1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/gfx/harfbuzz/src/hb-ot-shape-complex-arabic.cc Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,374 @@
1.4 +/*
1.5 + * Copyright © 2010,2012 Google, Inc.
1.6 + *
1.7 + * This is part of HarfBuzz, a text shaping library.
1.8 + *
1.9 + * Permission is hereby granted, without written agreement and without
1.10 + * license or royalty fees, to use, copy, modify, and distribute this
1.11 + * software and its documentation for any purpose, provided that the
1.12 + * above copyright notice and the following two paragraphs appear in
1.13 + * all copies of this software.
1.14 + *
1.15 + * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
1.16 + * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
1.17 + * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
1.18 + * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
1.19 + * DAMAGE.
1.20 + *
1.21 + * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
1.22 + * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
1.23 + * FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS
1.24 + * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
1.25 + * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
1.26 + *
1.27 + * Google Author(s): Behdad Esfahbod
1.28 + */
1.29 +
1.30 +#include "hb-ot-shape-complex-private.hh"
1.31 +#include "hb-ot-shape-private.hh"
1.32 +
1.33 +
1.34 +/* buffer var allocations */
1.35 +#define arabic_shaping_action() complex_var_u8_0() /* arabic shaping action */
1.36 +
1.37 +
1.38 +/*
1.39 + * Bits used in the joining tables
1.40 + */
1.41 +enum {
1.42 + JOINING_TYPE_U = 0,
1.43 + JOINING_TYPE_L = 1,
1.44 + JOINING_TYPE_R = 2,
1.45 + JOINING_TYPE_D = 3,
1.46 + JOINING_TYPE_C = JOINING_TYPE_D,
1.47 + JOINING_GROUP_ALAPH = 4,
1.48 + JOINING_GROUP_DALATH_RISH = 5,
1.49 + NUM_STATE_MACHINE_COLS = 6,
1.50 +
1.51 + JOINING_TYPE_T = 7,
1.52 + JOINING_TYPE_X = 8 /* means: use general-category to choose between U or T. */
1.53 +};
1.54 +
1.55 +/*
1.56 + * Joining types:
1.57 + */
1.58 +
1.59 +#include "hb-ot-shape-complex-arabic-table.hh"
1.60 +
1.61 +static unsigned int get_joining_type (hb_codepoint_t u, hb_unicode_general_category_t gen_cat)
1.62 +{
1.63 + if (likely (hb_in_range<hb_codepoint_t> (u, JOINING_TABLE_FIRST, JOINING_TABLE_LAST))) {
1.64 + unsigned int j_type = joining_table[u - JOINING_TABLE_FIRST];
1.65 + if (likely (j_type != JOINING_TYPE_X))
1.66 + return j_type;
1.67 + }
1.68 +
1.69 + /* Mongolian joining data is not in ArabicJoining.txt yet. */
1.70 + if (unlikely (hb_in_range<hb_codepoint_t> (u, 0x1800, 0x18AF)))
1.71 + {
1.72 + if (unlikely (hb_in_range<hb_codepoint_t> (u, 0x1880, 0x1886)))
1.73 + return JOINING_TYPE_U;
1.74 +
1.75 + /* All letters, SIBE SYLLABLE BOUNDARY MARKER, and NIRUGU are D */
1.76 + if ((FLAG(gen_cat) & (FLAG (HB_UNICODE_GENERAL_CATEGORY_OTHER_LETTER) |
1.77 + FLAG (HB_UNICODE_GENERAL_CATEGORY_MODIFIER_LETTER)))
1.78 + || u == 0x1807 || u == 0x180A)
1.79 + return JOINING_TYPE_D;
1.80 + }
1.81 +
1.82 + /* 'Phags-pa joining data is not in ArabicJoining.txt yet. */
1.83 + if (unlikely (hb_in_range<hb_codepoint_t> (u, 0xA840, 0xA872)))
1.84 + {
1.85 + if (unlikely (u == 0xA872))
1.86 + return JOINING_TYPE_L;
1.87 +
1.88 + return JOINING_TYPE_D;
1.89 + }
1.90 +
1.91 + if (unlikely (hb_in_range<hb_codepoint_t> (u, 0x200C, 0x200D)))
1.92 + {
1.93 + return u == 0x200C ? JOINING_TYPE_U : JOINING_TYPE_C;
1.94 + }
1.95 +
1.96 + return (FLAG(gen_cat) & (FLAG(HB_UNICODE_GENERAL_CATEGORY_NON_SPACING_MARK) | FLAG(HB_UNICODE_GENERAL_CATEGORY_ENCLOSING_MARK) | FLAG(HB_UNICODE_GENERAL_CATEGORY_FORMAT))) ?
1.97 + JOINING_TYPE_T : JOINING_TYPE_U;
1.98 +}
1.99 +
1.100 +static const hb_tag_t arabic_features[] =
1.101 +{
1.102 + HB_TAG('i','n','i','t'),
1.103 + HB_TAG('m','e','d','i'),
1.104 + HB_TAG('f','i','n','a'),
1.105 + HB_TAG('i','s','o','l'),
1.106 + /* Syriac */
1.107 + HB_TAG('m','e','d','2'),
1.108 + HB_TAG('f','i','n','2'),
1.109 + HB_TAG('f','i','n','3'),
1.110 + HB_TAG_NONE
1.111 +};
1.112 +
1.113 +
1.114 +/* Same order as the feature array */
1.115 +enum {
1.116 + INIT,
1.117 + MEDI,
1.118 + FINA,
1.119 + ISOL,
1.120 +
1.121 + /* Syriac */
1.122 + MED2,
1.123 + FIN2,
1.124 + FIN3,
1.125 +
1.126 + NONE,
1.127 +
1.128 + ARABIC_NUM_FEATURES = NONE
1.129 +};
1.130 +
1.131 +static const struct arabic_state_table_entry {
1.132 + uint8_t prev_action;
1.133 + uint8_t curr_action;
1.134 + uint16_t next_state;
1.135 +} arabic_state_table[][NUM_STATE_MACHINE_COLS] =
1.136 +{
1.137 + /* jt_U, jt_L, jt_R, jt_D, jg_ALAPH, jg_DALATH_RISH */
1.138 +
1.139 + /* State 0: prev was U, not willing to join. */
1.140 + { {NONE,NONE,0}, {NONE,ISOL,2}, {NONE,ISOL,1}, {NONE,ISOL,2}, {NONE,ISOL,1}, {NONE,ISOL,6}, },
1.141 +
1.142 + /* State 1: prev was R or ISOL/ALAPH, not willing to join. */
1.143 + { {NONE,NONE,0}, {NONE,ISOL,2}, {NONE,ISOL,1}, {NONE,ISOL,2}, {NONE,FIN2,5}, {NONE,ISOL,6}, },
1.144 +
1.145 + /* State 2: prev was D/L in ISOL form, willing to join. */
1.146 + { {NONE,NONE,0}, {NONE,ISOL,2}, {INIT,FINA,1}, {INIT,FINA,3}, {INIT,FINA,4}, {INIT,FINA,6}, },
1.147 +
1.148 + /* State 3: prev was D in FINA form, willing to join. */
1.149 + { {NONE,NONE,0}, {NONE,ISOL,2}, {MEDI,FINA,1}, {MEDI,FINA,3}, {MEDI,FINA,4}, {MEDI,FINA,6}, },
1.150 +
1.151 + /* State 4: prev was FINA ALAPH, not willing to join. */
1.152 + { {NONE,NONE,0}, {NONE,ISOL,2}, {MED2,ISOL,1}, {MED2,ISOL,2}, {MED2,FIN2,5}, {MED2,ISOL,6}, },
1.153 +
1.154 + /* State 5: prev was FIN2/FIN3 ALAPH, not willing to join. */
1.155 + { {NONE,NONE,0}, {NONE,ISOL,2}, {ISOL,ISOL,1}, {ISOL,ISOL,2}, {ISOL,FIN2,5}, {ISOL,ISOL,6}, },
1.156 +
1.157 + /* State 6: prev was DALATH/RISH, not willing to join. */
1.158 + { {NONE,NONE,0}, {NONE,ISOL,2}, {NONE,ISOL,1}, {NONE,ISOL,2}, {NONE,FIN3,5}, {NONE,ISOL,6}, }
1.159 +};
1.160 +
1.161 +
1.162 +static void
1.163 +nuke_joiners (const hb_ot_shape_plan_t *plan,
1.164 + hb_font_t *font,
1.165 + hb_buffer_t *buffer);
1.166 +
1.167 +static void
1.168 +arabic_fallback_shape (const hb_ot_shape_plan_t *plan,
1.169 + hb_font_t *font,
1.170 + hb_buffer_t *buffer);
1.171 +
1.172 +static void
1.173 +collect_features_arabic (hb_ot_shape_planner_t *plan)
1.174 +{
1.175 + hb_ot_map_builder_t *map = &plan->map;
1.176 +
1.177 + /* For Language forms (in ArabicOT speak), we do the iso/fina/medi/init together,
1.178 + * then rlig and calt each in their own stage. This makes IranNastaliq's ALLAH
1.179 + * ligature work correctly. It's unfortunate though...
1.180 + *
1.181 + * This also makes Arial Bold in Windows7 work. See:
1.182 + * https://bugzilla.mozilla.org/show_bug.cgi?id=644184
1.183 + *
1.184 + * TODO: Add test cases for these two.
1.185 + */
1.186 +
1.187 + map->add_gsub_pause (nuke_joiners);
1.188 +
1.189 + map->add_global_bool_feature (HB_TAG('c','c','m','p'));
1.190 + map->add_global_bool_feature (HB_TAG('l','o','c','l'));
1.191 +
1.192 + map->add_gsub_pause (NULL);
1.193 +
1.194 + for (unsigned int i = 0; i < ARABIC_NUM_FEATURES; i++)
1.195 + map->add_feature (arabic_features[i], 1, i < 4 ? F_HAS_FALLBACK : F_NONE); /* The first four features have fallback. */
1.196 +
1.197 + map->add_gsub_pause (NULL);
1.198 +
1.199 + map->add_feature (HB_TAG('r','l','i','g'), 1, F_GLOBAL|F_HAS_FALLBACK);
1.200 + map->add_gsub_pause (arabic_fallback_shape);
1.201 +
1.202 + map->add_global_bool_feature (HB_TAG('c','a','l','t'));
1.203 + map->add_gsub_pause (NULL);
1.204 +
1.205 + map->add_global_bool_feature (HB_TAG('m','s','e','t'));
1.206 +}
1.207 +
1.208 +#include "hb-ot-shape-complex-arabic-fallback.hh"
1.209 +
1.210 +struct arabic_shape_plan_t
1.211 +{
1.212 + ASSERT_POD ();
1.213 +
1.214 + /* The "+ 1" in the next array is to accommodate for the "NONE" command,
1.215 + * which is not an OpenType feature, but this simplifies the code by not
1.216 + * having to do a "if (... < NONE) ..." and just rely on the fact that
1.217 + * mask_array[NONE] == 0. */
1.218 + hb_mask_t mask_array[ARABIC_NUM_FEATURES + 1];
1.219 +
1.220 + bool do_fallback;
1.221 + arabic_fallback_plan_t *fallback_plan;
1.222 +};
1.223 +
1.224 +static void *
1.225 +data_create_arabic (const hb_ot_shape_plan_t *plan)
1.226 +{
1.227 + arabic_shape_plan_t *arabic_plan = (arabic_shape_plan_t *) calloc (1, sizeof (arabic_shape_plan_t));
1.228 + if (unlikely (!arabic_plan))
1.229 + return NULL;
1.230 +
1.231 + arabic_plan->do_fallback = plan->props.script == HB_SCRIPT_ARABIC;
1.232 + for (unsigned int i = 0; i < ARABIC_NUM_FEATURES; i++) {
1.233 + arabic_plan->mask_array[i] = plan->map.get_1_mask (arabic_features[i]);
1.234 + if (i < 4)
1.235 + arabic_plan->do_fallback = arabic_plan->do_fallback && plan->map.needs_fallback (arabic_features[i]);
1.236 + }
1.237 +
1.238 + return arabic_plan;
1.239 +}
1.240 +
1.241 +static void
1.242 +data_destroy_arabic (void *data)
1.243 +{
1.244 + arabic_shape_plan_t *arabic_plan = (arabic_shape_plan_t *) data;
1.245 +
1.246 + arabic_fallback_plan_destroy (arabic_plan->fallback_plan);
1.247 +
1.248 + free (data);
1.249 +}
1.250 +
1.251 +static void
1.252 +arabic_joining (hb_buffer_t *buffer)
1.253 +{
1.254 + unsigned int count = buffer->len;
1.255 + unsigned int prev = (unsigned int) -1, state = 0;
1.256 +
1.257 + HB_BUFFER_ALLOCATE_VAR (buffer, arabic_shaping_action);
1.258 +
1.259 + /* Check pre-context */
1.260 + if (!(buffer->flags & HB_BUFFER_FLAG_BOT))
1.261 + for (unsigned int i = 0; i < buffer->context_len[0]; i++)
1.262 + {
1.263 + unsigned int this_type = get_joining_type (buffer->context[0][i], buffer->unicode->general_category (buffer->context[0][i]));
1.264 +
1.265 + if (unlikely (this_type == JOINING_TYPE_T))
1.266 + continue;
1.267 +
1.268 + const arabic_state_table_entry *entry = &arabic_state_table[state][this_type];
1.269 + state = entry->next_state;
1.270 + break;
1.271 + }
1.272 +
1.273 + for (unsigned int i = 0; i < count; i++)
1.274 + {
1.275 + unsigned int this_type = get_joining_type (buffer->info[i].codepoint, _hb_glyph_info_get_general_category (&buffer->info[i]));
1.276 +
1.277 + if (unlikely (this_type == JOINING_TYPE_T)) {
1.278 + buffer->info[i].arabic_shaping_action() = NONE;
1.279 + continue;
1.280 + }
1.281 +
1.282 + const arabic_state_table_entry *entry = &arabic_state_table[state][this_type];
1.283 +
1.284 + if (entry->prev_action != NONE && prev != (unsigned int) -1)
1.285 + for (; prev < i; prev++)
1.286 + buffer->info[prev].arabic_shaping_action() = entry->prev_action;
1.287 +
1.288 + buffer->info[i].arabic_shaping_action() = entry->curr_action;
1.289 +
1.290 + prev = i;
1.291 + state = entry->next_state;
1.292 + }
1.293 +
1.294 + if (!(buffer->flags & HB_BUFFER_FLAG_EOT))
1.295 + for (unsigned int i = 0; i < buffer->context_len[1]; i++)
1.296 + {
1.297 + unsigned int this_type = get_joining_type (buffer->context[1][i], buffer->unicode->general_category (buffer->context[1][i]));
1.298 +
1.299 + if (unlikely (this_type == JOINING_TYPE_T))
1.300 + continue;
1.301 +
1.302 + const arabic_state_table_entry *entry = &arabic_state_table[state][this_type];
1.303 + if (entry->prev_action != NONE && prev != (unsigned int) -1)
1.304 + buffer->info[prev].arabic_shaping_action() = entry->prev_action;
1.305 + break;
1.306 + }
1.307 +
1.308 +
1.309 + HB_BUFFER_DEALLOCATE_VAR (buffer, arabic_shaping_action);
1.310 +}
1.311 +
1.312 +static void
1.313 +setup_masks_arabic (const hb_ot_shape_plan_t *plan,
1.314 + hb_buffer_t *buffer,
1.315 + hb_font_t *font HB_UNUSED)
1.316 +{
1.317 + const arabic_shape_plan_t *arabic_plan = (const arabic_shape_plan_t *) plan->data;
1.318 +
1.319 + arabic_joining (buffer);
1.320 + unsigned int count = buffer->len;
1.321 + for (unsigned int i = 0; i < count; i++)
1.322 + buffer->info[i].mask |= arabic_plan->mask_array[buffer->info[i].arabic_shaping_action()];
1.323 +}
1.324 +
1.325 +
1.326 +static void
1.327 +nuke_joiners (const hb_ot_shape_plan_t *plan HB_UNUSED,
1.328 + hb_font_t *font HB_UNUSED,
1.329 + hb_buffer_t *buffer)
1.330 +{
1.331 + unsigned int count = buffer->len;
1.332 + for (unsigned int i = 0; i < count; i++)
1.333 + if (_hb_glyph_info_is_zwj (&buffer->info[i]))
1.334 + _hb_glyph_info_flip_joiners (&buffer->info[i]);
1.335 +}
1.336 +
1.337 +static void
1.338 +arabic_fallback_shape (const hb_ot_shape_plan_t *plan,
1.339 + hb_font_t *font,
1.340 + hb_buffer_t *buffer)
1.341 +{
1.342 + const arabic_shape_plan_t *arabic_plan = (const arabic_shape_plan_t *) plan->data;
1.343 +
1.344 + if (!arabic_plan->do_fallback)
1.345 + return;
1.346 +
1.347 +retry:
1.348 + arabic_fallback_plan_t *fallback_plan = (arabic_fallback_plan_t *) hb_atomic_ptr_get (&arabic_plan->fallback_plan);
1.349 + if (unlikely (!fallback_plan))
1.350 + {
1.351 + /* This sucks. We need a font to build the fallback plan... */
1.352 + fallback_plan = arabic_fallback_plan_create (plan, font);
1.353 + if (unlikely (!hb_atomic_ptr_cmpexch (&(const_cast<arabic_shape_plan_t *> (arabic_plan))->fallback_plan, NULL, fallback_plan))) {
1.354 + arabic_fallback_plan_destroy (fallback_plan);
1.355 + goto retry;
1.356 + }
1.357 + }
1.358 +
1.359 + arabic_fallback_plan_shape (fallback_plan, font, buffer);
1.360 +}
1.361 +
1.362 +
1.363 +const hb_ot_complex_shaper_t _hb_ot_complex_shaper_arabic =
1.364 +{
1.365 + "arabic",
1.366 + collect_features_arabic,
1.367 + NULL, /* override_features */
1.368 + data_create_arabic,
1.369 + data_destroy_arabic,
1.370 + NULL, /* preprocess_text_arabic */
1.371 + HB_OT_SHAPE_NORMALIZATION_MODE_DEFAULT,
1.372 + NULL, /* decompose */
1.373 + NULL, /* compose */
1.374 + setup_masks_arabic,
1.375 + HB_OT_SHAPE_ZERO_WIDTH_MARKS_BY_GDEF_LATE,
1.376 + true, /* fallback_position */
1.377 +};
