1 /*
2 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
3 *
4 * This code is free software; you can redistribute it and/or modify it
5 * under the terms of the GNU General Public License version 2 only, as
6 * published by the Free Software Foundation. Oracle designates this
7 * particular file as subject to the "Classpath" exception as provided
8 * by Oracle in the LICENSE file that accompanied this code.
9 *
10 * This code is distributed in the hope that it will be useful, but WITHOUT
11 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
12 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
13 * version 2 for more details (a copy is included in the LICENSE file that
14 * accompanied this code).
15 *
16 * You should have received a copy of the GNU General Public License version
17 * 2 along with this work; if not, write to the Free Software Foundation,
18 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
19 *
20 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
21 * or visit www.oracle.com if you need additional information or have any
22 * questions.
23 *
24 */
25
26 /*
27 *
28 * (C) Copyright IBM Corp. 1998-2010 - All Rights Reserved
29 *
30 * Developed at DIT - Government of Bhutan
31 *
32 * Contact person: Pema Geyleg - <pema_geyleg@druknet.bt>
33 *
34 * This file is a modification of the ICU file KhmerReordering.cpp
35 * by Jens Herden and Javier Sola who have given all their possible rights to IBM and the Governement of Bhutan
36 * A first module for Dzongkha was developed by Karunakar under Panlocalisation funding.
37 * Assistance for this module has been received from Namgay Thinley, Christopher Fynn and Javier Sola
38 *
39 */
40
41 //#include <stdio.h>
42 #include "LETypes.h"
43 #include "OpenTypeTables.h"
44 #include "TibetanReordering.h"
45 #include "LEGlyphStorage.h"
46
47
48 U_NAMESPACE_BEGIN
49
50 // Characters that get referred to by name...
51 enum
52 {
53 C_DOTTED_CIRCLE = 0x25CC,
54 C_PRE_NUMBER_MARK = 0x0F3F
55 };
56
57
58 enum
59 {
60 // simple classes, they are used in the statetable (in this file) to control the length of a syllable
61 // they are also used to know where a character should be placed (location in reference to the base character)
62 // and also to know if a character, when independtly displayed, should be displayed with a dotted-circle to
63 // indicate error in syllable construction
64 _xx = TibetanClassTable::CC_RESERVED,
65 _ba = TibetanClassTable::CC_BASE,
66 _sj = TibetanClassTable::CC_SUBJOINED | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
67 _tp = TibetanClassTable::CC_TSA_PHRU | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
68 _ac = TibetanClassTable::CC_A_CHUNG | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
69 _cs = TibetanClassTable::CC_COMP_SANSKRIT | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
70 _ha = TibetanClassTable::CC_HALANTA | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
71 _bv = TibetanClassTable::CC_BELOW_VOWEL | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
72 _av = TibetanClassTable::CC_ABOVE_VOWEL | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
73 _an = TibetanClassTable::CC_ANUSVARA | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
74 _cb = TibetanClassTable::CC_CANDRABINDU | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
75 _vs = TibetanClassTable::CC_VISARGA | TibetanClassTable::CF_DOTTED_CIRCLE| TibetanClassTable::CF_POS_AFTER,
76 _as = TibetanClassTable::CC_ABOVE_S_MARK | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
77 _bs = TibetanClassTable::CC_BELOW_S_MARK | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
78 _di = TibetanClassTable::CC_DIGIT | TibetanClassTable::CF_DIGIT,
79 _pd = TibetanClassTable::CC_PRE_DIGIT_MARK | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_PREDIGIT | TibetanClassTable::CF_POS_BEFORE ,
80 _bd = TibetanClassTable::CC_POST_BELOW_DIGIT_M | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_AFTER
81 };
82
83
84 // Character class tables
85 //_xx Non Combining characters
86 //_ba Base Consonants
87 //_sj Subjoined consonants
88 //_tp Tsa - phru
89 //_ac A-chung, Vowel Lengthening mark
90 //_cs Precomposed Sanskrit vowel + subjoined consonants
91 //_ha Halanta/Virama
92 //_bv Below vowel
93 //_av above vowel
94 //_an Anusvara
95 //_cb Candrabindu
96 //_vs Visaraga/Post mark
97 //_as Upper Stress marks
98 //_bs Lower Stress marks
99 //_di Digit
100 //_pd Number pre combining, Needs reordering
101 //_bd Other number combining marks
102
103 static const TibetanClassTable::CharClass tibetanCharClasses[] =
104 {
105 // 0 1 2 3 4 5 6 7 8 9 a b c d e f
106 _xx, _ba, _xx, _xx, _ba, _ba, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0F00 - 0F0F 0
107 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _bd, _bd, _xx, _xx, _xx, _xx, _xx, _xx, // 0F10 - 0F1F 1
108 _di, _di, _di, _di, _di, _di, _di, _di, _di, _di, _xx, _xx, _xx, _xx, _xx, _xx, // 0F20 - 0F2F 2
109 _xx, _xx, _xx, _xx, _xx, _bs, _xx, _bs, _xx, _tp, _xx, _xx, _xx, _xx, _bd, _pd, // 0F30 - 0F3F 3
110 _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _xx, _ba, _ba, _ba, _ba, _ba, _ba, _ba, // 0F40 - 0F4F 4
111 _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, // 0F50 - 0F5F 5
112 _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _xx, _xx, _xx, _xx, _xx, // 0F60 - 0F6F 6
113 _xx, _ac, _av, _cs, _bv, _bv, _cs, _cs, _cs, _cs, _av, _av, _av, _av, _an, _vs, // 0F70 - 0F7F 7
114 _av, _cs, _cb, _cb, _ha, _xx, _as, _as, _ba, _ba, _ba, _ba, _xx, _xx, _xx, _xx, // 0F80 - 0F8F 8
115 _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _xx, _sj, _sj, _sj, _sj, _sj, _sj, _sj, // 0F90 - 0F9F 9
116 _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, // 0FA0 - 0FAF a
117 _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _xx, _sj, _sj, // 0FB0 - 0FBF b
118 _xx, _xx, _xx, _xx, _xx, _xx, _bs, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0FC0 - 0FCF c
119 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx,// 0FD0 - 0FDF d
120 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0FE0 - 0FEF e
121 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0FF0 - 0FFF f
122 };
123
124
125 //
126 // Tibetan Class Tables
127 //
128
129 //
130 // The range of characters defined in the above table is defined here. For Tibetan 0F00 to 0FFF
131 // Even if the Tibetan range is bigger, most of the characters are not combinable, and therefore treated
132 // as _xx
133 static const TibetanClassTable tibetanClassTable = {0x0F00, 0x0FFF, tibetanCharClasses};
134
135
136 // Below we define how a character in the input string is either in the tibetanCharClasses table
137 // (in which case we get its type back), or an unknown object in which case we get _xx (CC_RESERVED) back
getCharClass(LEUnicode ch) const138 TibetanClassTable::CharClass TibetanClassTable::getCharClass(LEUnicode ch) const
139 {
140 if (ch < firstChar || ch > lastChar) {
141 return CC_RESERVED;
142 }
143
144 return classTable[ch - firstChar];
145 }
146
getTibetanClassTable()147 const TibetanClassTable *TibetanClassTable::getTibetanClassTable()
148 {
149 return &tibetanClassTable;
150 }
151
152
153
154 class TibetanReorderingOutput : public UMemory {
155 private:
156 le_int32 fSyllableCount;
157 le_int32 fOutIndex;
158 LEUnicode *fOutChars;
159
160 LEGlyphStorage &fGlyphStorage;
161
162
163 public:
TibetanReorderingOutput(LEUnicode * outChars,LEGlyphStorage & glyphStorage)164 TibetanReorderingOutput(LEUnicode *outChars, LEGlyphStorage &glyphStorage)
165 : fSyllableCount(0), fOutIndex(0), fOutChars(outChars), fGlyphStorage(glyphStorage)
166 {
167 // nothing else to do...
168 }
169
~TibetanReorderingOutput()170 ~TibetanReorderingOutput()
171 {
172 // nothing to do here...
173 }
174
reset()175 void reset()
176 {
177 fSyllableCount += 1;
178 }
179
writeChar(LEUnicode ch,le_uint32 charIndex,FeatureMask featureMask)180 void writeChar(LEUnicode ch, le_uint32 charIndex, FeatureMask featureMask)
181 {
182 LEErrorCode success = LE_NO_ERROR;
183
184 fOutChars[fOutIndex] = ch;
185
186 fGlyphStorage.setCharIndex(fOutIndex, charIndex, success);
187 fGlyphStorage.setAuxData(fOutIndex, featureMask, success);
188
189 fOutIndex += 1;
190 }
191
getOutputIndex()192 le_int32 getOutputIndex()
193 {
194 return fOutIndex;
195 }
196 };
197
198
199 //TODO remove unused flags
200 #define ccmpFeatureTag LE_CCMP_FEATURE_TAG
201 #define blwfFeatureTag LE_BLWF_FEATURE_TAG
202 #define pstfFeatureTag LE_PSTF_FEATURE_TAG
203 #define presFeatureTag LE_PRES_FEATURE_TAG
204 #define blwsFeatureTag LE_BLWS_FEATURE_TAG
205 #define abvsFeatureTag LE_ABVS_FEATURE_TAG
206 #define pstsFeatureTag LE_PSTS_FEATURE_TAG
207
208 #define blwmFeatureTag LE_BLWM_FEATURE_TAG
209 #define abvmFeatureTag LE_ABVM_FEATURE_TAG
210 #define distFeatureTag LE_DIST_FEATURE_TAG
211
212 #define prefFeatureTag LE_PREF_FEATURE_TAG
213 #define abvfFeatureTag LE_ABVF_FEATURE_TAG
214 #define cligFeatureTag LE_CLIG_FEATURE_TAG
215 #define mkmkFeatureTag LE_MKMK_FEATURE_TAG
216
217 // Shaping features
218 #define prefFeatureMask 0x80000000UL
219 #define blwfFeatureMask 0x40000000UL
220 #define abvfFeatureMask 0x20000000UL
221 #define pstfFeatureMask 0x10000000UL
222 #define presFeatureMask 0x08000000UL
223 #define blwsFeatureMask 0x04000000UL
224 #define abvsFeatureMask 0x02000000UL
225 #define pstsFeatureMask 0x01000000UL
226 #define cligFeatureMask 0x00800000UL
227 #define ccmpFeatureMask 0x00040000UL
228
229 // Positioning features
230 #define distFeatureMask 0x00400000UL
231 #define blwmFeatureMask 0x00200000UL
232 #define abvmFeatureMask 0x00100000UL
233 #define mkmkFeatureMask 0x00080000UL
234
235 #define tagPref (ccmpFeatureMask | prefFeatureMask | presFeatureMask | cligFeatureMask | distFeatureMask)
236 #define tagAbvf (ccmpFeatureMask | abvfFeatureMask | abvsFeatureMask | cligFeatureMask | distFeatureMask | abvmFeatureMask | mkmkFeatureMask)
237 #define tagPstf (ccmpFeatureMask | blwfFeatureMask | blwsFeatureMask | prefFeatureMask | presFeatureMask | pstfFeatureMask | pstsFeatureMask | cligFeatureMask | distFeatureMask | blwmFeatureMask)
238 #define tagBlwf (ccmpFeatureMask | blwfFeatureMask | blwsFeatureMask | cligFeatureMask | distFeatureMask | blwmFeatureMask | mkmkFeatureMask)
239 #define tagDefault (ccmpFeatureMask | prefFeatureMask | blwfFeatureMask | presFeatureMask | blwsFeatureMask | cligFeatureMask | distFeatureMask | abvmFeatureMask | blwmFeatureMask | mkmkFeatureMask)
240
241
242
243 // These are in the order in which the features need to be applied
244 // for correct processing
245 static const FeatureMap featureMap[] =
246 {
247 // Shaping features
248 {ccmpFeatureTag, ccmpFeatureMask},
249 {prefFeatureTag, prefFeatureMask},
250 {blwfFeatureTag, blwfFeatureMask},
251 {abvfFeatureTag, abvfFeatureMask},
252 {pstfFeatureTag, pstfFeatureMask},
253 {presFeatureTag, presFeatureMask},
254 {blwsFeatureTag, blwsFeatureMask},
255 {abvsFeatureTag, abvsFeatureMask},
256 {pstsFeatureTag, pstsFeatureMask},
257 {cligFeatureTag, cligFeatureMask},
258
259 // Positioning features
260 {distFeatureTag, distFeatureMask},
261 {blwmFeatureTag, blwmFeatureMask},
262 {abvmFeatureTag, abvmFeatureMask},
263 {mkmkFeatureTag, mkmkFeatureMask},
264 };
265
266 static const le_int32 featureMapCount = LE_ARRAY_SIZE(featureMap);
267
268 // The stateTable is used to calculate the end (the length) of a well
269 // formed Tibetan Syllable.
270 //
271 // Each horizontal line is ordered exactly the same way as the values in TibetanClassTable
272 // CharClassValues in TibetanReordering.h This coincidence of values allows the
273 // follow up of the table.
274 //
275 // Each line corresponds to a state, which does not necessarily need to be a type
276 // of component... for example, state 2 is a base, with is always a first character
277 // in the syllable, but the state could be produced a consonant of any type when
278 // it is the first character that is analysed (in ground state).
279 //
280 static const le_int8 tibetanStateTable[][TibetanClassTable::CC_COUNT] =
281 {
282
283
284 //Dzongkha state table
285 //xx ba sj tp ac cs ha bv av an cb vs as bs di pd bd
286 { 1, 2, 4, 3, 8, 7, 9, 10, 14, 13, 17, 18, 19, 19, 20, 21, 21,}, // 0 - ground state
287 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,}, // 1 - exit state (or sign to the right of the syllable)
288 {-1, -1, 4, 3, 8, 7, 9, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 2 - Base consonant
289 {-1, -1, 5, -1, 8, 7, -1, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 3 - Tsa phru after base
290 {-1, -1, 4, 6, 8, 7, 9, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 4 - Subjoined consonant after base
291 {-1, -1, 5, -1, 8, 7, -1, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 5 - Subjoined consonant after tsa phru
292 {-1, -1, -1, -1, 8, 7, -1, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 6 - Tsa phru after subjoined consonant
293 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 19, 19, -1, -1, -1,}, // 7 - Pre Composed Sanskrit
294 {-1, -1, -1, -1, -1, -1, -1, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 8 - A-chung
295 {-1, -1, -1, -1, -1, -1, -1, -1, 14, 13, 17, -1, 19, 19, -1, -1, -1,}, // 9 - Halanta
296 {-1, -1, -1, -1, -1, -1, -1, 11, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 10 - below vowel 1
297 {-1, -1, -1, -1, -1, -1, -1, 12, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 11 - below vowel 2
298 {-1, -1, -1, -1, -1, -1, -1, -1, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 12 - below vowel 3
299 {-1, -1, -1, -1, -1, -1, -1, -1, 14, 17, 17, 18, 19, 19, -1, -1, -1,}, // 13 - Anusvara before vowel
300 {-1, -1, -1, -1, -1, -1, -1, -1, 15, 17, 17, 18, 19, 19, -1, -1, -1,}, // 14 - above vowel 1
301 {-1, -1, -1, -1, -1, -1, -1, -1, 16, 17, 17, 18, 19, 19, -1, -1, -1,}, // 15 - above vowel 2
302 {-1, -1, -1, -1, -1, -1, -1, -1, -1, 17, 17, 18, 19, 19, -1, -1, -1,}, // 16 - above vowel 3
303 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 18, 19, 19, -1, -1, -1,}, // 17 - Anusvara or Candrabindu after vowel
304 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 19, 19, -1, -1, -1,}, // 18 - Visarga
305 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,}, // 19 - strss mark
306 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 21, 21,}, // 20 - digit
307 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,}, // 21 - digit mark
308
309
310 };
311
312
getFeatureMap(le_int32 & count)313 const FeatureMap *TibetanReordering::getFeatureMap(le_int32 &count)
314 {
315 count = featureMapCount;
316
317 return featureMap;
318 }
319
320
321 // Given an input string of characters and a location in which to start looking
322 // calculate, using the state table, which one is the last character of the syllable
323 // that starts in the starting position.
findSyllable(const TibetanClassTable * classTable,const LEUnicode * chars,le_int32 prev,le_int32 charCount)324 le_int32 TibetanReordering::findSyllable(const TibetanClassTable *classTable, const LEUnicode *chars, le_int32 prev, le_int32 charCount)
325 {
326 le_int32 cursor = prev;
327 le_int8 state = 0;
328
329 while (cursor < charCount) {
330 TibetanClassTable::CharClass charClass = (classTable->getCharClass(chars[cursor]) & TibetanClassTable::CF_CLASS_MASK);
331
332 state = tibetanStateTable[state][charClass];
333
334 if (state < 0) {
335 break;
336 }
337
338 cursor += 1;
339 }
340
341 return cursor;
342 }
343
344
345 // This is the real reordering function as applied to the Tibetan language
346
reorder(const LEUnicode * chars,le_int32 charCount,le_int32,LEUnicode * outChars,LEGlyphStorage & glyphStorage)347 le_int32 TibetanReordering::reorder(const LEUnicode *chars, le_int32 charCount, le_int32,
348 LEUnicode *outChars, LEGlyphStorage &glyphStorage)
349 {
350 const TibetanClassTable *classTable = TibetanClassTable::getTibetanClassTable();
351
352 TibetanReorderingOutput output(outChars, glyphStorage);
353 TibetanClassTable::CharClass charClass;
354 le_int32 i, prev = 0;
355
356 // This loop only exits when we reach the end of a run, which may contain
357 // several syllables.
358 while (prev < charCount) {
359 le_int32 syllable = findSyllable(classTable, chars, prev, charCount);
360
361 output.reset();
362
363 // shall we add a dotted circle?
364 // If in the position in which the base should be (first char in the string) there is
365 // a character that has the Dotted circle flag (a character that cannot be a base)
366 // then write a dotted circle
367 if (classTable->getCharClass(chars[prev]) & TibetanClassTable::CF_DOTTED_CIRCLE) {
368 output.writeChar(C_DOTTED_CIRCLE, prev, tagDefault);
369 }
370
371 // copy the rest to output, inverting the pre-number mark if present after a digit.
372 for (i = prev; i < syllable; i += 1) {
373 charClass = classTable->getCharClass(chars[i]);
374
375 if ((TibetanClassTable::CF_DIGIT & charClass)
376 && ( classTable->getCharClass(chars[i+1]) & TibetanClassTable::CF_PREDIGIT))
377 {
378 output.writeChar(C_PRE_NUMBER_MARK, i, tagPref);
379 output.writeChar(chars[i], i+1 , tagPref);
380 i += 1;
381 } else {
382 switch (charClass & TibetanClassTable::CF_POS_MASK) {
383
384 // If the present character is a number, and the next character is a pre-number combining mark
385 // then the two characters are reordered
386
387 case TibetanClassTable::CF_POS_ABOVE :
388 output.writeChar(chars[i], i, tagAbvf);
389 break;
390
391 case TibetanClassTable::CF_POS_AFTER :
392 output.writeChar(chars[i], i, tagPstf);
393 break;
394
395 case TibetanClassTable::CF_POS_BELOW :
396 output.writeChar(chars[i], i, tagBlwf);
397 break;
398
399 default:
400 // default - any other characters
401 output.writeChar(chars[i], i, tagDefault);
402 break;
403 } // switch
404 } // if
405 } // for
406
407 prev = syllable; // move the pointer to the start of next syllable
408 }
409
410 return output.getOutputIndex();
411 }
412
413
414 U_NAMESPACE_END
415