Changes
=======

0.00013_02 30 Mar 2010
  * BACKWARDS COMPAT BREAKING CHANGE!
    * The format method for all Formatters now received the language id as
      its first argument. The auto localizer passes 'auto' as the language id.
      (Dave Rolsky)
    * Gettext method arguments are now passed as %1, %2, etc.

0.00013_01 24 Mar 2010
  * Added MultiLevel localizer that allows you to write a Rails-ish
    i18n: $loc->localize('my.key' => {arg => $value});
  * Fixed tests
  * BACKWARDS COMPAT BREAKING CHANGE!
    -> For Data::Localize::Localizer consumers only.
       End users should see no change.
    * Data::Localize::Localizer is no longer a role.
    * 'style' parameter has been deprecated
    * Message formatting is now pluggable.

0.00013 23 Mar 2010
  * Change all_* accessors back to what they were in 0.00011
  * Separate out BerkeleyDB tests, fix them
  * Namespace.pm no longer attempts to load files which failed to
    load previously. If you want to reflect changes, you need to
    recreate the object.
  * DEPRECATION WARNING!
    - Gettext.pm
        * path_add        -> add_path
        * lexicon_get     -> get_lexicon
        * lexicon_set     -> set_lexicon
        * lexicon_merge   -> merge_lexicon
        * lexicon_map_get -> get_lexicon_map
        * lexicon_map_set -> set_lexicon_map
    - Namespace.pm
        * lexicon_get     -> get_lexicon
    - Auto.pm
        * lexicon_get     -> get_lexicon

0.00012 22 Mar 2010
  * Remove the use of AttributeHelpers -- doing this in Moose/Mouse is a pain
  * Move POD tests to xt
  * Factor out Gettext parser (Dave Rolsky)
  * Only load Gettext lexicons for new paths in path_add (Dave Rolsky)
  * Protect $_ when loading Gettext lexicons (Dave Rolsky)
  * Predeclare class_type (Dave Rolsky)
  * Silence test Warnings (Dave Rolsky)

0.00011 15 Dec 2009
  * Don't attempt to load Namespace lexicons once we failed to load them.
  * Allow method calls for Maketext formats (UNTESTED -- will do more hacking)

0.00010 17 Aug 2009
  * Allow method calls to be made for Gettext formats. For example,
    %foo(1,2,3) will call $localizer->foo(undef, ['1', '2', '3']), and use
    its return value as the localized text (UNTESTED -- will do more hacking)

0.00009 05 Aug 2009
  * Do not allow superfluous entries in the lexicon hash for gettext

0.00008 04 Aug 2009
  * Add auto_style to determine the auto-fallback style

0.00007 15 Apr 2009
  * Properly handle BerkeleyDB dependency (it's not required)

0.00006 14 Apr 2009
  * Properly cleanse $lang before using it as a namespace

0.00006_01 25 Mar 2009
  * Add a storage backend. This is mainly to cover cases where storing
    many lexicons on memory is not a practical choice
  * API change: lexicon_get() now accepts 2 arguments, and expects
    the translated lexicon back

0.00005 09 Mar 2009
  * Any::Moose me! Requires MouseX::AttributeHandlers and Mouse 0.19

0.00004 04 Mar 2009
  * Silence a minor warning
  * Fix up POD tests

0.00003 25 Feb 2009
  * format_string() was not properly formatting for cases where the argument
    was a logical false, such as 0. This has been fixed, but note that undef
    will not show up on the formatted string.

0.00002 22 Feb 2009
  * Change the eval call, to see if we can trap failures like this
    http://www.nntp.perl.org/group/perl.cpan.testers/2009/02/msg3321095.html
  * Remove Data::Dumper from debug output
  * Clear %ENV so that I18N::LangTags doesn't pick up the system language

0.00001 20 Feb 2009
  * Add find_localizers()

0.00001_02 20 Feb 2009
  * Handle 'localizers' in the constructor
  * Add (stole) metadata parsing for Gettext

0.00001_01 15 Feb 2009
  * Initial release