Encode a string in python 3.4

Question

I am getting this error when I run this script in python 3.4

Traceback (most recent call last):
File "qcdlcomm.py", line 23, in <module>
if not call('whoami')[0] == 'root':
File "qcdlcomm.py", line 20, in call
return  subprocess.Popen(cmd,shell=True,stdout=subprocess.PIPE).stdout.read().strip().split("\n")
TypeError: 'str' does not support the buffer interface

I think the answer is here TypeError: 'str' does not support the buffer interface but I don't know how to implement it.

qcdlcomm.py

Answer 1

bytes.split() method does not accept str (Unicode type in Python 3):

>>> b'abc'.split("\n")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: Type str doesn't support the buffer API

The error message is improved in Python 3.5:

>>> b"abc".split("\n")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: a bytes-like object is required, not 'str'

"\n" (str type) is a Unicode string (text) that is not bytes-like (binary data) in Python 3.

---

To get the output of whoami command as a Unicode string:

#!/usr/bin/env python
from subprocess import check_output

username = check_output(['whoami'], universal_newlines=True).rstrip("\n")

universal_newlines enables text mode. check_output() redirects child's stdout automatically and raises an exception on its nonzero exit status.

Note: shell=True is unnecessary here (you don't need the shell, to run whoami).

---

Unrelated: to find out whether you are root in Python, you could use geteuid():

import os

if os.geteuid() == 0:
   # I'm root (or equivalent e.g., `setuid`)

If you need to find out what is the current user name in Python:

import getpass

print('User name', getpass.getuser())

Beware: don't use getuser() for access control purposes!

Answer 2

Strings are treated differently in python3. Strings are not the same as bytes. What you get from calling whoami is a byte array. To fix the call function, simply decode the bytes into a string:

def call(cmd,hide_stderr=True):                                                 
    return subprocess.Popen(cmd,shell=True,stdout=subprocess.PIPE).stdout.read().decode("utf-8").strip().split("\n")

This is a similar issue as mentioned in Convert bytes to a Python string.

Note that this is only a quick fix, but will work for your use case. Simply insert .decode('utf-8') into line 20 between .read() and .strip().

Answer 3

Generally if you want to use text mode with Popen in Python 3 (and assuming that you use cPython implementation), you can pass either encoding, error or universal_newlines argument to Popen constructor. For example:

return subprocess.Popen(cmd, shell=True, stdout=subprocess.PIPE, universal_newlines=True).stdout.read().strip().split("\n")

Thats because in cPython implementation of the __init__ in Popen there is a following line:

text_mode = encoding or errors or universal_newlines

This is also what check_output in cPython internally does - it assures that input keyword argument (if present and None) is of appropriate type and passes the keyword arguments to run (which later calls Popen):

def check_output(*popenargs, timeout=None, **kwargs):
    if 'stdout' in kwargs:
        raise ValueError('stdout argument not allowed, it will be overridden.')

    if 'input' in kwargs and kwargs['input'] is None:
        # Explicitly passing input=None was previously equivalent to passing an
        # empty string. That is maintained here for backwards compatibility.
        kwargs['input'] = '' if kwargs.get('universal_newlines', False) else b''

    return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
**kwargs).stdout

See more in subprocess.py source.