1 | # Default Unicode mappings
|
---|
2 | #
|
---|
3 | # Copyright 2000 Alexandre Julliard
|
---|
4 | #
|
---|
5 | # This library is free software; you can redistribute it and/or
|
---|
6 | # modify it under the terms of the GNU Lesser General Public
|
---|
7 | # License as published by the Free Software Foundation; either
|
---|
8 | # version 2.1 of the License, or (at your option) any later version.
|
---|
9 | #
|
---|
10 | # This library is distributed in the hope that it will be useful,
|
---|
11 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
12 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
---|
13 | # Lesser General Public License for more details.
|
---|
14 | #
|
---|
15 | # You should have received a copy of the GNU Lesser General Public
|
---|
16 | # License along with this library; if not, write to the Free Software
|
---|
17 | # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
|
---|
18 | #
|
---|
19 | # NOTES:
|
---|
20 | #
|
---|
21 | # This file contains default mappings for Unicode chars that aren't
|
---|
22 | # defined directly in the code page file. It is used by the cpmap.pl
|
---|
23 | # tool to build codepage tables.
|
---|
24 | #
|
---|
25 | # The first column defines equivalence sets; if any character in the set
|
---|
26 | # is defined in the codepage file, all others in the set will map to the
|
---|
27 | # same value.
|
---|
28 | #
|
---|
29 | # The second column specifies the default char to map the whole set to,
|
---|
30 | # if no characters in the set are defined by the codepage file.
|
---|
31 | #
|
---|
32 | # Characters that have an explicit decomposition in the UnicodeData.txt
|
---|
33 | # file (for instance 00c0 LATIN CAPITAL LETTER A WITH GRAVE -> 0041 0300)
|
---|
34 | # are handled automatically and do not have to be defined here. They can
|
---|
35 | # still be defined if necessary to override the unicode decomposition.
|
---|
36 | #
|
---|
37 |
|
---|
38 | #
|
---|
39 | # Sun LGPL Disclaimer: For the avoidance of doubt, except that if any license choice
|
---|
40 | # other than GPL or LGPL is available it will apply instead, Sun elects to use only
|
---|
41 | # the Lesser General Public License version 2.1 (LGPLv2) at this time for any software where
|
---|
42 | # a choice of LGPL license versions is made available with the language indicating
|
---|
43 | # that LGPLv2 or any later version may be used, or where a choice of which version
|
---|
44 | # of the LGPL is applied is otherwise unspecified.
|
---|
45 | #
|
---|
46 |
|
---|
47 | # letters
|
---|
48 |
|
---|
49 | 00a2,20a1 'c' # CENT SIGN, COLON SIGN
|
---|
50 | 00a3,20a4 'L' # POUND SIGN, LIRA SIGN
|
---|
51 | 00a5 'Y' # YEN SIGN
|
---|
52 | 00a6 '|' # BROKEN BAR
|
---|
53 | 00a9 'c' # COPYRIGHT SIGN
|
---|
54 | 00ae 'r' # REGISTERED SIGN
|
---|
55 | 00b5,03bc 'u' # MICRO SIGN, GREEK SMALL LETTER MU
|
---|
56 | 00c6 'A' # LATIN CAPITAL LETTER AE
|
---|
57 | 00d0,0110,0189 'D' # LATIN CAPITAL LETTER D WITH STROKE/AFRICAN D/ETH
|
---|
58 | 00d7 'x' # MULTIPLICATION SIGN
|
---|
59 | 00d8,2205 'O' # LATIN CAPITAL LETTER O WITH STROKE, EMPTY SET
|
---|
60 | 00df,03b2 none # LATIN SMALL LETTER SHARP S, GREEK SMALL LETTER BETA
|
---|
61 | 00e6 'a' # LATIN SMALL LETTER AE
|
---|
62 | 00f0 'd' # LATIN SMALL LETTER ETH
|
---|
63 | 00f8 'o' # LATIN SMALL LETTER O WITH STROKE
|
---|
64 | 0111 'd' # LATIN SMALL LETTER D WITH STROKE
|
---|
65 | 0126 'H' # LATIN CAPITAL LETTER H WITH STROKE
|
---|
66 | 0127 'h' # LATIN SMALL LETTER H WITH STROKE
|
---|
67 | 0131 'i' # LATIN SMALL LETTER DOTLESS I
|
---|
68 | 0141 'L' # LATIN CAPITAL LETTER L WITH STROKE
|
---|
69 | 0142 'l' # LATIN SMALL LETTER L WITH STROKE
|
---|
70 | 0152 'O' # LATIN CAPITAL LIGATURE OE
|
---|
71 | 0153 'o' # LATIN SMALL LIGATURE OE
|
---|
72 | 0166 'T' # LATIN CAPITAL LETTER T WITH STROKE
|
---|
73 | 0167 't' # LATIN SMALL LETTER T WITH STROKE
|
---|
74 | 0180 'b' # LATIN SMALL LETTER B WITH STROKE
|
---|
75 | 0190 'E' # LATIN CAPITAL LETTER OPEN E
|
---|
76 | 0191 'F' # LATIN CAPITAL LETTER F WITH HOOK
|
---|
77 | 0192 'f' # LATIN SMALL LETTER F WITH HOOK
|
---|
78 | 0197 'I' # LATIN CAPITAL LETTER I WITH STROKE
|
---|
79 | 019a 'l' # LATIN SMALL LETTER L WITH BAR
|
---|
80 | 019f 'O' # LATIN CAPITAL LETTER O WITH MIDDLE TILDE
|
---|
81 | 01ab 't' # LATIN SMALL LETTER T WITH PALATAL HOOK
|
---|
82 | 01ae 'T' # LATIN CAPITAL LETTER T WITH RETROFLEX HOOK
|
---|
83 | 01b6 'z' # LATIN SMALL LETTER Z WITH STROKE
|
---|
84 | 0391 'A' # GREEK CAPITAL LETTER ALPHA
|
---|
85 | 0392 'B' # GREEK CAPITAL LETTER BETA
|
---|
86 | 0395 'E' # GREEK CAPITAL LETTER EPSILON
|
---|
87 | 0396 'Z' # GREEK CAPITAL LETTER ZETA
|
---|
88 | 0397 'H' # GREEK CAPITAL LETTER ETA
|
---|
89 | 0399 'I' # GREEK CAPITAL LETTER IOTA
|
---|
90 | 039a 'K' # GREEK CAPITAL LETTER KAPPA
|
---|
91 | 039c 'M' # GREEK CAPITAL LETTER MU
|
---|
92 | 039d 'N' # GREEK CAPITAL LETTER NU
|
---|
93 | 039f 'O' # GREEK CAPITAL LETTER OMICRON
|
---|
94 | 03a1 'P' # GREEK CAPITAL LETTER RHO
|
---|
95 | 03a4 'T' # GREEK CAPITAL LETTER TAU
|
---|
96 | 03a5 'Y' # GREEK CAPITAL LETTER UPSILON
|
---|
97 | 03a7 'X' # GREEK CAPITAL LETTER CHI
|
---|
98 | 01e4 'G' # LATIN CAPITAL LETTER G WITH STROKE
|
---|
99 | 01e5 'g' # LATIN SMALL LETTER G WITH STROKE
|
---|
100 | 0261 'g' # LATIN SMALL LETTER SCRIPT G
|
---|
101 | 2118 'P' # SCRIPT CAPITAL P
|
---|
102 | 212e 'e' # ESTIMATED SYMBOL
|
---|
103 |
|
---|
104 | # accents
|
---|
105 |
|
---|
106 | 00a8,0308 '"' # DIAERESIS
|
---|
107 | 00af,02c9,0304 2014 # MACRON -> EM DASH
|
---|
108 | 00b4,02b9,02ca,0301,2032 ''' # ACUTE ACCENT, PRIME
|
---|
109 | 00b8,0327 ',' # CEDILLA
|
---|
110 | 02ba,02dd,030b,2033 '"' # DOUBLE ACUTE ACCENT, DOUBLE PRIME
|
---|
111 | 02c4,2303 '^' # UP ARROWHEAD
|
---|
112 | 02c6,0302 '^' # CIRCUMFLEX ACCENT
|
---|
113 | 02c7,030c 'v' # COMBINING CARON -> CARON
|
---|
114 | 02c8 ''' # MODIFIER LETTER VERTICAL LINE
|
---|
115 | 02cb,0300 '`' # GRAVE ACCENT
|
---|
116 | 02cd,0331,0332 '_' # MODIFIER LETTER LOW MACRON, COMBINING MACRON BELOW, COMBINING LOW LINE
|
---|
117 | 02d8,0306 none # BREVE
|
---|
118 | 02d9,0307 none # DOT ABOVE
|
---|
119 | 02da,030a,2070,2218 00b0 # RING ABOVE, SUPERSCRIPT ZERO, RING OPERATOR -> DEGREE SIGN
|
---|
120 | 02db,0328 none # OGONEK
|
---|
121 | 02dc,0303 '~' # SMALL TILDE
|
---|
122 | 0305,203e 00af # OVERLINE -> MACRON
|
---|
123 | 030e '"' # COMBINING DOUBLE VERTICAL LINE ABOVE
|
---|
124 | 0333 '_' # COMBINING DOUBLE LOW LINE
|
---|
125 |
|
---|
126 | # mathematical symbols
|
---|
127 |
|
---|
128 | 00b1,2213 none # PLUS-MINUS SIGN, MINUS-OR-PLUS SIGN
|
---|
129 | 2044,2215 '/' # FRACTION/DIVISION SLASH
|
---|
130 | 2216 '\' # SET MINUS
|
---|
131 | 2217 '*' # ASTERISK OPERATOR
|
---|
132 | 221a 'V' # SQUARE ROOT
|
---|
133 | 221f 'L' # RIGHT ANGLE
|
---|
134 | 2223 '|' # DIVIDES
|
---|
135 | 2229 'n' # INTERSECTION
|
---|
136 | 2236 ':' # RATIO
|
---|
137 | 2248 02DC # ALMOST EQUAL TO -> SMALL TILDE
|
---|
138 | 2261,2263 '=' # IDENTICAL TO, STRICTLY EQUIVALENT TO
|
---|
139 | 226a 00ab # MUCH LESS-THAN -> LEFT-POINTING DOUBLE ANGLE QUOTATION MARK
|
---|
140 | 226b 00bb # MUCH GREATER-THAN -> RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK
|
---|
141 | 22c5 00b7 # DOT OPERATOR -> MIDDLE DOT
|
---|
142 |
|
---|
143 | # misc symbols
|
---|
144 |
|
---|
145 | 00a1 '!' # INVERTED EXCLAMATION MARK
|
---|
146 | 00ab,300a '<' # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK, LEFT DOUBLE ANGLE BRACKET
|
---|
147 | 00ad '-' # SOFT HYPHEN
|
---|
148 | 00bb,300b '>' # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK, RIGHT DOUBLE ANGLE BRACKET
|
---|
149 | 01c0 2502 # LATIN LETTER DENTAL CLICK -> BOX DRAWINGS LIGHT VERTICAL
|
---|
150 | 01c3 '!' # LATIN LETTER RETROFLEX CLICK
|
---|
151 | 02bb 2018 # MODIFIER LETTER TURNED COMMA -> LEFT SINGLE QUOTATION MARK
|
---|
152 | 02bc,2019 ''' # RIGHT SINGLE QUOTATION MARK
|
---|
153 | 2010,2011,2212 '-' # HYPHEN, MINUS SIGN
|
---|
154 | 2013,2014,2015 '-' # EN DASH, EM DASH, HORIZONTAL BAR, MINUS SIGN
|
---|
155 | 2018,201b,2035 '`' # LEFT SINGLE QUOTATION MARK, SINGLE HIGH-REVERSED-9 QUOTATION MARK, REVERSED PRIME
|
---|
156 | 201a ',' # SINGLE LOW-9 QUOTATION MARK
|
---|
157 | 201c,301d '"' # LEFT DOUBLE QUOTATION MARK, REVERSED DOUBLE PRIME QUOTATION MARK
|
---|
158 | 201d,301e '"' # RIGHT DOUBLE QUOTATION MARK, DOUBLE PRIME QUOTATION MARK
|
---|
159 | 201e,301f ',' # LOW DOUBLE PRIME QUOTATION MARK, DOUBLE LOW-9 QUOTATION MARK
|
---|
160 | 2022,2219 none # BULLET, BULLET OPERATOR
|
---|
161 | 2039,3008 '<' # SINGLE LEFT-POINTING ANGLE QUOTATION MARK, LEFT ANGLE BRACKET
|
---|
162 | 203a,3009 '>' # SINGLE RIGHT-POINTING ANGLE QUOTATION MARK, RIGHT ANGLE BRACKET
|
---|
163 | 203c '!' # DOUBLE EXCLAMATION MARK
|
---|
164 | 2190 2039 # LEFTWARDS ARROW -> SINGLE LEFT-POINTING ANGLE QUOTATION MARK
|
---|
165 | 2191 02c6 # UPWARDS ARROW -> MODIFIER LETTER CIRCUMFLEX ACCENT
|
---|
166 | 2192 203a # RIGHTWARDS ARROW -> SINGLE RIGHT-POINTING ANGLE QUOTATION MARK
|
---|
167 | 2193 02c7 # DOWNWARDS ARROW -> CARON
|
---|
168 | 2194 '-' # LEFT RIGHT ARROW
|
---|
169 | 2195 '|' # UP DOWN ARROW
|
---|
170 | 223c '~' # TILDE
|
---|
171 | 301a '[' # LEFT WHITE SQUARE BRACKET
|
---|
172 | 301b ']' # RIGHT WHITE SQUARE BRACKET
|
---|
173 | fe49,fe4a,fe4b,fe4c 203e # DOUBLE WAVY OVERLINE -> OVERLINE
|
---|
174 | fe4d,fe4e,fe4f '_' # DASHED/CENTRELINE/WAVY LOW LINE
|
---|
175 |
|
---|
176 | # box drawing chars
|
---|
177 |
|
---|
178 | 2500,2501 '-' # BOX DRAWINGS LIGHT/HEAVY HORIZONTAL
|
---|
179 | 2502,2503 '|' # BOX DRAWINGS LIGHT/HEAVY VERTICAL
|
---|
180 | 2504,2505 '-' # BOX DRAWINGS LIGHT/HEAVY TRIPLE DASH HORIZONTAL
|
---|
181 | 2506,2507 00a6 # BOX DRAWINGS LIGHT/HEAVY TRIPLE DASH VERTICAL -> BROKEN BAR
|
---|
182 | 2508,2509 '-' # BOX DRAWINGS LIGHT/HEAVY QUADRUPLE DASH HORIZONTAL
|
---|
183 | 250a,250b 00a6 # BOX DRAWINGS LIGHT/HEAVY QUADRUPLE DASH VERTICAL -> BROKEN BAR
|
---|
184 | 250c,250d,250e,250f '+' # BOX DRAWINGS LIGHT/HEAVY DOWN AND RIGHT
|
---|
185 | 2510,2511,2512,2513 00ac # BOX DRAWINGS LIGHT/HEAVY DOWN AND LEFT -> NOT SIGN
|
---|
186 | 2514,2515,2516,2517 'L' # BOX DRAWINGS LIGHT/HEAVY UP AND RIGHT
|
---|
187 | 2518,2519,251a,251b '+' # BOX DRAWINGS LIGHT/HEAVY UP AND LEFT
|
---|
188 | 251c,251d,251e,251f,2520,2521,2522,2523 '+' # BOX DRAWINGS LIGHT/HEAVY VERTICAL AND RIGHT
|
---|
189 | 2524,2525,2526,2527,2528,2529,252a,252b '+' # BOX DRAWINGS LIGHT/HEAVY VERTICAL AND LEFT
|
---|
190 | 252c,252d,252e,252f,2530,2531,2532,2533 'T' # BOX DRAWINGS LIGHT/HEAVY DOWN AND HORIZONTAL
|
---|
191 | 2534,2535,2536,2537,2538,2539,253a,253b '+' # BOX DRAWINGS LIGHT/HEAVY UP AND HORIZONTAL
|
---|
192 | 253c,253d,253e,253f,2540,2541,2542,2543,2544,2545,2546,2547,2548,2549,254a,254b '+' # BOX DRAWINGS LIGHT/HEAVY VERTICAL AND HORIZONTAL
|
---|
193 | 254c,254d '-' # BOX DRAWINGS LIGHT/HEAVY DOUBLE DASH HORIZONTAL
|
---|
194 | 254e,254f 00a6 # BOX DRAWINGS LIGHT/HEAVY DOUBLE DASH VERTICAL -> BROKEN BAR
|
---|
195 | 2550 '=' # BOX DRAWINGS DOUBLE HORIZONTAL
|
---|
196 | 2551 '|' # BOX DRAWINGS DOUBLE VERTICAL
|
---|
197 | 2552,2553,2554 '+' # BOX DRAWINGS DOWN AND RIGHT
|
---|
198 | 2555,2556,2557 00ac # BOX DRAWINGS DOWN AND LEFT -> NOT SIGN
|
---|
199 | 2558,2559,255a 'L' # BOX DRAWINGS UP AND RIGHT
|
---|
200 | 255b,255c,255d '+' # BOX DRAWINGS UP AND LEFT
|
---|
201 | 255e,255f,2560 '+' # BOX DRAWINGS VERTICAL AND RIGHT
|
---|
202 | 2561,2562,2563 '+' # BOX DRAWINGS VERTICAL AND LEFT
|
---|
203 | 2564,2565,2566 'T' # BOX DRAWINGS DOWN AND HORIZONTAL
|
---|
204 | 2567,2568,2569 '+' # BOX DRAWINGS UP AND HORIZONTAL
|
---|
205 | 256a,256b,256c '+' # BOX DRAWINGS VERTICAL AND HORIZONTAL
|
---|
206 | 2571 '/' # BOX DRAWINGS LIGHT DIAGONAL UPPER RIGHT TO LOWER LEFT
|
---|
207 | 2572 '\' # BOX DRAWINGS LIGHT DIAGONAL UPPER LEFT TO LOWER RIGHT
|
---|
208 | 2573 'X' # BOX DRAWINGS LIGHT DIAGONAL CROSS
|
---|
209 | 2574,2576,2578,257a,257c,257e '-' # BOX DRAWINGS LIGHT/HEAVY LEFT/RIGHT
|
---|
210 | 2575,2577,2579,257b,257d,257f '|' # BOX DRAWINGS LIGHT/HEAVY UP/DOWN
|
---|
211 | 25cb,25ef 'O' # WHITE/LARGE CIRCLE
|
---|