in python 2 and python 3

[Skaperen]

in python 2 and python 3 i want to test if a variable refers to any kind of string.  in python 2 i can do if isinstance( var, (str,unicode) ): or if isinstance( var, basestring ): and in python 3 i can do if isinstance( var, str ): but i want to have one coding that works in python 2 and in python 3 without installing anything extra (so someone can just download this code and run it as-is).

i'd also like to find a way to test for any integer (in python 2, including long).

it seems like i will need to test for the python version somehow.

[stranac]

If you want to write code that works on both pyzhon 2 and python 3, you should be using six
In this particular case, six.string_types is what you want.

If you really want to do it without using third-party modules, you will have to manually check for the version, and use if/else blocks to run the appropriate code every time you use something that is different between the two versions.
IMO, it's not worth it at all.

[wavic]

If you check in the beginning of the code the Python version and do something similar to this:

if sys.version_info[0] == 2:
    range = xrange

Maybe inherit the unicode class in str class in Python 2 and predefine __str__ and __repr__ so awlays to return Unicode. I might be not so clear but I just start to learn classes in Python.

[metulburr]

it seems like i will need to test for the python version somehow.

I will often do this for my games, as people usually do not give it a second chance or want to hassle much before giving up on your game and moving on. So i usually make sure my programs runs in both py2 and py3.

You can also define a function based on which version the code is being ran under. And example of importing different module renames and defining a the same function under both versions. You can put this stuff in the beginning of your script, or if its a large program, i usually make a separate module for version checking and version definitions

import sys
if sys.version[0] == '3':
    from urllib.request import urlopen
    def do_something():
        #do something via py3 method
        pass 
else:
    from urllib2 import urlopen
    range = xrange 
    def do_something()
        #do something via py2 method
        pass 

[ichabod801]

You can often manage this without importing sys, depending on what you're doing and how complicated it is:

try:
    input = raw_input
except NameError:
    pass

[snippsat]

As mention six is more elegant than import sys  try: except stuff.
A lot is under six.moves.
Also use future then can take quite a lot.

from __future__ import absolute_import, division, print_function, unicode_literals
from six.moves import input
from six.moves.urllib.request import urlopen
from six.moves import range

I think is best to not write code for Python 2.x anymore.
To move faster forward,Python 3.6 aslo has a lot features that can not be back ported.
Django 2.0 will drop Python 2.7 and 3.4,i think this is the right move.

Django 2.0 supports Python 3.5+. Since Django 1.11, support for Python 2.7 and 3.4 is removed.

[Skaperen]

the script people download needs to work without them downloading or installing anything else, nor having done so in the past, whether they run it on a recent python 2 or run it on a recent python 3.  importing sys and testing sys.version_info is very reasonable to do.  i have done the range = xrange thing wavic suggested, before.  i have no problem with that, either.  but i don't understand ichabod801's suggestion.

[metulburr]

the script people download needs to work without them downloading or installing anything else, nor having done so in the past, whether they run it on a recent python 2 or run it on a recent python 3.

the module six, and sys are in the standard library so they wont have to install anything extra either way.

but i don't understand ichabod801's suggestion.

he is saying to utilize python3.x where raw_input is no longer defined to determine python version

You can often manage this without importing sys, depending on what you're doing and how complicated it is:

try:
    input = raw_input
except NameError:
    pass

python 2.x

metulburr@ubuntu:~$ python
Python 2.7.12 (default, Nov 19 2016, 06:48:10) 
[GCC 5.4.0 20160609] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> input = raw_input
>>>

python3.x

metulburr@ubuntu:~$ python3
Python 3.5.2 (default, Nov 17 2016, 17:05:23) 
[GCC 5.4.0 20160609] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> input = raw_input
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'raw_input' is not defined
>>> 

in py2 both input and raw_input are defined, whereas py3 only input is. So if you catch a NameError, it will pass in both versions, but it gives you an option to throw version specific code in the try/except block. 

try:
    input = raw_input
    #your py2 code here
except NameError:
    #your py3 code here

*Of course if you get a NameError with whatever py2 specific code you have, then that will mess things up

[wavic]

@ichabod801 suggested what I did but in try statement. "Easier to ask for forgiveness than permission"

[ichabod801]

he is saying to utilize python3.x where raw_input is no longer defined to determine python version

Well, not exactly. What I meant to say was that in some cases, you don't need sys to determine the version. In that case I had a script that I wanted to work in either 2.x or 3.x, but I wanted to use the 3.x version of input (which is 2.x's raw_input) no matter what. I wasn't suggesting that you use raw_input's existence as a way to test for version if you wanted to do something else across the 2/3 barrier.