Some testing around OS I/O encodings in python

encodings-tester.py

import sys, locale

e_grave = u'\xe8'
param = sys.argv[1]
sys.stderr.write('Type: ')
inp = raw_input()

encodings = (('locale.getpreferredencoding()', locale.getpreferredencoding()),
             ('sys.stdout.encoding', sys.stdout.encoding),
             ('utf8', 'utf8'))
             
for name, encoding in encodings:
    print ''
    print '***', name, ':', encoding
    if encoding:
        print "u'\\xe8'          ", e_grave.encode(encoding, 'replace')
        print 'repr(sys.argv[1])', repr(param.decode(encoding, 'replace'))
        print 'repr(raw_input())', repr(inp.decode(encoding, 'replace'))

Ubuntu-console.txt

*** locale.getpreferredencoding() : UTF-8
u'\xe8'           è
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\xe8'

*** sys.stdout.encoding : UTF-8
u'\xe8'           è
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\xe8'

*** utf8 : utf8
u'\xe8'           è
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\xe8'

Ubuntu-pipe.txt

*** locale.getpreferredencoding() : UTF-8
u'\xe8'           è
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\xe8'

*** sys.stdout.encoding : UTF-8
u'\xe8'           è
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\xe8'

*** utf8 : utf8
u'\xe8'           è
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\xe8'

Windows-console.txt

*** locale.getpreferredencoding() : cp1252
u'\xe8'           Þ
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\u0160'

*** sys.stdout.encoding : cp850
u'\xe8'           è
repr(sys.argv[1]) u'\xde'
repr(raw_input()) u'\xe8'

*** utf8 : utf8
u'\xe8'           è
repr(sys.argv[1]) u'\ufffd'
repr(raw_input()) u'\ufffd'

Windows-IDLE.txt

*** locale.getpreferredencoding() : cp1252
u'\xe8'           è
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\xe8'

*** sys.stdout.encoding : cp1252
u'\xe8'           è
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\xe8'

*** utf8 : utf8
u'\xe8'           è
repr(sys.argv[1]) u'\ufffd'
repr(raw_input()) u'\ufffd'

Windows-pipe.txt

*** locale.getpreferredencoding() : cp1252
u'\xe8'           è
repr(sys.argv[1]) u'\xe8'
repr(raw_input()) u'\u0160'

*** sys.stdout.encoding : None

*** utf8 : utf8
u'\xe8'           è
repr(sys.argv[1]) u'\ufffd'
repr(raw_input()) u'\ufffd'