2018-01-05 15:54:31 +01:00
|
|
|
/*
|
|
|
|
* Copyright © 2011,2012 Google, Inc.
|
|
|
|
*
|
|
|
|
* 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.
|
|
|
|
*
|
|
|
|
* Google Author(s): Behdad Esfahbod
|
|
|
|
*/
|
|
|
|
|
2019-06-26 22:21:03 +02:00
|
|
|
#include "hb.hh"
|
|
|
|
|
|
|
|
#ifndef HB_NO_OT_SHAPE
|
|
|
|
|
2018-08-26 07:36:36 +02:00
|
|
|
#include "hb-ot-shape-complex-khmer.hh"
|
2021-01-29 04:27:59 +01:00
|
|
|
#include "hb-ot-shape-complex-khmer-machine.hh"
|
2018-08-26 07:36:36 +02:00
|
|
|
#include "hb-ot-layout.hh"
|
2018-01-05 15:54:31 +01:00
|
|
|
|
|
|
|
|
|
|
|
/*
|
2018-01-05 18:01:37 +01:00
|
|
|
* Khmer shaper.
|
2018-01-05 15:54:31 +01:00
|
|
|
*/
|
|
|
|
|
2018-09-24 23:55:03 +02:00
|
|
|
static const hb_ot_map_feature_t
|
2018-01-05 15:54:31 +01:00
|
|
|
khmer_features[] =
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Basic features.
|
2021-08-03 18:11:27 +02:00
|
|
|
* These features are applied all at once, before reordering, constrained
|
|
|
|
* to the syllable.
|
2018-01-05 15:54:31 +01:00
|
|
|
*/
|
2018-09-30 11:43:22 +02:00
|
|
|
{HB_TAG('p','r','e','f'), F_MANUAL_JOINERS},
|
|
|
|
{HB_TAG('b','l','w','f'), F_MANUAL_JOINERS},
|
|
|
|
{HB_TAG('a','b','v','f'), F_MANUAL_JOINERS},
|
|
|
|
{HB_TAG('p','s','t','f'), F_MANUAL_JOINERS},
|
|
|
|
{HB_TAG('c','f','a','r'), F_MANUAL_JOINERS},
|
2018-01-05 15:54:31 +01:00
|
|
|
/*
|
|
|
|
* Other features.
|
2018-10-27 00:40:12 +02:00
|
|
|
* These features are applied all at once after clearing syllables.
|
2018-01-05 15:54:31 +01:00
|
|
|
*/
|
2018-09-30 11:43:22 +02:00
|
|
|
{HB_TAG('p','r','e','s'), F_GLOBAL_MANUAL_JOINERS},
|
|
|
|
{HB_TAG('a','b','v','s'), F_GLOBAL_MANUAL_JOINERS},
|
|
|
|
{HB_TAG('b','l','w','s'), F_GLOBAL_MANUAL_JOINERS},
|
|
|
|
{HB_TAG('p','s','t','s'), F_GLOBAL_MANUAL_JOINERS},
|
2018-01-05 15:54:31 +01:00
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Must be in the same order as the khmer_features array.
|
|
|
|
*/
|
|
|
|
enum {
|
2019-07-03 00:17:56 +02:00
|
|
|
KHMER_PREF,
|
|
|
|
KHMER_BLWF,
|
|
|
|
KHMER_ABVF,
|
|
|
|
KHMER_PSTF,
|
|
|
|
KHMER_CFAR,
|
2018-01-05 15:54:31 +01:00
|
|
|
|
2019-07-03 00:17:56 +02:00
|
|
|
_KHMER_PRES,
|
|
|
|
_KHMER_ABVS,
|
|
|
|
_KHMER_BLWS,
|
|
|
|
_KHMER_PSTS,
|
2018-09-24 04:33:38 +02:00
|
|
|
|
2018-01-05 17:13:09 +01:00
|
|
|
KHMER_NUM_FEATURES,
|
2019-07-03 00:17:56 +02:00
|
|
|
KHMER_BASIC_FEATURES = _KHMER_PRES, /* Don't forget to update this! */
|
2018-01-05 15:54:31 +01:00
|
|
|
};
|
|
|
|
|
|
|
|
static void
|
2019-07-02 23:42:45 +02:00
|
|
|
setup_syllables_khmer (const hb_ot_shape_plan_t *plan,
|
|
|
|
hb_font_t *font,
|
|
|
|
hb_buffer_t *buffer);
|
2018-01-05 15:54:31 +01:00
|
|
|
static void
|
2019-07-02 23:42:45 +02:00
|
|
|
reorder_khmer (const hb_ot_shape_plan_t *plan,
|
|
|
|
hb_font_t *font,
|
|
|
|
hb_buffer_t *buffer);
|
2018-01-05 15:54:31 +01:00
|
|
|
|
|
|
|
static void
|
|
|
|
collect_features_khmer (hb_ot_shape_planner_t *plan)
|
|
|
|
{
|
|
|
|
hb_ot_map_builder_t *map = &plan->map;
|
|
|
|
|
|
|
|
/* Do this before any lookups have been applied. */
|
2019-07-02 23:42:45 +02:00
|
|
|
map->add_gsub_pause (setup_syllables_khmer);
|
|
|
|
map->add_gsub_pause (reorder_khmer);
|
2018-01-05 15:54:31 +01:00
|
|
|
|
2018-07-31 23:10:31 +02:00
|
|
|
/* Testing suggests that Uniscribe does NOT pause between basic
|
2018-07-31 22:35:10 +02:00
|
|
|
* features. Test with KhmerUI.ttf and the following three
|
|
|
|
* sequences:
|
2018-07-31 21:49:02 +02:00
|
|
|
*
|
|
|
|
* U+1789,U+17BC
|
|
|
|
* U+1789,U+17D2,U+1789
|
|
|
|
* U+1789,U+17D2,U+1789,U+17BC
|
|
|
|
*
|
|
|
|
* https://github.com/harfbuzz/harfbuzz/issues/974
|
|
|
|
*/
|
2018-09-25 00:01:53 +02:00
|
|
|
map->enable_feature (HB_TAG('l','o','c','l'));
|
|
|
|
map->enable_feature (HB_TAG('c','c','m','p'));
|
2018-07-31 22:35:10 +02:00
|
|
|
|
|
|
|
unsigned int i = 0;
|
2018-09-24 04:33:38 +02:00
|
|
|
for (; i < KHMER_BASIC_FEATURES; i++)
|
2018-09-24 23:55:03 +02:00
|
|
|
map->add_feature (khmer_features[i]);
|
2018-07-31 22:35:10 +02:00
|
|
|
|
2019-07-02 23:42:45 +02:00
|
|
|
map->add_gsub_pause (_hb_clear_syllables);
|
2018-07-31 21:49:02 +02:00
|
|
|
|
2018-09-24 04:33:38 +02:00
|
|
|
for (; i < KHMER_NUM_FEATURES; i++)
|
2018-09-24 23:55:03 +02:00
|
|
|
map->add_feature (khmer_features[i]);
|
2018-01-05 15:54:31 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
override_features_khmer (hb_ot_shape_planner_t *plan)
|
|
|
|
{
|
2018-11-29 18:30:14 +01:00
|
|
|
hb_ot_map_builder_t *map = &plan->map;
|
|
|
|
|
|
|
|
/* Khmer spec has 'clig' as part of required shaping features:
|
|
|
|
* "Apply feature 'clig' to form ligatures that are desired for
|
|
|
|
* typographical correctness.", hence in overrides... */
|
|
|
|
map->enable_feature (HB_TAG('c','l','i','g'));
|
|
|
|
|
2018-01-05 15:54:31 +01:00
|
|
|
/* Uniscribe does not apply 'kern' in Khmer. */
|
|
|
|
if (hb_options ().uniscribe_bug_compatible)
|
|
|
|
{
|
2018-11-29 18:30:14 +01:00
|
|
|
map->disable_feature (HB_TAG('k','e','r','n'));
|
2018-01-05 15:54:31 +01:00
|
|
|
}
|
|
|
|
|
2018-11-29 18:30:14 +01:00
|
|
|
map->disable_feature (HB_TAG('l','i','g','a'));
|
2018-01-05 15:54:31 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
struct khmer_shape_plan_t
|
|
|
|
{
|
2018-01-05 17:13:09 +01:00
|
|
|
hb_mask_t mask_array[KHMER_NUM_FEATURES];
|
2018-01-05 15:54:31 +01:00
|
|
|
};
|
|
|
|
|
|
|
|
static void *
|
|
|
|
data_create_khmer (const hb_ot_shape_plan_t *plan)
|
|
|
|
{
|
2021-07-08 18:58:50 +02:00
|
|
|
khmer_shape_plan_t *khmer_plan = (khmer_shape_plan_t *) hb_calloc (1, sizeof (khmer_shape_plan_t));
|
2018-01-05 15:54:31 +01:00
|
|
|
if (unlikely (!khmer_plan))
|
|
|
|
return nullptr;
|
|
|
|
|
|
|
|
for (unsigned int i = 0; i < ARRAY_LENGTH (khmer_plan->mask_array); i++)
|
|
|
|
khmer_plan->mask_array[i] = (khmer_features[i].flags & F_GLOBAL) ?
|
|
|
|
0 : plan->map.get_1_mask (khmer_features[i].tag);
|
|
|
|
|
|
|
|
return khmer_plan;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
data_destroy_khmer (void *data)
|
|
|
|
{
|
2021-07-08 18:58:50 +02:00
|
|
|
hb_free (data);
|
2018-01-05 15:54:31 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
setup_masks_khmer (const hb_ot_shape_plan_t *plan HB_UNUSED,
|
|
|
|
hb_buffer_t *buffer,
|
|
|
|
hb_font_t *font HB_UNUSED)
|
|
|
|
{
|
|
|
|
HB_BUFFER_ALLOCATE_VAR (buffer, khmer_category);
|
|
|
|
|
|
|
|
/* We cannot setup masks here. We save information about characters
|
|
|
|
* and setup masks later on in a pause-callback. */
|
|
|
|
|
|
|
|
unsigned int count = buffer->len;
|
|
|
|
hb_glyph_info_t *info = buffer->info;
|
|
|
|
for (unsigned int i = 0; i < count; i++)
|
|
|
|
set_khmer_properties (info[i]);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2019-07-02 23:42:45 +02:00
|
|
|
setup_syllables_khmer (const hb_ot_shape_plan_t *plan HB_UNUSED,
|
|
|
|
hb_font_t *font HB_UNUSED,
|
|
|
|
hb_buffer_t *buffer)
|
2018-01-05 15:54:31 +01:00
|
|
|
{
|
2019-07-02 23:42:45 +02:00
|
|
|
find_syllables_khmer (buffer);
|
2018-01-05 15:54:31 +01:00
|
|
|
foreach_syllable (buffer, start, end)
|
|
|
|
buffer->unsafe_to_break (start, end);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* Rules from:
|
2018-04-12 11:10:45 +02:00
|
|
|
* https://docs.microsoft.com/en-us/typography/script-development/devanagari */
|
2018-01-05 15:54:31 +01:00
|
|
|
|
|
|
|
static void
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
reorder_consonant_syllable (const hb_ot_shape_plan_t *plan,
|
2018-10-27 06:01:11 +02:00
|
|
|
hb_face_t *face HB_UNUSED,
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
hb_buffer_t *buffer,
|
|
|
|
unsigned int start, unsigned int end)
|
2018-01-05 15:54:31 +01:00
|
|
|
{
|
|
|
|
const khmer_shape_plan_t *khmer_plan = (const khmer_shape_plan_t *) plan->data;
|
|
|
|
hb_glyph_info_t *info = buffer->info;
|
|
|
|
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
/* Setup masks. */
|
2018-01-05 15:54:31 +01:00
|
|
|
{
|
|
|
|
/* Post-base */
|
2019-07-03 00:17:56 +02:00
|
|
|
hb_mask_t mask = khmer_plan->mask_array[KHMER_BLWF] |
|
|
|
|
khmer_plan->mask_array[KHMER_ABVF] |
|
|
|
|
khmer_plan->mask_array[KHMER_PSTF];
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
for (unsigned int i = start + 1; i < end; i++)
|
2018-01-05 15:54:31 +01:00
|
|
|
info[i].mask |= mask;
|
|
|
|
}
|
|
|
|
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
unsigned int num_coengs = 0;
|
|
|
|
for (unsigned int i = start + 1; i < end; i++)
|
2018-01-05 15:54:31 +01:00
|
|
|
{
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
/* """
|
|
|
|
* When a COENG + (Cons | IndV) combination are found (and subscript count
|
|
|
|
* is less than two) the character combination is handled according to the
|
|
|
|
* subscript type of the character following the COENG.
|
|
|
|
*
|
|
|
|
* ...
|
|
|
|
*
|
|
|
|
* Subscript Type 2 - The COENG + RO characters are reordered to immediately
|
|
|
|
* before the base glyph. Then the COENG + RO characters are assigned to have
|
|
|
|
* the 'pref' OpenType feature applied to them.
|
|
|
|
* """
|
|
|
|
*/
|
|
|
|
if (info[i].khmer_category() == OT_Coeng && num_coengs <= 2 && i + 1 < end)
|
|
|
|
{
|
|
|
|
num_coengs++;
|
|
|
|
|
|
|
|
if (info[i + 1].khmer_category() == OT_Ra)
|
2018-01-05 15:54:31 +01:00
|
|
|
{
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
for (unsigned int j = 0; j < 2; j++)
|
2019-07-03 00:17:56 +02:00
|
|
|
info[i + j].mask |= khmer_plan->mask_array[KHMER_PREF];
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
|
|
|
|
/* Move the Coeng,Ro sequence to the start. */
|
|
|
|
buffer->merge_clusters (start, i + 2);
|
|
|
|
hb_glyph_info_t t0 = info[i];
|
|
|
|
hb_glyph_info_t t1 = info[i + 1];
|
|
|
|
memmove (&info[start + 2], &info[start], (i - start) * sizeof (info[0]));
|
|
|
|
info[start] = t0;
|
|
|
|
info[start + 1] = t1;
|
2018-01-05 15:54:31 +01:00
|
|
|
|
|
|
|
/* Mark the subsequent stuff with 'cfar'. Used in Khmer.
|
|
|
|
* Read the feature spec.
|
|
|
|
* This allows distinguishing the following cases with MS Khmer fonts:
|
|
|
|
* U+1784,U+17D2,U+179A,U+17D2,U+1782
|
|
|
|
* U+1784,U+17D2,U+1782,U+17D2,U+179A
|
|
|
|
*/
|
2019-07-03 00:17:56 +02:00
|
|
|
if (khmer_plan->mask_array[KHMER_CFAR])
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
for (unsigned int j = i + 2; j < end; j++)
|
2019-07-03 00:17:56 +02:00
|
|
|
info[j].mask |= khmer_plan->mask_array[KHMER_CFAR];
|
2018-01-05 15:54:31 +01:00
|
|
|
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
num_coengs = 2; /* Done. */
|
2018-01-05 15:54:31 +01:00
|
|
|
}
|
|
|
|
}
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
|
|
|
|
/* Reorder left matra piece. */
|
2018-10-01 19:09:58 +02:00
|
|
|
else if (info[i].khmer_category() == OT_VPre)
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
{
|
|
|
|
/* Move to the start. */
|
|
|
|
buffer->merge_clusters (start, i + 1);
|
|
|
|
hb_glyph_info_t t = info[i];
|
|
|
|
memmove (&info[start + 1], &info[start], (i - start) * sizeof (info[0]));
|
|
|
|
info[start] = t;
|
|
|
|
}
|
2018-01-05 15:54:31 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2019-07-03 00:23:00 +02:00
|
|
|
reorder_syllable_khmer (const hb_ot_shape_plan_t *plan,
|
|
|
|
hb_face_t *face,
|
|
|
|
hb_buffer_t *buffer,
|
|
|
|
unsigned int start, unsigned int end)
|
2018-01-05 15:54:31 +01:00
|
|
|
{
|
2019-07-02 23:42:45 +02:00
|
|
|
khmer_syllable_type_t syllable_type = (khmer_syllable_type_t) (buffer->info[start].syllable() & 0x0F);
|
2018-01-05 15:54:31 +01:00
|
|
|
switch (syllable_type)
|
|
|
|
{
|
2019-07-02 23:42:45 +02:00
|
|
|
case khmer_broken_cluster: /* We already inserted dotted-circles, so just call the consonant_syllable. */
|
|
|
|
case khmer_consonant_syllable:
|
[khmer] Rewrite most of shaper to better follow spec
Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering. Implement that. Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence. That said, some combinations are now
reordered differently from Uniscribe. Not clear if that's intended
or a bug in Uniscribe. The following two sequences render the same
in Uniscribe whereas we reorder them differently:
U+17A0,U+17D2,U+179A,U+17C2
U+17A0,U+17C2,U+17D2,U+179A
For that reason, our test suite numbers regressed slightly. Used
to be at 34 for fails, now at:
KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)
But generally a good change, and removed lots of code.
Fixes https://github.com/harfbuzz/harfbuzz/issues/1026
2018-07-31 20:45:32 +02:00
|
|
|
reorder_consonant_syllable (plan, face, buffer, start, end);
|
2018-01-05 15:54:31 +01:00
|
|
|
break;
|
|
|
|
|
2019-07-02 23:42:45 +02:00
|
|
|
case khmer_non_khmer_cluster:
|
2018-01-05 15:54:31 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2019-07-02 23:42:45 +02:00
|
|
|
reorder_khmer (const hb_ot_shape_plan_t *plan,
|
|
|
|
hb_font_t *font,
|
|
|
|
hb_buffer_t *buffer)
|
2018-01-05 15:54:31 +01:00
|
|
|
{
|
2021-01-16 03:13:47 +01:00
|
|
|
if (buffer->message (font, "start reordering khmer"))
|
|
|
|
{
|
|
|
|
hb_syllabic_insert_dotted_circles (font, buffer,
|
|
|
|
khmer_broken_cluster,
|
|
|
|
OT_DOTTEDCIRCLE,
|
|
|
|
OT_Repha);
|
2018-01-05 15:54:31 +01:00
|
|
|
|
2020-09-17 15:25:30 +02:00
|
|
|
foreach_syllable (buffer, start, end)
|
|
|
|
reorder_syllable_khmer (plan, font->face, buffer, start, end);
|
2020-09-18 17:24:47 +02:00
|
|
|
(void) buffer->message (font, "end reordering khmer");
|
2020-09-17 15:25:30 +02:00
|
|
|
}
|
2018-01-05 15:54:31 +01:00
|
|
|
HB_BUFFER_DEALLOCATE_VAR (buffer, khmer_category);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static bool
|
|
|
|
decompose_khmer (const hb_ot_shape_normalize_context_t *c,
|
|
|
|
hb_codepoint_t ab,
|
|
|
|
hb_codepoint_t *a,
|
|
|
|
hb_codepoint_t *b)
|
|
|
|
{
|
|
|
|
switch (ab)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Decompose split matras that don't have Unicode decompositions.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* Khmer */
|
|
|
|
case 0x17BEu : *a = 0x17C1u; *b= 0x17BEu; return true;
|
|
|
|
case 0x17BFu : *a = 0x17C1u; *b= 0x17BFu; return true;
|
|
|
|
case 0x17C0u : *a = 0x17C1u; *b= 0x17C0u; return true;
|
|
|
|
case 0x17C4u : *a = 0x17C1u; *b= 0x17C4u; return true;
|
|
|
|
case 0x17C5u : *a = 0x17C1u; *b= 0x17C5u; return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
return (bool) c->unicode->decompose (ab, a, b);
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool
|
|
|
|
compose_khmer (const hb_ot_shape_normalize_context_t *c,
|
|
|
|
hb_codepoint_t a,
|
|
|
|
hb_codepoint_t b,
|
|
|
|
hb_codepoint_t *ab)
|
|
|
|
{
|
|
|
|
/* Avoid recomposing split matras. */
|
|
|
|
if (HB_UNICODE_GENERAL_CATEGORY_IS_MARK (c->unicode->general_category (a)))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
return (bool) c->unicode->compose (a, b, ab);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
const hb_ot_complex_shaper_t _hb_ot_complex_shaper_khmer =
|
|
|
|
{
|
|
|
|
collect_features_khmer,
|
|
|
|
override_features_khmer,
|
|
|
|
data_create_khmer,
|
|
|
|
data_destroy_khmer,
|
|
|
|
nullptr, /* preprocess_text */
|
|
|
|
nullptr, /* postprocess_glyphs */
|
|
|
|
HB_OT_SHAPE_NORMALIZATION_MODE_COMPOSED_DIACRITICS_NO_SHORT_CIRCUIT,
|
|
|
|
decompose_khmer,
|
|
|
|
compose_khmer,
|
|
|
|
setup_masks_khmer,
|
2018-10-02 14:17:42 +02:00
|
|
|
HB_TAG_NONE, /* gpos_tag */
|
2018-01-05 15:54:31 +01:00
|
|
|
nullptr, /* reorder_marks */
|
|
|
|
HB_OT_SHAPE_ZERO_WIDTH_MARKS_NONE,
|
|
|
|
false, /* fallback_position */
|
|
|
|
};
|
2019-06-26 22:21:03 +02:00
|
|
|
|
|
|
|
|
|
|
|
#endif
|