1 /* Copyright 2002, 2003 Elliotte Rusty Harold 2 3 This library is free software; you can redistribute it and/or modify 4 it under the terms of version 2.1 of the GNU Lesser General Public 5 License as published by the Free Software Foundation. 6 7 This library is distributed in the hope that it will be useful, 8 but WITHOUT ANY WARRANTY; without even the implied warranty of 9 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 10 GNU Lesser General Public License for more details. 11 12 You should have received a copy of the GNU Lesser General Public 13 License along with this library; if not, write to the 14 Free Software Foundation, Inc., 59 Temple Place, Suite 330, 15 Boston, MA 02111-1307 USA 16 17 You can contact Elliotte Rusty Harold by sending e-mail to 18 elharo@ibiblio.org. Please include the word "XOM" in the 19 subject line. The XOM home page is located at http://www.xom.nu/ 20 */ 21 22 package nu.xom; 23 24 import java.io.Writer; 25 26 /** 27 * <p> 28 * ISO-8859-13, for Latvian and other Baltic languages. 29 * </p> 30 * 31 * @author Elliotte Rusty Harold 32 * @version 1.0 33 * 34 */ 35 class Latin7Writer extends TextWriter { 36 Latin7Writer(Writer out, String encoding)37 Latin7Writer(Writer out, String encoding) { 38 super(out, encoding); 39 } 40 41 /** 42 * @see nu.xom.TextWriter#needsEscaping(char) 43 */ needsEscaping(char c)44 boolean needsEscaping(char c) { 45 if (c <= 0xA0) return false; 46 switch (c) { // Latin-1 overlap 47 case 0x00A1: return true; // place holder to enable table lookup 48 case 0x00A2: return false; // CENT SIGN 49 case 0x00A3: return false; // POUND SIGN 50 case 0x00A4: return false; // CURRENCY SIGN 51 case 0x00A5: return true; // place holder to enable table lookup 52 case 0x00A6: return false; // BROKEN BAR 53 case 0x00A7: return false; // SECTION SIGN 54 case 0x00A8: return true; // place holder to enable table lookup 55 case 0x00A9: return false; // COPYRIGHT SIGN 56 case 0x00AA: return true; // place holder to enable table lookup 57 case 0x00AB: return false; // LEFT-POINTING DOUBLE ANGLE QUOTATION MARK 58 case 0x00AC: return false; // NOT SIGN 59 case 0x00AD: return false; // SOFT HYPHEN 60 case 0x00AE: return false; // REGISTERED SIGN 61 case 0x00AF: return true; // place holder to enable table lookup 62 case 0x00B0: return false; // DEGREE SIGN 63 case 0x00B1: return false; // PLUS-MINUS SIGN 64 case 0x00B2: return false; // SUPERSCRIPT TWO 65 case 0x00B3: return false; // SUPERSCRIPT THREE 66 case 0x00B4: return true; // place holder to enable table lookup 67 case 0x00B5: return false; // MICRO SIGN 68 case 0x00B6: return false; // PILCROW SIGN 69 case 0x00B7: return false; // MIDDLE DOT 70 case 0x00B8: return true; // place holder to enable table lookup 71 case 0x00B9: return false; // SUPERSCRIPT ONE 72 case 0x00BA: return true; // place holder to enable table lookup 73 case 0x00BB: return false; // RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK 74 case 0x00BC: return false; // VULGAR FRACTION ONE QUARTER 75 case 0x00BD: return false; // VULGAR FRACTION ONE HALF 76 case 0x00BE: return false; // VULGAR FRACTION THREE QUARTERS 77 case 0x00BF: return true; // place holder to enable table lookup 78 case 0x00C0: return true; // place holder to enable table lookup 79 case 0x00C1: return true; // place holder to enable table lookup 80 case 0x00C2: return true; // place holder to enable table lookup 81 case 0x00C3: return true; // place holder to enable table lookup 82 case 0x00C4: return false; // LATIN CAPITAL LETTER A WITH DIAERESIS 83 case 0x00C5: return false; // LATIN CAPITAL LETTER A WITH RING ABOVE 84 case 0x00C6: return false; // LATIN CAPITAL LETTER AE 85 case 0x00C7: return true; // place holder to enable table lookup 86 case 0x00C8: return true; // place holder to enable table lookup 87 case 0x00C9: return false; // LATIN CAPITAL LETTER E WITH ACUTE 88 case 0x00CA: return true; // place holder to enable table lookup 89 case 0x00CB: return true; // place holder to enable table lookup 90 case 0x00CC: return true; // place holder to enable table lookup 91 case 0x00CD: return true; // place holder to enable table lookup 92 case 0x00CE: return true; // place holder to enable table lookup 93 case 0x00CF: return true; // place holder to enable table lookup 94 case 0x00D0: return true; // place holder to enable table lookup 95 case 0x00D1: return true; // place holder to enable table lookup 96 case 0x00D2: return true; // place holder to enable table lookup 97 case 0x00D3: return false; // LATIN CAPITAL LETTER O WITH ACUTE 98 case 0x00D4: return true; // place holder to enable table lookup 99 case 0x00D5: return false; // LATIN CAPITAL LETTER O WITH TILDE 100 case 0x00D6: return false; // LATIN CAPITAL LETTER O WITH DIAERESIS 101 case 0x00D7: return false; // MULTIPLICATION SIGN 102 case 0x00D8: return false; // LATIN CAPITAL LETTER O WITH STROKE 103 case 0x00D9: return true; // place holder to enable table lookup 104 case 0x00DA: return true; // place holder to enable table lookup 105 case 0x00DB: return true; // place holder to enable table lookup 106 case 0x00DC: return false; // LATIN CAPITAL LETTER U WITH DIAERESIS 107 case 0x00DD: return true; // place holder to enable table lookup 108 case 0x00DE: return true; // place holder to enable table lookup 109 case 0x00DF: return false; // LATIN SMALL LETTER SHARP S (German) 110 case 0x00E0: return true; // place holder to enable table lookup 111 case 0x00E1: return true; // place holder to enable table lookup 112 case 0x00E2: return true; // place holder to enable table lookup 113 case 0x00E3: return true; // place holder to enable table lookup 114 case 0x00E4: return false; // LATIN SMALL LETTER A WITH DIAERESIS 115 case 0x00E5: return false; // LATIN SMALL LETTER A WITH RING ABOVE 116 case 0x00E6: return false; // LATIN SMALL LETTER AE 117 case 0x00E7: return true; // place holder to enable table lookup 118 case 0x00E8: return true; // place holder to enable table lookup 119 case 0x00E9: return false; // LATIN SMALL LETTER E WITH ACUTE 120 case 0x00EA: return true; // place holder to enable table lookup 121 case 0x00EB: return true; // place holder to enable table lookup 122 case 0x00EC: return true; // place holder to enable table lookup 123 case 0x00ED: return true; // place holder to enable table lookup 124 case 0x00EE: return true; // place holder to enable table lookup 125 case 0x00EF: return true; // place holder to enable table lookup 126 case 0x00F0: return true; // place holder to enable table lookup 127 case 0x00F1: return true; // place holder to enable table lookup 128 case 0x00F2: return true; // place holder to enable table lookup 129 case 0x00F3: return false; // LATIN SMALL LETTER O WITH ACUTE 130 case 0x00F4: return true; // place holder to enable table lookup 131 case 0x00F5: return false; // LATIN SMALL LETTER O WITH TILDE 132 case 0x00F6: return false; // LATIN SMALL LETTER O WITH DIAERESIS 133 case 0x00F7: return false; // DIVISION SIGN 134 case 0x00F8: return false; // LATIN SMALL LETTER O WITH STROKE 135 case 0x00F9: return true; // place holder to enable table lookup 136 case 0x00FA: return true; // place holder to enable table lookup 137 case 0x00FB: return true; // place holder to enable table lookup 138 case 0x00FC: return false; // LATIN SMALL LETTER U WITH DIAERESIS 139 case 0x00FD: return true; // place holder to enable table lookup 140 case 0x00FE: return true; // place holder to enable table lookup 141 case 0x00FF: return true; // place holder to enable table lookup 142 case 0x0100: return false; // LATIN CAPITAL LETTER A WITH MACRON 143 case 0x0101: return false; // LATIN SMALL LETTER A WITH MACRON 144 case 0x0102: return true; // place holder to enable table lookup 145 case 0x0103: return true; // place holder to enable table lookup 146 case 0x0104: return false; // LATIN CAPITAL LETTER A WITH OGONEK 147 case 0x0105: return false; // LATIN SMALL LETTER A WITH OGONEK 148 case 0x0106: return false; // LATIN CAPITAL LETTER C WITH ACUTE 149 case 0x0107: return false; // LATIN SMALL LETTER C WITH ACUTE 150 case 0x0108: return true; // place holder to enable table lookup 151 case 0x0109: return true; // place holder to enable table lookup 152 case 0x010A: return true; // place holder to enable table lookup 153 case 0x010B: return true; // place holder to enable table lookup 154 case 0x010C: return false; // LATIN CAPITAL LETTER C WITH CARON 155 case 0x010D: return false; // LATIN SMALL LETTER C WITH CARON 156 case 0x010E: return true; // place holder to enable table lookup 157 case 0x010F: return true; // place holder to enable table lookup 158 case 0x0110: return true; // place holder to enable table lookup 159 case 0x0111: return true; // place holder to enable table lookup 160 case 0x0112: return false; // LATIN CAPITAL LETTER E WITH MACRON 161 case 0x0113: return false; // LATIN SMALL LETTER E WITH MACRON 162 case 0x0114: return true; // place holder to enable table lookup 163 case 0x0115: return true; // place holder to enable table lookup 164 case 0x0116: return false; // LATIN CAPITAL LETTER E WITH DOT ABOVE 165 case 0x0117: return false; // LATIN SMALL LETTER E WITH DOT ABOVE 166 case 0x0118: return false; // LATIN CAPITAL LETTER E WITH OGONEK 167 case 0x0119: return false; // LATIN SMALL LETTER E WITH OGONEK 168 case 0x011A: return true; // place holder to enable table lookup 169 case 0x011B: return true; // place holder to enable table lookup 170 case 0x011C: return true; // place holder to enable table lookup 171 case 0x011D: return true; // place holder to enable table lookup 172 case 0x011E: return true; // place holder to enable table lookup 173 case 0x011F: return true; // place holder to enable table lookup 174 case 0x0120: return true; // place holder to enable table lookup 175 case 0x0121: return true; // place holder to enable table lookup 176 case 0x0122: return false; // LATIN CAPITAL LETTER G WITH CEDILLA 177 case 0x0123: return false; // LATIN SMALL LETTER G WITH CEDILLA 178 case 0x0124: return true; // place holder to enable table lookup 179 case 0x0125: return true; // place holder to enable table lookup 180 case 0x0126: return true; // place holder to enable table lookup 181 case 0x0127: return true; // place holder to enable table lookup 182 case 0x0128: return true; // place holder to enable table lookup 183 case 0x0129: return true; // place holder to enable table lookup 184 case 0x012A: return false; // LATIN CAPITAL LETTER I WITH MACRON 185 case 0x012B: return false; // LATIN SMALL LETTER I WITH MACRON 186 case 0x012C: return true; // place holder to enable table lookup 187 case 0x012D: return true; // place holder to enable table lookup 188 case 0x012E: return false; // LATIN CAPITAL LETTER I WITH OGONEK 189 case 0x012F: return false; // LATIN SMALL LETTER I WITH OGONEK 190 case 0x0130: return true; // place holder to enable table lookup 191 case 0x0131: return true; // place holder to enable table lookup 192 case 0x0132: return true; // place holder to enable table lookup 193 case 0x0133: return true; // place holder to enable table lookup 194 case 0x0134: return true; // place holder to enable table lookup 195 case 0x0135: return true; // place holder to enable table lookup 196 case 0x0136: return false; // LATIN CAPITAL LETTER K WITH CEDILLA 197 case 0x0137: return false; // LATIN SMALL LETTER K WITH CEDILLA 198 case 0x0138: return true; // place holder to enable table lookup 199 case 0x0139: return true; // place holder to enable table lookup 200 case 0x013A: return true; // place holder to enable table lookup 201 case 0x013B: return false; // LATIN CAPITAL LETTER L WITH CEDILLA 202 case 0x013C: return false; // LATIN SMALL LETTER L WITH CEDILLA 203 case 0x013D: return true; // place holder to enable table lookup 204 case 0x013E: return true; // place holder to enable table lookup 205 case 0x013F: return true; // place holder to enable table lookup 206 case 0x0140: return true; // place holder to enable table lookup 207 case 0x0141: return false; // LATIN CAPITAL LETTER L WITH STROKE 208 case 0x0142: return false; // LATIN SMALL LETTER L WITH STROKE 209 case 0x0143: return false; // LATIN CAPITAL LETTER N WITH ACUTE 210 case 0x0144: return false; // LATIN SMALL LETTER N WITH ACUTE 211 case 0x0145: return false; // LATIN CAPITAL LETTER N WITH CEDILLA 212 case 0x0146: return false; // LATIN SMALL LETTER N WITH CEDILLA 213 case 0x0147: return true; // place holder to enable table lookup 214 case 0x0148: return true; // place holder to enable table lookup 215 case 0x0149: return true; // place holder to enable table lookup 216 case 0x014A: return true; // place holder to enable table lookup 217 case 0x014B: return true; // place holder to enable table lookup 218 case 0x014C: return false; // LATIN CAPITAL LETTER O WITH MACRON 219 case 0x014D: return false; // LATIN SMALL LETTER O WITH MACRON 220 case 0x014E: return true; // place holder to enable table lookup 221 case 0x014F: return true; // place holder to enable table lookup 222 case 0x0150: return true; // place holder to enable table lookup 223 case 0x0151: return true; // place holder to enable table lookup 224 case 0x0152: return true; // place holder to enable table lookup 225 case 0x0153: return true; // place holder to enable table lookup 226 case 0x0154: return true; // place holder to enable table lookup 227 case 0x0155: return true; // place holder to enable table lookup 228 case 0x0156: return false; // LATIN CAPITAL LETTER R WITH CEDILLA 229 case 0x0157: return false; // LATIN SMALL LETTER R WITH CEDILLA 230 case 0x0158: return true; // place holder to enable table lookup 231 case 0x0159: return true; // place holder to enable table lookup 232 case 0x015A: return false; // LATIN CAPITAL LETTER S WITH ACUTE 233 case 0x015B: return false; // LATIN SMALL LETTER S WITH ACUTE 234 case 0x015C: return true; // place holder to enable table lookup 235 case 0x015D: return true; // place holder to enable table lookup 236 case 0x015E: return true; // place holder to enable table lookup 237 case 0x015F: return true; // place holder to enable table lookup 238 case 0x0160: return false; // LATIN CAPITAL LETTER S WITH CARON 239 case 0x0161: return false; // LATIN SMALL LETTER S WITH CARON 240 case 0x0162: return true; // place holder to enable table lookup 241 case 0x0163: return true; // place holder to enable table lookup 242 case 0x0164: return true; // place holder to enable table lookup 243 case 0x0165: return true; // place holder to enable table lookup 244 case 0x0166: return true; // place holder to enable table lookup 245 case 0x0167: return true; // place holder to enable table lookup 246 case 0x0168: return true; // place holder to enable table lookup 247 case 0x0169: return true; // place holder to enable table lookup 248 case 0x016A: return false; // LATIN CAPITAL LETTER U WITH MACRON 249 case 0x016B: return false; // LATIN SMALL LETTER U WITH MACRON 250 case 0x016C: return true; // place holder to enable table lookup 251 case 0x016D: return true; // place holder to enable table lookup 252 case 0x016E: return true; // place holder to enable table lookup 253 case 0x016F: return true; // place holder to enable table lookup 254 case 0x0170: return true; // place holder to enable table lookup 255 case 0x0171: return true; // place holder to enable table lookup 256 case 0x0172: return false; // LATIN CAPITAL LETTER U WITH OGONEK 257 case 0x0173: return false; // LATIN SMALL LETTER U WITH OGONEK 258 case 0x0174: return true; // place holder to enable table lookup 259 case 0x0175: return true; // place holder to enable table lookup 260 case 0x0176: return true; // place holder to enable table lookup 261 case 0x0177: return true; // place holder to enable table lookup 262 case 0x0178: return true; // place holder to enable table lookup 263 case 0x0179: return false; // LATIN CAPITAL LETTER Z WITH ACUTE 264 case 0x017A: return false; // LATIN SMALL LETTER Z WITH ACUTE 265 case 0x017B: return false; // LATIN CAPITAL LETTER Z WITH DOT ABOVE 266 case 0x017C: return false; // LATIN SMALL LETTER Z WITH DOT ABOVE 267 case 0x017D: return false; // LATIN CAPITAL LETTER Z WITH CARON 268 case 0x017E: return false; // LATIN SMALL LETTER Z WITH CARON 269 } 270 switch (c) { 271 case 0x2019: return false; // RIGHT SINGLE QUOTATION MARK 272 case 0x201A: return true; // place holder to enable table lookup 273 case 0x201B: return true; // place holder to enable table lookup 274 case 0x201C: return false; // LEFT DOUBLE QUOTATION MARK 275 case 0x201D: return false; // RIGHT DOUBLE QUOTATION MARK 276 case 0x201E: return false; // DOUBLE LOW-9 QUOTATION MARK 277 } 278 279 return true; 280 281 } 282 283 }