Home | History | Annotate | Download | only in src 
      1 /*
      2  * Copyright  2010,2012  Google, Inc.
      3  *
      4  *  This is part of HarfBuzz, a text shaping library.
      5  *
      6  * Permission is hereby granted, without written agreement and without
      7  * license or royalty fees, to use, copy, modify, and distribute this
      8  * software and its documentation for any purpose, provided that the
      9  * above copyright notice and the following two paragraphs appear in
     10  * all copies of this software.
     11  *
     12  * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
     13  * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
     14  * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
     15  * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
     16  * DAMAGE.
     17  *
     18  * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
     19  * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
     20  * FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
     21  * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
     22  * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
     23  *
     24  * Google Author(s): Behdad Esfahbod
     25  */
     26 
     27 #include "hb-ot-shape-complex-private.hh"
     28 
     29 
     30 static bool
     31 compose_hebrew (const hb_ot_shape_normalize_context_t *c,
     32 		hb_codepoint_t  a,
     33 		hb_codepoint_t  b,
     34 		hb_codepoint_t *ab)
     35 {
     36   /* Hebrew presentation-form shaping.
     37    * https://bugzilla.mozilla.org/show_bug.cgi?id=728866
     38    * Hebrew presentation forms with dagesh, for characters U+05D0..05EA;
     39    * Note that some letters do not have a dagesh presForm encoded.
     40    */
     41   static const hb_codepoint_t sDageshForms[0x05EAu - 0x05D0u + 1] = {
     42     0xFB30u, /* ALEF */
     43     0xFB31u, /* BET */
     44     0xFB32u, /* GIMEL */
     45     0xFB33u, /* DALET */
     46     0xFB34u, /* HE */
     47     0xFB35u, /* VAV */
     48     0xFB36u, /* ZAYIN */
     49     0x0000u, /* HET */
     50     0xFB38u, /* TET */
     51     0xFB39u, /* YOD */
     52     0xFB3Au, /* FINAL KAF */
     53     0xFB3Bu, /* KAF */
     54     0xFB3Cu, /* LAMED */
     55     0x0000u, /* FINAL MEM */
     56     0xFB3Eu, /* MEM */
     57     0x0000u, /* FINAL NUN */
     58     0xFB40u, /* NUN */
     59     0xFB41u, /* SAMEKH */
     60     0x0000u, /* AYIN */
     61     0xFB43u, /* FINAL PE */
     62     0xFB44u, /* PE */
     63     0x0000u, /* FINAL TSADI */
     64     0xFB46u, /* TSADI */
     65     0xFB47u, /* QOF */
     66     0xFB48u, /* RESH */
     67     0xFB49u, /* SHIN */
     68     0xFB4Au /* TAV */
     69   };
     70 
     71   bool found = c->unicode->compose (a, b, ab);
     72 
     73   if (!found && !c->plan->has_mark)
     74   {
     75       /* Special-case Hebrew presentation forms that are excluded from
     76        * standard normalization, but wanted for old fonts. */
     77       switch (b) {
     78       case 0x05B4u: /* HIRIQ */
     79 	  if (a == 0x05D9u) { /* YOD */
     80 	      *ab = 0xFB1Du;
     81 	      found = true;
     82 	  }
     83 	  break;
     84       case 0x05B7u: /* patah */
     85 	  if (a == 0x05F2u) { /* YIDDISH YOD YOD */
     86 	      *ab = 0xFB1Fu;
     87 	      found = true;
     88 	  } else if (a == 0x05D0u) { /* ALEF */
     89 	      *ab = 0xFB2Eu;
     90 	      found = true;
     91 	  }
     92 	  break;
     93       case 0x05B8u: /* QAMATS */
     94 	  if (a == 0x05D0u) { /* ALEF */
     95 	      *ab = 0xFB2Fu;
     96 	      found = true;
     97 	  }
     98 	  break;
     99       case 0x05B9u: /* HOLAM */
    100 	  if (a == 0x05D5u) { /* VAV */
    101 	      *ab = 0xFB4Bu;
    102 	      found = true;
    103 	  }
    104 	  break;
    105       case 0x05BCu: /* DAGESH */
    106 	  if (a >= 0x05D0u && a <= 0x05EAu) {
    107 	      *ab = sDageshForms[a - 0x05D0u];
    108 	      found = (*ab != 0);
    109 	  } else if (a == 0xFB2Au) { /* SHIN WITH SHIN DOT */
    110 	      *ab = 0xFB2Cu;
    111 	      found = true;
    112 	  } else if (a == 0xFB2Bu) { /* SHIN WITH SIN DOT */
    113 	      *ab = 0xFB2Du;
    114 	      found = true;
    115 	  }
    116 	  break;
    117       case 0x05BFu: /* RAFE */
    118 	  switch (a) {
    119 	  case 0x05D1u: /* BET */
    120 	      *ab = 0xFB4Cu;
    121 	      found = true;
    122 	      break;
    123 	  case 0x05DBu: /* KAF */
    124 	      *ab = 0xFB4Du;
    125 	      found = true;
    126 	      break;
    127 	  case 0x05E4u: /* PE */
    128 	      *ab = 0xFB4Eu;
    129 	      found = true;
    130 	      break;
    131 	  }
    132 	  break;
    133       case 0x05C1u: /* SHIN DOT */
    134 	  if (a == 0x05E9u) { /* SHIN */
    135 	      *ab = 0xFB2Au;
    136 	      found = true;
    137 	  } else if (a == 0xFB49u) { /* SHIN WITH DAGESH */
    138 	      *ab = 0xFB2Cu;
    139 	      found = true;
    140 	  }
    141 	  break;
    142       case 0x05C2u: /* SIN DOT */
    143 	  if (a == 0x05E9u) { /* SHIN */
    144 	      *ab = 0xFB2Bu;
    145 	      found = true;
    146 	  } else if (a == 0xFB49u) { /* SHIN WITH DAGESH */
    147 	      *ab = 0xFB2Du;
    148 	      found = true;
    149 	  }
    150 	  break;
    151       }
    152   }
    153 
    154   return found;
    155 }
    156 
    157 
    158 const hb_ot_complex_shaper_t _hb_ot_complex_shaper_hebrew =
    159 {
    160   "hebrew",
    161   NULL, /* collect_features */
    162   NULL, /* override_features */
    163   NULL, /* data_create */
    164   NULL, /* data_destroy */
    165   NULL, /* preprocess_text */
    166   HB_OT_SHAPE_NORMALIZATION_MODE_DEFAULT,
    167   NULL, /* decompose */
    168   compose_hebrew,
    169   NULL, /* setup_masks */
    170   HB_OT_SHAPE_ZERO_WIDTH_MARKS_BY_GDEF_LATE,
    171   true, /* fallback_position */
    172 };
    173