Search code examples
rubyarrayspackunpack

Why is hex -> base64 so different from base64 -> hex using pack and unpack?


I got this code working, which converts from hex to base64, and vice versa. I got to_base64 from another SO question, and I wrote to_hex with some guesswork and trial and error.

class String

  def to_base64
    [[self].pack("H*")].pack("m0")
  end

  def to_hex
    self.unpack("m0").first.unpack("H*").first
  end
end

But I don't really grok the pack and unpack methods, even after reading the docs. Specifically, I'm confused by the asymmetry between the two implementations. Conceptually, in both cases, we take a string encoded in some base (16 or 64), and we wish to convert it to another base. So why can't we implement to_hex like this:

def to_hex
  [[self].pack("m0")].pack("H*")
end

or to_base64 using unpack? Why does the base we chose completely change the method we need to use to accomplish conversions?


Solution

  • to_hex is the exact inverse of to_base64:

    to_base64

    1. put string in an array: [self]
    2. call pack with H*: [self].pack("H*")
    3. put string in an array: [[self].pack("H*")]
    4. call pack with m0: [[self].pack("H*")].pack("m0")

    to_hex

    1. call unpack with m0: self.unpack("m0")
    2. extract string from array: self.unpack("m0").first
    3. call unpack with H*: self.unpack("m0").first.unpack("H*")
    4. extract string from array: self.unpack("m0").first.unpack("H*").first

    That's how you undo operations, by applying the inverse operations:

    a = 5
    (a + 4) * 3
    #=> 27
    

    And the other way around:

    a = 27
    (a / 3) - 4
    #=> 5
    

    a.pack is the inverse of a.unpack and a.first is the inverse of [a]