Here is my approach:

1. Treat regular expression as a finite state machine.
2. Draw states and transitions accordingly.
3. Eliminate intermediate states.

This kata has only a few states, which can be solved by pen and paper.

I am totally agreed with semiessessi. It's a shame to call this piece of **** best practice.

This what I got when i submit.

File "", line 74     
   	Test.it("Testing ASCII85 encoding of %s" % `unit`)
                                            ^
SyntaxError: invalid syntax

Please fix this testcase for Python 3.4.3

I was stuck at input functionality test where the first two testcases have same Whitespace code and input, but expect different output:
Code："ssstntnttssstsntnttsssttntnttsssttntttssstsntttssstnttttnsttnsttnstnnn"
Input : "1","2","3"
Expect output 1 : "123"(passed)
Expect output 2 : "867"(failed)
My output : "123"
Language : Python

Did I miss something? Or there's something wrong with the testcases?