| 1 | /* |
| 2 | * Copyright (C) 2015 The Qt Company Ltd |
| 3 | * |
| 4 | * This is part of HarfBuzz, an OpenType Layout engine 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 | |
| 25 | #include "harfbuzz-shaper.h" |
| 26 | #include "harfbuzz-shaper-private.h" |
| 27 | |
| 28 | #include <assert.h> |
| 29 | #include <stdio.h> |
| 30 | |
| 31 | enum MymrCharClassValues |
| 32 | { |
| 33 | Mymr_CC_RESERVED = 0, |
| 34 | Mymr_CC_CONSONANT = 1, /* Consonant of type 1, that has subscript form */ |
| 35 | Mymr_CC_CONSONANT2 = 2, /* Consonant of type 2, that has no subscript form */ |
| 36 | Mymr_CC_NGA = 3, /* Consonant NGA */ |
| 37 | Mymr_CC_YA = 4, /* Consonant YA */ |
| 38 | Mymr_CC_RA = 5, /* Consonant RA */ |
| 39 | Mymr_CC_WA = 6, /* Consonant WA */ |
| 40 | Mymr_CC_HA = 7, /* Consonant HA */ |
| 41 | Mymr_CC_IND_VOWEL = 8, /* Independent vowel */ |
| 42 | Mymr_CC_ZERO_WIDTH_NJ_MARK = 9, /* Zero Width non joiner character (0x200C) */ |
| 43 | Mymr_CC_VIRAMA = 10, /* Subscript consonant combining character */ |
| 44 | Mymr_CC_PRE_VOWEL = 11, /* Dependent vowel, prebase (Vowel e) */ |
| 45 | Mymr_CC_BELOW_VOWEL = 12, /* Dependent vowel, prebase (Vowel u, uu) */ |
| 46 | Mymr_CC_ABOVE_VOWEL = 13, /* Dependent vowel, prebase (Vowel i, ii, ai) */ |
| 47 | Mymr_CC_POST_VOWEL = 14, /* Dependent vowel, prebase (Vowel aa) */ |
| 48 | Mymr_CC_SIGN_ABOVE = 15, |
| 49 | Mymr_CC_SIGN_BELOW = 16, |
| 50 | Mymr_CC_SIGN_AFTER = 17, |
| 51 | Mymr_CC_ZERO_WIDTH_J_MARK = 18, /* Zero width joiner character */ |
| 52 | Mymr_CC_COUNT = 19 /* This is the number of character classes */ |
| 53 | }; |
| 54 | |
| 55 | enum MymrCharClassFlags |
| 56 | { |
| 57 | Mymr_CF_CLASS_MASK = 0x0000FFFF, |
| 58 | |
| 59 | Mymr_CF_CONSONANT = 0x01000000, /* flag to speed up comparing */ |
| 60 | Mymr_CF_MEDIAL = 0x02000000, /* flag to speed up comparing */ |
| 61 | Mymr_CF_IND_VOWEL = 0x04000000, /* flag to speed up comparing */ |
| 62 | Mymr_CF_DEP_VOWEL = 0x08000000, /* flag to speed up comparing */ |
| 63 | Mymr_CF_DOTTED_CIRCLE = 0x10000000, /* add a dotted circle if a character with this flag is the first in a syllable */ |
| 64 | Mymr_CF_VIRAMA = 0x20000000, /* flag to speed up comparing */ |
| 65 | |
| 66 | /* position flags */ |
| 67 | Mymr_CF_POS_BEFORE = 0x00080000, |
| 68 | Mymr_CF_POS_BELOW = 0x00040000, |
| 69 | Mymr_CF_POS_ABOVE = 0x00020000, |
| 70 | Mymr_CF_POS_AFTER = 0x00010000, |
| 71 | Mymr_CF_POS_MASK = 0x000f0000, |
| 72 | |
| 73 | Mymr_CF_AFTER_KINZI = 0x00100000 |
| 74 | }; |
| 75 | |
| 76 | /* Characters that get refrered to by name */ |
| 77 | enum MymrChar |
| 78 | { |
| 79 | Mymr_C_SIGN_ZWNJ = 0x200C, |
| 80 | Mymr_C_SIGN_ZWJ = 0x200D, |
| 81 | Mymr_C_DOTTED_CIRCLE = 0x25CC, |
| 82 | Mymr_C_RA = 0x101B, |
| 83 | Mymr_C_YA = 0x101A, |
| 84 | Mymr_C_NGA = 0x1004, |
| 85 | Mymr_C_VOWEL_E = 0x1031, |
| 86 | Mymr_C_VIRAMA = 0x1039 |
| 87 | }; |
| 88 | |
| 89 | enum |
| 90 | { |
| 91 | Mymr_xx = Mymr_CC_RESERVED, |
| 92 | Mymr_c1 = Mymr_CC_CONSONANT | Mymr_CF_CONSONANT | Mymr_CF_POS_BELOW, |
| 93 | Mymr_c2 = Mymr_CC_CONSONANT2 | Mymr_CF_CONSONANT, |
| 94 | Mymr_ng = Mymr_CC_NGA | Mymr_CF_CONSONANT | Mymr_CF_POS_ABOVE, |
| 95 | Mymr_ya = Mymr_CC_YA | Mymr_CF_CONSONANT | Mymr_CF_MEDIAL | Mymr_CF_POS_AFTER | Mymr_CF_AFTER_KINZI, |
| 96 | Mymr_ra = Mymr_CC_RA | Mymr_CF_CONSONANT | Mymr_CF_MEDIAL | Mymr_CF_POS_BEFORE, |
| 97 | Mymr_wa = Mymr_CC_WA | Mymr_CF_CONSONANT | Mymr_CF_MEDIAL | Mymr_CF_POS_BELOW, |
| 98 | Mymr_ha = Mymr_CC_HA | Mymr_CF_CONSONANT | Mymr_CF_MEDIAL | Mymr_CF_POS_BELOW, |
| 99 | Mymr_id = Mymr_CC_IND_VOWEL | Mymr_CF_IND_VOWEL, |
| 100 | Mymr_vi = Mymr_CC_VIRAMA | Mymr_CF_VIRAMA | Mymr_CF_POS_ABOVE | Mymr_CF_DOTTED_CIRCLE, |
| 101 | Mymr_dl = Mymr_CC_PRE_VOWEL | Mymr_CF_DEP_VOWEL | Mymr_CF_POS_BEFORE | Mymr_CF_DOTTED_CIRCLE | Mymr_CF_AFTER_KINZI, |
| 102 | Mymr_db = Mymr_CC_BELOW_VOWEL | Mymr_CF_DEP_VOWEL | Mymr_CF_POS_BELOW | Mymr_CF_DOTTED_CIRCLE | Mymr_CF_AFTER_KINZI, |
| 103 | Mymr_da = Mymr_CC_ABOVE_VOWEL | Mymr_CF_DEP_VOWEL | Mymr_CF_POS_ABOVE | Mymr_CF_DOTTED_CIRCLE | Mymr_CF_AFTER_KINZI, |
| 104 | Mymr_dr = Mymr_CC_POST_VOWEL | Mymr_CF_DEP_VOWEL | Mymr_CF_POS_AFTER | Mymr_CF_DOTTED_CIRCLE | Mymr_CF_AFTER_KINZI, |
| 105 | Mymr_sa = Mymr_CC_SIGN_ABOVE | Mymr_CF_DOTTED_CIRCLE | Mymr_CF_POS_ABOVE | Mymr_CF_AFTER_KINZI, |
| 106 | Mymr_sb = Mymr_CC_SIGN_BELOW | Mymr_CF_DOTTED_CIRCLE | Mymr_CF_POS_BELOW | Mymr_CF_AFTER_KINZI, |
| 107 | Mymr_sp = Mymr_CC_SIGN_AFTER | Mymr_CF_DOTTED_CIRCLE | Mymr_CF_AFTER_KINZI |
| 108 | }; |
| 109 | |
| 110 | |
| 111 | typedef int MymrCharClass; |
| 112 | |
| 113 | |
| 114 | static const MymrCharClass mymrCharClasses[] = |
| 115 | { |
| 116 | Mymr_c1, Mymr_c1, Mymr_c1, Mymr_c1, Mymr_ng, Mymr_c1, Mymr_c1, Mymr_c1, |
| 117 | Mymr_c1, Mymr_c1, Mymr_c2, Mymr_c1, Mymr_c1, Mymr_c1, Mymr_c1, Mymr_c1, /* 1000 - 100F */ |
| 118 | Mymr_c1, Mymr_c1, Mymr_c1, Mymr_c1, Mymr_c1, Mymr_c1, Mymr_c1, Mymr_c1, |
| 119 | Mymr_c1, Mymr_c1, Mymr_ya, Mymr_ra, Mymr_c1, Mymr_wa, Mymr_c1, Mymr_ha, /* 1010 - 101F */ |
| 120 | Mymr_c2, Mymr_c2, Mymr_xx, Mymr_id, Mymr_id, Mymr_id, Mymr_id, Mymr_id, |
| 121 | Mymr_xx, Mymr_id, Mymr_id, Mymr_xx, Mymr_dr, Mymr_da, Mymr_da, Mymr_db, /* 1020 - 102F */ |
| 122 | Mymr_db, Mymr_dl, Mymr_da, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_sa, Mymr_sb, |
| 123 | Mymr_sp, Mymr_vi, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, /* 1030 - 103F */ |
| 124 | Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, |
| 125 | Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, /* 1040 - 104F */ |
| 126 | Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, |
| 127 | Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, Mymr_xx, /* 1050 - 105F */ |
| 128 | }; |
| 129 | |
| 130 | static MymrCharClass |
| 131 | getMyanmarCharClass (HB_UChar16 ch) |
| 132 | { |
| 133 | if (ch == Mymr_C_SIGN_ZWJ) |
| 134 | return Mymr_CC_ZERO_WIDTH_J_MARK; |
| 135 | |
| 136 | if (ch == Mymr_C_SIGN_ZWNJ) |
| 137 | return Mymr_CC_ZERO_WIDTH_NJ_MARK; |
| 138 | |
| 139 | if (ch < 0x1000 || ch > 0x105f) |
| 140 | return Mymr_CC_RESERVED; |
| 141 | |
| 142 | return mymrCharClasses[ch - 0x1000]; |
| 143 | } |
| 144 | |
| 145 | static const signed char mymrStateTable[][Mymr_CC_COUNT] = |
| 146 | { |
| 147 | /* xx c1, c2 ng ya ra wa ha id zwnj vi dl db da dr sa sb sp zwj */ |
| 148 | { 1, 4, 4, 2, 4, 4, 4, 4, 24, 1, 27, 17, 18, 19, 20, 21, 1, 1, 4}, /* 0 - ground state */ |
| 149 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1}, /* 1 - exit state (or sp to the right of the syllable) */ |
| 150 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 3, 17, 18, 19, 20, 21, -1, -1, 4}, /* 2 - NGA */ |
| 151 | {-1, 4, 4, 4, 4, 4, 4, 4, -1, 23, -1, -1, -1, -1, -1, -1, -1, -1, -1}, /* 3 - Virama after NGA */ |
| 152 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 5, 17, 18, 19, 20, 21, 1, 1, -1}, /* 4 - Base consonant */ |
| 153 | {-2, 6, -2, -2, 7, 8, 9, 10, -2, 23, -2, -2, -2, -2, -2, -2, -2, -2, -2}, /* 5 - First virama */ |
| 154 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 25, 17, 18, 19, 20, 21, -1, -1, -1}, /* 6 - c1 after virama */ |
| 155 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 12, 17, 18, 19, 20, 21, -1, -1, -1}, /* 7 - ya after virama */ |
| 156 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 12, 17, 18, 19, 20, 21, -1, -1, -1}, /* 8 - ra after virama */ |
| 157 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 12, 17, 18, 19, 20, 21, -1, -1, -1}, /* 9 - wa after virama */ |
| 158 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 17, 18, 19, 20, 21, -1, -1, -1}, /* 10 - ha after virama */ |
| 159 | {-1, -1, -1, -1, 7, 8, 9, 10, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1}, /* 11 - Virama after NGA+zwj */ |
| 160 | {-2, -2, -2, -2, -2, -2, 13, 14, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2}, /* 12 - Second virama */ |
| 161 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 15, 17, 18, 19, 20, 21, -1, -1, -1}, /* 13 - wa after virama */ |
| 162 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 17, 18, 19, 20, 21, -1, -1, -1}, /* 14 - ha after virama */ |
| 163 | {-2, -2, -2, -2, -2, -2, -2, 16, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2}, /* 15 - Third virama */ |
| 164 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 17, 18, 19, 20, 21, -1, -1, -1}, /* 16 - ha after virama */ |
| 165 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 20, 21, 1, 1, -1}, /* 17 - dl, Dependent vowel e */ |
| 166 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 19, -1, 21, 1, 1, -1}, /* 18 - db, Dependent vowel u,uu */ |
| 167 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, -1}, /* 19 - da, Dependent vowel i,ii,ai */ |
| 168 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 22, -1, -1, -1, -1, -1, 1, 1, -1}, /* 20 - dr, Dependent vowel aa */ |
| 169 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, -1}, /* 21 - sa, Sign anusvara */ |
| 170 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, 23, -1, -1, -1, -1, -1, -1, -1, -1, -1}, /* 22 - atha */ |
| 171 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, -1}, /* 23 - zwnj for atha */ |
| 172 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, -1}, /* 24 - Independent vowel */ |
| 173 | {-2, -2, -2, -2, 26, 26, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2}, /* 25 - Virama after subscript consonant */ |
| 174 | {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 12, 17, 18, 19, 20, 21, -1, 1, -1}, /* 26 - ra/ya after subscript consonant + virama */ |
| 175 | {-1, 6, -1, -1, 7, 8, 9, 10, -1, 23, -1, -1, -1, -1, -1, -1, -1, -1, -1}, /* 27 - Virama after ground state */ |
| 176 | /* exit state -2 is for invalid order of medials and combination of invalids |
| 177 | with virama where virama should treat as start of next syllable |
| 178 | */ |
| 179 | }; |
| 180 | |
| 181 | |
| 182 | |
| 183 | /*#define MYANMAR_DEBUG */ |
| 184 | #ifdef MYANMAR_DEBUG |
| 185 | #define MMDEBUG qDebug |
| 186 | #else |
| 187 | #define MMDEBUG if(0) printf |
| 188 | #endif |
| 189 | |
| 190 | /* |
| 191 | // Given an input string of characters and a location in which to start looking |
| 192 | // calculate, using the state table, which one is the last character of the syllable |
| 193 | // that starts in the starting position. |
| 194 | */ |
| 195 | static int myanmar_nextSyllableBoundary(const HB_UChar16 *s, int start, int end, HB_Bool *invalid) |
| 196 | { |
| 197 | const HB_UChar16 *uc = s + start; |
| 198 | int state = 0; |
| 199 | int pos = start; |
| 200 | *invalid = FALSE; |
| 201 | |
| 202 | while (pos < end) { |
| 203 | MymrCharClass charClass = getMyanmarCharClass(ch: *uc); |
| 204 | state = mymrStateTable[state][charClass & Mymr_CF_CLASS_MASK]; |
| 205 | if (pos == start) |
| 206 | *invalid = (HB_Bool)(charClass & Mymr_CF_DOTTED_CIRCLE); |
| 207 | |
| 208 | MMDEBUG(format: "state[%d]=%d class=%8x (uc=%4x)" , pos - start, state, charClass, *uc); |
| 209 | |
| 210 | if (state < 0) { |
| 211 | if (state < -1) |
| 212 | --pos; |
| 213 | break; |
| 214 | } |
| 215 | ++uc; |
| 216 | ++pos; |
| 217 | } |
| 218 | return pos; |
| 219 | } |
| 220 | |
| 221 | #ifndef NO_OPENTYPE |
| 222 | /* ###### might have to change order of above and below forms and substitutions, |
| 223 | but according to Unicode below comes before above */ |
| 224 | static const HB_OpenTypeFeature myanmar_features[] = { |
| 225 | { HB_MAKE_TAG('p', 'r', 'e', 'f'), .property: PreFormProperty }, |
| 226 | { HB_MAKE_TAG('b', 'l', 'w', 'f'), .property: BelowFormProperty }, |
| 227 | { HB_MAKE_TAG('a', 'b', 'v', 'f'), .property: AboveFormProperty }, |
| 228 | { HB_MAKE_TAG('p', 's', 't', 'f'), .property: PostFormProperty }, |
| 229 | { HB_MAKE_TAG('p', 'r', 'e', 's'), .property: PreSubstProperty }, |
| 230 | { HB_MAKE_TAG('b', 'l', 'w', 's'), .property: BelowSubstProperty }, |
| 231 | { HB_MAKE_TAG('a', 'b', 'v', 's'), .property: AboveSubstProperty }, |
| 232 | { HB_MAKE_TAG('p', 's', 't', 's'), .property: PostSubstProperty }, |
| 233 | { HB_MAKE_TAG('r', 'l', 'i', 'g'), .property: CligProperty }, /* Myanmar1 uses this instead of the other features */ |
| 234 | { .tag: 0, .property: 0 } |
| 235 | }; |
| 236 | #endif |
| 237 | |
| 238 | |
| 239 | /* |
| 240 | // Visual order before shaping should be: |
| 241 | // |
| 242 | // [Vowel Mark E] |
| 243 | // [Virama + Medial Ra] |
| 244 | // [Base] |
| 245 | // [Virama + Consonant] |
| 246 | // [Nga + Virama] (Kinzi) ### should probably come before post forms (medial ya) |
| 247 | // [Vowels] |
| 248 | // [Marks] |
| 249 | // |
| 250 | // This means that we can keep the logical order apart from having to |
| 251 | // move the pre vowel, medial ra and kinzi |
| 252 | */ |
| 253 | |
| 254 | static HB_Bool myanmar_shape_syllable(HB_Bool openType, HB_ShaperItem *item, HB_Bool invalid) |
| 255 | { |
| 256 | /* |
| 257 | // MMDEBUG("\nsyllable from %d len %d, str='%s'", item->item.pos, item->item.length, |
| 258 | // item->string->mid(item->from, item->length).toUtf8().data()); |
| 259 | */ |
| 260 | |
| 261 | #ifndef NO_OPENTYPE |
| 262 | const int availableGlyphs = item->num_glyphs; |
| 263 | #endif |
| 264 | const HB_UChar16 *uc = item->string + item->item.pos; |
| 265 | int vowel_e = -1; |
| 266 | int kinzi = -1; |
| 267 | int medial_ra = -1; |
| 268 | int base = -1; |
| 269 | int i; |
| 270 | int len = 0; |
| 271 | unsigned short reordered[32]; |
| 272 | unsigned char properties[32]; |
| 273 | enum { |
| 274 | AboveForm = 0x01, |
| 275 | PreForm = 0x02, |
| 276 | PostForm = 0x04, |
| 277 | BelowForm = 0x08 |
| 278 | }; |
| 279 | HB_Bool lastWasVirama = FALSE; |
| 280 | int basePos = -1; |
| 281 | |
| 282 | memset(s: properties, c: 0, n: 32*sizeof(unsigned char)); |
| 283 | |
| 284 | /* according to the table the max length of a syllable should be around 14 chars */ |
| 285 | assert(item->item.length < 32); |
| 286 | |
| 287 | #ifdef MYANMAR_DEBUG |
| 288 | printf("original:" ); |
| 289 | for (i = 0; i < (int)item->item.length; i++) { |
| 290 | printf(" %d: %4x" , i, uc[i]); |
| 291 | } |
| 292 | #endif |
| 293 | for (i = 0; i < (int)item->item.length; ++i) { |
| 294 | HB_UChar16 chr = uc[i]; |
| 295 | |
| 296 | if (chr == Mymr_C_VOWEL_E) { |
| 297 | vowel_e = i; |
| 298 | continue; |
| 299 | } |
| 300 | if (i == 0 |
| 301 | && chr == Mymr_C_NGA |
| 302 | && i + 2 < (int)item->item.length |
| 303 | && uc[i+1] == Mymr_C_VIRAMA) { |
| 304 | int mc = getMyanmarCharClass(ch: uc[i+2]); |
| 305 | /*MMDEBUG("maybe kinzi: mc=%x", mc);*/ |
| 306 | if ((mc & Mymr_CF_CONSONANT) == Mymr_CF_CONSONANT) { |
| 307 | kinzi = i; |
| 308 | continue; |
| 309 | } |
| 310 | } |
| 311 | if (base >= 0 |
| 312 | && chr == Mymr_C_VIRAMA |
| 313 | && i + 1 < (int)item->item.length |
| 314 | && uc[i+1] == Mymr_C_RA) { |
| 315 | medial_ra = i; |
| 316 | continue; |
| 317 | } |
| 318 | if (base < 0) |
| 319 | base = i; |
| 320 | } |
| 321 | |
| 322 | MMDEBUG(format: "\n base=%d, vowel_e=%d, kinzi=%d, medial_ra=%d" , base, vowel_e, kinzi, medial_ra); |
| 323 | /* write vowel_e if found */ |
| 324 | if (vowel_e >= 0) { |
| 325 | reordered[0] = Mymr_C_VOWEL_E; |
| 326 | len = 1; |
| 327 | } |
| 328 | /* write medial_ra */ |
| 329 | if (medial_ra >= 0) { |
| 330 | reordered[len] = Mymr_C_VIRAMA; |
| 331 | reordered[len+1] = Mymr_C_RA; |
| 332 | properties[len] = PreForm; |
| 333 | properties[len+1] = PreForm; |
| 334 | len += 2; |
| 335 | } |
| 336 | |
| 337 | /* shall we add a dotted circle? |
| 338 | If in the position in which the base should be (first char in the string) there is |
| 339 | a character that has the Dotted circle flag (a character that cannot be a base) |
| 340 | then write a dotted circle */ |
| 341 | if (invalid) { |
| 342 | reordered[len] = C_DOTTED_CIRCLE; |
| 343 | ++len; |
| 344 | } |
| 345 | |
| 346 | /* copy the rest of the syllable to the output, inserting the kinzi |
| 347 | at the correct place */ |
| 348 | for (i = 0; i < (int)item->item.length; ++i) { |
| 349 | hb_uint16 chr = uc[i]; |
| 350 | MymrCharClass cc; |
| 351 | if (i == vowel_e) |
| 352 | continue; |
| 353 | if (i == medial_ra || i == kinzi) { |
| 354 | ++i; |
| 355 | continue; |
| 356 | } |
| 357 | |
| 358 | cc = getMyanmarCharClass(ch: uc[i]); |
| 359 | if (kinzi >= 0 && i > base && (cc & Mymr_CF_AFTER_KINZI)) { |
| 360 | reordered[len] = Mymr_C_NGA; |
| 361 | reordered[len+1] = Mymr_C_VIRAMA; |
| 362 | if (len > 0) |
| 363 | properties[len-1] = AboveForm; |
| 364 | properties[len] = AboveForm; |
| 365 | len += 2; |
| 366 | kinzi = -1; |
| 367 | } |
| 368 | |
| 369 | if (lastWasVirama) { |
| 370 | int prop = 0; |
| 371 | switch(cc & Mymr_CF_POS_MASK) { |
| 372 | case Mymr_CF_POS_BEFORE: |
| 373 | prop = PreForm; |
| 374 | break; |
| 375 | case Mymr_CF_POS_BELOW: |
| 376 | prop = BelowForm; |
| 377 | break; |
| 378 | case Mymr_CF_POS_ABOVE: |
| 379 | prop = AboveForm; |
| 380 | break; |
| 381 | case Mymr_CF_POS_AFTER: |
| 382 | prop = PostForm; |
| 383 | break; |
| 384 | default: |
| 385 | break; |
| 386 | } |
| 387 | properties[len-1] = prop; |
| 388 | properties[len] = prop; |
| 389 | if(basePos >= 0 && basePos == len-2) |
| 390 | properties[len-2] = prop; |
| 391 | } |
| 392 | lastWasVirama = (chr == Mymr_C_VIRAMA); |
| 393 | if(i == base) |
| 394 | basePos = len; |
| 395 | |
| 396 | if ((chr != Mymr_C_SIGN_ZWNJ && chr != Mymr_C_SIGN_ZWJ) || !len) { |
| 397 | reordered[len] = chr; |
| 398 | ++len; |
| 399 | } |
| 400 | } |
| 401 | if (kinzi >= 0) { |
| 402 | reordered[len] = Mymr_C_NGA; |
| 403 | reordered[len+1] = Mymr_C_VIRAMA; |
| 404 | properties[len] = AboveForm; |
| 405 | properties[len+1] = AboveForm; |
| 406 | len += 2; |
| 407 | } |
| 408 | |
| 409 | if (!item->font->klass->convertStringToGlyphIndices(item->font, |
| 410 | reordered, len, |
| 411 | item->glyphs, &item->num_glyphs, |
| 412 | item->item.bidiLevel % 2)) |
| 413 | return FALSE; |
| 414 | |
| 415 | MMDEBUG(format: "after shaping: len=%d" , len); |
| 416 | for (i = 0; i < len; i++) { |
| 417 | item->attributes[i].mark = FALSE; |
| 418 | item->attributes[i].clusterStart = FALSE; |
| 419 | item->attributes[i].justification = 0; |
| 420 | item->attributes[i].zeroWidth = FALSE; |
| 421 | MMDEBUG(format: " %d: %4x property=%x" , i, reordered[i], properties[i]); |
| 422 | } |
| 423 | |
| 424 | /* now we have the syllable in the right order, and can start running it through open type. */ |
| 425 | |
| 426 | #ifndef NO_OPENTYPE |
| 427 | if (openType) { |
| 428 | hb_uint32 where[32]; |
| 429 | |
| 430 | for (i = 0; i < len; ++i) { |
| 431 | where[i] = ~(PreSubstProperty |
| 432 | | BelowSubstProperty |
| 433 | | AboveSubstProperty |
| 434 | | PostSubstProperty |
| 435 | | CligProperty |
| 436 | | PositioningProperties); |
| 437 | if (properties[i] & PreForm) |
| 438 | where[i] &= ~PreFormProperty; |
| 439 | if (properties[i] & BelowForm) |
| 440 | where[i] &= ~BelowFormProperty; |
| 441 | if (properties[i] & AboveForm) |
| 442 | where[i] &= ~AboveFormProperty; |
| 443 | if (properties[i] & PostForm) |
| 444 | where[i] &= ~PostFormProperty; |
| 445 | } |
| 446 | |
| 447 | HB_OpenTypeShape(item, properties: where); |
| 448 | if (!HB_OpenTypePosition(item, availableGlyphs, /*doLogClusters*/FALSE)) |
| 449 | return FALSE; |
| 450 | } else |
| 451 | #endif |
| 452 | { |
| 453 | MMDEBUG(format: "Not using openType" ); |
| 454 | HB_HeuristicPosition(item); |
| 455 | } |
| 456 | |
| 457 | item->attributes[0].clusterStart = TRUE; |
| 458 | return TRUE; |
| 459 | } |
| 460 | |
| 461 | HB_Bool HB_MyanmarShape(HB_ShaperItem *item) |
| 462 | { |
| 463 | HB_Bool openType = FALSE; |
| 464 | unsigned short *logClusters = item->log_clusters; |
| 465 | |
| 466 | HB_ShaperItem syllable = *item; |
| 467 | int first_glyph = 0; |
| 468 | |
| 469 | int sstart = item->item.pos; |
| 470 | int end = sstart + item->item.length; |
| 471 | int i = 0; |
| 472 | |
| 473 | assert(item->item.script == HB_Script_Myanmar); |
| 474 | #ifndef NO_OPENTYPE |
| 475 | openType = HB_SelectScript(shaper_item: item, features: myanmar_features); |
| 476 | #endif |
| 477 | |
| 478 | MMDEBUG(format: "myanmar_shape: from %d length %d" , item->item.pos, item->item.length); |
| 479 | while (sstart < end) { |
| 480 | HB_Bool invalid; |
| 481 | int send = myanmar_nextSyllableBoundary(s: item->string, start: sstart, end, invalid: &invalid); |
| 482 | MMDEBUG(format: "syllable from %d, length %d, invalid=%s" , sstart, send-sstart, |
| 483 | invalid ? "TRUE" : "FALSE" ); |
| 484 | syllable.item.pos = sstart; |
| 485 | syllable.item.length = send-sstart; |
| 486 | syllable.glyphs = item->glyphs + first_glyph; |
| 487 | syllable.attributes = item->attributes + first_glyph; |
| 488 | syllable.advances = item->advances + first_glyph; |
| 489 | syllable.offsets = item->offsets + first_glyph; |
| 490 | syllable.num_glyphs = item->num_glyphs - first_glyph; |
| 491 | if (!myanmar_shape_syllable(openType, item: &syllable, invalid)) { |
| 492 | MMDEBUG(format: "syllable shaping failed, syllable requests %d glyphs" , syllable.num_glyphs); |
| 493 | item->num_glyphs += syllable.num_glyphs; |
| 494 | return FALSE; |
| 495 | } |
| 496 | |
| 497 | /* fix logcluster array */ |
| 498 | MMDEBUG(format: "syllable:" ); |
| 499 | for (i = first_glyph; i < first_glyph + (int)syllable.num_glyphs; ++i) |
| 500 | MMDEBUG(format: " %d -> glyph %x" , i, item->glyphs[i]); |
| 501 | MMDEBUG(format: " logclusters:" ); |
| 502 | for (i = sstart; i < send; ++i) { |
| 503 | MMDEBUG(format: " %d -> glyph %d" , i, first_glyph); |
| 504 | logClusters[i-item->item.pos] = first_glyph; |
| 505 | } |
| 506 | sstart = send; |
| 507 | first_glyph += syllable.num_glyphs; |
| 508 | } |
| 509 | item->num_glyphs = first_glyph; |
| 510 | return TRUE; |
| 511 | } |
| 512 | |
| 513 | void HB_MyanmarAttributes(HB_Script script, const HB_UChar16 *text, hb_uint32 from, hb_uint32 len, HB_CharAttributes *attributes) |
| 514 | { |
| 515 | int end = from + len; |
| 516 | const HB_UChar16 *uc = text + from; |
| 517 | hb_uint32 i = 0; |
| 518 | HB_UNUSED(script); |
| 519 | attributes += from; |
| 520 | while (i < len) { |
| 521 | HB_Bool invalid; |
| 522 | hb_uint32 boundary = myanmar_nextSyllableBoundary(s: text, start: from+i, end, invalid: &invalid) - from; |
| 523 | |
| 524 | attributes[i].graphemeBoundary = TRUE; |
| 525 | attributes[i].lineBreak = TRUE; |
| 526 | |
| 527 | if (boundary > len-1) |
| 528 | boundary = len; |
| 529 | i++; |
| 530 | while (i < boundary) { |
| 531 | attributes[i].graphemeBoundary = FALSE; |
| 532 | ++uc; |
| 533 | ++i; |
| 534 | } |
| 535 | assert(i == boundary); |
| 536 | } |
| 537 | } |
| 538 | |
| 539 | |