src/hb-ot-shape-complex-indic-private.hh

   1 /*
   2  * Copyright © 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 #ifndef HB_OT_SHAPE_COMPLEX_INDIC_PRIVATE_HH
  28 #define HB_OT_SHAPE_COMPLEX_INDIC_PRIVATE_HH
  29
  30 #include "hb-private.hh"
  31
  32
  33 #include "hb-ot-shape-complex-private.hh"
  34
  35
  36 /* buffer var allocations */
  37 #define indic_category() complex_var_persistent_u8_0() /* indic_category_t */
  38 #define indic_position() complex_var_persistent_u8_1() /* indic_matra_category_t */
  39
  40 #define INDIC_TABLE_ELEMENT_TYPE uint8_t
  41
  42 /* Cateories used in the OpenType spec:
  43  * https://www.microsoft.com/typography/otfntdev/devanot/shaping.aspx
  44  */
  45 /* Note: This enum is duplicated in the -machine.rl source file.
  46  * Not sure how to avoid duplication. */
  47 enum indic_category_t {
  48   OT_X = 0,
  49   OT_C,
  50   OT_Ra, /* Not explicitly listed in the OT spec, but used in the grammar. */
  51   OT_V,
  52   OT_N,
  53   OT_H,
  54   OT_ZWNJ,
  55   OT_ZWJ,
  56   OT_M,
  57   OT_SM,
  58   OT_VD,
  59   OT_A,
  60   OT_NBSP,
  61   OT_DOTTEDCIRCLE /* Not in the spec, but special in Uniscribe. /Very very/ special! */
  62 };
  63
  64 /* Visual positions in a syllable from left to right. */
  65 enum indic_position_t {
  66   POS_RA_TO_BECOME_REPH,
  67   POS_PRE_M,
  68   POS_PRE_C,
  69   POS_BASE_C,
  70   POS_ABOVE_C,
  71   POS_BELOW_C,
  72   POS_ABOVE_M,
  73   POS_BELOW_M,
  74   POS_POST_C,
  75   POS_POST_M,
  76   POS_SMVD
  77 };
  78
  79 /* Categories used in IndicSyllabicCategory.txt from UCD. */
  80 enum indic_syllabic_category_t {
  81   INDIC_SYLLABIC_CATEGORY_OTHER                 = OT_X,
  82
  83   INDIC_SYLLABIC_CATEGORY_AVAGRAHA              = OT_X,
  84   INDIC_SYLLABIC_CATEGORY_BINDU                 = OT_SM,
  85   INDIC_SYLLABIC_CATEGORY_CONSONANT             = OT_C,
  86   INDIC_SYLLABIC_CATEGORY_CONSONANT_DEAD        = OT_C,
  87   INDIC_SYLLABIC_CATEGORY_CONSONANT_FINAL       = OT_C,
  88   INDIC_SYLLABIC_CATEGORY_CONSONANT_HEAD_LETTER = OT_C,
  89   INDIC_SYLLABIC_CATEGORY_CONSONANT_MEDIAL      = OT_C,
  90   INDIC_SYLLABIC_CATEGORY_CONSONANT_PLACEHOLDER = OT_NBSP,
  91   INDIC_SYLLABIC_CATEGORY_CONSONANT_SUBJOINED   = OT_C,
  92   INDIC_SYLLABIC_CATEGORY_CONSONANT_REPHA       = OT_C,
  93   INDIC_SYLLABIC_CATEGORY_MODIFYING_LETTER      = OT_X,
  94   INDIC_SYLLABIC_CATEGORY_NUKTA                 = OT_N,
  95   INDIC_SYLLABIC_CATEGORY_REGISTER_SHIFTER      = OT_X,
  96   INDIC_SYLLABIC_CATEGORY_TONE_LETTER           = OT_X,
  97   INDIC_SYLLABIC_CATEGORY_TONE_MARK             = OT_X,
  98   INDIC_SYLLABIC_CATEGORY_VIRAMA                = OT_H,
  99   INDIC_SYLLABIC_CATEGORY_VISARGA               = OT_SM,
 100   INDIC_SYLLABIC_CATEGORY_VOWEL                 = OT_V,
 101   INDIC_SYLLABIC_CATEGORY_VOWEL_DEPENDENT       = OT_M,
 102   INDIC_SYLLABIC_CATEGORY_VOWEL_INDEPENDENT     = OT_V
 103 };
 104
 105 /* Categories used in IndicSMatraCategory.txt from UCD */
 106 enum indic_matra_category_t {
 107   INDIC_MATRA_CATEGORY_NOT_APPLICABLE           = POS_BASE_C,
 108
 109   INDIC_MATRA_CATEGORY_LEFT                     = POS_PRE_M,
 110   INDIC_MATRA_CATEGORY_TOP                      = POS_ABOVE_M,
 111   INDIC_MATRA_CATEGORY_BOTTOM                   = POS_BELOW_M,
 112   INDIC_MATRA_CATEGORY_RIGHT                    = POS_POST_M,
 113
 114   /* We don't really care much about these since we decompose them
 115    * in the generic pre-shaping layer.  They will only be used if
 116    * the font does not cover the decomposition.  In which case, we
 117    * define these as aliases to the place we want the split-matra
 118    * glyph to show up.  Quite arbitrary.
 119    *
 120    * TODO: There are some split matras without Unicode decompositions.
 121    * We have to figure out what to do with them.
 122    */
 123   INDIC_MATRA_CATEGORY_BOTTOM_AND_RIGHT         = POS_POST_M,
 124   INDIC_MATRA_CATEGORY_LEFT_AND_RIGHT           = POS_PRE_M,
 125   INDIC_MATRA_CATEGORY_TOP_AND_BOTTOM           = POS_BELOW_M,
 126   INDIC_MATRA_CATEGORY_TOP_AND_BOTTOM_AND_RIGHT = POS_POST_M,
 127   INDIC_MATRA_CATEGORY_TOP_AND_LEFT             = POS_PRE_M,
 128   INDIC_MATRA_CATEGORY_TOP_AND_LEFT_AND_RIGHT   = POS_PRE_M,
 129   INDIC_MATRA_CATEGORY_TOP_AND_RIGHT            = POS_POST_M,
 130
 131   INDIC_MATRA_CATEGORY_INVISIBLE                = INDIC_MATRA_CATEGORY_NOT_APPLICABLE,
 132   INDIC_MATRA_CATEGORY_OVERSTRUCK               = INDIC_MATRA_CATEGORY_NOT_APPLICABLE,
 133   INDIC_MATRA_CATEGORY_VISUAL_ORDER_LEFT        = INDIC_MATRA_CATEGORY_NOT_APPLICABLE
 134 };
 135
 136 /* Note: We use ASSERT_STATIC_EXPR_ZERO() instead of ASSERT_STATIC_EXPR() and the comma operation
 137  * because gcc fails to optimize the latter and fills the table in at runtime. */
 138 #define INDIC_COMBINE_CATEGORIES(S,M) \
 139   (ASSERT_STATIC_EXPR_ZERO (M == INDIC_MATRA_CATEGORY_NOT_APPLICABLE || (S == INDIC_SYLLABIC_CATEGORY_VIRAMA || S == INDIC_SYLLABIC_CATEGORY_VOWEL_DEPENDENT)) + \
 140    ASSERT_STATIC_EXPR_ZERO (S < 16 && M < 16) + \
 141    ((M << 4) | S))
 142
 143
 144 #include "hb-ot-shape-complex-indic-table.hh"
 145
 146 /* XXX
 147  * This is a hack for now.  We should:
 148  * 1. Move this data into the main Indic table,
 149  * and/or
 150  * 2. Probe font lookups to determine consonant positions.
 151  */
 152 static const struct consonant_position_t {
 153   hb_codepoint_t u;
 154   indic_position_t position;
 155 } consonant_positions[] = {
 156   {0x0930, POS_BELOW_C},
 157   {0x09AC, POS_BELOW_C},
 158   {0x09AF, POS_POST_C},
 159   {0x09B0, POS_BELOW_C},
 160   {0x09F0, POS_BELOW_C},
 161   {0x0A2F, POS_POST_C},
 162   {0x0A30, POS_BELOW_C},
 163   {0x0A35, POS_BELOW_C},
 164   {0x0A39, POS_BELOW_C},
 165   {0x0AB0, POS_BELOW_C},
 166   {0x0B24, POS_BELOW_C},
 167   {0x0B28, POS_BELOW_C},
 168   {0x0B2C, POS_BELOW_C},
 169   {0x0B2D, POS_BELOW_C},
 170   {0x0B2E, POS_BELOW_C},
 171   {0x0B2F, POS_POST_C},
 172   {0x0B30, POS_BELOW_C},
 173   {0x0B32, POS_BELOW_C},
 174   {0x0B33, POS_BELOW_C},
 175   {0x0B5F, POS_POST_C},
 176   {0x0B71, POS_BELOW_C},
 177   {0x0C15, POS_BELOW_C},
 178   {0x0C16, POS_BELOW_C},
 179   {0x0C17, POS_BELOW_C},
 180   {0x0C18, POS_BELOW_C},
 181   {0x0C19, POS_BELOW_C},
 182   {0x0C1A, POS_BELOW_C},
 183   {0x0C1B, POS_BELOW_C},
 184   {0x0C1C, POS_BELOW_C},
 185   {0x0C1D, POS_BELOW_C},
 186   {0x0C1E, POS_BELOW_C},
 187   {0x0C1F, POS_BELOW_C},
 188   {0x0C20, POS_BELOW_C},
 189   {0x0C21, POS_BELOW_C},
 190   {0x0C22, POS_BELOW_C},
 191   {0x0C23, POS_BELOW_C},
 192   {0x0C24, POS_BELOW_C},
 193   {0x0C25, POS_BELOW_C},
 194   {0x0C26, POS_BELOW_C},
 195   {0x0C27, POS_BELOW_C},
 196   {0x0C28, POS_BELOW_C},
 197   {0x0C2A, POS_BELOW_C},
 198   {0x0C2B, POS_BELOW_C},
 199   {0x0C2C, POS_BELOW_C},
 200   {0x0C2D, POS_BELOW_C},
 201   {0x0C2E, POS_BELOW_C},
 202   {0x0C2F, POS_BELOW_C},
 203   {0x0C30, POS_BELOW_C},
 204   {0x0C32, POS_BELOW_C},
 205   {0x0C33, POS_BELOW_C},
 206   {0x0C35, POS_BELOW_C},
 207   {0x0C36, POS_BELOW_C},
 208   {0x0C37, POS_BELOW_C},
 209   {0x0C38, POS_BELOW_C},
 210   {0x0C39, POS_BELOW_C},
 211   {0x0C95, POS_BELOW_C},
 212   {0x0C96, POS_BELOW_C},
 213   {0x0C97, POS_BELOW_C},
 214   {0x0C98, POS_BELOW_C},
 215   {0x0C99, POS_BELOW_C},
 216   {0x0C9A, POS_BELOW_C},
 217   {0x0C9B, POS_BELOW_C},
 218   {0x0C9C, POS_BELOW_C},
 219   {0x0C9D, POS_BELOW_C},
 220   {0x0C9E, POS_BELOW_C},
 221   {0x0C9F, POS_BELOW_C},
 222   {0x0CA0, POS_BELOW_C},
 223   {0x0CA1, POS_BELOW_C},
 224   {0x0CA2, POS_BELOW_C},
 225   {0x0CA3, POS_BELOW_C},
 226   {0x0CA4, POS_BELOW_C},
 227   {0x0CA5, POS_BELOW_C},
 228   {0x0CA6, POS_BELOW_C},
 229   {0x0CA7, POS_BELOW_C},
 230   {0x0CA8, POS_BELOW_C},
 231   {0x0CAA, POS_BELOW_C},
 232   {0x0CAB, POS_BELOW_C},
 233   {0x0CAC, POS_BELOW_C},
 234   {0x0CAD, POS_BELOW_C},
 235   {0x0CAE, POS_BELOW_C},
 236   {0x0CAF, POS_BELOW_C},
 237   {0x0CB0, POS_BELOW_C},
 238   {0x0CB2, POS_BELOW_C},
 239   {0x0CB3, POS_BELOW_C},
 240   {0x0CB5, POS_BELOW_C},
 241   {0x0CB6, POS_BELOW_C},
 242   {0x0CB7, POS_BELOW_C},
 243   {0x0CB8, POS_BELOW_C},
 244   {0x0CB9, POS_BELOW_C},
 245   {0x0CDE, POS_BELOW_C},
 246   {0x0D2F, POS_POST_C},
 247   {0x0D30, POS_POST_C},
 248   {0x0D32, POS_BELOW_C},
 249   {0x0D35, POS_POST_C},
 250 };
 251
 252 /* XXX
 253  * This is a hack for now.  We should move this data into the main Indic table.
 254  * Or completely remove it and just check in the tables.
 255  */
 256 static const hb_codepoint_t ra_chars[] = {
 257   0x0930, /* Devanagari */
 258   0x09B0, /* Bengali */
 259   0x09F0, /* Bengali */
 260   0x0A30, /* Gurmukhi */        /* No Reph */
 261   0x0AB0, /* Gujarati */
 262   0x0B30, /* Oriya */
 263   0x0BB0, /* Tamil */           /* No Reph */
 264   0x0C30, /* Telugu */          /* No Reph */
 265   0x0CB0, /* Kannada */
 266   0x0D30, /* Malayalam */       /* No Reph */
 267 };
 268
 269
 270 #endif /* HB_OT_SHAPE_COMPLEX_INDIC_PRIVATE_HH */