What's changed?

Unicode / bytes

Print function

Lazy default with iterators (e.g. zip, dict.items, map)

Library rearrangement (e.g. urllib)

Integer / float division

Exceptions

http://python3porting.com/differences.html

Examples

# Python 2.7.6
>>> print u'\u0040'
@
>>> print '\u0040'
\u0040
>>> print xrange(10)
xrange(10)
>>> print range(10)
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

# Python 3.4.0
>>> print('\u0040')
@
>>> print(b'\u0040')
b'\\u0040'
>>> print(range(10))
range(0, 10)
>>> print(list(range(10)))
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

A bit about Unicode

bytestrings are strings of 8-bit characters:  the glyph they map to is determined by encoding.

unicode strings are strings of code points/glyphs:  the raw bytes are abstracted away.

encoding takes an encoded string of glyphs and returns a bytestring.

decoding takes a bytestring and returns the glyphs respective to the encoding that they are in.

codecs provide methods to encode or decode bytestrings.

Differences in Unicode Implementation

Python 2:

bytestring default

print uses local encoding

implicit encoding/decoding