1 /* 2 * Copyright © 2009 Red Hat, Inc. 3 * Copyright © 2011 Codethink Limited 4 * Copyright © 2010,2011,2012 Google, Inc. 5 * 6 * This is part of HarfBuzz, a text shaping library. 7 * 8 * Permission is hereby granted, without written agreement and without 9 * license or royalty fees, to use, copy, modify, and distribute this 10 * software and its documentation for any purpose, provided that the 11 * above copyright notice and the following two paragraphs appear in 12 * all copies of this software. 13 * 14 * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR 15 * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES 16 * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN 17 * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH 18 * DAMAGE. 19 * 20 * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING, 21 * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND 22 * FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS 23 * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO 24 * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS. 25 * 26 * Red Hat Author(s): Behdad Esfahbod 27 * Codethink Author(s): Ryan Lortie 28 * Google Author(s): Behdad Esfahbod 29 */ 30 31 #ifndef HB_UNICODE_PRIVATE_HH 32 #define HB_UNICODE_PRIVATE_HH 33 34 #include "hb-private.hh" 35 36 #include "hb-unicode.h" 37 #include "hb-object-private.hh" 38 39 40 extern HB_INTERNAL const uint8_t _hb_modified_combining_class[256]; 41 42 /* 43 * hb_unicode_funcs_t 44 */ 45 46 #define HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS \ 47 HB_UNICODE_FUNC_IMPLEMENT (combining_class) \ 48 HB_UNICODE_FUNC_IMPLEMENT (eastasian_width) \ 49 HB_UNICODE_FUNC_IMPLEMENT (general_category) \ 50 HB_UNICODE_FUNC_IMPLEMENT (mirroring) \ 51 HB_UNICODE_FUNC_IMPLEMENT (script) \ 52 HB_UNICODE_FUNC_IMPLEMENT (compose) \ 53 HB_UNICODE_FUNC_IMPLEMENT (decompose) \ 54 HB_UNICODE_FUNC_IMPLEMENT (decompose_compatibility) \ 55 /* ^--- Add new callbacks here */ 56 57 /* Simple callbacks are those taking a hb_codepoint_t and returning a hb_codepoint_t */ 58 #define HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE \ 59 HB_UNICODE_FUNC_IMPLEMENT (hb_unicode_combining_class_t, combining_class) \ 60 HB_UNICODE_FUNC_IMPLEMENT (unsigned int, eastasian_width) \ 61 HB_UNICODE_FUNC_IMPLEMENT (hb_unicode_general_category_t, general_category) \ 62 HB_UNICODE_FUNC_IMPLEMENT (hb_codepoint_t, mirroring) \ 63 HB_UNICODE_FUNC_IMPLEMENT (hb_script_t, script) \ 64 /* ^--- Add new simple callbacks here */ 65 66 struct hb_unicode_funcs_t { 67 hb_object_header_t header; 68 ASSERT_POD (); 69 70 hb_unicode_funcs_t *parent; 71 72 bool immutable; 73 74 #define HB_UNICODE_FUNC_IMPLEMENT(return_type, name) \ 75 inline return_type name (hb_codepoint_t unicode) { return func.name (this, unicode, user_data.name); } 76 HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE 77 #undef HB_UNICODE_FUNC_IMPLEMENT 78 composehb_unicode_funcs_t79 inline hb_bool_t compose (hb_codepoint_t a, hb_codepoint_t b, 80 hb_codepoint_t *ab) 81 { 82 *ab = 0; 83 if (unlikely (!a || !b)) return false; 84 return func.compose (this, a, b, ab, user_data.compose); 85 } 86 decomposehb_unicode_funcs_t87 inline hb_bool_t decompose (hb_codepoint_t ab, 88 hb_codepoint_t *a, hb_codepoint_t *b) 89 { 90 *a = ab; *b = 0; 91 return func.decompose (this, ab, a, b, user_data.decompose); 92 } 93 decompose_compatibilityhb_unicode_funcs_t94 inline unsigned int decompose_compatibility (hb_codepoint_t u, 95 hb_codepoint_t *decomposed) 96 { 97 unsigned int ret = func.decompose_compatibility (this, u, decomposed, user_data.decompose_compatibility); 98 if (ret == 1 && u == decomposed[0]) { 99 decomposed[0] = 0; 100 return 0; 101 } 102 decomposed[ret] = 0; 103 return ret; 104 } 105 106 107 unsigned int modified_combining_classhb_unicode_funcs_t108 modified_combining_class (hb_codepoint_t unicode) 109 { 110 /* XXX This hack belongs to the Myanmar shaper. */ 111 if (unicode == 0x1037) unicode = 0x103A; 112 113 return _hb_modified_combining_class[combining_class (unicode)]; 114 } 115 116 inline hb_bool_t is_variation_selectorhb_unicode_funcs_t117 is_variation_selector (hb_codepoint_t unicode) 118 { 119 return unlikely (hb_in_ranges<hb_codepoint_t> (unicode, 120 0x180B, 0x180D, /* MONGOLIAN FREE VARIATION SELECTOR ONE..THREE */ 121 0xFE00, 0xFE0F, /* VARIATION SELECTOR-1..16 */ 122 0xE0100, 0xE01EF)); /* VARIATION SELECTOR-17..256 */ 123 } 124 125 /* Default_Ignorable codepoints: 126 * 127 * Note that as of Oct 2012 (Unicode 6.2), U+180E MONGOLIAN VOWEL SEPARATOR 128 * is NOT Default_Ignorable, but it really behaves in a way that it should 129 * be. That has been reported to the Unicode Technical Committee for 130 * consideration. As such, we include it here, since Uniscribe removes it. 131 * 132 * Note: While U+115F and U+1160 are Default_Ignorable, we do NOT want to 133 * hide them, as the way Uniscribe has implemented them is with regular 134 * spacing glyphs, and that's the way fonts are made to work. As such, 135 * we make exceptions for those two. 136 * 137 * Gathered from: 138 * http://unicode.org/cldr/utility/list-unicodeset.jsp?a=[:DI:]&abb=on&ucd=on&esc=on 139 * 140 * Last updated to the page with the following versions: 141 * Version 3.6; ICU version: 50.0.1.0; Unicode version: 6.1.0.0 142 * 143 * 4,167 Code Points 144 * 145 * [\u00AD\u034F\u115F\u1160\u17B4\u17B5\u180B-\u180D\u200B-\u200F\u202A-\u202E\u2060-\u206F\u3164\uFE00-\uFE0F\uFEFF\uFFA0\uFFF0-\uFFF8\U0001D173-\U0001D17A\U000E0000-\U000E0FFF] 146 * 147 * 00AD ;SOFT HYPHEN 148 * 034F ;COMBINING GRAPHEME JOINER 149 * #115F ;HANGUL CHOSEONG FILLER 150 * #1160 ;HANGUL JUNGSEONG FILLER 151 * 17B4 ;KHMER VOWEL INHERENT AQ 152 * 17B5 ;KHMER VOWEL INHERENT AA 153 * 180B..180D ;MONGOLIAN FREE VARIATION SELECTOR THREE 154 * 200B..200F ;RIGHT-TO-LEFT MARK 155 * 202A..202E ;RIGHT-TO-LEFT OVERRIDE 156 * 2060..206F ;NOMINAL DIGIT SHAPES 157 * 3164 ;HANGUL FILLER 158 * FE00..FE0F ;VARIATION SELECTOR-16 159 * FEFF ;ZERO WIDTH NO-BREAK SPACE 160 * FFA0 ;HALFWIDTH HANGUL FILLER 161 * FFF0..FFF8 ;<unassigned-FFF8> 162 * 1D173..1D17A ;MUSICAL SYMBOL END PHRASE 163 * E0000..E0FFF ;<unassigned-E0FFF> 164 */ 165 inline hb_bool_t is_default_ignorablehb_unicode_funcs_t166 is_default_ignorable (hb_codepoint_t ch) 167 { 168 hb_codepoint_t plane = ch >> 16; 169 if (likely (plane == 0)) 170 { 171 /* BMP */ 172 hb_codepoint_t page = ch >> 8; 173 switch (page) { 174 case 0x00: return unlikely (ch == 0x00AD); 175 case 0x03: return unlikely (ch == 0x034F); 176 case 0x17: return hb_in_range<hb_codepoint_t> (ch, 0x17B4, 0x17B5); 177 case 0x18: return hb_in_range<hb_codepoint_t> (ch, 0x180B, 0x180E); 178 case 0x20: return hb_in_ranges<hb_codepoint_t> (ch, 0x200B, 0x200F, 179 0x202A, 0x202E, 180 0x2060, 0x206F); 181 case 0x31: return unlikely (ch == 0x3164); 182 case 0xFE: return hb_in_range<hb_codepoint_t> (ch, 0xFE00, 0xFE0F) || ch == 0xFEFF; 183 case 0xFF: return hb_in_range<hb_codepoint_t> (ch, 0xFFF0, 0xFFF8) || ch == 0xFFA0; 184 default: return false; 185 } 186 } 187 else 188 { 189 /* Other planes */ 190 switch (plane) { 191 case 0x01: return hb_in_range<hb_codepoint_t> (ch, 0x0001D173, 0x0001D17A); 192 case 0x0E: return hb_in_range<hb_codepoint_t> (ch, 0x000E0000, 0x000E0FFF); 193 default: return false; 194 } 195 } 196 } 197 198 199 struct { 200 #define HB_UNICODE_FUNC_IMPLEMENT(name) hb_unicode_##name##_func_t name; 201 HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS 202 #undef HB_UNICODE_FUNC_IMPLEMENT 203 } func; 204 205 struct { 206 #define HB_UNICODE_FUNC_IMPLEMENT(name) void *name; 207 HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS 208 #undef HB_UNICODE_FUNC_IMPLEMENT 209 } user_data; 210 211 struct { 212 #define HB_UNICODE_FUNC_IMPLEMENT(name) hb_destroy_func_t name; 213 HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS 214 #undef HB_UNICODE_FUNC_IMPLEMENT 215 } destroy; 216 }; 217 218 219 extern HB_INTERNAL const hb_unicode_funcs_t _hb_unicode_funcs_nil; 220 221 222 /* Modified combining marks */ 223 224 /* Hebrew 225 * 226 * We permute the "fixed-position" classes 10-26 into the order 227 * described in the SBL Hebrew manual: 228 * 229 * http://www.sbl-site.org/Fonts/SBLHebrewUserManual1.5x.pdf 230 * 231 * (as recommended by: 232 * http://forum.fontlab.com/archive-old-microsoft-volt-group/vista-and-diacritic-ordering-t6751.0.html) 233 * 234 * More details here: 235 * https://bugzilla.mozilla.org/show_bug.cgi?id=662055 236 */ 237 #define HB_MODIFIED_COMBINING_CLASS_CCC10 22 /* sheva */ 238 #define HB_MODIFIED_COMBINING_CLASS_CCC11 15 /* hataf segol */ 239 #define HB_MODIFIED_COMBINING_CLASS_CCC12 16 /* hataf patah */ 240 #define HB_MODIFIED_COMBINING_CLASS_CCC13 17 /* hataf qamats */ 241 #define HB_MODIFIED_COMBINING_CLASS_CCC14 23 /* hiriq */ 242 #define HB_MODIFIED_COMBINING_CLASS_CCC15 18 /* tsere */ 243 #define HB_MODIFIED_COMBINING_CLASS_CCC16 19 /* segol */ 244 #define HB_MODIFIED_COMBINING_CLASS_CCC17 20 /* patah */ 245 #define HB_MODIFIED_COMBINING_CLASS_CCC18 21 /* qamats */ 246 #define HB_MODIFIED_COMBINING_CLASS_CCC19 14 /* holam */ 247 #define HB_MODIFIED_COMBINING_CLASS_CCC20 24 /* qubuts */ 248 #define HB_MODIFIED_COMBINING_CLASS_CCC21 12 /* dagesh */ 249 #define HB_MODIFIED_COMBINING_CLASS_CCC22 25 /* meteg */ 250 #define HB_MODIFIED_COMBINING_CLASS_CCC23 13 /* rafe */ 251 #define HB_MODIFIED_COMBINING_CLASS_CCC24 10 /* shin dot */ 252 #define HB_MODIFIED_COMBINING_CLASS_CCC25 11 /* sin dot */ 253 #define HB_MODIFIED_COMBINING_CLASS_CCC26 26 /* point varika */ 254 255 /* 256 * Arabic 257 * 258 * Modify to move Shadda (ccc=33) before other marks. See: 259 * http://unicode.org/faq/normalization.html#8 260 * http://unicode.org/faq/normalization.html#9 261 */ 262 #define HB_MODIFIED_COMBINING_CLASS_CCC27 28 /* fathatan */ 263 #define HB_MODIFIED_COMBINING_CLASS_CCC28 29 /* dammatan */ 264 #define HB_MODIFIED_COMBINING_CLASS_CCC29 30 /* kasratan */ 265 #define HB_MODIFIED_COMBINING_CLASS_CCC30 31 /* fatha */ 266 #define HB_MODIFIED_COMBINING_CLASS_CCC31 32 /* damma */ 267 #define HB_MODIFIED_COMBINING_CLASS_CCC32 33 /* kasra */ 268 #define HB_MODIFIED_COMBINING_CLASS_CCC33 27 /* shadda */ 269 #define HB_MODIFIED_COMBINING_CLASS_CCC34 34 /* sukun */ 270 #define HB_MODIFIED_COMBINING_CLASS_CCC35 35 /* superscript alef */ 271 272 /* Syriac */ 273 #define HB_MODIFIED_COMBINING_CLASS_CCC36 36 /* superscript alaph */ 274 275 /* Telugu 276 * 277 * Modify Telugu length marks (ccc=84, ccc=91). 278 * These are the only matras in the main Indic scripts range that have 279 * a non-zero ccc. That makes them reorder with the Halant that is 280 * ccc=9. Just zero them, we don't need them in our Indic shaper. 281 */ 282 #define HB_MODIFIED_COMBINING_CLASS_CCC84 0 /* length mark */ 283 #define HB_MODIFIED_COMBINING_CLASS_CCC91 0 /* ai length mark */ 284 285 /* Thai 286 * 287 * Modify U+0E38 and U+0E39 (ccc=103) to be reordered before U+0E3A (ccc=9). 288 * Assign 3, which is unassigned otherwise. 289 * Uniscribe does this reordering too. 290 */ 291 #define HB_MODIFIED_COMBINING_CLASS_CCC103 3 /* sara u / sara uu */ 292 #define HB_MODIFIED_COMBINING_CLASS_CCC107 107 /* mai * */ 293 294 /* Lao */ 295 #define HB_MODIFIED_COMBINING_CLASS_CCC118 118 /* sign u / sign uu */ 296 #define HB_MODIFIED_COMBINING_CLASS_CCC122 122 /* mai * */ 297 298 /* Tibetan */ 299 #define HB_MODIFIED_COMBINING_CLASS_CCC129 129 /* sign aa */ 300 #define HB_MODIFIED_COMBINING_CLASS_CCC130 130 /* sign i */ 301 #define HB_MODIFIED_COMBINING_CLASS_CCC132 132 /* sign u */ 302 303 304 /* Misc */ 305 306 #define HB_UNICODE_GENERAL_CATEGORY_IS_MARK(gen_cat) \ 307 (FLAG (gen_cat) & \ 308 (FLAG (HB_UNICODE_GENERAL_CATEGORY_SPACING_MARK) | \ 309 FLAG (HB_UNICODE_GENERAL_CATEGORY_ENCLOSING_MARK) | \ 310 FLAG (HB_UNICODE_GENERAL_CATEGORY_NON_SPACING_MARK))) 311 312 313 #endif /* HB_UNICODE_PRIVATE_HH */ 314