Python 3 fix and documentation updates

• The description for org_by_name incorrectly directed people to use
  name_by_addr rather than name_by_name.
• Previously GeoIP.country_names was populated from GeoIP_country_name in
  the libGeoIP C API. Some versions of the libGeoIP include non-ASCII
  ISO-8859-1 characters in these names, causing encoding errors under Python
  1. When installed under Python 3, this API now uses UTF-8 country names to
     populate this dict. Reported and fixed by Sergey Farbotka. GitHub PR #15.

Fix for broken strings in Python 3

Strings with non-ASCII characters would throw a UTF-8 decoding error. In Python 3, all strings from the database are in UTF-8 and using set_charset to set the character set to something other than UTF-8 will throw an ValueError exception.

Python 3 Support

• Python 3 support and significant code cleanup. ( Zack Weinberg )
• The methods name_by_addr, name_by_addr_v6, name_by_name, and
  name_by_name_v6 were added for ISP, Org, and ASNum lookups.
  ( Carlos Rodrigues )
• Portability fixes.
• Expose GEOIP_MMAP_CACHE from the C library. ( George Macon )
• Unit tests were added.

1.2.9: Packaging fix. PyPI-only release.

v1.2.9

Packaging fix. PyPI-only release.

1.2.8

• Remove unused code. ( Boris Zentner )
• Fix low memory error handling and refcount issues on error reported by Matt Domsch

1.2.7

• Add example test_netspeedcell.py ( Boris Zentner )

1.2.6

• Add new methods record_by_name_v6 and record_by_addr_v6 to support
  GeoIPCityv6 databases. ( Boris Zentner )

1.2.5

• Remove Confidence and Accuracy Database it is unsued anyway. ( Boris
  Zentner )
• Add new methods enable_teredo, teredo, time_zone_by_country_and_region and lib_version ( Boris Zentner )
• Add support for Confidence and Accuracy Database.
• Add support for Netspeed Database. via:
  id_by_addr
  id_by_name
• Add IPv6 glue *** needs libGeoIP 1.4.7 ***
  country_code_by_name_v6
  country_name_by_name_v6
  country_code_by_addr_v6
  country_name_by_addr_v6 ( Boris Zentner )