hb-unicode.c

   1 /*
   2 Port from c++ is protected by a GNU Lesser GPLv3
   3 Copyright © 2013 Sylvain BERTRAND <sylvain.bertrand@gmail.com>
   4                                   <sylware@legeek.net>
   5 */
   6 #include <stdlib.h>
   7
   8 #include "hb.h"
   9 #include "hb-private.h"
  10 #include "hb-atomic-private.h"
  11 #include "hb-unicode-private.h"
  12
  13 static hb_unicode_combining_class_t
  14 hb_unicode_combining_class_nil(hb_unicode_funcs_t * ufuncs HB_UNUSED,
  15                                hb_codepoint_t unicode HB_UNUSED,
  16                                void *user_data HB_UNUSED)
  17 {
  18         return HB_UNICODE_COMBINING_CLASS_NOT_REORDERED;
  19 }
  20
  21 static unsigned int hb_unicode_eastasian_width_nil(hb_unicode_funcs_t *
  22                                                    ufuncs HB_UNUSED,
  23                                                    hb_codepoint_t unicode
  24                                                    HB_UNUSED,
  25                                                    void *user_data HB_UNUSED)
  26 {
  27         return 1;
  28 }
  29
  30 static hb_unicode_general_category_t
  31 hb_unicode_general_category_nil(hb_unicode_funcs_t * ufuncs HB_UNUSED,
  32                                 hb_codepoint_t unicode HB_UNUSED,
  33                                 void *user_data HB_UNUSED)
  34 {
  35         return HB_UNICODE_GENERAL_CATEGORY_OTHER_LETTER;
  36 }
  37
  38 static hb_codepoint_t hb_unicode_mirroring_nil(hb_unicode_funcs_t *
  39                                                ufuncs HB_UNUSED,
  40                                                hb_codepoint_t unicode HB_UNUSED,
  41                                                void *user_data HB_UNUSED)
  42 {
  43         return unicode;
  44 }
  45
  46 static hb_script_t hb_unicode_script_nil(hb_unicode_funcs_t * ufuncs HB_UNUSED,
  47                                          hb_codepoint_t unicode HB_UNUSED,
  48                                          void *user_data HB_UNUSED)
  49 {
  50         return HB_SCRIPT_UNKNOWN;
  51 }
  52
  53 static hb_bool_t hb_unicode_compose_nil(hb_unicode_funcs_t * ufuncs HB_UNUSED,
  54                                         hb_codepoint_t a HB_UNUSED,
  55                                         hb_codepoint_t b HB_UNUSED,
  56                                         hb_codepoint_t * ab HB_UNUSED,
  57                                         void *user_data HB_UNUSED)
  58 {
  59         return FALSE;
  60 }
  61
  62 static hb_bool_t hb_unicode_decompose_nil(hb_unicode_funcs_t * ufuncs HB_UNUSED,
  63                                           hb_codepoint_t ab HB_UNUSED,
  64                                           hb_codepoint_t * a HB_UNUSED,
  65                                           hb_codepoint_t * b HB_UNUSED,
  66                                           void *user_data HB_UNUSED)
  67 {
  68         return FALSE;
  69 }
  70
  71 static unsigned int hb_unicode_decompose_compatibility_nil(hb_unicode_funcs_t *
  72                                                            ufuncs HB_UNUSED,
  73                                                            hb_codepoint_t u
  74                                                            HB_UNUSED,
  75                                                            hb_codepoint_t *
  76                                                            decomposed HB_UNUSED,
  77                                                            void *user_data
  78                                                            HB_UNUSED)
  79 {
  80         return 0;
  81 }
  82
  83 /*must be public*/
  84 hb_unicode_funcs_t _hb_unicode_funcs_nil = {
  85         REF_CNT_INVALID_VAL,    /*ref_cnt */
  86         NULL,                   /*parent */
  87         TRUE,                   /*immutable */
  88         {                       /*func */
  89          hb_unicode_combining_class_nil,
  90          hb_unicode_eastasian_width_nil,
  91          hb_unicode_general_category_nil,
  92          hb_unicode_mirroring_nil,
  93          hb_unicode_script_nil,
  94          hb_unicode_compose_nil,
  95          hb_unicode_decompose_nil,
  96          hb_unicode_decompose_compatibility_nil}
  97         ,
  98         {                       /*user_data */
  99          NULL,                  /*combining_class */
 100          NULL,                  /*eastasian_width */
 101          NULL,                  /*general_category */
 102          NULL,                  /*mirroring */
 103          NULL,                  /*script */
 104          NULL,                  /*compose */
 105          NULL,                  /*decompose */
 106          NULL                   /*decompose_compatibility */
 107          }
 108         ,
 109         {                       /*destroy */
 110          NULL,                  /*combining_class */
 111          NULL,                  /*eastasian_width */
 112          NULL,                  /*general_category */
 113          NULL,                  /*mirroring */
 114          NULL,                  /*script */
 115          NULL,                  /*compose */
 116          NULL,                  /*decompose */
 117          NULL                   /*decompose_compatibility */
 118          }
 119 };
 120
 121 /*
 122 Default_Ignorable codepoints:
 123
 124 Note that as of Oct 2012 (Unicode 6.2), U+180E MONGOLIAN VOWEL SEPARATOR
 125 is NOT Default_Ignorable, but it really behaves in a way that it should
 126 be.  That has been reported to the Unicode Technical Committee for
 127 consideration.  As such, we include it here, since Uniscribe removes it.
 128 It *is* in Unicode 6.3 however.  U+061C ARABIC LETTER MARK from Unicode
 129 6.3 is also added manually.  The new Unicode 6.3 bidi formatting
 130 characters are encoded in a block that was Default_Ignorable already.
 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
 166 hb_bool_t hb_unicode_is_default_ignorable(hb_codepoint_t ch)
 167 {
 168         hb_codepoint_t plane;
 169
 170         plane = ch >> 16;
 171         if (plane == 0) {
 172                 hb_codepoint_t page;
 173                  /*BMP*/ page = ch >> 8;
 174                 switch (page) {
 175                 case 0x00:
 176                         return ch == 0x00AD;
 177                 case 0x03:
 178                         return ch == 0x034F;
 179                 case 0x06:
 180                         return ch == 0x061C;
 181                 case 0x17:
 182                         return hb_codepoint_in_range(ch, 0x17B4, 0x17B5);
 183                 case 0x18:
 184                         return hb_codepoint_in_range(ch, 0x180B, 0x180E);
 185                 case 0x20:
 186                         return hb_codepoint_in_ranges(ch, 0x200B, 0x200F,
 187                                                       0x202A, 0x202E, 0x2060,
 188                                                       0x206F);
 189                 case 0x31:
 190                         return ch == 0x3164;
 191                 case 0xFE:
 192                         return hb_codepoint_in_range(ch, 0xFE00, 0xFE0F)
 193                             || ch == 0xFEFF;
 194                 case 0xFF:
 195                         return hb_codepoint_in_range(ch, 0xFFF0, 0xFFF8)
 196                             || ch == 0xFFA0;
 197                 default:
 198                         return FALSE;
 199                 }
 200         } else {
 201                 /*Other planes */
 202                 switch (plane) {
 203                 case 0x01:
 204                         return hb_codepoint_in_range(ch, 0x0001D173,
 205                                                      0x0001D17A);
 206                 case 0x0E:
 207                         return hb_codepoint_in_range(ch, 0x000E0000,
 208                                                      0x000E0FFF);
 209                 default:
 210                         return FALSE;
 211                 }
 212         }
 213 }
 214
 215 hb_unicode_funcs_t *hb_unicode_funcs_get_empty(void)
 216 {
 217         return &_hb_unicode_funcs_nil;
 218 }
 219
 220 hb_unicode_funcs_t *hb_unicode_funcs_reference(hb_unicode_funcs_t * ufuncs)
 221 {
 222         if (hb_atomic_int32_get(&ufuncs->ref_cnt) != REF_CNT_INVALID_VAL)
 223                 hb_atomic_int32_add(&ufuncs->ref_cnt, 1);
 224         return ufuncs;
 225 }
 226
 227 extern hb_unicode_funcs_t *hb_glib_get_unicode_funcs(void);
 228 extern hb_unicode_funcs_t *hb_nil_get_unicode_funcs(void);
 229
 230 hb_unicode_funcs_t *hb_unicode_funcs_get_default(void)
 231 {
 232 #ifdef HAVE_GLIB
 233         return hb_glib_get_unicode_funcs();
 234 #else
 235         return hb_unicode_funcs_get_empty();
 236 #endif
 237 }
 238
 239 void hb_unicode_funcs_destroy(hb_unicode_funcs_t * ufuncs)
 240 {
 241         if (!ufuncs)
 242                 return;
 243         if (hb_atomic_int32_get(&ufuncs->ref_cnt) == REF_CNT_INVALID_VAL)
 244                 return;
 245         hb_atomic_int32_add(&ufuncs->ref_cnt, -1);
 246         if (hb_atomic_int32_get(&ufuncs->ref_cnt) > 0)
 247                 return;
 248         hb_atomic_int32_set(&ufuncs->ref_cnt, REF_CNT_INVALID_VAL);
 249
 250         if (ufuncs->destroy.combining_class)
 251                 ufuncs->destroy.combining_class(ufuncs->user_data.
 252                                                 combining_class);
 253         if (ufuncs->destroy.eastasian_width)
 254                 ufuncs->destroy.eastasian_width(ufuncs->user_data.
 255                                                 eastasian_width);
 256         if (ufuncs->destroy.general_category)
 257                 ufuncs->destroy.general_category(ufuncs->user_data.
 258                                                  general_category);
 259         if (ufuncs->destroy.mirroring)
 260                 ufuncs->destroy.mirroring(ufuncs->user_data.mirroring);
 261         if (ufuncs->destroy.script)
 262                 ufuncs->destroy.script(ufuncs->user_data.script);
 263         if (ufuncs->destroy.compose)
 264                 ufuncs->destroy.compose(ufuncs->user_data.compose);
 265         if (ufuncs->destroy.decompose)
 266                 ufuncs->destroy.decompose(ufuncs->user_data.decompose);
 267         if (ufuncs->destroy.decompose_compatibility)
 268                 ufuncs->destroy.decompose_compatibility(ufuncs->user_data.
 269                                                         decompose_compatibility);
 270
 271         hb_unicode_funcs_destroy(ufuncs->parent);
 272         free(ufuncs);
 273 }