changelog

Changes

0.3.1 release
=============
- follow up on changes to the HanDeDict & CFDICT download
- fix issues with MySQL
- work with SQLAlchemy >= 0.6.0
- updates glyph data: decomposition, stroke order, locale mapping
  (Uriah Eisenstein)
- augment glyph entries with locale dependent decompositions, so that for
  each locale the locale-specific component is used in a decomposition
- add script to produce locale dependent decomposition entries (see above)
- add option "-y" to cjknife to search the dictionary for headword
  substrings (annotation of text)
- initial support for dico, exposing supported dictionaries to a dict server

0.3 release
===========
- easy dictionary access for EDICT, CEDICT and others
- multiple database support
- easy dictionary installation
- support for Win32 and OS X, improved package installation
- support for narrow Python builds (UCS-2)
- documentation ported to Sphinx
- pronunciation data for Shanghainese and new reading ShanghaineseIPA
  (Kellen Parker and Allan Simon)
- pronunciation information from Unihan's kMandarin is not used anymore due to
  unknown provenance, see thread under
  http://www.unicode.org/mail-arch/unicode-ml/y2010-m01/0246.html
- stroke order corrections and new entries (Uriah Eisenstein, Hugo Lopez)
- character decomposition corrections (Uriah Eisenstein)
- fix stroke order calculation for scheme ⿷ that affects characters with
  radicals 匚 and ⼖, ⿺ for characters other than 辶, 廴, 乙 and include 
  automatic calculation for ⿴ (Uriah Eisenstein)
- fix reading converters not using dbconnector of ReadingFactory
- added BIG5HKSCS character domain for zh_HK
- added JISX0208 and JISX0208_0213 character domains for ja
- possibility to return partial stroke order information, now shown in cjknife
- show character domains for option '-i' in cjknife
- new examples check-strokes.py for showing stroke order coverage and
  icutransliterator.py for exposing cjklib's conversions as ICU transliterators.

0.2 release
===========
- initial stable release