[pyicu-dev] towards release 1.0

[Andi Vajda]

The list of changes since version 0.9 was released in January reads:

  - added support for UnicodeSet, UnicodeFunctor, UnicodeFilter, UnicodeMatcher
  - added support for RegexPattern and RegexMatcher
  - added support for Normalizer
  - added support for UTransPosition (Christoph Burgmer)
  - added support for UTransPosition and UnicodeFilter methods to Transliterator
  - fixed bug in UnicodeString single character assignment method
  - added support for extending Transliterator from Python (Christoph Burgmer)
  - improved support for in-place operations on UnicodeStrings
  - added support for Transliterator::registerInstance (Christoph Burgmer)
  - added support for ICU 4.4
  - added support for Normalizer2, FilteredNormalizer2 if ICU version >= 4.4
  - added support for SelectFormat if ICU version >= 4.4
  - added support for TimeUnitFormat if ICU version >= 4.2
  - added support for PluralRules and PluralFormat if ICU version >= 4.0
  - added support for DateInterval, DateIntervalFormat if ICU version >= 4.0
  - added support for CurrencyPluralInfo if ICU version >= 4.2
  - added support for UnicodeSetIterator
  - added support for SearchIterator and StringSearch
  - added some missing Collator and RuleBasedCollator methods

Are there any ICU wrappers still missing from PyICU preventing a 1.0 
release (besides LayoutEngine et al.) ?

I would like a 1.0 release to not have any glaring omissions. ICU is a large 
library and I've only used small amounts of it myself. I don't have enough 
perspective to decide what is important or not on my own. Are there any 
pieces, important to readers of this list, PyICU users, that are not 
supported by PyICU yet ?

Also, for a 1.0 release, I'd like to have many more unit tests, ideally 
testing each and every one of the wrapped APIs. This is a daunting task, to 
say the least. Any contribution would be greatly appreciated.

Another good way to test the PyICU bindings is to port the ICU C++ samples 
to Python. Again, any help in porting these would be welcome. I started by 
porting the break.cpp sample and found bugs with UnicodeString-converted 
text parameters in iterator setters (which are now fixed).

Andi..