Changeset 11202
 Timestamp:
 02/16/15 23:26:25 (4 years ago)
 File:

 1 edited
Legend:
 Unmodified
 Added
 Removed

trunk/common/collation/yi.xml
r10671 r11202 14 14 <collations> 15 15 <collation type="search" > 16 <settings normalization="on"/> 17 <! root search, suppress contractions for Thai, Lao > 18 <suppress_contractions>[เไ ເໄ]</suppress_contractions> 16 <import source="und" type="search"/> 19 17 <cr><![CDATA[ 20 # root search rules for Symbols 21 &'='<'≠' 22 # root search rules for Arabic, Hebrew, plus two Hebrewspecific additions 23 &ا # 0627 ARABIC LETTER ALEF 24 <<<ﺎ<<<ﺍ # FE8E, FE8D: FINAL FORM, ISOLATED FORM 25 <<آ # 0622 ARABIC LETTER ALEF WITH MADDA ABOVE 26 <<<ﺂ<<<ﺁ # FE82, FE81: FINAL FORM, ISOLATED FORM 27 <<أ # 0623 ARABIC LETTER ALEF WITH HAMZA ABOVE 28 <<<ﺄ<<<ﺃ # FE84, FE83: FINAL FORM, ISOLATED FORM 29 <<إ # 0625 ARABIC LETTER ALEF WITH HAMZA BELOW 30 <<<ﺈ<<<ﺇ # FE88, FE87: FINAL FORM, ISOLATED FORM 31 &و # 0648 ARABIC LETTER WAW 32 <<<ۥ # 06E5: SMALL WAW 33 <<<ﻮ<<<ﻭ # FEEE, FEED: FINAL FORM, ISOLATED FORM 34 <<ؤ # 0624 ARABIC LETTER WAW WITH HAMZA ABOVE 35 <<<ﺆ<<<ﺅ # FE86, FE85: FINAL FORM, ISOLATED FORM 36 &ي # 064A ARABIC LETTER YEH 37 <<<ۦ # 06E6: ARABIC SMALL YEH 38 <<<ﻳ<<<ﻴ<<<ﻲ<<<ﻱ # FEF3, FEF4, FEF2, FEF1: INITIAL FORM, MEDIAL FORM, FINAL FORM, ISOLATED FORM 39 <<ئ # 0626 ARABIC LETTER YEH WITH HAMZA ABOVE 40 <<<ﺋ<<<ﺌ<<<ﺊ<<<ﺉ # FE8B, FE8C, FE8A, FE89: INITIAL FORM, MEDIAL FORM. FINAL FORM, ISOLATED FORM 41 <<ى # 0649 ARABIC LETTER ALEF MAKSURA 42 <<<ﯨ<<<ﯩ # FBE8, FBE9: UIGHUR KAZAKH KIRGHIZ ALEF MAKSURA INITIAL FORM, MEDIAL FORM 43 <<<ﻰ<<<ﻯ # FEF0, FEEF: FINAL FORM, ISOLATED FORM 44 &ه # 0647 ARABIC LETTER HEH 45 <<<ﻫ<<<ﻬ<<<ﻪ<<<ﻩ # FEEB, FEEC, FEEA, FEE9: INITIAL FORM, MEDIAL FORM, FINAL FORM;, ISOLATED FORM 46 <<ة # 0629 ARABIC LETTER TEH MARBUTA 47 <<<ﺔ<<<ﺓ # FE94, FE93: FINAL FORM, ISOLATED FORM 48 &[last primary ignorable]<<׳ # 05F3 HEBREW PUNCTUATION GERESH, ignorable 49 18 &׳ # 05F3 HEBREW PUNCTUATION GERESH, ignorable 50 19 <<'' # APOSTROPHE, Hebrewspecific addition, should match GERESH 51 <<״ # 05F4 HEBREW PUNCTUATION GERSHAYIM, ignorable20 &״ # 05F4 HEBREW PUNCTUATION GERSHAYIM, ignorable 52 21 <<'\"' # QUOTATION MARK, Hebrewspecific addition, should match GERSHAYIM 53 <<ـ # 0640 ARABIC TATWEEL54 # Don't need explicit entries for 064B  0652 ARABIC FATHATAN  ARABIC SUKUN;55 # these are already ignorable at level 1, and are not involved in contractions56 <<ฺ # 0E3A THAI CHARACTER PHINTHU57 # root search rules for modern Korean jamos58 # Korean modern complex consonants, decompose: x(L) = CHOSEONG x, x(T) = JONGSEONG x59 &ᄀ # 1100 KIYEOK(L) = 11A8 KIYEOK(T)60 =ᆨ61 &ᄀᄀ # 1100 KIYEOK(L) + 1100 KIYEOK(L) = 1101 SSANGKIYEOK(L) = 11A9 SSANGKIYEOK(T)62 =ᄁ=ᆩ63 &ᄀᄉ # 1100 KIYEOK(L) + 1109 SIOS(L) = 11AA KIYEOKSIOS(T)64 =ᆪ65 &ᄂ # 1102 NIEUN(L) = 11AB NIEUN(T)66 =ᆫ67 &ᄂᄌ # 1102 NIEUN(L) + 110C CIEUC(L) = 11AC NIEUNCIEUC(T), also archaic 115C NIEUNCIEUC(L)68 =ᆬ69 &ᄂᄒ # 1102 NIEUN(L) + 1112 HIEUH(L) = 11AD NIEUNHIEUH(T), also archaic 115D NIEUNHIEUH(L)70 =ᆭ71 &ᄃ # 1103 TIKEUT(L) = 11AE TIKEUT(T)72 =ᆮ73 &ᄃᄃ # 1103 TIKEUT(L) + 1103 TIKEUT(L) = 1104 SSANGTIKEUT(L)74 =ᄄ75 &ᄅ # 1105 RIEUL(L) = 11AF RIEUL(T)76 =ᆯ77 &ᄅᄀ # 1105 RIEUL(L) + 1100 KIYEOK(L) = 11B0 RIEULKIYEOK(T)78 =ᆰ79 &ᄅᄆ # 1105 RIEUL(L) + 1106 MIEUM(L) = 11B1 RIEULMIEUM(T)80 =ᆱ81 &ᄅᄇ # 1105 RIEUL(L) + 1107 PIEUP(L) = 11B2 RIEULPIEUP(T)82 =ᆲ83 &ᄅᄉ # 1105 RIEUL(L) + 1109 SIOS(L) = 11B3 RIEULSIOS(T)84 =ᆳ85 &ᄅᄐ # 1105 RIEUL(L) + 1110 THIEUTH(L) = 11B4 RIEULTHIEUTH(T)86 =ᆴ87 &ᄅᄑ # 1105 RIEUL(L) + 1111 PHIEUPH(L) = 11B5 RIEULPHIEUPH(T)88 =ᆵ89 &ᄅᄒ # 1105 RIEUL(L) + 1112 HIEUH(L) = 11B6 RIEULHIEUH(T), also archic 111A RIEULHIEUH(L)90 =ᆶ91 &ᄆ # 1106 MIEUM(L) = 11B7 MIEUM(T)92 =ᆷ93 &ᄇ # 1107 PIEUP(L) = 11B8 PIEUP(T)94 =ᆸ95 &ᄇᄇ # 1107 PIEUP(L) + 1107 PIEUP(L) = 1108 SSANGPIEUP(L)96 =ᄈ97 &ᄇᄉ # 1107 PIEUP(L) + 1109 SIOS(L) = 11B9 PIEUPSIOS(T), also archaic 1121 PIEUPSIOS(L)98 =ᆹ99 &ᄉ # 1109 SIOS(L) = 11BA SIOS(T)100 =ᆺ101 &ᄉᄉ # 1109 SIOS(L) + 1109 SIOS(L) = 110A SSANGSIOS(L) = 11BB SSANGSIOS(T)102 =ᄊ=ᆻ103 &ᄋ # 110B IEUNG(L) = 11BC IEUNG(T)104 =ᆼ105 &ᄌ # 110C CIEUC(L) = 11BD CIEUC(T)106 =ᆽ107 &ᄌᄌ # 110C CIEUC(L) + 110C CIEUC(L) = 110D SSANGCIEUC(L)108 =ᄍ109 &ᄎ # 110E CHIEUCH(L) = 11BE CHIEUCH(T)110 =ᆾ111 &ᄏ # 110F KHIEUKH(L) = 11BF KHIEUKH(T)112 =ᆿ113 &ᄐ # 1110 THIEUTH(L) = 11C0 THIEUTH(T)114 =ᇀ115 &ᄑ # 1111 PHIEUPH(L) = 11C1 PHIEUPH(T)116 =ᇁ117 &ᄒ # 1112 HIEUH(L) = 11C2 HIEUH(T)118 =ᇂ119 # Korean modern complex vowels, decompose120 &ᅡᅵ # 1161 A(V) + 1175 I(V) = 1162 AE(V)121 =ᅢ122 &ᅣᅵ # 1163 YA(V) + 1175 I(V) = 1164 YAE(V)123 =ᅤ124 &ᅥᅵ # 1165 EO(V) + 1175 I(V) = 1166 E(V)125 =ᅦ126 &ᅧᅵ # 1167 YEO(V) + 1175 I(V) = 1168 YE(V)127 =ᅨ128 &ᅩᅡ # 1169 O(V) + 1161 A(V) = 116A WA(V)129 =ᅪ130 &ᅩᅡᅵ # 1169 O(V) + 1161 A(V) + 1175 I(V) = 116B WAE(V)131 =ᅫ132 &ᅩᅵ # 1169 O(V) + 1175 I(V) = 116C OE(V)133 =ᅬ134 &ᅮᅴ # 116E U(V) + 1174 YI(V) = 116F WEO(V)135 =ᅯ136 &ᅮᅴᅵ # 116E U(V) + 1174 YI(V) + 1175 I(V) = 1170 WE(V)137 =ᅰ138 &ᅮᅵ # 116E U(V) + 1175 I(V) = 1171 WI(V)139 =ᅱ140 22 ]]></cr> 141 23 </collation>
Note: See TracChangeset
for help on using the changeset viewer.