Home | History | Annotate | Download | only in src 
      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_INDIC_MACHINE_HH
     28 #define HB_OT_SHAPE_COMPLEX_INDIC_MACHINE_HH
     29 
     30 #include "hb-private.hh"
     31 
     32 %%{
     33   machine indic_syllable_machine;
     34   alphtype unsigned char;
     35   write data;
     36 }%%
     37 
     38 %%{
     39 
     40 # Same order as enum indic_category_t.  Not sure how to avoid duplication.
     41 X    = 0;
     42 C    = 1;
     43 V    = 2;
     44 N    = 3;
     45 H    = 4;
     46 ZWNJ = 5;
     47 ZWJ  = 6;
     48 M    = 7;
     49 SM   = 8;
     50 VD   = 9;
     51 A    = 10;
     52 PLACEHOLDER = 11;
     53 DOTTEDCIRCLE = 12;
     54 RS    = 13;
     55 Coeng = 14;
     56 Repha = 15;
     57 Ra    = 16;
     58 CM    = 17;
     59 Symbol= 18;
     60 CM2   = 31;
     61 
     62 c = (C | Ra);			# is_consonant
     63 n = ((ZWNJ?.RS)? (N.N?)?);	# is_consonant_modifier
     64 z = ZWJ|ZWNJ;			# is_joiner
     65 h = H | Coeng;			# is_halant_or_coeng
     66 reph = (Ra H | Repha);		# possible reph
     67 
     68 cn = c.ZWJ?.n?;
     69 forced_rakar = ZWJ H ZWJ Ra;
     70 symbol = Symbol.N?;
     71 matra_group = z{0,3}.M.N?.(H | forced_rakar)?;
     72 syllable_tail = (SM.SM?.ZWNJ?)? A{0,3}? VD{0,2};
     73 place_holder = PLACEHOLDER | DOTTEDCIRCLE;
     74 halant_group = (z?.h.(ZWJ.N?)?);
     75 final_halant_group = halant_group | h.ZWNJ;
     76 medial_group = CM?.CM2?;
     77 halant_or_matra_group = (final_halant_group | (h.ZWJ)? matra_group{0,4}) (Coeng (cn|V))?;
     78 
     79 
     80 consonant_syllable =	Repha? (cn.halant_group){0,4} cn medial_group halant_or_matra_group syllable_tail;
     81 vowel_syllable =	reph? V.n? (ZWJ | (halant_group.cn){0,4} medial_group halant_or_matra_group syllable_tail);
     82 standalone_cluster =	(Repha? PLACEHOLDER | reph? DOTTEDCIRCLE).n? (halant_group.cn){0,4} medial_group halant_or_matra_group syllable_tail;
     83 symbol_cluster = 	symbol syllable_tail;
     84 broken_cluster =	reph? n? (halant_group.cn){0,4} medial_group halant_or_matra_group syllable_tail;
     85 other =			any;
     86 
     87 main := |*
     88 	consonant_syllable	=> { found_syllable (consonant_syllable); };
     89 	vowel_syllable		=> { found_syllable (vowel_syllable); };
     90 	standalone_cluster	=> { found_syllable (standalone_cluster); };
     91 	symbol_cluster		=> { found_syllable (symbol_cluster); };
     92 	broken_cluster		=> { found_syllable (broken_cluster); };
     93 	other			=> { found_syllable (non_indic_cluster); };
     94 *|;
     95 
     96 
     97 }%%
     98 
     99 #define found_syllable(syllable_type) \
    100   HB_STMT_START { \
    101     if (0) fprintf (stderr, "syllable %d..%d %s\n", last, p+1, #syllable_type); \
    102     for (unsigned int i = last; i < p+1; i++) \
    103       info[i].syllable() = (syllable_serial << 4) | syllable_type; \
    104     last = p+1; \
    105     syllable_serial++; \
    106     if (unlikely (syllable_serial == 16)) syllable_serial = 1; \
    107   } HB_STMT_END
    108 
    109 static void
    110 find_syllables (hb_buffer_t *buffer)
    111 {
    112   unsigned int p, pe, eof, ts HB_UNUSED, te HB_UNUSED, act HB_UNUSED;
    113   int cs;
    114   hb_glyph_info_t *info = buffer->info;
    115   %%{
    116     write init;
    117     getkey info[p].indic_category();
    118   }%%
    119 
    120   p = 0;
    121   pe = eof = buffer->len;
    122 
    123   unsigned int last = 0;
    124   unsigned int syllable_serial = 1;
    125   %%{
    126     write exec;
    127   }%%
    128 }
    129 
    130 #endif /* HB_OT_SHAPE_COMPLEX_INDIC_MACHINE_HH */
    131