third_party/harfbuzz-ng/src/hb-ot-shape-complex-arabic-fallback.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_ARABIC_FALLBACK_HH
  28 #define HB_OT_SHAPE_COMPLEX_ARABIC_FALLBACK_HH
  29
  30 #include "hb-private.hh"
  31
  32 #include "hb-ot-shape-private.hh"
  33 #include "hb-ot-layout-gsub-table.hh"
  34
  35
  36 /* Features ordered the same as the entries in shaping_table rows,
  37  * followed by rlig.  Don't change. */
  38 static const hb_tag_t arabic_fallback_features[] =
  39 {
  40   HB_TAG('i','n','i','t'),
  41   HB_TAG('m','e','d','i'),
  42   HB_TAG('f','i','n','a'),
  43   HB_TAG('i','s','o','l'),
  44   HB_TAG('r','l','i','g'),
  45 };
  46
  47 static OT::SubstLookup *
  48 arabic_fallback_synthesize_lookup_single (const hb_ot_shape_plan_t *plan HB_UNUSED,
  49                                           hb_font_t *font,
  50                                           unsigned int feature_index)
  51 {
  52   OT::GlyphID glyphs[SHAPING_TABLE_LAST - SHAPING_TABLE_FIRST + 1];
  53   OT::GlyphID substitutes[SHAPING_TABLE_LAST - SHAPING_TABLE_FIRST + 1];
  54   unsigned int num_glyphs = 0;
  55
  56   /* Populate arrays */
  57   for (hb_codepoint_t u = SHAPING_TABLE_FIRST; u < SHAPING_TABLE_LAST + 1; u++)
  58   {
  59     hb_codepoint_t s = shaping_table[u - SHAPING_TABLE_FIRST][feature_index];
  60     hb_codepoint_t u_glyph, s_glyph;
  61
  62     if (!s ||
  63         !hb_font_get_glyph (font, u, 0, &u_glyph) ||
  64         !hb_font_get_glyph (font, s, 0, &s_glyph) ||
  65         u_glyph == s_glyph ||
  66         u_glyph > 0xFFFFu || s_glyph > 0xFFFFu)
  67       continue;
  68
  69     glyphs[num_glyphs].set (u_glyph);
  70     substitutes[num_glyphs].set (s_glyph);
  71
  72     num_glyphs++;
  73   }
  74
  75   if (!num_glyphs)
  76     return NULL;
  77
  78   /* Bubble-sort!
  79    * May not be good-enough for presidential candidate interviews, but good-enough for us... */
  80   hb_bubble_sort (&glyphs[0], num_glyphs, OT::GlyphID::cmp, &substitutes[0]);
  81
  82   OT::Supplier<OT::GlyphID> glyphs_supplier      (glyphs, num_glyphs);
  83   OT::Supplier<OT::GlyphID> substitutes_supplier (substitutes, num_glyphs);
  84
  85   /* Each glyph takes four bytes max, and there's some overhead. */
  86   char buf[(SHAPING_TABLE_LAST - SHAPING_TABLE_FIRST + 1) * 4 + 128];
  87   OT::hb_serialize_context_t c (buf, sizeof (buf));
  88   OT::SubstLookup *lookup = c.start_serialize<OT::SubstLookup> ();
  89   bool ret = lookup->serialize_single (&c,
  90                                        OT::LookupFlag::IgnoreMarks,
  91                                        glyphs_supplier,
  92                                        substitutes_supplier,
  93                                        num_glyphs);
  94   c.end_serialize ();
  95   /* TODO sanitize the results? */
  96
  97   return ret ? c.copy<OT::SubstLookup> () : NULL;
  98 }
  99
 100 static OT::SubstLookup *
 101 arabic_fallback_synthesize_lookup_ligature (const hb_ot_shape_plan_t *plan HB_UNUSED,
 102                                             hb_font_t *font)
 103 {
 104   OT::GlyphID first_glyphs[ARRAY_LENGTH_CONST (ligature_table)];
 105   unsigned int first_glyphs_indirection[ARRAY_LENGTH_CONST (ligature_table)];
 106   unsigned int ligature_per_first_glyph_count_list[ARRAY_LENGTH_CONST (first_glyphs)];
 107   unsigned int num_first_glyphs = 0;
 108
 109   /* We know that all our ligatures are 2-component */
 110   OT::GlyphID ligature_list[ARRAY_LENGTH_CONST (first_glyphs) * ARRAY_LENGTH_CONST(ligature_table[0].ligatures)];
 111   unsigned int component_count_list[ARRAY_LENGTH_CONST (ligature_list)];
 112   OT::GlyphID component_list[ARRAY_LENGTH_CONST (ligature_list) * 1/* One extra component per ligature */];
 113   unsigned int num_ligatures = 0;
 114
 115   /* Populate arrays */
 116
 117   /* Sort out the first-glyphs */
 118   for (unsigned int first_glyph_idx = 0; first_glyph_idx < ARRAY_LENGTH (first_glyphs); first_glyph_idx++)
 119   {
 120     hb_codepoint_t first_u = ligature_table[first_glyph_idx].first;
 121     hb_codepoint_t first_glyph;
 122     if (!hb_font_get_glyph (font, first_u, 0, &first_glyph))
 123       continue;
 124     first_glyphs[num_first_glyphs].set (first_glyph);
 125     ligature_per_first_glyph_count_list[num_first_glyphs] = 0;
 126     first_glyphs_indirection[num_first_glyphs] = first_glyph_idx;
 127     num_first_glyphs++;
 128   }
 129   hb_bubble_sort (&first_glyphs[0], num_first_glyphs, OT::GlyphID::cmp, &first_glyphs_indirection[0]);
 130
 131   /* Now that the first-glyphs are sorted, walk again, populate ligatures. */
 132   for (unsigned int i = 0; i < num_first_glyphs; i++)
 133   {
 134     unsigned int first_glyph_idx = first_glyphs_indirection[i];
 135
 136     for (unsigned int second_glyph_idx = 0; second_glyph_idx < ARRAY_LENGTH (ligature_table[0].ligatures); second_glyph_idx++)
 137     {
 138       hb_codepoint_t second_u   = ligature_table[first_glyph_idx].ligatures[second_glyph_idx].second;
 139       hb_codepoint_t ligature_u = ligature_table[first_glyph_idx].ligatures[second_glyph_idx].ligature;
 140       hb_codepoint_t second_glyph, ligature_glyph;
 141       if (!second_u ||
 142           !hb_font_get_glyph (font, second_u,   0, &second_glyph) ||
 143           !hb_font_get_glyph (font, ligature_u, 0, &ligature_glyph))
 144         continue;
 145
 146       ligature_per_first_glyph_count_list[i]++;
 147
 148       ligature_list[num_ligatures].set (ligature_glyph);
 149       component_count_list[num_ligatures] = 2;
 150       component_list[num_ligatures].set (second_glyph);
 151       num_ligatures++;
 152     }
 153   }
 154
 155   if (!num_ligatures)
 156     return NULL;
 157
 158   OT::Supplier<OT::GlyphID>   first_glyphs_supplier                      (first_glyphs, num_first_glyphs);
 159   OT::Supplier<unsigned int > ligature_per_first_glyph_count_supplier    (ligature_per_first_glyph_count_list, num_first_glyphs);
 160   OT::Supplier<OT::GlyphID>   ligatures_supplier                         (ligature_list, num_ligatures);
 161   OT::Supplier<unsigned int > component_count_supplier                   (component_count_list, num_ligatures);
 162   OT::Supplier<OT::GlyphID>   component_supplier                         (component_list, num_ligatures);
 163
 164   /* 16 bytes per ligature ought to be enough... */
 165   char buf[ARRAY_LENGTH_CONST (ligature_list) * 16 + 128];
 166   OT::hb_serialize_context_t c (buf, sizeof (buf));
 167   OT::SubstLookup *lookup = c.start_serialize<OT::SubstLookup> ();
 168   bool ret = lookup->serialize_ligature (&c,
 169                                          OT::LookupFlag::IgnoreMarks,
 170                                          first_glyphs_supplier,
 171                                          ligature_per_first_glyph_count_supplier,
 172                                          num_first_glyphs,
 173                                          ligatures_supplier,
 174                                          component_count_supplier,
 175                                          component_supplier);
 176
 177   c.end_serialize ();
 178   /* TODO sanitize the results? */
 179
 180   return ret ? c.copy<OT::SubstLookup> () : NULL;
 181 }
 182
 183 static OT::SubstLookup *
 184 arabic_fallback_synthesize_lookup (const hb_ot_shape_plan_t *plan,
 185                                    hb_font_t *font,
 186                                    unsigned int feature_index)
 187 {
 188   if (feature_index < 4)
 189     return arabic_fallback_synthesize_lookup_single (plan, font, feature_index);
 190   else
 191     return arabic_fallback_synthesize_lookup_ligature (plan, font);
 192 }
 193
 194 #define ARABIC_FALLBACK_MAX_LOOKUPS 5
 195
 196 struct arabic_fallback_plan_t
 197 {
 198   ASSERT_POD ();
 199
 200   unsigned int num_lookups;
 201   bool free_lookups;
 202
 203   hb_mask_t mask_array[ARABIC_FALLBACK_MAX_LOOKUPS];
 204   OT::SubstLookup *lookup_array[ARABIC_FALLBACK_MAX_LOOKUPS];
 205   hb_ot_layout_lookup_accelerator_t accel_array[ARABIC_FALLBACK_MAX_LOOKUPS];
 206 };
 207
 208 static const arabic_fallback_plan_t arabic_fallback_plan_nil = {};
 209
 210 #if (defined(_WIN32) || defined(__CYGWIN__)) && !defined(HB_NO_WIN1256)
 211 #define HB_WITH_WIN1256
 212 #endif
 213
 214 #ifdef HB_WITH_WIN1256
 215 #include "hb-ot-shape-complex-arabic-win1256.hh"
 216 #endif
 217
 218 struct ManifestLookup {
 219   OT::Tag tag;
 220   OT::OffsetTo<OT::SubstLookup> lookupOffset;
 221 };
 222 typedef OT::ArrayOf<ManifestLookup> Manifest;
 223
 224 static bool
 225 arabic_fallback_plan_init_win1256 (arabic_fallback_plan_t *fallback_plan,
 226                                    const hb_ot_shape_plan_t *plan,
 227                                    hb_font_t *font)
 228 {
 229 #ifdef HB_WITH_WIN1256
 230   /* Does this font look like it's Windows-1256-encoded? */
 231   hb_codepoint_t g;
 232   if (!(hb_font_get_glyph (font, 0x0627u, 0, &g) && g == 199 /* ALEF */ &&
 233         hb_font_get_glyph (font, 0x0644u, 0, &g) && g == 225 /* LAM */ &&
 234         hb_font_get_glyph (font, 0x0649u, 0, &g) && g == 236 /* ALEF MAKSURA */ &&
 235         hb_font_get_glyph (font, 0x064Au, 0, &g) && g == 237 /* YEH */ &&
 236         hb_font_get_glyph (font, 0x0652u, 0, &g) && g == 250 /* SUKUN */))
 237     return false;
 238
 239   const Manifest &manifest = reinterpret_cast<const Manifest&> (arabic_win1256_gsub_lookups.manifest);
 240   ASSERT_STATIC (sizeof (arabic_win1256_gsub_lookups.manifestData) / sizeof (ManifestLookup)
 241                  <= ARABIC_FALLBACK_MAX_LOOKUPS);
 242   /* TODO sanitize the table? */
 243
 244   unsigned j = 0;
 245   unsigned int count = manifest.len;
 246   for (unsigned int i = 0; i < count; i++)
 247   {
 248     fallback_plan->mask_array[j] = plan->map.get_1_mask (manifest[i].tag);
 249     if (fallback_plan->mask_array[j])
 250     {
 251       fallback_plan->lookup_array[j] = const_cast<OT::SubstLookup*> (&(&manifest+manifest[i].lookupOffset));
 252       if (fallback_plan->lookup_array[j])
 253       {
 254         fallback_plan->accel_array[j].init (*fallback_plan->lookup_array[j]);
 255         j++;
 256       }
 257     }
 258   }
 259
 260   fallback_plan->num_lookups = j;
 261   fallback_plan->free_lookups = false;
 262
 263   return j > 0;
 264 #else
 265   return false;
 266 #endif
 267 }
 268
 269 static bool
 270 arabic_fallback_plan_init_unicode (arabic_fallback_plan_t *fallback_plan,
 271                                    const hb_ot_shape_plan_t *plan,
 272                                    hb_font_t *font)
 273 {
 274   ASSERT_STATIC (ARRAY_LENGTH_CONST(arabic_fallback_features) <= ARABIC_FALLBACK_MAX_LOOKUPS);
 275   unsigned int j = 0;
 276   for (unsigned int i = 0; i < ARRAY_LENGTH(arabic_fallback_features) ; i++)
 277   {
 278     fallback_plan->mask_array[j] = plan->map.get_1_mask (arabic_fallback_features[i]);
 279     if (fallback_plan->mask_array[j])
 280     {
 281       fallback_plan->lookup_array[j] = arabic_fallback_synthesize_lookup (plan, font, i);
 282       if (fallback_plan->lookup_array[j])
 283       {
 284         fallback_plan->accel_array[j].init (*fallback_plan->lookup_array[j]);
 285         j++;
 286       }
 287     }
 288   }
 289
 290   fallback_plan->num_lookups = j;
 291   fallback_plan->free_lookups = true;
 292
 293   return j > 0;
 294 }
 295
 296 static arabic_fallback_plan_t *
 297 arabic_fallback_plan_create (const hb_ot_shape_plan_t *plan,
 298                              hb_font_t *font)
 299 {
 300   arabic_fallback_plan_t *fallback_plan = (arabic_fallback_plan_t *) calloc (1, sizeof (arabic_fallback_plan_t));
 301   if (unlikely (!fallback_plan))
 302     return const_cast<arabic_fallback_plan_t *> (&arabic_fallback_plan_nil);
 303
 304   fallback_plan->num_lookups = 0;
 305   fallback_plan->free_lookups = false;
 306
 307   /* Try synthesizing GSUB table using Unicode Arabic Presentation Forms,
 308    * in case the font has cmap entries for the presentation-forms characters. */
 309   if (arabic_fallback_plan_init_unicode (fallback_plan, plan, font))
 310     return fallback_plan;
 311
 312   /* See if this looks like a Windows-1256-encoded font.  If it does, use a
 313    * hand-coded GSUB table. */
 314   if (arabic_fallback_plan_init_win1256 (fallback_plan, plan, font))
 315     return fallback_plan;
 316
 317   free (fallback_plan);
 318   return const_cast<arabic_fallback_plan_t *> (&arabic_fallback_plan_nil);
 319 }
 320
 321 static void
 322 arabic_fallback_plan_destroy (arabic_fallback_plan_t *fallback_plan)
 323 {
 324   if (!fallback_plan || fallback_plan == &arabic_fallback_plan_nil)
 325     return;
 326
 327   for (unsigned int i = 0; i < fallback_plan->num_lookups; i++)
 328     if (fallback_plan->lookup_array[i])
 329     {
 330       fallback_plan->accel_array[i].fini (fallback_plan->lookup_array[i]);
 331       if (fallback_plan->free_lookups)
 332         free (fallback_plan->lookup_array[i]);
 333     }
 334
 335   free (fallback_plan);
 336 }
 337
 338 static void
 339 arabic_fallback_plan_shape (arabic_fallback_plan_t *fallback_plan,
 340                             hb_font_t *font,
 341                             hb_buffer_t *buffer)
 342 {
 343   OT::hb_apply_context_t c (0, font, buffer);
 344   for (unsigned int i = 0; i < fallback_plan->num_lookups; i++)
 345     if (fallback_plan->lookup_array[i]) {
 346       c.set_lookup_mask (fallback_plan->mask_array[i]);
 347       hb_ot_layout_substitute_lookup (&c,
 348                                       *fallback_plan->lookup_array[i],
 349                                       fallback_plan->accel_array[i]);
 350     }
 351 }
 352
 353
 354 #endif /* HB_OT_SHAPE_COMPLEX_ARABIC_FALLBACK_HH */