Home | History | Annotate | Download | only in src
      1 /*
      2  * Copyright  2009  Red Hat, Inc.
      3  * Copyright  2011  Google, Inc.
      4  *
      5  *  This is part of HarfBuzz, a text shaping library.
      6  *
      7  * Permission is hereby granted, without written agreement and without
      8  * license or royalty fees, to use, copy, modify, and distribute this
      9  * software and its documentation for any purpose, provided that the
     10  * above copyright notice and the following two paragraphs appear in
     11  * all copies of this software.
     12  *
     13  * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
     14  * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
     15  * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
     16  * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
     17  * DAMAGE.
     18  *
     19  * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
     20  * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
     21  * FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
     22  * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
     23  * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
     24  *
     25  * Red Hat Author(s): Behdad Esfahbod
     26  * Google Author(s): Behdad Esfahbod
     27  */
     28 
     29 #include "hb-private.hh"
     30 
     31 #include "hb-glib.h"
     32 
     33 #include "hb-unicode-private.hh"
     34 
     35 
     36 #if !GLIB_CHECK_VERSION(2,29,14)
     37 static const hb_script_t
     38 glib_script_to_script[] =
     39 {
     40   HB_SCRIPT_COMMON,
     41   HB_SCRIPT_INHERITED,
     42   HB_SCRIPT_ARABIC,
     43   HB_SCRIPT_ARMENIAN,
     44   HB_SCRIPT_BENGALI,
     45   HB_SCRIPT_BOPOMOFO,
     46   HB_SCRIPT_CHEROKEE,
     47   HB_SCRIPT_COPTIC,
     48   HB_SCRIPT_CYRILLIC,
     49   HB_SCRIPT_DESERET,
     50   HB_SCRIPT_DEVANAGARI,
     51   HB_SCRIPT_ETHIOPIC,
     52   HB_SCRIPT_GEORGIAN,
     53   HB_SCRIPT_GOTHIC,
     54   HB_SCRIPT_GREEK,
     55   HB_SCRIPT_GUJARATI,
     56   HB_SCRIPT_GURMUKHI,
     57   HB_SCRIPT_HAN,
     58   HB_SCRIPT_HANGUL,
     59   HB_SCRIPT_HEBREW,
     60   HB_SCRIPT_HIRAGANA,
     61   HB_SCRIPT_KANNADA,
     62   HB_SCRIPT_KATAKANA,
     63   HB_SCRIPT_KHMER,
     64   HB_SCRIPT_LAO,
     65   HB_SCRIPT_LATIN,
     66   HB_SCRIPT_MALAYALAM,
     67   HB_SCRIPT_MONGOLIAN,
     68   HB_SCRIPT_MYANMAR,
     69   HB_SCRIPT_OGHAM,
     70   HB_SCRIPT_OLD_ITALIC,
     71   HB_SCRIPT_ORIYA,
     72   HB_SCRIPT_RUNIC,
     73   HB_SCRIPT_SINHALA,
     74   HB_SCRIPT_SYRIAC,
     75   HB_SCRIPT_TAMIL,
     76   HB_SCRIPT_TELUGU,
     77   HB_SCRIPT_THAANA,
     78   HB_SCRIPT_THAI,
     79   HB_SCRIPT_TIBETAN,
     80   HB_SCRIPT_CANADIAN_SYLLABICS,
     81   HB_SCRIPT_YI,
     82   HB_SCRIPT_TAGALOG,
     83   HB_SCRIPT_HANUNOO,
     84   HB_SCRIPT_BUHID,
     85   HB_SCRIPT_TAGBANWA,
     86 
     87   /* Unicode-4.0 additions */
     88   HB_SCRIPT_BRAILLE,
     89   HB_SCRIPT_CYPRIOT,
     90   HB_SCRIPT_LIMBU,
     91   HB_SCRIPT_OSMANYA,
     92   HB_SCRIPT_SHAVIAN,
     93   HB_SCRIPT_LINEAR_B,
     94   HB_SCRIPT_TAI_LE,
     95   HB_SCRIPT_UGARITIC,
     96 
     97   /* Unicode-4.1 additions */
     98   HB_SCRIPT_NEW_TAI_LUE,
     99   HB_SCRIPT_BUGINESE,
    100   HB_SCRIPT_GLAGOLITIC,
    101   HB_SCRIPT_TIFINAGH,
    102   HB_SCRIPT_SYLOTI_NAGRI,
    103   HB_SCRIPT_OLD_PERSIAN,
    104   HB_SCRIPT_KHAROSHTHI,
    105 
    106   /* Unicode-5.0 additions */
    107   HB_SCRIPT_UNKNOWN,
    108   HB_SCRIPT_BALINESE,
    109   HB_SCRIPT_CUNEIFORM,
    110   HB_SCRIPT_PHOENICIAN,
    111   HB_SCRIPT_PHAGS_PA,
    112   HB_SCRIPT_NKO,
    113 
    114   /* Unicode-5.1 additions */
    115   HB_SCRIPT_KAYAH_LI,
    116   HB_SCRIPT_LEPCHA,
    117   HB_SCRIPT_REJANG,
    118   HB_SCRIPT_SUNDANESE,
    119   HB_SCRIPT_SAURASHTRA,
    120   HB_SCRIPT_CHAM,
    121   HB_SCRIPT_OL_CHIKI,
    122   HB_SCRIPT_VAI,
    123   HB_SCRIPT_CARIAN,
    124   HB_SCRIPT_LYCIAN,
    125   HB_SCRIPT_LYDIAN,
    126 
    127   /* Unicode-5.2 additions */
    128   HB_SCRIPT_AVESTAN,
    129   HB_SCRIPT_BAMUM,
    130   HB_SCRIPT_EGYPTIAN_HIEROGLYPHS,
    131   HB_SCRIPT_IMPERIAL_ARAMAIC,
    132   HB_SCRIPT_INSCRIPTIONAL_PAHLAVI,
    133   HB_SCRIPT_INSCRIPTIONAL_PARTHIAN,
    134   HB_SCRIPT_JAVANESE,
    135   HB_SCRIPT_KAITHI,
    136   HB_SCRIPT_TAI_THAM,
    137   HB_SCRIPT_LISU,
    138   HB_SCRIPT_MEETEI_MAYEK,
    139   HB_SCRIPT_OLD_SOUTH_ARABIAN,
    140   HB_SCRIPT_OLD_TURKIC,
    141   HB_SCRIPT_SAMARITAN,
    142   HB_SCRIPT_TAI_VIET,
    143 
    144   /* Unicode-6.0 additions */
    145   HB_SCRIPT_BATAK,
    146   HB_SCRIPT_BRAHMI,
    147   HB_SCRIPT_MANDAIC,
    148 
    149   /* Unicode-6.1 additions */
    150   HB_SCRIPT_CHAKMA,
    151   HB_SCRIPT_MEROITIC_CURSIVE,
    152   HB_SCRIPT_MEROITIC_HIEROGLYPHS,
    153   HB_SCRIPT_MIAO,
    154   HB_SCRIPT_SHARADA,
    155   HB_SCRIPT_SORA_SOMPENG,
    156   HB_SCRIPT_TAKRI
    157 };
    158 #endif
    159 
    160 hb_script_t
    161 hb_glib_script_to_script (GUnicodeScript script)
    162 {
    163 #if GLIB_CHECK_VERSION(2,29,14)
    164   return (hb_script_t) g_unicode_script_to_iso15924 (script);
    165 #else
    166   if (likely ((unsigned int) script < ARRAY_LENGTH (glib_script_to_script)))
    167     return glib_script_to_script[script];
    168 
    169   if (unlikely (script == G_UNICODE_SCRIPT_INVALID_CODE))
    170     return HB_SCRIPT_INVALID;
    171 
    172   return HB_SCRIPT_UNKNOWN;
    173 #endif
    174 }
    175 
    176 GUnicodeScript
    177 hb_glib_script_from_script (hb_script_t script)
    178 {
    179 #if GLIB_CHECK_VERSION(2,29,14)
    180   return g_unicode_script_from_iso15924 (script);
    181 #else
    182   unsigned int count = ARRAY_LENGTH (glib_script_to_script);
    183   for (unsigned int i = 0; i < count; i++)
    184     if (glib_script_to_script[i] == script)
    185       return (GUnicodeScript) i;
    186 
    187   if (unlikely (script == HB_SCRIPT_INVALID))
    188     return G_UNICODE_SCRIPT_INVALID_CODE;
    189 
    190   return G_UNICODE_SCRIPT_UNKNOWN;
    191 #endif
    192 }
    193 
    194 
    195 static hb_unicode_combining_class_t
    196 hb_glib_unicode_combining_class (hb_unicode_funcs_t *ufuncs HB_UNUSED,
    197 				 hb_codepoint_t      unicode,
    198 				 void               *user_data HB_UNUSED)
    199 
    200 {
    201   return (hb_unicode_combining_class_t) g_unichar_combining_class (unicode);
    202 }
    203 
    204 static unsigned int
    205 hb_glib_unicode_eastasian_width (hb_unicode_funcs_t *ufuncs HB_UNUSED,
    206 				 hb_codepoint_t      unicode,
    207 				 void               *user_data HB_UNUSED)
    208 {
    209   return g_unichar_iswide (unicode) ? 2 : 1;
    210 }
    211 
    212 static hb_unicode_general_category_t
    213 hb_glib_unicode_general_category (hb_unicode_funcs_t *ufuncs HB_UNUSED,
    214 				  hb_codepoint_t      unicode,
    215 				  void               *user_data HB_UNUSED)
    216 
    217 {
    218   /* hb_unicode_general_category_t and GUnicodeType are identical */
    219   return (hb_unicode_general_category_t) g_unichar_type (unicode);
    220 }
    221 
    222 static hb_codepoint_t
    223 hb_glib_unicode_mirroring (hb_unicode_funcs_t *ufuncs HB_UNUSED,
    224 			   hb_codepoint_t      unicode,
    225 			   void               *user_data HB_UNUSED)
    226 {
    227   g_unichar_get_mirror_char (unicode, &unicode);
    228   return unicode;
    229 }
    230 
    231 static hb_script_t
    232 hb_glib_unicode_script (hb_unicode_funcs_t *ufuncs HB_UNUSED,
    233 			hb_codepoint_t      unicode,
    234 			void               *user_data HB_UNUSED)
    235 {
    236   return hb_glib_script_to_script (g_unichar_get_script (unicode));
    237 }
    238 
    239 static hb_bool_t
    240 hb_glib_unicode_compose (hb_unicode_funcs_t *ufuncs HB_UNUSED,
    241 			 hb_codepoint_t      a,
    242 			 hb_codepoint_t      b,
    243 			 hb_codepoint_t     *ab,
    244 			 void               *user_data HB_UNUSED)
    245 {
    246 #if GLIB_CHECK_VERSION(2,29,12)
    247   return g_unichar_compose (a, b, ab);
    248 #endif
    249 
    250   /* We don't ifdef-out the fallback code such that compiler always
    251    * sees it and makes sure it's compilable. */
    252 
    253   gchar utf8[12];
    254   gchar *normalized;
    255   int len;
    256   hb_bool_t ret;
    257 
    258   len = g_unichar_to_utf8 (a, utf8);
    259   len += g_unichar_to_utf8 (b, utf8 + len);
    260   normalized = g_utf8_normalize (utf8, len, G_NORMALIZE_NFC);
    261   len = g_utf8_strlen (normalized, -1);
    262   if (unlikely (!len))
    263     return false;
    264 
    265   if (len == 1) {
    266     *ab = g_utf8_get_char (normalized);
    267     ret = true;
    268   } else {
    269     ret = false;
    270   }
    271 
    272   g_free (normalized);
    273   return ret;
    274 }
    275 
    276 static hb_bool_t
    277 hb_glib_unicode_decompose (hb_unicode_funcs_t *ufuncs HB_UNUSED,
    278 			   hb_codepoint_t      ab,
    279 			   hb_codepoint_t     *a,
    280 			   hb_codepoint_t     *b,
    281 			   void               *user_data HB_UNUSED)
    282 {
    283 #if GLIB_CHECK_VERSION(2,29,12)
    284   return g_unichar_decompose (ab, a, b);
    285 #endif
    286 
    287   /* We don't ifdef-out the fallback code such that compiler always
    288    * sees it and makes sure it's compilable. */
    289 
    290   gchar utf8[6];
    291   gchar *normalized;
    292   int len;
    293   hb_bool_t ret;
    294 
    295   len = g_unichar_to_utf8 (ab, utf8);
    296   normalized = g_utf8_normalize (utf8, len, G_NORMALIZE_NFD);
    297   len = g_utf8_strlen (normalized, -1);
    298   if (unlikely (!len))
    299     return false;
    300 
    301   if (len == 1) {
    302     *a = g_utf8_get_char (normalized);
    303     *b = 0;
    304     ret = *a != ab;
    305   } else if (len == 2) {
    306     *a = g_utf8_get_char (normalized);
    307     *b = g_utf8_get_char (g_utf8_next_char (normalized));
    308     /* Here's the ugly part: if ab decomposes to a single character and
    309      * that character decomposes again, we have to detect that and undo
    310      * the second part :-(. */
    311     gchar *recomposed = g_utf8_normalize (normalized, -1, G_NORMALIZE_NFC);
    312     hb_codepoint_t c = g_utf8_get_char (recomposed);
    313     if (c != ab && c != *a) {
    314       *a = c;
    315       *b = 0;
    316     }
    317     g_free (recomposed);
    318     ret = true;
    319   } else {
    320     /* If decomposed to more than two characters, take the last one,
    321      * and recompose the rest to get the first component. */
    322     gchar *end = g_utf8_offset_to_pointer (normalized, len - 1);
    323     gchar *recomposed;
    324     *b = g_utf8_get_char (end);
    325     recomposed = g_utf8_normalize (normalized, end - normalized, G_NORMALIZE_NFC);
    326     /* We expect that recomposed has exactly one character now. */
    327     *a = g_utf8_get_char (recomposed);
    328     g_free (recomposed);
    329     ret = true;
    330   }
    331 
    332   g_free (normalized);
    333   return ret;
    334 }
    335 
    336 static unsigned int
    337 hb_glib_unicode_decompose_compatibility (hb_unicode_funcs_t *ufuncs HB_UNUSED,
    338 					 hb_codepoint_t      u,
    339 					 hb_codepoint_t     *decomposed,
    340 					 void               *user_data HB_UNUSED)
    341 {
    342 #if GLIB_CHECK_VERSION(2,29,12)
    343   return g_unichar_fully_decompose (u, TRUE, decomposed, HB_UNICODE_MAX_DECOMPOSITION_LEN);
    344 #endif
    345 
    346   /* If the user doesn't have GLib >= 2.29.12 we have to perform
    347    * a round trip to UTF-8 and the associated memory management dance. */
    348   gchar utf8[6];
    349   gchar *utf8_decomposed, *c;
    350   gsize utf8_len, utf8_decomposed_len, i;
    351 
    352   /* Convert @u to UTF-8 and normalise it in NFKD mode. This performs the compatibility decomposition. */
    353   utf8_len = g_unichar_to_utf8 (u, utf8);
    354   utf8_decomposed = g_utf8_normalize (utf8, utf8_len, G_NORMALIZE_NFKD);
    355   utf8_decomposed_len = g_utf8_strlen (utf8_decomposed, -1);
    356 
    357   assert (utf8_decomposed_len <= HB_UNICODE_MAX_DECOMPOSITION_LEN);
    358 
    359   for (i = 0, c = utf8_decomposed; i < utf8_decomposed_len; i++, c = g_utf8_next_char (c))
    360     *decomposed++ = g_utf8_get_char (c);
    361 
    362   g_free (utf8_decomposed);
    363 
    364   return utf8_decomposed_len;
    365 }
    366 
    367 hb_unicode_funcs_t *
    368 hb_glib_get_unicode_funcs (void)
    369 {
    370   static const hb_unicode_funcs_t _hb_glib_unicode_funcs = {
    371     HB_OBJECT_HEADER_STATIC,
    372 
    373     NULL, /* parent */
    374     true, /* immutable */
    375     {
    376 #define HB_UNICODE_FUNC_IMPLEMENT(name) hb_glib_unicode_##name,
    377       HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
    378 #undef HB_UNICODE_FUNC_IMPLEMENT
    379     }
    380   };
    381 
    382   return const_cast<hb_unicode_funcs_t *> (&_hb_glib_unicode_funcs);
    383 }
    384 
    385