wow, awesome blog article. Keep writing….

Is is possible to add some quick-options to convert code to 2.x newest style (based on current interpreter), using all possible __future__ imports (so no str, byte, newstyle, and print for 2.7)?

It’s released under Apache License 2.0.

Please don’t take this as pressure – as someone who’s been on sabbatical from work and coding for most of the last year, I respect your “you time”.

Thx

Is there any way to NOT “translate” some parts of my code ? I’m using the struct module, and it doesn’t like the string to unicode translation (‘…’ to u’…’ I mean), because it needs real ascii string.

I don’t have much time to read your code and understand it, because I’m already crossing the deadline of my work, so I will correct by hand the wrong parts for the moment.

It gives me:
ushort = Struct(u’<H')
TypeError: Struct() argument 1 must be string, not unicode

Thx for your work, and sorry for my english

PS: I cooked a cludgy module to make my code work under python2 with 2to3, to use it I add the line "from compat import *" at the beginning of my modules. Here it is:

compat.py
—
# Compatibility for Python 2 & 3
# open (compat_open) and re.compile (compat_regexp)

import sys
import re

if sys.version_info[0] == 3:

import builtins

__builtin__ = builtins

def open(filename, mode='r'):
return __builtin__.open(filename, mode, encoding='utf-8', newline='')

def re_compile(pat, flags=None):
if flags is not None: return re._compileorig(pat, flags)
else: return re._compileorig(pat)

else:

from future_builtins import *
import codecs
import __builtin__

def open(filename, mode='r'):
return codecs.open(filename, mode, encoding='utf-8')

def re_compile(pat, flags=0):
return re._compileorig(pat, flags|re.UNICODE)

#__builtin__.__open_orig = __builtin__.open
#__builtin__.open = _compat_open

re._compileorig = re.compile
re.compile = re_compile