1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/gfx/harfbuzz/src/hb-unicode-private.hh Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,318 @@
1.4 +/*
1.5 + * Copyright © 2009 Red Hat, Inc.
1.6 + * Copyright © 2011 Codethink Limited
1.7 + * Copyright © 2010,2011,2012 Google, Inc.
1.8 + *
1.9 + * This is part of HarfBuzz, a text shaping library.
1.10 + *
1.11 + * Permission is hereby granted, without written agreement and without
1.12 + * license or royalty fees, to use, copy, modify, and distribute this
1.13 + * software and its documentation for any purpose, provided that the
1.14 + * above copyright notice and the following two paragraphs appear in
1.15 + * all copies of this software.
1.16 + *
1.17 + * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
1.18 + * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
1.19 + * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
1.20 + * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
1.21 + * DAMAGE.
1.22 + *
1.23 + * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
1.24 + * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
1.25 + * FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS
1.26 + * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
1.27 + * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
1.28 + *
1.29 + * Red Hat Author(s): Behdad Esfahbod
1.30 + * Codethink Author(s): Ryan Lortie
1.31 + * Google Author(s): Behdad Esfahbod
1.32 + */
1.33 +
1.34 +#ifndef HB_UNICODE_PRIVATE_HH
1.35 +#define HB_UNICODE_PRIVATE_HH
1.36 +
1.37 +#include "hb-private.hh"
1.38 +#include "hb-object-private.hh"
1.39 +
1.40 +
1.41 +extern HB_INTERNAL const uint8_t _hb_modified_combining_class[256];
1.42 +
1.43 +/*
1.44 + * hb_unicode_funcs_t
1.45 + */
1.46 +
1.47 +#define HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS \
1.48 + HB_UNICODE_FUNC_IMPLEMENT (combining_class) \
1.49 + HB_UNICODE_FUNC_IMPLEMENT (eastasian_width) \
1.50 + HB_UNICODE_FUNC_IMPLEMENT (general_category) \
1.51 + HB_UNICODE_FUNC_IMPLEMENT (mirroring) \
1.52 + HB_UNICODE_FUNC_IMPLEMENT (script) \
1.53 + HB_UNICODE_FUNC_IMPLEMENT (compose) \
1.54 + HB_UNICODE_FUNC_IMPLEMENT (decompose) \
1.55 + HB_UNICODE_FUNC_IMPLEMENT (decompose_compatibility) \
1.56 + /* ^--- Add new callbacks here */
1.57 +
1.58 +/* Simple callbacks are those taking a hb_codepoint_t and returning a hb_codepoint_t */
1.59 +#define HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE \
1.60 + HB_UNICODE_FUNC_IMPLEMENT (hb_unicode_combining_class_t, combining_class) \
1.61 + HB_UNICODE_FUNC_IMPLEMENT (unsigned int, eastasian_width) \
1.62 + HB_UNICODE_FUNC_IMPLEMENT (hb_unicode_general_category_t, general_category) \
1.63 + HB_UNICODE_FUNC_IMPLEMENT (hb_codepoint_t, mirroring) \
1.64 + HB_UNICODE_FUNC_IMPLEMENT (hb_script_t, script) \
1.65 + /* ^--- Add new simple callbacks here */
1.66 +
1.67 +struct hb_unicode_funcs_t {
1.68 + hb_object_header_t header;
1.69 + ASSERT_POD ();
1.70 +
1.71 + hb_unicode_funcs_t *parent;
1.72 +
1.73 + bool immutable;
1.74 +
1.75 +#define HB_UNICODE_FUNC_IMPLEMENT(return_type, name) \
1.76 + inline return_type name (hb_codepoint_t unicode) { return func.name (this, unicode, user_data.name); }
1.77 +HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE
1.78 +#undef HB_UNICODE_FUNC_IMPLEMENT
1.79 +
1.80 + inline hb_bool_t compose (hb_codepoint_t a, hb_codepoint_t b,
1.81 + hb_codepoint_t *ab)
1.82 + {
1.83 + *ab = 0;
1.84 + if (unlikely (!a || !b)) return false;
1.85 + return func.compose (this, a, b, ab, user_data.compose);
1.86 + }
1.87 +
1.88 + inline hb_bool_t decompose (hb_codepoint_t ab,
1.89 + hb_codepoint_t *a, hb_codepoint_t *b)
1.90 + {
1.91 + *a = ab; *b = 0;
1.92 + return func.decompose (this, ab, a, b, user_data.decompose);
1.93 + }
1.94 +
1.95 + inline unsigned int decompose_compatibility (hb_codepoint_t u,
1.96 + hb_codepoint_t *decomposed)
1.97 + {
1.98 + unsigned int ret = func.decompose_compatibility (this, u, decomposed, user_data.decompose_compatibility);
1.99 + if (ret == 1 && u == decomposed[0]) {
1.100 + decomposed[0] = 0;
1.101 + return 0;
1.102 + }
1.103 + decomposed[ret] = 0;
1.104 + return ret;
1.105 + }
1.106 +
1.107 +
1.108 + unsigned int
1.109 + modified_combining_class (hb_codepoint_t unicode)
1.110 + {
1.111 + /* XXX This hack belongs to the Myanmar shaper. */
1.112 + if (unlikely (unicode == 0x1037)) unicode = 0x103A;
1.113 +
1.114 + /* XXX This hack belongs to the SEA shaper (for Tai Tham):
1.115 + * Reorder SAKOT to ensure it comes after any tone marks. */
1.116 + if (unlikely (unicode == 0x1A60)) return 254;
1.117 +
1.118 + return _hb_modified_combining_class[combining_class (unicode)];
1.119 + }
1.120 +
1.121 + inline hb_bool_t
1.122 + is_variation_selector (hb_codepoint_t unicode)
1.123 + {
1.124 + return unlikely (hb_in_ranges<hb_codepoint_t> (unicode,
1.125 + 0x180B, 0x180D, /* MONGOLIAN FREE VARIATION SELECTOR ONE..THREE */
1.126 + 0xFE00, 0xFE0F, /* VARIATION SELECTOR-1..16 */
1.127 + 0xE0100, 0xE01EF)); /* VARIATION SELECTOR-17..256 */
1.128 + }
1.129 +
1.130 + /* Default_Ignorable codepoints:
1.131 + *
1.132 + * Note that as of Oct 2012 (Unicode 6.2), U+180E MONGOLIAN VOWEL SEPARATOR
1.133 + * is NOT Default_Ignorable, but it really behaves in a way that it should
1.134 + * be. That has been reported to the Unicode Technical Committee for
1.135 + * consideration. As such, we include it here, since Uniscribe removes it.
1.136 + * It *is* in Unicode 6.3 however. U+061C ARABIC LETTER MARK from Unicode
1.137 + * 6.3 is also added manually. The new Unicode 6.3 bidi formatting
1.138 + * characters are encoded in a block that was Default_Ignorable already.
1.139 + *
1.140 + * Note: While U+115F, U+1160, U+3164 and U+FFA0 are Default_Ignorable,
1.141 + * we do NOT want to hide them, as the way Uniscribe has implemented them
1.142 + * is with regular spacing glyphs, and that's the way fonts are made to work.
1.143 + * As such, we make exceptions for those four.
1.144 + *
1.145 + * Gathered from:
1.146 + * http://unicode.org/cldr/utility/list-unicodeset.jsp?a=[:DI:]&abb=on&ucd=on&esc=on
1.147 + *
1.148 + * Last updated to the page with the following versions:
1.149 + * Version 3.6; ICU version: 50.0.1.0; Unicode version: 6.1.0.0
1.150 + *
1.151 + * 4,167 Code Points
1.152 + *
1.153 + * [\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]
1.154 + *
1.155 + * 00AD ;SOFT HYPHEN
1.156 + * 034F ;COMBINING GRAPHEME JOINER
1.157 + * #115F ;HANGUL CHOSEONG FILLER
1.158 + * #1160 ;HANGUL JUNGSEONG FILLER
1.159 + * 17B4 ;KHMER VOWEL INHERENT AQ
1.160 + * 17B5 ;KHMER VOWEL INHERENT AA
1.161 + * 180B..180D ;MONGOLIAN FREE VARIATION SELECTOR THREE
1.162 + * 200B..200F ;RIGHT-TO-LEFT MARK
1.163 + * 202A..202E ;RIGHT-TO-LEFT OVERRIDE
1.164 + * 2060..206F ;NOMINAL DIGIT SHAPES
1.165 + * #3164 ;HANGUL FILLER
1.166 + * FE00..FE0F ;VARIATION SELECTOR-16
1.167 + * FEFF ;ZERO WIDTH NO-BREAK SPACE
1.168 + * #FFA0 ;HALFWIDTH HANGUL FILLER
1.169 + * FFF0..FFF8 ;<unassigned-FFF8>
1.170 + * 1D173..1D17A ;MUSICAL SYMBOL END PHRASE
1.171 + * E0000..E0FFF ;<unassigned-E0FFF>
1.172 + */
1.173 + inline hb_bool_t
1.174 + is_default_ignorable (hb_codepoint_t ch)
1.175 + {
1.176 + hb_codepoint_t plane = ch >> 16;
1.177 + if (likely (plane == 0))
1.178 + {
1.179 + /* BMP */
1.180 + hb_codepoint_t page = ch >> 8;
1.181 + switch (page) {
1.182 + case 0x00: return unlikely (ch == 0x00AD);
1.183 + case 0x03: return unlikely (ch == 0x034F);
1.184 + case 0x06: return unlikely (ch == 0x061C);
1.185 + case 0x17: return hb_in_range<hb_codepoint_t> (ch, 0x17B4, 0x17B5);
1.186 + case 0x18: return hb_in_range<hb_codepoint_t> (ch, 0x180B, 0x180E);
1.187 + case 0x20: return hb_in_ranges<hb_codepoint_t> (ch, 0x200B, 0x200F,
1.188 + 0x202A, 0x202E,
1.189 + 0x2060, 0x206F);
1.190 + case 0xFE: return hb_in_range<hb_codepoint_t> (ch, 0xFE00, 0xFE0F) || ch == 0xFEFF;
1.191 + case 0xFF: return hb_in_range<hb_codepoint_t> (ch, 0xFFF0, 0xFFF8);
1.192 + default: return false;
1.193 + }
1.194 + }
1.195 + else
1.196 + {
1.197 + /* Other planes */
1.198 + switch (plane) {
1.199 + case 0x01: return hb_in_range<hb_codepoint_t> (ch, 0x0001D173, 0x0001D17A);
1.200 + case 0x0E: return hb_in_range<hb_codepoint_t> (ch, 0x000E0000, 0x000E0FFF);
1.201 + default: return false;
1.202 + }
1.203 + }
1.204 + }
1.205 +
1.206 +
1.207 + struct {
1.208 +#define HB_UNICODE_FUNC_IMPLEMENT(name) hb_unicode_##name##_func_t name;
1.209 + HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
1.210 +#undef HB_UNICODE_FUNC_IMPLEMENT
1.211 + } func;
1.212 +
1.213 + struct {
1.214 +#define HB_UNICODE_FUNC_IMPLEMENT(name) void *name;
1.215 + HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
1.216 +#undef HB_UNICODE_FUNC_IMPLEMENT
1.217 + } user_data;
1.218 +
1.219 + struct {
1.220 +#define HB_UNICODE_FUNC_IMPLEMENT(name) hb_destroy_func_t name;
1.221 + HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
1.222 +#undef HB_UNICODE_FUNC_IMPLEMENT
1.223 + } destroy;
1.224 +};
1.225 +
1.226 +
1.227 +extern HB_INTERNAL const hb_unicode_funcs_t _hb_unicode_funcs_nil;
1.228 +
1.229 +
1.230 +/* Modified combining marks */
1.231 +
1.232 +/* Hebrew
1.233 + *
1.234 + * We permute the "fixed-position" classes 10-26 into the order
1.235 + * described in the SBL Hebrew manual:
1.236 + *
1.237 + * http://www.sbl-site.org/Fonts/SBLHebrewUserManual1.5x.pdf
1.238 + *
1.239 + * (as recommended by:
1.240 + * http://forum.fontlab.com/archive-old-microsoft-volt-group/vista-and-diacritic-ordering-t6751.0.html)
1.241 + *
1.242 + * More details here:
1.243 + * https://bugzilla.mozilla.org/show_bug.cgi?id=662055
1.244 + */
1.245 +#define HB_MODIFIED_COMBINING_CLASS_CCC10 22 /* sheva */
1.246 +#define HB_MODIFIED_COMBINING_CLASS_CCC11 15 /* hataf segol */
1.247 +#define HB_MODIFIED_COMBINING_CLASS_CCC12 16 /* hataf patah */
1.248 +#define HB_MODIFIED_COMBINING_CLASS_CCC13 17 /* hataf qamats */
1.249 +#define HB_MODIFIED_COMBINING_CLASS_CCC14 23 /* hiriq */
1.250 +#define HB_MODIFIED_COMBINING_CLASS_CCC15 18 /* tsere */
1.251 +#define HB_MODIFIED_COMBINING_CLASS_CCC16 19 /* segol */
1.252 +#define HB_MODIFIED_COMBINING_CLASS_CCC17 20 /* patah */
1.253 +#define HB_MODIFIED_COMBINING_CLASS_CCC18 21 /* qamats */
1.254 +#define HB_MODIFIED_COMBINING_CLASS_CCC19 14 /* holam */
1.255 +#define HB_MODIFIED_COMBINING_CLASS_CCC20 24 /* qubuts */
1.256 +#define HB_MODIFIED_COMBINING_CLASS_CCC21 12 /* dagesh */
1.257 +#define HB_MODIFIED_COMBINING_CLASS_CCC22 25 /* meteg */
1.258 +#define HB_MODIFIED_COMBINING_CLASS_CCC23 13 /* rafe */
1.259 +#define HB_MODIFIED_COMBINING_CLASS_CCC24 10 /* shin dot */
1.260 +#define HB_MODIFIED_COMBINING_CLASS_CCC25 11 /* sin dot */
1.261 +#define HB_MODIFIED_COMBINING_CLASS_CCC26 26 /* point varika */
1.262 +
1.263 +/*
1.264 + * Arabic
1.265 + *
1.266 + * Modify to move Shadda (ccc=33) before other marks. See:
1.267 + * http://unicode.org/faq/normalization.html#8
1.268 + * http://unicode.org/faq/normalization.html#9
1.269 + */
1.270 +#define HB_MODIFIED_COMBINING_CLASS_CCC27 28 /* fathatan */
1.271 +#define HB_MODIFIED_COMBINING_CLASS_CCC28 29 /* dammatan */
1.272 +#define HB_MODIFIED_COMBINING_CLASS_CCC29 30 /* kasratan */
1.273 +#define HB_MODIFIED_COMBINING_CLASS_CCC30 31 /* fatha */
1.274 +#define HB_MODIFIED_COMBINING_CLASS_CCC31 32 /* damma */
1.275 +#define HB_MODIFIED_COMBINING_CLASS_CCC32 33 /* kasra */
1.276 +#define HB_MODIFIED_COMBINING_CLASS_CCC33 27 /* shadda */
1.277 +#define HB_MODIFIED_COMBINING_CLASS_CCC34 34 /* sukun */
1.278 +#define HB_MODIFIED_COMBINING_CLASS_CCC35 35 /* superscript alef */
1.279 +
1.280 +/* Syriac */
1.281 +#define HB_MODIFIED_COMBINING_CLASS_CCC36 36 /* superscript alaph */
1.282 +
1.283 +/* Telugu
1.284 + *
1.285 + * Modify Telugu length marks (ccc=84, ccc=91).
1.286 + * These are the only matras in the main Indic scripts range that have
1.287 + * a non-zero ccc. That makes them reorder with the Halant that is
1.288 + * ccc=9. Just zero them, we don't need them in our Indic shaper.
1.289 + */
1.290 +#define HB_MODIFIED_COMBINING_CLASS_CCC84 0 /* length mark */
1.291 +#define HB_MODIFIED_COMBINING_CLASS_CCC91 0 /* ai length mark */
1.292 +
1.293 +/* Thai
1.294 + *
1.295 + * Modify U+0E38 and U+0E39 (ccc=103) to be reordered before U+0E3A (ccc=9).
1.296 + * Assign 3, which is unassigned otherwise.
1.297 + * Uniscribe does this reordering too.
1.298 + */
1.299 +#define HB_MODIFIED_COMBINING_CLASS_CCC103 3 /* sara u / sara uu */
1.300 +#define HB_MODIFIED_COMBINING_CLASS_CCC107 107 /* mai * */
1.301 +
1.302 +/* Lao */
1.303 +#define HB_MODIFIED_COMBINING_CLASS_CCC118 118 /* sign u / sign uu */
1.304 +#define HB_MODIFIED_COMBINING_CLASS_CCC122 122 /* mai * */
1.305 +
1.306 +/* Tibetan */
1.307 +#define HB_MODIFIED_COMBINING_CLASS_CCC129 129 /* sign aa */
1.308 +#define HB_MODIFIED_COMBINING_CLASS_CCC130 130 /* sign i */
1.309 +#define HB_MODIFIED_COMBINING_CLASS_CCC132 132 /* sign u */
1.310 +
1.311 +
1.312 +/* Misc */
1.313 +
1.314 +#define HB_UNICODE_GENERAL_CATEGORY_IS_MARK(gen_cat) \
1.315 + (FLAG (gen_cat) & \
1.316 + (FLAG (HB_UNICODE_GENERAL_CATEGORY_SPACING_MARK) | \
1.317 + FLAG (HB_UNICODE_GENERAL_CATEGORY_ENCLOSING_MARK) | \
1.318 + FLAG (HB_UNICODE_GENERAL_CATEGORY_NON_SPACING_MARK)))
1.319 +
1.320 +
1.321 +#endif /* HB_UNICODE_PRIVATE_HH */
