harfbuzz/src/hb-ot-map.cc

313 lines
10 KiB
C++
Raw Normal View History

2010-10-09 02:14:57 +02:00
/*
2011-04-21 23:14:28 +02:00
* Copyright © 2009,2010 Red Hat, Inc.
* Copyright © 2010,2011,2013 Google, Inc.
2010-10-09 02:14:57 +02:00
*
* This is part of HarfBuzz, a text shaping library.
*
* Permission is hereby granted, without written agreement and without
* license or royalty fees, to use, copy, modify, and distribute this
* software and its documentation for any purpose, provided that the
* above copyright notice and the following two paragraphs appear in
* all copies of this software.
*
* IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
* DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
* ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
* IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
* DAMAGE.
*
* THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
* BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
* FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS
* ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
* PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
*
* Red Hat Author(s): Behdad Esfahbod
* Google Author(s): Behdad Esfahbod
*/
#include "hb-ot-map-private.hh"
void
2010-10-12 22:00:21 +02:00
hb_ot_map_t::add_lookups (hb_face_t *face,
2010-10-12 21:35:45 +02:00
unsigned int table_index,
unsigned int feature_index,
[Indic-like] Disable automatic joiner handling for basic shaping features Not for Arabic, but for Indic-like scripts. ZWJ/ZWNJ have special meanings in those scripts, so let font lookups take full control. This undoes the regression caused by automatic-joiners handling introduced two commits ago. We only disable automatic joiner handling for the "basic shaping features" of Indic, Myanmar, and SEAsian shapers. The "presentation forms" and other features are still applied with automatic-joiner handling. This change also changes the test suite failure statistics, such that a few scripts show more "failures". The most affected is Kannada. However, upon inspection, we believe that in most, if not all, of the new failures, we are producing results superior to Uniscribe. Hard to count those! Here's an example of what is fixed by the recent joiner-handling changes: https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers, for future reference: BENGALI: 353892 out of 354188 tests passed. 296 failed (0.0835714%) DEVANAGARI: 707336 out of 707394 tests passed. 58 failed (0.00819911%) GUJARATI: 366262 out of 366457 tests passed. 195 failed (0.0532122%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 950680 out of 951913 tests passed. 1233 failed (0.129529%) KHMER: 299074 out of 299124 tests passed. 50 failed (0.0167155%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1047983 out of 1048334 tests passed. 351 failed (0.0334817%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271539 out of 271847 tests passed. 308 failed (0.113299%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-02-14 16:40:12 +01:00
hb_mask_t mask,
[Indic] Futher adjust ZWJ handling in Indic-like shapers After the Ngapi hackfest work, we were assuming that fonts won't use presentation features to choose specific forms (eg. conjuncts). As such, we were using auto-joiner behavior for such features. It proved to be troublesome as many fonts used presentation forms ('pres') for example to form conjuncts, which need to be disabled when a ZWJ is inserted. Two examples: U+0D2F,U+200D,U+0D4D,U+0D2F with kartika.ttf U+0995,U+09CD,U+200D,U+09B7 with vrinda.ttf What we do now is to never do magic to ZWJ during GSUB's main input match for Indic-style shapers. Note that backtrack/lookahead are still matched liberally, as is GPOS. This seems to be an acceptable compromise. As to the bug that initially started this work, that one needs to be fixed differently: Bug 58714 - Kannada u+0cb0 u+200d u+0ccd u+0c95 u+0cbe does not provide same results as Windows8 https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers: BENGALI: 353689 out of 354188 tests passed. 499 failed (0.140886%) DEVANAGARI: 707305 out of 707394 tests passed. 89 failed (0.0125814%) GUJARATI: 366349 out of 366457 tests passed. 108 failed (0.0294714%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 951030 out of 951913 tests passed. 883 failed (0.0927606%) KHMER: 299070 out of 299124 tests passed. 54 failed (0.0180527%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1048102 out of 1048334 tests passed. 232 failed (0.0221304%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271666 out of 271847 tests passed. 181 failed (0.0665816%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-03-19 10:53:26 +01:00
bool auto_zwj)
2010-10-12 21:35:45 +02:00
{
unsigned int lookup_indices[32];
unsigned int offset, len;
offset = 0;
do {
len = ARRAY_LENGTH (lookup_indices);
hb_ot_layout_feature_get_lookups (face,
table_tags[table_index],
feature_index,
offset, &len,
lookup_indices);
for (unsigned int i = 0; i < len; i++) {
hb_ot_map_t::lookup_map_t *lookup = lookups[table_index].push ();
if (unlikely (!lookup))
return;
lookup->mask = mask;
lookup->index = lookup_indices[i];
[Indic] Futher adjust ZWJ handling in Indic-like shapers After the Ngapi hackfest work, we were assuming that fonts won't use presentation features to choose specific forms (eg. conjuncts). As such, we were using auto-joiner behavior for such features. It proved to be troublesome as many fonts used presentation forms ('pres') for example to form conjuncts, which need to be disabled when a ZWJ is inserted. Two examples: U+0D2F,U+200D,U+0D4D,U+0D2F with kartika.ttf U+0995,U+09CD,U+200D,U+09B7 with vrinda.ttf What we do now is to never do magic to ZWJ during GSUB's main input match for Indic-style shapers. Note that backtrack/lookahead are still matched liberally, as is GPOS. This seems to be an acceptable compromise. As to the bug that initially started this work, that one needs to be fixed differently: Bug 58714 - Kannada u+0cb0 u+200d u+0ccd u+0c95 u+0cbe does not provide same results as Windows8 https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers: BENGALI: 353689 out of 354188 tests passed. 499 failed (0.140886%) DEVANAGARI: 707305 out of 707394 tests passed. 89 failed (0.0125814%) GUJARATI: 366349 out of 366457 tests passed. 108 failed (0.0294714%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 951030 out of 951913 tests passed. 883 failed (0.0927606%) KHMER: 299070 out of 299124 tests passed. 54 failed (0.0180527%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1048102 out of 1048334 tests passed. 232 failed (0.0221304%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271666 out of 271847 tests passed. 181 failed (0.0665816%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-03-19 10:53:26 +01:00
lookup->auto_zwj = auto_zwj;
}
2010-10-12 21:35:45 +02:00
offset += len;
} while (len == ARRAY_LENGTH (lookup_indices));
2010-10-12 21:35:45 +02:00
}
hb_ot_map_builder_t::hb_ot_map_builder_t (hb_face_t *face_,
const hb_segment_properties_t *props_)
{
memset (this, 0, sizeof (*this));
face = face_;
props = *props_;
/* Fetch script/language indices for GSUB/GPOS. We need these later to skip
* features not available in either table and not waste precious bits for them. */
hb_tag_t script_tags[3] = {HB_TAG_NONE, HB_TAG_NONE, HB_TAG_NONE};
hb_tag_t language_tag;
hb_ot_tags_from_script (props.script, &script_tags[0], &script_tags[1]);
language_tag = hb_ot_tag_from_language (props.language);
for (unsigned int table_index = 0; table_index < 2; table_index++) {
hb_tag_t table_tag = table_tags[table_index];
2012-11-15 01:24:05 +01:00
found_script[table_index] = hb_ot_layout_table_choose_script (face, table_tag, script_tags, &script_index[table_index], &chosen_script[table_index]);
hb_ot_layout_script_find_language (face, table_tag, script_index[table_index], language_tag, &language_index[table_index]);
}
}
2010-10-12 21:35:45 +02:00
void hb_ot_map_builder_t::add_feature (hb_tag_t tag, unsigned int value,
hb_ot_map_feature_flags_t flags)
2011-05-28 00:15:56 +02:00
{
feature_info_t *info = feature_infos.push();
if (unlikely (!info)) return;
info->tag = tag;
info->seq = feature_infos.len;
info->max_value = value;
info->flags = flags;
info->default_value = (flags & F_GLOBAL) ? value : 0;
info->stage[0] = current_stage[0];
info->stage[1] = current_stage[1];
}
2012-07-30 08:38:39 +02:00
/* Keep the next two functions in sync. */
void hb_ot_map_t::substitute (const hb_ot_shape_plan_t *plan, hb_font_t *font, hb_buffer_t *buffer) const
{
2012-07-30 08:38:39 +02:00
const unsigned int table_index = 0;
unsigned int i = 0;
for (unsigned int stage_index = 0; stage_index < stages[table_index].len; stage_index++) {
const stage_map_t *stage = &stages[table_index][stage_index];
for (; i < stage->num_lookups; i++)
[Indic-like] Disable automatic joiner handling for basic shaping features Not for Arabic, but for Indic-like scripts. ZWJ/ZWNJ have special meanings in those scripts, so let font lookups take full control. This undoes the regression caused by automatic-joiners handling introduced two commits ago. We only disable automatic joiner handling for the "basic shaping features" of Indic, Myanmar, and SEAsian shapers. The "presentation forms" and other features are still applied with automatic-joiner handling. This change also changes the test suite failure statistics, such that a few scripts show more "failures". The most affected is Kannada. However, upon inspection, we believe that in most, if not all, of the new failures, we are producing results superior to Uniscribe. Hard to count those! Here's an example of what is fixed by the recent joiner-handling changes: https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers, for future reference: BENGALI: 353892 out of 354188 tests passed. 296 failed (0.0835714%) DEVANAGARI: 707336 out of 707394 tests passed. 58 failed (0.00819911%) GUJARATI: 366262 out of 366457 tests passed. 195 failed (0.0532122%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 950680 out of 951913 tests passed. 1233 failed (0.129529%) KHMER: 299074 out of 299124 tests passed. 50 failed (0.0167155%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1047983 out of 1048334 tests passed. 351 failed (0.0334817%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271539 out of 271847 tests passed. 308 failed (0.113299%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-02-14 16:40:12 +01:00
hb_ot_layout_substitute_lookup (font, buffer,
lookups[table_index][i].index,
lookups[table_index][i].mask,
[Indic] Futher adjust ZWJ handling in Indic-like shapers After the Ngapi hackfest work, we were assuming that fonts won't use presentation features to choose specific forms (eg. conjuncts). As such, we were using auto-joiner behavior for such features. It proved to be troublesome as many fonts used presentation forms ('pres') for example to form conjuncts, which need to be disabled when a ZWJ is inserted. Two examples: U+0D2F,U+200D,U+0D4D,U+0D2F with kartika.ttf U+0995,U+09CD,U+200D,U+09B7 with vrinda.ttf What we do now is to never do magic to ZWJ during GSUB's main input match for Indic-style shapers. Note that backtrack/lookahead are still matched liberally, as is GPOS. This seems to be an acceptable compromise. As to the bug that initially started this work, that one needs to be fixed differently: Bug 58714 - Kannada u+0cb0 u+200d u+0ccd u+0c95 u+0cbe does not provide same results as Windows8 https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers: BENGALI: 353689 out of 354188 tests passed. 499 failed (0.140886%) DEVANAGARI: 707305 out of 707394 tests passed. 89 failed (0.0125814%) GUJARATI: 366349 out of 366457 tests passed. 108 failed (0.0294714%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 951030 out of 951913 tests passed. 883 failed (0.0927606%) KHMER: 299070 out of 299124 tests passed. 54 failed (0.0180527%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1048102 out of 1048334 tests passed. 232 failed (0.0221304%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271666 out of 271847 tests passed. 181 failed (0.0665816%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-03-19 10:53:26 +01:00
lookups[table_index][i].auto_zwj);
2012-07-30 08:38:39 +02:00
if (stage->callback)
2013-04-21 21:13:57 +02:00
{
buffer->clear_output ();
stage->callback (plan, font, buffer);
2013-04-21 21:13:57 +02:00
}
2012-07-30 08:38:39 +02:00
}
}
void hb_ot_map_t::position (const hb_ot_shape_plan_t *plan, hb_font_t *font, hb_buffer_t *buffer) const
2012-07-30 08:38:39 +02:00
{
const unsigned int table_index = 1;
unsigned int i = 0;
for (unsigned int stage_index = 0; stage_index < stages[table_index].len; stage_index++) {
const stage_map_t *stage = &stages[table_index][stage_index];
for (; i < stage->num_lookups; i++)
[Indic-like] Disable automatic joiner handling for basic shaping features Not for Arabic, but for Indic-like scripts. ZWJ/ZWNJ have special meanings in those scripts, so let font lookups take full control. This undoes the regression caused by automatic-joiners handling introduced two commits ago. We only disable automatic joiner handling for the "basic shaping features" of Indic, Myanmar, and SEAsian shapers. The "presentation forms" and other features are still applied with automatic-joiner handling. This change also changes the test suite failure statistics, such that a few scripts show more "failures". The most affected is Kannada. However, upon inspection, we believe that in most, if not all, of the new failures, we are producing results superior to Uniscribe. Hard to count those! Here's an example of what is fixed by the recent joiner-handling changes: https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers, for future reference: BENGALI: 353892 out of 354188 tests passed. 296 failed (0.0835714%) DEVANAGARI: 707336 out of 707394 tests passed. 58 failed (0.00819911%) GUJARATI: 366262 out of 366457 tests passed. 195 failed (0.0532122%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 950680 out of 951913 tests passed. 1233 failed (0.129529%) KHMER: 299074 out of 299124 tests passed. 50 failed (0.0167155%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1047983 out of 1048334 tests passed. 351 failed (0.0334817%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271539 out of 271847 tests passed. 308 failed (0.113299%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-02-14 16:40:12 +01:00
hb_ot_layout_position_lookup (font, buffer, lookups[table_index][i].index,
lookups[table_index][i].mask,
[Indic] Futher adjust ZWJ handling in Indic-like shapers After the Ngapi hackfest work, we were assuming that fonts won't use presentation features to choose specific forms (eg. conjuncts). As such, we were using auto-joiner behavior for such features. It proved to be troublesome as many fonts used presentation forms ('pres') for example to form conjuncts, which need to be disabled when a ZWJ is inserted. Two examples: U+0D2F,U+200D,U+0D4D,U+0D2F with kartika.ttf U+0995,U+09CD,U+200D,U+09B7 with vrinda.ttf What we do now is to never do magic to ZWJ during GSUB's main input match for Indic-style shapers. Note that backtrack/lookahead are still matched liberally, as is GPOS. This seems to be an acceptable compromise. As to the bug that initially started this work, that one needs to be fixed differently: Bug 58714 - Kannada u+0cb0 u+200d u+0ccd u+0c95 u+0cbe does not provide same results as Windows8 https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers: BENGALI: 353689 out of 354188 tests passed. 499 failed (0.140886%) DEVANAGARI: 707305 out of 707394 tests passed. 89 failed (0.0125814%) GUJARATI: 366349 out of 366457 tests passed. 108 failed (0.0294714%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 951030 out of 951913 tests passed. 883 failed (0.0927606%) KHMER: 299070 out of 299124 tests passed. 54 failed (0.0180527%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1048102 out of 1048334 tests passed. 232 failed (0.0221304%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271666 out of 271847 tests passed. 181 failed (0.0665816%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-03-19 10:53:26 +01:00
lookups[table_index][i].auto_zwj);
if (stage->callback)
stage->callback (plan, font, buffer);
}
}
void hb_ot_map_t::collect_lookups (unsigned int table_index, hb_set_t *lookups_out) const
2012-04-23 23:21:14 +02:00
{
for (unsigned int i = 0; i < lookups[table_index].len; i++)
hb_set_add (lookups_out, lookups[table_index][i].index);
2012-04-23 23:21:14 +02:00
}
2012-08-02 15:44:18 +02:00
void hb_ot_map_builder_t::add_pause (unsigned int table_index, hb_ot_map_t::pause_func_t pause_func)
{
stage_info_t *s = stages[table_index].push ();
if (likely (s)) {
s->index = current_stage[table_index];
s->callback = pause_func;
}
current_stage[table_index]++;
2011-05-28 00:15:56 +02:00
}
2010-10-12 21:35:45 +02:00
void
hb_ot_map_builder_t::compile (hb_ot_map_t &m)
2010-10-09 02:14:57 +02:00
{
m.global_mask = 1;
2012-11-15 01:24:05 +01:00
for (unsigned int table_index = 0; table_index < 2; table_index++) {
m.chosen_script[table_index] = chosen_script[table_index];
2012-11-15 01:24:05 +01:00
m.found_script[table_index] = found_script[table_index];
}
2010-10-09 02:14:57 +02:00
if (!feature_infos.len)
2010-10-09 02:14:57 +02:00
return;
/* Sort features and merge duplicates */
2011-06-15 15:49:58 +02:00
{
feature_infos.sort ();
unsigned int j = 0;
for (unsigned int i = 1; i < feature_infos.len; i++)
if (feature_infos[i].tag != feature_infos[j].tag)
feature_infos[++j] = feature_infos[i];
2010-10-09 02:14:57 +02:00
else {
if (feature_infos[i].flags & F_GLOBAL) {
feature_infos[j].flags |= F_GLOBAL;
feature_infos[j].max_value = feature_infos[i].max_value;
feature_infos[j].default_value = feature_infos[i].default_value;
} else {
feature_infos[j].flags &= ~F_GLOBAL;
2011-06-15 15:49:58 +02:00
feature_infos[j].max_value = MAX (feature_infos[j].max_value, feature_infos[i].max_value);
/* Inherit default_value from j */
2011-06-15 15:49:58 +02:00
}
feature_infos[j].flags |= (feature_infos[i].flags & F_HAS_FALLBACK);
feature_infos[j].stage[0] = MIN (feature_infos[j].stage[0], feature_infos[i].stage[0]);
feature_infos[j].stage[1] = MIN (feature_infos[j].stage[1], feature_infos[i].stage[1]);
2010-10-09 02:14:57 +02:00
}
2011-06-15 15:49:58 +02:00
feature_infos.shrink (j + 1);
}
2010-10-09 02:14:57 +02:00
/* Allocate bits now */
unsigned int next_bit = 1;
for (unsigned int i = 0; i < feature_infos.len; i++) {
2010-10-09 02:14:57 +02:00
const feature_info_t *info = &feature_infos[i];
unsigned int bits_needed;
if ((info->flags & F_GLOBAL) && info->max_value == 1)
2010-10-09 02:14:57 +02:00
/* Uses the global bit */
bits_needed = 0;
else
bits_needed = _hb_bit_storage (info->max_value);
2010-10-09 02:14:57 +02:00
if (!info->max_value || next_bit + bits_needed > 8 * sizeof (hb_mask_t))
2010-10-09 02:14:57 +02:00
continue; /* Feature disabled, or not enough bits. */
bool found = false;
unsigned int feature_index[2];
for (unsigned int table_index = 0; table_index < 2; table_index++)
2010-10-12 22:00:21 +02:00
found |= hb_ot_layout_language_find_feature (face,
2010-10-09 02:14:57 +02:00
table_tags[table_index],
script_index[table_index],
language_index[table_index],
info->tag,
&feature_index[table_index]);
if (!found && !(info->flags & F_HAS_FALLBACK))
2010-10-09 02:14:57 +02:00
continue;
hb_ot_map_t::feature_map_t *map = m.features.push ();
if (unlikely (!map))
break;
2010-10-09 02:14:57 +02:00
map->tag = info->tag;
map->index[0] = feature_index[0];
map->index[1] = feature_index[1];
map->stage[0] = info->stage[0];
map->stage[1] = info->stage[1];
[Indic] Futher adjust ZWJ handling in Indic-like shapers After the Ngapi hackfest work, we were assuming that fonts won't use presentation features to choose specific forms (eg. conjuncts). As such, we were using auto-joiner behavior for such features. It proved to be troublesome as many fonts used presentation forms ('pres') for example to form conjuncts, which need to be disabled when a ZWJ is inserted. Two examples: U+0D2F,U+200D,U+0D4D,U+0D2F with kartika.ttf U+0995,U+09CD,U+200D,U+09B7 with vrinda.ttf What we do now is to never do magic to ZWJ during GSUB's main input match for Indic-style shapers. Note that backtrack/lookahead are still matched liberally, as is GPOS. This seems to be an acceptable compromise. As to the bug that initially started this work, that one needs to be fixed differently: Bug 58714 - Kannada u+0cb0 u+200d u+0ccd u+0c95 u+0cbe does not provide same results as Windows8 https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers: BENGALI: 353689 out of 354188 tests passed. 499 failed (0.140886%) DEVANAGARI: 707305 out of 707394 tests passed. 89 failed (0.0125814%) GUJARATI: 366349 out of 366457 tests passed. 108 failed (0.0294714%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 951030 out of 951913 tests passed. 883 failed (0.0927606%) KHMER: 299070 out of 299124 tests passed. 54 failed (0.0180527%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1048102 out of 1048334 tests passed. 232 failed (0.0221304%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271666 out of 271847 tests passed. 181 failed (0.0665816%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-03-19 10:53:26 +01:00
map->auto_zwj = !(info->flags & F_MANUAL_ZWJ);
if ((info->flags & F_GLOBAL) && info->max_value == 1) {
2010-10-09 02:14:57 +02:00
/* Uses the global bit */
map->shift = 0;
map->mask = 1;
} else {
map->shift = next_bit;
map->mask = (1 << (next_bit + bits_needed)) - (1 << next_bit);
next_bit += bits_needed;
m.global_mask |= (info->default_value << map->shift) & map->mask;
2010-10-09 02:14:57 +02:00
}
map->_1_mask = (1 << map->shift) & map->mask;
map->needs_fallback = !found;
2010-10-09 02:14:57 +02:00
}
feature_infos.shrink (0); /* Done with these */
2010-10-09 02:14:57 +02:00
2012-08-02 15:44:18 +02:00
add_gsub_pause (NULL);
add_gpos_pause (NULL);
2010-10-09 02:14:57 +02:00
for (unsigned int table_index = 0; table_index < 2; table_index++) {
hb_tag_t table_tag = table_tags[table_index];
/* Collect lookup indices for features */
unsigned int required_feature_index;
2010-10-12 22:00:21 +02:00
if (hb_ot_layout_language_get_required_feature_index (face,
2010-10-09 02:14:57 +02:00
table_tag,
script_index[table_index],
language_index[table_index],
&required_feature_index))
[Indic-like] Disable automatic joiner handling for basic shaping features Not for Arabic, but for Indic-like scripts. ZWJ/ZWNJ have special meanings in those scripts, so let font lookups take full control. This undoes the regression caused by automatic-joiners handling introduced two commits ago. We only disable automatic joiner handling for the "basic shaping features" of Indic, Myanmar, and SEAsian shapers. The "presentation forms" and other features are still applied with automatic-joiner handling. This change also changes the test suite failure statistics, such that a few scripts show more "failures". The most affected is Kannada. However, upon inspection, we believe that in most, if not all, of the new failures, we are producing results superior to Uniscribe. Hard to count those! Here's an example of what is fixed by the recent joiner-handling changes: https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers, for future reference: BENGALI: 353892 out of 354188 tests passed. 296 failed (0.0835714%) DEVANAGARI: 707336 out of 707394 tests passed. 58 failed (0.00819911%) GUJARATI: 366262 out of 366457 tests passed. 195 failed (0.0532122%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 950680 out of 951913 tests passed. 1233 failed (0.129529%) KHMER: 299074 out of 299124 tests passed. 50 failed (0.0167155%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1047983 out of 1048334 tests passed. 351 failed (0.0334817%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271539 out of 271847 tests passed. 308 failed (0.113299%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-02-14 16:40:12 +01:00
m.add_lookups (face, table_index, required_feature_index, 1, true);
2010-10-09 02:14:57 +02:00
unsigned int stage_index = 0;
unsigned int last_num_lookups = 0;
for (unsigned stage = 0; stage < current_stage[table_index]; stage++)
2010-10-09 02:14:57 +02:00
{
for (unsigned i = 0; i < m.features.len; i++)
if (m.features[i].stage[table_index] == stage)
[Indic-like] Disable automatic joiner handling for basic shaping features Not for Arabic, but for Indic-like scripts. ZWJ/ZWNJ have special meanings in those scripts, so let font lookups take full control. This undoes the regression caused by automatic-joiners handling introduced two commits ago. We only disable automatic joiner handling for the "basic shaping features" of Indic, Myanmar, and SEAsian shapers. The "presentation forms" and other features are still applied with automatic-joiner handling. This change also changes the test suite failure statistics, such that a few scripts show more "failures". The most affected is Kannada. However, upon inspection, we believe that in most, if not all, of the new failures, we are producing results superior to Uniscribe. Hard to count those! Here's an example of what is fixed by the recent joiner-handling changes: https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers, for future reference: BENGALI: 353892 out of 354188 tests passed. 296 failed (0.0835714%) DEVANAGARI: 707336 out of 707394 tests passed. 58 failed (0.00819911%) GUJARATI: 366262 out of 366457 tests passed. 195 failed (0.0532122%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 950680 out of 951913 tests passed. 1233 failed (0.129529%) KHMER: 299074 out of 299124 tests passed. 50 failed (0.0167155%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1047983 out of 1048334 tests passed. 351 failed (0.0334817%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271539 out of 271847 tests passed. 308 failed (0.113299%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-02-14 16:40:12 +01:00
m.add_lookups (face, table_index,
m.features[i].index[table_index],
m.features[i].mask,
[Indic] Futher adjust ZWJ handling in Indic-like shapers After the Ngapi hackfest work, we were assuming that fonts won't use presentation features to choose specific forms (eg. conjuncts). As such, we were using auto-joiner behavior for such features. It proved to be troublesome as many fonts used presentation forms ('pres') for example to form conjuncts, which need to be disabled when a ZWJ is inserted. Two examples: U+0D2F,U+200D,U+0D4D,U+0D2F with kartika.ttf U+0995,U+09CD,U+200D,U+09B7 with vrinda.ttf What we do now is to never do magic to ZWJ during GSUB's main input match for Indic-style shapers. Note that backtrack/lookahead are still matched liberally, as is GPOS. This seems to be an acceptable compromise. As to the bug that initially started this work, that one needs to be fixed differently: Bug 58714 - Kannada u+0cb0 u+200d u+0ccd u+0c95 u+0cbe does not provide same results as Windows8 https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers: BENGALI: 353689 out of 354188 tests passed. 499 failed (0.140886%) DEVANAGARI: 707305 out of 707394 tests passed. 89 failed (0.0125814%) GUJARATI: 366349 out of 366457 tests passed. 108 failed (0.0294714%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 951030 out of 951913 tests passed. 883 failed (0.0927606%) KHMER: 299070 out of 299124 tests passed. 54 failed (0.0180527%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1048102 out of 1048334 tests passed. 232 failed (0.0221304%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271666 out of 271847 tests passed. 181 failed (0.0665816%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-03-19 10:53:26 +01:00
m.features[i].auto_zwj);
/* Sort lookups and merge duplicates */
if (last_num_lookups < m.lookups[table_index].len)
{
m.lookups[table_index].sort (last_num_lookups, m.lookups[table_index].len);
unsigned int j = last_num_lookups;
for (unsigned int i = j + 1; i < m.lookups[table_index].len; i++)
if (m.lookups[table_index][i].index != m.lookups[table_index][j].index)
m.lookups[table_index][++j] = m.lookups[table_index][i];
else
[Indic-like] Disable automatic joiner handling for basic shaping features Not for Arabic, but for Indic-like scripts. ZWJ/ZWNJ have special meanings in those scripts, so let font lookups take full control. This undoes the regression caused by automatic-joiners handling introduced two commits ago. We only disable automatic joiner handling for the "basic shaping features" of Indic, Myanmar, and SEAsian shapers. The "presentation forms" and other features are still applied with automatic-joiner handling. This change also changes the test suite failure statistics, such that a few scripts show more "failures". The most affected is Kannada. However, upon inspection, we believe that in most, if not all, of the new failures, we are producing results superior to Uniscribe. Hard to count those! Here's an example of what is fixed by the recent joiner-handling changes: https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers, for future reference: BENGALI: 353892 out of 354188 tests passed. 296 failed (0.0835714%) DEVANAGARI: 707336 out of 707394 tests passed. 58 failed (0.00819911%) GUJARATI: 366262 out of 366457 tests passed. 195 failed (0.0532122%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 950680 out of 951913 tests passed. 1233 failed (0.129529%) KHMER: 299074 out of 299124 tests passed. 50 failed (0.0167155%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1047983 out of 1048334 tests passed. 351 failed (0.0334817%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271539 out of 271847 tests passed. 308 failed (0.113299%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-02-14 16:40:12 +01:00
{
m.lookups[table_index][j].mask |= m.lookups[table_index][i].mask;
[Indic] Futher adjust ZWJ handling in Indic-like shapers After the Ngapi hackfest work, we were assuming that fonts won't use presentation features to choose specific forms (eg. conjuncts). As such, we were using auto-joiner behavior for such features. It proved to be troublesome as many fonts used presentation forms ('pres') for example to form conjuncts, which need to be disabled when a ZWJ is inserted. Two examples: U+0D2F,U+200D,U+0D4D,U+0D2F with kartika.ttf U+0995,U+09CD,U+200D,U+09B7 with vrinda.ttf What we do now is to never do magic to ZWJ during GSUB's main input match for Indic-style shapers. Note that backtrack/lookahead are still matched liberally, as is GPOS. This seems to be an acceptable compromise. As to the bug that initially started this work, that one needs to be fixed differently: Bug 58714 - Kannada u+0cb0 u+200d u+0ccd u+0c95 u+0cbe does not provide same results as Windows8 https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers: BENGALI: 353689 out of 354188 tests passed. 499 failed (0.140886%) DEVANAGARI: 707305 out of 707394 tests passed. 89 failed (0.0125814%) GUJARATI: 366349 out of 366457 tests passed. 108 failed (0.0294714%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 951030 out of 951913 tests passed. 883 failed (0.0927606%) KHMER: 299070 out of 299124 tests passed. 54 failed (0.0180527%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1048102 out of 1048334 tests passed. 232 failed (0.0221304%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271666 out of 271847 tests passed. 181 failed (0.0665816%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-03-19 10:53:26 +01:00
m.lookups[table_index][j].auto_zwj &= m.lookups[table_index][i].auto_zwj;
[Indic-like] Disable automatic joiner handling for basic shaping features Not for Arabic, but for Indic-like scripts. ZWJ/ZWNJ have special meanings in those scripts, so let font lookups take full control. This undoes the regression caused by automatic-joiners handling introduced two commits ago. We only disable automatic joiner handling for the "basic shaping features" of Indic, Myanmar, and SEAsian shapers. The "presentation forms" and other features are still applied with automatic-joiner handling. This change also changes the test suite failure statistics, such that a few scripts show more "failures". The most affected is Kannada. However, upon inspection, we believe that in most, if not all, of the new failures, we are producing results superior to Uniscribe. Hard to count those! Here's an example of what is fixed by the recent joiner-handling changes: https://bugs.freedesktop.org/show_bug.cgi?id=58714 New numbers, for future reference: BENGALI: 353892 out of 354188 tests passed. 296 failed (0.0835714%) DEVANAGARI: 707336 out of 707394 tests passed. 58 failed (0.00819911%) GUJARATI: 366262 out of 366457 tests passed. 195 failed (0.0532122%) GURMUKHI: 60706 out of 60747 tests passed. 41 failed (0.067493%) KANNADA: 950680 out of 951913 tests passed. 1233 failed (0.129529%) KHMER: 299074 out of 299124 tests passed. 50 failed (0.0167155%) LAO: 53611 out of 53644 tests passed. 33 failed (0.0615167%) MALAYALAM: 1047983 out of 1048334 tests passed. 351 failed (0.0334817%) ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%) SINHALA: 271539 out of 271847 tests passed. 308 failed (0.113299%) TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%) TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%) TIBETAN: 208469 out of 208469 tests passed. 0 failed (0%)
2013-02-14 16:40:12 +01:00
}
m.lookups[table_index].shrink (j + 1);
}
last_num_lookups = m.lookups[table_index].len;
if (stage_index < stages[table_index].len && stages[table_index][stage_index].index == stage) {
hb_ot_map_t::stage_map_t *pause_map = m.stages[table_index].push ();
if (likely (pause_map)) {
pause_map->num_lookups = last_num_lookups;
pause_map->callback = stages[table_index][stage_index].callback;
}
stage_index++;
}
2010-10-09 02:14:57 +02:00
}
}
}