
# with bit stream "string" (MSB)
bits = message.unpack('B*')[0]

org_len = bits.size

bits << '1' # adding a single 1 bit
bits << '0' while !((bits.size + 64) % 512 == 0) # padding with zeros
bits << (org_len % 2 ** 64).to_s(2).rjust(64, '0')


message32 = [bits].pack('B*').unpack('V*') # ?
# 1. bits.scan(/(.{8})/).flatten.map { |b| b.to_i(2).to_s(16).rjust(2, '0') }.each_slice(16) { |c| puts c.join(' ')} => for test
# 2. [bits].pack('B*').unpack('N*') == bits.scan(/(.{32})/).flatten.map { |b| b.to_i(2).to_s(16).rjust(8, '0').to_i(16) } => true



# custom operations for wrapping the results as modulo 2 ** 32
class Integer
  def r
    self & 0xFFFFFFFF
  end

  def rotate_l(count)
    (self << count).r | (self >> (32 - count))
  end
end

# Process the message in successive 512-bit chunks:
message32.each_slice(16).each do |m|
  a = a0
  b = b0
  c = c0
  d = d0

  0.upto(63) do |i|
    if i < 16
      f = d ^ (b & (c ^ d))
      g = i
    elsif i < 32
      f = c ^ (d & (b ^ c))
      g = (5 * i + 1) % 16
    elsif i < 48
      f = b ^ c ^ d
      g = (3 * i + 5) % 16
    elsif i < 64
      f = c ^ (b | ~d)
      g = (7 * i) % 16
    end

    f = (f + a + k[i] + m[g]).r
    a = d
    d = c
    c = b
    b = (b + f.rotate_l(s[i])).r
  end

  a0 = (a0 + a).r
  b0 = (b0 + b).r
  c0 = (c0 + c).r
  d0 = (d0 + d).r
end

puts [a0, b0, c0, d0].pack('V*').unpack('H*')

我正在使用消息进行测试,众所周知,仅在一个块中发生冲突:

它们产生相同的值,但不正确:

❯ ruby md5.rb message1.bin
816922b82e2f8d5bd3abf90777ad72c9

❯ ruby md5.rb message2.bin
816922b82e2f8d5bd3abf90777ad72c9

❯ md5 message*
MD5 (/Users/hansuk/Downloads/message1.bin) = 008ee33a9d58b51cfeb425b0959121c9
MD5 (/Users/hansuk/Downloads/message2.bin) = 008ee33a9d58b51cfeb425b0959121c9

我对预处理步骤有不确定性。

我检查了预处理后的比特流,第 34 和 35 行的注释、原始消息的写法和填充位都是正确的:

❯ hexdump message1.bin
0000000 4d c9 68 ff 0e e3 5c 20 95 72 d4 77 7b 72 15 87
0000010 d3 6f a7 b2 1b dc 56 b7 4a 3d c0 78 3e 7b 95 18
0000020 af bf a2 00 a8 28 4b f3 6e 8e 4b 55 b3 5f 42 75
0000030 93 d8 49 67 6d a0 d1 55 5d 83 60 fb 5f 07 fe a2
0000040
(byebug) bits.scan(/(.{8})/).flatten.map { |b| b.to_i(2).to_s(16).rjust(2, '0') }.each_slice(16) { |c| puts c.join(' ')}
4d c9 68 ff 0e e3 5c 20 95 72 d4 77 7b 72 15 87
d3 6f a7 b2 1b dc 56 b7 4a 3d c0 78 3e 7b 95 18
af bf a2 00 a8 28 4b f3 6e 8e 4b 55 b3 5f 42 75
93 d8 49 67 6d a0 d1 55 5d 83 60 fb 5f 07 fe a2
80 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 02 00
nil

我错过了什么?

I am trying to implement MD5 in Ruby, following the pseudo code written in the wiki.

Here is the codes, not working well:

# : All variables are unsigned 32 bit and wrap modulo 2^32 when calculating
# s specifies the per-round shift amounts
s = []
s.concat([7, 12, 17, 22,  7, 12, 17, 22,  7, 12, 17, 22,  7, 12, 17, 22])
s.concat([5,  9, 14, 20,  5,  9, 14, 20,  5,  9, 14, 20,  5,  9, 14, 20])
s.concat([4, 11, 16, 23,  4, 11, 16, 23,  4, 11, 16, 23,  4, 11, 16, 23])
s.concat([6, 10, 15, 21,  6, 10, 15, 21,  6, 10, 15, 21,  6, 10, 15, 21])

# Use binary integer part of the sines of integers (Radians) as constants:
k = 0.upto(63).map do |i|
  (Math.sin(i+1).abs * 2 ** 32).floor
end

# Initialize variables:
a0 = 0x67452301 # A
b0 = 0xefcdab89 # B
c0 = 0x98badcfe # C
d0 = 0x10325476 # D

message = File.read(ARGV[0])

# Pre-processing
# with bit stream "string" (MSB)
bits = message.unpack('B*')[0]

org_len = bits.size

bits << '1' # adding a single 1 bit
bits << '0' while !((bits.size + 64) % 512 == 0) # padding with zeros
bits << (org_len % 2 ** 64).to_s(2).rjust(64, '0')


message32 = [bits].pack('B*').unpack('V*') # ?
# 1. bits.scan(/(.{8})/).flatten.map { |b| b.to_i(2).to_s(16).rjust(2, '0') }.each_slice(16) { |c| puts c.join(' ')} => for test
# 2. [bits].pack('B*').unpack('N*') == bits.scan(/(.{32})/).flatten.map { |b| b.to_i(2).to_s(16).rjust(8, '0').to_i(16) } => true



# custom operations for wrapping the results as modulo 2 ** 32
class Integer
  def r
    self & 0xFFFFFFFF
  end

  def rotate_l(count)
    (self << count).r | (self >> (32 - count))
  end
end

# Process the message in successive 512-bit chunks:
message32.each_slice(16).each do |m|
  a = a0
  b = b0
  c = c0
  d = d0

  0.upto(63) do |i|
    if i < 16
      f = d ^ (b & (c ^ d))
      g = i
    elsif i < 32
      f = c ^ (d & (b ^ c))
      g = (5 * i + 1) % 16
    elsif i < 48
      f = b ^ c ^ d
      g = (3 * i + 5) % 16
    elsif i < 64
      f = c ^ (b | ~d)
      g = (7 * i) % 16
    end

    f = (f + a + k[i] + m[g]).r
    a = d
    d = c
    c = b
    b = (b + f.rotate_l(s[i])).r
  end

  a0 = (a0 + a).r
  b0 = (b0 + b).r
  c0 = (c0 + c).r
  d0 = (d0 + d).r
end

puts [a0, b0, c0, d0].pack('V*').unpack('H*')

I'm testing with messages, well known for collision in just one block:

They are resulted in the same value, but not correct:

❯ ruby md5.rb message1.bin
816922b82e2f8d5bd3abf90777ad72c9

❯ ruby md5.rb message2.bin
816922b82e2f8d5bd3abf90777ad72c9

❯ md5 message*
MD5 (/Users/hansuk/Downloads/message1.bin) = 008ee33a9d58b51cfeb425b0959121c9
MD5 (/Users/hansuk/Downloads/message2.bin) = 008ee33a9d58b51cfeb425b0959121c9

I have an uncertainty about pre-processing steps.

I checked the bit stream after pre-processing, with the comments on the line 34 and 35, the original message written in same and the padding bits are right:

❯ hexdump message1.bin
0000000 4d c9 68 ff 0e e3 5c 20 95 72 d4 77 7b 72 15 87
0000010 d3 6f a7 b2 1b dc 56 b7 4a 3d c0 78 3e 7b 95 18
0000020 af bf a2 00 a8 28 4b f3 6e 8e 4b 55 b3 5f 42 75
0000030 93 d8 49 67 6d a0 d1 55 5d 83 60 fb 5f 07 fe a2
0000040
(byebug) bits.scan(/(.{8})/).flatten.map { |b| b.to_i(2).to_s(16).rjust(2, '0') }.each_slice(16) { |c| puts c.join(' ')}
4d c9 68 ff 0e e3 5c 20 95 72 d4 77 7b 72 15 87
d3 6f a7 b2 1b dc 56 b7 4a 3d c0 78 3e 7b 95 18
af bf a2 00 a8 28 4b f3 6e 8e 4b 55 b3 5f 42 75
93 d8 49 67 6d a0 d1 55 5d 83 60 fb 5f 07 fe a2
80 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 02 00
nil

What am I missed?

如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

扫码二维码加入Web技术交流群

发布评论

需要 登录 才能够评论, 你可以免费 注册 一个本站的账号。

评论(1

時窥 2025-01-20 16:33:31

实现 MD5 时最典型的错误是破坏字节序:填充的消息和消息长度将根据 little-endian 约定,以便将 ASCII 格式的消息“abc”(0x61 0x62 0x63) 转换为 16 字填充m[0]=0x80636261、m[14]=0x18 和 m[1…13,15]=0 的消息块。

我从来没有用 Ruby 写过任何东西,但我感觉代码会产生 m[0]=0x61626380、m[15]=0x18 和 m[1…14]=0。

另外:根据小端约定,4 字结果也将转换为 16 字节。

The one most classical mistake in implementing MD5 is botching endianness: the padded message and the message length are to be turned to 32-bit words per little-endian convention, so that message 'abc' in ASCII (0x61 0x62 0x63) is turned to a 16-word padded message block with m[0]=0x80636261, m[14]=0x18, and m[1…13,15]=0.

I never wrote anything in Ruby, but I get the feeling the code yields m[0]=0x61626380, m[15]=0x18, and m[1…14]=0.

Also: the 4-word result is to be turned to 16 bytes per little-endian convention too.

~没有更多了~
我们使用 Cookies 和其他技术来定制您的体验包括您的登录状态等。通过阅读我们的 隐私政策 了解更多相关信息。 单击 接受 或继续使用网站,即表示您同意使用 Cookies 和您的相关数据。
原文