Boost-Commit :

Subject: [Boost-commit] svn:boost r54045 - in sandbox/SOC/2009/unicode: boost boost/iterator boost/unicode boost/unicode/ucd boost/unicode/ucd/detail libs/unicode/data_parser libs/unicode/doc libs/unicode/example libs/unicode/src libs/unicode/src/ucd
From: loufoque_at_[hidden]
Date: 2009-06-17 21:26:57

• Next message: christian.schladetsch_at_[hidden]: "[Boost-commit] svn:boost r54046 - sandbox/monotonic/libs/monotonic/test"
• Previous message: jewillco_at_[hidden]: "[Boost-commit] svn:boost r54044 - trunk/boost/graph"

Author: mgaunard
Date: 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
New Revision: 54045
URL: http://svn.boost.org/trac/boost/changeset/54045

Log:
Block support, general cleanup and automatic doxygen documentation
Added:
   sandbox/SOC/2009/unicode/boost/unicode/surrogates.hpp (contents, props changed)
   sandbox/SOC/2009/unicode/boost/unicode/ucd/block_types.hpp (contents, props changed)
   sandbox/SOC/2009/unicode/boost/unicode/ucd/detail/
   sandbox/SOC/2009/unicode/boost/unicode/ucd/detail/unichar_data.hpp (contents, props changed)
   sandbox/SOC/2009/unicode/boost/unicode/ucd/properties.hpp (contents, props changed)
   sandbox/SOC/2009/unicode/boost/unicode/ucd/properties_types.hpp (contents, props changed)
Removed:
   sandbox/SOC/2009/unicode/boost/unicode/ucd/unicode_block_types.hpp
   sandbox/SOC/2009/unicode/boost/unicode/unicode_properties.hpp
   sandbox/SOC/2009/unicode/boost/unicode/unicode_properties_types.hpp
   sandbox/SOC/2009/unicode/boost/unicode/unicode_proprieties.hpp
Text files modified:
   sandbox/SOC/2009/unicode/boost/cuchar.hpp | 10
   sandbox/SOC/2009/unicode/boost/iterator/pipe_iterator.hpp | 2
   sandbox/SOC/2009/unicode/boost/unicode/utf_conversion.hpp | 54
   sandbox/SOC/2009/unicode/libs/unicode/data_parser/main.cpp | 2
   sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties.cpp | 2
   sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties.hpp | 2
   sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties_unicodedata.cpp | 36
   sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_collation_data.hpp | 2
   sandbox/SOC/2009/unicode/libs/unicode/data_parser/write_character_properties.cpp | 27
   sandbox/SOC/2009/unicode/libs/unicode/data_parser/write_character_properties.hpp | 1
   sandbox/SOC/2009/unicode/libs/unicode/doc/Jamfile.v2 | 24
   sandbox/SOC/2009/unicode/libs/unicode/doc/users_manual.qbk | 229
   sandbox/SOC/2009/unicode/libs/unicode/example/test.cpp | 4
   sandbox/SOC/2009/unicode/libs/unicode/src/Jamfile.v2 | 2
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data.cpp | 346
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_1.ipp | 1452
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_2.ipp | 464
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_3.ipp | 3126 +-
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_4.ipp | 1348
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_5.ipp | 1480
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_6.ipp | 2182
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_7.ipp | 3194 +-
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_8.ipp | 6876 ++--
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_9.ipp | 990
   sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_sort_data.ipp | 52608 ++++++++++++++++++++++++++++++++-------
   sandbox/SOC/2009/unicode/libs/unicode/src/unicode_properties.cpp | 51
   26 files changed, 53323 insertions(+), 21191 deletions(-)

Modified: sandbox/SOC/2009/unicode/boost/cuchar.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/boost/cuchar.hpp (original)
+++ sandbox/SOC/2009/unicode/boost/cuchar.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -5,14 +5,20 @@
 
 namespace boost
 {
-#ifdef BOOST_NO_CHAR16_T
+#if defined(BOOST_NO_CHAR16_T)
     typedef boost::uint_least16_t char16;
+#elif defined(BOOST_UNICODE_DOXYGEN_INVOKED)
+ /** Typedef to C++0x char16_t, or uint_least16_t from cstdint if not available. **/
+ typedef detail::char16 char16;
 #else
         typedef char16_t char16;
 #endif
         
-#ifdef BOOST_NO_CHAR32_T
+#if defined(BOOST_NO_CHAR32_T)
         typedef boost::uint_least32_t char32;
+#elif defined(BOOST_UNICODE_DOXYGEN_INVOKED)
+ /** Typedef to C++0x char32_t, or uint_least32_t from cstdint if not available. **/
+ typedef detail::char32 char32;
 #else
         typedef char32_t char32;
 #endif

Modified: sandbox/SOC/2009/unicode/boost/iterator/pipe_iterator.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/boost/iterator/pipe_iterator.hpp (original)
+++ sandbox/SOC/2009/unicode/boost/iterator/pipe_iterator.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -145,7 +145,7 @@
         {
                 return pos;
         }
-
+
 private:
         typedef typename Pipe::output_type T;
 

Added: sandbox/SOC/2009/unicode/boost/unicode/surrogates.hpp
==============================================================================
--- (empty file)
+++ sandbox/SOC/2009/unicode/boost/unicode/surrogates.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -0,0 +1,29 @@
+#ifndef BOOST_UNICODE_SURROGATES_HPP
+#define BOOST_UNICODE_SURROGATES_HPP
+
+#include <boost/cuchar.hpp>
+
+namespace boost
+{
+namespace unicode
+{
+
+inline bool is_high_surrogate(char32 v)
+{
+ return (v & 0xFC00u) == 0xd800u;
+}
+
+inline bool is_low_surrogate(char32 v)
+{
+ return (v & 0xFC00u) == 0xdc00u;
+}
+
+inline bool is_surrogate(char32 v)
+{
+ return (v & 0xF800u) == 0xd800;
+}
+
+} // namespace unicode
+} // namespace boost
+
+#endif

Added: sandbox/SOC/2009/unicode/boost/unicode/ucd/block_types.hpp
==============================================================================
--- (empty file)
+++ sandbox/SOC/2009/unicode/boost/unicode/ucd/block_types.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -0,0 +1,212 @@
+// Though this file is under the Boost license, it is NOT (or not yet) part of
+// Boost!
+
+// Copyright Graham Barnett, Rogier van Dalen 2005.
+// Use, modification, and distribution are subject to the Boost Software
+// License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at
+// http://www.boost.org/LICENSE_1_0.txt)
+
+// This file was created using information from the
+// www.unicode.org web site
+// License http://www.unicode.org/copyright.html
+
+/**** This file should be included in any project that uses ****/
+/**** the boost Unicode character interface ****/
+/**** This file is automatically generated and should not be modified.****/
+
+/**** WARNING !! The block enum is a sparse enum to allow for ****/
+/**** new values to be added to the unicode spec without affecting ****/
+/**** existing code ****/
+
+
+#ifndef BOOST_UNICODE_UNI_UCD_CHARACTER_PROPERTIES_HPP_INCLUDED
+#define BOOST_UNICODE_UNI_UCD_CHARACTER_PROPERTIES_HPP_INCLUDED
+
+
+namespace boost { namespace unicode { namespace ucd {
+
+ struct block
+ {
+ enum type
+ {
+ none = -1,
+
+ basic_latin = 0x0,
+ latin_1_supplement = 0x80,
+ latin_extended_a = 0x100,
+ latin_extended_b = 0x180,
+ ipa_extensions = 0x250,
+ spacing_modifier_letters = 0x2b0,
+ combining_diacritical_marks = 0x300,
+ greek_and_coptic = 0x370,
+ cyrillic = 0x400,
+ cyrillic_supplement = 0x500,
+ armenian = 0x530,
+ hebrew = 0x590,
+ arabic = 0x600,
+ syriac = 0x700,
+ arabic_supplement = 0x750,
+ thaana = 0x780,
+ nko = 0x7c0,
+ devanagari = 0x900,
+ bengali = 0x980,
+ gurmukhi = 0xa00,
+ gujarati = 0xa80,
+ oriya = 0xb00,
+ tamil = 0xb80,
+ telugu = 0xc00,
+ kannada = 0xc80,
+ malayalam = 0xd00,
+ sinhala = 0xd80,
+ thai = 0xe00,
+ lao = 0xe80,
+ tibetan = 0xf00,
+ myanmar = 0x1000,
+ georgian = 0x10a0,
+ hangul_jamo = 0x1100,
+ ethiopic = 0x1200,
+ ethiopic_supplement = 0x1380,
+ cherokee = 0x13a0,
+ unified_canadian_aboriginal_syllabics = 0x1400,
+ ogham = 0x1680,
+ runic = 0x16a0,
+ tagalog = 0x1700,
+ hanunoo = 0x1720,
+ buhid = 0x1740,
+ tagbanwa = 0x1760,
+ khmer = 0x1780,
+ mongolian = 0x1800,
+ limbu = 0x1900,
+ tai_le = 0x1950,
+ new_tai_lue = 0x1980,
+ khmer_symbols = 0x19e0,
+ buginese = 0x1a00,
+ balinese = 0x1b00,
+ sundanese = 0x1b80,
+ lepcha = 0x1c00,
+ ol_chiki = 0x1c50,
+ phonetic_extensions = 0x1d00,
+ phonetic_extensions_supplement = 0x1d80,
+ combining_diacritical_marks_supplement = 0x1dc0,
+ latin_extended_additional = 0x1e00,
+ greek_extended = 0x1f00,
+ general_punctuation = 0x2000,
+ superscripts_and_subscripts = 0x2070,
+ currency_symbols = 0x20a0,
+ combining_diacritical_marks_for_symbols = 0x20d0,
+ letterlike_symbols = 0x2100,
+ number_forms = 0x2150,
+ arrows = 0x2190,
+ mathematical_operators = 0x2200,
+ miscellaneous_technical = 0x2300,
+ control_pictures = 0x2400,
+ optical_character_recognition = 0x2440,
+ enclosed_alphanumerics = 0x2460,
+ box_drawing = 0x2500,
+ block_elements = 0x2580,
+ geometric_shapes = 0x25a0,
+ miscellaneous_symbols = 0x2600,
+ dingbats = 0x2700,
+ miscellaneous_mathematical_symbols_a = 0x27c0,
+ supplemental_arrows_a = 0x27f0,
+ braille_patterns = 0x2800,
+ supplemental_arrows_b = 0x2900,
+ miscellaneous_mathematical_symbols_b = 0x2980,
+ supplemental_mathematical_operators = 0x2a00,
+ miscellaneous_symbols_and_arrows = 0x2b00,
+ glagolitic = 0x2c00,
+ latin_extended_c = 0x2c60,
+ coptic = 0x2c80,
+ georgian_supplement = 0x2d00,
+ tifinagh = 0x2d30,
+ ethiopic_extended = 0x2d80,
+ cyrillic_extended_a = 0x2de0,
+ supplemental_punctuation = 0x2e00,
+ cjk_radicals_supplement = 0x2e80,
+ kangxi_radicals = 0x2f00,
+ ideographic_description_characters = 0x2ff0,
+ cjk_symbols_and_punctuation = 0x3000,
+ hiragana = 0x3040,
+ katakana = 0x30a0,
+ bopomofo = 0x3100,
+ hangul_compatibility_jamo = 0x3130,
+ kanbun = 0x3190,
+ bopomofo_extended = 0x31a0,
+ cjk_strokes = 0x31c0,
+ katakana_phonetic_extensions = 0x31f0,
+ enclosed_cjk_letters_and_months = 0x3200,
+ cjk_compatibility = 0x3300,
+ cjk_unified_ideographs_extension_a = 0x3400,
+ yijing_hexagram_symbols = 0x4dc0,
+ cjk_unified_ideographs = 0x4e00,
+ yi_syllables = 0xa000,
+ yi_radicals = 0xa490,
+ vai = 0xa500,
+ cyrillic_extended_b = 0xa640,
+ modifier_tone_letters = 0xa700,
+ latin_extended_d = 0xa720,
+ syloti_nagri = 0xa800,
+ phags_pa = 0xa840,
+ saurashtra = 0xa880,
+ kayah_li = 0xa900,
+ rejang = 0xa930,
+ cham = 0xaa00,
+ hangul_syllables = 0xac00,
+ high_surrogates = 0xd800,
+ high_private_use_surrogates = 0xdb80,
+ low_surrogates = 0xdc00,
+ private_use_area = 0xe000,
+ cjk_compatibility_ideographs = 0xf900,
+ alphabetic_presentation_forms = 0xfb00,
+ arabic_presentation_forms_a = 0xfb50,
+ variation_selectors = 0xfe00,
+ vertical_forms = 0xfe10,
+ combining_half_marks = 0xfe20,
+ cjk_compatibility_forms = 0xfe30,
+ small_form_variants = 0xfe50,
+ arabic_presentation_forms_b = 0xfe70,
+ halfwidth_and_fullwidth_forms = 0xff00,
+ specials = 0xfff0,
+ linear_b_syllabary = 0x10000,
+ linear_b_ideograms = 0x10080,
+ aegean_numbers = 0x10100,
+ ancient_greek_numbers = 0x10140,
+ ancient_symbols = 0x10190,
+ phaistos_disc = 0x101d0,
+ lycian = 0x10280,
+ carian = 0x102a0,
+ old_italic = 0x10300,
+ gothic = 0x10330,
+ ugaritic = 0x10380,
+ old_persian = 0x103a0,
+ deseret = 0x10400,
+ shavian = 0x10450,
+ osmanya = 0x10480,
+ cypriot_syllabary = 0x10800,
+ phoenician = 0x10900,
+ lydian = 0x10920,
+ kharoshthi = 0x10a00,
+ cuneiform = 0x12000,
+ cuneiform_numbers_and_punctuation = 0x12400,
+ byzantine_musical_symbols = 0x1d000,
+ musical_symbols = 0x1d100,
+ ancient_greek_musical_notation = 0x1d200,
+ tai_xuan_jing_symbols = 0x1d300,
+ counting_rod_numerals = 0x1d360,
+ mathematical_alphanumeric_symbols = 0x1d400,
+ mahjong_tiles = 0x1f000,
+ domino_tiles = 0x1f030,
+ cjk_unified_ideographs_extension_b = 0x20000,
+ cjk_compatibility_ideographs_supplement = 0x2f800,
+ tags = 0xe0000,
+ variation_selectors_supplement = 0xe0100,
+ supplementary_private_use_area_a = 0xf0000,
+ supplementary_private_use_area_b = 0x100000,
+ };
+ };
+
+ const char* as_string(block::type);
+
+}}} // namespaces
+
+#endif // BOOST_UNICODE_UNI_UCD_CHARACTER_PROPERTIES_HPP_INCLUDED

Added: sandbox/SOC/2009/unicode/boost/unicode/ucd/detail/unichar_data.hpp
==============================================================================
--- (empty file)
+++ sandbox/SOC/2009/unicode/boost/unicode/ucd/detail/unichar_data.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -0,0 +1,154 @@
+#ifndef BOOST_UNICODE_UNICHAR_DATA_HPP
+#define BOOST_UNICODE_UNICHAR_DATA_HPP
+
+#include <boost/assert.hpp>
+#include <boost/mpl/int.hpp>
+
+#include <boost/cuchar.hpp>
+#include <boost/unicode/ucd/properties_types.hpp>
+
+namespace boost
+{
+namespace unicode
+{
+namespace ucd
+{
+ namespace detail
+ {
+ template<int X, int Y>
+ struct pow : boost::mpl::int_< X * pow<X, Y-1>::value >
+ {
+ };
+
+ template<int X>
+ struct pow<X, 0> : boost::mpl::int_<1>
+ {
+ };
+ } // namespace detail
+
+ /***************************************************************
+ *** These structure are for internal use only and should not be
+ **** used by any developer unless they are providing a custom
+ **** implementation of the character properties interface
+ **** Changes to this must be reflected
+ **** in write_character_properties.cpp:
+ **** fnct: operator << (std::fstream& file, const write_entry& data)
+ ****************************************************************/
+ const int complex_case_size_const = 4;
+ const int block_size_bits_const = 7;
+ const int block_size_const = detail::pow<2, block_size_bits_const>::value;
+
+ struct unichar_sort_data_entry
+ {
+ // to optimise the structure across different targets we move the
+ // pointers to the front
+ const uint16_t * sort_data_begin;
+ const uint16_t * sort_data_end;
+ const char32 * following_chars;
+ const uint16_t following_chars_count;
+ };
+
+ struct unichar_complex_case_internal
+ {
+ const int length_uppercase;
+ const char32 uppercase[complex_case_size_const];
+ const int length_lowercase;
+ const char32 lowercase[complex_case_size_const];
+ const int length_titlecase;
+ const char32 titlecase[complex_case_size_const];
+ // There may be several sequential entries for complex case
+ // for a single character
+ // Entries will be parsed in order until an entry with
+ // finalEntry is found
+ const bool final_entry;
+ // TODO local and conditions
+ };
+
+ struct sort_type
+ {
+ enum type
+ {
+ // this enum is used to reduce the data size required
+ // to represent the sort data from the uca
+
+ // sort data is default value for the character range for the character
+ default_ = 0,
+
+ // sort data is defined in the unichar_sort_data_entry table
+ // and sort_index_or_data1 is the index into the table
+ is_index,
+
+ // sort data is 0000.data1.data2.char32
+ zero_data1_data2_cp,
+
+ // sort data is data1.0020.data2.char32
+ data1_0x0020_data2_cp,
+
+ _count
+ };
+ enum { data_0x0020_0x000D_cp_offset = 5 };
+ };
+ const char* as_string(sort_type::type);
+
+ struct unichar_data_internal
+ {
+ // to optimise the structure across different targets we move the
+ // pointers to the front
+ const char * name;
+ const char32 * decomp;
+ const unichar_complex_case_internal* complex_case;
+
+ struct
+ {
+ uint32_t category : 5;
+ uint32_t join_type : 3;
+ uint32_t word_break : 3;
+ uint32_t unknown_char : 1;
+ // first value for sort_variable
+ uint32_t sort_variable : 1;
+ uint32_t sort_data_type : 2;
+ uint32_t sort_data2 : 16;
+ uint32_t : 0; // alignment
+
+ uint32_t bidi_class : 5;
+ uint32_t decomposition_type : 5;
+ uint32_t break_class : 6;
+ uint32_t combining : 8;
+ uint32_t sentence_break : 4;
+ uint32_t grapheme_cluster_break : 4;
+ uint32_t : 0; // alignment
+ };
+ // the meaning of this is controlled by predefined_sort
+ const uint16_t sort_index_or_data1;
+ const char32 uppercase;
+ const char32 lowercase;
+ const char32 titlecase;
+ };
+
+ struct unichar_blocks_internal
+ {
+ const char32 first;
+ const char32 last;
+ const char* name;
+ };
+
+ extern const boost::unicode::ucd::unichar_data_internal* __uni_char_data[];
+ extern const boost::unicode::ucd::unichar_blocks_internal __uni_block_data[];
+ extern const boost::unicode::ucd::unichar_sort_data_entry __uni_sort_entry[];
+
+ extern const size_t __uni_block_data_size;
+
+ inline const unichar_data_internal& get_data_internal(char32 ch)
+ {
+ BOOST_ASSERT(ch <= 0x10FFFD);
+
+ return __uni_char_data
+ [ch >> block_size_bits_const]
+ [ch & (block_size_const-1)];
+ }
+
+} // namespace ucd
+} // namespace unicode
+} // namespace boost
+
+#endif

Added: sandbox/SOC/2009/unicode/boost/unicode/ucd/properties.hpp
==============================================================================
--- (empty file)
+++ sandbox/SOC/2009/unicode/boost/unicode/ucd/properties.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -0,0 +1,60 @@
+#ifndef BOOST_UNICODE_PROPRIETIES_HPP
+#define BOOST_UNICODE_PROPRIETIES_HPP
+
+#include <boost/cuchar.hpp>
+#include <boost/unicode/ucd/properties_types.hpp>
+#include <boost/unicode/ucd/block_types.hpp>
+
+#include <boost/unicode/ucd/detail/unichar_data.hpp>
+
+#include <algorithm>
+#include <iostream>
+
+namespace boost
+{
+namespace unicode
+{
+
+/** INTERNAL ONLY **/
+#define BOOST_UNICODE_GET_PROPERTY_DEF(Name) \
+inline Name::type get_ ## Name(char32 ch) \
+{ \
+ return (Name::type)ucd::get_data_internal(ch).Name; \
+}
+
+namespace ucd
+{
+
+BOOST_UNICODE_GET_PROPERTY_DEF(category)
+BOOST_UNICODE_GET_PROPERTY_DEF(join_type)
+BOOST_UNICODE_GET_PROPERTY_DEF(bidi_class)
+BOOST_UNICODE_GET_PROPERTY_DEF(decomposition_type)
+BOOST_UNICODE_GET_PROPERTY_DEF(break_class)
+BOOST_UNICODE_GET_PROPERTY_DEF(grapheme_cluster_break)
+BOOST_UNICODE_GET_PROPERTY_DEF(word_break)
+BOOST_UNICODE_GET_PROPERTY_DEF(sentence_break)
+
+inline bool is_unknown(char32 ch)
+{
+ return ucd::get_data_internal(ch).unknown_char;
+}
+
+inline const char* get_name(char32 ch)
+{
+ return ucd::get_data_internal(ch).name;
+}
+
+inline int get_combining_class(char32 ch)
+{
+ return ucd::get_data_internal(ch).combining;
+}
+
+/** Returns the block the code point designated by \c ch is in, or
+ * \c block::none if the code point does not lie in any block. */
+block::type get_block(char32 ch);
+
+} // namespace ucd
+} // namespace unicode
+} // namespace boost
+
+#endif

Added: sandbox/SOC/2009/unicode/boost/unicode/ucd/properties_types.hpp
==============================================================================
--- (empty file)
+++ sandbox/SOC/2009/unicode/boost/unicode/ucd/properties_types.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -0,0 +1,270 @@
+// Define Unicode character properties.
+// Though this file is under the Boost license, it is NOT (or not yet) part of
+// Boost!
+
+// Copyright Graham Barnett, Rogier van Dalen 2005.
+// Use, modification, and distribution are subject to the Boost Software
+// License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at
+// http://www.boost.org/LICENSE_1_0.txt)
+
+#ifndef BOOST_UNICODE_CHARACTER_PROPERTIES_HPP_INCLUDED
+#define BOOST_UNICODE_CHARACTER_PROPERTIES_HPP_INCLUDED
+
+namespace boost
+{
+namespace unicode
+{
+
+ namespace ucd
+ {
+ /*
+ Unicode character types
+ */
+
+ struct category
+ {
+ enum type // maximum 32 values (5 bits)
+ {
+ letter_uppercase = 0,
+ letter_lowercase,
+ letter_titlecase,
+ letter_modifier,
+ letter_other,
+ mark_nonspacing,
+ mark_spacing_combining,
+ mark_enclosing,
+ number_decimal_digit,
+ number_letter,
+ number_other,
+ punctuation_connector,
+ punctuation_dash,
+ punctuation_open,
+ punctuation_close,
+ punctuation_initial_quote,
+ punctuation_final_quote,
+ punctuation_other,
+ symbol_math,
+ symbol_currency,
+ symbol_modifier,
+ symbol_other,
+ separator_space,
+ separator_line,
+ separator_paragraph,
+ other_control,
+ other_format,
+ other_surrogate,
+ other_private_use,
+ // Todo: collapse these values into one?
+ // It seems too subtle a difference.
+ other_not_assigned,
+ unknown, // default value for unknown characters
+
+ _count
+ };
+ };
+ const char* as_string(category::type);
+
+ struct join_type
+ {
+ enum type // maximum 8 values (3 bits)
+ {
+ none = 0, // default value for unknown characters
+ right,
+ left,
+ dual,
+ causing,
+ transparent,
+ _count
+ };
+ };
+ const char* as_string(join_type::type);
+
+ struct bidi_class
+ {
+ enum type // maximum 32 values (5 bits)
+ {
+ // default value for unknown characters
+ strong_left_to_right = 1,
+ strong_left_to_right_embedding,
+ strong_left_to_right_override,
+
+ strong_right_to_left, // = 4
+ strong_right_to_left_arabic,
+ strong_right_to_left_embedding,
+ strong_right_to_left_override,
+
+ weak_pop_direction_format, // = 8
+ weak_european_digits,
+ weak_european_number_separator,
+ weak_european_number_terminator,
+ weak_arabic_number,
+ weak_common_number_separator,
+ weak_non_spacing_mark,
+ weak_boundary_neutral,
+
+ neutral_paragraph_separator, // = 16
+ neutral_segment_separator,
+ neutral_whitespace,
+ neutral_other,
+
+ _count
+ };
+ };
+ const char* as_string(bidi_class::type);
+
+ struct break_class
+ {
+ enum type // maximum 64 values (6 bits)
+ {
+ mandatory = 0,
+ carriage_return,
+ line_feed,
+ combining_mark,
+ next_line,
+ surrogate,
+ word_joiner,
+ zero_width_space,
+ non_breaking,
+ contingent_break_opportunity,
+ space,
+ break_opportunity_before_after,
+ break_opportunity_after,
+ break_opportunity_before,
+ hyphen,
+ closing_punctuation,
+ exclamation_interrogation,
+ inseparable,
+ non_starter,
+ opening_punctuation,
+ ambiguous_quotation,
+ infix_separator,
+ numeric,
+ postfix_numeric,
+ prefix_numeric,
+ symbol_allowing_break,
+ ambiguous,
+ ordinary_alpabetic_or_symbol_character,
+ hangul_lv_syllable,
+ hangul_lvt_syllable,
+ ideograph,
+ hangul_l_jamo,
+ hangul_v_jamo,
+ hangul_t_jamo,
+ complex_context,
+ unknown, // default value for unknown characters
+
+ _count
+ };
+ };
+ const char* as_string(break_class::type);
+
+ struct break_action // maximum 8 values (3 bits)
+ {
+ enum type
+ {
+ direct = 0,
+ indirect = 1, // default value for unknown characters
+ combining_indirect = 2,
+ combining_prohibited = 3,
+ prohibited = 4,
+ always = 5,
+
+ _count
+ };
+ };
+ const char* as_string(break_action::type);
+
+ struct decomposition_type
+ {
+ enum type // maximum 32 values (5 bits)
+ {
+ font = 0,
+ no_break,
+ initial,
+ medial,
+ final,
+ isolated,
+ circle,
+ super,
+ sub,
+ vertical,
+ wide,
+ narrow,
+ small,
+ square,
+ fraction,
+ compat,
+ canonical,
+ none, // default value for unknown characters
+
+ _count
+ };
+ };
+ const char* as_string(decomposition_type::type);
+
+ struct grapheme_cluster_break
+ {
+ enum type // maximum 16 values (4 bits)
+ {
+ cr = 0, // value fixed
+ lf = 1, // value fixed
+ control = 2, // value fixed
+ l = 3, // value fixed
+ v = 4, // value fixed
+ t = 5, // value fixed
+ lv = 6, // value fixed
+ lvt = 7, // value fixed
+ extend = 8, // value fixed
+ // default value for unknown characters
+ any = 9, // value fixed
+
+ _count
+ };
+ };
+ const char* as_string(grapheme_cluster_break::type);
+
+ struct word_break
+ {
+ enum type // maximum 8 values (3 bits)
+ {
+ format = 0,
+ katakana,
+ aletter,
+ midletter,
+ midnum,
+ numeric,
+ extendnumlet,
+ any, // default value for unknown characters
+
+ _count
+ };
+ };
+ const char* as_string(word_break::type);
+
+ struct sentence_break
+ {
+ enum type // maximum 16 values (4 bits)
+ {
+ sep = 0,
+ format,
+ sp,
+ lower,
+ upper,
+ oletter,
+ numeric,
+ aterm,
+ sterm,
+ close,
+ any, // default value for unknown characters
+
+ _count
+ };
+ };
+ const char* as_string(sentence_break::type);
+
+ } // namespace ucd
+
+} // namespace unicode
+} // namespace boost
+
+#endif // BOOST_UNICODE_CHARACTER_PROPERTIES_HPP_INCLUDED

Deleted: sandbox/SOC/2009/unicode/boost/unicode/ucd/unicode_block_types.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/boost/unicode/ucd/unicode_block_types.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
+++ (empty file)
@@ -1,207 +0,0 @@
-// Though this file is under the Boost license, it is NOT (or not yet) part of
-// Boost!
-
-// Copyright Graham Barnett, Rogier van Dalen 2005.
-// Use, modification, and distribution are subject to the Boost Software
-// License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at
-// http://www.boost.org/LICENSE_1_0.txt)
-
-// This file was created using information from the
-// www.unicode.org web site
-// License http://www.unicode.org/copyright.html
-
-/**** This file should be included in any project that uses ****/
-/**** the boost Unicode character interface ****/
-/**** This file is automatically generated and should not be modified.****/
-
-/**** WARNING !! The block enum is a sparse enum to allow for ****/
-/**** new values to be added to the unicode spec without affecting ****/
-/**** existing code ****/
-
-
-#ifndef BOOST_UNICODE_UNI_UCD_CHARACTER_PROPERTIES_HPP_INCLUDED
-#define BOOST_UNICODE_UNI_UCD_CHARACTER_PROPERTIES_HPP_INCLUDED
-
-
-namespace boost { namespace unicode {
-
- struct block
- {
- enum type
- {
- basic_latin = 0x0,
- latin_1_supplement = 0x80,
- latin_extended_a = 0x100,
- latin_extended_b = 0x180,
- ipa_extensions = 0x250,
- spacing_modifier_letters = 0x2b0,
- combining_diacritical_marks = 0x300,
- greek_and_coptic = 0x370,
- cyrillic = 0x400,
- cyrillic_supplement = 0x500,
- armenian = 0x530,
- hebrew = 0x590,
- arabic = 0x600,
- syriac = 0x700,
- arabic_supplement = 0x750,
- thaana = 0x780,
- nko = 0x7c0,
- devanagari = 0x900,
- bengali = 0x980,
- gurmukhi = 0xa00,
- gujarati = 0xa80,
- oriya = 0xb00,
- tamil = 0xb80,
- telugu = 0xc00,
- kannada = 0xc80,
- malayalam = 0xd00,
- sinhala = 0xd80,
- thai = 0xe00,
- lao = 0xe80,
- tibetan = 0xf00,
- myanmar = 0x1000,
- georgian = 0x10a0,
- hangul_jamo = 0x1100,
- ethiopic = 0x1200,
- ethiopic_supplement = 0x1380,
- cherokee = 0x13a0,
- unified_canadian_aboriginal_syllabics = 0x1400,
- ogham = 0x1680,
- runic = 0x16a0,
- tagalog = 0x1700,
- hanunoo = 0x1720,
- buhid = 0x1740,
- tagbanwa = 0x1760,
- khmer = 0x1780,
- mongolian = 0x1800,
- limbu = 0x1900,
- tai_le = 0x1950,
- new_tai_lue = 0x1980,
- khmer_symbols = 0x19e0,
- buginese = 0x1a00,
- balinese = 0x1b00,
- sundanese = 0x1b80,
- lepcha = 0x1c00,
- ol_chiki = 0x1c50,
- phonetic_extensions = 0x1d00,
- phonetic_extensions_supplement = 0x1d80,
- combining_diacritical_marks_supplement = 0x1dc0,
- latin_extended_additional = 0x1e00,
- greek_extended = 0x1f00,
- general_punctuation = 0x2000,
- superscripts_and_subscripts = 0x2070,
- currency_symbols = 0x20a0,
- combining_diacritical_marks_for_symbols = 0x20d0,
- letterlike_symbols = 0x2100,
- number_forms = 0x2150,
- arrows = 0x2190,
- mathematical_operators = 0x2200,
- miscellaneous_technical = 0x2300,
- control_pictures = 0x2400,
- optical_character_recognition = 0x2440,
- enclosed_alphanumerics = 0x2460,
- box_drawing = 0x2500,
- block_elements = 0x2580,
- geometric_shapes = 0x25a0,
- miscellaneous_symbols = 0x2600,
- dingbats = 0x2700,
- miscellaneous_mathematical_symbols_a = 0x27c0,
- supplemental_arrows_a = 0x27f0,
- braille_patterns = 0x2800,
- supplemental_arrows_b = 0x2900,
- miscellaneous_mathematical_symbols_b = 0x2980,
- supplemental_mathematical_operators = 0x2a00,
- miscellaneous_symbols_and_arrows = 0x2b00,
- glagolitic = 0x2c00,
- latin_extended_c = 0x2c60,
- coptic = 0x2c80,
- georgian_supplement = 0x2d00,
- tifinagh = 0x2d30,
- ethiopic_extended = 0x2d80,
- cyrillic_extended_a = 0x2de0,
- supplemental_punctuation = 0x2e00,
- cjk_radicals_supplement = 0x2e80,
- kangxi_radicals = 0x2f00,
- ideographic_description_characters = 0x2ff0,
- cjk_symbols_and_punctuation = 0x3000,
- hiragana = 0x3040,
- katakana = 0x30a0,
- bopomofo = 0x3100,
- hangul_compatibility_jamo = 0x3130,
- kanbun = 0x3190,
- bopomofo_extended = 0x31a0,
- cjk_strokes = 0x31c0,
- katakana_phonetic_extensions = 0x31f0,
- enclosed_cjk_letters_and_months = 0x3200,
- cjk_compatibility = 0x3300,
- cjk_unified_ideographs_extension_a = 0x3400,
- yijing_hexagram_symbols = 0x4dc0,
- cjk_unified_ideographs = 0x4e00,
- yi_syllables = 0xa000,
- yi_radicals = 0xa490,
- vai = 0xa500,
- cyrillic_extended_b = 0xa640,
- modifier_tone_letters = 0xa700,
- latin_extended_d = 0xa720,
- syloti_nagri = 0xa800,
- phags_pa = 0xa840,
- saurashtra = 0xa880,
- kayah_li = 0xa900,
- rejang = 0xa930,
- cham = 0xaa00,
- hangul_syllables = 0xac00,
- high_surrogates = 0xd800,
- high_private_use_surrogates = 0xdb80,
- low_surrogates = 0xdc00,
- private_use_area = 0xe000,
- cjk_compatibility_ideographs = 0xf900,
- alphabetic_presentation_forms = 0xfb00,
- arabic_presentation_forms_a = 0xfb50,
- variation_selectors = 0xfe00,
- vertical_forms = 0xfe10,
- combining_half_marks = 0xfe20,
- cjk_compatibility_forms = 0xfe30,
- small_form_variants = 0xfe50,
- arabic_presentation_forms_b = 0xfe70,
- halfwidth_and_fullwidth_forms = 0xff00,
- specials = 0xfff0,
- linear_b_syllabary = 0x10000,
- linear_b_ideograms = 0x10080,
- aegean_numbers = 0x10100,
- ancient_greek_numbers = 0x10140,
- ancient_symbols = 0x10190,
- phaistos_disc = 0x101d0,
- lycian = 0x10280,
- carian = 0x102a0,
- old_italic = 0x10300,
- gothic = 0x10330,
- ugaritic = 0x10380,
- old_persian = 0x103a0,
- deseret = 0x10400,
- shavian = 0x10450,
- osmanya = 0x10480,
- cypriot_syllabary = 0x10800,
- phoenician = 0x10900,
- lydian = 0x10920,
- kharoshthi = 0x10a00,
- cuneiform = 0x12000,
- cuneiform_numbers_and_punctuation = 0x12400,
- byzantine_musical_symbols = 0x1d000,
- musical_symbols = 0x1d100,
- ancient_greek_musical_notation = 0x1d200,
- tai_xuan_jing_symbols = 0x1d300,
- counting_rod_numerals = 0x1d360,
- mathematical_alphanumeric_symbols = 0x1d400,
- mahjong_tiles = 0x1f000,
- domino_tiles = 0x1f030,
- cjk_unified_ideographs_extension_b = 0x20000,
- cjk_compatibility_ideographs_supplement = 0x2f800,
- tags = 0xe0000,
- variation_selectors_supplement = 0xe0100,
- supplementary_private_use_area_a = 0xf0000,
- supplementary_private_use_area_b = 0x100000,
- };
- };
-}} // namespaces
-
-#endif // BOOST_UNICODE_UNI_UCD_CHARACTER_PROPERTIES_HPP_INCLUDED

Deleted: sandbox/SOC/2009/unicode/boost/unicode/unicode_properties.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/boost/unicode/unicode_properties.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
+++ (empty file)
@@ -1,200 +0,0 @@
-#ifndef BOOST_UNICODE_PROPRIETIES
-#define BOOST_UNICODE_PROPRIETIES
-
-#include <boost/cuchar.hpp>
-#include <boost/assert.hpp>
-#include <boost/preprocessor/cat.hpp>
-#include <boost/mpl/int.hpp>
-
-#include <boost/unicode/unicode_properties_types.hpp>
-
-namespace boost
-{
-namespace unicode
-{
-
-inline bool is_high_surrogate(char32 v)
-{
- return (v & 0xFC00u) == 0xd800u;
-}
-
-inline bool is_low_surrogate(char32 v)
-{
- return (v & 0xFC00u) == 0xdc00u;
-}
-
-inline bool is_surrogate(char32 v)
-{
- return (v & 0xF800u) == 0xd800;
-}
-
-namespace detail
-{
- template<int X, int Y>
- struct pow : boost::mpl::int_< X * pow<X, Y-1>::value >
- {
- };
-
- template<int X>
- struct pow<X, 0> : boost::mpl::int_<1>
- {
- };
-}
-
-namespace ucd
-{
- /***************************************************************
- *** These structure are for internal use only and should not be
- **** used by any developer unless they are providing a custom
- **** implementation of the character properties interface
- **** Changes to this must be reflected
- **** in write_character_properties.cpp:
- **** fnct: operator << (std::fstream& file, const write_entry& data)
- ****************************************************************/
- const int complex_case_size_const = 4;
- const int block_size_bits_const = 7;
- const int block_size_const = detail::pow<2, block_size_bits_const>::value;
-
- struct unichar_sort_data_entry
- {
- // to optimise the structure across different targets we move the
- // pointers to the front
- const uint16_t * sort_data_begin;
- const uint16_t * sort_data_end;
- const char32 * following_chars;
- const uint16_t following_chars_count;
- };
-
- struct unichar_complex_case_internal
- {
- const int length_uppercase;
- const char32 uppercase[complex_case_size_const];
- const int length_lowercase;
- const char32 lowercase[complex_case_size_const];
- const int length_titlecase;
- const char32 titlecase[complex_case_size_const];
- // There may be several sequential entries for complex case
- // for a single character
- // Entries will be parsed in order until an entry with
- // finalEntry is found
- const bool final_entry;
- // TODO local and conditions
- };
-
- struct sort_type
- {
- enum type
- {
- // this enum is used to reduce the data size required
- // to represent the sort data from the uca
-
- // sort data is default value for the character range for the character
- default_ = 0,
-
- // sort data is defined in the unichar_sort_data_entry table
- // and sort_index_or_data1 is the index into the table
- is_index,
-
- // sort data is 0000.data1.data2.char32
- zero_data1_data2_cp,
-
- // sort data is data1.0020.data2.char32
- data1_0x0020_data2_cp,
-
- _count
- };
- enum { data_0x0020_0x000D_cp_offset = 5 };
- };
- const char* as_string(sort_type::type);
-
- struct unichar_data_internal
- {
- // to optimise the structure across different targets we move the
- // pointers to the front
- const char * name;
- const char32 * decomp;
- const unichar_complex_case_internal* complex_case;
-
- struct
- {
- uint32_t category : 5;
- uint32_t join_type : 3;
- uint32_t word_break : 3;
- uint32_t unknown_char : 1;
- // first value for sort_variable
- uint32_t sort_variable : 1;
- uint32_t sort_data_type : 2;
- uint32_t sort_data2 : 16;
- uint32_t : 0; // alignment
-
- uint32_t bidi_class : 5;
- uint32_t decomposition_type : 5;
- uint32_t break_class : 6;
- uint32_t combining : 8;
- uint32_t sentence_break : 4;
- uint32_t grapheme_cluster_break : 4;
- uint32_t : 0; // alignment
- };
- // the meaning of this is controlled by predefined_sort
- const uint16_t sort_index_or_data1;
- const char32 uppercase;
- const char32 lowercase;
- const char32 titlecase;
- };
-
- struct unichar_blocks_internal
- {
- const char32 first;
- const char32 last;
- const char* name;
- };
-
- extern const boost::unicode::ucd::unichar_data_internal* __uni_char_data[];
- extern const boost::unicode::ucd::unichar_blocks_internal __uni_block_data[];
- extern const boost::unicode::ucd::unichar_sort_data_entry __uni_sort_entry[];
-
- inline const unichar_data_internal& get_data_internal(char32 ch)
- {
- BOOST_ASSERT(ch <= 0x10FFFD);
-
- return __uni_char_data
- [ch >> block_size_bits_const]
- [ch & (block_size_const-1)];
- }
-
-} // namespace ucd
-
-#define BOOST_UNICODE_GET_PROPERTY_DEF(Name) \
-inline Name::type BOOST_PP_CAT(get_, Name)(char32 ch) \
-{ \
- return (Name::type)ucd::get_data_internal(ch).Name; \
-}
-
-BOOST_UNICODE_GET_PROPERTY_DEF(category)
-BOOST_UNICODE_GET_PROPERTY_DEF(join_type)
-BOOST_UNICODE_GET_PROPERTY_DEF(bidi_class)
-BOOST_UNICODE_GET_PROPERTY_DEF(decomposition_type)
-BOOST_UNICODE_GET_PROPERTY_DEF(break_class)
-BOOST_UNICODE_GET_PROPERTY_DEF(grapheme_cluster_break)
-BOOST_UNICODE_GET_PROPERTY_DEF(word_break)
-BOOST_UNICODE_GET_PROPERTY_DEF(sentence_break)
-
-inline bool is_unknown(char32 ch)
-{
- return ucd::get_data_internal(ch).unknown_char;
-}
-
-inline const char* get_name(char32 ch)
-{
- return ucd::get_data_internal(ch).name;
-}
-
-inline int get_combining_class(char32 ch)
-{
- return ucd::get_data_internal(ch).combining;
-}
-
-} // namespace unicode
-} // namespace boost
-
-#endif

Deleted: sandbox/SOC/2009/unicode/boost/unicode/unicode_properties_types.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/boost/unicode/unicode_properties_types.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
+++ (empty file)
@@ -1,269 +0,0 @@
-// Define Unicode character properties.
-// Though this file is under the Boost license, it is NOT (or not yet) part of
-// Boost!
-
-// Copyright Graham Barnett, Rogier van Dalen 2005.
-// Use, modification, and distribution are subject to the Boost Software
-// License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at
-// http://www.boost.org/LICENSE_1_0.txt)
-
-#ifndef BOOST_UNICODE_CHARACTER_PROPERTIES_HPP_INCLUDED
-#define BOOST_UNICODE_CHARACTER_PROPERTIES_HPP_INCLUDED
-
-#include <boost/unicode/ucd/unicode_block_types.hpp>
-
-namespace boost
-{
-
- namespace unicode
- {
- /*
- Unicode character types
- */
-
- struct category
- {
- enum type // maximum 32 values (5 bits)
- {
- letter_uppercase = 0,
- letter_lowercase,
- letter_titlecase,
- letter_modifier,
- letter_other,
- mark_nonspacing,
- mark_spacing_combining,
- mark_enclosing,
- number_decimal_digit,
- number_letter,
- number_other,
- punctuation_connector,
- punctuation_dash,
- punctuation_open,
- punctuation_close,
- punctuation_initial_quote,
- punctuation_final_quote,
- punctuation_other,
- symbol_math,
- symbol_currency,
- symbol_modifier,
- symbol_other,
- separator_space,
- separator_line,
- separator_paragraph,
- other_control,
- other_format,
- other_surrogate,
- other_private_use,
- // Todo: collapse these values into one?
- // It seems too subtle a difference.
- other_not_assigned,
- unknown, // default value for unknown characters
-
- _count
- };
- };
- const char* as_string(category::type);
-
- struct join_type
- {
- enum type // maximum 8 values (3 bits)
- {
- none = 0, // default value for unknown characters
- right,
- left,
- dual,
- causing,
- transparent,
- _count
- };
- };
- const char* as_string(join_type::type);
-
- struct bidi_class
- {
- enum type // maximum 32 values (5 bits)
- {
- // default value for unknown characters
- strong_left_to_right = 1,
- strong_left_to_right_embedding,
- strong_left_to_right_override,
-
- strong_right_to_left, // = 4
- strong_right_to_left_arabic,
- strong_right_to_left_embedding,
- strong_right_to_left_override,
-
- weak_pop_direction_format, // = 8
- weak_european_digits,
- weak_european_number_separator,
- weak_european_number_terminator,
- weak_arabic_number,
- weak_common_number_separator,
- weak_non_spacing_mark,
- weak_boundary_neutral,
-
- neutral_paragraph_separator, // = 16
- neutral_segment_separator,
- neutral_whitespace,
- neutral_other,
-
- _count
- };
- };
- const char* as_string(bidi_class::type);
-
- struct break_class
- {
- enum type // maximum 64 values (6 bits)
- {
- mandatory = 0,
- carriage_return,
- line_feed,
- combining_mark,
- next_line,
- surrogate,
- word_joiner,
- zero_width_space,
- non_breaking,
- contingent_break_opportunity,
- space,
- break_opportunity_before_after,
- break_opportunity_after,
- break_opportunity_before,
- hyphen,
- closing_punctuation,
- exclamation_interrogation,
- inseparable,
- non_starter,
- opening_punctuation,
- ambiguous_quotation,
- infix_separator,
- numeric,
- postfix_numeric,
- prefix_numeric,
- symbol_allowing_break,
- ambiguous,
- ordinary_alpabetic_or_symbol_character,
- hangul_lv_syllable,
- hangul_lvt_syllable,
- ideograph,
- hangul_l_jamo,
- hangul_v_jamo,
- hangul_t_jamo,
- complex_context,
- unknown, // default value for unknown characters
-
- _count
- };
- };
- const char* as_string(break_class::type);
-
- struct break_action // maximum 8 values (3 bits)
- {
- enum type
- {
- direct = 0,
- indirect = 1, // default value for unknown characters
- combining_indirect = 2,
- combining_prohibited = 3,
- prohibited = 4,
- always = 5,
-
- _count
- };
- };
- const char* as_string(break_action::type);
-
- struct decomposition_type
- {
- enum type // maximum 32 values (5 bits)
- {
- dct_font = 0,
- dct_no_break,
- dct_initial,
- dct_medial,
- dct_final,
- dct_isolated,
- dct_circle,
- dct_super,
- dct_sub,
- dct_vertical,
- dct_wide,
- dct_narrow,
- dct_small,
- dct_square,
- dct_fraction,
- dct_compat,
- dct_canonical,
- dct_none, // default value for unknown characters
-
- _count
- };
- };
- const char* as_string(decomposition_type::type);
-
- struct grapheme_cluster_break
- {
- enum type // maximum 16 values (4 bits)
- {
- cr = 0, // value fixed
- lf = 1, // value fixed
- control = 2, // value fixed
- l = 3, // value fixed
- v = 4, // value fixed
- t = 5, // value fixed
- lv = 6, // value fixed
- lvt = 7, // value fixed
- extend = 8, // value fixed
- // default value for unknown characters
- any = 9, // value fixed
-
- _count
- };
- };
- const char* as_string(grapheme_cluster_break::type);
-
- struct word_break
- {
- enum type // maximum 8 values (3 bits)
- {
- format = 0,
- katakana,
- aletter,
- midletter,
- midnum,
- numeric,
- extendnumlet,
- any, // default value for unknown characters
-
- _count
- };
- };
- const char* as_string(word_break::type);
-
- struct sentence_break
- {
- enum type // maximum 16 values (4 bits)
- {
- sep = 0,
- format,
- sp,
- lower,
- upper,
- oletter,
- numeric,
- aterm,
- sterm,
- close,
- any, // default value for unknown characters
-
- _count
- };
- };
- const char* as_string(sentence_break::type);
-
- } // namespace unicode
-
-} // namespace boost
-
-#endif // BOOST_UNICODE_CHARACTER_PROPERTIES_HPP_INCLUDED

Deleted: sandbox/SOC/2009/unicode/boost/unicode/unicode_proprieties.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/boost/unicode/unicode_proprieties.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
+++ (empty file)
@@ -1,29 +0,0 @@
-#ifndef BOOST_UNICODE_PROPRIETIES
-#define BOOST_UNICODE_PROPRIETIES
-
-#include <boost/cuchar.hpp>
-
-namespace boost
-{
-namespace unicode
-{
-
-inline bool is_high_surrogate(char32 v)
-{
- return (v & 0xFC00u) == 0xd800u;
-}
-
-inline bool is_low_surrogate(char32 v)
-{
- return (v & 0xFC00u) == 0xdc00u;
-}
-
-inline bool is_surrogate(char32 v)
-{
- return (v & 0xF800u) == 0xd800;
-}
-
-} // namespace unicode
-} // namespace boost
-
-#endif

Modified: sandbox/SOC/2009/unicode/boost/unicode/utf_conversion.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/boost/unicode/utf_conversion.hpp (original)
+++ sandbox/SOC/2009/unicode/boost/unicode/utf_conversion.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -10,11 +10,14 @@
 #endif
 
 #include <boost/iterator/pipe_iterator.hpp>
-#include <boost/unicode/unicode_properties.hpp>
+#include <boost/unicode/surrogates.hpp>
 
 namespace boost
 {
 
+namespace unicode
+{
+
 namespace detail
 {
     
@@ -69,6 +72,8 @@
         throw_exception(e);
 }
 
+} // namespace detail
+
 struct u16_packer
 {
         typedef char16 output_type;
@@ -155,18 +160,18 @@
                 if(unicode::is_low_surrogate(value))
                 {
                         if(it == begin)
- invalid_utf_sequence(begin, end);
+ detail::invalid_utf_sequence(begin, end);
                         --it;
                         
                         char16 hi = *it;
                  if(!unicode::is_high_surrogate(hi))
- invalid_code_point(hi);
+ detail::invalid_code_point(hi);
                         
                         value = code_point(hi, value);
               }
               // postcondition; result must not be a surrogate:
               if(unicode::is_surrogate(value))
- invalid_code_point(static_cast<char16>(value));
+ detail::invalid_code_point(static_cast<char16>(value));
                         
                 *out++ = value;
                 
@@ -297,56 +302,63 @@
         }
 };
 
-} // namespace detail
+} // namespace unicode
 
+/** \brief test
+ * blabla
+ * \param range a big ass range
+ * \return something
+ * \throw bla
+ * \pre precondition
+ * \post postcondition **/
 template<typename Range>
 std::pair<
- pipe_iterator<typename range_iterator<const Range>::type, one_many_pipe<detail::u16_packer> >,
- pipe_iterator<typename range_iterator<const Range>::type, one_many_pipe<detail::u16_packer> >
+ pipe_iterator<typename range_iterator<const Range>::type, one_many_pipe<unicode::u16_packer> >,
+ pipe_iterator<typename range_iterator<const Range>::type, one_many_pipe<unicode::u16_packer> >
> make_u32_to_u16_range(const Range& range)
 {
- return make_pipe_range(range, make_one_many_pipe(detail::u16_packer()));
+ return make_pipe_range(range, make_one_many_pipe(unicode::u16_packer()));
 }
 
 template<typename Range>
 std::pair<
- pipe_iterator<typename range_iterator<const Range>::type, detail::u16_unpacker>,
- pipe_iterator<typename range_iterator<const Range>::type, detail::u16_unpacker>
+ pipe_iterator<typename range_iterator<const Range>::type, unicode::u16_unpacker>,
+ pipe_iterator<typename range_iterator<const Range>::type, unicode::u16_unpacker>
> make_u16_to_u32_range(const Range& range)
 {
- return make_pipe_range(range, detail::u16_unpacker());
+ return make_pipe_range(range, unicode::u16_unpacker());
 }
 
 template<typename Range>
 std::pair<
- pipe_iterator<typename range_iterator<const Range>::type, one_many_pipe<detail::u8_packer> >,
- pipe_iterator<typename range_iterator<const Range>::type, one_many_pipe<detail::u8_packer> >
+ pipe_iterator<typename range_iterator<const Range>::type, one_many_pipe<unicode::u8_packer> >,
+ pipe_iterator<typename range_iterator<const Range>::type, one_many_pipe<unicode::u8_packer> >
> make_u32_to_u8_range(const Range& range)
 {
- return make_pipe_range(range, make_one_many_pipe(detail::u8_packer()));
+ return make_pipe_range(range, make_one_many_pipe(unicode::u8_packer()));
 }
 
 template<typename Range>
 std::pair<
- pipe_iterator<typename range_iterator<const Range>::type, detail::u8_unpacker>,
- pipe_iterator<typename range_iterator<const Range>::type, detail::u8_unpacker>
+ pipe_iterator<typename range_iterator<const Range>::type, unicode::u8_unpacker>,
+ pipe_iterator<typename range_iterator<const Range>::type, unicode::u8_unpacker>
> make_u8_to_u32_range(const Range& range)
 {
- return make_pipe_range(range, detail::u8_unpacker());
+ return make_pipe_range(range, unicode::u8_unpacker());
 }
 
 template<typename OutputIterator>
-pipe_output_iterator<OutputIterator, one_many_pipe<detail::u8_packer> >
+pipe_output_iterator<OutputIterator, one_many_pipe<unicode::u8_packer> >
 make_u8_output_iterator(OutputIterator out)
 {
- return make_pipe_output_iterator(out, make_one_many_pipe(detail::u8_packer()));
+ return make_pipe_output_iterator(out, make_one_many_pipe(unicode::u8_packer()));
 }
 
 template<typename OutputIterator>
-pipe_output_iterator<OutputIterator, one_many_pipe<detail::u16_packer> >
+pipe_output_iterator<OutputIterator, one_many_pipe<unicode::u16_packer> >
 make_u16_output_iterator(OutputIterator out)
 {
- return make_pipe_output_iterator(out, make_one_many_pipe(detail::u16_packer()));
+ return make_pipe_output_iterator(out, make_one_many_pipe(unicode::u16_packer()));
 }
 
 } // namespace boost

Modified: sandbox/SOC/2009/unicode/libs/unicode/data_parser/main.cpp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/data_parser/main.cpp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/data_parser/main.cpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -8,7 +8,7 @@
 // http://www.boost.org/LICENSE_1_0.txt)
 
 
-#include <boost/unicode/unicode_properties.hpp>
+#include <boost/unicode/ucd/detail/unichar_data.hpp>
 #include "read_character_properties.hpp"
 #include "read_character_properties_files.hpp"
 #include "write_character_properties.hpp"

Modified: sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties.cpp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties.cpp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties.cpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -48,7 +48,7 @@
         char_not_present.general_category = category::unknown;
         char_not_present.combining = 0;
         char_not_present.bidi = bidi_class::strong_left_to_right;
- char_not_present.decomposition_kind = decomposition_type::dct_none;
+ char_not_present.decomposition_kind = decomposition_type::none;
         char_not_present.line_break = break_class::unknown;
         char_not_present.joining = join_type::none;
         char_not_present.unknown_char = true;

Modified: sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties.hpp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -15,7 +15,7 @@
 #include <vector>
 #include <map>
 
-#include <boost/unicode/unicode_properties.hpp>
+#include <boost/unicode/ucd/detail/unichar_data.hpp>
 
 namespace boost
 {

Modified: sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties_unicodedata.cpp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties_unicodedata.cpp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_character_properties_unicodedata.cpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -221,60 +221,60 @@
                 // If none is indicated, it's a canonical decomposition_type.
                 (((str_p ("<font>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_font)] |
+ decomposition_type::font)] |
                 str_p ("<noBreak>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_no_break)] |
+ decomposition_type::no_break)] |
                 str_p ("<initial>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_initial)] |
+ decomposition_type::initial)] |
                 str_p ("<medial>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_medial)] |
+ decomposition_type::medial)] |
                 str_p ("<final>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_final)] |
+ decomposition_type::final)] |
                 str_p ("<isolated>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_isolated)] |
+ decomposition_type::isolated)] |
                 str_p ("<circle>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_circle)] |
+ decomposition_type::circle)] |
                 str_p ("<super>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_super)] |
+ decomposition_type::super)] |
                 str_p ("<sub>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_sub)] |
+ decomposition_type::sub)] |
                 str_p ("<vertical>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_vertical)] |
+ decomposition_type::vertical)] |
                 str_p ("<wide>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_wide)] |
+ decomposition_type::wide)] |
                 str_p ("<narrow>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_narrow)] |
+ decomposition_type::narrow)] |
                 str_p ("<small>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_small)] |
+ decomposition_type::small)] |
                 str_p ("<square>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_square)] |
+ decomposition_type::square)] |
                 str_p ("<fraction>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_fraction)] |
+ decomposition_type::fraction)] |
                 str_p ("<compat>")
                         [assign_a (prop.decomposition_kind,
- decomposition_type::dct_compat)] |
+ decomposition_type::compat)] |
                 eps_p [assign_a (prop.decomposition_kind,
- decomposition_type::dct_canonical)]
+ decomposition_type::canonical)]
                 ) >>
                 // composition
                 +(hex_p [push_back_a (prop.decomposition)])
                 ) |
                 // or no composition at all
- eps_p [assign_a (prop.decomposition_kind, decomposition_type::dct_none)]
+ eps_p [assign_a (prop.decomposition_kind, decomposition_type::none)]
                 ) >> ';' >>
 
                 // numeric type is skipped

Modified: sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_collation_data.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_collation_data.hpp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/data_parser/read_collation_data.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -22,7 +22,7 @@
 #include <vector>
 #include <map>
 
-#include <boost/unicode/unicode_properties.hpp>
+#include <boost/unicode/ucd/detail/unichar_data.hpp>
 #include "read_character_properties.hpp"
 
 namespace boost

Modified: sandbox/SOC/2009/unicode/libs/unicode/data_parser/write_character_properties.cpp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/data_parser/write_character_properties.cpp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/data_parser/write_character_properties.cpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -310,8 +310,8 @@
 std::fstream& operator << (std::fstream& file, const read_block& data)
 {
         file << "\t{ ";
- file << "0x" << data.first << ", ";
- file << "0x" << data.last << ", ";
+ file << "0x" << std::hex << data.first << ", ";
+ file << "0x" << std::hex << data.last << ", ";
         file << "\"" << data.name << "\" ";
         file << "},\n";
 
@@ -437,7 +437,7 @@
     
         file << "\n\n#include <boost/assert.hpp>\n";
         file << "#include <boost/static_assert.hpp>\n";
- file << "#include <boost/unicode/unicode_properties.hpp>\n";
+ file << "#include <boost/unicode/ucd/properties.hpp>\n";
         size_t n;
         for (n = 0; n < nFiles; n++)
         {
@@ -455,14 +455,14 @@
                 file << tbl_block[n];
         }
 
- // write a terminating block
+ /*// write a terminating block
     read_block terminating_data;
     terminating_data.first = 0;
     terminating_data.last = 0;
     terminating_data.name = "";
- file << terminating_data;
+ file << terminating_data;*/
 
- file << "};\n\n\n";
+ file << "};\n\nconst size_t __uni_block_data_size = sizeof __uni_block_data / sizeof __uni_block_data[0];\n\n";
 
         // ---- block table -------------------------------------------------------
         
@@ -488,12 +488,12 @@
         const std::vector<read_block>& tbl_block, const char * dest_path_ucd_props)
 {
         std::cout << " writing " << dest_path_ucd_props
- << "unicode_block_types.hpp" << std::endl;
+ << "block_types.hpp" << std::endl;
 
         // open file
         std::stringstream fss;
         fss << dest_path_ucd_props;
- fss << "unicode_block_types.hpp";
+ fss << "block_types.hpp";
 
         std::fstream file;
         file.open(fss.str().c_str(), std::ios_base::out);
@@ -509,7 +509,7 @@
         
     file << "\n\n#ifndef BOOST_UNICODE_UNI_UCD_CHARACTER_PROPERTIES_HPP_INCLUDED\n";
     file << "#define BOOST_UNICODE_UNI_UCD_CHARACTER_PROPERTIES_HPP_INCLUDED\n";
- file << "\n\nnamespace boost { namespace unicode { \n\n";
+ file << "\n\nnamespace boost { namespace unicode { namespace ucd { \n\n";
 
         // ---- block table -------------------------------------------------------
 
@@ -517,6 +517,7 @@
         file << "\t\t{\n";
         file << "\t\t\tenum type\n";
         file << "\t\t\t{\n";
+ file << "\t\t\t\tnone = -1,\n\n";
 
         for (size_t n = 0; n < tbl_block.size(); n++)
         {
@@ -538,8 +539,10 @@
 
         file << "\t\t\t};\n";
         file << "\t\t};\n";
+
+ file << "\n\t\tconst char* as_string(block::type);\n\n";
 
- file << "}} // namespaces\n\n";
+ file << "}}} // namespaces\n\n";
     file << "#endif // BOOST_UNICODE_UNI_UCD_CHARACTER_PROPERTIES_HPP_INCLUDED\n";
         
         file.flush();
@@ -670,7 +673,7 @@
                     collation_entry sort_entry;
                     get_default_sort_characteristics(cp, sort_entry);
                     // this optimisation requires default to be of size 1
- assert(sort_entry.data.size() == 1);
+ //BUG assert(sort_entry.data.size() == 1);
                     tbl_coll.push_back(sort_entry.data[0]);
                 }
                 break;
@@ -694,7 +697,7 @@
                 break;
             default:
                 // invalid enum
- assert(iter_char->second.sort_type == (size_t)-1);
+ assert(iter_char->second.sort_type == (sort_type::type)-1);
                 break;
             }
         }

Modified: sandbox/SOC/2009/unicode/libs/unicode/data_parser/write_character_properties.hpp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/data_parser/write_character_properties.hpp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/data_parser/write_character_properties.hpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -27,7 +27,6 @@
 #endif
 #include <boost/tuple/tuple.hpp>
 
-#include <boost/unicode/unicode_properties.hpp>
 #include "read_character_properties.hpp"
 
 namespace boost { namespace unicode { namespace ucd {

Modified: sandbox/SOC/2009/unicode/libs/unicode/doc/Jamfile.v2
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/doc/Jamfile.v2 (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/doc/Jamfile.v2 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -9,13 +9,33 @@
 import boostbook : boostbook ;
 using quickbook ;
 
+import toolset ;
+toolset.using doxygen ;
+
 path-constant images : html ;
 
+import path ;
+
 boostbook quickbook
     :
- users_manual.qbk
+ users_manual.qbk autodoc
     :
+ <format>pdf:<xsl:param>boost.url.prefix=http://www.boost.org/doc/libs/release/doc/html
         <format>pdf:<xsl:param>img.src.path=$(images)/
     ;
 
-
+doxygen autodoc
+ :
+ [ path.glob-tree ../../../boost/unicode ../../../boost/iterator : *.hpp : .svn detail ]
+ ../../../boost/cuchar.hpp
+ :
+ <doxygen:param>EXTRACT_ALL=YES
+ <doxygen:param>"PREDEFINED=\"BOOST_UNICODE_DOXYGEN_INVOKED\""
+ <doxygen:param>HIDE_UNDOC_MEMBERS=NO
+ <doxygen:param>EXTRACT_PRIVATE=NO
+ <doxygen:param>ENABLE_PREPROCESSING=YES
+ <doxygen:param>MACRO_EXPANSION=YES
+# <doxygen:param>EXPAND_ONLY_PREDEF=YES
+ <doxygen:param>SEARCH_INCLUDES=YES
+ <doxygen:param>INCLUDE_PATH=$(BOOST_ROOT)
+ ;

Modified: sandbox/SOC/2009/unicode/libs/unicode/doc/users_manual.qbk
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/doc/users_manual.qbk (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/doc/users_manual.qbk 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -1,6 +1,6 @@
 [library Unicode
     [quickbook 1.3]
- [version 1.0]
+ [version 0.1a2]
     [authors [Gaunard, Mathias]]
     [copyright 2009 Mathias Gaunard]
     [category string-text]
@@ -67,23 +67,26 @@
 Since microprocessors usually deal with integers whose capacity are multiples of 8 bits, a naive usage would be to use 32 bits per code point,
 which is quite a waste, especially since most daily-used characters lie in the Basic Multilingual Plane, which fits on 16 bits.
 
-That is why variable-width encodings were designed, where each code point is represented by a variable number of *code values*.
+That is why variable-width encodings were designed, where each code point is represented by a variable number of *code units*, formely also known as code values.
 [endsect]
 
 [section Encodings]
 
-The UTF-X family of encodings encode a single *code point* into a variable number of *code values*, each of which does X bits.
+The UTF-X family of encodings encode a single *code point* into a variable number of *code units*, each of which does X bits.
 
 [heading UTF-32]
 
-This encoding is fixed-width, each code value is simply a code point.
+This encoding is fixed-width, each code unit is simply a code point.
+
+This encoding isn't really recommended
 
 [heading UTF-16]
 
-Every code point is encoded by one or two code values. If the code point lies within the BMP, it is represented by exactly that code point.
+Every code point is encoded by one or two code units. If the code point lies within the BMP, it is represented by exactly that code point.
 Otherwise, the code point is represented by two values which both lie in the surrogate category of Unicode code points.
 
-This is the recommended encoding for dealing with Unicode.
+This is the recommended encoding for dealing with Unicode internally for general purposes, since it has fairly low processing overhead
+compared to UTF-8 and doesn't waste as much memory as UTF-32.
 
 [heading UTF-8]
 
@@ -92,6 +95,8 @@
 Every code point within ASCII is represented as exactly that ASCII character, others are represented as a variable-sized sequence from
 two to four bytes, all of which are non-ASCII.
 
+This encoding is popular for data storage and interchange, but can also be useful for compatibility with byte-oriented string manipulation.
+
 [endsect]
 
 [section Composite characters]
@@ -116,190 +121,118 @@
 the normalized form assumed by the XML standard.
 [endsect]
 
-[section Other]
-[endsect]
+[section Other operations]
 The Unicode standard also specifies various features such as a collation algorithm in __tr10__ for comparison and ordering of strings with
 a locale-specific criterion, as well as mechanisms to iterate over words, sentences and lines in __tr29__.
 
-Those features are not implemented by this library.
+Those features are not implemented by the current version of the library.
 [endsect]
 
-[section Core types]
-The library provides the following core types in the boost namespace:
-
-``uchar8_t
-uchar16_t
-uchar32_t``
-
-[^uchar/X/_t] is a /X/-bit character used as a *code value* in UTF-/X/.
-
-[endsect]
+[section Character properties]
 
-[section Concepts]
+Unicode also provides a database of character properties called the Unicode Character Database (UCD), which consists of a set of files describing
+the following properties:
 
-This library uses ranges to represent Unicode text, and thus refines the __boost_range__ concepts.
+* Name.
+* General category (classification as letters, numbers, symbols, punctuation, etc.).
+* Other important general characteristics (white space, dash, ideographic, alphabetic, non character, deprecated, etc.).
+* Character shaping (bidi category, shaping, mirroring, width, etc.).
+* Case (upper, lower, title, folding; both simple and full).
+* Numeric values and types (for digits).
+* Script and block.
+* Normalization properties (decompositions, decomposition type, canonical combining class, composition exclusions, etc.).
+* Age (version of the standard in which the code point was first designated).
+* Boundaries (grapheme cluster, word, line and sentence).
+* Standardized variants.
 
-[section =UnicodeRange=]
-refinement of =SinglePassRange=.
-
-A model of =UnicodeRange= is a range of Unicode *code values* whose *encoding is valid* and which is in *Normalized Form C*.
-As such, it does nothing more than a =SinglePassRange= except assume additional invariants.
-
-The encoding depends on the value type of the range: UTF-8 for =uchar8_t=, UTF-16 for =uchar16_t= and UTF-32 for =uchar32_t=.
-
-For any =X= model of =UnicodeRange=, the meta-function `boost::is_unicode_range<X>` evaluates to =true=.
+The database is useful for Unicode implementation in general, as it is the base for most algorithms, but can also be of interest to the library user that wants to
+implement facilities not provided by the library core.
 
 [endsect]
-[section =UnicodeCPRange=]
-refinement of =UnicodeRange=.
-
-An =UnicodeCPRange= is an =UnicodeRange= whose value type is =uchar32_t=. Every *code value* is thus a *code point*.
-
-For any =X= model of =UnicodeCPRange=, the meta-function `boost::is_unicode_cp_range<X>` evaluates to =true=.
-
-[endsect]
-[section =UnicodeGrapheme=]
-refinement of =SinglePassRange=.
-
-A model of =UnicodeGrapheme= is a range of Unicode *code points* that is a single *grapheme cluster* in *Normalized Form C*.
-
-For any =X= model of =UnicodeGrapheme=, the meta-function `boost::is_unicode_grapheme<X>` evaluates to =true=.
-
-
-[endsect]
-[endsect]
-
-[section Type erasure]
-Type erasure types can be constructed from objects of any type that model a certain concept. They wrap that object
-while erasing its type information.
-
-``template<typename Value>
-struct unicode_range;``
-
-`unicode_range<Value>` is a model of =UnicodeRange= whose value type is =Value=.
-
-``struct unicode_grapheme;``
-
-`unicode_grapheme` is a model of =UnicodeGrapheme=.
 [endsect]
 
-[section Range adaptors]
+[section Linking the library]
 
-C++0x notation is used for simplification.
+As has been stated in [link unicode.introduction_to_unicode.character_properties Introduction to Unicode], several Unicode algorithms require the usage of a large
+database of information which, as of version 0.2 of this library, is 2.6 MB big on x86.
 
-All iterator adaptors have a =base()= member function returning the adapted iterator.
+For this reason, features that can avoid dependency on that database do so; it is not required for conversions for example. All algorithms that depend on the Unicode
+Character Database are documented as such. All other features are header-only.
 
-[section Putting invariants in place]
+[heading UCD generation]
 
-``template<SinglePassRange Range>
-unspecified assume_utf8(Range&& range);``
+The Unicode Character Database can be generated using a parser present in the source distribution of this library to analyze
+[@http://www.unicode.org/Public/ the data provided by Unicode.org].
 
-Assumes range =range= is a properly encoded UTF-8 range in Normalization Form C. The behaviour is undefined if it isn't.
+Note however that the parser itself needs to be updated whenever new properties are added, and the layout of the table adapted to accomodate for the storage
+required to encode those new properties; otherwise those properties will fallback to the default value.
 
-Return type is a model of =UnicodeRange= whose value type is =uchar8_t=.
+[heading Binary compatibility]
 
+This library does not provide any kind of binary compatibility of the UCD so that applications compiled with version X of the library may actually
+link to version Y of the libray, with Y >= X, partially due to performance considerations.
 
-``template<SinglePassRange Range>
-unspecified make_utf8(Range&& range);``
+This may change in the future once proper benchmarking has been done.
 
-Assumes range =range= is a properly encoded UTF-8 range in Normalization Form C. Iterating the range may throw an exception if it isn't.
+[heading Alternate databases]
 
-Return type is a model of =UnicodeRange= whose value type is =uchar8_t=.
-
-``template<SinglePassRange Range>
-unspecified assume_utf16(Range&& range);``
-
-Assumes range =range= is a properly encoded UTF-16 range in Normalization Form C. The behaviour is undefined if it isn't.
-
-Return type is a model of =UnicodeRange= whose value type is =uchar16_t=.
-
-``template<SinglePassRange Range>
-unspecified make_utf16(Range&& range);``
-
-Assumes range =range= is a properly encoded UTF-16 range in Normalization Form C. Iterating the range may throw an exception if it isn't.
-
-Return type is a model of =UnicodeRange= whose value type is =uchar16_t=.
-
-``template<SinglePassRange Range>
-unspecified assume_utf32(Range&& range);``
-
-Assumes range =range= is a properly encoded UTF-32 range in Normalization Form C. The behaviour is undefined if it isn't.
-
-Return type is a model of =UnicodeRange= whose value type is =uchar32_t=.
-
-``template<SinglePassRange Range>
-unspecified make_utf32(Range&& range);``
-
-Assumes range =range= is a properly encoded UTF-32 range in Normalization Form C. Iterating the range throws an exception if it isn't.
-
-Return type is a model of =UnicodeRange= whose value type is =uchar32_t=.
+Future versions of this library may provide alternate implementations of this database as a thin layer over a database provided by another library or environment
+to prevent duplication of data.
 
 [endsect]
 
-[section On-the-fly UTF conversion]
+[section Conversions]
 
-[section Input]
-Read-only range adaptors.
+``auto concept Pipe<typename T>
+{
+ typename output_type = T::output_type;
+ static const int T::max_output; // optional
 
-``template<UnicodeCPRange Range>
-unspecified as_utf16(Range&& range);``
+ template<typename In, typename Out>
+ std::pair<In, Out> T::ltr(In begin, In end, Out out);
+
+ template<typename In, typename Out>
+ std::pair<In, Out> T::rtl(In begin, In end, Out out);
+};``
 
-Return type is a model of =UnicodeRange= whose value type is =uchar16_t=.
+models: [classref boost::unicode::u8_unpacker], [classref boost::unicode::u8_packer],
+[classref boost::unicode::u16_unpacker], [classref boost::unicode::u16_packer].
 
-``template<UnicodeCPRange Range>
-unspecified as_utf8(Range&& range);``
+Then pipe_iterator for on-the-fly conversion.
 
-Return type is a model of =UnicodeRange= whose value type is =uchar8_t=.
-
-``template<UnicodeRange Range>
-unspecified as_code_points(Range&& range);``
-
-Return type is a model of =UnicodeCPRange=.
 [endsect]
 
-[section Output]
-Output iterators that convert any *code point* to a sequence of *code values*.
-
-``template<typename OutputIterator>
-struct utf8_output_iterator;``
+[section Boundaries]
 
-``template<typename OutputIterator>
-struct utf16_output_iterator;``
+``auto concept BoundaryChecker<typename T>
+{
+ template<typename Iterator>
+ bool T::operator()(Iterator begin, Iterator end, Iterator pos);
+};``
 
-[endsect]
+``auto concept BoundaryConsumer<typename T>
+{
+ template<typename Iterator>
+ Iterator T::ltr(Iterator begin, Iterator end);
+
+ template<typename Iterator>
+ Iterator T::rtl(Iterator begin, Iterator end);
+}``
 
-[endsect]
+basically the same as Pipe without the output.
 
-[section Iterating Graphemes]
-``template<UnicodeCPRange Range>
-unspecified as_graphemes(Range&& range);``
+BoundaryConsumer can be defined in terms of BoundaryChecker.
 
-Return type is a read-only range whose value type is a model of =UnicodeGrapheme=.
-[endsect]
 
 [endsect]
 
-[section Character proprieties]
+[xinclude autodoc.xml]
 
-The library provides ways to check for certain *code point* proprieties within the =boost::unicode= namespace.
-All functions take any integer and return a boolean.
-
-``is_surrogate
-is_high_surrogate
-is_low_surrogate
-
-is_prepend
-is_hangul_syllable
-is_control
-is_grapheme_extend
-is_spacing_mark``
-
-[endsect]
-
-[section Normalization]
+[section Acknowledgements]
 
-[section String algorithms]
+Eric Niebler for mentoring this project, John Maddock for contributing preliminary on-the-fly UTF conversion, Graham Barnett and Rogier van Dalen for their work
+on Unicode character properties.
 
-It is expected that algorithms would take models of =UnicodeRange= as input.
+Beman Dawes and other members of the mailing list for their suggestions and support.
 
 [endsect]

Modified: sandbox/SOC/2009/unicode/libs/unicode/example/test.cpp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/example/test.cpp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/example/test.cpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -1,6 +1,7 @@
 #include <iostream>
 #include <boost/foreach.hpp>
 #include <boost/unicode/utf_conversion.hpp>
+#include <boost/unicode/ucd/properties.hpp>
 
 template<typename Range>
 std::pair<
@@ -38,5 +39,6 @@
         BOOST_FOREACH(char cp, make_reverse_range(boost::make_u32_to_u8_range(v)))
                 std::cout << std::hex << (int)(unsigned char)cp << std::endl;
         
- std::cout << "\n" << boost::unicode::get_name(0xE9) << std::endl;
+ std::cout << "\n" << boost::unicode::ucd::get_name(0xE9) << std::endl;
+ std::cout << boost::unicode::ucd::as_string(boost::unicode::ucd::get_block(0xE9)) << std::endl;
 }

Modified: sandbox/SOC/2009/unicode/libs/unicode/src/Jamfile.v2
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/src/Jamfile.v2 (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/src/Jamfile.v2 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -12,4 +12,4 @@
       <include>$(BOOST_ROOT)
     ;
 
-lib boost-unicode : unicode_properties.cpp ucd/uni_ucd_interface_impl_data.cpp ;
+lib boost-unicode : unicode_properties.cpp unicode_blocks.cpp ucd/uni_ucd_interface_impl_data.cpp ;

Modified: sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data.cpp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data.cpp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data.cpp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -19,7 +19,7 @@
 
 #include <boost/assert.hpp>
 #include <boost/static_assert.hpp>
-#include <boost/unicode/unicode_properties.hpp>
+#include <boost/unicode/ucd/properties.hpp>
 #include "uni_ucd_interface_impl_data_1.ipp"
 #include "uni_ucd_interface_impl_data_2.ipp"
 #include "uni_ucd_interface_impl_data_3.ipp"
@@ -36,180 +36,180 @@
 
 const unichar_blocks_internal __uni_block_data[]=
 {
- { 0x0, 0x127, "Basic Latin" },
- { 0x128, 0x255, "Latin-1 Supplement" },
- { 0x256, 0x383, "Latin Extended-A" },
- { 0x384, 0x591, "Latin Extended-B" },
- { 0x592, 0x687, "IPA Extensions" },
- { 0x688, 0x767, "Spacing Modifier Letters" },
- { 0x768, 0x879, "Combining Diacritical Marks" },
- { 0x880, 0x1023, "Greek and Coptic" },
- { 0x1024, 0x1279, "Cyrillic" },
- { 0x1280, 0x1327, "Cyrillic Supplement" },
- { 0x1328, 0x1423, "Armenian" },
- { 0x1424, 0x1535, "Hebrew" },
- { 0x1536, 0x1791, "Arabic" },
- { 0x1792, 0x1871, "Syriac" },
- { 0x1872, 0x1919, "Arabic Supplement" },
- { 0x1920, 0x1983, "Thaana" },
- { 0x1984, 0x2047, "NKo" },
- { 0x2304, 0x2431, "Devanagari" },
- { 0x2432, 0x2559, "Bengali" },
- { 0x2560, 0x2687, "Gurmukhi" },
- { 0x2688, 0x2815, "Gujarati" },
- { 0x2816, 0x2943, "Oriya" },
- { 0x2944, 0x3071, "Tamil" },
- { 0x3072, 0x3199, "Telugu" },
- { 0x3200, 0x3327, "Kannada" },
- { 0x3328, 0x3455, "Malayalam" },
- { 0x3456, 0x3583, "Sinhala" },
- { 0x3584, 0x3711, "Thai" },
- { 0x3712, 0x3839, "Lao" },
- { 0x3840, 0x4095, "Tibetan" },
- { 0x4096, 0x4255, "Myanmar" },
- { 0x4256, 0x4351, "Georgian" },
- { 0x4352, 0x4607, "Hangul Jamo" },
- { 0x4608, 0x4991, "Ethiopic" },
- { 0x4992, 0x5023, "Ethiopic Supplement" },
- { 0x5024, 0x5119, "Cherokee" },
- { 0x5120, 0x5759, "Unified Canadian Aboriginal Syllabics" },
- { 0x5760, 0x5791, "Ogham" },
- { 0x5792, 0x5887, "Runic" },
- { 0x5888, 0x5919, "Tagalog" },
- { 0x5920, 0x5951, "Hanunoo" },
- { 0x5952, 0x5983, "Buhid" },
- { 0x5984, 0x6015, "Tagbanwa" },
- { 0x6016, 0x6143, "Khmer" },
- { 0x6144, 0x6319, "Mongolian" },
- { 0x6400, 0x6479, "Limbu" },
- { 0x6480, 0x6527, "Tai Le" },
- { 0x6528, 0x6623, "New Tai Lue" },
- { 0x6624, 0x6655, "Khmer Symbols" },
- { 0x6656, 0x6687, "Buginese" },
- { 0x6912, 0x7039, "Balinese" },
- { 0x7040, 0x7103, "Sundanese" },
- { 0x7168, 0x7247, "Lepcha" },
- { 0x7248, 0x7295, "Ol Chiki" },
- { 0x7424, 0x7551, "Phonetic Extensions" },
- { 0x7552, 0x7615, "Phonetic Extensions Supplement" },
- { 0x7616, 0x7679, "Combining Diacritical Marks Supplement" },
- { 0x7680, 0x7935, "Latin Extended Additional" },
- { 0x7936, 0x8191, "Greek Extended" },
- { 0x8192, 0x8303, "General Punctuation" },
- { 0x8304, 0x8351, "Superscripts and Subscripts" },
- { 0x8352, 0x8399, "Currency Symbols" },
- { 0x8400, 0x8447, "Combining Diacritical Marks for Symbols" },
- { 0x8448, 0x8527, "Letterlike Symbols" },
- { 0x8528, 0x8591, "Number Forms" },
- { 0x8592, 0x8703, "Arrows" },
- { 0x8704, 0x8959, "Mathematical Operators" },
- { 0x8960, 0x9215, "Miscellaneous Technical" },
- { 0x9216, 0x9279, "Control Pictures" },
- { 0x9280, 0x9311, "Optical Character Recognition" },
- { 0x9312, 0x9471, "Enclosed Alphanumerics" },
- { 0x9472, 0x9599, "Box Drawing" },
- { 0x9600, 0x9631, "Block Elements" },
- { 0x9632, 0x9727, "Geometric Shapes" },
- { 0x9728, 0x9983, "Miscellaneous Symbols" },
- { 0x9984, 0x10175, "Dingbats" },
- { 0x10176, 0x10223, "Miscellaneous Mathematical Symbols-A" },
- { 0x10224, 0x10239, "Supplemental Arrows-A" },
- { 0x10240, 0x10495, "Braille Patterns" },
- { 0x10496, 0x10623, "Supplemental Arrows-B" },
- { 0x10624, 0x10751, "Miscellaneous Mathematical Symbols-B" },
- { 0x10752, 0x11007, "Supplemental Mathematical Operators" },
- { 0x11008, 0x11263, "Miscellaneous Symbols and Arrows" },
- { 0x11264, 0x11359, "Glagolitic" },
- { 0x11360, 0x11391, "Latin Extended-C" },
- { 0x11392, 0x11519, "Coptic" },
- { 0x11520, 0x11567, "Georgian Supplement" },
- { 0x11568, 0x11647, "Tifinagh" },
- { 0x11648, 0x11743, "Ethiopic Extended" },
- { 0x11744, 0x11775, "Cyrillic Extended-A" },
- { 0x11776, 0x11903, "Supplemental Punctuation" },
- { 0x11904, 0x12031, "CJK Radicals Supplement" },
- { 0x12032, 0x12255, "Kangxi Radicals" },
- { 0x12272, 0x12287, "Ideographic Description Characters" },
- { 0x12288, 0x12351, "CJK Symbols and Punctuation" },
- { 0x12352, 0x12447, "Hiragana" },
- { 0x12448, 0x12543, "Katakana" },
- { 0x12544, 0x12591, "Bopomofo" },
- { 0x12592, 0x12687, "Hangul Compatibility Jamo" },
- { 0x12688, 0x12703, "Kanbun" },
- { 0x12704, 0x12735, "Bopomofo Extended" },
- { 0x12736, 0x12783, "CJK Strokes" },
- { 0x12784, 0x12799, "Katakana Phonetic Extensions" },
- { 0x12800, 0x13055, "Enclosed CJK Letters and Months" },
- { 0x13056, 0x13311, "CJK Compatibility" },
- { 0x13312, 0x19903, "CJK Unified Ideographs Extension A" },
- { 0x19904, 0x19967, "Yijing Hexagram Symbols" },
- { 0x19968, 0x40959, "CJK Unified Ideographs" },
- { 0x40960, 0x42127, "Yi Syllables" },
- { 0x42128, 0x42191, "Yi Radicals" },
- { 0x42240, 0x42559, "Vai" },
- { 0x42560, 0x42655, "Cyrillic Extended-B" },
- { 0x42752, 0x42783, "Modifier Tone Letters" },
- { 0x42784, 0x43007, "Latin Extended-D" },
- { 0x43008, 0x43055, "Syloti Nagri" },
- { 0x43072, 0x43135, "Phags-pa" },
- { 0x43136, 0x43231, "Saurashtra" },
- { 0x43264, 0x43311, "Kayah Li" },
- { 0x43312, 0x43359, "Rejang" },
- { 0x43520, 0x43615, "Cham" },
- { 0x44032, 0x55215, "Hangul Syllables" },
- { 0x55296, 0x56191, "High Surrogates" },
- { 0x56192, 0x56319, "High Private Use Surrogates" },
- { 0x56320, 0x57343, "Low Surrogates" },
- { 0x57344, 0x63743, "Private Use Area" },
- { 0x63744, 0x64255, "CJK Compatibility Ideographs" },
- { 0x64256, 0x64335, "Alphabetic Presentation Forms" },
- { 0x64336, 0x65023, "Arabic Presentation Forms-A" },
- { 0x65024, 0x65039, "Variation Selectors" },
- { 0x65040, 0x65055, "Vertical Forms" },
- { 0x65056, 0x65071, "Combining Half Marks" },
- { 0x65072, 0x65103, "CJK Compatibility Forms" },
- { 0x65104, 0x65135, "Small Form Variants" },
- { 0x65136, 0x65279, "Arabic Presentation Forms-B" },
- { 0x65280, 0x65519, "Halfwidth and Fullwidth Forms" },
- { 0x65520, 0x65535, "Specials" },
- { 0x65536, 0x65663, "Linear B Syllabary" },
- { 0x65664, 0x65791, "Linear B Ideograms" },
- { 0x65792, 0x65855, "Aegean Numbers" },
- { 0x65856, 0x65935, "Ancient Greek Numbers" },
- { 0x65936, 0x65999, "Ancient Symbols" },
- { 0x66000, 0x66047, "Phaistos Disc" },
- { 0x66176, 0x66207, "Lycian" },
- { 0x66208, 0x66271, "Carian" },
- { 0x66304, 0x66351, "Old Italic" },
- { 0x66352, 0x66383, "Gothic" },
- { 0x66432, 0x66463, "Ugaritic" },
- { 0x66464, 0x66527, "Old Persian" },
- { 0x66560, 0x66639, "Deseret" },
- { 0x66640, 0x66687, "Shavian" },
- { 0x66688, 0x66735, "Osmanya" },
- { 0x67584, 0x67647, "Cypriot Syllabary" },
- { 0x67840, 0x67871, "Phoenician" },
- { 0x67872, 0x67903, "Lydian" },
- { 0x68096, 0x68191, "Kharoshthi" },
- { 0x73728, 0x74751, "Cuneiform" },
- { 0x74752, 0x74879, "Cuneiform Numbers and Punctuation" },
- { 0x118784, 0x119039, "Byzantine Musical Symbols" },
- { 0x119040, 0x119295, "Musical Symbols" },
- { 0x119296, 0x119375, "Ancient Greek Musical Notation" },
- { 0x119552, 0x119647, "Tai Xuan Jing Symbols" },
- { 0x119648, 0x119679, "Counting Rod Numerals" },
- { 0x119808, 0x120831, "Mathematical Alphanumeric Symbols" },
- { 0x126976, 0x127023, "Mahjong Tiles" },
- { 0x127024, 0x127135, "Domino Tiles" },
- { 0x131072, 0x173791, "CJK Unified Ideographs Extension B" },
- { 0x194560, 0x195103, "CJK Compatibility Ideographs Supplement" },
- { 0x917504, 0x917631, "Tags" },
- { 0x917760, 0x917999, "Variation Selectors Supplement" },
- { 0x983040, 0x1048575, "Supplementary Private Use Area-A" },
- { 0x1048576, 0x1114111, "Supplementary Private Use Area-B" },
- { 0x0, 0x0, "" },
+ { 0x0, 0x7f, "Basic Latin" },
+ { 0x80, 0xff, "Latin-1 Supplement" },
+ { 0x100, 0x17f, "Latin Extended-A" },
+ { 0x180, 0x24f, "Latin Extended-B" },
+ { 0x250, 0x2af, "IPA Extensions" },
+ { 0x2b0, 0x2ff, "Spacing Modifier Letters" },
+ { 0x300, 0x36f, "Combining Diacritical Marks" },
+ { 0x370, 0x3ff, "Greek and Coptic" },
+ { 0x400, 0x4ff, "Cyrillic" },
+ { 0x500, 0x52f, "Cyrillic Supplement" },
+ { 0x530, 0x58f, "Armenian" },
+ { 0x590, 0x5ff, "Hebrew" },
+ { 0x600, 0x6ff, "Arabic" },
+ { 0x700, 0x74f, "Syriac" },
+ { 0x750, 0x77f, "Arabic Supplement" },
+ { 0x780, 0x7bf, "Thaana" },
+ { 0x7c0, 0x7ff, "NKo" },
+ { 0x900, 0x97f, "Devanagari" },
+ { 0x980, 0x9ff, "Bengali" },
+ { 0xa00, 0xa7f, "Gurmukhi" },
+ { 0xa80, 0xaff, "Gujarati" },
+ { 0xb00, 0xb7f, "Oriya" },
+ { 0xb80, 0xbff, "Tamil" },
+ { 0xc00, 0xc7f, "Telugu" },
+ { 0xc80, 0xcff, "Kannada" },
+ { 0xd00, 0xd7f, "Malayalam" },
+ { 0xd80, 0xdff, "Sinhala" },
+ { 0xe00, 0xe7f, "Thai" },
+ { 0xe80, 0xeff, "Lao" },
+ { 0xf00, 0xfff, "Tibetan" },
+ { 0x1000, 0x109f, "Myanmar" },
+ { 0x10a0, 0x10ff, "Georgian" },
+ { 0x1100, 0x11ff, "Hangul Jamo" },
+ { 0x1200, 0x137f, "Ethiopic" },
+ { 0x1380, 0x139f, "Ethiopic Supplement" },
+ { 0x13a0, 0x13ff, "Cherokee" },
+ { 0x1400, 0x167f, "Unified Canadian Aboriginal Syllabics" },
+ { 0x1680, 0x169f, "Ogham" },
+ { 0x16a0, 0x16ff, "Runic" },
+ { 0x1700, 0x171f, "Tagalog" },
+ { 0x1720, 0x173f, "Hanunoo" },
+ { 0x1740, 0x175f, "Buhid" },
+ { 0x1760, 0x177f, "Tagbanwa" },
+ { 0x1780, 0x17ff, "Khmer" },
+ { 0x1800, 0x18af, "Mongolian" },
+ { 0x1900, 0x194f, "Limbu" },
+ { 0x1950, 0x197f, "Tai Le" },
+ { 0x1980, 0x19df, "New Tai Lue" },
+ { 0x19e0, 0x19ff, "Khmer Symbols" },
+ { 0x1a00, 0x1a1f, "Buginese" },
+ { 0x1b00, 0x1b7f, "Balinese" },
+ { 0x1b80, 0x1bbf, "Sundanese" },
+ { 0x1c00, 0x1c4f, "Lepcha" },
+ { 0x1c50, 0x1c7f, "Ol Chiki" },
+ { 0x1d00, 0x1d7f, "Phonetic Extensions" },
+ { 0x1d80, 0x1dbf, "Phonetic Extensions Supplement" },
+ { 0x1dc0, 0x1dff, "Combining Diacritical Marks Supplement" },
+ { 0x1e00, 0x1eff, "Latin Extended Additional" },
+ { 0x1f00, 0x1fff, "Greek Extended" },
+ { 0x2000, 0x206f, "General Punctuation" },
+ { 0x2070, 0x209f, "Superscripts and Subscripts" },
+ { 0x20a0, 0x20cf, "Currency Symbols" },
+ { 0x20d0, 0x20ff, "Combining Diacritical Marks for Symbols" },
+ { 0x2100, 0x214f, "Letterlike Symbols" },
+ { 0x2150, 0x218f, "Number Forms" },
+ { 0x2190, 0x21ff, "Arrows" },
+ { 0x2200, 0x22ff, "Mathematical Operators" },
+ { 0x2300, 0x23ff, "Miscellaneous Technical" },
+ { 0x2400, 0x243f, "Control Pictures" },
+ { 0x2440, 0x245f, "Optical Character Recognition" },
+ { 0x2460, 0x24ff, "Enclosed Alphanumerics" },
+ { 0x2500, 0x257f, "Box Drawing" },
+ { 0x2580, 0x259f, "Block Elements" },
+ { 0x25a0, 0x25ff, "Geometric Shapes" },
+ { 0x2600, 0x26ff, "Miscellaneous Symbols" },
+ { 0x2700, 0x27bf, "Dingbats" },
+ { 0x27c0, 0x27ef, "Miscellaneous Mathematical Symbols-A" },
+ { 0x27f0, 0x27ff, "Supplemental Arrows-A" },
+ { 0x2800, 0x28ff, "Braille Patterns" },
+ { 0x2900, 0x297f, "Supplemental Arrows-B" },
+ { 0x2980, 0x29ff, "Miscellaneous Mathematical Symbols-B" },
+ { 0x2a00, 0x2aff, "Supplemental Mathematical Operators" },
+ { 0x2b00, 0x2bff, "Miscellaneous Symbols and Arrows" },
+ { 0x2c00, 0x2c5f, "Glagolitic" },
+ { 0x2c60, 0x2c7f, "Latin Extended-C" },
+ { 0x2c80, 0x2cff, "Coptic" },
+ { 0x2d00, 0x2d2f, "Georgian Supplement" },
+ { 0x2d30, 0x2d7f, "Tifinagh" },
+ { 0x2d80, 0x2ddf, "Ethiopic Extended" },
+ { 0x2de0, 0x2dff, "Cyrillic Extended-A" },
+ { 0x2e00, 0x2e7f, "Supplemental Punctuation" },
+ { 0x2e80, 0x2eff, "CJK Radicals Supplement" },
+ { 0x2f00, 0x2fdf, "Kangxi Radicals" },
+ { 0x2ff0, 0x2fff, "Ideographic Description Characters" },
+ { 0x3000, 0x303f, "CJK Symbols and Punctuation" },
+ { 0x3040, 0x309f, "Hiragana" },
+ { 0x30a0, 0x30ff, "Katakana" },
+ { 0x3100, 0x312f, "Bopomofo" },
+ { 0x3130, 0x318f, "Hangul Compatibility Jamo" },
+ { 0x3190, 0x319f, "Kanbun" },
+ { 0x31a0, 0x31bf, "Bopomofo Extended" },
+ { 0x31c0, 0x31ef, "CJK Strokes" },
+ { 0x31f0, 0x31ff, "Katakana Phonetic Extensions" },
+ { 0x3200, 0x32ff, "Enclosed CJK Letters and Months" },
+ { 0x3300, 0x33ff, "CJK Compatibility" },
+ { 0x3400, 0x4dbf, "CJK Unified Ideographs Extension A" },
+ { 0x4dc0, 0x4dff, "Yijing Hexagram Symbols" },
+ { 0x4e00, 0x9fff, "CJK Unified Ideographs" },
+ { 0xa000, 0xa48f, "Yi Syllables" },
+ { 0xa490, 0xa4cf, "Yi Radicals" },
+ { 0xa500, 0xa63f, "Vai" },
+ { 0xa640, 0xa69f, "Cyrillic Extended-B" },
+ { 0xa700, 0xa71f, "Modifier Tone Letters" },
+ { 0xa720, 0xa7ff, "Latin Extended-D" },
+ { 0xa800, 0xa82f, "Syloti Nagri" },
+ { 0xa840, 0xa87f, "Phags-pa" },
+ { 0xa880, 0xa8df, "Saurashtra" },
+ { 0xa900, 0xa92f, "Kayah Li" },
+ { 0xa930, 0xa95f, "Rejang" },
+ { 0xaa00, 0xaa5f, "Cham" },
+ { 0xac00, 0xd7af, "Hangul Syllables" },
+ { 0xd800, 0xdb7f, "High Surrogates" },
+ { 0xdb80, 0xdbff, "High Private Use Surrogates" },
+ { 0xdc00, 0xdfff, "Low Surrogates" },
+ { 0xe000, 0xf8ff, "Private Use Area" },
+ { 0xf900, 0xfaff, "CJK Compatibility Ideographs" },
+ { 0xfb00, 0xfb4f, "Alphabetic Presentation Forms" },
+ { 0xfb50, 0xfdff, "Arabic Presentation Forms-A" },
+ { 0xfe00, 0xfe0f, "Variation Selectors" },
+ { 0xfe10, 0xfe1f, "Vertical Forms" },
+ { 0xfe20, 0xfe2f, "Combining Half Marks" },
+ { 0xfe30, 0xfe4f, "CJK Compatibility Forms" },
+ { 0xfe50, 0xfe6f, "Small Form Variants" },
+ { 0xfe70, 0xfeff, "Arabic Presentation Forms-B" },
+ { 0xff00, 0xffef, "Halfwidth and Fullwidth Forms" },
+ { 0xfff0, 0xffff, "Specials" },
+ { 0x10000, 0x1007f, "Linear B Syllabary" },
+ { 0x10080, 0x100ff, "Linear B Ideograms" },
+ { 0x10100, 0x1013f, "Aegean Numbers" },
+ { 0x10140, 0x1018f, "Ancient Greek Numbers" },
+ { 0x10190, 0x101cf, "Ancient Symbols" },
+ { 0x101d0, 0x101ff, "Phaistos Disc" },
+ { 0x10280, 0x1029f, "Lycian" },
+ { 0x102a0, 0x102df, "Carian" },
+ { 0x10300, 0x1032f, "Old Italic" },
+ { 0x10330, 0x1034f, "Gothic" },
+ { 0x10380, 0x1039f, "Ugaritic" },
+ { 0x103a0, 0x103df, "Old Persian" },
+ { 0x10400, 0x1044f, "Deseret" },
+ { 0x10450, 0x1047f, "Shavian" },
+ { 0x10480, 0x104af, "Osmanya" },
+ { 0x10800, 0x1083f, "Cypriot Syllabary" },
+ { 0x10900, 0x1091f, "Phoenician" },
+ { 0x10920, 0x1093f, "Lydian" },
+ { 0x10a00, 0x10a5f, "Kharoshthi" },
+ { 0x12000, 0x123ff, "Cuneiform" },
+ { 0x12400, 0x1247f, "Cuneiform Numbers and Punctuation" },
+ { 0x1d000, 0x1d0ff, "Byzantine Musical Symbols" },
+ { 0x1d100, 0x1d1ff, "Musical Symbols" },
+ { 0x1d200, 0x1d24f, "Ancient Greek Musical Notation" },
+ { 0x1d300, 0x1d35f, "Tai Xuan Jing Symbols" },
+ { 0x1d360, 0x1d37f, "Counting Rod Numerals" },
+ { 0x1d400, 0x1d7ff, "Mathematical Alphanumeric Symbols" },
+ { 0x1f000, 0x1f02f, "Mahjong Tiles" },
+ { 0x1f030, 0x1f09f, "Domino Tiles" },
+ { 0x20000, 0x2a6df, "CJK Unified Ideographs Extension B" },
+ { 0x2f800, 0x2fa1f, "CJK Compatibility Ideographs Supplement" },
+ { 0xe0000, 0xe007f, "Tags" },
+ { 0xe0100, 0xe01ef, "Variation Selectors Supplement" },
+ { 0xf0000, 0xfffff, "Supplementary Private Use Area-A" },
+ { 0x100000, 0x10ffff, "Supplementary Private Use Area-B" },
 };
 
+const size_t __uni_block_data_size = sizeof __uni_block_data / sizeof __uni_block_data[0];
 
 const unichar_data_internal* __uni_char_data[]=
 {

Modified: sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_1.ipp
==============================================================================
--- sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_1.ipp (original)
+++ sandbox/SOC/2009/unicode/libs/unicode/src/ucd/uni_ucd_interface_impl_data_1.ipp 2009-06-17 21:26:43 EDT (Wed, 17 Jun 2009)
@@ -3964,8 +3964,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 27,
                         bidi_class::weak_common_number_separator,
                         decomposition_type::no_break,
                         break_class::non_breaking,
@@ -3973,7 +3973,7 @@
                         sentence_break::sp,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 521,
                 0x0,
                 0x0,
                 0x0,
@@ -4156,8 +4156,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ambiguous,
@@ -4165,7 +4165,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 533,
                 0x0,
                 0x0,
                 0x0,
@@ -4204,8 +4204,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ambiguous,
@@ -4213,7 +4213,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 4417,
                 0x0,
                 0x0,
                 0x0,
@@ -4324,8 +4324,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -4333,7 +4333,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 529,
                 0x0,
                 0x0,
                 0x0,
@@ -4396,8 +4396,8 @@
                         word_break::any,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::weak_european_digits,
                         decomposition_type::super,
                         break_class::ambiguous,
@@ -4405,7 +4405,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 4409,
                 0x0,
                 0x0,
                 0x0,
@@ -4420,8 +4420,8 @@
                         word_break::any,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::weak_european_digits,
                         decomposition_type::super,
                         break_class::ambiguous,
@@ -4429,7 +4429,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 4410,
                 0x0,
                 0x0,
                 0x0,
@@ -4444,8 +4444,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::break_opportunity_before,
@@ -4453,7 +4453,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 526,
                 0x0,
                 0x0,
                 0x0,
@@ -4540,8 +4540,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ambiguous,
@@ -4549,7 +4549,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 538,
                 0x0,
                 0x0,
                 0x0,
@@ -4564,8 +4564,8 @@
                         word_break::any,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::weak_european_digits,
                         decomposition_type::super,
                         break_class::ambiguous,
@@ -4573,7 +4573,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 4408,
                 0x0,
                 0x0,
                 0x0,
@@ -4588,8 +4588,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ambiguous,
@@ -4597,7 +4597,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 4800,
                 0x0,
                 0x0,
                 0x0,
@@ -4669,7 +4669,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 66,
                 0x0,
                 0x0,
                 0x0,
@@ -4693,7 +4693,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 67,
                 0x0,
                 0x0,
                 0x0,
@@ -4741,7 +4741,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 68,
                 0x0,
                 0xe0,
                 0x0,
@@ -4765,7 +4765,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 69,
                 0x0,
                 0xe1,
                 0x0,
@@ -4789,7 +4789,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 70,
                 0x0,
                 0xe2,
                 0x0,
@@ -4813,7 +4813,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 71,
                 0x0,
                 0xe3,
                 0x0,
@@ -4837,7 +4837,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 72,
                 0x0,
                 0xe4,
                 0x0,
@@ -4861,7 +4861,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 73,
                 0x0,
                 0xe5,
                 0x0,
@@ -4885,7 +4885,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 65,
+ 74,
                 0x0,
                 0xe6,
                 0x0,
@@ -4909,7 +4909,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 75,
                 0x0,
                 0xe7,
                 0x0,
@@ -4933,7 +4933,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 76,
                 0x0,
                 0xe8,
                 0x0,
@@ -4957,7 +4957,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 77,
                 0x0,
                 0xe9,
                 0x0,
@@ -4981,7 +4981,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 78,
                 0x0,
                 0xea,
                 0x0,
@@ -5005,7 +5005,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 79,
                 0x0,
                 0xeb,
                 0x0,
@@ -5029,7 +5029,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 80,
                 0x0,
                 0xec,
                 0x0,
@@ -5053,7 +5053,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 81,
                 0x0,
                 0xed,
                 0x0,
@@ -5077,7 +5077,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 82,
                 0x0,
                 0xee,
                 0x0,
@@ -5101,7 +5101,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 83,
                 0x0,
                 0xef,
                 0x0,
@@ -5125,7 +5125,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 66,
+ 84,
                 0x0,
                 0xf0,
                 0x0,
@@ -5149,7 +5149,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 67,
+ 85,
                 0x0,
                 0xf1,
                 0x0,
@@ -5173,7 +5173,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 67,
+ 86,
                 0x0,
                 0xf2,
                 0x0,
@@ -5197,7 +5197,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 67,
+ 87,
                 0x0,
                 0xf3,
                 0x0,
@@ -5221,7 +5221,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 67,
+ 88,
                 0x0,
                 0xf4,
                 0x0,
@@ -5245,7 +5245,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 67,
+ 89,
                 0x0,
                 0xf5,
                 0x0,
@@ -5269,7 +5269,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 67,
+ 90,
                 0x0,
                 0xf6,
                 0x0,
@@ -5317,7 +5317,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 67,
+ 91,
                 0x0,
                 0xf8,
                 0x0,
@@ -5341,7 +5341,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 68,
+ 92,
                 0x0,
                 0xf9,
                 0x0,
@@ -5365,7 +5365,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 68,
+ 93,
                 0x0,
                 0xfa,
                 0x0,
@@ -5389,7 +5389,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 68,
+ 94,
                 0x0,
                 0xfb,
                 0x0,
@@ -5413,7 +5413,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 68,
+ 95,
                 0x0,
                 0xfc,
                 0x0,
@@ -5437,7 +5437,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 68,
+ 96,
                 0x0,
                 0xfd,
                 0x0,
@@ -5485,7 +5485,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 68,
+ 97,
                 0x0,
                 0xdf,
                 0x0,
@@ -5509,7 +5509,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 69,
+ 98,
                 0xc0,
                 0x0,
                 0xc0,
@@ -5533,7 +5533,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 69,
+ 99,
                 0xc1,
                 0x0,
                 0xc1,
@@ -5557,7 +5557,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 69,
+ 100,
                 0xc2,
                 0x0,
                 0xc2,
@@ -5581,7 +5581,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 69,
+ 101,
                 0xc3,
                 0x0,
                 0xc3,
@@ -5605,7 +5605,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 69,
+ 102,
                 0xc4,
                 0x0,
                 0xc4,
@@ -5629,7 +5629,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 69,
+ 103,
                 0xc5,
                 0x0,
                 0xc5,
@@ -5653,7 +5653,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 69,
+ 104,
                 0xc6,
                 0x0,
                 0xc6,
@@ -5677,7 +5677,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 105,
                 0xc7,
                 0x0,
                 0xc7,
@@ -5701,7 +5701,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 106,
                 0xc8,
                 0x0,
                 0xc8,
@@ -5725,7 +5725,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 107,
                 0xc9,
                 0x0,
                 0xc9,
@@ -5749,7 +5749,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 108,
                 0xca,
                 0x0,
                 0xca,
@@ -5773,7 +5773,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 109,
                 0xcb,
                 0x0,
                 0xcb,
@@ -5797,7 +5797,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 110,
                 0xcc,
                 0x0,
                 0xcc,
@@ -5821,7 +5821,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 111,
                 0xcd,
                 0x0,
                 0xcd,
@@ -5845,7 +5845,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 112,
                 0xce,
                 0x0,
                 0xce,
@@ -5869,7 +5869,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 113,
                 0xcf,
                 0x0,
                 0xcf,
@@ -5893,7 +5893,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 70,
+ 114,
                 0xd0,
                 0x0,
                 0xd0,
@@ -5917,7 +5917,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 71,
+ 115,
                 0xd1,
                 0x0,
                 0xd1,
@@ -5941,7 +5941,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 71,
+ 116,
                 0xd2,
                 0x0,
                 0xd2,
@@ -5965,7 +5965,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 71,
+ 117,
                 0xd3,
                 0x0,
                 0xd3,
@@ -5989,7 +5989,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 71,
+ 118,
                 0xd4,
                 0x0,
                 0xd4,
@@ -6013,7 +6013,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 71,
+ 119,
                 0xd5,
                 0x0,
                 0xd5,
@@ -6037,7 +6037,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 71,
+ 120,
                 0xd6,
                 0x0,
                 0xd6,
@@ -6085,7 +6085,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 71,
+ 121,
                 0xd8,
                 0x0,
                 0xd8,
@@ -6109,7 +6109,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 122,
                 0xd9,
                 0x0,
                 0xd9,
@@ -6133,7 +6133,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 123,
                 0xda,
                 0x0,
                 0xda,
@@ -6157,7 +6157,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 124,
                 0xdb,
                 0x0,
                 0xdb,
@@ -6181,7 +6181,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 125,
                 0xdc,
                 0x0,
                 0xdc,
@@ -6205,7 +6205,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 126,
                 0xdd,
                 0x0,
                 0xdd,
@@ -6253,7 +6253,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 127,
                 0x178,
                 0x0,
                 0x178,
@@ -6422,7 +6422,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 128,
                 0x0,
                 0x101,
                 0x0,
@@ -6446,7 +6446,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 129,
                 0x100,
                 0x0,
                 0x100,
@@ -6470,7 +6470,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 130,
                 0x0,
                 0x103,
                 0x0,
@@ -6494,7 +6494,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 131,
                 0x102,
                 0x0,
                 0x102,
@@ -6518,7 +6518,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 132,
                 0x0,
                 0x105,
                 0x0,
@@ -6542,7 +6542,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 133,
                 0x104,
                 0x0,
                 0x104,
@@ -6566,7 +6566,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 134,
                 0x0,
                 0x107,
                 0x0,
@@ -6590,7 +6590,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 135,
                 0x106,
                 0x0,
                 0x106,
@@ -6614,7 +6614,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 136,
                 0x0,
                 0x109,
                 0x0,
@@ -6638,7 +6638,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 137,
                 0x108,
                 0x0,
                 0x108,
@@ -6662,7 +6662,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 138,
                 0x0,
                 0x10b,
                 0x0,
@@ -6686,7 +6686,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 139,
                 0x10a,
                 0x0,
                 0x10a,
@@ -6710,7 +6710,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 140,
                 0x0,
                 0x10d,
                 0x0,
@@ -6734,7 +6734,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 141,
                 0x10c,
                 0x0,
                 0x10c,
@@ -6758,7 +6758,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 142,
                 0x0,
                 0x10f,
                 0x0,
@@ -6782,7 +6782,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 143,
                 0x10e,
                 0x0,
                 0x10e,
@@ -6806,7 +6806,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 72,
+ 144,
                 0x0,
                 0x111,
                 0x0,
@@ -6830,7 +6830,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 73,
+ 145,
                 0x110,
                 0x0,
                 0x110,
@@ -6854,7 +6854,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 146,
                 0x0,
                 0x113,
                 0x0,
@@ -6878,7 +6878,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 147,
                 0x112,
                 0x0,
                 0x112,
@@ -6902,7 +6902,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 148,
                 0x0,
                 0x115,
                 0x0,
@@ -6926,7 +6926,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 149,
                 0x114,
                 0x0,
                 0x114,
@@ -6950,7 +6950,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 150,
                 0x0,
                 0x117,
                 0x0,
@@ -6974,7 +6974,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 151,
                 0x116,
                 0x0,
                 0x116,
@@ -6998,7 +6998,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 152,
                 0x0,
                 0x119,
                 0x0,
@@ -7022,7 +7022,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 153,
                 0x118,
                 0x0,
                 0x118,
@@ -7046,7 +7046,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 154,
                 0x0,
                 0x11b,
                 0x0,
@@ -7070,7 +7070,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 155,
                 0x11a,
                 0x0,
                 0x11a,
@@ -7094,7 +7094,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 156,
                 0x0,
                 0x11d,
                 0x0,
@@ -7118,7 +7118,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 157,
                 0x11c,
                 0x0,
                 0x11c,
@@ -7142,7 +7142,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 158,
                 0x0,
                 0x11f,
                 0x0,
@@ -7166,7 +7166,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 159,
                 0x11e,
                 0x0,
                 0x11e,
@@ -7190,7 +7190,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 160,
                 0x0,
                 0x121,
                 0x0,
@@ -7214,7 +7214,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 161,
                 0x120,
                 0x0,
                 0x120,
@@ -7238,7 +7238,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 162,
                 0x0,
                 0x123,
                 0x0,
@@ -7262,7 +7262,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 163,
                 0x122,
                 0x0,
                 0x122,
@@ -7286,7 +7286,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 164,
                 0x0,
                 0x125,
                 0x0,
@@ -7310,7 +7310,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 165,
                 0x124,
                 0x0,
                 0x124,
@@ -7334,7 +7334,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 74,
+ 166,
                 0x0,
                 0x127,
                 0x0,
@@ -7358,7 +7358,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 75,
+ 167,
                 0x126,
                 0x0,
                 0x126,
@@ -7382,7 +7382,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 168,
                 0x0,
                 0x129,
                 0x0,
@@ -7406,7 +7406,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 169,
                 0x128,
                 0x0,
                 0x128,
@@ -7430,7 +7430,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 170,
                 0x0,
                 0x12b,
                 0x0,
@@ -7454,7 +7454,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 171,
                 0x12a,
                 0x0,
                 0x12a,
@@ -7478,7 +7478,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 172,
                 0x0,
                 0x12d,
                 0x0,
@@ -7502,7 +7502,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 173,
                 0x12c,
                 0x0,
                 0x12c,
@@ -7526,7 +7526,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 174,
                 0x0,
                 0x12f,
                 0x0,
@@ -7550,7 +7550,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 175,
                 0x12e,
                 0x0,
                 0x12e,
@@ -7574,7 +7574,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 176,
                 0x130,
                 0x69,
                 0x130,
@@ -7622,7 +7622,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 177,
                 0x0,
                 0x133,
                 0x0,
@@ -7646,7 +7646,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 178,
                 0x132,
                 0x0,
                 0x132,
@@ -7670,7 +7670,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 179,
                 0x0,
                 0x135,
                 0x0,
@@ -7694,7 +7694,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 180,
                 0x134,
                 0x0,
                 0x134,
@@ -7718,7 +7718,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 181,
                 0x0,
                 0x137,
                 0x0,
@@ -7742,7 +7742,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 182,
                 0x136,
                 0x0,
                 0x136,
@@ -7790,7 +7790,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 76,
+ 183,
                 0x0,
                 0x13a,
                 0x0,
@@ -7814,7 +7814,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 77,
+ 184,
                 0x139,
                 0x0,
                 0x139,
@@ -7838,7 +7838,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 78,
+ 185,
                 0x0,
                 0x13c,
                 0x0,
@@ -7862,7 +7862,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 79,
+ 186,
                 0x13b,
                 0x0,
                 0x13b,
@@ -7886,7 +7886,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 80,
+ 187,
                 0x0,
                 0x13e,
                 0x0,
@@ -7910,7 +7910,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 81,
+ 188,
                 0x13d,
                 0x0,
                 0x13d,
@@ -7934,7 +7934,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 82,
+ 189,
                 0x0,
                 0x140,
                 0x0,
@@ -7958,7 +7958,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 83,
+ 190,
                 0x13f,
                 0x0,
                 0x13f,
@@ -7982,7 +7982,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 84,
+ 191,
                 0x0,
                 0x142,
                 0x0,
@@ -8006,7 +8006,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 85,
+ 192,
                 0x141,
                 0x0,
                 0x141,
@@ -8030,7 +8030,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 86,
+ 193,
                 0x0,
                 0x144,
                 0x0,
@@ -8054,7 +8054,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 86,
+ 194,
                 0x143,
                 0x0,
                 0x143,
@@ -8078,7 +8078,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 86,
+ 195,
                 0x0,
                 0x146,
                 0x0,
@@ -8102,7 +8102,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 86,
+ 196,
                 0x145,
                 0x0,
                 0x145,
@@ -8126,7 +8126,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 86,
+ 197,
                 0x0,
                 0x148,
                 0x0,
@@ -8150,7 +8150,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 86,
+ 198,
                 0x147,
                 0x0,
                 0x147,
@@ -8174,7 +8174,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 86,
+ 199,
                 0x0,
                 0x149,
                 0x0,
@@ -8246,7 +8246,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 87,
+ 200,
                 0x0,
                 0x14d,
                 0x0,
@@ -8270,7 +8270,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 87,
+ 201,
                 0x14c,
                 0x0,
                 0x14c,
@@ -8294,7 +8294,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 87,
+ 202,
                 0x0,
                 0x14f,
                 0x0,
@@ -8318,7 +8318,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 87,
+ 203,
                 0x14e,
                 0x0,
                 0x14e,
@@ -8342,7 +8342,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 87,
+ 204,
                 0x0,
                 0x151,
                 0x0,
@@ -8366,7 +8366,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 87,
+ 205,
                 0x150,
                 0x0,
                 0x150,
@@ -8390,7 +8390,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 87,
+ 206,
                 0x0,
                 0x153,
                 0x0,
@@ -8414,7 +8414,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 88,
+ 207,
                 0x152,
                 0x0,
                 0x152,
@@ -8438,7 +8438,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 208,
                 0x0,
                 0x155,
                 0x0,
@@ -8462,7 +8462,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 209,
                 0x154,
                 0x0,
                 0x154,
@@ -8486,7 +8486,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 210,
                 0x0,
                 0x157,
                 0x0,
@@ -8510,7 +8510,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 211,
                 0x156,
                 0x0,
                 0x156,
@@ -8534,7 +8534,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 212,
                 0x0,
                 0x159,
                 0x0,
@@ -8558,7 +8558,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 213,
                 0x158,
                 0x0,
                 0x158,
@@ -8582,7 +8582,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 214,
                 0x0,
                 0x15b,
                 0x0,
@@ -8606,7 +8606,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 215,
                 0x15a,
                 0x0,
                 0x15a,
@@ -8630,7 +8630,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 216,
                 0x0,
                 0x15d,
                 0x0,
@@ -8654,7 +8654,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 217,
                 0x15c,
                 0x0,
                 0x15c,
@@ -8678,7 +8678,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 218,
                 0x0,
                 0x15f,
                 0x0,
@@ -8702,7 +8702,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 219,
                 0x15e,
                 0x0,
                 0x15e,
@@ -8726,7 +8726,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 220,
                 0x0,
                 0x161,
                 0x0,
@@ -8750,7 +8750,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 221,
                 0x160,
                 0x0,
                 0x160,
@@ -8774,7 +8774,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 222,
                 0x0,
                 0x163,
                 0x0,
@@ -8798,7 +8798,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 223,
                 0x162,
                 0x0,
                 0x162,
@@ -8822,7 +8822,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 224,
                 0x0,
                 0x165,
                 0x0,
@@ -8846,7 +8846,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 225,
                 0x164,
                 0x0,
                 0x164,
@@ -8918,7 +8918,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 226,
                 0x0,
                 0x169,
                 0x0,
@@ -8942,7 +8942,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 227,
                 0x168,
                 0x0,
                 0x168,
@@ -8966,7 +8966,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 228,
                 0x0,
                 0x16b,
                 0x0,
@@ -8990,7 +8990,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 229,
                 0x16a,
                 0x0,
                 0x16a,
@@ -9014,7 +9014,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 230,
                 0x0,
                 0x16d,
                 0x0,
@@ -9038,7 +9038,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 231,
                 0x16c,
                 0x0,
                 0x16c,
@@ -9062,7 +9062,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 232,
                 0x0,
                 0x16f,
                 0x0,
@@ -9086,7 +9086,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 233,
                 0x16e,
                 0x0,
                 0x16e,
@@ -9110,7 +9110,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 234,
                 0x0,
                 0x171,
                 0x0,
@@ -9134,7 +9134,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 235,
                 0x170,
                 0x0,
                 0x170,
@@ -9158,7 +9158,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 236,
                 0x0,
                 0x173,
                 0x0,
@@ -9182,7 +9182,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 237,
                 0x172,
                 0x0,
                 0x172,
@@ -9206,7 +9206,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 238,
                 0x0,
                 0x175,
                 0x0,
@@ -9230,7 +9230,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 239,
                 0x174,
                 0x0,
                 0x174,
@@ -9254,7 +9254,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 240,
                 0x0,
                 0x177,
                 0x0,
@@ -9278,7 +9278,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 241,
                 0x176,
                 0x0,
                 0x176,
@@ -9302,7 +9302,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 242,
                 0x0,
                 0xff,
                 0x0,
@@ -9326,7 +9326,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 243,
                 0x0,
                 0x17a,
                 0x0,
@@ -9350,7 +9350,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 244,
                 0x179,
                 0x0,
                 0x179,
@@ -9374,7 +9374,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 245,
                 0x0,
                 0x17c,
                 0x0,
@@ -9398,7 +9398,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 246,
                 0x17b,
                 0x0,
                 0x17b,
@@ -9422,7 +9422,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 247,
                 0x0,
                 0x17e,
                 0x0,
@@ -9446,7 +9446,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 248,
                 0x17d,
                 0x0,
                 0x17d,
@@ -9470,7 +9470,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 249,
                 0x53,
                 0x0,
                 0x53,
@@ -9885,7 +9885,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 89,
+ 250,
                 0x0,
                 0x0,
                 0x0,
@@ -10341,7 +10341,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 90,
+ 251,
                 0x0,
                 0x1a1,
                 0x0,
@@ -10365,7 +10365,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 90,
+ 252,
                 0x1a0,
                 0x0,
                 0x1a0,
@@ -10701,7 +10701,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 90,
+ 253,
                 0x0,
                 0x1b0,
                 0x0,
@@ -10725,7 +10725,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 90,
+ 254,
                 0x1af,
                 0x0,
                 0x1af,
@@ -11061,7 +11061,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 90,
+ 255,
                 0x0,
                 0x0,
                 0x0,
@@ -11205,7 +11205,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 91,
+ 256,
                 0x0,
                 0x1c6,
                 0x1c5,
@@ -11229,7 +11229,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 91,
+ 257,
                 0x1c4,
                 0x1c6,
                 0x1c5,
@@ -11253,7 +11253,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 91,
+ 258,
                 0x1c4,
                 0x0,
                 0x1c5,
@@ -11277,7 +11277,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 91,
+ 259,
                 0x0,
                 0x1c9,
                 0x1c8,
@@ -11301,7 +11301,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 92,
+ 260,
                 0x1c7,
                 0x1c9,
                 0x1c8,
@@ -11325,7 +11325,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 93,
+ 261,
                 0x1c7,
                 0x0,
                 0x1c8,
@@ -11349,7 +11349,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 262,
                 0x0,
                 0x1cc,
                 0x1cb,
@@ -11373,7 +11373,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 263,
                 0x1ca,
                 0x1cc,
                 0x1cb,
@@ -11397,7 +11397,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 264,
                 0x1ca,
                 0x0,
                 0x1cb,
@@ -11421,7 +11421,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 265,
                 0x0,
                 0x1ce,
                 0x0,
@@ -11445,7 +11445,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 266,
                 0x1cd,
                 0x0,
                 0x1cd,
@@ -11469,7 +11469,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 267,
                 0x0,
                 0x1d0,
                 0x0,
@@ -11493,7 +11493,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 268,
                 0x1cf,
                 0x0,
                 0x1cf,
@@ -11517,7 +11517,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 269,
                 0x0,
                 0x1d2,
                 0x0,
@@ -11541,7 +11541,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 270,
                 0x1d1,
                 0x0,
                 0x1d1,
@@ -11565,7 +11565,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 271,
                 0x0,
                 0x1d4,
                 0x0,
@@ -11589,7 +11589,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 272,
                 0x1d3,
                 0x0,
                 0x1d3,
@@ -11613,7 +11613,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x0,
                 0x1d6,
                 0x0,
@@ -11637,7 +11637,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x1d5,
                 0x0,
                 0x1d5,
@@ -11661,7 +11661,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x0,
                 0x1d8,
                 0x0,
@@ -11685,7 +11685,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x1d7,
                 0x0,
                 0x1d7,
@@ -11709,7 +11709,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x0,
                 0x1da,
                 0x0,
@@ -11733,7 +11733,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x1d9,
                 0x0,
                 0x1d9,
@@ -11757,7 +11757,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x0,
                 0x1dc,
                 0x0,
@@ -11781,7 +11781,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x1db,
                 0x0,
                 0x1db,
@@ -11829,7 +11829,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x0,
                 0x1df,
                 0x0,
@@ -11853,7 +11853,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x1de,
                 0x0,
                 0x1de,
@@ -11877,7 +11877,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x0,
                 0x1e1,
                 0x0,
@@ -11901,7 +11901,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x1e0,
                 0x0,
                 0x1e0,
@@ -11925,7 +11925,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 94,
+ 273,
                 0x0,
                 0x1e3,
                 0x0,
@@ -11949,7 +11949,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 95,
+ 274,
                 0x1e2,
                 0x0,
                 0x1e2,
@@ -12021,7 +12021,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 275,
                 0x0,
                 0x1e7,
                 0x0,
@@ -12045,7 +12045,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 276,
                 0x1e6,
                 0x0,
                 0x1e6,
@@ -12069,7 +12069,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 277,
                 0x0,
                 0x1e9,
                 0x0,
@@ -12093,7 +12093,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 278,
                 0x1e8,
                 0x0,
                 0x1e8,
@@ -12117,7 +12117,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 279,
                 0x0,
                 0x1eb,
                 0x0,
@@ -12141,7 +12141,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 280,
                 0x1ea,
                 0x0,
                 0x1ea,
@@ -12165,7 +12165,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 281,
                 0x0,
                 0x1ed,
                 0x0,
@@ -12189,7 +12189,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 281,
                 0x1ec,
                 0x0,
                 0x1ec,
@@ -12213,7 +12213,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 281,
                 0x0,
                 0x1ef,
                 0x0,
@@ -12237,7 +12237,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 282,
                 0x1ee,
                 0x0,
                 0x1ee,
@@ -12261,7 +12261,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 282,
                 0x0,
                 0x1f0,
                 0x0,
@@ -12285,7 +12285,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 283,
                 0x0,
                 0x1f3,
                 0x1f2,
@@ -12309,7 +12309,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 284,
                 0x1f1,
                 0x1f3,
                 0x1f2,
@@ -12333,7 +12333,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 285,
                 0x1f1,
                 0x0,
                 0x1f2,
@@ -12357,7 +12357,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 286,
                 0x0,
                 0x1f5,
                 0x0,
@@ -12381,7 +12381,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 287,
                 0x1f4,
                 0x0,
                 0x1f4,
@@ -12453,7 +12453,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 288,
                 0x0,
                 0x1f9,
                 0x0,
@@ -12477,7 +12477,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 289,
                 0x1f8,
                 0x0,
                 0x1f8,
@@ -12501,7 +12501,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 290,
                 0x0,
                 0x1fb,
                 0x0,
@@ -12525,7 +12525,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 290,
                 0x1fa,
                 0x0,
                 0x1fa,
@@ -12549,7 +12549,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 96,
+ 290,
                 0x0,
                 0x1fd,
                 0x0,
@@ -12573,7 +12573,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 97,
+ 291,
                 0x1fc,
                 0x0,
                 0x1fc,
@@ -12597,7 +12597,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 292,
                 0x0,
                 0x1ff,
                 0x0,
@@ -12621,7 +12621,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 292,
                 0x1fe,
                 0x0,
                 0x1fe,
@@ -12698,7 +12698,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 292,
                 0x0,
                 0x201,
                 0x0,
@@ -12722,7 +12722,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 293,
                 0x200,
                 0x0,
                 0x200,
@@ -12746,7 +12746,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 294,
                 0x0,
                 0x203,
                 0x0,
@@ -12770,7 +12770,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 295,
                 0x202,
                 0x0,
                 0x202,
@@ -12794,7 +12794,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 296,
                 0x0,
                 0x205,
                 0x0,
@@ -12818,7 +12818,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 297,
                 0x204,
                 0x0,
                 0x204,
@@ -12842,7 +12842,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 298,
                 0x0,
                 0x207,
                 0x0,
@@ -12866,7 +12866,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 299,
                 0x206,
                 0x0,
                 0x206,
@@ -12890,7 +12890,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 300,
                 0x0,
                 0x209,
                 0x0,
@@ -12914,7 +12914,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 301,
                 0x208,
                 0x0,
                 0x208,
@@ -12938,7 +12938,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 302,
                 0x0,
                 0x20b,
                 0x0,
@@ -12962,7 +12962,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 303,
                 0x20a,
                 0x0,
                 0x20a,
@@ -12986,7 +12986,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 304,
                 0x0,
                 0x20d,
                 0x0,
@@ -13010,7 +13010,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 305,
                 0x20c,
                 0x0,
                 0x20c,
@@ -13034,7 +13034,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 306,
                 0x0,
                 0x20f,
                 0x0,
@@ -13058,7 +13058,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 307,
                 0x20e,
                 0x0,
                 0x20e,
@@ -13082,7 +13082,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 308,
                 0x0,
                 0x211,
                 0x0,
@@ -13106,7 +13106,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 309,
                 0x210,
                 0x0,
                 0x210,
@@ -13130,7 +13130,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 310,
                 0x0,
                 0x213,
                 0x0,
@@ -13154,7 +13154,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 311,
                 0x212,
                 0x0,
                 0x212,
@@ -13178,7 +13178,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 312,
                 0x0,
                 0x215,
                 0x0,
@@ -13202,7 +13202,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 313,
                 0x214,
                 0x0,
                 0x214,
@@ -13226,7 +13226,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 314,
                 0x0,
                 0x217,
                 0x0,
@@ -13250,7 +13250,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 315,
                 0x216,
                 0x0,
                 0x216,
@@ -13274,7 +13274,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 316,
                 0x0,
                 0x219,
                 0x0,
@@ -13298,7 +13298,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 317,
                 0x218,
                 0x0,
                 0x218,
@@ -13322,7 +13322,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 318,
                 0x0,
                 0x21b,
                 0x0,
@@ -13346,7 +13346,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 319,
                 0x21a,
                 0x0,
                 0x21a,
@@ -13418,7 +13418,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 320,
                 0x0,
                 0x21f,
                 0x0,
@@ -13442,7 +13442,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 321,
                 0x21e,
                 0x0,
                 0x21e,
@@ -13610,7 +13610,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 322,
                 0x0,
                 0x227,
                 0x0,
@@ -13634,7 +13634,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 323,
                 0x226,
                 0x0,
                 0x226,
@@ -13658,7 +13658,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 324,
                 0x0,
                 0x229,
                 0x0,
@@ -13682,7 +13682,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 325,
                 0x228,
                 0x0,
                 0x228,
@@ -13706,7 +13706,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 326,
                 0x0,
                 0x22b,
                 0x0,
@@ -13730,7 +13730,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 326,
                 0x22a,
                 0x0,
                 0x22a,
@@ -13754,7 +13754,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 326,
                 0x0,
                 0x22d,
                 0x0,
@@ -13778,7 +13778,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 326,
                 0x22c,
                 0x0,
                 0x22c,
@@ -13802,7 +13802,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 326,
                 0x0,
                 0x22f,
                 0x0,
@@ -13826,7 +13826,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 327,
                 0x22e,
                 0x0,
                 0x22e,
@@ -13850,7 +13850,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 328,
                 0x0,
                 0x231,
                 0x0,
@@ -13874,7 +13874,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 328,
                 0x230,
                 0x0,
                 0x230,
@@ -13898,7 +13898,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 328,
                 0x0,
                 0x233,
                 0x0,
@@ -13922,7 +13922,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 329,
                 0x232,
                 0x0,
                 0x232,
@@ -14042,7 +14042,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 98,
+ 330,
                 0x0,
                 0x0,
                 0x0,
@@ -14066,7 +14066,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 99,
+ 331,
                 0x0,
                 0x0,
                 0x0,
@@ -16639,7 +16639,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 100,
+ 332,
                 0x0,
                 0x0,
                 0x0,
@@ -16663,7 +16663,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 101,
+ 333,
                 0x0,
                 0x0,
                 0x0,
@@ -16687,7 +16687,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 102,
+ 334,
                 0x0,
                 0x0,
                 0x0,
@@ -16711,7 +16711,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 103,
+ 335,
                 0x0,
                 0x0,
                 0x0,
@@ -16735,7 +16735,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 104,
+ 336,
                 0x0,
                 0x0,
                 0x0,
@@ -16759,7 +16759,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 105,
+ 337,
                 0x0,
                 0x0,
                 0x0,
@@ -16783,7 +16783,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 106,
+ 338,
                 0x0,
                 0x0,
                 0x0,
@@ -16807,7 +16807,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 107,
+ 339,
                 0x0,
                 0x0,
                 0x0,
@@ -16831,7 +16831,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 108,
+ 340,
                 0x0,
                 0x0,
                 0x0,
@@ -16942,8 +16942,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -16951,7 +16951,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 4613,
                 0x0,
                 0x0,
                 0x0,
@@ -16966,8 +16966,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -16975,7 +16975,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 4625,
                 0x0,
                 0x0,
                 0x0,
@@ -16990,8 +16990,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -16999,7 +16999,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 4663,
                 0x0,
                 0x0,
                 0x0,
@@ -17014,8 +17014,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -17023,7 +17023,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 4876,
                 0x0,
                 0x0,
                 0x0,
@@ -17038,8 +17038,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -17047,7 +17047,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 4891,
                 0x0,
                 0x0,
                 0x0,
@@ -17062,8 +17062,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -17071,7 +17071,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 4901,
                 0x0,
                 0x0,
                 0x0,
@@ -17086,8 +17086,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -17095,7 +17095,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 4923,
                 0x0,
                 0x0,
                 0x0,
@@ -17110,8 +17110,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -17119,7 +17119,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 5055,
                 0x0,
                 0x0,
                 0x0,
@@ -17134,8 +17134,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -17143,7 +17143,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 5070,
                 0x0,
                 0x0,
                 0x0,
@@ -17902,8 +17902,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ambiguous,
@@ -17911,7 +17911,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 531,
                 0x0,
                 0x0,
                 0x0,
@@ -17926,8 +17926,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ambiguous,
@@ -17935,7 +17935,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 532,
                 0x0,
                 0x0,
                 0x0,
@@ -17950,8 +17950,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ambiguous,
@@ -17959,7 +17959,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 534,
                 0x0,
                 0x0,
                 0x0,
@@ -17974,8 +17974,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ambiguous,
@@ -17983,7 +17983,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 539,
                 0x0,
                 0x0,
                 0x0,
@@ -17998,8 +17998,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -18007,7 +18007,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 527,
                 0x0,
                 0x0,
                 0x0,
@@ -18022,8 +18022,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ambiguous,
@@ -18031,7 +18031,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 535,
                 0x0,
                 0x0,
                 0x0,
@@ -18094,8 +18094,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -18103,7 +18103,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 4605,
                 0x0,
                 0x0,
                 0x0,
@@ -18127,7 +18127,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 109,
+ 341,
                 0x0,
                 0x0,
                 0x0,
@@ -18142,8 +18142,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -18151,7 +18151,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 4930,
                 0x0,
                 0x0,
                 0x0,
@@ -18166,8 +18166,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -18175,7 +18175,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 5065,
                 0x0,
                 0x0,
                 0x0,
@@ -18190,8 +18190,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 20,
                         bidi_class::strong_left_to_right,
                         decomposition_type::super,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -18199,7 +18199,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 5186,
                 0x0,
                 0x0,
                 0x0,
@@ -20423,7 +20423,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::extend,
                 },
- 110,
+ 342,
                 0x0,
                 0x0,
                 0x0,
@@ -20447,7 +20447,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::extend,
                 },
- 110,
+ 342,
                 0x0,
                 0x0,
                 0x0,
@@ -20495,7 +20495,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::extend,
                 },
- 110,
+ 342,
                 0x0,
                 0x0,
                 0x0,
@@ -20519,7 +20519,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::extend,
                 },
- 110,
+ 342,
                 0x0,
                 0x0,
                 0x0,
@@ -21671,7 +21671,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 343,
                 0x0,
                 0x0,
                 0x0,
@@ -21806,8 +21806,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 4,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -21815,7 +21815,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 5252,
                 0x0,
                 0x0,
                 0x0,
@@ -21911,7 +21911,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 343,
                 0x0,
                 0x0,
                 0x0,
@@ -22113,8 +22113,8 @@
                         word_break::any,
                         false,
                         true,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 2,
                         bidi_class::neutral_other,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -22122,7 +22122,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 526,
                 0x0,
                 0x0,
                 0x0,
@@ -22146,7 +22146,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 343,
                 0x0,
                 0x0,
                 0x0,
@@ -22170,7 +22170,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 344,
                 0x0,
                 0x3ac,
                 0x0,
@@ -22194,7 +22194,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 345,
                 0x0,
                 0x0,
                 0x0,
@@ -22218,7 +22218,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 345,
                 0x0,
                 0x3ad,
                 0x0,
@@ -22242,7 +22242,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 346,
                 0x0,
                 0x3ae,
                 0x0,
@@ -22266,7 +22266,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 347,
                 0x0,
                 0x3af,
                 0x0,
@@ -22314,7 +22314,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 348,
                 0x0,
                 0x3cc,
                 0x0,
@@ -22362,7 +22362,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 349,
                 0x0,
                 0x3cd,
                 0x0,
@@ -22386,7 +22386,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 350,
                 0x0,
                 0x3ce,
                 0x0,
@@ -22410,7 +22410,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 351,
                 0x0,
                 0x390,
                 0x0,
@@ -23034,7 +23034,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 352,
                 0x0,
                 0x3ca,
                 0x0,
@@ -23058,7 +23058,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 353,
                 0x0,
                 0x3cb,
                 0x0,
@@ -23082,7 +23082,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 354,
                 0x386,
                 0x0,
                 0x386,
@@ -23106,7 +23106,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 355,
                 0x388,
                 0x0,
                 0x388,
@@ -23130,7 +23130,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 356,
                 0x389,
                 0x0,
                 0x389,
@@ -23154,7 +23154,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 357,
                 0x38a,
                 0x0,
                 0x38a,
@@ -23178,7 +23178,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 358,
                 0x0,
                 0x3b0,
                 0x0,
@@ -23802,7 +23802,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 359,
                 0x3aa,
                 0x0,
                 0x3aa,
@@ -23826,7 +23826,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 360,
                 0x3ab,
                 0x0,
                 0x3ab,
@@ -23850,7 +23850,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 361,
                 0x38c,
                 0x0,
                 0x38c,
@@ -23874,7 +23874,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 362,
                 0x38e,
                 0x0,
                 0x38e,
@@ -23898,7 +23898,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 363,
                 0x38f,
                 0x0,
                 0x38f,
@@ -23922,7 +23922,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 110,
+ 364,
                 0x0,
                 0x3d7,
                 0x0,
@@ -23937,8 +23937,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 4,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -23946,7 +23946,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 111,
+ 5240,
                 0x392,
                 0x0,
                 0x392,
@@ -23961,8 +23961,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 4,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -23970,7 +23970,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 111,
+ 5251,
                 0x398,
                 0x0,
                 0x398,
@@ -23985,8 +23985,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 10,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -23994,7 +23994,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 111,
+ 5274,
                 0x0,
                 0x0,
                 0x0,
@@ -24018,7 +24018,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 111,
+ 365,
                 0x0,
                 0x0,
                 0x0,
@@ -24042,7 +24042,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 111,
+ 365,
                 0x0,
                 0x0,
                 0x0,
@@ -24057,8 +24057,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 4,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -24066,7 +24066,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 111,
+ 5275,
                 0x3a6,
                 0x0,
                 0x3a6,
@@ -24081,8 +24081,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 4,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -24090,7 +24090,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 111,
+ 5261,
                 0x3a0,
                 0x0,
                 0x3a0,
@@ -24114,7 +24114,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 111,
+ 365,
                 0x3cf,
                 0x0,
                 0x3cf,
@@ -24705,8 +24705,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 4,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -24714,7 +24714,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 112,
+ 5254,
                 0x39a,
                 0x0,
                 0x39a,
@@ -24729,8 +24729,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 4,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -24738,7 +24738,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 112,
+ 5266,
                 0x3a1,
                 0x0,
                 0x3a1,
@@ -24753,8 +24753,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 4,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -24762,7 +24762,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 112,
+ 5269,
                 0x3f9,
                 0x0,
                 0x3f9,
@@ -24801,8 +24801,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 10,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -24810,7 +24810,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 112,
+ 5251,
                 0x0,
                 0x3b8,
                 0x0,
@@ -24825,8 +24825,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 4,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -24834,7 +24834,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 112,
+ 5244,
                 0x395,
                 0x0,
                 0x395,
@@ -24921,8 +24921,8 @@
                         word_break::aletter,
                         false,
                         false,
- ucd::sort_type::is_index,
- 0,
+ ucd::sort_type::data1_0x0020_data2_cp,
+ 10,
                         bidi_class::strong_left_to_right,
                         decomposition_type::compat,
                         break_class::ordinary_alpabetic_or_symbol_character,
@@ -24930,7 +24930,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 112,
+ 5269,
                 0x0,
                 0x3f2,
                 0x0,
@@ -25125,7 +25125,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 112,
+ 366,
                 0x0,
                 0x450,
                 0x0,
@@ -25149,7 +25149,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 113,
+ 367,
                 0x0,
                 0x451,
                 0x0,
@@ -25197,7 +25197,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 114,
+ 368,
                 0x0,
                 0x453,
                 0x0,
@@ -25269,7 +25269,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 115,
+ 369,
                 0x0,
                 0x456,
                 0x0,
@@ -25293,7 +25293,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 117,
+ 371,
                 0x0,
                 0x457,
                 0x0,
@@ -25413,7 +25413,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 118,
+ 372,
                 0x0,
                 0x45c,
                 0x0,
@@ -25437,7 +25437,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 119,
+ 373,
                 0x0,
                 0x45d,
                 0x0,
@@ -25461,7 +25461,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 120,
+ 374,
                 0x0,
                 0x45e,
                 0x0,
@@ -25509,7 +25509,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 121,
+ 375,
                 0x0,
                 0x430,
                 0x0,
@@ -25581,7 +25581,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 124,
+ 378,
                 0x0,
                 0x433,
                 0x0,
@@ -25629,7 +25629,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 127,
+ 381,
                 0x0,
                 0x435,
                 0x0,
@@ -25653,7 +25653,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 129,
+ 383,
                 0x0,
                 0x436,
                 0x0,
@@ -25677,7 +25677,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 131,
+ 385,
                 0x0,
                 0x437,
                 0x0,
@@ -25701,7 +25701,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 133,
+ 387,
                 0x0,
                 0x438,
                 0x0,
@@ -25725,7 +25725,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 136,
+ 390,
                 0x0,
                 0x439,
                 0x0,
@@ -25749,7 +25749,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 137,
+ 391,
                 0x0,
                 0x43a,
                 0x0,
@@ -25845,7 +25845,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 140,
+ 394,
                 0x0,
                 0x43e,
                 0x0,
@@ -25965,7 +25965,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 142,
+ 396,
                 0x0,
                 0x443,
                 0x0,
@@ -26061,7 +26061,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 146,
+ 400,
                 0x0,
                 0x447,
                 0x0,
@@ -26157,7 +26157,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 148,
+ 402,
                 0x0,
                 0x44b,
                 0x0,
@@ -26205,7 +26205,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 150,
+ 404,
                 0x0,
                 0x44d,
                 0x0,
@@ -26277,7 +26277,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 152,
+ 406,
                 0x410,
                 0x0,
                 0x410,
@@ -26349,7 +26349,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 155,
+ 409,
                 0x413,
                 0x0,
                 0x413,
@@ -26397,7 +26397,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 158,
+ 412,
                 0x415,
                 0x0,
                 0x415,
@@ -26421,7 +26421,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 160,
+ 414,
                 0x416,
                 0x0,
                 0x416,
@@ -26445,7 +26445,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 162,
+ 416,
                 0x417,
                 0x0,
                 0x417,
@@ -26469,7 +26469,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 164,
+ 418,
                 0x418,
                 0x0,
                 0x418,
@@ -26493,7 +26493,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 167,
+ 421,
                 0x419,
                 0x0,
                 0x419,
@@ -26517,7 +26517,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 168,
+ 422,
                 0x41a,
                 0x0,
                 0x41a,
@@ -26613,7 +26613,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 171,
+ 425,
                 0x41e,
                 0x0,
                 0x41e,
@@ -26733,7 +26733,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 173,
+ 427,
                 0x423,
                 0x0,
                 0x423,
@@ -26829,7 +26829,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 177,
+ 431,
                 0x427,
                 0x0,
                 0x427,
@@ -26925,7 +26925,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 179,
+ 433,
                 0x42b,
                 0x0,
                 0x42b,
@@ -26973,7 +26973,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 181,
+ 435,
                 0x42d,
                 0x0,
                 0x42d,
@@ -27045,7 +27045,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 183,
+ 437,
                 0x400,
                 0x0,
                 0x400,
@@ -27069,7 +27069,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 184,
+ 438,
                 0x401,
                 0x0,
                 0x401,
@@ -27117,7 +27117,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 185,
+ 439,
                 0x403,
                 0x0,
                 0x403,
@@ -27189,7 +27189,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 186,
+ 440,
                 0x406,
                 0x0,
                 0x406,
@@ -27213,7 +27213,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 188,
+ 442,
                 0x407,
                 0x0,
                 0x407,
@@ -27333,7 +27333,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 189,
+ 443,
                 0x40c,
                 0x0,
                 0x40c,
@@ -27357,7 +27357,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 190,
+ 444,
                 0x40d,
                 0x0,
                 0x40d,
@@ -27381,7 +27381,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 191,
+ 445,
                 0x40e,
                 0x0,
                 0x40e,
@@ -27909,7 +27909,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 192,
+ 446,
                 0x0,
                 0x475,
                 0x0,
@@ -27933,7 +27933,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 194,
+ 448,
                 0x474,
                 0x0,
                 0x474,
@@ -27957,7 +27957,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 196,
+ 450,
                 0x0,
                 0x477,
                 0x0,
@@ -27981,7 +27981,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 197,
+ 451,
                 0x476,
                 0x0,
                 0x476,
@@ -28624,7 +28624,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 198,
+ 452,
                 0x0,
                 0x491,
                 0x0,
@@ -28648,7 +28648,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 199,
+ 453,
                 0x490,
                 0x0,
                 0x490,
@@ -29800,7 +29800,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 200,
+ 454,
                 0x0,
                 0x4c2,
                 0x0,
@@ -29824,7 +29824,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 201,
+ 455,
                 0x4c1,
                 0x0,
                 0x4c1,
@@ -30160,7 +30160,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 202,
+ 456,
                 0x0,
                 0x4d1,
                 0x0,
@@ -30184,7 +30184,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 203,
+ 457,
                 0x4d0,
                 0x0,
                 0x4d0,
@@ -30208,7 +30208,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 204,
+ 458,
                 0x0,
                 0x4d3,
                 0x0,
@@ -30232,7 +30232,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 205,
+ 459,
                 0x4d2,
                 0x0,
                 0x4d2,
@@ -30304,7 +30304,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 206,
+ 460,
                 0x0,
                 0x4d7,
                 0x0,
@@ -30328,7 +30328,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 207,
+ 461,
                 0x4d6,
                 0x0,
                 0x4d6,
@@ -30352,7 +30352,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 208,
+ 462,
                 0x0,
                 0x4d9,
                 0x0,
@@ -30376,7 +30376,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 210,
+ 464,
                 0x4d8,
                 0x0,
                 0x4d8,
@@ -30400,7 +30400,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 212,
+ 466,
                 0x0,
                 0x4db,
                 0x0,
@@ -30424,7 +30424,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 213,
+ 467,
                 0x4da,
                 0x0,
                 0x4da,
@@ -30448,7 +30448,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 214,
+ 468,
                 0x0,
                 0x4dd,
                 0x0,
@@ -30472,7 +30472,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 215,
+ 469,
                 0x4dc,
                 0x0,
                 0x4dc,
@@ -30496,7 +30496,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 216,
+ 470,
                 0x0,
                 0x4df,
                 0x0,
@@ -30520,7 +30520,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 217,
+ 471,
                 0x4de,
                 0x0,
                 0x4de,
@@ -30592,7 +30592,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 218,
+ 472,
                 0x0,
                 0x4e3,
                 0x0,
@@ -30616,7 +30616,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 219,
+ 473,
                 0x4e2,
                 0x0,
                 0x4e2,
@@ -30640,7 +30640,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 220,
+ 474,
                 0x0,
                 0x4e5,
                 0x0,
@@ -30664,7 +30664,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 221,
+ 475,
                 0x4e4,
                 0x0,
                 0x4e4,
@@ -30688,7 +30688,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 222,
+ 476,
                 0x0,
                 0x4e7,
                 0x0,
@@ -30712,7 +30712,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 223,
+ 477,
                 0x4e6,
                 0x0,
                 0x4e6,
@@ -30736,7 +30736,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 224,
+ 478,
                 0x0,
                 0x4e9,
                 0x0,
@@ -30760,7 +30760,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 226,
+ 480,
                 0x4e8,
                 0x0,
                 0x4e8,
@@ -30784,7 +30784,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 228,
+ 482,
                 0x0,
                 0x4eb,
                 0x0,
@@ -30808,7 +30808,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 229,
+ 483,
                 0x4ea,
                 0x0,
                 0x4ea,
@@ -30832,7 +30832,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 230,
+ 484,
                 0x0,
                 0x4ed,
                 0x0,
@@ -30856,7 +30856,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 231,
+ 485,
                 0x4ec,
                 0x0,
                 0x4ec,
@@ -30880,7 +30880,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 232,
+ 486,
                 0x0,
                 0x4ef,
                 0x0,
@@ -30904,7 +30904,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 233,
+ 487,
                 0x4ee,
                 0x0,
                 0x4ee,
@@ -30928,7 +30928,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 234,
+ 488,
                 0x0,
                 0x4f1,
                 0x0,
@@ -30952,7 +30952,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 235,
+ 489,
                 0x4f0,
                 0x0,
                 0x4f0,
@@ -30976,7 +30976,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 236,
+ 490,
                 0x0,
                 0x4f3,
                 0x0,
@@ -31000,7 +31000,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 237,
+ 491,
                 0x4f2,
                 0x0,
                 0x4f2,
@@ -31024,7 +31024,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 238,
+ 492,
                 0x0,
                 0x4f5,
                 0x0,
@@ -31048,7 +31048,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 239,
+ 493,
                 0x4f4,
                 0x0,
                 0x4f4,
@@ -31120,7 +31120,7 @@
                         sentence_break::upper,
                         grapheme_cluster_break::any,
                 },
- 240,
+ 494,
                 0x0,
                 0x4f9,
                 0x0,
@@ -31144,7 +31144,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 241,
+ 495,
                 0x4f8,
                 0x0,
                 0x4f8,
@@ -34582,7 +34582,7 @@
                         sentence_break::lower,
                         grapheme_cluster_break::any,
                 },
- 242,
+ 496,
                 0x0,
                 0x587,
                 0x0,
@@ -37102,7 +37102,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 242,
+ 497,
                 0x0,
                 0x0,
                 0x0,
@@ -37126,7 +37126,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 243,
+ 498,
                 0x0,
                 0x0,
                 0x0,
@@ -37150,7 +37150,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 244,
+ 499,
                 0x0,
                 0x0,
                 0x0,
@@ -38320,7 +38320,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 245,
+ 500,
                 0x0,
                 0x0,
                 0x0,
@@ -38344,7 +38344,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 246,
+ 501,
                 0x0,
                 0x0,
                 0x0,
@@ -38368,7 +38368,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 247,
+ 502,
                 0x0,
                 0x0,
                 0x0,
@@ -38392,7 +38392,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 248,
+ 503,
                 0x0,
                 0x0,
                 0x0,
@@ -38416,7 +38416,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 249,
+ 504,
                 0x0,
                 0x0,
                 0x0,
@@ -38440,7 +38440,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 250,
+ 505,
                 0x0,
                 0x0,
                 0x0,
@@ -39232,7 +39232,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 254,
+ 509,
                 0x0,
                 0x0,
                 0x0,
@@ -39280,7 +39280,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 256,
+ 511,
                 0x0,
                 0x0,
                 0x0,
@@ -39808,7 +39808,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 258,
+ 513,
                 0x0,
                 0x0,
                 0x0,
@@ -39832,7 +39832,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 259,
+ 514,
                 0x0,
                 0x0,
                 0x0,
@@ -39856,7 +39856,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 260,
+ 515,
                 0x0,
                 0x0,
                 0x0,
@@ -39880,7 +39880,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 261,
+ 516,
                 0x0,
                 0x0,
                 0x0,
@@ -39904,7 +39904,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 262,
+ 517,
                 0x0,
                 0x0,
                 0x0,
@@ -39928,7 +39928,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 263,
+ 518,
                 0x0,
                 0x0,
                 0x0,
@@ -39952,7 +39952,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 264,
+ 519,
                 0x0,
                 0x0,
                 0x0,
@@ -39976,7 +39976,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 265,
+ 520,
                 0x0,
                 0x0,
                 0x0,
@@ -40000,7 +40000,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 266,
+ 521,
                 0x0,
                 0x0,
                 0x0,
@@ -40024,7 +40024,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 267,
+ 522,
                 0x0,
                 0x0,
                 0x0,
@@ -40312,7 +40312,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 268,
+ 523,
                 0x0,
                 0x0,
                 0x0,
@@ -40336,7 +40336,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 269,
+ 524,
                 0x0,
                 0x0,
                 0x0,
@@ -40360,7 +40360,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 270,
+ 525,
                 0x0,
                 0x0,
                 0x0,
@@ -40384,7 +40384,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 271,
+ 526,
                 0x0,
                 0x0,
                 0x0,
@@ -42124,7 +42124,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 272,
+ 527,
                 0x0,
                 0x0,
                 0x0,
@@ -42172,7 +42172,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 272,
+ 528,
                 0x0,
                 0x0,
                 0x0,
@@ -42580,7 +42580,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 272,
+ 529,
                 0x0,
                 0x0,
                 0x0,
@@ -43276,7 +43276,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 272,
+ 530,
                 0x0,
                 0x0,
                 0x0,
@@ -43300,7 +43300,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 273,
+ 531,
                 0x0,
                 0x0,
                 0x0,
@@ -43324,7 +43324,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 274,
+ 532,
                 0x0,
                 0x0,
                 0x0,
@@ -43348,7 +43348,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 275,
+ 533,
                 0x0,
                 0x0,
                 0x0,
@@ -43372,7 +43372,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 276,
+ 534,
                 0x0,
                 0x0,
                 0x0,
@@ -43396,7 +43396,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 277,
+ 535,
                 0x0,
                 0x0,
                 0x0,
@@ -43420,7 +43420,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 278,
+ 536,
                 0x0,
                 0x0,
                 0x0,
@@ -43444,7 +43444,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 279,
+ 537,
                 0x0,
                 0x0,
                 0x0,
@@ -43468,7 +43468,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 280,
+ 538,
                 0x0,
                 0x0,
                 0x0,
@@ -43492,7 +43492,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 281,
+ 539,
                 0x0,
                 0x0,
                 0x0,
@@ -43588,7 +43588,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 282,
+ 540,
                 0x0,
                 0x0,
                 0x0,
@@ -43612,7 +43612,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 283,
+ 541,
                 0x0,
                 0x0,
                 0x0,
@@ -44029,7 +44029,7 @@
                         sentence_break::format,
                         grapheme_cluster_break::control,
                 },
- 284,
+ 542,
                 0x0,
                 0x0,
                 0x0,
@@ -44149,7 +44149,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 285,
+ 543,
                 0x0,
                 0x0,
                 0x0,
@@ -44341,7 +44341,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 286,
+ 544,
                 0x0,
                 0x0,
                 0x0,
@@ -44605,7 +44605,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 287,
+ 545,
                 0x0,
                 0x0,
                 0x0,
@@ -44749,7 +44749,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 288,
+ 546,
                 0x0,
                 0x0,
                 0x0,
@@ -44773,7 +44773,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 289,
+ 547,
                 0x0,
                 0x0,
                 0x0,
@@ -44797,7 +44797,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 290,
+ 548,
                 0x0,
                 0x0,
                 0x0,
@@ -48286,7 +48286,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 291,
+ 549,
                 0x0,
                 0x0,
                 0x0,
@@ -48310,7 +48310,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 292,
+ 550,
                 0x0,
                 0x0,
                 0x0,
@@ -48334,7 +48334,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 293,
+ 551,
                 0x0,
                 0x0,
                 0x0,
@@ -48358,7 +48358,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 294,
+ 552,
                 0x0,
                 0x0,
                 0x0,
@@ -48382,7 +48382,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 295,
+ 553,
                 0x0,
                 0x0,
                 0x0,
@@ -48406,7 +48406,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 296,
+ 554,
                 0x0,
                 0x0,
                 0x0,
@@ -48430,7 +48430,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 297,
+ 555,
                 0x0,
                 0x0,
                 0x0,
@@ -48454,7 +48454,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 298,
+ 556,
                 0x0,
                 0x0,
                 0x0,
@@ -48478,7 +48478,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 299,
+ 557,
                 0x0,
                 0x0,
                 0x0,
@@ -48502,7 +48502,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 300,
+ 558,
                 0x0,
                 0x0,
                 0x0,
@@ -49246,7 +49246,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 301,
+ 559,
                 0x0,
                 0x0,
                 0x0,
@@ -49270,7 +49270,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 302,
+ 560,
                 0x0,
                 0x0,
                 0x0,
@@ -49294,7 +49294,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 303,
+ 561,
                 0x0,
                 0x0,
                 0x0,
@@ -53907,7 +53907,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 562,
                 0x0,
                 0x0,
                 0x0,
@@ -54099,7 +54099,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 563,
                 0x0,
                 0x0,
                 0x0,
@@ -54171,7 +54171,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 564,
                 0x0,
                 0x0,
                 0x0,
@@ -55035,7 +55035,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 565,
                 0x0,
                 0x0,
                 0x0,
@@ -55059,7 +55059,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 566,
                 0x0,
                 0x0,
                 0x0,
@@ -55083,7 +55083,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 567,
                 0x0,
                 0x0,
                 0x0,
@@ -55107,7 +55107,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 568,
                 0x0,
                 0x0,
                 0x0,
@@ -55131,7 +55131,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 569,
                 0x0,
                 0x0,
                 0x0,
@@ -55155,7 +55155,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 570,
                 0x0,
                 0x0,
                 0x0,
@@ -55179,7 +55179,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 571,
                 0x0,
                 0x0,
                 0x0,
@@ -55203,7 +55203,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 572,
                 0x0,
                 0x0,
                 0x0,
@@ -55371,7 +55371,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 304,
+ 573,
                 0x0,
                 0x0,
                 0x0,
@@ -55395,7 +55395,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 305,
+ 574,
                 0x0,
                 0x0,
                 0x0,
@@ -55419,7 +55419,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 306,
+ 575,
                 0x0,
                 0x0,
                 0x0,
@@ -55443,7 +55443,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 307,
+ 576,
                 0x0,
                 0x0,
                 0x0,
@@ -55467,7 +55467,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 308,
+ 577,
                 0x0,
                 0x0,
                 0x0,
@@ -55491,7 +55491,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 309,
+ 578,
                 0x0,
                 0x0,
                 0x0,
@@ -55515,7 +55515,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 310,
+ 579,
                 0x0,
                 0x0,
                 0x0,
@@ -55539,7 +55539,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 311,
+ 580,
                 0x0,
                 0x0,
                 0x0,
@@ -55563,7 +55563,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 312,
+ 581,
                 0x0,
                 0x0,
                 0x0,
@@ -55587,7 +55587,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 313,
+ 582,
                 0x0,
                 0x0,
                 0x0,
@@ -57713,7 +57713,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 314,
+ 583,
                 0x0,
                 0x0,
                 0x0,
@@ -57809,7 +57809,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 317,
+ 586,
                 0x0,
                 0x0,
                 0x0,
@@ -57833,7 +57833,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 318,
+ 587,
                 0x0,
                 0x0,
                 0x0,
@@ -57881,7 +57881,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 319,
+ 588,
                 0x0,
                 0x0,
                 0x0,
@@ -58217,7 +58217,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 320,
+ 589,
                 0x0,
                 0x0,
                 0x0,
@@ -58241,7 +58241,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 320,
+ 590,
                 0x0,
                 0x0,
                 0x0,
@@ -58289,7 +58289,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 320,
+ 591,
                 0x0,
                 0x0,
                 0x0,
@@ -58457,7 +58457,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 320,
+ 592,
                 0x0,
                 0x0,
                 0x0,
@@ -58481,7 +58481,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 321,
+ 593,
                 0x0,
                 0x0,
                 0x0,
@@ -58505,7 +58505,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 322,
+ 594,
                 0x0,
                 0x0,
                 0x0,
@@ -58529,7 +58529,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 323,
+ 595,
                 0x0,
                 0x0,
                 0x0,
@@ -58553,7 +58553,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 324,
+ 596,
                 0x0,
                 0x0,
                 0x0,
@@ -58577,7 +58577,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 325,
+ 597,
                 0x0,
                 0x0,
                 0x0,
@@ -58601,7 +58601,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 326,
+ 598,
                 0x0,
                 0x0,
                 0x0,
@@ -58625,7 +58625,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 327,
+ 599,
                 0x0,
                 0x0,
                 0x0,
@@ -58649,7 +58649,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 328,
+ 600,
                 0x0,
                 0x0,
                 0x0,
@@ -58673,7 +58673,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 329,
+ 601,
                 0x0,
                 0x0,
                 0x0,
@@ -58793,7 +58793,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 330,
+ 602,
                 0x0,
                 0x0,
                 0x0,
@@ -58817,7 +58817,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 331,
+ 603,
                 0x0,
                 0x0,
                 0x0,
@@ -58841,7 +58841,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 332,
+ 604,
                 0x0,
                 0x0,
                 0x0,
@@ -58865,7 +58865,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 333,
+ 605,
                 0x0,
                 0x0,
                 0x0,
@@ -60320,7 +60320,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 334,
+ 606,
                 0x0,
                 0x0,
                 0x0,
@@ -60392,7 +60392,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 334,
+ 607,
                 0x0,
                 0x0,
                 0x0,
@@ -61232,7 +61232,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 334,
+ 607,
                 0x0,
                 0x0,
                 0x0,
@@ -61256,7 +61256,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 334,
+ 608,
                 0x0,
                 0x0,
                 0x0,
@@ -61280,7 +61280,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 334,
+ 609,
                 0x0,
                 0x0,
                 0x0,
@@ -61352,7 +61352,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 334,
+ 610,
                 0x0,
                 0x0,
                 0x0,
@@ -61544,7 +61544,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 334,
+ 611,
                 0x0,
                 0x0,
                 0x0,
@@ -61568,7 +61568,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 335,
+ 612,
                 0x0,
                 0x0,
                 0x0,
@@ -61592,7 +61592,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 336,
+ 613,
                 0x0,
                 0x0,
                 0x0,
@@ -61616,7 +61616,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 337,
+ 614,
                 0x0,
                 0x0,
                 0x0,
@@ -61640,7 +61640,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 338,
+ 615,
                 0x0,
                 0x0,
                 0x0,
@@ -61664,7 +61664,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 339,
+ 616,
                 0x0,
                 0x0,
                 0x0,
@@ -61688,7 +61688,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 340,
+ 617,
                 0x0,
                 0x0,
                 0x0,
@@ -61712,7 +61712,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 341,
+ 618,
                 0x0,
                 0x0,
                 0x0,
@@ -61736,7 +61736,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 342,
+ 619,
                 0x0,
                 0x0,
                 0x0,
@@ -61760,7 +61760,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 343,
+ 620,
                 0x0,
                 0x0,
                 0x0,
@@ -64625,7 +64625,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 344,
+ 621,
                 0x0,
                 0x0,
                 0x0,
@@ -64649,7 +64649,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 345,
+ 622,
                 0x0,
                 0x0,
                 0x0,
@@ -64673,7 +64673,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 346,
+ 623,
                 0x0,
                 0x0,
                 0x0,
@@ -64697,7 +64697,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 347,
+ 624,
                 0x0,
                 0x0,
                 0x0,
@@ -64721,7 +64721,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 348,
+ 625,
                 0x0,
                 0x0,
                 0x0,
@@ -64745,7 +64745,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 349,
+ 626,
                 0x0,
                 0x0,
                 0x0,
@@ -64769,7 +64769,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 350,
+ 627,
                 0x0,
                 0x0,
                 0x0,
@@ -64793,7 +64793,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 351,
+ 628,
                 0x0,
                 0x0,
                 0x0,
@@ -64817,7 +64817,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 352,
+ 629,
                 0x0,
                 0x0,
                 0x0,
@@ -64841,7 +64841,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 353,
+ 630,
                 0x0,
                 0x0,
                 0x0,
@@ -66967,7 +66967,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 354,
+ 631,
                 0x0,
                 0x0,
                 0x0,
@@ -66991,7 +66991,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 358,
+ 635,
                 0x0,
                 0x0,
                 0x0,
@@ -67063,7 +67063,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 359,
+ 636,
                 0x0,
                 0x0,
                 0x0,
@@ -67087,7 +67087,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 360,
+ 637,
                 0x0,
                 0x0,
                 0x0,
@@ -67471,7 +67471,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 361,
+ 638,
                 0x0,
                 0x0,
                 0x0,
@@ -67495,7 +67495,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 361,
+ 639,
                 0x0,
                 0x0,
                 0x0,
@@ -67711,7 +67711,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 361,
+ 640,
                 0x0,
                 0x0,
                 0x0,
@@ -67735,7 +67735,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 362,
+ 641,
                 0x0,
                 0x0,
                 0x0,
@@ -67759,7 +67759,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 363,
+ 642,
                 0x0,
                 0x0,
                 0x0,
@@ -67783,7 +67783,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 364,
+ 643,
                 0x0,
                 0x0,
                 0x0,
@@ -67807,7 +67807,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 365,
+ 644,
                 0x0,
                 0x0,
                 0x0,
@@ -67831,7 +67831,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 366,
+ 645,
                 0x0,
                 0x0,
                 0x0,
@@ -67855,7 +67855,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 367,
+ 646,
                 0x0,
                 0x0,
                 0x0,
@@ -67879,7 +67879,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 368,
+ 647,
                 0x0,
                 0x0,
                 0x0,
@@ -67903,7 +67903,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 369,
+ 648,
                 0x0,
                 0x0,
                 0x0,
@@ -67927,7 +67927,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 370,
+ 649,
                 0x0,
                 0x0,
                 0x0,
@@ -68780,7 +68780,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 371,
+ 650,
                 0x0,
                 0x0,
                 0x0,
@@ -68828,7 +68828,7 @@
                         sentence_break::oletter,
                         grapheme_cluster_break::any,
                 },
- 373,
+ 652,
                 0x0,
                 0x0,
                 0x0,
@@ -70028,7 +70028,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 374,
+ 653,
                 0x0,
                 0x0,
                 0x0,
@@ -70052,7 +70052,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 377,
+ 656,
                 0x0,
                 0x0,
                 0x0,
@@ -70124,7 +70124,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 379,
+ 658,
                 0x0,
                 0x0,
                 0x0,
@@ -70148,7 +70148,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 380,
+ 659,
                 0x0,
                 0x0,
                 0x0,
@@ -70172,7 +70172,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 381,
+ 660,
                 0x0,
                 0x0,
                 0x0,
@@ -70796,7 +70796,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 382,
+ 661,
                 0x0,
                 0x0,
                 0x0,
@@ -70820,7 +70820,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 383,
+ 662,
                 0x0,
                 0x0,
                 0x0,
@@ -70844,7 +70844,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 384,
+ 663,
                 0x0,
                 0x0,
                 0x0,
@@ -70868,7 +70868,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 385,
+ 664,
                 0x0,
                 0x0,
                 0x0,
@@ -70892,7 +70892,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 386,
+ 665,
                 0x0,
                 0x0,
                 0x0,
@@ -70916,7 +70916,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 387,
+ 666,
                 0x0,
                 0x0,
                 0x0,
@@ -70940,7 +70940,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 388,
+ 667,
                 0x0,
                 0x0,
                 0x0,
@@ -70964,7 +70964,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 389,
+ 668,
                 0x0,
                 0x0,
                 0x0,
@@ -70988,7 +70988,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 390,
+ 669,
                 0x0,
                 0x0,
                 0x0,
@@ -71012,7 +71012,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 391,
+ 670,
                 0x0,
                 0x0,
                 0x0,
@@ -73110,7 +73110,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::extend,
                 },
- 392,
+ 671,
                 0x0,
                 0x0,
                 0x0,
@@ -73158,7 +73158,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::extend,
                 },
- 394,
+ 673,
                 0x0,
                 0x0,
                 0x0,
@@ -73878,7 +73878,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 395,
+ 674,
                 0x0,
                 0x0,
                 0x0,
@@ -73902,7 +73902,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 396,
+ 675,
                 0x0,
                 0x0,
                 0x0,
@@ -73926,7 +73926,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 397,
+ 676,
                 0x0,
                 0x0,
                 0x0,
@@ -73950,7 +73950,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 398,
+ 677,
                 0x0,
                 0x0,
                 0x0,
@@ -73974,7 +73974,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 399,
+ 678,
                 0x0,
                 0x0,
                 0x0,
@@ -73998,7 +73998,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 400,
+ 679,
                 0x0,
                 0x0,
                 0x0,
@@ -74022,7 +74022,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 401,
+ 680,
                 0x0,
                 0x0,
                 0x0,
@@ -74046,7 +74046,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 402,
+ 681,
                 0x0,
                 0x0,
                 0x0,
@@ -74070,7 +74070,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 403,
+ 682,
                 0x0,
                 0x0,
                 0x0,
@@ -74094,7 +74094,7 @@
                         sentence_break::numeric,
                         grapheme_cluster_break::any,
                 },
- 404,
+ 683,
                 0x0,
                 0x0,
                 0x0,
@@ -74310,7 +74310,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 405,
+ 684,
                 0x0,
                 0x0,
                 0x0,
@@ -74334,7 +74334,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 406,
+ 685,
                 0x0,
                 0x0,
                 0x0,
@@ -74358,7 +74358,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 407,
+ 686,
                 0x0,
                 0x0,
                 0x0,
@@ -74382,7 +74382,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 408,
+ 687,
                 0x0,
                 0x0,
                 0x0,
@@ -74406,7 +74406,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 409,
+ 688,
                 0x0,
                 0x0,
                 0x0,
@@ -74430,7 +74430,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 410,
+ 689,
                 0x0,
                 0x0,
                 0x0,
@@ -74454,7 +74454,7 @@
                         sentence_break::any,
                         grapheme_cluster_break::any,
                 },
- 411,
+ 690,
                 0x0,
                 0x0,
                 0x0,