hb-ot-shape-complex-indic-machine.rl revision 25bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2
1867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod/*
227aba594c90b4444c35273a38f5fedc8e09d9a88Behdad Esfahbod * Copyright © 2011,2012  Google, Inc.
3867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod *
4867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod *  This is part of HarfBuzz, a text shaping library.
5867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod *
6867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * Permission is hereby granted, without written agreement and without
7867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * license or royalty fees, to use, copy, modify, and distribute this
8867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * software and its documentation for any purpose, provided that the
9867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * above copyright notice and the following two paragraphs appear in
10867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * all copies of this software.
11867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod *
12867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
13867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
14867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
15867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
16867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * DAMAGE.
17867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod *
18867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
19867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
20867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
21867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
22867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
23867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod *
24867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod * Google Author(s): Behdad Esfahbod
25867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod */
26867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
27867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod#ifndef HB_OT_SHAPE_COMPLEX_INDIC_MACHINE_HH
28867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod#define HB_OT_SHAPE_COMPLEX_INDIC_MACHINE_HH
29867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
30867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod#include "hb-private.hh"
31867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
32867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod%%{
33867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  machine indic_syllable_machine;
34867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  alphtype unsigned char;
35867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  write data;
36867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod}%%
37867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
38867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod%%{
39867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
40c7fe56a1d5d3e969b6ec51cd9ecd471706a19568Behdad Esfahbod# Same order as enum indic_category_t.  Not sure how to avoid duplication.
41867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodX    = 0;
42867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodC    = 1;
43867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodRa   = 2;
44867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodV    = 3;
45867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodN    = 4;
46867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodH    = 5;
47867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodZWNJ = 6;
48867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodZWJ  = 7;
49867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodM    = 8;
50867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodSM   = 9;
51867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodVD   = 10;
52867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodA    = 11;
53867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad EsfahbodNBSP = 12;
5418c06e189bd078affbb84c3bb5bb80687a227c5eBehdad EsfahbodDOTTEDCIRCLE = 13;
557d09c98a1fff97127e48eae48d380dc9fcff288eBehdad EsfahbodRS   = 14;
56deb521dee4fdca8c2124cfb39a205e6269d4a70dBehdad EsfahbodCoeng = 15;
57867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
5825bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2Behdad Esfahbodc = C | Ra;			# is_consonant
5925bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2Behdad Esfahbodn = (N.N? | ZWNJ?.RS);		# is_consonant_modifier
6025bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2Behdad Esfahbodz = ZWJ|ZWNJ;			# is_joiner
6125bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2Behdad Esfahbodh = H | Coeng;			# is_halant_or_coeng
6225bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2Behdad Esfahbodmatra_group = M.N?.H?;
6325bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2Behdad Esfahbodsyllable_tail = SM? (Coeng (c|V))? (VD VD?)?;
6418c06e189bd078affbb84c3bb5bb80687a227c5eBehdad Esfahbodplace_holder = NBSP | DOTTEDCIRCLE;
65867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
66867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
6725bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2Behdad Esfahbodconsonant_syllable =	(c.n? (h.z?|z.h))* c.n? A? (h.z? | matra_group*)? syllable_tail;
6811e2a601b19861b05dbb2051d2d078c3cfd75b29Behdad Esfahbodvowel_syllable =	(Ra H)? V.n? (z?.h.c | ZWJ.c)* matra_group* syllable_tail;
6911e2a601b19861b05dbb2051d2d078c3cfd75b29Behdad Esfahbodstandalone_cluster =	(Ra H)? place_holder.n? (z? h c)* matra_group* syllable_tail;
704be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbodother =			any;
71867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
724be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbodmain := |*
734be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod	consonant_syllable	=> { process_syllable (consonant_syllable); };
744be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod	vowel_syllable		=> { process_syllable (vowel_syllable); };
754be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod	standalone_cluster	=> { process_syllable (standalone_cluster); };
764be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod	other			=> { process_syllable (non_indic); };
774be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod*|;
78867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
79867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
80867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod}%%
81867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
824be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod#define process_syllable(func) \
834be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod  HB_STMT_START { \
849c09928989316e2befe00d52ed66e055637ccd36Behdad Esfahbod    /* fprintf (stderr, "syllable %d..%d %s\n", last, p+1, #func); */ \
854be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod    for (unsigned int i = last; i < p+1; i++) \
864be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod      info[i].syllable() = syllable_serial; \
874be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod    PASTE (initial_reordering_, func) (map, buffer, mask_array, last, p+1); \
884be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod    last = p+1; \
894be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod    syllable_serial++; \
90ff24d1081af08a887895975285d7e38f5d07bc37Behdad Esfahbod    if (unlikely (!syllable_serial)) syllable_serial++; \
914be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod  } HB_STMT_END
92867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
93743807a3ce1b2229e5307a8aea074a7544623d8dBehdad Esfahbodstatic void
94743807a3ce1b2229e5307a8aea074a7544623d8dBehdad Esfahbodfind_syllables (const hb_ot_map_t *map, hb_buffer_t *buffer, hb_mask_t *mask_array)
95867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod{
964be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod  unsigned int p, pe, eof, ts, te, act;
97867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  int cs;
98ef24cc8c8e2478a6352c340f4611a617646de4ccBehdad Esfahbod  hb_glyph_info_t *info = buffer->info;
99867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  %%{
100867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod    write init;
101ef24cc8c8e2478a6352c340f4611a617646de4ccBehdad Esfahbod    getkey info[p].indic_category();
102867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  }%%
103867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
104867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  p = 0;
10576f76812ac7cca8ac6935952a2360d5e151480faBehdad Esfahbod  pe = eof = buffer->len;
106867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
107d69d5ceaa0ad30e8d4b9783507c59c6d4221de4fBehdad Esfahbod  unsigned int last = 0;
108ff24d1081af08a887895975285d7e38f5d07bc37Behdad Esfahbod  uint8_t syllable_serial = 1;
109867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  %%{
110867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod    write exec;
111867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  }%%
112867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod}
113867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
114867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod#endif /* HB_OT_SHAPE_COMPLEX_INDIC_MACHINE_HH */
115