hb-ot-shape-complex-indic-machine.rl revision e7562f53fe6a506d2c6d59d6688e4fa468bba462
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;
43db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodV    = 2;
44db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodN    = 3;
45db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodH    = 4;
46db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodZWNJ = 5;
47db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodZWJ  = 6;
48db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodM    = 7;
49db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodSM   = 8;
50db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodVD   = 9;
51db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodA    = 10;
52db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodNBSP = 11;
53db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodDOTTEDCIRCLE = 12;
54db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodRS   = 13;
55db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodCoeng = 14;
56db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodRepha = 15;
57db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad EsfahbodRa    = 16;
589cac1338c4bc3e9034cbfa104291ed0329debefeBehdad EsfahbodCM    = 17;
59867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
609cac1338c4bc3e9034cbfa104291ed0329debefeBehdad Esfahbodc = (C | Ra)CM*;		# is_consonant
61552d19b7a11f7dff888587fce4d56d9f8e47e819Behdad Esfahbodn = ((ZWNJ?.RS)? (N.N?)?);	# is_consonant_modifier
6225bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2Behdad Esfahbodz = ZWJ|ZWNJ;			# is_joiner
6325bc489498ef7d0beb8fe9ab663e3f0b2f52c9c2Behdad Esfahbodh = H | Coeng;			# is_halant_or_coeng
64db8981f1e0e8625714568c6d0f11f0b317b11d0aBehdad Esfahbodreph = (Ra H | Repha);		# possible reph
654c3691d2a32ca7e54a54f7c08098fd96fa7af39eBehdad Esfahbod
664ed717ef61813fa16cf74f2874848e9feb81568fBehdad Esfahbodcn = c.ZWJ?.n?;
67422ecd2d3c198a36d07d409341cb82ea57c7ad6bBehdad Esfahbodforced_rakar = ZWJ H ZWJ Ra;
68dff0ece11d61978c04e839501f179a5c3077f340Behdad Esfahbodmatra_group = z{0,3}.M.N?.(H | forced_rakar)?;
69aa7141efe49991a1160489106984e95163fe2ab8Behdad Esfahbodsyllable_tail =  (Coeng (cn|V))? (SM.ZWNJ?)? (VD VD?)?;
7018c06e189bd078affbb84c3bb5bb80687a227c5eBehdad Esfahbodplace_holder = NBSP | DOTTEDCIRCLE;
714ed717ef61813fa16cf74f2874848e9feb81568fBehdad Esfahbodhalant_group = (z?.h.(ZWJ.N?)?);
729e4f94a72cea6d65a6a7ba5a47db92e00dbfbb91Behdad Esfahbodfinal_halant_group = halant_group | h.ZWNJ;
73f0b8ed1b6dd9f1d2b9084c101a6fc5dee0cc22a8Behdad Esfahbodhalant_or_matra_group = (final_halant_group | (h.ZWJ)? matra_group{0,4});
74867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
75867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
769fa052733eb93a3ce1205f63ff8f74cb295cbe99Behdad Esfahbodconsonant_syllable =	Repha? (cn.halant_group){0,4} cn A? halant_or_matra_group? syllable_tail;
7773d71cc527d28fd5519c5d965c272ea1fb149a0eBehdad Esfahbodvowel_syllable =	reph? V.n? (ZWJ | (halant_group.cn){0,4} halant_or_matra_group? syllable_tail);
789fa052733eb93a3ce1205f63ff8f74cb295cbe99Behdad Esfahbodstandalone_cluster =	reph? place_holder.n? (halant_group.cn){0,4} halant_or_matra_group? syllable_tail;
79596740db04e7c1dadae0d8be6e401089fcaffc2dBehdad Esfahbodbroken_cluster =	reph? n? (halant_group.cn){0,4} halant_or_matra_group syllable_tail;
804be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbodother =			any;
81867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
824be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbodmain := |*
83327d14ef188396006d54af976506ab6f8bb2869aBehdad Esfahbod	consonant_syllable	=> { found_syllable (consonant_syllable); };
84327d14ef188396006d54af976506ab6f8bb2869aBehdad Esfahbod	vowel_syllable		=> { found_syllable (vowel_syllable); };
85327d14ef188396006d54af976506ab6f8bb2869aBehdad Esfahbod	standalone_cluster	=> { found_syllable (standalone_cluster); };
86166b5cf7ec2d37fb54a909c437ccdec1edaf37aaBehdad Esfahbod	broken_cluster		=> { found_syllable (broken_cluster); };
87327d14ef188396006d54af976506ab6f8bb2869aBehdad Esfahbod	other			=> { found_syllable (non_indic_cluster); };
884be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod*|;
89867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
90867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
91867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod}%%
92867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
93327d14ef188396006d54af976506ab6f8bb2869aBehdad Esfahbod#define found_syllable(syllable_type) \
944be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod  HB_STMT_START { \
95327d14ef188396006d54af976506ab6f8bb2869aBehdad Esfahbod    if (0) fprintf (stderr, "syllable %d..%d %s\n", last, p+1, #syllable_type); \
964be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod    for (unsigned int i = last; i < p+1; i++) \
97327d14ef188396006d54af976506ab6f8bb2869aBehdad Esfahbod      info[i].syllable() = (syllable_serial << 4) | syllable_type; \
984be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod    last = p+1; \
994be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod    syllable_serial++; \
100327d14ef188396006d54af976506ab6f8bb2869aBehdad Esfahbod    if (unlikely (syllable_serial == 16)) syllable_serial = 1; \
1014be46bade26faf13f7b4d447e9cc608e183955dcBehdad Esfahbod  } HB_STMT_END
102867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
103743807a3ce1b2229e5307a8aea074a7544623d8dBehdad Esfahbodstatic void
104166b5cf7ec2d37fb54a909c437ccdec1edaf37aaBehdad Esfahbodfind_syllables (hb_buffer_t *buffer)
105867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod{
106e7562f53fe6a506d2c6d59d6688e4fa468bba462Behdad Esfahbod  unsigned int p, pe, eof, ts HB_UNUSED, te HB_UNUSED, act HB_UNUSED;
107867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  int cs;
108ef24cc8c8e2478a6352c340f4611a617646de4ccBehdad Esfahbod  hb_glyph_info_t *info = buffer->info;
109867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  %%{
110867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod    write init;
111ef24cc8c8e2478a6352c340f4611a617646de4ccBehdad Esfahbod    getkey info[p].indic_category();
112867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  }%%
113867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
114867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  p = 0;
11576f76812ac7cca8ac6935952a2360d5e151480faBehdad Esfahbod  pe = eof = buffer->len;
116867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
117d69d5ceaa0ad30e8d4b9783507c59c6d4221de4fBehdad Esfahbod  unsigned int last = 0;
118327d14ef188396006d54af976506ab6f8bb2869aBehdad Esfahbod  unsigned int syllable_serial = 1;
119867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  %%{
120867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod    write exec;
121867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod  }%%
122867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod}
123867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod
124867361c3ad39629a8d5b7dc48d558a1c19e37d43Behdad Esfahbod#endif /* HB_OT_SHAPE_COMPLEX_INDIC_MACHINE_HH */
125