• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1/*
2 * Copyright © 2011,2012  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_KHMER_MACHINE_HH
28#define HB_OT_SHAPE_COMPLEX_KHMER_MACHINE_HH
29
30#include "hb.hh"
31
32enum khmer_syllable_type_t {
33  khmer_consonant_syllable,
34  khmer_broken_cluster,
35  khmer_non_khmer_cluster,
36};
37
38%%{
39  machine khmer_syllable_machine;
40  alphtype unsigned char;
41  write exports;
42  write data;
43}%%
44
45%%{
46
47export C    = 1;
48export V    = 2;
49export ZWNJ = 5;
50export ZWJ  = 6;
51export PLACEHOLDER = 11;
52export DOTTEDCIRCLE = 12;
53export Coeng= 14;
54export Ra   = 16;
55export Robatic = 20;
56export Xgroup  = 21;
57export Ygroup  = 22;
58export VAbv = 26;
59export VBlw = 27;
60export VPre = 28;
61export VPst = 29;
62
63c = (C | Ra | V);
64cn = c.((ZWJ|ZWNJ)?.Robatic)?;
65joiner = (ZWJ | ZWNJ);
66xgroup = (joiner*.Xgroup)*;
67ygroup = Ygroup*;
68
69# This grammar was experimentally extracted from what Uniscribe allows.
70
71matra_group = VPre? xgroup VBlw? xgroup (joiner?.VAbv)? xgroup VPst?;
72syllable_tail = xgroup matra_group xgroup (Coeng.c)? ygroup;
73
74
75broken_cluster =	(Coeng.cn)* (Coeng | syllable_tail);
76consonant_syllable =	(cn|PLACEHOLDER|DOTTEDCIRCLE) broken_cluster;
77other =			any;
78
79main := |*
80	consonant_syllable	=> { found_syllable (khmer_consonant_syllable); };
81	broken_cluster		=> { found_syllable (khmer_broken_cluster); };
82	other			=> { found_syllable (khmer_non_khmer_cluster); };
83*|;
84
85
86}%%
87
88#define found_syllable(syllable_type) \
89  HB_STMT_START { \
90    if (0) fprintf (stderr, "syllable %d..%d %s\n", ts, te, #syllable_type); \
91    for (unsigned int i = ts; i < te; i++) \
92      info[i].syllable() = (syllable_serial << 4) | syllable_type; \
93    syllable_serial++; \
94    if (unlikely (syllable_serial == 16)) syllable_serial = 1; \
95  } HB_STMT_END
96
97static void
98find_syllables_khmer (hb_buffer_t *buffer)
99{
100  unsigned int p, pe, eof, ts, te, act HB_UNUSED;
101  int cs;
102  hb_glyph_info_t *info = buffer->info;
103  %%{
104    write init;
105    getkey info[p].khmer_category();
106  }%%
107
108  p = 0;
109  pe = eof = buffer->len;
110
111  unsigned int syllable_serial = 1;
112  %%{
113    write exec;
114  }%%
115}
116
117#undef found_syllable
118
119#endif /* HB_OT_SHAPE_COMPLEX_KHMER_MACHINE_HH */
120