diff options
| author | nobu <nobu@b2dd03c8-39d4-4d8f-98ff-823fe69b080e> | 2017-12-02 03:12:51 +0000 |
|---|---|---|
| committer | nobu <nobu@b2dd03c8-39d4-4d8f-98ff-823fe69b080e> | 2017-12-02 03:12:51 +0000 |
| commit | 01830719f641ee9b56e0a6f5ef24759202761476 (patch) | |
| tree | 0789c4d448287ca4873f4fefee065006d9dbed1b | |
| parent | 8b180dd74ef7604bc896275ce2a13a1d8c2c2565 (diff) | |
fix for emoji-data.txt
* common.mk: download emoji-data.txt. As emoji data files are
located in a separate directory in Unicode.org site, reearranged
Unicode data files directories same as the site.
* tool/enc-unicode.rb (get_file): search emoji data files in the
second argument path.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60977 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
| -rw-r--r-- | common.mk | 23 | ||||
| -rw-r--r-- | enc/unicode/10.0.0/casefold.h | 48 | ||||
| -rw-r--r-- | enc/unicode/10.0.0/name2ctype.h | 990 | ||||
| -rw-r--r-- | template/unicode_norm_gen.tmpl | 2 | ||||
| -rwxr-xr-x | tool/enc-unicode.rb | 6 |
5 files changed, 677 insertions, 392 deletions
@@ -16,12 +16,14 @@ gnumake_recursive = enable_shared = $(ENABLE_SHARED:no=) UNICODE_VERSION = 10.0.0 +UNICODE_EMOJI_VERSION = 5.0 ### set the following environment variable or uncomment the line if ### the Unicode data files should be updated completely on every update ('make up',...). # ALWAYS_UPDATE_UNICODE = yes -UNICODE_DATA_DIR = enc/unicode/data/$(UNICODE_VERSION) +UNICODE_DATA_DIR = enc/unicode/data/$(UNICODE_VERSION)/ucd UNICODE_SRC_DATA_DIR = $(srcdir)/$(UNICODE_DATA_DIR) +UNICODE_SRC_EMOJI_DATA_DIR = $(srcdir)/enc/unicode/data/emoji/$(UNICODE_EMOJI_VERSION) UNICODE_HDR_DIR = $(srcdir)/enc/unicode/$(UNICODE_VERSION) UNICODE_DATA_HEADERS = \ $(UNICODE_HDR_DIR)/casefold.h \ @@ -1207,21 +1209,34 @@ UNICODE_PROPERTY_FILES = \ $(UNICODE_SRC_DATA_DIR)/auxiliary/GraphemeBreakProperty.txt \ $(empty) +UNICODE_EMOJI_FILES = \ + $(UNICODE_SRC_EMOJI_DATA_DIR)/emoji-data.txt \ + $(empty) + update-unicode: $(UNICODE_FILES) CACHE_DIR = $(srcdir)/.downloaded-cache UNICODE_DOWNLOAD = \ $(BASERUBY) $(srcdir)/tool/downloader.rb \ --cache-dir=$(CACHE_DIR) \ - -d $(srcdir)/$(UNICODE_DATA_DIR) \ + -d $(UNICODE_SRC_DATA_DIR) \ -p $(UNICODE_VERSION)/ucd \ -e $(ALWAYS_UPDATE_UNICODE:yes=-a) unicode +UNICODE_EMOJI_DOWNLOAD = \ + $(BASERUBY) $(srcdir)/tool/downloader.rb \ + --cache-dir=$(CACHE_DIR) \ + -d $(UNICODE_SRC_EMOJI_DATA_DIR) \ + -p emoji/$(UNICODE_EMOJI_VERSION) \ + -e $(ALWAYS_UPDATE_UNICODE:yes=-a) unicode $(UNICODE_PROPERTY_FILES): update-unicode-property-files update-unicode-property-files: $(ECHO) Downloading Unicode $(UNICODE_VERSION) property files... $(Q) $(MAKEDIRS) "$(UNICODE_SRC_DATA_DIR)/auxiliary" $(Q) $(UNICODE_DOWNLOAD) $(UNICODE_PROPERTY_FILES) + $(ECHO) Downloading Unicode emoji $(UNICODE_VERSION) files... + $(Q) $(MAKEDIRS) "$(UNICODE_SRC_EMOJI_DATA_DIR)" + $(Q) $(UNICODE_EMOJI_DOWNLOAD) $(UNICODE_EMOJI_FILES) $(UNICODE_FILES): update-unicode-files update-unicode-files: @@ -1259,7 +1274,9 @@ $(UNICODE_HDR_DIR)/$(ALWAYS_UPDATE_UNICODE:yes=name2ctype.h): \ $(UNICODE_HDR_DIR)/name2ctype.h: $(MAKEDIRS) $(@D) - $(BOOTSTRAPRUBY) $(srcdir)/tool/enc-unicode.rb --header $(UNICODE_SRC_DATA_DIR) > $@ + $(BOOTSTRAPRUBY) $(srcdir)/tool/enc-unicode.rb --header \ + $(UNICODE_SRC_DATA_DIR) $(UNICODE_SRC_EMOJI_DATA_DIR) > $@.new + $(MV) $@.new $@ # the next non-comment line was: # $(UNICODE_HDR_DIR)/casefold.h: $(srcdir)/enc/unicode/case-folding.rb \ diff --git a/enc/unicode/10.0.0/casefold.h b/enc/unicode/10.0.0/casefold.h index 714ebe6f73..a8f5a24e24 100644 --- a/enc/unicode/10.0.0/casefold.h +++ b/enc/unicode/10.0.0/casefold.h @@ -1419,7 +1419,7 @@ static const CaseFold_11_Type CaseFold_11_Table[] = { {0x0130, {2|F|D, {0x0069, 0x0307}}}, }; -/* C code produced by gperf version 3.0.4 */ +/* ANSI-C code produced by gperf version 3.1 */ /* Command-line: gperf -7 -k1,2,3 -F,-1 -c -j1 -i1 -t -T -E -C -H onigenc_unicode_CaseFold_11_hash -N onigenc_unicode_CaseFold_11_lookup -n */ /* maximum key range = 3623, duplicates = 0 */ @@ -1462,12 +1462,6 @@ onigenc_unicode_CaseFold_11_hash(const OnigCodePoint code) return asso_values[bits_of(code, 2)+81] + asso_values[bits_of(code, 1)+2] + asso_values[bits_of(code, 0)]; } -#ifdef __GNUC__ -__inline -#if defined __GNUC_STDC_INLINE__ || defined __GNUC_GNU_INLINE__ -__attribute__ ((__gnu_inline__)) -#endif -#endif static const CodePointList3 * onigenc_unicode_CaseFold_11_lookup(const OnigCodePoint code) { @@ -3583,9 +3577,9 @@ onigenc_unicode_CaseFold_11_lookup(const OnigCodePoint code) if (code <= MAX_CODE_VALUE && code >= MIN_CODE_VALUE) { - register int key = onigenc_unicode_CaseFold_11_hash(code); + register unsigned int key = onigenc_unicode_CaseFold_11_hash(code); - if (key <= MAX_HASH_VALUE && key >= 0) + if (key <= MAX_HASH_VALUE) { register short s = wordlist[key]; @@ -4868,7 +4862,7 @@ static const CaseUnfold_11_Type CaseUnfold_11_Table[] = { {0x0069, {1|U, {0x0049}}}, }; -/* C code produced by gperf version 3.0.4 */ +/* ANSI-C code produced by gperf version 3.1 */ /* Command-line: gperf -7 -k1,2,3 -F,-1 -c -j1 -i1 -t -T -E -C -H onigenc_unicode_CaseUnfold_11_hash -N onigenc_unicode_CaseUnfold_11_lookup -n */ /* maximum key range = 2216, duplicates = 0 */ @@ -4910,12 +4904,6 @@ onigenc_unicode_CaseUnfold_11_hash(const OnigCodePoint code) return asso_values[bits_of(code, 2)+66] + asso_values[bits_of(code, 1)+4] + asso_values[bits_of(code, 0)]; } -#ifdef __GNUC__ -__inline -#if defined __GNUC_STDC_INLINE__ || defined __GNUC_GNU_INLINE__ -__attribute__ ((__gnu_inline__)) -#endif -#endif static const CodePointList3 * onigenc_unicode_CaseUnfold_11_lookup(const OnigCodePoint code) { @@ -6602,9 +6590,9 @@ onigenc_unicode_CaseUnfold_11_lookup(const OnigCodePoint code) if (code <= MAX_CODE_VALUE && code >= MIN_CODE_VALUE) { - register int key = onigenc_unicode_CaseUnfold_11_hash(code); + register unsigned int key = onigenc_unicode_CaseUnfold_11_hash(code); - if (key <= MAX_HASH_VALUE && key >= 0) + if (key <= MAX_HASH_VALUE) { register short s = wordlist[key]; @@ -6679,7 +6667,7 @@ static const CaseUnfold_12_Type CaseUnfold_12_Table[] = { {{0x0069, 0x0307}, {1, {0x0130}}}, }; -/* C code produced by gperf version 3.0.4 */ +/* ANSI-C code produced by gperf version 3.1 */ /* Command-line: gperf -7 -k1,2,3,4,5,6 -F,-1 -c -j1 -i1 -t -T -E -C -H onigenc_unicode_CaseUnfold_12_hash -N onigenc_unicode_CaseUnfold_12_lookup -n */ /* maximum key range = 71, duplicates = 0 */ @@ -6714,12 +6702,6 @@ onigenc_unicode_CaseUnfold_12_hash(const OnigCodePoint *codes) return asso_values[bits_at(codes, 5)] + asso_values[bits_at(codes, 4)] + asso_values[bits_at(codes, 3)] + asso_values[bits_at(codes, 2)] + asso_values[bits_at(codes, 1)] + asso_values[bits_at(codes, 0)]; } -#ifdef __GNUC__ -__inline -#if defined __GNUC_STDC_INLINE__ || defined __GNUC_GNU_INLINE__ -__attribute__ ((__gnu_inline__)) -#endif -#endif static const CodePointList2 * onigenc_unicode_CaseUnfold_12_lookup(const OnigCodePoint *codes) { @@ -6804,9 +6786,9 @@ onigenc_unicode_CaseUnfold_12_lookup(const OnigCodePoint *codes) if (codes[0] <= MAX_CODE_VALUE && codes[0] >= MIN_CODE_VALUE && codes[1] <= MAX_CODE_VALUE && codes[1] >= MIN_CODE_VALUE) { - register int key = onigenc_unicode_CaseUnfold_12_hash(codes); + register unsigned int key = onigenc_unicode_CaseUnfold_12_hash(codes); - if (key <= MAX_HASH_VALUE && key >= 0) + if (key <= MAX_HASH_VALUE) { register short s = wordlist[key]; @@ -6835,7 +6817,7 @@ static const CaseUnfold_13_Type CaseUnfold_13_Table[] = { {{0x03c9, 0x0342, 0x03b9}, {1, {0x1ff7}}}, }; -/* C code produced by gperf version 3.0.4 */ +/* ANSI-C code produced by gperf version 3.1 */ /* Command-line: gperf -7 -k1,2,3,4,5,6,7,8,9 -F,-1 -c -j1 -i1 -t -T -E -C -H onigenc_unicode_CaseUnfold_13_hash -N onigenc_unicode_CaseUnfold_13_lookup -n */ /* maximum key range = 20, duplicates = 0 */ @@ -6870,12 +6852,6 @@ onigenc_unicode_CaseUnfold_13_hash(const OnigCodePoint *codes) return asso_values[bits_at(codes, 8)] + asso_values[bits_at(codes, 7)] + asso_values[bits_at(codes, 6)] + asso_values[bits_at(codes, 5)] + asso_values[bits_at(codes, 4)] + asso_values[bits_at(codes, 3)] + asso_values[bits_at(codes, 2)] + asso_values[bits_at(codes, 1)] + asso_values[bits_at(codes, 0)]; } -#ifdef __GNUC__ -__inline -#if defined __GNUC_STDC_INLINE__ || defined __GNUC_GNU_INLINE__ -__attribute__ ((__gnu_inline__)) -#endif -#endif static const CodePointList2 * onigenc_unicode_CaseUnfold_13_lookup(const OnigCodePoint *codes) { @@ -6918,9 +6894,9 @@ onigenc_unicode_CaseUnfold_13_lookup(const OnigCodePoint *codes) codes[1] <= MAX_CODE_VALUE && codes[1] >= MIN_CODE_VALUE && codes[2] <= MAX_CODE_VALUE && codes[2] >= MIN_CODE_VALUE) { - register int key = onigenc_unicode_CaseUnfold_13_hash(codes); + register unsigned int key = onigenc_unicode_CaseUnfold_13_hash(codes); - if (key <= MAX_HASH_VALUE && key >= 0) + if (key <= MAX_HASH_VALUE) { register short s = wordlist[key]; diff --git a/enc/unicode/10.0.0/name2ctype.h b/enc/unicode/10.0.0/name2ctype.h index 94b7c83f6e..6aab2cb3ae 100644 --- a/enc/unicode/10.0.0/name2ctype.h +++ b/enc/unicode/10.0.0/name2ctype.h @@ -1,4 +1,4 @@ -/* C code produced by gperf version 3.0.4 */ +/* ANSI-C code produced by gperf version 3.1 */ /* Command-line: gperf -7 -c -j1 -i1 -t -C -P -T -H uniname2ctype_hash -Q uniname2ctype_pool -N uniname2ctype_p */ #ifndef USE_UNICODE_PROPERTIES /* Computed positions: -k'1,3' */ @@ -30,7 +30,7 @@ && ('w' == 119) && ('x' == 120) && ('y' == 121) && ('z' == 122) \ && ('{' == 123) && ('|' == 124) && ('}' == 125) && ('~' == 126)) /* The character set is not based on ISO-646. */ -error "gperf generated tables don't work with this execution character set. Please report a bug to <bug-gnu-gperf@gnu.org>." +#error "gperf generated tables don't work with this execution character set. Please report a bug to <bug-gperf@gnu.org>." #endif @@ -21317,6 +21317,289 @@ static const OnigCodePoint CR_Regional_Indicator[] = { 0x1f1e6, 0x1f1ff, }; /* CR_Regional_Indicator */ +/* 'Emoji': Emoji */ +static const OnigCodePoint CR_Emoji[] = { + 145, + 0x0023, 0x0023, + 0x002a, 0x002a, + 0x0030, 0x0039, + 0x00a9, 0x00a9, + 0x00ae, 0x00ae, + 0x203c, 0x203c, + 0x2049, 0x2049, + 0x2122, 0x2122, + 0x2139, 0x2139, + 0x2194, 0x2199, + 0x21a9, 0x21aa, + 0x231a, 0x231b, + 0x2328, 0x2328, + 0x23cf, 0x23cf, + 0x23e9, 0x23f3, + 0x23f8, 0x23fa, + 0x24c2, 0x24c2, + 0x25aa, 0x25ab, + 0x25b6, 0x25b6, + 0x25c0, 0x25c0, + 0x25fb, 0x25fe, + 0x2600, 0x2604, + 0x260e, 0x260e, + 0x2611, 0x2611, + 0x2614, 0x2615, + 0x2618, 0x2618, + 0x261d, 0x261d, + 0x2620, 0x2620, + 0x2622, 0x2623, + 0x2626, 0x2626, + 0x262a, 0x262a, + 0x262e, 0x262f, + 0x2638, 0x263a, + 0x2640, 0x2640, + 0x2642, 0x2642, + 0x2648, 0x2653, + 0x2660, 0x2660, + 0x2663, 0x2663, + 0x2665, 0x2666, + 0x2668, 0x2668, + 0x267b, 0x267b, + 0x267f, 0x267f, + 0x2692, 0x2697, + 0x2699, 0x2699, + 0x269b, 0x269c, + 0x26a0, 0x26a1, + 0x26aa, 0x26ab, + 0x26b0, 0x26b1, + 0x26bd, 0x26be, + 0x26c4, 0x26c5, + 0x26c8, 0x26c8, + 0x26ce, 0x26cf, + 0x26d1, 0x26d1, + 0x26d3, 0x26d4, + 0x26e9, 0x26ea, + 0x26f0, 0x26f5, + 0x26f7, 0x26fa, + 0x26fd, 0x26fd, + 0x2702, 0x2702, + 0x2705, 0x2705, + 0x2708, 0x270d, + 0x270f, 0x270f, + 0x2712, 0x2712, + 0x2714, 0x2714, + 0x2716, 0x2716, + 0x271d, 0x271d, + 0x2721, 0x2721, + 0x2728, 0x2728, + 0x2733, 0x2734, + 0x2744, 0x2744, + 0x2747, 0x2747, + 0x274c, 0x274c, + 0x274e, 0x274e, + 0x2753, 0x2755, + 0x2757, 0x2757, + 0x2763, 0x2764, + 0x2795, 0x2797, + 0x27a1, 0x27a1, + 0x27b0, 0x27b0, + 0x27bf, 0x27bf, + 0x2934, 0x2935, + 0x2b05, 0x2b07, + 0x2b1b, 0x2b1c, + 0x2b50, 0x2b50, + 0x2b55, 0x2b55, + 0x3030, 0x3030, + 0x303d, 0x303d, + 0x3297, 0x3297, + 0x3299, 0x3299, + 0x1f004, 0x1f004, + 0x1f0cf, 0x1f0cf, + 0x1f170, 0x1f171, + 0x1f17e, 0x1f17f, + 0x1f18e, 0x1f18e, + 0x1f191, 0x1f19a, + 0x1f1e6, 0x1f1ff, + 0x1f201, 0x1f202, + 0x1f21a, 0x1f21a, + 0x1f22f, 0x1f22f, + 0x1f232, 0x1f23a, + 0x1f250, 0x1f251, + 0x1f300, 0x1f321, + 0x1f324, 0x1f393, + 0x1f396, 0x1f397, + 0x1f399, 0x1f39b, + 0x1f39e, 0x1f3f0, + 0x1f3f3, 0x1f3f5, + 0x1f3f7, 0x1f4fd, + 0x1f4ff, 0x1f53d, + 0x1f549, 0x1f54e, + 0x1f550, 0x1f567, + 0x1f56f, 0x1f570, + 0x1f573, 0x1f57a, + 0x1f587, 0x1f587, + 0x1f58a, 0x1f58d, + 0x1f590, 0x1f590, + 0x1f595, 0x1f596, + 0x1f5a4, 0x1f5a5, + 0x1f5a8, 0x1f5a8, + 0x1f5b1, 0x1f5b2, + 0x1f5bc, 0x1f5bc, + 0x1f5c2, 0x1f5c4, + 0x1f5d1, 0x1f5d3, + 0x1f5dc, 0x1f5de, + 0x1f5e1, 0x1f5e1, + 0x1f5e3, 0x1f5e3, + 0x1f5e8, 0x1f5e8, + 0x1f5ef, 0x1f5ef, + 0x1f5f3, 0x1f5f3, + 0x1f5fa, 0x1f64f, + 0x1f680, 0x1f6c5, + 0x1f6cb, 0x1f6d2, + 0x1f6e0, 0x1f6e5, + 0x1f6e9, 0x1f6e9, + 0x1f6eb, 0x1f6ec, + 0x1f6f0, 0x1f6f0, + 0x1f6f3, 0x1f6f8, + 0x1f910, 0x1f93a, + 0x1f93c, 0x1f93e, + 0x1f940, 0x1f945, + 0x1f947, 0x1f94c, + 0x1f950, 0x1f96b, + 0x1f980, 0x1f997, + 0x1f9c0, 0x1f9c0, + 0x1f9d0, 0x1f9e6, +}; /* CR_Emoji */ + +/* 'Emoji_Presentation': Emoji */ +static const OnigCodePoint CR_Emoji_Presentation[] = { + 75, + 0x231a, 0x231b, + 0x23e9, 0x23ec, + 0x23f0, 0x23f0, + 0x23f3, 0x23f3, + 0x25fd, 0x25fe, + 0x2614, 0x2615, + 0x2648, 0x2653, + 0x267f, 0x267f, + 0x2693, 0x2693, + 0x26a1, 0x26a1, + 0x26aa, 0x26ab, + 0x26bd, 0x26be, + 0x26c4, 0x26c5, + 0x26ce, 0x26ce, + 0x26d4, 0x26d4, + 0x26ea, 0x26ea, + 0x26f2, 0x26f3, + 0x26f5, 0x26f5, + 0x26fa, 0x26fa, + 0x26fd, 0x26fd, + 0x2705, 0x2705, + 0x270a, 0x270b, + 0x2728, 0x2728, + 0x274c, 0x274c, + 0x274e, 0x274e, + 0x2753, 0x2755, + 0x2757, 0x2757, + 0x2795, 0x2797, + 0x27b0, 0x27b0, + 0x27bf, 0x27bf, + 0x2b1b, 0x2b1c, + 0x2b50, 0x2b50, + 0x2b55, 0x2b55, + 0x1f004, 0x1f004, + 0x1f0cf, 0x1f0cf, + 0x1f18e, 0x1f18e, + 0x1f191, 0x1f19a, + 0x1f1e6, 0x1f1ff, + 0x1f201, 0x1f201, + 0x1f21a, 0x1f21a, + 0x1f22f, 0x1f22f, + 0x1f232, 0x1f236, + 0x1f238, 0x1f23a, + 0x1f250, 0x1f251, + 0x1f300, 0x1f320, + 0x1f32d, 0x1f335, + 0x1f337, 0x1f37c, + 0x1f37e, 0x1f393, + 0x1f3a0, 0x1f3ca, + 0x1f3cf, 0x1f3d3, + 0x1f3e0, 0x1f3f0, + 0x1f3f4, 0x1f3f4, + 0x1f3f8, 0x1f43e, + 0x1f440, 0x1f440, + 0x1f442, 0x1f4fc, + 0x1f4ff, 0x1f53d, + 0x1f54b, 0x1f54e, + 0x1f550, 0x1f567, + 0x1f57a, 0x1f57a, + 0x1f595, 0x1f596, + 0x1f5a4, 0x1f5a4, + 0x1f5fb, 0x1f64f, + 0x1f680, 0x1f6c5, + 0x1f6cc, 0x1f6cc, + 0x1f6d0, 0x1f6d2, + 0x1f6eb, 0x1f6ec, + 0x1f6f4, 0x1f6f8, + 0x1f910, 0x1f93a, + 0x1f93c, 0x1f93e, + 0x1f940, 0x1f945, + 0x1f947, 0x1f94c, + 0x1f950, 0x1f96b, + 0x1f980, 0x1f997, + 0x1f9c0, 0x1f9c0, + 0x1f9d0, 0x1f9e6, +}; /* CR_Emoji_Presentation */ + +/* 'Emoji_Modifier': Emoji */ +static const OnigCodePoint CR_Emoji_Modifier[] = { + 1, + 0x1f3fb, 0x1f3ff, +}; /* CR_Emoji_Modifier */ + +/* 'Emoji_Modifier_Base': Emoji */ +static const OnigCodePoint CR_Emoji_Modifier_Base[] = { + 32, + 0x261d, 0x261d, + 0x26f9, 0x26f9, + 0x270a, 0x270d, + 0x1f385, 0x1f385, + 0x1f3c2, 0x1f3c4, + 0x1f3c7, 0x1f3c7, + 0x1f3ca, 0x1f3cc, + 0x1f442, 0x1f443, + 0x1f446, 0x1f450, + 0x1f466, 0x1f469, + 0x1f46e, 0x1f46e, + 0x1f470, 0x1f478, + 0x1f47c, 0x1f47c, + 0x1f481, 0x1f483, + 0x1f485, 0x1f487, + 0x1f4aa, 0x1f4aa, + 0x1f574, 0x1f575, + 0x1f57a, 0x1f57a, + 0x1f590, 0x1f590, + 0x1f595, 0x1f596, + 0x1f645, 0x1f647, + 0x1f64b, 0x1f64f, + 0x1f6a3, 0x1f6a3, + 0x1f6b4, 0x1f6b6, + 0x1f6c0, 0x1f6c0, + 0x1f6cc, 0x1f6cc, + 0x1f918, 0x1f91c, + 0x1f91e, 0x1f91f, + 0x1f926, 0x1f926, + 0x1f930, 0x1f939, + 0x1f93d, 0x1f93e, + 0x1f9d1, 0x1f9dd, +}; /* CR_Emoji_Modifier_Base */ + +/* 'Emoji_Component': Emoji */ +static const OnigCodePoint CR_Emoji_Component[] = { + 5, + 0x0023, 0x0023, + 0x002a, 0x002a, + 0x0030, 0x0039, + 0x1f1e6, 0x1f1ff, + 0x1f3fb, 0x1f3ff, +}; /* CR_Emoji_Component */ + /* 'Unknown': Script */ static const OnigCodePoint CR_Unknown[] = { 647, @@ -32146,10 +32429,7 @@ static const OnigCodePoint CR_Grapheme_Cluster_Break_E_Base[] = { }; /* CR_Grapheme_Cluster_Break_E_Base */ /* 'Grapheme_Cluster_Break_E_Modifier': Grapheme_Cluster_Break=E_Modifier */ -static const OnigCodePoint CR_Grapheme_Cluster_Break_E_Modifier[] = { - 1, - 0x1f3fb, 0x1f3ff, -}; /* CR_Grapheme_Cluster_Break_E_Modifier */ +#define CR_Grapheme_Cluster_Break_E_Modifier CR_Emoji_Modifier /* 'Grapheme_Cluster_Break_ZWJ': Grapheme_Cluster_Break=ZWJ */ static const OnigCodePoint CR_Grapheme_Cluster_Break_ZWJ[] = { @@ -34143,6 +34423,11 @@ static const OnigCodePoint* const CodeRanges[] = { CR_Pattern_Syntax, CR_Prepended_Concatenation_Mark, CR_Regional_Indicator, + CR_Emoji, + CR_Emoji_Presentation, + CR_Emoji_Modifier, + CR_Emoji_Modifier_Base, + CR_Emoji_Component, CR_Unknown, #ifdef USE_UNICODE_AGE_PROPERTIES CR_Age_1_1, @@ -34482,9 +34767,9 @@ static const struct uniname2ctype_struct *uniname2ctype_p(const char *, unsigned /* maximum key range = 15, duplicates = 0 */ #else /* USE_UNICODE_PROPERTIES */ #ifndef USE_UNICODE_AGE_PROPERTIES -#define TOTAL_KEYWORDS 770 +#define TOTAL_KEYWORDS 775 #else /* USE_UNICODE_AGE_PROPERTIES */ -#define TOTAL_KEYWORDS 789 +#define TOTAL_KEYWORDS 794 #endif /* USE_UNICODE_AGE_PROPERTIES */ #define MIN_WORD_LENGTH 1 #define MAX_WORD_LENGTH 44 @@ -34501,9 +34786,7 @@ inline #endif #endif static unsigned int -uniname2ctype_hash (str, len) - register const char *str; - register unsigned int len; +uniname2ctype_hash (register const char *str, register unsigned int len) { #ifndef USE_UNICODE_PROPERTIES static const unsigned char asso_values[] = @@ -34549,7 +34832,7 @@ uniname2ctype_hash (str, len) #ifndef USE_UNICODE_PROPERTIES return len + asso_values[(unsigned char)str[2]] + asso_values[(unsigned char)str[0]]; #else /* USE_UNICODE_PROPERTIES */ - register int hval = len; + register unsigned int hval = len; switch (hval) { @@ -34992,6 +35275,7 @@ struct uniname2ctype_pool_t char uniname2ctype_pool_str1355[sizeof("intibetan")]; char uniname2ctype_pool_str1364[sizeof("inlepcha")]; char uniname2ctype_pool_str1365[sizeof("inbatak")]; + char uniname2ctype_pool_str1367[sizeof("emoji")]; char uniname2ctype_pool_str1368[sizeof("osma")]; char uniname2ctype_pool_str1369[sizeof("bhks")]; char uniname2ctype_pool_str1372[sizeof("inmongolian")]; @@ -35058,6 +35342,7 @@ struct uniname2ctype_pool_t char uniname2ctype_pool_str1470[sizeof("number")]; char uniname2ctype_pool_str1471[sizeof("insundanesesupplement")]; char uniname2ctype_pool_str1485[sizeof("incopticepactnumbers")]; + char uniname2ctype_pool_str1490[sizeof("emojimodifier")]; char uniname2ctype_pool_str1496[sizeof("zp")]; char uniname2ctype_pool_str1497[sizeof("asciihexdigit")]; char uniname2ctype_pool_str1500[sizeof("indevanagari")]; @@ -35094,6 +35379,7 @@ struct uniname2ctype_pool_t char uniname2ctype_pool_str1631[sizeof("enclosingmark")]; char uniname2ctype_pool_str1632[sizeof("java")]; char uniname2ctype_pool_str1636[sizeof("tale")]; + char uniname2ctype_pool_str1638[sizeof("emojimodifierbase")]; char uniname2ctype_pool_str1645[sizeof("inphoneticextensions")]; char uniname2ctype_pool_str1647[sizeof("lowercase")]; char uniname2ctype_pool_str1648[sizeof("inverticalforms")]; @@ -35122,6 +35408,7 @@ struct uniname2ctype_pool_t char uniname2ctype_pool_str1736[sizeof("graph")]; char uniname2ctype_pool_str1737[sizeof("olchiki")]; char uniname2ctype_pool_str1741[sizeof("inphoneticextensionssupplement")]; + char uniname2ctype_pool_str1743[sizeof("emojicomponent")]; char uniname2ctype_pool_str1746[sizeof("inunifiedcanadianaboriginalsyllabicsextended")]; char uniname2ctype_pool_str1748[sizeof("ingeometricshapes")]; char uniname2ctype_pool_str1752[sizeof("invariationselectorssupplement")]; @@ -35263,6 +35550,7 @@ struct uniname2ctype_pool_t char uniname2ctype_pool_str2372[sizeof("syriac")]; char uniname2ctype_pool_str2380[sizeof("ethi")]; char uniname2ctype_pool_str2386[sizeof("otheralphabetic")]; + char uniname2ctype_pool_str2393[sizeof("emojipresentation")]; char uniname2ctype_pool_str2399[sizeof("inarabicmathematicalalphabeticsymbols")]; char uniname2ctype_pool_str2408[sizeof("tang")]; char uniname2ctype_pool_str2409[sizeof("buhid")]; @@ -35814,6 +36102,7 @@ static const struct uniname2ctype_pool_t uniname2ctype_pool_contents = "intibetan", "inlepcha", "inbatak", + "emoji", "osma", "bhks", "inmongolian", @@ -35880,6 +36169,7 @@ static const struct uniname2ctype_pool_t uniname2ctype_pool_contents = "number", "insundanesesupplement", "incopticepactnumbers", + "emojimodifier", "zp", "asciihexdigit", "indevanagari", @@ -35920,6 +36210,7 @@ static const struct uniname2ctype_pool_t uniname2ctype_pool_contents = "enclosingmark", "java", "tale", + "emojimodifierbase", "inphoneticextensions", "lowercase", "inverticalforms", @@ -35950,6 +36241,7 @@ static const struct uniname2ctype_pool_t uniname2ctype_pool_contents = #ifdef USE_UNICODE_PROPERTIES "olchiki", "inphoneticextensionssupplement", + "emojicomponent", "inunifiedcanadianaboriginalsyllabicsextended", "ingeometricshapes", "invariationselectorssupplement", @@ -36091,6 +36383,7 @@ static const struct uniname2ctype_pool_t uniname2ctype_pool_contents = "syriac", "ethi", "otheralphabetic", + "emojipresentation", "inarabicmathematicalalphabeticsymbols", "tang", "buhid", @@ -36241,16 +36534,8 @@ static const struct uniname2ctype_pool_t uniname2ctype_pool_contents = #endif /* USE_UNICODE_PROPERTIES */ }; #define uniname2ctype_pool ((const char *) &uniname2ctype_pool_contents) -#ifdef __GNUC__ -__inline -#if defined __GNUC_STDC_INLINE__ || defined __GNUC_GNU_INLINE__ -__attribute__ ((__gnu_inline__)) -#endif -#endif const struct uniname2ctype_struct * -uniname2ctype_p (str, len) - register const char *str; - register unsigned int len; +uniname2ctype_p (register const char *str, register unsigned int len) { static const struct uniname2ctype_struct wordlist[] = { @@ -36284,11 +36569,11 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str47), 44}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str51), 315}, + {uniname2ctype_offset(str51), 320}, {uniname2ctype_offset(str52), 181}, {-1}, - {uniname2ctype_offset(str54), 303}, - {uniname2ctype_offset(str55), 250}, + {uniname2ctype_offset(str54), 308}, + {uniname2ctype_offset(str55), 255}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str59), 42}, {uniname2ctype_offset(str60), 149}, @@ -36315,33 +36600,33 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str93), 212}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str97), 479}, + {uniname2ctype_offset(str97), 484}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str105), 170}, - {uniname2ctype_offset(str106), 482}, + {uniname2ctype_offset(str106), 487}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str111), 18}, {-1}, - {uniname2ctype_offset(str113), 297}, + {uniname2ctype_offset(str113), 302}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str117), 75}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str122), 465}, - {uniname2ctype_offset(str123), 521}, + {uniname2ctype_offset(str122), 470}, + {uniname2ctype_offset(str123), 526}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str129), 524}, + {uniname2ctype_offset(str129), 529}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str137), 169}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str142), 321}, + {uniname2ctype_offset(str142), 326}, {uniname2ctype_offset(str143), 27}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str150), 248}, {-1}, {-1}, - {uniname2ctype_offset(str153), 525}, - {uniname2ctype_offset(str154), 493}, + {uniname2ctype_offset(str153), 530}, + {uniname2ctype_offset(str154), 498}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str160), 334}, + {uniname2ctype_offset(str160), 339}, {uniname2ctype_offset(str161), 194}, {uniname2ctype_offset(str162), 40}, {-1}, @@ -36349,15 +36634,15 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str167), 31}, {uniname2ctype_offset(str168), 171}, - {uniname2ctype_offset(str169), 329}, - {uniname2ctype_offset(str170), 456}, + {uniname2ctype_offset(str169), 334}, + {uniname2ctype_offset(str170), 461}, {-1}, - {uniname2ctype_offset(str172), 516}, + {uniname2ctype_offset(str172), 521}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str184), 507}, + {uniname2ctype_offset(str184), 512}, {-1}, - {uniname2ctype_offset(str186), 326}, + {uniname2ctype_offset(str186), 331}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str191), 115}, {-1}, @@ -36366,69 +36651,69 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str205), 25}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str210), 515}, + {uniname2ctype_offset(str210), 520}, {-1}, {-1}, {uniname2ctype_offset(str213), 19}, {-1}, {uniname2ctype_offset(str215), 79}, - {uniname2ctype_offset(str216), 483}, + {uniname2ctype_offset(str216), 488}, {uniname2ctype_offset(str217), 161}, {-1}, {uniname2ctype_offset(str219), 32}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str223), 79}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str227), 291}, + {uniname2ctype_offset(str227), 296}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str238), 513}, + {uniname2ctype_offset(str238), 518}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str243), 40}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str250), 495}, + {uniname2ctype_offset(str250), 500}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str255), 53}, {uniname2ctype_offset(str256), 212}, {-1}, {-1}, - {uniname2ctype_offset(str259), 363}, + {uniname2ctype_offset(str259), 368}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str268), 249}, {-1}, - {uniname2ctype_offset(str270), 424}, + {uniname2ctype_offset(str270), 429}, {uniname2ctype_offset(str271), 76}, - {uniname2ctype_offset(str272), 295}, + {uniname2ctype_offset(str272), 300}, {uniname2ctype_offset(str273), 76}, {-1}, {-1}, - {uniname2ctype_offset(str276), 509}, + {uniname2ctype_offset(str276), 514}, {-1}, {-1}, - {uniname2ctype_offset(str279), 304}, + {uniname2ctype_offset(str279), 309}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str283), 248}, {-1}, - {uniname2ctype_offset(str285), 514}, + {uniname2ctype_offset(str285), 519}, {-1}, {-1}, {uniname2ctype_offset(str288), 221}, {uniname2ctype_offset(str289), 129}, {-1}, {uniname2ctype_offset(str291), 160}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str295), 318}, + {uniname2ctype_offset(str295), 323}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str300), 163}, - {uniname2ctype_offset(str301), 302}, + {uniname2ctype_offset(str301), 307}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str309), 491}, + {uniname2ctype_offset(str309), 496}, {-1}, {uniname2ctype_offset(str311), 44}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str315), 394}, + {uniname2ctype_offset(str315), 399}, {-1}, - {uniname2ctype_offset(str317), 340}, + {uniname2ctype_offset(str317), 345}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str343), 300}, - {uniname2ctype_offset(str344), 506}, + {uniname2ctype_offset(str343), 305}, + {uniname2ctype_offset(str344), 511}, {uniname2ctype_offset(str345), 163}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -36451,16 +36736,16 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str383), 127}, {uniname2ctype_offset(str384), 70}, {-1}, - {uniname2ctype_offset(str386), 425}, + {uniname2ctype_offset(str386), 430}, {uniname2ctype_offset(str387), 14}, {-1}, {uniname2ctype_offset(str389), 46}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str393), 529}, + {uniname2ctype_offset(str393), 534}, {-1}, {-1}, - {uniname2ctype_offset(str396), 431}, + {uniname2ctype_offset(str396), 436}, {-1}, {-1}, - {uniname2ctype_offset(str399), 489}, + {uniname2ctype_offset(str399), 494}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str408), 37}, {-1}, {-1}, {-1}, @@ -36469,34 +36754,34 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {uniname2ctype_offset(str425), 91}, {-1}, {-1}, - {uniname2ctype_offset(str428), 481}, + {uniname2ctype_offset(str428), 486}, {-1}, - {uniname2ctype_offset(str430), 391}, + {uniname2ctype_offset(str430), 396}, {-1}, - {uniname2ctype_offset(str432), 504}, + {uniname2ctype_offset(str432), 509}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str440), 444}, + {uniname2ctype_offset(str440), 449}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str445), 438}, + {uniname2ctype_offset(str445), 443}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str453), 70}, {-1}, - {uniname2ctype_offset(str455), 541}, + {uniname2ctype_offset(str455), 546}, {-1}, {-1}, {uniname2ctype_offset(str458), 168}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str480), 458}, + {uniname2ctype_offset(str480), 463}, {-1}, - {uniname2ctype_offset(str482), 374}, + {uniname2ctype_offset(str482), 379}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str486), 448}, - {uniname2ctype_offset(str487), 550}, + {uniname2ctype_offset(str486), 453}, + {uniname2ctype_offset(str487), 555}, {uniname2ctype_offset(str488), 246}, {uniname2ctype_offset(str489), 113}, {-1}, {-1}, - {uniname2ctype_offset(str492), 552}, + {uniname2ctype_offset(str492), 557}, {uniname2ctype_offset(str493), 24}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -36510,15 +36795,15 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str526), 69}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str530), 415}, - {uniname2ctype_offset(str531), 305}, + {uniname2ctype_offset(str530), 420}, + {uniname2ctype_offset(str531), 310}, {-1}, - {uniname2ctype_offset(str533), 469}, + {uniname2ctype_offset(str533), 474}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str541), 562}, + {uniname2ctype_offset(str541), 567}, {-1}, {-1}, - {uniname2ctype_offset(str544), 319}, - {uniname2ctype_offset(str545), 345}, + {uniname2ctype_offset(str544), 324}, + {uniname2ctype_offset(str545), 350}, {-1}, {uniname2ctype_offset(str547), 106}, {-1}, {-1}, {-1}, {-1}, @@ -36526,7 +36811,7 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str554), 68}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str561), 453}, + {uniname2ctype_offset(str561), 458}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str573), 146}, @@ -36535,15 +36820,15 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str576), 72}, {uniname2ctype_offset(str577), 19}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str586), 528}, + {uniname2ctype_offset(str586), 533}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str591), 367}, + {uniname2ctype_offset(str591), 372}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str596), 542}, + {uniname2ctype_offset(str596), 547}, {-1}, {uniname2ctype_offset(str598), 51}, {-1}, {-1}, - {uniname2ctype_offset(str601), 480}, + {uniname2ctype_offset(str601), 485}, {uniname2ctype_offset(str602), 49}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str606), 7}, @@ -36552,26 +36837,26 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str609), 192}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str618), 167}, - {uniname2ctype_offset(str619), 545}, - {uniname2ctype_offset(str620), 342}, + {uniname2ctype_offset(str619), 550}, + {uniname2ctype_offset(str620), 347}, {-1}, - {uniname2ctype_offset(str622), 455}, + {uniname2ctype_offset(str622), 460}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str629), 175}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str636), 167}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str642), 412}, + {uniname2ctype_offset(str642), 417}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str657), 246}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str665), 316}, + {uniname2ctype_offset(str665), 321}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str670), 399}, + {uniname2ctype_offset(str670), 404}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str681), 366}, + {uniname2ctype_offset(str681), 371}, {-1}, {-1}, {uniname2ctype_offset(str684), 156}, {-1}, {-1}, {-1}, {-1}, @@ -36579,7 +36864,7 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {uniname2ctype_offset(str693), 156}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str699), 534}, + {uniname2ctype_offset(str699), 539}, {uniname2ctype_offset(str700), 67}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str705), 41}, @@ -36589,27 +36874,27 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str715), 193}, {uniname2ctype_offset(str716), 218}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str725), 450}, + {uniname2ctype_offset(str725), 455}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str730), 68}, {-1}, - {uniname2ctype_offset(str732), 535}, + {uniname2ctype_offset(str732), 540}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str737), 487}, + {uniname2ctype_offset(str737), 492}, {uniname2ctype_offset(str738), 13}, - {uniname2ctype_offset(str739), 461}, + {uniname2ctype_offset(str739), 466}, {-1}, {uniname2ctype_offset(str741), 47}, - {uniname2ctype_offset(str742), 372}, + {uniname2ctype_offset(str742), 377}, {uniname2ctype_offset(str743), 35}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str748), 157}, {-1}, {-1}, - {uniname2ctype_offset(str751), 380}, + {uniname2ctype_offset(str751), 385}, {-1}, - {uniname2ctype_offset(str753), 414}, + {uniname2ctype_offset(str753), 419}, {-1}, {-1}, - {uniname2ctype_offset(str756), 549}, + {uniname2ctype_offset(str756), 554}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str770), 48}, @@ -36624,39 +36909,39 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str785), 71}, {-1}, {uniname2ctype_offset(str787), 235}, - {uniname2ctype_offset(str788), 338}, + {uniname2ctype_offset(str788), 343}, {-1}, {-1}, {uniname2ctype_offset(str791), 140}, {uniname2ctype_offset(str792), 140}, {uniname2ctype_offset(str793), 228}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str797), 475}, + {uniname2ctype_offset(str797), 480}, {-1}, - {uniname2ctype_offset(str799), 289}, + {uniname2ctype_offset(str799), 294}, {-1}, {-1}, - {uniname2ctype_offset(str802), 463}, + {uniname2ctype_offset(str802), 468}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str807), 429}, + {uniname2ctype_offset(str807), 434}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str812), 468}, + {uniname2ctype_offset(str812), 473}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str819), 417}, + {uniname2ctype_offset(str819), 422}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str823), 427}, + {uniname2ctype_offset(str823), 432}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str834), 67}, {uniname2ctype_offset(str835), 136}, {-1}, - {uniname2ctype_offset(str837), 292}, + {uniname2ctype_offset(str837), 297}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str848), 85}, - {uniname2ctype_offset(str849), 313}, + {uniname2ctype_offset(str849), 318}, {-1}, - {uniname2ctype_offset(str851), 498}, + {uniname2ctype_offset(str851), 503}, {uniname2ctype_offset(str852), 245}, - {uniname2ctype_offset(str853), 476}, + {uniname2ctype_offset(str853), 481}, {-1}, {uniname2ctype_offset(str855), 136}, {uniname2ctype_offset(str856), 142}, @@ -36676,10 +36961,10 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {uniname2ctype_offset(str881), 69}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str885), 454}, + {uniname2ctype_offset(str885), 459}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str897), 452}, + {uniname2ctype_offset(str897), 457}, {-1}, {uniname2ctype_offset(str899), 102}, {-1}, {-1}, {-1}, @@ -36693,17 +36978,17 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str911), 138}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str918), 208}, - {uniname2ctype_offset(str919), 393}, + {uniname2ctype_offset(str919), 398}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str926), 517}, + {uniname2ctype_offset(str926), 522}, {-1}, {uniname2ctype_offset(str928), 73}, {-1}, {uniname2ctype_offset(str930), 151}, - {uniname2ctype_offset(str931), 471}, + {uniname2ctype_offset(str931), 476}, {uniname2ctype_offset(str932), 166}, {-1}, {-1}, - {uniname2ctype_offset(str935), 503}, + {uniname2ctype_offset(str935), 508}, {-1}, {uniname2ctype_offset(str937), 228}, {-1}, {-1}, {-1}, {-1}, @@ -36711,15 +36996,15 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str944), 74}, {-1}, {-1}, - {uniname2ctype_offset(str947), 499}, - {uniname2ctype_offset(str948), 494}, + {uniname2ctype_offset(str947), 504}, + {uniname2ctype_offset(str948), 499}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str956), 183}, - {uniname2ctype_offset(str957), 328}, + {uniname2ctype_offset(str957), 333}, {-1}, {-1}, {uniname2ctype_offset(str960), 135}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str966), 537}, + {uniname2ctype_offset(str966), 542}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str975), 186}, {uniname2ctype_offset(str976), 129}, @@ -36727,24 +37012,24 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str980), 244}, {-1}, {-1}, - {uniname2ctype_offset(str983), 324}, + {uniname2ctype_offset(str983), 329}, {-1}, {uniname2ctype_offset(str985), 26}, {-1}, {-1}, - {uniname2ctype_offset(str988), 348}, + {uniname2ctype_offset(str988), 353}, {-1}, {-1}, - {uniname2ctype_offset(str991), 382}, + {uniname2ctype_offset(str991), 387}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str996), 53}, {-1}, {uniname2ctype_offset(str998), 206}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1016), 296}, + {uniname2ctype_offset(str1016), 301}, {-1}, {-1}, - {uniname2ctype_offset(str1019), 387}, + {uniname2ctype_offset(str1019), 392}, {-1}, {-1}, - {uniname2ctype_offset(str1022), 430}, + {uniname2ctype_offset(str1022), 435}, {uniname2ctype_offset(str1023), 36}, {-1}, {uniname2ctype_offset(str1025), 105}, @@ -36756,8 +37041,8 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str1034), 104}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1048), 536}, - {uniname2ctype_offset(str1049), 317}, + {uniname2ctype_offset(str1048), 541}, + {uniname2ctype_offset(str1049), 322}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1053), 126}, {-1}, {-1}, @@ -36771,35 +37056,35 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str1069), 206}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1076), 484}, - {uniname2ctype_offset(str1077), 485}, + {uniname2ctype_offset(str1076), 489}, + {uniname2ctype_offset(str1077), 490}, {uniname2ctype_offset(str1078), 93}, {-1}, {uniname2ctype_offset(str1080), 215}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1086), 403}, + {uniname2ctype_offset(str1086), 408}, {-1}, {uniname2ctype_offset(str1088), 203}, {uniname2ctype_offset(str1089), 221}, - {uniname2ctype_offset(str1090), 560}, + {uniname2ctype_offset(str1090), 565}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1094), 473}, + {uniname2ctype_offset(str1094), 478}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1098), 512}, + {uniname2ctype_offset(str1098), 517}, {-1}, - {uniname2ctype_offset(str1100), 467}, + {uniname2ctype_offset(str1100), 472}, {-1}, {-1}, - {uniname2ctype_offset(str1103), 397}, + {uniname2ctype_offset(str1103), 402}, {uniname2ctype_offset(str1104), 164}, {-1}, {-1}, {uniname2ctype_offset(str1107), 109}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1114), 122}, - {uniname2ctype_offset(str1115), 349}, + {uniname2ctype_offset(str1115), 354}, {uniname2ctype_offset(str1116), 131}, {-1}, {-1}, {uniname2ctype_offset(str1119), 135}, - {uniname2ctype_offset(str1120), 462}, + {uniname2ctype_offset(str1120), 467}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1124), 113}, {-1}, @@ -36812,12 +37097,12 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str1136), 120}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1147), 365}, + {uniname2ctype_offset(str1147), 370}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1152), 110}, {-1}, {-1}, {uniname2ctype_offset(str1155), 110}, - {uniname2ctype_offset(str1156), 409}, + {uniname2ctype_offset(str1156), 414}, {uniname2ctype_offset(str1157), 141}, {uniname2ctype_offset(str1158), 121}, {-1}, @@ -36825,16 +37110,16 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str1161), 181}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1167), 117}, - {uniname2ctype_offset(str1168), 474}, + {uniname2ctype_offset(str1168), 479}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1172), 88}, - {uniname2ctype_offset(str1173), 502}, + {uniname2ctype_offset(str1173), 507}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1178), 86}, {uniname2ctype_offset(str1179), 166}, {uniname2ctype_offset(str1180), 145}, {-1}, - {uniname2ctype_offset(str1182), 558}, + {uniname2ctype_offset(str1182), 563}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1186), 117}, {-1}, {-1}, @@ -36843,26 +37128,26 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str1192), 176}, {uniname2ctype_offset(str1193), 107}, {-1}, - {uniname2ctype_offset(str1195), 299}, + {uniname2ctype_offset(str1195), 304}, {uniname2ctype_offset(str1196), 109}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1201), 236}, {uniname2ctype_offset(str1202), 86}, - {uniname2ctype_offset(str1203), 497}, + {uniname2ctype_offset(str1203), 502}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1207), 81}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1211), 120}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1218), 460}, + {uniname2ctype_offset(str1218), 465}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1222), 168}, {-1}, {-1}, - {uniname2ctype_offset(str1225), 488}, + {uniname2ctype_offset(str1225), 493}, {-1}, - {uniname2ctype_offset(str1227), 532}, + {uniname2ctype_offset(str1227), 537}, {-1}, - {uniname2ctype_offset(str1229), 383}, + {uniname2ctype_offset(str1229), 388}, {-1}, {-1}, {uniname2ctype_offset(str1232), 240}, {-1}, {-1}, {-1}, {-1}, @@ -36870,11 +37155,11 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1243), 115}, {-1}, - {uniname2ctype_offset(str1245), 405}, + {uniname2ctype_offset(str1245), 410}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1249), 222}, {-1}, {-1}, - {uniname2ctype_offset(str1252), 392}, + {uniname2ctype_offset(str1252), 397}, {uniname2ctype_offset(str1253), 72}, {uniname2ctype_offset(str1254), 114}, {-1}, {-1}, @@ -36882,58 +37167,59 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str1259), 137}, {-1}, {-1}, - {uniname2ctype_offset(str1262), 306}, + {uniname2ctype_offset(str1262), 311}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1267), 287}, + {uniname2ctype_offset(str1267), 292}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1275), 114}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1281), 548}, + {uniname2ctype_offset(str1281), 553}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1285), 217}, {-1}, {-1}, {uniname2ctype_offset(str1288), 42}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1294), 355}, + {uniname2ctype_offset(str1294), 360}, {-1}, - {uniname2ctype_offset(str1296), 440}, + {uniname2ctype_offset(str1296), 445}, {-1}, {uniname2ctype_offset(str1298), 73}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1306), 106}, {uniname2ctype_offset(str1307), 201}, - {uniname2ctype_offset(str1308), 368}, - {uniname2ctype_offset(str1309), 500}, + {uniname2ctype_offset(str1308), 373}, + {uniname2ctype_offset(str1309), 505}, {uniname2ctype_offset(str1310), 98}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1317), 210}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1325), 466}, - {uniname2ctype_offset(str1326), 357}, + {uniname2ctype_offset(str1325), 471}, + {uniname2ctype_offset(str1326), 362}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1331), 145}, {-1}, {-1}, - {uniname2ctype_offset(str1334), 464}, + {uniname2ctype_offset(str1334), 469}, {uniname2ctype_offset(str1335), 9}, {uniname2ctype_offset(str1336), 203}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1340), 472}, - {uniname2ctype_offset(str1341), 519}, + {uniname2ctype_offset(str1340), 477}, + {uniname2ctype_offset(str1341), 524}, {-1}, {-1}, {uniname2ctype_offset(str1344), 32}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1348), 1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1354), 322}, - {uniname2ctype_offset(str1355), 320}, + {uniname2ctype_offset(str1354), 327}, + {uniname2ctype_offset(str1355), 325}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1364), 347}, - {uniname2ctype_offset(str1365), 346}, - {-1}, {-1}, + {uniname2ctype_offset(str1364), 352}, + {uniname2ctype_offset(str1365), 351}, + {-1}, + {uniname2ctype_offset(str1367), 250}, {uniname2ctype_offset(str1368), 125}, {uniname2ctype_offset(str1369), 207}, {-1}, {-1}, - {uniname2ctype_offset(str1372), 335}, + {uniname2ctype_offset(str1372), 340}, {uniname2ctype_offset(str1373), 245}, {uniname2ctype_offset(str1374), 127}, {-1}, {-1}, {-1}, {-1}, @@ -36943,78 +37229,78 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1388), 138}, {-1}, - {uniname2ctype_offset(str1390), 370}, + {uniname2ctype_offset(str1390), 375}, #ifndef USE_UNICODE_AGE_PROPERTIES {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, #else /* USE_UNICODE_AGE_PROPERTIES */ - {uniname2ctype_offset(str1391), 251}, - {uniname2ctype_offset(str1392), 258}, - {uniname2ctype_offset(str1393), 257}, - {uniname2ctype_offset(str1394), 269}, - {uniname2ctype_offset(str1395), 253}, - {uniname2ctype_offset(str1396), 252}, - {uniname2ctype_offset(str1397), 263}, - {uniname2ctype_offset(str1398), 262}, - {uniname2ctype_offset(str1399), 268}, - {uniname2ctype_offset(str1400), 267}, - {uniname2ctype_offset(str1401), 264}, - {uniname2ctype_offset(str1402), 266}, + {uniname2ctype_offset(str1391), 256}, + {uniname2ctype_offset(str1392), 263}, + {uniname2ctype_offset(str1393), 262}, + {uniname2ctype_offset(str1394), 274}, + {uniname2ctype_offset(str1395), 258}, + {uniname2ctype_offset(str1396), 257}, + {uniname2ctype_offset(str1397), 268}, + {uniname2ctype_offset(str1398), 267}, + {uniname2ctype_offset(str1399), 273}, + {uniname2ctype_offset(str1400), 272}, + {uniname2ctype_offset(str1401), 269}, + {uniname2ctype_offset(str1402), 271}, #endif /* USE_UNICODE_AGE_PROPERTIES */ - {uniname2ctype_offset(str1403), 547}, - {uniname2ctype_offset(str1404), 362}, + {uniname2ctype_offset(str1403), 552}, + {uniname2ctype_offset(str1404), 367}, #ifndef USE_UNICODE_AGE_PROPERTIES {-1}, {-1}, #else /* USE_UNICODE_AGE_PROPERTIES */ - {uniname2ctype_offset(str1405), 260}, - {uniname2ctype_offset(str1406), 259}, + {uniname2ctype_offset(str1405), 265}, + {uniname2ctype_offset(str1406), 264}, #endif /* USE_UNICODE_AGE_PROPERTIES */ {uniname2ctype_offset(str1407), 36}, {uniname2ctype_offset(str1408), 52}, #ifndef USE_UNICODE_AGE_PROPERTIES {-1}, {-1}, #else /* USE_UNICODE_AGE_PROPERTIES */ - {uniname2ctype_offset(str1409), 261}, + {uniname2ctype_offset(str1409), 266}, {-1}, #endif /* USE_UNICODE_AGE_PROPERTIES */ - {uniname2ctype_offset(str1411), 310}, + {uniname2ctype_offset(str1411), 315}, {-1}, - {uniname2ctype_offset(str1413), 559}, + {uniname2ctype_offset(str1413), 564}, {uniname2ctype_offset(str1414), 153}, {-1}, {uniname2ctype_offset(str1416), 93}, #ifndef USE_UNICODE_AGE_PROPERTIES {-1}, {-1}, {-1}, #else /* USE_UNICODE_AGE_PROPERTIES */ - {uniname2ctype_offset(str1417), 255}, - {uniname2ctype_offset(str1418), 254}, + {uniname2ctype_offset(str1417), 260}, + {uniname2ctype_offset(str1418), 259}, {-1}, #endif /* USE_UNICODE_AGE_PROPERTIES */ {uniname2ctype_offset(str1420), 202}, #ifndef USE_UNICODE_AGE_PROPERTIES {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, #else /* USE_UNICODE_AGE_PROPERTIES */ - {uniname2ctype_offset(str1421), 256}, + {uniname2ctype_offset(str1421), 261}, {-1}, - {uniname2ctype_offset(str1423), 265}, + {uniname2ctype_offset(str1423), 270}, {-1}, {-1}, {-1}, #endif /* USE_UNICODE_AGE_PROPERTIES */ {uniname2ctype_offset(str1427), 20}, {-1}, {uniname2ctype_offset(str1429), 124}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1434), 496}, + {uniname2ctype_offset(str1434), 501}, {uniname2ctype_offset(str1435), 137}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1439), 423}, + {uniname2ctype_offset(str1439), 428}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1444), 39}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1449), 307}, + {uniname2ctype_offset(str1449), 312}, {uniname2ctype_offset(str1450), 202}, - {uniname2ctype_offset(str1451), 331}, + {uniname2ctype_offset(str1451), 336}, {-1}, - {uniname2ctype_offset(str1453), 413}, + {uniname2ctype_offset(str1453), 418}, {-1}, {uniname2ctype_offset(str1455), 27}, {uniname2ctype_offset(str1456), 238}, @@ -37022,38 +37308,39 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str1458), 225}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1463), 241}, - {uniname2ctype_offset(str1464), 533}, + {uniname2ctype_offset(str1464), 538}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1469), 141}, {uniname2ctype_offset(str1470), 35}, - {uniname2ctype_offset(str1471), 350}, + {uniname2ctype_offset(str1471), 355}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1485), 457}, - {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {-1}, + {uniname2ctype_offset(str1485), 462}, + {-1}, {-1}, {-1}, {-1}, + {uniname2ctype_offset(str1490), 252}, + {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1496), 54}, {uniname2ctype_offset(str1497), 225}, {-1}, {-1}, - {uniname2ctype_offset(str1500), 308}, + {uniname2ctype_offset(str1500), 313}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1508), 501}, + {uniname2ctype_offset(str1508), 506}, {-1}, {uniname2ctype_offset(str1510), 154}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1521), 492}, + {uniname2ctype_offset(str1521), 497}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1535), 378}, - {uniname2ctype_offset(str1536), 376}, + {uniname2ctype_offset(str1535), 383}, + {uniname2ctype_offset(str1536), 381}, {-1}, {-1}, {uniname2ctype_offset(str1539), 34}, - {uniname2ctype_offset(str1540), 294}, + {uniname2ctype_offset(str1540), 299}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1545), 8}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1553), 508}, + {uniname2ctype_offset(str1553), 513}, {-1}, {uniname2ctype_offset(str1555), 243}, {-1}, {-1}, {-1}, {-1}, @@ -37063,26 +37350,26 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1572), 6}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1580), 327}, + {uniname2ctype_offset(str1580), 332}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1585), 234}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1589), 344}, + {uniname2ctype_offset(str1589), 349}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1593), 531}, - {uniname2ctype_offset(str1594), 567}, + {uniname2ctype_offset(str1593), 536}, + {uniname2ctype_offset(str1594), 572}, {uniname2ctype_offset(str1595), 43}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1602), 312}, + {uniname2ctype_offset(str1602), 317}, {-1}, - {uniname2ctype_offset(str1604), 400}, + {uniname2ctype_offset(str1604), 405}, {-1}, - {uniname2ctype_offset(str1606), 419}, + {uniname2ctype_offset(str1606), 424}, {uniname2ctype_offset(str1607), 199}, - {uniname2ctype_offset(str1608), 530}, + {uniname2ctype_offset(str1608), 535}, {uniname2ctype_offset(str1609), 142}, {-1}, - {uniname2ctype_offset(str1611), 477}, + {uniname2ctype_offset(str1611), 482}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1617), 85}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37095,17 +37382,19 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str1632), 159}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1636), 121}, - {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1645), 352}, + {-1}, + {uniname2ctype_offset(str1638), 253}, + {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, + {uniname2ctype_offset(str1645), 357}, {-1}, {uniname2ctype_offset(str1647), 58}, - {uniname2ctype_offset(str1648), 442}, + {uniname2ctype_offset(str1648), 447}, {-1}, {uniname2ctype_offset(str1650), 159}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1654), 244}, - {uniname2ctype_offset(str1655), 356}, - {uniname2ctype_offset(str1656), 441}, + {uniname2ctype_offset(str1655), 361}, + {uniname2ctype_offset(str1656), 446}, {uniname2ctype_offset(str1657), 66}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1663), 148}, @@ -37116,14 +37405,14 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str1671), 215}, {uniname2ctype_offset(str1672), 162}, - {uniname2ctype_offset(str1673), 351}, - {uniname2ctype_offset(str1674), 390}, + {uniname2ctype_offset(str1673), 356}, + {uniname2ctype_offset(str1674), 395}, {-1}, {-1}, - {uniname2ctype_offset(str1677), 426}, + {uniname2ctype_offset(str1677), 431}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1682), 97}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1690), 389}, + {uniname2ctype_offset(str1690), 394}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1695), 97}, {-1}, @@ -37135,21 +37424,23 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str1725), 37}, {-1}, - {uniname2ctype_offset(str1727), 325}, + {uniname2ctype_offset(str1727), 330}, {uniname2ctype_offset(str1728), 143}, {-1}, - {uniname2ctype_offset(str1730), 428}, + {uniname2ctype_offset(str1730), 433}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1736), 5}, {uniname2ctype_offset(str1737), 143}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1741), 353}, - {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1746), 336}, + {uniname2ctype_offset(str1741), 358}, {-1}, - {uniname2ctype_offset(str1748), 371}, + {uniname2ctype_offset(str1743), 254}, + {-1}, {-1}, + {uniname2ctype_offset(str1746), 341}, + {-1}, + {uniname2ctype_offset(str1748), 376}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1752), 564}, + {uniname2ctype_offset(str1752), 569}, {-1}, {uniname2ctype_offset(str1754), 87}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37158,7 +37449,7 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str1766), 87}, {uniname2ctype_offset(str1767), 232}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1772), 339}, + {uniname2ctype_offset(str1772), 344}, {-1}, {uniname2ctype_offset(str1774), 131}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37166,69 +37457,69 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str1783), 174}, {-1}, - {uniname2ctype_offset(str1785), 510}, + {uniname2ctype_offset(str1785), 515}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1789), 411}, + {uniname2ctype_offset(str1789), 416}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1800), 288}, + {uniname2ctype_offset(str1800), 293}, {-1}, {uniname2ctype_offset(str1802), 219}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1806), 373}, + {uniname2ctype_offset(str1806), 378}, {-1}, {uniname2ctype_offset(str1808), 55}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1813), 139}, {uniname2ctype_offset(str1814), 29}, - {uniname2ctype_offset(str1815), 402}, + {uniname2ctype_offset(str1815), 407}, {-1}, {-1}, - {uniname2ctype_offset(str1818), 526}, - {uniname2ctype_offset(str1819), 293}, + {uniname2ctype_offset(str1818), 531}, + {uniname2ctype_offset(str1819), 298}, {-1}, - {uniname2ctype_offset(str1821), 408}, + {uniname2ctype_offset(str1821), 413}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1825), 196}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1829), 360}, + {uniname2ctype_offset(str1829), 365}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1840), 177}, {-1}, {-1}, {uniname2ctype_offset(str1843), 237}, {-1}, {-1}, - {uniname2ctype_offset(str1846), 418}, + {uniname2ctype_offset(str1846), 423}, {uniname2ctype_offset(str1847), 190}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1857), 395}, - {uniname2ctype_offset(str1858), 561}, + {uniname2ctype_offset(str1857), 400}, + {uniname2ctype_offset(str1858), 566}, {-1}, {-1}, - {uniname2ctype_offset(str1861), 478}, + {uniname2ctype_offset(str1861), 483}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1866), 247}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1885), 358}, + {uniname2ctype_offset(str1885), 363}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1894), 150}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1900), 150}, {-1}, {-1}, - {uniname2ctype_offset(str1903), 563}, + {uniname2ctype_offset(str1903), 568}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1909), 314}, - {uniname2ctype_offset(str1910), 385}, + {uniname2ctype_offset(str1909), 319}, + {uniname2ctype_offset(str1910), 390}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1914), 554}, - {uniname2ctype_offset(str1915), 354}, + {uniname2ctype_offset(str1914), 559}, + {uniname2ctype_offset(str1915), 359}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1922), 84}, {-1}, - {uniname2ctype_offset(str1924), 437}, + {uniname2ctype_offset(str1924), 442}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1932), 84}, {uniname2ctype_offset(str1933), 232}, - {uniname2ctype_offset(str1934), 490}, + {uniname2ctype_offset(str1934), 495}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1943), 26}, {uniname2ctype_offset(str1944), 12}, @@ -37237,7 +37528,7 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str1975), 61}, - {uniname2ctype_offset(str1976), 407}, + {uniname2ctype_offset(str1976), 412}, {-1}, {-1}, {uniname2ctype_offset(str1979), 239}, {-1}, @@ -37245,26 +37536,26 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str1983), 83}, {-1}, - {uniname2ctype_offset(str1985), 343}, + {uniname2ctype_offset(str1985), 348}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1989), 538}, + {uniname2ctype_offset(str1989), 543}, {-1}, {uniname2ctype_offset(str1991), 83}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str1995), 551}, + {uniname2ctype_offset(str1995), 556}, {-1}, {uniname2ctype_offset(str1997), 134}, {uniname2ctype_offset(str1998), 21}, - {uniname2ctype_offset(str1999), 375}, + {uniname2ctype_offset(str1999), 380}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2003), 511}, + {uniname2ctype_offset(str2003), 516}, {-1}, {uniname2ctype_offset(str2005), 64}, {-1}, {uniname2ctype_offset(str2007), 198}, {uniname2ctype_offset(str2008), 185}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2013), 445}, + {uniname2ctype_offset(str2013), 450}, {-1}, {uniname2ctype_offset(str2015), 198}, {-1}, @@ -37276,9 +37567,9 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2047), 240}, {-1}, - {uniname2ctype_offset(str2049), 416}, + {uniname2ctype_offset(str2049), 421}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2056), 386}, + {uniname2ctype_offset(str2056), 391}, {-1}, {-1}, {uniname2ctype_offset(str2059), 152}, {-1}, @@ -37286,18 +37577,18 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2068), 191}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2072), 410}, + {uniname2ctype_offset(str2072), 415}, {uniname2ctype_offset(str2073), 183}, {-1}, {-1}, {uniname2ctype_offset(str2076), 118}, {-1}, {-1}, {uniname2ctype_offset(str2079), 191}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2084), 439}, + {uniname2ctype_offset(str2084), 444}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2091), 175}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2095), 555}, + {uniname2ctype_offset(str2095), 560}, {uniname2ctype_offset(str2096), 194}, {uniname2ctype_offset(str2097), 78}, {-1}, {-1}, {-1}, @@ -37306,20 +37597,20 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str2112), 172}, {-1}, - {uniname2ctype_offset(str2114), 301}, + {uniname2ctype_offset(str2114), 306}, {-1}, {-1}, {uniname2ctype_offset(str2117), 195}, {-1}, - {uniname2ctype_offset(str2119), 486}, + {uniname2ctype_offset(str2119), 491}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2134), 92}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2139), 557}, + {uniname2ctype_offset(str2139), 562}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2146), 192}, {-1}, - {uniname2ctype_offset(str2148), 379}, + {uniname2ctype_offset(str2148), 384}, {uniname2ctype_offset(str2149), 92}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2154), 214}, @@ -37328,39 +37619,39 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str2161), 139}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2165), 282}, + {uniname2ctype_offset(str2165), 287}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2172), 286}, + {uniname2ctype_offset(str2172), 291}, {-1}, - {uniname2ctype_offset(str2174), 323}, + {uniname2ctype_offset(str2174), 328}, {-1}, {uniname2ctype_offset(str2176), 128}, {-1}, {-1}, - {uniname2ctype_offset(str2179), 276}, + {uniname2ctype_offset(str2179), 281}, {-1}, {-1}, - {uniname2ctype_offset(str2182), 421}, + {uniname2ctype_offset(str2182), 426}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2187), 48}, {-1}, {uniname2ctype_offset(str2189), 89}, {-1}, {-1}, - {uniname2ctype_offset(str2192), 271}, + {uniname2ctype_offset(str2192), 276}, {uniname2ctype_offset(str2193), 130}, {-1}, {-1}, {uniname2ctype_offset(str2196), 128}, {uniname2ctype_offset(str2197), 90}, - {uniname2ctype_offset(str2198), 384}, - {uniname2ctype_offset(str2199), 283}, + {uniname2ctype_offset(str2198), 389}, + {uniname2ctype_offset(str2199), 288}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2207), 275}, + {uniname2ctype_offset(str2207), 280}, {-1}, {-1}, - {uniname2ctype_offset(str2210), 337}, + {uniname2ctype_offset(str2210), 342}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2224), 401}, + {uniname2ctype_offset(str2224), 406}, {uniname2ctype_offset(str2225), 211}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2229), 364}, + {uniname2ctype_offset(str2229), 369}, {-1}, {-1}, {uniname2ctype_offset(str2232), 209}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37369,27 +37660,27 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str2252), 80}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2260), 332}, + {uniname2ctype_offset(str2260), 337}, {-1}, {-1}, - {uniname2ctype_offset(str2263), 539}, + {uniname2ctype_offset(str2263), 544}, {uniname2ctype_offset(str2264), 82}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2271), 239}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2276), 522}, + {uniname2ctype_offset(str2276), 527}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2286), 243}, {-1}, {-1}, - {uniname2ctype_offset(str2289), 361}, + {uniname2ctype_offset(str2289), 366}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2296), 229}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2300), 18}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2306), 505}, + {uniname2ctype_offset(str2306), 510}, {uniname2ctype_offset(str2307), 38}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2311), 422}, + {uniname2ctype_offset(str2311), 427}, {-1}, {uniname2ctype_offset(str2313), 112}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37398,40 +37689,41 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2349), 446}, - {uniname2ctype_offset(str2350), 436}, + {uniname2ctype_offset(str2349), 451}, + {uniname2ctype_offset(str2350), 441}, {-1}, {-1}, - {uniname2ctype_offset(str2353), 443}, + {uniname2ctype_offset(str2353), 448}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2360), 309}, + {uniname2ctype_offset(str2360), 314}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2364), 65}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2369), 341}, + {uniname2ctype_offset(str2369), 346}, {-1}, {-1}, {uniname2ctype_offset(str2372), 82}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2380), 100}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2386), 226}, - {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2399), 543}, + {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, + {uniname2ctype_offset(str2393), 251}, + {-1}, {-1}, {-1}, {-1}, {-1}, + {uniname2ctype_offset(str2399), 548}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2408), 211}, {uniname2ctype_offset(str2409), 118}, {-1}, {-1}, - {uniname2ctype_offset(str2412), 279}, + {uniname2ctype_offset(str2412), 284}, {uniname2ctype_offset(str2413), 229}, - {uniname2ctype_offset(str2414), 281}, + {uniname2ctype_offset(str2414), 286}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2422), 119}, {uniname2ctype_offset(str2423), 99}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2433), 359}, + {uniname2ctype_offset(str2433), 364}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2449), 311}, + {uniname2ctype_offset(str2449), 316}, {-1}, {uniname2ctype_offset(str2451), 54}, {uniname2ctype_offset(str2452), 96}, @@ -37448,18 +37740,18 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str2475), 23}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2481), 520}, + {uniname2ctype_offset(str2481), 525}, {-1}, - {uniname2ctype_offset(str2483), 420}, - {uniname2ctype_offset(str2484), 527}, + {uniname2ctype_offset(str2483), 425}, + {uniname2ctype_offset(str2484), 532}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2488), 182}, - {uniname2ctype_offset(str2489), 274}, - {uniname2ctype_offset(str2490), 270}, + {uniname2ctype_offset(str2489), 279}, + {uniname2ctype_offset(str2490), 275}, {-1}, {-1}, {uniname2ctype_offset(str2493), 178}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2500), 459}, + {uniname2ctype_offset(str2500), 464}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2507), 57}, {-1}, @@ -37471,14 +37763,14 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {uniname2ctype_offset(str2523), 153}, {-1}, {-1}, - {uniname2ctype_offset(str2526), 388}, + {uniname2ctype_offset(str2526), 393}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2539), 41}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2545), 546}, + {uniname2ctype_offset(str2545), 551}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2550), 451}, + {uniname2ctype_offset(str2550), 456}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37488,7 +37780,7 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str2579), 179}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2587), 449}, + {uniname2ctype_offset(str2587), 454}, {-1}, {uniname2ctype_offset(str2589), 126}, {-1}, {-1}, {-1}, {-1}, @@ -37500,14 +37792,14 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str2617), 63}, {uniname2ctype_offset(str2618), 213}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2622), 333}, + {uniname2ctype_offset(str2622), 338}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2630), 133}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2643), 277}, + {uniname2ctype_offset(str2643), 282}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2649), 273}, + {uniname2ctype_offset(str2649), 278}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2653), 165}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37523,16 +37815,16 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str2705), 49}, {-1}, {-1}, - {uniname2ctype_offset(str2708), 553}, + {uniname2ctype_offset(str2708), 558}, {-1}, - {uniname2ctype_offset(str2710), 565}, + {uniname2ctype_offset(str2710), 570}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2717), 88}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2751), 544}, + {uniname2ctype_offset(str2751), 549}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2758), 64}, {-1}, {-1}, {-1}, {-1}, @@ -37541,7 +37833,7 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str2775), 230}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2779), 381}, + {uniname2ctype_offset(str2779), 386}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2791), 242}, @@ -37551,11 +37843,11 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2823), 432}, + {uniname2ctype_offset(str2823), 437}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2829), 180}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2834), 330}, + {uniname2ctype_offset(str2834), 335}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2850), 231}, @@ -37570,17 +37862,17 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str2888), 133}, {-1}, {uniname2ctype_offset(str2890), 124}, - {uniname2ctype_offset(str2891), 523}, + {uniname2ctype_offset(str2891), 528}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2901), 447}, + {uniname2ctype_offset(str2901), 452}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2905), 290}, + {uniname2ctype_offset(str2905), 295}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2926), 125}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2930), 272}, + {uniname2ctype_offset(str2930), 277}, {-1}, {uniname2ctype_offset(str2932), 51}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37589,7 +37881,7 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str2953), 71}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str2967), 470}, + {uniname2ctype_offset(str2967), 475}, {uniname2ctype_offset(str2968), 197}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str2972), 146}, @@ -37597,14 +37889,14 @@ uniname2ctype_p (str, len) {-1}, {-1}, {uniname2ctype_offset(str2984), 204}, {-1}, {-1}, - {uniname2ctype_offset(str2987), 250}, + {uniname2ctype_offset(str2987), 255}, {-1}, - {uniname2ctype_offset(str2989), 404}, + {uniname2ctype_offset(str2989), 409}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str3007), 180}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3016), 398}, + {uniname2ctype_offset(str3016), 403}, {-1}, {-1}, {uniname2ctype_offset(str3019), 62}, {uniname2ctype_offset(str3020), 45}, @@ -37613,7 +37905,7 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str3038), 10}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3052), 377}, + {uniname2ctype_offset(str3052), 382}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str3066), 190}, @@ -37648,7 +37940,7 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3235), 433}, + {uniname2ctype_offset(str3235), 438}, {-1}, {-1}, {uniname2ctype_offset(str3238), 100}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37658,15 +37950,15 @@ uniname2ctype_p (str, len) {-1}, {uniname2ctype_offset(str3276), 177}, {-1}, {-1}, - {uniname2ctype_offset(str3279), 278}, - {uniname2ctype_offset(str3280), 280}, + {uniname2ctype_offset(str3279), 283}, + {uniname2ctype_offset(str3280), 285}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str3285), 214}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3296), 284}, + {uniname2ctype_offset(str3296), 289}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3305), 285}, + {uniname2ctype_offset(str3305), 290}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str3315), 205}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37687,9 +37979,9 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3438), 298}, + {uniname2ctype_offset(str3438), 303}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3442), 435}, + {uniname2ctype_offset(str3442), 440}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str3448), 46}, {-1}, {-1}, {-1}, {-1}, @@ -37702,7 +37994,7 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3496), 406}, + {uniname2ctype_offset(str3496), 411}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37729,9 +38021,9 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str3676), 201}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3689), 540}, + {uniname2ctype_offset(str3689), 545}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3697), 556}, + {uniname2ctype_offset(str3697), 561}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37740,11 +38032,11 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3763), 566}, + {uniname2ctype_offset(str3763), 571}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3791), 518}, + {uniname2ctype_offset(str3791), 523}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37785,7 +38077,7 @@ uniname2ctype_p (str, len) {uniname2ctype_offset(str3960), 147}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str3971), 396}, + {uniname2ctype_offset(str3971), 401}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -37849,7 +38141,7 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str4460), 237}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str4464), 369}, + {uniname2ctype_offset(str4464), 374}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {uniname2ctype_offset(str4483), 182}, @@ -37867,7 +38159,7 @@ uniname2ctype_p (str, len) {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, - {uniname2ctype_offset(str4608), 434}, + {uniname2ctype_offset(str4608), 439}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, {-1}, @@ -38050,9 +38342,9 @@ uniname2ctype_p (str, len) if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH) { - register int key = uniname2ctype_hash (str, len); + register unsigned int key = uniname2ctype_hash (str, len); - if (key <= MAX_HASH_VALUE && key >= 0) + if (key <= MAX_HASH_VALUE) { register int o = wordlist[key].name; if (o >= 0) diff --git a/template/unicode_norm_gen.tmpl b/template/unicode_norm_gen.tmpl index 74c361fe2a..64acfc8d18 100644 --- a/template/unicode_norm_gen.tmpl +++ b/template/unicode_norm_gen.tmpl @@ -7,7 +7,7 @@ # Constants for input and ouput directory InputDataDir = ARGV[0] || 'enc/unicode/data' -unicode_version = InputDataDir[/[\d.]+\z/] +unicode_version = InputDataDir[/.*\/(\d+\.\d+\.\d+)(?=\/|\z)/, 1] # convenience methods class Integer diff --git a/tool/enc-unicode.rb b/tool/enc-unicode.rb index 222573b3a8..50fa07e7c9 100755 --- a/tool/enc-unicode.rb +++ b/tool/enc-unicode.rb @@ -14,8 +14,8 @@ if ARGV[0] == "--header" header = true ARGV.shift end -unless ARGV.size == 1 - abort "Usage: #{$0} data_directory" +unless ARGV.size == 2 + abort "Usage: #{$0} data_directory emoji_data_directory" end $unicode_version = File.basename(ARGV[0])[/\A[.\d]+\z/] @@ -302,7 +302,7 @@ def constantize_blockname(name) end def get_file(name) - File.join(ARGV[0], name) + File.join(ARGV[name.start_with?("emoji-") ? 1 : 0], name) end def data_foreach(name, &block) |