include/grub/unicode.h

   1 /*
   2  *  GRUB  --  GRand Unified Bootloader
   3  *  Copyright (C) 2010  Free Software Foundation, Inc.
   4  *
   5  *  GRUB is free software: you can redistribute it and/or modify
   6  *  it under the terms of the GNU General Public License as published by
   7  *  the Free Software Foundation, either version 3 of the License, or
   8  *  (at your option) any later version.
   9  *
  10  *  GRUB is distributed in the hope that it will be useful,
  11  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13  *  GNU General Public License for more details.
  14  *
  15  *  You should have received a copy of the GNU General Public License
  16  *  along with GRUB.  If not, see <http://www.gnu.org/licenses/>.
  17  */
  18
  19 #ifndef GRUB_BIDI_HEADER
  20 #define GRUB_BIDI_HEADER        1
  21
  22 #include <grub/types.h>
  23 #include <grub/mm.h>
  24 #include <grub/misc.h>
  25
  26 struct grub_unicode_bidi_pair
  27 {
  28   grub_uint32_t key;
  29   grub_uint32_t replace;
  30 };
  31
  32 struct grub_unicode_compact_range
  33 {
  34   unsigned start:21;
  35   unsigned len:9;
  36   unsigned bidi_type:5;
  37   unsigned comb_type:8;
  38   unsigned bidi_mirror:1;
  39   unsigned join_type:3;
  40 } __attribute__ ((packed));
  41
  42 /* Old-style Arabic shaping. Used for "visual UTF-8" and
  43    in grub-mkfont to find variant glyphs in absence of GPOS tables.  */
  44 struct grub_unicode_arabic_shape
  45 {
  46   grub_uint32_t code;
  47   grub_uint32_t isolated;
  48   grub_uint32_t right_linked;
  49   grub_uint32_t both_linked;
  50   grub_uint32_t left_linked;
  51 };
  52
  53 extern struct grub_unicode_arabic_shape grub_unicode_arabic_shapes[];
  54
  55 enum grub_bidi_type
  56   {
  57     GRUB_BIDI_TYPE_L = 0,
  58     GRUB_BIDI_TYPE_LRE,
  59     GRUB_BIDI_TYPE_LRO,
  60     GRUB_BIDI_TYPE_R,
  61     GRUB_BIDI_TYPE_AL,
  62     GRUB_BIDI_TYPE_RLE,
  63     GRUB_BIDI_TYPE_RLO,
  64     GRUB_BIDI_TYPE_PDF,
  65     GRUB_BIDI_TYPE_EN,
  66     GRUB_BIDI_TYPE_ES,
  67     GRUB_BIDI_TYPE_ET,
  68     GRUB_BIDI_TYPE_AN,
  69     GRUB_BIDI_TYPE_CS,
  70     GRUB_BIDI_TYPE_NSM,
  71     GRUB_BIDI_TYPE_BN,
  72     GRUB_BIDI_TYPE_B,
  73     GRUB_BIDI_TYPE_S,
  74     GRUB_BIDI_TYPE_WS,
  75     GRUB_BIDI_TYPE_ON
  76   };
  77
  78 enum grub_join_type
  79   {
  80     GRUB_JOIN_TYPE_NONJOINING = 0,
  81     GRUB_JOIN_TYPE_LEFT = 1,
  82     GRUB_JOIN_TYPE_RIGHT = 2,
  83     GRUB_JOIN_TYPE_DUAL = 3,
  84     GRUB_JOIN_TYPE_CAUSING = 4,
  85     GRUB_JOIN_TYPE_TRANSPARENT = 5
  86   };
  87
  88 enum grub_comb_type
  89   {
  90     GRUB_UNICODE_COMB_NONE = 0,
  91     GRUB_UNICODE_COMB_OVERLAY = 1,
  92     GRUB_UNICODE_COMB_HEBREW_SHEVA = 10,
  93     GRUB_UNICODE_COMB_HEBREW_HATAF_SEGOL = 11,
  94     GRUB_UNICODE_COMB_HEBREW_HATAF_PATAH = 12,
  95     GRUB_UNICODE_COMB_HEBREW_HATAF_QAMATS = 13,
  96     GRUB_UNICODE_COMB_HEBREW_HIRIQ = 14,
  97     GRUB_UNICODE_COMB_HEBREW_TSERE = 15,
  98     GRUB_UNICODE_COMB_HEBREW_SEGOL = 16,
  99     GRUB_UNICODE_COMB_HEBREW_PATAH = 17,
 100     GRUB_UNICODE_COMB_HEBREW_QAMATS = 18,
 101     GRUB_UNICODE_COMB_HEBREW_HOLAM = 19,
 102     GRUB_UNICODE_COMB_HEBREW_QUBUTS = 20,
 103     GRUB_UNICODE_COMB_HEBREW_DAGESH = 21,
 104     GRUB_UNICODE_COMB_HEBREW_METEG = 22,
 105     GRUB_UNICODE_COMB_HEBREW_RAFE = 23,
 106     GRUB_UNICODE_COMB_HEBREW_SHIN_DOT = 24,
 107     GRUB_UNICODE_COMB_HEBREW_SIN_DOT = 25,
 108     GRUB_UNICODE_COMB_HEBREW_VARIKA = 26,
 109     GRUB_UNICODE_COMB_ARABIC_FATHATAN = 27,
 110     GRUB_UNICODE_COMB_ARABIC_DAMMATAN = 28,
 111     GRUB_UNICODE_COMB_ARABIC_KASRATAN = 29,
 112     GRUB_UNICODE_COMB_ARABIC_FATHAH = 30,
 113     GRUB_UNICODE_COMB_ARABIC_DAMMAH = 31,
 114     GRUB_UNICODE_COMB_ARABIC_KASRA = 32,
 115     GRUB_UNICODE_COMB_ARABIC_SHADDA = 33,
 116     GRUB_UNICODE_COMB_ARABIC_SUKUN = 34,
 117     GRUB_UNICODE_COMB_ARABIC_SUPERSCRIPT_ALIF = 35,
 118     GRUB_UNICODE_COMB_SYRIAC_SUPERSCRIPT_ALAPH = 36,
 119     GRUB_UNICODE_STACK_ATTACHED_BELOW = 202,
 120     GRUB_UNICODE_STACK_ATTACHED_ABOVE = 214,
 121     GRUB_UNICODE_COMB_ATTACHED_ABOVE_RIGHT = 216,
 122     GRUB_UNICODE_STACK_BELOW = 220,
 123     GRUB_UNICODE_COMB_BELOW_RIGHT = 222,
 124     GRUB_UNICODE_COMB_ABOVE_LEFT = 228,
 125     GRUB_UNICODE_STACK_ABOVE = 230,
 126     GRUB_UNICODE_COMB_ABOVE_RIGHT = 232,
 127     GRUB_UNICODE_COMB_YPOGEGRAMMENI = 240,
 128     /* If combining nature is indicated only by class and
 129        not "combining type".  */
 130     GRUB_UNICODE_COMB_ME = 253,
 131     GRUB_UNICODE_COMB_MC = 254,
 132     GRUB_UNICODE_COMB_MN = 255,
 133   };
 134
 135 /* This structure describes a glyph as opposed to character.  */
 136 struct grub_unicode_glyph
 137 {
 138   grub_uint32_t base;
 139   grub_uint16_t variant:9;
 140   grub_uint8_t attributes:5;
 141   grub_size_t ncomb;
 142   grub_size_t orig_pos;
 143   struct grub_unicode_combining {
 144     grub_uint32_t code;
 145     enum grub_comb_type type;
 146   } *combining;
 147   /* Hint by unicode subsystem how wide this character usually is.
 148      Real width is determined by font. Set only in UTF-8 stream.  */
 149   int estimated_width;
 150 };
 151
 152 #define GRUB_UNICODE_GLYPH_ATTRIBUTE_MIRROR 0x1
 153 #define GRUB_UNICODE_GLYPH_ATTRIBUTES_JOIN_LEFT_TO_RIGHT_SHIFT 1
 154 #define GRUB_UNICODE_GLYPH_ATTRIBUTE_LEFT_JOINED 0x2
 155 #define GRUB_UNICODE_GLYPH_ATTRIBUTE_RIGHT_JOINED \
 156   (GRUB_UNICODE_GLYPH_ATTRIBUTE_LEFT_JOINED \
 157    << GRUB_UNICODE_GLYPH_ATTRIBUTES_JOIN_LEFT_TO_RIGHT_SHIFT)
 158 /* Set iff the corresponding joining flags come from ZWJ or ZWNJ.  */
 159 #define GRUB_UNICODE_GLYPH_ATTRIBUTE_LEFT_JOINED_EXPLICIT 0x8
 160 #define GRUB_UNICODE_GLYPH_ATTRIBUTE_RIGHT_JOINED_EXPLICIT \
 161   (GRUB_UNICODE_GLYPH_ATTRIBUTE_LEFT_JOINED_EXPLICIT \
 162    << GRUB_UNICODE_GLYPH_ATTRIBUTES_JOIN_LEFT_TO_RIGHT_SHIFT)
 163 #define GRUB_UNICODE_GLYPH_ATTRIBUTES_JOIN \
 164   (GRUB_UNICODE_GLYPH_ATTRIBUTE_LEFT_JOINED \
 165    | GRUB_UNICODE_GLYPH_ATTRIBUTE_RIGHT_JOINED \
 166    | GRUB_UNICODE_GLYPH_ATTRIBUTE_LEFT_JOINED_EXPLICIT \
 167    | GRUB_UNICODE_GLYPH_ATTRIBUTE_RIGHT_JOINED_EXPLICIT)
 168
 169 enum
 170   {
 171     GRUB_UNICODE_DOTLESS_LOWERCASE_I       = 0x0131,
 172     GRUB_UNICODE_DOTLESS_LOWERCASE_J       = 0x0237,
 173     GRUB_UNICODE_COMBINING_GRAPHEME_JOINER = 0x034f,
 174     GRUB_UNICODE_HEBREW_WAW                = 0x05d5,
 175     GRUB_UNICODE_ARABIC_START              = 0x0600,
 176     GRUB_UNICODE_ARABIC_END                = 0x0700,
 177     GRUB_UNICODE_THAANA_ABAFILI            = 0x07a6,
 178     GRUB_UNICODE_THAANA_AABAAFILI          = 0x07a7,
 179     GRUB_UNICODE_THAANA_IBIFILI            = 0x07a8,
 180     GRUB_UNICODE_THAANA_EEBEEFILI          = 0x07a9,
 181     GRUB_UNICODE_THAANA_UBUFILI            = 0x07aa,
 182     GRUB_UNICODE_THAANA_OOBOOFILI          = 0x07ab,
 183     GRUB_UNICODE_THAANA_EBEFILI            = 0x07ac,
 184     GRUB_UNICODE_THAANA_EYBEYFILI          = 0x07ad,
 185     GRUB_UNICODE_THAANA_OBOFILI            = 0x07ae,
 186     GRUB_UNICODE_THAANA_OABOAFILI          = 0x07af,
 187     GRUB_UNICODE_THAANA_SUKUN              = 0x07b0,
 188     GRUB_UNICODE_ZWNJ                      = 0x200c,
 189     GRUB_UNICODE_ZWJ                       = 0x200d,
 190     GRUB_UNICODE_LRM                       = 0x200e,
 191     GRUB_UNICODE_RLM                       = 0x200f,
 192     GRUB_UNICODE_LRE                       = 0x202a,
 193     GRUB_UNICODE_RLE                       = 0x202b,
 194     GRUB_UNICODE_PDF                       = 0x202c,
 195     GRUB_UNICODE_LRO                       = 0x202d,
 196     GRUB_UNICODE_RLO                       = 0x202e,
 197     GRUB_UNICODE_LEFTARROW                 = 0x2190,
 198     GRUB_UNICODE_UPARROW                   = 0x2191,
 199     GRUB_UNICODE_RIGHTARROW                = 0x2192,
 200     GRUB_UNICODE_DOWNARROW                 = 0x2193,
 201     GRUB_UNICODE_LIGHT_HLINE               = 0x2500,
 202     GRUB_UNICODE_HLINE                     = 0x2501,
 203     GRUB_UNICODE_LIGHT_VLINE               = 0x2502,
 204     GRUB_UNICODE_VLINE                     = 0x2503,
 205     GRUB_UNICODE_LIGHT_CORNER_UL           = 0x250c,
 206     GRUB_UNICODE_CORNER_UL                 = 0x250f,
 207     GRUB_UNICODE_LIGHT_CORNER_UR           = 0x2510,
 208     GRUB_UNICODE_CORNER_UR                 = 0x2513,
 209     GRUB_UNICODE_LIGHT_CORNER_LL           = 0x2514,
 210     GRUB_UNICODE_CORNER_LL                 = 0x2517,
 211     GRUB_UNICODE_LIGHT_CORNER_LR           = 0x2518,
 212     GRUB_UNICODE_CORNER_LR                 = 0x251b,
 213     GRUB_UNICODE_BLACK_UP_TRIANGLE         = 0x25b2,
 214     GRUB_UNICODE_BLACK_RIGHT_TRIANGLE      = 0x25ba,
 215     GRUB_UNICODE_BLACK_DOWN_TRIANGLE       = 0x25bc,
 216     GRUB_UNICODE_BLACK_LEFT_TRIANGLE       = 0x25c4,
 217     GRUB_UNICODE_VARIATION_SELECTOR_1      = 0xfe00,
 218     GRUB_UNICODE_VARIATION_SELECTOR_16     = 0xfe0f,
 219     GRUB_UNICODE_TAG_START                 = 0xe0000,
 220     GRUB_UNICODE_TAG_END                   = 0xe007f,
 221     GRUB_UNICODE_VARIATION_SELECTOR_17     = 0xe0100,
 222     GRUB_UNICODE_VARIATION_SELECTOR_256    = 0xe01ef,
 223     GRUB_UNICODE_LAST_VALID                = 0x10ffff
 224   };
 225
 226 extern struct grub_unicode_compact_range grub_unicode_compact[];
 227 extern struct grub_unicode_bidi_pair grub_unicode_bidi_pairs[];
 228
 229 #define GRUB_UNICODE_MAX_CACHED_CHAR 0x20000
 230 /*  Unicode mandates an arbitrary limit.  */
 231 #define GRUB_BIDI_MAX_EXPLICIT_LEVEL 61
 232
 233 struct grub_term_pos
 234 {
 235   unsigned valid:1;
 236   unsigned x:15, y:16;
 237 };
 238
 239 grub_ssize_t
 240 grub_bidi_logical_to_visual (const grub_uint32_t *logical,
 241                              grub_size_t logical_len,
 242                              struct grub_unicode_glyph **visual_out,
 243                              grub_ssize_t (*getcharwidth) (const struct grub_unicode_glyph *visual),
 244                              grub_size_t max_width,
 245                              grub_size_t start_width, grub_uint32_t codechar,
 246                              struct grub_term_pos *pos,
 247                              int primitive_wrap);
 248
 249 enum grub_comb_type
 250 grub_unicode_get_comb_type (grub_uint32_t c);
 251 grub_size_t
 252 grub_unicode_aglomerate_comb (const grub_uint32_t *in, grub_size_t inlen,
 253                               struct grub_unicode_glyph *out);
 254
 255 static inline struct grub_unicode_glyph *
 256 grub_unicode_glyph_dup (const struct grub_unicode_glyph *in)
 257 {
 258   struct grub_unicode_glyph *out = grub_malloc (sizeof (*out));
 259   if (!out)
 260     return NULL;
 261   grub_memcpy (out, in, sizeof (*in));
 262   if (in->combining)
 263     {
 264       out->combining = grub_malloc (in->ncomb * sizeof (out->combining[0]));
 265       if (!out->combining)
 266         {
 267           grub_free (out);
 268           return NULL;
 269         }
 270       grub_memcpy (out->combining, in->combining,
 271                    in->ncomb * sizeof (out->combining[0]));
 272     }
 273   return out;
 274 }
 275
 276 static inline struct grub_unicode_glyph *
 277 grub_unicode_glyph_from_code (grub_uint32_t code)
 278 {
 279   struct grub_unicode_glyph *ret;
 280   ret = grub_zalloc (sizeof (*ret));
 281   if (!ret)
 282     return NULL;
 283
 284   ret->base = code;
 285
 286   return ret;
 287 }
 288
 289 grub_uint32_t
 290 grub_unicode_mirror_code (grub_uint32_t in);
 291 grub_uint32_t
 292 grub_unicode_shape_code (grub_uint32_t in, grub_uint8_t attr);
 293
 294 const grub_uint32_t *
 295 grub_unicode_get_comb_end (const grub_uint32_t *end,
 296                            const grub_uint32_t *cur);
 297
 298 #endif