Question

I have a Python command that generates the URL-safe base-64 encoding of the SHA-1 hash of a string:

>>> import base64
>>> import sha
>>> base64.urlsafe_b64encode((sha.new("abc").digest()))
'qZk-NkcGgWq6PiVxeFDCbJzQ2J0='

I want to do the same thing in the bash shell, but I’m running into problems:

me:~$ echo -n "abc" | sha1sum | sed 's/ .*//' 
a9993e364706816aba3e25717850c26c9cd0d89d
me:~$ echo -n "abc" | sha1sum | sed 's/ .*//' | base64
YTk5OTNlMzY0NzA2ODE2YWJhM2UyNTcxNzg1MGMyNmM5Y2QwZDg5ZAo=

I suspect that this is because of the way that sha1sum prints out (in hex format). I think base64 is reading 40 bytes, but it really should only be reading 20. I’ve tried piping it through iconv, but no success there:

me:~$ echo -n "abc" | sha1sum | sed 's/ .*//' 
a9993e364706816aba3e25717850c26c9cd0d89d
me:~$ echo -n "abc" | sha1sum | sed 's/ .*//' | base64
YTk5OTNlMzY0NzA2ODE2YWJhM2UyNTcxNzg1MGMyNmM5Y2QwZDg5ZAo=

What should I be doing here to get the output I’m expecting?

Thanks in advance for the help!

Answer 1

digest() is not the same as hexdigest()

base64.urlsafe_b64encode((sha.new("abc").hexdigest()))
'YTk5OTNlMzY0NzA2ODE2YWJhM2UyNTcxNzg1MGMyNmM5Y2QwZDg5ZA=='

Docstring for digest:

Type:       builtin_function_or_method
Base Class: <type 'builtin_function_or_method'>
String Form:<built-in method digest of _hashlib.HASH object at 0x1c88940>
Namespace:  Interactive
Docstring:  Return the digest value as a **string of binary data**.

EDIT: Try:

echo -n "abc"| openssl sha1 -binary | base64
qZk+NkcGgWq6PiVxeFDCbJzQ2J0=