1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/gfx/harfbuzz/src/hb-glib.cc Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,384 @@
1.4 +/*
1.5 + * Copyright © 2009 Red Hat, Inc.
1.6 + * Copyright © 2011 Google, Inc.
1.7 + *
1.8 + * This is part of HarfBuzz, a text shaping library.
1.9 + *
1.10 + * Permission is hereby granted, without written agreement and without
1.11 + * license or royalty fees, to use, copy, modify, and distribute this
1.12 + * software and its documentation for any purpose, provided that the
1.13 + * above copyright notice and the following two paragraphs appear in
1.14 + * all copies of this software.
1.15 + *
1.16 + * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
1.17 + * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
1.18 + * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
1.19 + * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
1.20 + * DAMAGE.
1.21 + *
1.22 + * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
1.23 + * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
1.24 + * FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS
1.25 + * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
1.26 + * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
1.27 + *
1.28 + * Red Hat Author(s): Behdad Esfahbod
1.29 + * Google Author(s): Behdad Esfahbod
1.30 + */
1.31 +
1.32 +#include "hb-private.hh"
1.33 +
1.34 +#include "hb-glib.h"
1.35 +
1.36 +#include "hb-unicode-private.hh"
1.37 +
1.38 +
1.39 +#if !GLIB_CHECK_VERSION(2,29,14)
1.40 +static const hb_script_t
1.41 +glib_script_to_script[] =
1.42 +{
1.43 + HB_SCRIPT_COMMON,
1.44 + HB_SCRIPT_INHERITED,
1.45 + HB_SCRIPT_ARABIC,
1.46 + HB_SCRIPT_ARMENIAN,
1.47 + HB_SCRIPT_BENGALI,
1.48 + HB_SCRIPT_BOPOMOFO,
1.49 + HB_SCRIPT_CHEROKEE,
1.50 + HB_SCRIPT_COPTIC,
1.51 + HB_SCRIPT_CYRILLIC,
1.52 + HB_SCRIPT_DESERET,
1.53 + HB_SCRIPT_DEVANAGARI,
1.54 + HB_SCRIPT_ETHIOPIC,
1.55 + HB_SCRIPT_GEORGIAN,
1.56 + HB_SCRIPT_GOTHIC,
1.57 + HB_SCRIPT_GREEK,
1.58 + HB_SCRIPT_GUJARATI,
1.59 + HB_SCRIPT_GURMUKHI,
1.60 + HB_SCRIPT_HAN,
1.61 + HB_SCRIPT_HANGUL,
1.62 + HB_SCRIPT_HEBREW,
1.63 + HB_SCRIPT_HIRAGANA,
1.64 + HB_SCRIPT_KANNADA,
1.65 + HB_SCRIPT_KATAKANA,
1.66 + HB_SCRIPT_KHMER,
1.67 + HB_SCRIPT_LAO,
1.68 + HB_SCRIPT_LATIN,
1.69 + HB_SCRIPT_MALAYALAM,
1.70 + HB_SCRIPT_MONGOLIAN,
1.71 + HB_SCRIPT_MYANMAR,
1.72 + HB_SCRIPT_OGHAM,
1.73 + HB_SCRIPT_OLD_ITALIC,
1.74 + HB_SCRIPT_ORIYA,
1.75 + HB_SCRIPT_RUNIC,
1.76 + HB_SCRIPT_SINHALA,
1.77 + HB_SCRIPT_SYRIAC,
1.78 + HB_SCRIPT_TAMIL,
1.79 + HB_SCRIPT_TELUGU,
1.80 + HB_SCRIPT_THAANA,
1.81 + HB_SCRIPT_THAI,
1.82 + HB_SCRIPT_TIBETAN,
1.83 + HB_SCRIPT_CANADIAN_SYLLABICS,
1.84 + HB_SCRIPT_YI,
1.85 + HB_SCRIPT_TAGALOG,
1.86 + HB_SCRIPT_HANUNOO,
1.87 + HB_SCRIPT_BUHID,
1.88 + HB_SCRIPT_TAGBANWA,
1.89 +
1.90 + /* Unicode-4.0 additions */
1.91 + HB_SCRIPT_BRAILLE,
1.92 + HB_SCRIPT_CYPRIOT,
1.93 + HB_SCRIPT_LIMBU,
1.94 + HB_SCRIPT_OSMANYA,
1.95 + HB_SCRIPT_SHAVIAN,
1.96 + HB_SCRIPT_LINEAR_B,
1.97 + HB_SCRIPT_TAI_LE,
1.98 + HB_SCRIPT_UGARITIC,
1.99 +
1.100 + /* Unicode-4.1 additions */
1.101 + HB_SCRIPT_NEW_TAI_LUE,
1.102 + HB_SCRIPT_BUGINESE,
1.103 + HB_SCRIPT_GLAGOLITIC,
1.104 + HB_SCRIPT_TIFINAGH,
1.105 + HB_SCRIPT_SYLOTI_NAGRI,
1.106 + HB_SCRIPT_OLD_PERSIAN,
1.107 + HB_SCRIPT_KHAROSHTHI,
1.108 +
1.109 + /* Unicode-5.0 additions */
1.110 + HB_SCRIPT_UNKNOWN,
1.111 + HB_SCRIPT_BALINESE,
1.112 + HB_SCRIPT_CUNEIFORM,
1.113 + HB_SCRIPT_PHOENICIAN,
1.114 + HB_SCRIPT_PHAGS_PA,
1.115 + HB_SCRIPT_NKO,
1.116 +
1.117 + /* Unicode-5.1 additions */
1.118 + HB_SCRIPT_KAYAH_LI,
1.119 + HB_SCRIPT_LEPCHA,
1.120 + HB_SCRIPT_REJANG,
1.121 + HB_SCRIPT_SUNDANESE,
1.122 + HB_SCRIPT_SAURASHTRA,
1.123 + HB_SCRIPT_CHAM,
1.124 + HB_SCRIPT_OL_CHIKI,
1.125 + HB_SCRIPT_VAI,
1.126 + HB_SCRIPT_CARIAN,
1.127 + HB_SCRIPT_LYCIAN,
1.128 + HB_SCRIPT_LYDIAN,
1.129 +
1.130 + /* Unicode-5.2 additions */
1.131 + HB_SCRIPT_AVESTAN,
1.132 + HB_SCRIPT_BAMUM,
1.133 + HB_SCRIPT_EGYPTIAN_HIEROGLYPHS,
1.134 + HB_SCRIPT_IMPERIAL_ARAMAIC,
1.135 + HB_SCRIPT_INSCRIPTIONAL_PAHLAVI,
1.136 + HB_SCRIPT_INSCRIPTIONAL_PARTHIAN,
1.137 + HB_SCRIPT_JAVANESE,
1.138 + HB_SCRIPT_KAITHI,
1.139 + HB_SCRIPT_TAI_THAM,
1.140 + HB_SCRIPT_LISU,
1.141 + HB_SCRIPT_MEETEI_MAYEK,
1.142 + HB_SCRIPT_OLD_SOUTH_ARABIAN,
1.143 + HB_SCRIPT_OLD_TURKIC,
1.144 + HB_SCRIPT_SAMARITAN,
1.145 + HB_SCRIPT_TAI_VIET,
1.146 +
1.147 + /* Unicode-6.0 additions */
1.148 + HB_SCRIPT_BATAK,
1.149 + HB_SCRIPT_BRAHMI,
1.150 + HB_SCRIPT_MANDAIC,
1.151 +
1.152 + /* Unicode-6.1 additions */
1.153 + HB_SCRIPT_CHAKMA,
1.154 + HB_SCRIPT_MEROITIC_CURSIVE,
1.155 + HB_SCRIPT_MEROITIC_HIEROGLYPHS,
1.156 + HB_SCRIPT_MIAO,
1.157 + HB_SCRIPT_SHARADA,
1.158 + HB_SCRIPT_SORA_SOMPENG,
1.159 + HB_SCRIPT_TAKRI
1.160 +};
1.161 +#endif
1.162 +
1.163 +hb_script_t
1.164 +hb_glib_script_to_script (GUnicodeScript script)
1.165 +{
1.166 +#if GLIB_CHECK_VERSION(2,29,14)
1.167 + return (hb_script_t) g_unicode_script_to_iso15924 (script);
1.168 +#else
1.169 + if (likely ((unsigned int) script < ARRAY_LENGTH (glib_script_to_script)))
1.170 + return glib_script_to_script[script];
1.171 +
1.172 + if (unlikely (script == G_UNICODE_SCRIPT_INVALID_CODE))
1.173 + return HB_SCRIPT_INVALID;
1.174 +
1.175 + return HB_SCRIPT_UNKNOWN;
1.176 +#endif
1.177 +}
1.178 +
1.179 +GUnicodeScript
1.180 +hb_glib_script_from_script (hb_script_t script)
1.181 +{
1.182 +#if GLIB_CHECK_VERSION(2,29,14)
1.183 + return g_unicode_script_from_iso15924 (script);
1.184 +#else
1.185 + unsigned int count = ARRAY_LENGTH (glib_script_to_script);
1.186 + for (unsigned int i = 0; i < count; i++)
1.187 + if (glib_script_to_script[i] == script)
1.188 + return (GUnicodeScript) i;
1.189 +
1.190 + if (unlikely (script == HB_SCRIPT_INVALID))
1.191 + return G_UNICODE_SCRIPT_INVALID_CODE;
1.192 +
1.193 + return G_UNICODE_SCRIPT_UNKNOWN;
1.194 +#endif
1.195 +}
1.196 +
1.197 +
1.198 +static hb_unicode_combining_class_t
1.199 +hb_glib_unicode_combining_class (hb_unicode_funcs_t *ufuncs HB_UNUSED,
1.200 + hb_codepoint_t unicode,
1.201 + void *user_data HB_UNUSED)
1.202 +
1.203 +{
1.204 + return (hb_unicode_combining_class_t) g_unichar_combining_class (unicode);
1.205 +}
1.206 +
1.207 +static unsigned int
1.208 +hb_glib_unicode_eastasian_width (hb_unicode_funcs_t *ufuncs HB_UNUSED,
1.209 + hb_codepoint_t unicode,
1.210 + void *user_data HB_UNUSED)
1.211 +{
1.212 + return g_unichar_iswide (unicode) ? 2 : 1;
1.213 +}
1.214 +
1.215 +static hb_unicode_general_category_t
1.216 +hb_glib_unicode_general_category (hb_unicode_funcs_t *ufuncs HB_UNUSED,
1.217 + hb_codepoint_t unicode,
1.218 + void *user_data HB_UNUSED)
1.219 +
1.220 +{
1.221 + /* hb_unicode_general_category_t and GUnicodeType are identical */
1.222 + return (hb_unicode_general_category_t) g_unichar_type (unicode);
1.223 +}
1.224 +
1.225 +static hb_codepoint_t
1.226 +hb_glib_unicode_mirroring (hb_unicode_funcs_t *ufuncs HB_UNUSED,
1.227 + hb_codepoint_t unicode,
1.228 + void *user_data HB_UNUSED)
1.229 +{
1.230 + g_unichar_get_mirror_char (unicode, &unicode);
1.231 + return unicode;
1.232 +}
1.233 +
1.234 +static hb_script_t
1.235 +hb_glib_unicode_script (hb_unicode_funcs_t *ufuncs HB_UNUSED,
1.236 + hb_codepoint_t unicode,
1.237 + void *user_data HB_UNUSED)
1.238 +{
1.239 + return hb_glib_script_to_script (g_unichar_get_script (unicode));
1.240 +}
1.241 +
1.242 +static hb_bool_t
1.243 +hb_glib_unicode_compose (hb_unicode_funcs_t *ufuncs HB_UNUSED,
1.244 + hb_codepoint_t a,
1.245 + hb_codepoint_t b,
1.246 + hb_codepoint_t *ab,
1.247 + void *user_data HB_UNUSED)
1.248 +{
1.249 +#if GLIB_CHECK_VERSION(2,29,12)
1.250 + return g_unichar_compose (a, b, ab);
1.251 +#endif
1.252 +
1.253 + /* We don't ifdef-out the fallback code such that compiler always
1.254 + * sees it and makes sure it's compilable. */
1.255 +
1.256 + gchar utf8[12];
1.257 + gchar *normalized;
1.258 + int len;
1.259 + hb_bool_t ret;
1.260 +
1.261 + len = g_unichar_to_utf8 (a, utf8);
1.262 + len += g_unichar_to_utf8 (b, utf8 + len);
1.263 + normalized = g_utf8_normalize (utf8, len, G_NORMALIZE_NFC);
1.264 + len = g_utf8_strlen (normalized, -1);
1.265 + if (unlikely (!len))
1.266 + return false;
1.267 +
1.268 + if (len == 1) {
1.269 + *ab = g_utf8_get_char (normalized);
1.270 + ret = true;
1.271 + } else {
1.272 + ret = false;
1.273 + }
1.274 +
1.275 + g_free (normalized);
1.276 + return ret;
1.277 +}
1.278 +
1.279 +static hb_bool_t
1.280 +hb_glib_unicode_decompose (hb_unicode_funcs_t *ufuncs HB_UNUSED,
1.281 + hb_codepoint_t ab,
1.282 + hb_codepoint_t *a,
1.283 + hb_codepoint_t *b,
1.284 + void *user_data HB_UNUSED)
1.285 +{
1.286 +#if GLIB_CHECK_VERSION(2,29,12)
1.287 + return g_unichar_decompose (ab, a, b);
1.288 +#endif
1.289 +
1.290 + /* We don't ifdef-out the fallback code such that compiler always
1.291 + * sees it and makes sure it's compilable. */
1.292 +
1.293 + gchar utf8[6];
1.294 + gchar *normalized;
1.295 + int len;
1.296 + hb_bool_t ret;
1.297 +
1.298 + len = g_unichar_to_utf8 (ab, utf8);
1.299 + normalized = g_utf8_normalize (utf8, len, G_NORMALIZE_NFD);
1.300 + len = g_utf8_strlen (normalized, -1);
1.301 + if (unlikely (!len))
1.302 + return false;
1.303 +
1.304 + if (len == 1) {
1.305 + *a = g_utf8_get_char (normalized);
1.306 + *b = 0;
1.307 + ret = *a != ab;
1.308 + } else if (len == 2) {
1.309 + *a = g_utf8_get_char (normalized);
1.310 + *b = g_utf8_get_char (g_utf8_next_char (normalized));
1.311 + /* Here's the ugly part: if ab decomposes to a single character and
1.312 + * that character decomposes again, we have to detect that and undo
1.313 + * the second part :-(. */
1.314 + gchar *recomposed = g_utf8_normalize (normalized, -1, G_NORMALIZE_NFC);
1.315 + hb_codepoint_t c = g_utf8_get_char (recomposed);
1.316 + if (c != ab && c != *a) {
1.317 + *a = c;
1.318 + *b = 0;
1.319 + }
1.320 + g_free (recomposed);
1.321 + ret = true;
1.322 + } else {
1.323 + /* If decomposed to more than two characters, take the last one,
1.324 + * and recompose the rest to get the first component. */
1.325 + gchar *end = g_utf8_offset_to_pointer (normalized, len - 1);
1.326 + gchar *recomposed;
1.327 + *b = g_utf8_get_char (end);
1.328 + recomposed = g_utf8_normalize (normalized, end - normalized, G_NORMALIZE_NFC);
1.329 + /* We expect that recomposed has exactly one character now. */
1.330 + *a = g_utf8_get_char (recomposed);
1.331 + g_free (recomposed);
1.332 + ret = true;
1.333 + }
1.334 +
1.335 + g_free (normalized);
1.336 + return ret;
1.337 +}
1.338 +
1.339 +static unsigned int
1.340 +hb_glib_unicode_decompose_compatibility (hb_unicode_funcs_t *ufuncs HB_UNUSED,
1.341 + hb_codepoint_t u,
1.342 + hb_codepoint_t *decomposed,
1.343 + void *user_data HB_UNUSED)
1.344 +{
1.345 +#if GLIB_CHECK_VERSION(2,29,12)
1.346 + return g_unichar_fully_decompose (u, TRUE, decomposed, HB_UNICODE_MAX_DECOMPOSITION_LEN);
1.347 +#endif
1.348 +
1.349 + /* If the user doesn't have GLib >= 2.29.12 we have to perform
1.350 + * a round trip to UTF-8 and the associated memory management dance. */
1.351 + gchar utf8[6];
1.352 + gchar *utf8_decomposed, *c;
1.353 + gsize utf8_len, utf8_decomposed_len, i;
1.354 +
1.355 + /* Convert @u to UTF-8 and normalise it in NFKD mode. This performs the compatibility decomposition. */
1.356 + utf8_len = g_unichar_to_utf8 (u, utf8);
1.357 + utf8_decomposed = g_utf8_normalize (utf8, utf8_len, G_NORMALIZE_NFKD);
1.358 + utf8_decomposed_len = g_utf8_strlen (utf8_decomposed, -1);
1.359 +
1.360 + assert (utf8_decomposed_len <= HB_UNICODE_MAX_DECOMPOSITION_LEN);
1.361 +
1.362 + for (i = 0, c = utf8_decomposed; i < utf8_decomposed_len; i++, c = g_utf8_next_char (c))
1.363 + *decomposed++ = g_utf8_get_char (c);
1.364 +
1.365 + g_free (utf8_decomposed);
1.366 +
1.367 + return utf8_decomposed_len;
1.368 +}
1.369 +
1.370 +hb_unicode_funcs_t *
1.371 +hb_glib_get_unicode_funcs (void)
1.372 +{
1.373 + static const hb_unicode_funcs_t _hb_glib_unicode_funcs = {
1.374 + HB_OBJECT_HEADER_STATIC,
1.375 +
1.376 + NULL, /* parent */
1.377 + true, /* immutable */
1.378 + {
1.379 +#define HB_UNICODE_FUNC_IMPLEMENT(name) hb_glib_unicode_##name,
1.380 + HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
1.381 +#undef HB_UNICODE_FUNC_IMPLEMENT
1.382 + }
1.383 + };
1.384 +
1.385 + return const_cast<hb_unicode_funcs_t *> (&_hb_glib_unicode_funcs);
1.386 +}
1.387 +
