Wed, 31 Dec 2014 07:22:50 +0100
Correct previous dual key logic pending first delivery installment.
michael@0 | 1 | /* |
michael@0 | 2 | ***************************************************************** |
michael@0 | 3 | * Copyright (c) 2002-2011, International Business Machines Corporation |
michael@0 | 4 | * and others. All Rights Reserved. |
michael@0 | 5 | ***************************************************************** |
michael@0 | 6 | * Date Name Description |
michael@0 | 7 | * 06/06/2002 aliu Creation. |
michael@0 | 8 | ***************************************************************** |
michael@0 | 9 | */ |
michael@0 | 10 | |
michael@0 | 11 | #include "unicode/utypes.h" |
michael@0 | 12 | |
michael@0 | 13 | #if !UCONFIG_NO_TRANSLITERATION |
michael@0 | 14 | |
michael@0 | 15 | #include "unicode/uobject.h" |
michael@0 | 16 | #include "unicode/uscript.h" |
michael@0 | 17 | #include "nultrans.h" |
michael@0 | 18 | #include "anytrans.h" |
michael@0 | 19 | #include "uvector.h" |
michael@0 | 20 | #include "tridpars.h" |
michael@0 | 21 | #include "hash.h" |
michael@0 | 22 | #include "putilimp.h" |
michael@0 | 23 | #include "uinvchar.h" |
michael@0 | 24 | |
michael@0 | 25 | //------------------------------------------------------------ |
michael@0 | 26 | // Constants |
michael@0 | 27 | |
michael@0 | 28 | static const UChar TARGET_SEP = 45; // '-' |
michael@0 | 29 | static const UChar VARIANT_SEP = 47; // '/' |
michael@0 | 30 | static const UChar ANY[] = {65,110,121,0}; // "Any" |
michael@0 | 31 | static const UChar NULL_ID[] = {78,117,108,108,0}; // "Null" |
michael@0 | 32 | static const UChar LATIN_PIVOT[] = {45,76,97,116,105,110,59,76,97,116,105,110,45,0}; // "-Latin;Latin-" |
michael@0 | 33 | |
michael@0 | 34 | //------------------------------------------------------------ |
michael@0 | 35 | |
michael@0 | 36 | U_CDECL_BEGIN |
michael@0 | 37 | /** |
michael@0 | 38 | * Deleter function for Transliterator*. |
michael@0 | 39 | */ |
michael@0 | 40 | static void U_CALLCONV |
michael@0 | 41 | _deleteTransliterator(void *obj) { |
michael@0 | 42 | delete (icu::Transliterator*) obj; |
michael@0 | 43 | } |
michael@0 | 44 | U_CDECL_END |
michael@0 | 45 | |
michael@0 | 46 | //------------------------------------------------------------ |
michael@0 | 47 | |
michael@0 | 48 | U_NAMESPACE_BEGIN |
michael@0 | 49 | |
michael@0 | 50 | //------------------------------------------------------------ |
michael@0 | 51 | // ScriptRunIterator |
michael@0 | 52 | |
michael@0 | 53 | /** |
michael@0 | 54 | * Returns a series of ranges corresponding to scripts. They will be |
michael@0 | 55 | * of the form: |
michael@0 | 56 | * |
michael@0 | 57 | * ccccSScSSccccTTcTcccc - c = common, S = first script, T = second |
michael@0 | 58 | * | | - first run (start, limit) |
michael@0 | 59 | * | | - second run (start, limit) |
michael@0 | 60 | * |
michael@0 | 61 | * That is, the runs will overlap. The reason for this is so that a |
michael@0 | 62 | * transliterator can consider common characters both before and after |
michael@0 | 63 | * the scripts. |
michael@0 | 64 | */ |
michael@0 | 65 | class ScriptRunIterator : public UMemory { |
michael@0 | 66 | private: |
michael@0 | 67 | const Replaceable& text; |
michael@0 | 68 | int32_t textStart; |
michael@0 | 69 | int32_t textLimit; |
michael@0 | 70 | |
michael@0 | 71 | public: |
michael@0 | 72 | /** |
michael@0 | 73 | * The code of the current run, valid after next() returns. May |
michael@0 | 74 | * be USCRIPT_INVALID_CODE if and only if the entire text is |
michael@0 | 75 | * COMMON/INHERITED. |
michael@0 | 76 | */ |
michael@0 | 77 | UScriptCode scriptCode; |
michael@0 | 78 | |
michael@0 | 79 | /** |
michael@0 | 80 | * The start of the run, inclusive, valid after next() returns. |
michael@0 | 81 | */ |
michael@0 | 82 | int32_t start; |
michael@0 | 83 | |
michael@0 | 84 | /** |
michael@0 | 85 | * The end of the run, exclusive, valid after next() returns. |
michael@0 | 86 | */ |
michael@0 | 87 | int32_t limit; |
michael@0 | 88 | |
michael@0 | 89 | /** |
michael@0 | 90 | * Constructs a run iterator over the given text from start |
michael@0 | 91 | * (inclusive) to limit (exclusive). |
michael@0 | 92 | */ |
michael@0 | 93 | ScriptRunIterator(const Replaceable& text, int32_t start, int32_t limit); |
michael@0 | 94 | |
michael@0 | 95 | /** |
michael@0 | 96 | * Returns TRUE if there are any more runs. TRUE is always |
michael@0 | 97 | * returned at least once. Upon return, the caller should |
michael@0 | 98 | * examine scriptCode, start, and limit. |
michael@0 | 99 | */ |
michael@0 | 100 | UBool next(); |
michael@0 | 101 | |
michael@0 | 102 | /** |
michael@0 | 103 | * Adjusts internal indices for a change in the limit index of the |
michael@0 | 104 | * given delta. A positive delta means the limit has increased. |
michael@0 | 105 | */ |
michael@0 | 106 | void adjustLimit(int32_t delta); |
michael@0 | 107 | |
michael@0 | 108 | private: |
michael@0 | 109 | ScriptRunIterator(const ScriptRunIterator &other); // forbid copying of this class |
michael@0 | 110 | ScriptRunIterator &operator=(const ScriptRunIterator &other); // forbid copying of this class |
michael@0 | 111 | }; |
michael@0 | 112 | |
michael@0 | 113 | ScriptRunIterator::ScriptRunIterator(const Replaceable& theText, |
michael@0 | 114 | int32_t myStart, int32_t myLimit) : |
michael@0 | 115 | text(theText) |
michael@0 | 116 | { |
michael@0 | 117 | textStart = myStart; |
michael@0 | 118 | textLimit = myLimit; |
michael@0 | 119 | limit = myStart; |
michael@0 | 120 | } |
michael@0 | 121 | |
michael@0 | 122 | UBool ScriptRunIterator::next() { |
michael@0 | 123 | UChar32 ch; |
michael@0 | 124 | UScriptCode s; |
michael@0 | 125 | UErrorCode ec = U_ZERO_ERROR; |
michael@0 | 126 | |
michael@0 | 127 | scriptCode = USCRIPT_INVALID_CODE; // don't know script yet |
michael@0 | 128 | start = limit; |
michael@0 | 129 | |
michael@0 | 130 | // Are we done? |
michael@0 | 131 | if (start == textLimit) { |
michael@0 | 132 | return FALSE; |
michael@0 | 133 | } |
michael@0 | 134 | |
michael@0 | 135 | // Move start back to include adjacent COMMON or INHERITED |
michael@0 | 136 | // characters |
michael@0 | 137 | while (start > textStart) { |
michael@0 | 138 | ch = text.char32At(start - 1); // look back |
michael@0 | 139 | s = uscript_getScript(ch, &ec); |
michael@0 | 140 | if (s == USCRIPT_COMMON || s == USCRIPT_INHERITED) { |
michael@0 | 141 | --start; |
michael@0 | 142 | } else { |
michael@0 | 143 | break; |
michael@0 | 144 | } |
michael@0 | 145 | } |
michael@0 | 146 | |
michael@0 | 147 | // Move limit ahead to include COMMON, INHERITED, and characters |
michael@0 | 148 | // of the current script. |
michael@0 | 149 | while (limit < textLimit) { |
michael@0 | 150 | ch = text.char32At(limit); // look ahead |
michael@0 | 151 | s = uscript_getScript(ch, &ec); |
michael@0 | 152 | if (s != USCRIPT_COMMON && s != USCRIPT_INHERITED) { |
michael@0 | 153 | if (scriptCode == USCRIPT_INVALID_CODE) { |
michael@0 | 154 | scriptCode = s; |
michael@0 | 155 | } else if (s != scriptCode) { |
michael@0 | 156 | break; |
michael@0 | 157 | } |
michael@0 | 158 | } |
michael@0 | 159 | ++limit; |
michael@0 | 160 | } |
michael@0 | 161 | |
michael@0 | 162 | // Return TRUE even if the entire text is COMMON / INHERITED, in |
michael@0 | 163 | // which case scriptCode will be USCRIPT_INVALID_CODE. |
michael@0 | 164 | return TRUE; |
michael@0 | 165 | } |
michael@0 | 166 | |
michael@0 | 167 | void ScriptRunIterator::adjustLimit(int32_t delta) { |
michael@0 | 168 | limit += delta; |
michael@0 | 169 | textLimit += delta; |
michael@0 | 170 | } |
michael@0 | 171 | |
michael@0 | 172 | //------------------------------------------------------------ |
michael@0 | 173 | // AnyTransliterator |
michael@0 | 174 | |
michael@0 | 175 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(AnyTransliterator) |
michael@0 | 176 | |
michael@0 | 177 | AnyTransliterator::AnyTransliterator(const UnicodeString& id, |
michael@0 | 178 | const UnicodeString& theTarget, |
michael@0 | 179 | const UnicodeString& theVariant, |
michael@0 | 180 | UScriptCode theTargetScript, |
michael@0 | 181 | UErrorCode& ec) : |
michael@0 | 182 | Transliterator(id, NULL), |
michael@0 | 183 | targetScript(theTargetScript) |
michael@0 | 184 | { |
michael@0 | 185 | cache = uhash_open(uhash_hashLong, uhash_compareLong, NULL, &ec); |
michael@0 | 186 | if (U_FAILURE(ec)) { |
michael@0 | 187 | return; |
michael@0 | 188 | } |
michael@0 | 189 | uhash_setValueDeleter(cache, _deleteTransliterator); |
michael@0 | 190 | |
michael@0 | 191 | target = theTarget; |
michael@0 | 192 | if (theVariant.length() > 0) { |
michael@0 | 193 | target.append(VARIANT_SEP).append(theVariant); |
michael@0 | 194 | } |
michael@0 | 195 | } |
michael@0 | 196 | |
michael@0 | 197 | AnyTransliterator::~AnyTransliterator() { |
michael@0 | 198 | uhash_close(cache); |
michael@0 | 199 | } |
michael@0 | 200 | |
michael@0 | 201 | /** |
michael@0 | 202 | * Copy constructor. |
michael@0 | 203 | */ |
michael@0 | 204 | AnyTransliterator::AnyTransliterator(const AnyTransliterator& o) : |
michael@0 | 205 | Transliterator(o), |
michael@0 | 206 | target(o.target), |
michael@0 | 207 | targetScript(o.targetScript) |
michael@0 | 208 | { |
michael@0 | 209 | // Don't copy the cache contents |
michael@0 | 210 | UErrorCode ec = U_ZERO_ERROR; |
michael@0 | 211 | cache = uhash_open(uhash_hashLong, uhash_compareLong, NULL, &ec); |
michael@0 | 212 | if (U_FAILURE(ec)) { |
michael@0 | 213 | return; |
michael@0 | 214 | } |
michael@0 | 215 | uhash_setValueDeleter(cache, _deleteTransliterator); |
michael@0 | 216 | } |
michael@0 | 217 | |
michael@0 | 218 | /** |
michael@0 | 219 | * Transliterator API. |
michael@0 | 220 | */ |
michael@0 | 221 | Transliterator* AnyTransliterator::clone() const { |
michael@0 | 222 | return new AnyTransliterator(*this); |
michael@0 | 223 | } |
michael@0 | 224 | |
michael@0 | 225 | /** |
michael@0 | 226 | * Implements {@link Transliterator#handleTransliterate}. |
michael@0 | 227 | */ |
michael@0 | 228 | void AnyTransliterator::handleTransliterate(Replaceable& text, UTransPosition& pos, |
michael@0 | 229 | UBool isIncremental) const { |
michael@0 | 230 | int32_t allStart = pos.start; |
michael@0 | 231 | int32_t allLimit = pos.limit; |
michael@0 | 232 | |
michael@0 | 233 | ScriptRunIterator it(text, pos.contextStart, pos.contextLimit); |
michael@0 | 234 | |
michael@0 | 235 | while (it.next()) { |
michael@0 | 236 | // Ignore runs in the ante context |
michael@0 | 237 | if (it.limit <= allStart) continue; |
michael@0 | 238 | |
michael@0 | 239 | // Try to instantiate transliterator from it.scriptCode to |
michael@0 | 240 | // our target or target/variant |
michael@0 | 241 | Transliterator* t = getTransliterator(it.scriptCode); |
michael@0 | 242 | |
michael@0 | 243 | if (t == NULL) { |
michael@0 | 244 | // We have no transliterator. Do nothing, but keep |
michael@0 | 245 | // pos.start up to date. |
michael@0 | 246 | pos.start = it.limit; |
michael@0 | 247 | continue; |
michael@0 | 248 | } |
michael@0 | 249 | |
michael@0 | 250 | // If the run end is before the transliteration limit, do |
michael@0 | 251 | // a non-incremental transliteration. Otherwise do an |
michael@0 | 252 | // incremental one. |
michael@0 | 253 | UBool incremental = isIncremental && (it.limit >= allLimit); |
michael@0 | 254 | |
michael@0 | 255 | pos.start = uprv_max(allStart, it.start); |
michael@0 | 256 | pos.limit = uprv_min(allLimit, it.limit); |
michael@0 | 257 | int32_t limit = pos.limit; |
michael@0 | 258 | t->filteredTransliterate(text, pos, incremental); |
michael@0 | 259 | int32_t delta = pos.limit - limit; |
michael@0 | 260 | allLimit += delta; |
michael@0 | 261 | it.adjustLimit(delta); |
michael@0 | 262 | |
michael@0 | 263 | // We're done if we enter the post context |
michael@0 | 264 | if (it.limit >= allLimit) break; |
michael@0 | 265 | } |
michael@0 | 266 | |
michael@0 | 267 | // Restore limit. pos.start is fine where the last transliterator |
michael@0 | 268 | // left it, or at the end of the last run. |
michael@0 | 269 | pos.limit = allLimit; |
michael@0 | 270 | } |
michael@0 | 271 | |
michael@0 | 272 | Transliterator* AnyTransliterator::getTransliterator(UScriptCode source) const { |
michael@0 | 273 | |
michael@0 | 274 | if (source == targetScript || source == USCRIPT_INVALID_CODE) { |
michael@0 | 275 | return NULL; |
michael@0 | 276 | } |
michael@0 | 277 | |
michael@0 | 278 | Transliterator* t = (Transliterator*) uhash_iget(cache, (int32_t) source); |
michael@0 | 279 | if (t == NULL) { |
michael@0 | 280 | UErrorCode ec = U_ZERO_ERROR; |
michael@0 | 281 | UnicodeString sourceName(uscript_getName(source), -1, US_INV); |
michael@0 | 282 | UnicodeString id(sourceName); |
michael@0 | 283 | id.append(TARGET_SEP).append(target); |
michael@0 | 284 | |
michael@0 | 285 | t = Transliterator::createInstance(id, UTRANS_FORWARD, ec); |
michael@0 | 286 | if (U_FAILURE(ec) || t == NULL) { |
michael@0 | 287 | delete t; |
michael@0 | 288 | |
michael@0 | 289 | // Try to pivot around Latin, our most common script |
michael@0 | 290 | id = sourceName; |
michael@0 | 291 | id.append(LATIN_PIVOT, -1).append(target); |
michael@0 | 292 | t = Transliterator::createInstance(id, UTRANS_FORWARD, ec); |
michael@0 | 293 | if (U_FAILURE(ec) || t == NULL) { |
michael@0 | 294 | delete t; |
michael@0 | 295 | t = NULL; |
michael@0 | 296 | } |
michael@0 | 297 | } |
michael@0 | 298 | |
michael@0 | 299 | if (t != NULL) { |
michael@0 | 300 | uhash_iput(cache, (int32_t) source, t, &ec); |
michael@0 | 301 | } |
michael@0 | 302 | } |
michael@0 | 303 | |
michael@0 | 304 | return t; |
michael@0 | 305 | } |
michael@0 | 306 | |
michael@0 | 307 | /** |
michael@0 | 308 | * Return the script code for a given name, or -1 if not found. |
michael@0 | 309 | */ |
michael@0 | 310 | static UScriptCode scriptNameToCode(const UnicodeString& name) { |
michael@0 | 311 | char buf[128]; |
michael@0 | 312 | UScriptCode code; |
michael@0 | 313 | UErrorCode ec = U_ZERO_ERROR; |
michael@0 | 314 | int32_t nameLen = name.length(); |
michael@0 | 315 | UBool isInvariant = uprv_isInvariantUString(name.getBuffer(), nameLen); |
michael@0 | 316 | |
michael@0 | 317 | if (isInvariant) { |
michael@0 | 318 | name.extract(0, nameLen, buf, (int32_t)sizeof(buf), US_INV); |
michael@0 | 319 | buf[127] = 0; // Make sure that we NULL terminate the string. |
michael@0 | 320 | } |
michael@0 | 321 | if (!isInvariant || uscript_getCode(buf, &code, 1, &ec) != 1 || U_FAILURE(ec)) |
michael@0 | 322 | { |
michael@0 | 323 | code = USCRIPT_INVALID_CODE; |
michael@0 | 324 | } |
michael@0 | 325 | return code; |
michael@0 | 326 | } |
michael@0 | 327 | |
michael@0 | 328 | /** |
michael@0 | 329 | * Registers standard transliterators with the system. Called by |
michael@0 | 330 | * Transliterator during initialization. Scan all current targets and |
michael@0 | 331 | * register those that are scripts T as Any-T/V. |
michael@0 | 332 | */ |
michael@0 | 333 | void AnyTransliterator::registerIDs() { |
michael@0 | 334 | |
michael@0 | 335 | UErrorCode ec = U_ZERO_ERROR; |
michael@0 | 336 | Hashtable seen(TRUE, ec); |
michael@0 | 337 | |
michael@0 | 338 | int32_t sourceCount = Transliterator::_countAvailableSources(); |
michael@0 | 339 | for (int32_t s=0; s<sourceCount; ++s) { |
michael@0 | 340 | UnicodeString source; |
michael@0 | 341 | Transliterator::_getAvailableSource(s, source); |
michael@0 | 342 | |
michael@0 | 343 | // Ignore the "Any" source |
michael@0 | 344 | if (source.caseCompare(ANY, 3, 0 /*U_FOLD_CASE_DEFAULT*/) == 0) continue; |
michael@0 | 345 | |
michael@0 | 346 | int32_t targetCount = Transliterator::_countAvailableTargets(source); |
michael@0 | 347 | for (int32_t t=0; t<targetCount; ++t) { |
michael@0 | 348 | UnicodeString target; |
michael@0 | 349 | Transliterator::_getAvailableTarget(t, source, target); |
michael@0 | 350 | |
michael@0 | 351 | // Only process each target once |
michael@0 | 352 | if (seen.geti(target) != 0) continue; |
michael@0 | 353 | ec = U_ZERO_ERROR; |
michael@0 | 354 | seen.puti(target, 1, ec); |
michael@0 | 355 | |
michael@0 | 356 | // Get the script code for the target. If not a script, ignore. |
michael@0 | 357 | UScriptCode targetScript = scriptNameToCode(target); |
michael@0 | 358 | if (targetScript == USCRIPT_INVALID_CODE) continue; |
michael@0 | 359 | |
michael@0 | 360 | int32_t variantCount = Transliterator::_countAvailableVariants(source, target); |
michael@0 | 361 | // assert(variantCount >= 1); |
michael@0 | 362 | for (int32_t v=0; v<variantCount; ++v) { |
michael@0 | 363 | UnicodeString variant; |
michael@0 | 364 | Transliterator::_getAvailableVariant(v, source, target, variant); |
michael@0 | 365 | |
michael@0 | 366 | UnicodeString id; |
michael@0 | 367 | TransliteratorIDParser::STVtoID(UnicodeString(TRUE, ANY, 3), target, variant, id); |
michael@0 | 368 | ec = U_ZERO_ERROR; |
michael@0 | 369 | AnyTransliterator* t = new AnyTransliterator(id, target, variant, |
michael@0 | 370 | targetScript, ec); |
michael@0 | 371 | if (U_FAILURE(ec)) { |
michael@0 | 372 | delete t; |
michael@0 | 373 | } else { |
michael@0 | 374 | Transliterator::_registerInstance(t); |
michael@0 | 375 | Transliterator::_registerSpecialInverse(target, UnicodeString(TRUE, NULL_ID, 4), FALSE); |
michael@0 | 376 | } |
michael@0 | 377 | } |
michael@0 | 378 | } |
michael@0 | 379 | } |
michael@0 | 380 | } |
michael@0 | 381 | |
michael@0 | 382 | U_NAMESPACE_END |
michael@0 | 383 | |
michael@0 | 384 | #endif /* #if !UCONFIG_NO_TRANSLITERATION */ |
michael@0 | 385 | |
michael@0 | 386 | //eof |