Supporting Python 3

• About this book
  • A note on terminology
• Foreword
• Welcome to Python 3
  • The time is nigh
  • What if I can’t switch right now?
  • Python and its versions
  • Further resources
• Migration strategies
  • Only supporting Python 3
  • Separate branches for Python 2 and Python 3
  • Converting to Python 3 with 2to3
    • Using Distribute to support the 2to3 conversion
  • Python 2 and Python 3 without conversion
  • Using 3to2
  • Which strategy is for you?
    • Applications
    • Python modules and packages
  • Summary
• Preparing for Python 3
  • Run under Python 2.7.
  • Use // instead of / when dividing integers
  • Use new-style classes
  • Separate binary data and strings
  • When sorting, use key instead of cmp
  • Use rich comparison operators
    • Comparatively tricky
    • Implementing __hash__()
  • Make sure you aren’t using any removed modules
  • Testing coverage and tox
  • Optional: Use the iterator-methods on dictionaries
• 2to3
  • Using 2to3
    • Explicit fixers
  • Distributing packages
    • Running 2to3 on install
  • Supporting multiple versions of Python with Distribute
    • Running tests with Distribute
    • Running 2to3 with Distribute
• Common migration problems
  • Incorrect imports
  • Relative import problems
  • Unorderable types, __cmp__ and cmp
  • Sorting
  • Sorting Unicode
  • Bytes, strings and Unicode
    • Byte literals
    • Binary data in Python 2 and Python 3
    • Nicer solutions
    • Manipulating binary data
    • Reading from files
  • Replacing UserDict
  • CSV API changes
  • Running doctests
    • write() has a return value
    • Types are now classes
    • Handling expected exceptions
    • String representation
    • Unicode representation
    • dict and set order
• Improving your code with modern idioms
  • Use sorted() instead of .sort()
  • Coding with context managers
  • Advanced string formatting
  • Class decorators
  • Set literals
  • yield to the generators
  • More comprehensions
  • The next next()
  • New modules
    • Abstract base classes
    • multiprocessing and futures
    • numbers and fractions
• Supporting Python 2 and 3 without 2to3 conversion
  • Supporting the print() function
  • Handling exceptions
  • Import errors
  • Integer incompatibilities
  • More bytes, strings and Unicode
  • Two times three is “six”
• Migrating C extensions
  • Before you start
  • Object initialization
  • Module initialization
  • Changes in Python
  • Strings and Unicode
• Extending 2to3 with your own fixers
  • When fixers are necessary
  • The Parse Tree
  • Creating a fixer
  • Modifying the Parse Tree
  • Finding the nodes with Patterns
• Language differences and workarounds
  • apply()
  • buffer()
  • callable()
  • Classes
  • Comparisons
  • coerce() and __coerce__
  • Dictionary methods
  • except
  • Exception objects
  • exec
  • execfile
  • file
  • filter()
  • Imports
  • Indentation
  • input() and raw_input()
  • Integer division
  • long
  • map()
  • Metaclasses
  • .next()
  • Parameter unpacking
  • print
  • raise
  • range() and xrange()
  • repr() as backticks.
  • Rounding behavior
  • Slice operator methods
  • Sorting
  • StandardError
  • String types
• Reorganizations and renamings
  • The standard library
    • urllib, urllib2 and urlparse
  • Removed modules
  • Moved builtins
  • string module removals
  • Function and method attribute renamings