I think this is a consequence of the new Unicode support in Python 3+

Here is code copied from C:\Python32\Tools\Scripts\crlf.py (on windows)
==================================================================
    for filename in os.listdir("."):
        if os.path.isdir(filename):
            print(filename, "Directory!")
            continue
        data = open(filename, "rb").read()
        if '\0' in data:
            print(filename, "Binary!")
            continue
        newdata = data.replace("\r\n", "\n")
        if newdata != data:
            print(filename)

===================================================================

When run, it produces this (run under the PyCharm debugger)

===================================================================
C:\Python32\python.exe C:/pydev/zob/zobtest.py
Traceback (most recent call last):
  File "C:/pydev/zob/zobtest.py", line 134, in <module>
    x() 
  File "C:/pydev/zob/zobtest.py", line 126, in x
    if '\0' in data:
TypeError: Type str doesn't support the buffer API

Process finished with exit code 1
===================================================================

Removing the test for "\0" produces this:
===================================================================
C:\Python32\python.exe C:/pydev/zob/zobtest.py
Traceback (most recent call last):
  File "C:/pydev/zob/zobtest.py", line 131, in <module>
    x()  
  File "C:/pydev/zob/zobtest.py", line 126, in x
    newdata = data.replace("\r\n", "\n")
TypeError: expected an object with the buffer interface

Process finished with exit code 1
===================================================================