Hi, first time contributing so please bear with me if I missed something.

When testing out some code based on the contents of the internals section, I believe I ran into a small mistake. The header should contain the size in bytes of the content, but content.length in the example returns the number of characters. Using bytesize returns the right value: https://ruby-doc.org/core-2.4.0/String.html#method-i-bytesize

The example works with its content (what is up, doc?) but the issue manifests with other characters

irb(main):002:0> content = "i have €5 in my pocket"
=> "i have €5 in my pocket"
irb(main):003:0> content.length
=> 22
irb(main):004:0> content.bytesize
=> 24
irb(main):05:0> sha_w_length = Digest::SHA1.hexdigest("blob #{content.length}\0" + content)
=> "9859c2c849cc5591aa2223a6fb697aeaa9a5f7fe"
irb(main):06:0> sha_w_bytesize = Digest::SHA1.hexdigest("blob #{content.bytesize}\0" + content)
=> "3d446f4f877e1bea82e603328a845ad7b036338e"

As expected they result in different sha's. The correct one is the one using bytesize

$ echo -n 'i have €5 in my pocket' | git hash-object --stdin
3d446f4f877e1bea82e603328a845ad7b036338e

This PR just tweaks the example to use bytesize instead of length