| 1 | // © 2016 and later: Unicode, Inc. and others. | 
|---|
| 2 | // License & terms of use: http://www.unicode.org/copyright.html | 
|---|
| 3 | // | 
|---|
| 4 | //  rbbisetb.cpp | 
|---|
| 5 | // | 
|---|
| 6 | /* | 
|---|
| 7 | *************************************************************************** | 
|---|
| 8 | *   Copyright (C) 2002-2008 International Business Machines Corporation   * | 
|---|
| 9 | *   and others. All rights reserved.                                      * | 
|---|
| 10 | *************************************************************************** | 
|---|
| 11 | */ | 
|---|
| 12 | // | 
|---|
| 13 | //  RBBISetBuilder   Handles processing of Unicode Sets from RBBI rules | 
|---|
| 14 | //                   (part of the rule building process.) | 
|---|
| 15 | // | 
|---|
| 16 | //      Starting with the rules parse tree from the scanner, | 
|---|
| 17 | // | 
|---|
| 18 | //                   -  Enumerate the set of UnicodeSets that are referenced | 
|---|
| 19 | //                      by the RBBI rules. | 
|---|
| 20 | //                   -  compute a set of non-overlapping character ranges | 
|---|
| 21 | //                      with all characters within a range belonging to the same | 
|---|
| 22 | //                      set of input unicode sets. | 
|---|
| 23 | //                   -  Derive a set of non-overlapping UnicodeSet (like things) | 
|---|
| 24 | //                      that will correspond to columns in the state table for | 
|---|
| 25 | //                      the RBBI execution engine.  All characters within one | 
|---|
| 26 | //                      of these sets belong to the same set of the original | 
|---|
| 27 | //                      UnicodeSets from the user's rules. | 
|---|
| 28 | //                   -  construct the trie table that maps input characters | 
|---|
| 29 | //                      to the index of the matching non-overlapping set of set from | 
|---|
| 30 | //                      the previous step. | 
|---|
| 31 | // | 
|---|
| 32 |  | 
|---|
| 33 | #include "unicode/utypes.h" | 
|---|
| 34 |  | 
|---|
| 35 | #if !UCONFIG_NO_BREAK_ITERATION | 
|---|
| 36 |  | 
|---|
| 37 | #include "unicode/uniset.h" | 
|---|
| 38 | #include "uvector.h" | 
|---|
| 39 | #include "uassert.h" | 
|---|
| 40 | #include "cmemory.h" | 
|---|
| 41 | #include "cstring.h" | 
|---|
| 42 |  | 
|---|
| 43 | #include "rbbisetb.h" | 
|---|
| 44 | #include "rbbinode.h" | 
|---|
| 45 |  | 
|---|
| 46 | U_NAMESPACE_BEGIN | 
|---|
| 47 |  | 
|---|
| 48 | const int32_t kMaxCharCategoriesFor8BitsTrie = 255; | 
|---|
| 49 | //------------------------------------------------------------------------ | 
|---|
| 50 | // | 
|---|
| 51 | //   Constructor | 
|---|
| 52 | // | 
|---|
| 53 | //------------------------------------------------------------------------ | 
|---|
| 54 | RBBISetBuilder::RBBISetBuilder(RBBIRuleBuilder *rb) | 
|---|
| 55 | { | 
|---|
| 56 | fRB             = rb; | 
|---|
| 57 | fStatus         = rb->fStatus; | 
|---|
| 58 | fRangeList      = nullptr; | 
|---|
| 59 | fMutableTrie    = nullptr; | 
|---|
| 60 | fTrie           = nullptr; | 
|---|
| 61 | fTrieSize       = 0; | 
|---|
| 62 | fGroupCount     = 0; | 
|---|
| 63 | fSawBOF         = false; | 
|---|
| 64 | } | 
|---|
| 65 |  | 
|---|
| 66 |  | 
|---|
| 67 | //------------------------------------------------------------------------ | 
|---|
| 68 | // | 
|---|
| 69 | //   Destructor | 
|---|
| 70 | // | 
|---|
| 71 | //------------------------------------------------------------------------ | 
|---|
| 72 | RBBISetBuilder::~RBBISetBuilder() | 
|---|
| 73 | { | 
|---|
| 74 | RangeDescriptor   *; | 
|---|
| 75 |  | 
|---|
| 76 | // Walk through & delete the linked list of RangeDescriptors | 
|---|
| 77 | for (nextRangeDesc = fRangeList; nextRangeDesc!=nullptr;) { | 
|---|
| 78 | RangeDescriptor *r = nextRangeDesc; | 
|---|
| 79 | nextRangeDesc      = r->fNext; | 
|---|
| 80 | delete r; | 
|---|
| 81 | } | 
|---|
| 82 |  | 
|---|
| 83 | ucptrie_close(fTrie); | 
|---|
| 84 | umutablecptrie_close(fMutableTrie); | 
|---|
| 85 | } | 
|---|
| 86 |  | 
|---|
| 87 |  | 
|---|
| 88 |  | 
|---|
| 89 |  | 
|---|
| 90 | //------------------------------------------------------------------------ | 
|---|
| 91 | // | 
|---|
| 92 | //   build          Build the list of non-overlapping character ranges | 
|---|
| 93 | //                  from the Unicode Sets. | 
|---|
| 94 | // | 
|---|
| 95 | //------------------------------------------------------------------------ | 
|---|
| 96 | void RBBISetBuilder::buildRanges() { | 
|---|
| 97 | RBBINode        *usetNode; | 
|---|
| 98 | RangeDescriptor *rlRange; | 
|---|
| 99 |  | 
|---|
| 100 | if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "usets")) {printSets();} | 
|---|
| 101 |  | 
|---|
| 102 | // | 
|---|
| 103 | //  Initialize the process by creating a single range encompassing all characters | 
|---|
| 104 | //  that is in no sets. | 
|---|
| 105 | // | 
|---|
| 106 | fRangeList                = new RangeDescriptor(*fStatus); // will check for status here | 
|---|
| 107 | if (fRangeList == nullptr) { | 
|---|
| 108 | *fStatus = U_MEMORY_ALLOCATION_ERROR; | 
|---|
| 109 | return; | 
|---|
| 110 | } | 
|---|
| 111 | fRangeList->fStartChar    = 0; | 
|---|
| 112 | fRangeList->fEndChar      = 0x10ffff; | 
|---|
| 113 |  | 
|---|
| 114 | if (U_FAILURE(*fStatus)) { | 
|---|
| 115 | return; | 
|---|
| 116 | } | 
|---|
| 117 |  | 
|---|
| 118 | // | 
|---|
| 119 | //  Find the set of non-overlapping ranges of characters | 
|---|
| 120 | // | 
|---|
| 121 | int  ni; | 
|---|
| 122 | for (ni=0; ; ni++) {        // Loop over each of the UnicodeSets encountered in the input rules | 
|---|
| 123 | usetNode = (RBBINode *)this->fRB->fUSetNodes->elementAt(ni); | 
|---|
| 124 | if (usetNode==nullptr) { | 
|---|
| 125 | break; | 
|---|
| 126 | } | 
|---|
| 127 |  | 
|---|
| 128 | UnicodeSet      *inputSet             = usetNode->fInputSet; | 
|---|
| 129 | int32_t          inputSetRangeCount   = inputSet->getRangeCount(); | 
|---|
| 130 | int              inputSetRangeIndex   = 0; | 
|---|
| 131 | rlRange              = fRangeList; | 
|---|
| 132 |  | 
|---|
| 133 | for (;;) { | 
|---|
| 134 | if (inputSetRangeIndex >= inputSetRangeCount) { | 
|---|
| 135 | break; | 
|---|
| 136 | } | 
|---|
| 137 | UChar32      inputSetRangeBegin  = inputSet->getRangeStart(inputSetRangeIndex); | 
|---|
| 138 | UChar32      inputSetRangeEnd    = inputSet->getRangeEnd(inputSetRangeIndex); | 
|---|
| 139 |  | 
|---|
| 140 | // skip over ranges from the range list that are completely | 
|---|
| 141 | //   below the current range from the input unicode set. | 
|---|
| 142 | while (rlRange->fEndChar < inputSetRangeBegin) { | 
|---|
| 143 | rlRange = rlRange->fNext; | 
|---|
| 144 | } | 
|---|
| 145 |  | 
|---|
| 146 | // If the start of the range from the range list is before with | 
|---|
| 147 | //   the start of the range from the unicode set, split the range list range | 
|---|
| 148 | //   in two, with one part being before (wholly outside of) the unicode set | 
|---|
| 149 | //   and the other containing the rest. | 
|---|
| 150 | //   Then continue the loop; the post-split current range will then be skipped | 
|---|
| 151 | //     over | 
|---|
| 152 | if (rlRange->fStartChar < inputSetRangeBegin) { | 
|---|
| 153 | rlRange->split(inputSetRangeBegin, *fStatus); | 
|---|
| 154 | if (U_FAILURE(*fStatus)) { | 
|---|
| 155 | return; | 
|---|
| 156 | } | 
|---|
| 157 | continue; | 
|---|
| 158 | } | 
|---|
| 159 |  | 
|---|
| 160 | // Same thing at the end of the ranges... | 
|---|
| 161 | // If the end of the range from the range list doesn't coincide with | 
|---|
| 162 | //   the end of the range from the unicode set, split the range list | 
|---|
| 163 | //   range in two.  The first part of the split range will be | 
|---|
| 164 | //   wholly inside the Unicode set. | 
|---|
| 165 | if (rlRange->fEndChar > inputSetRangeEnd) { | 
|---|
| 166 | rlRange->split(inputSetRangeEnd+1, *fStatus); | 
|---|
| 167 | if (U_FAILURE(*fStatus)) { | 
|---|
| 168 | return; | 
|---|
| 169 | } | 
|---|
| 170 | } | 
|---|
| 171 |  | 
|---|
| 172 | // The current rlRange is now entirely within the UnicodeSet range. | 
|---|
| 173 | // Add this unicode set to the list of sets for this rlRange | 
|---|
| 174 | if (rlRange->fIncludesSets->indexOf(usetNode) == -1) { | 
|---|
| 175 | rlRange->fIncludesSets->addElement(usetNode, *fStatus); | 
|---|
| 176 | if (U_FAILURE(*fStatus)) { | 
|---|
| 177 | return; | 
|---|
| 178 | } | 
|---|
| 179 | } | 
|---|
| 180 |  | 
|---|
| 181 | // Advance over ranges that we are finished with. | 
|---|
| 182 | if (inputSetRangeEnd == rlRange->fEndChar) { | 
|---|
| 183 | inputSetRangeIndex++; | 
|---|
| 184 | } | 
|---|
| 185 | rlRange = rlRange->fNext; | 
|---|
| 186 | } | 
|---|
| 187 | } | 
|---|
| 188 |  | 
|---|
| 189 | if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "range")) { printRanges();} | 
|---|
| 190 |  | 
|---|
| 191 | // | 
|---|
| 192 | //  Group the above ranges, with each group consisting of one or more | 
|---|
| 193 | //    ranges that are in exactly the same set of original UnicodeSets. | 
|---|
| 194 | //    The groups are numbered, and these group numbers are the set of | 
|---|
| 195 | //    input symbols recognized by the run-time state machine. | 
|---|
| 196 | // | 
|---|
| 197 | //    Numbering: # 0  (state table column 0) is unused. | 
|---|
| 198 | //               # 1  is reserved - table column 1 is for end-of-input | 
|---|
| 199 | //               # 2  is reserved - table column 2 is for beginning-of-input | 
|---|
| 200 | //               # 3  is the first range list. | 
|---|
| 201 | // | 
|---|
| 202 | RangeDescriptor *rlSearchRange; | 
|---|
| 203 | int32_t dictGroupCount = 0; | 
|---|
| 204 |  | 
|---|
| 205 | for (rlRange = fRangeList; rlRange!=nullptr; rlRange=rlRange->fNext) { | 
|---|
| 206 | for (rlSearchRange=fRangeList; rlSearchRange != rlRange; rlSearchRange=rlSearchRange->fNext) { | 
|---|
| 207 | if (rlRange->fIncludesSets->equals(*rlSearchRange->fIncludesSets)) { | 
|---|
| 208 | rlRange->fNum = rlSearchRange->fNum; | 
|---|
| 209 | rlRange->fIncludesDict = rlSearchRange->fIncludesDict; | 
|---|
| 210 | break; | 
|---|
| 211 | } | 
|---|
| 212 | } | 
|---|
| 213 | if (rlRange->fNum == 0) { | 
|---|
| 214 | rlRange->fFirstInGroup = true; | 
|---|
| 215 | if (rlRange->isDictionaryRange()) { | 
|---|
| 216 | rlRange->fNum = ++dictGroupCount; | 
|---|
| 217 | rlRange->fIncludesDict = true; | 
|---|
| 218 | } else { | 
|---|
| 219 | fGroupCount++; | 
|---|
| 220 | rlRange->fNum = fGroupCount+2; | 
|---|
| 221 | addValToSets(rlRange->fIncludesSets, rlRange->fNum); | 
|---|
| 222 | } | 
|---|
| 223 | } | 
|---|
| 224 | } | 
|---|
| 225 |  | 
|---|
| 226 | // Move the character category numbers for any dictionary ranges up, so that they | 
|---|
| 227 | // immediately follow the non-dictionary ranges. | 
|---|
| 228 |  | 
|---|
| 229 | fDictCategoriesStart = fGroupCount + 3; | 
|---|
| 230 | for (rlRange = fRangeList; rlRange!=nullptr; rlRange=rlRange->fNext) { | 
|---|
| 231 | if (rlRange->fIncludesDict) { | 
|---|
| 232 | rlRange->fNum += fDictCategoriesStart - 1; | 
|---|
| 233 | if (rlRange->fFirstInGroup) { | 
|---|
| 234 | addValToSets(rlRange->fIncludesSets, rlRange->fNum); | 
|---|
| 235 | } | 
|---|
| 236 | } | 
|---|
| 237 | } | 
|---|
| 238 | fGroupCount += dictGroupCount; | 
|---|
| 239 |  | 
|---|
| 240 |  | 
|---|
| 241 | // Handle input sets that contain the special string {eof}. | 
|---|
| 242 | //   Column 1 of the state table is reserved for EOF on input. | 
|---|
| 243 | //   Column 2 is reserved for before-the-start-input. | 
|---|
| 244 | //            (This column can be optimized away later if there are no rule | 
|---|
| 245 | //             references to {bof}.) | 
|---|
| 246 | //   Add this column value (1 or 2) to the equivalent expression | 
|---|
| 247 | //     subtree for each UnicodeSet that contains the string {eof} | 
|---|
| 248 | //   Because {bof} and {eof} are not characters in the normal sense, | 
|---|
| 249 | //   they don't affect the computation of the ranges or TRIE. | 
|---|
| 250 |  | 
|---|
| 251 | UnicodeString eofString(u "eof"); | 
|---|
| 252 | UnicodeString bofString(u "bof"); | 
|---|
| 253 | for (ni=0; ; ni++) {        // Loop over each of the UnicodeSets encountered in the input rules | 
|---|
| 254 | usetNode = (RBBINode *)this->fRB->fUSetNodes->elementAt(ni); | 
|---|
| 255 | if (usetNode==nullptr) { | 
|---|
| 256 | break; | 
|---|
| 257 | } | 
|---|
| 258 | UnicodeSet      *inputSet = usetNode->fInputSet; | 
|---|
| 259 | if (inputSet->contains(eofString)) { | 
|---|
| 260 | addValToSet(usetNode, 1); | 
|---|
| 261 | } | 
|---|
| 262 | if (inputSet->contains(bofString)) { | 
|---|
| 263 | addValToSet(usetNode, 2); | 
|---|
| 264 | fSawBOF = true; | 
|---|
| 265 | } | 
|---|
| 266 | } | 
|---|
| 267 |  | 
|---|
| 268 |  | 
|---|
| 269 | if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "rgroup")) {printRangeGroups();} | 
|---|
| 270 | if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "esets")) {printSets();} | 
|---|
| 271 | } | 
|---|
| 272 |  | 
|---|
| 273 |  | 
|---|
| 274 | // | 
|---|
| 275 | // Build the Trie table for mapping UChar32 values to the corresponding | 
|---|
| 276 | // range group number. | 
|---|
| 277 | // | 
|---|
| 278 | void RBBISetBuilder::buildTrie() { | 
|---|
| 279 | fMutableTrie = umutablecptrie_open( | 
|---|
| 280 | 0,       //  Initial value for all code points. | 
|---|
| 281 | 0,       //  Error value for out-of-range input. | 
|---|
| 282 | fStatus); | 
|---|
| 283 |  | 
|---|
| 284 | for (RangeDescriptor *range = fRangeList; range!=nullptr && U_SUCCESS(*fStatus); range=range->fNext) { | 
|---|
| 285 | umutablecptrie_setRange(fMutableTrie, | 
|---|
| 286 | range->fStartChar,     // Range start | 
|---|
| 287 | range->fEndChar,       // Range end (inclusive) | 
|---|
| 288 | range->fNum,           // value for range | 
|---|
| 289 | fStatus); | 
|---|
| 290 | } | 
|---|
| 291 | } | 
|---|
| 292 |  | 
|---|
| 293 |  | 
|---|
| 294 | void RBBISetBuilder::mergeCategories(IntPair categories) { | 
|---|
| 295 | U_ASSERT(categories.first >= 1); | 
|---|
| 296 | U_ASSERT(categories.second > categories.first); | 
|---|
| 297 | U_ASSERT((categories.first <  fDictCategoriesStart && categories.second <  fDictCategoriesStart) || | 
|---|
| 298 | (categories.first >= fDictCategoriesStart && categories.second >= fDictCategoriesStart)); | 
|---|
| 299 |  | 
|---|
| 300 | for (RangeDescriptor *rd = fRangeList; rd != nullptr; rd = rd->fNext) { | 
|---|
| 301 | int32_t rangeNum = rd->fNum; | 
|---|
| 302 | if (rangeNum == categories.second) { | 
|---|
| 303 | rd->fNum = categories.first; | 
|---|
| 304 | } else if (rangeNum > categories.second) { | 
|---|
| 305 | rd->fNum--; | 
|---|
| 306 | } | 
|---|
| 307 | } | 
|---|
| 308 | --fGroupCount; | 
|---|
| 309 | if (categories.second <= fDictCategoriesStart) { | 
|---|
| 310 | --fDictCategoriesStart; | 
|---|
| 311 | } | 
|---|
| 312 | } | 
|---|
| 313 |  | 
|---|
| 314 |  | 
|---|
| 315 | //----------------------------------------------------------------------------------- | 
|---|
| 316 | // | 
|---|
| 317 | //  getTrieSize()    Return the size that will be required to serialize the Trie. | 
|---|
| 318 | // | 
|---|
| 319 | //----------------------------------------------------------------------------------- | 
|---|
| 320 | int32_t RBBISetBuilder::getTrieSize()  { | 
|---|
| 321 | if (U_FAILURE(*fStatus)) { | 
|---|
| 322 | return 0; | 
|---|
| 323 | } | 
|---|
| 324 | if (fTrie == nullptr) { | 
|---|
| 325 | bool use8Bits = getNumCharCategories() <= kMaxCharCategoriesFor8BitsTrie; | 
|---|
| 326 | fTrie = umutablecptrie_buildImmutable( | 
|---|
| 327 | fMutableTrie, | 
|---|
| 328 | UCPTRIE_TYPE_FAST, | 
|---|
| 329 | use8Bits ? UCPTRIE_VALUE_BITS_8 : UCPTRIE_VALUE_BITS_16, | 
|---|
| 330 | fStatus); | 
|---|
| 331 | fTrieSize = ucptrie_toBinary(fTrie, nullptr, 0, fStatus); | 
|---|
| 332 | if (*fStatus == U_BUFFER_OVERFLOW_ERROR) { | 
|---|
| 333 | *fStatus = U_ZERO_ERROR; | 
|---|
| 334 | } | 
|---|
| 335 | } | 
|---|
| 336 | return fTrieSize; | 
|---|
| 337 | } | 
|---|
| 338 |  | 
|---|
| 339 |  | 
|---|
| 340 | //----------------------------------------------------------------------------------- | 
|---|
| 341 | // | 
|---|
| 342 | //  serializeTrie()   Put the serialized trie at the specified address. | 
|---|
| 343 | //                    Trust the caller to have given us enough memory. | 
|---|
| 344 | //                    getTrieSize() MUST be called first. | 
|---|
| 345 | // | 
|---|
| 346 | //----------------------------------------------------------------------------------- | 
|---|
| 347 | void RBBISetBuilder::serializeTrie(uint8_t *where) { | 
|---|
| 348 | ucptrie_toBinary(fTrie, | 
|---|
| 349 | where,                // Buffer | 
|---|
| 350 | fTrieSize,            // Capacity | 
|---|
| 351 | fStatus); | 
|---|
| 352 | } | 
|---|
| 353 |  | 
|---|
| 354 | //------------------------------------------------------------------------ | 
|---|
| 355 | // | 
|---|
| 356 | //  addValToSets     Add a runtime-mapped input value to each uset from a | 
|---|
| 357 | //                   list of uset nodes. (val corresponds to a state table column.) | 
|---|
| 358 | //                   For each of the original Unicode sets - which correspond | 
|---|
| 359 | //                   directly to uset nodes - a logically equivalent expression | 
|---|
| 360 | //                   is constructed in terms of the remapped runtime input | 
|---|
| 361 | //                   symbol set.  This function adds one runtime input symbol to | 
|---|
| 362 | //                   a list of sets. | 
|---|
| 363 | // | 
|---|
| 364 | //                   The "logically equivalent expression" is the tree for an | 
|---|
| 365 | //                   or-ing together of all of the symbols that go into the set. | 
|---|
| 366 | // | 
|---|
| 367 | //------------------------------------------------------------------------ | 
|---|
| 368 | void  RBBISetBuilder::addValToSets(UVector *sets, uint32_t val) { | 
|---|
| 369 | int32_t       ix; | 
|---|
| 370 |  | 
|---|
| 371 | for (ix=0; ix<sets->size(); ix++) { | 
|---|
| 372 | RBBINode *usetNode = (RBBINode *)sets->elementAt(ix); | 
|---|
| 373 | addValToSet(usetNode, val); | 
|---|
| 374 | } | 
|---|
| 375 | } | 
|---|
| 376 |  | 
|---|
| 377 | void  RBBISetBuilder::addValToSet(RBBINode *usetNode, uint32_t val) { | 
|---|
| 378 | RBBINode *leafNode = new RBBINode(RBBINode::leafChar); | 
|---|
| 379 | if (leafNode == nullptr) { | 
|---|
| 380 | *fStatus = U_MEMORY_ALLOCATION_ERROR; | 
|---|
| 381 | return; | 
|---|
| 382 | } | 
|---|
| 383 | leafNode->fVal = (unsigned short)val; | 
|---|
| 384 | if (usetNode->fLeftChild == nullptr) { | 
|---|
| 385 | usetNode->fLeftChild = leafNode; | 
|---|
| 386 | leafNode->fParent    = usetNode; | 
|---|
| 387 | } else { | 
|---|
| 388 | // There are already input symbols present for this set. | 
|---|
| 389 | // Set up an OR node, with the previous stuff as the left child | 
|---|
| 390 | //   and the new value as the right child. | 
|---|
| 391 | RBBINode *orNode = new RBBINode(RBBINode::opOr); | 
|---|
| 392 | if (orNode == nullptr) { | 
|---|
| 393 | *fStatus = U_MEMORY_ALLOCATION_ERROR; | 
|---|
| 394 | return; | 
|---|
| 395 | } | 
|---|
| 396 | orNode->fLeftChild  = usetNode->fLeftChild; | 
|---|
| 397 | orNode->fRightChild = leafNode; | 
|---|
| 398 | orNode->fLeftChild->fParent  = orNode; | 
|---|
| 399 | orNode->fRightChild->fParent = orNode; | 
|---|
| 400 | usetNode->fLeftChild = orNode; | 
|---|
| 401 | orNode->fParent = usetNode; | 
|---|
| 402 | } | 
|---|
| 403 | } | 
|---|
| 404 |  | 
|---|
| 405 |  | 
|---|
| 406 | //------------------------------------------------------------------------ | 
|---|
| 407 | // | 
|---|
| 408 | //   getNumCharCategories | 
|---|
| 409 | // | 
|---|
| 410 | //------------------------------------------------------------------------ | 
|---|
| 411 | int32_t  RBBISetBuilder::getNumCharCategories() const { | 
|---|
| 412 | return fGroupCount + 3; | 
|---|
| 413 | } | 
|---|
| 414 |  | 
|---|
| 415 |  | 
|---|
| 416 | //------------------------------------------------------------------------ | 
|---|
| 417 | // | 
|---|
| 418 | //   getDictCategoriesStart | 
|---|
| 419 | // | 
|---|
| 420 | //------------------------------------------------------------------------ | 
|---|
| 421 | int32_t  RBBISetBuilder::getDictCategoriesStart() const { | 
|---|
| 422 | return fDictCategoriesStart; | 
|---|
| 423 | } | 
|---|
| 424 |  | 
|---|
| 425 |  | 
|---|
| 426 | //------------------------------------------------------------------------ | 
|---|
| 427 | // | 
|---|
| 428 | //   sawBOF | 
|---|
| 429 | // | 
|---|
| 430 | //------------------------------------------------------------------------ | 
|---|
| 431 | UBool  RBBISetBuilder::sawBOF() const { | 
|---|
| 432 | return fSawBOF; | 
|---|
| 433 | } | 
|---|
| 434 |  | 
|---|
| 435 |  | 
|---|
| 436 | //------------------------------------------------------------------------ | 
|---|
| 437 | // | 
|---|
| 438 | //   getFirstChar      Given a runtime RBBI character category, find | 
|---|
| 439 | //                     the first UChar32 that is in the set of chars | 
|---|
| 440 | //                     in the category. | 
|---|
| 441 | //------------------------------------------------------------------------ | 
|---|
| 442 | UChar32  RBBISetBuilder::getFirstChar(int32_t category) const { | 
|---|
| 443 | RangeDescriptor   *rlRange; | 
|---|
| 444 | UChar32            retVal = (UChar32)-1; | 
|---|
| 445 | for (rlRange = fRangeList; rlRange!=nullptr; rlRange=rlRange->fNext) { | 
|---|
| 446 | if (rlRange->fNum == category) { | 
|---|
| 447 | retVal = rlRange->fStartChar; | 
|---|
| 448 | break; | 
|---|
| 449 | } | 
|---|
| 450 | } | 
|---|
| 451 | return retVal; | 
|---|
| 452 | } | 
|---|
| 453 |  | 
|---|
| 454 |  | 
|---|
| 455 | //------------------------------------------------------------------------ | 
|---|
| 456 | // | 
|---|
| 457 | //   printRanges        A debugging function. | 
|---|
| 458 | //                      dump out all of the range definitions. | 
|---|
| 459 | // | 
|---|
| 460 | //------------------------------------------------------------------------ | 
|---|
| 461 | #ifdef RBBI_DEBUG | 
|---|
| 462 | void RBBISetBuilder::printRanges() { | 
|---|
| 463 | RangeDescriptor       *rlRange; | 
|---|
| 464 | int                    i; | 
|---|
| 465 |  | 
|---|
| 466 | RBBIDebugPrintf( "\n\n Nonoverlapping Ranges ...\n"); | 
|---|
| 467 | for (rlRange = fRangeList; rlRange!=nullptr; rlRange=rlRange->fNext) { | 
|---|
| 468 | RBBIDebugPrintf( "%4x-%4x  ", rlRange->fStartChar, rlRange->fEndChar); | 
|---|
| 469 |  | 
|---|
| 470 | for (i=0; i<rlRange->fIncludesSets->size(); i++) { | 
|---|
| 471 | RBBINode       *usetNode    = (RBBINode *)rlRange->fIncludesSets->elementAt(i); | 
|---|
| 472 | UnicodeString   setName {u "anon"}; | 
|---|
| 473 | RBBINode       *setRef = usetNode->fParent; | 
|---|
| 474 | if (setRef != nullptr) { | 
|---|
| 475 | RBBINode *varRef = setRef->fParent; | 
|---|
| 476 | if (varRef != nullptr  &&  varRef->fType == RBBINode::varRef) { | 
|---|
| 477 | setName = varRef->fText; | 
|---|
| 478 | } | 
|---|
| 479 | } | 
|---|
| 480 | RBBI_DEBUG_printUnicodeString(setName); RBBIDebugPrintf( "  "); | 
|---|
| 481 | } | 
|---|
| 482 | RBBIDebugPrintf( "\n"); | 
|---|
| 483 | } | 
|---|
| 484 | } | 
|---|
| 485 | #endif | 
|---|
| 486 |  | 
|---|
| 487 |  | 
|---|
| 488 | //------------------------------------------------------------------------ | 
|---|
| 489 | // | 
|---|
| 490 | //   printRangeGroups     A debugging function. | 
|---|
| 491 | //                        dump out all of the range groups. | 
|---|
| 492 | // | 
|---|
| 493 | //------------------------------------------------------------------------ | 
|---|
| 494 | #ifdef RBBI_DEBUG | 
|---|
| 495 | void RBBISetBuilder::printRangeGroups() { | 
|---|
| 496 | int                    i; | 
|---|
| 497 |  | 
|---|
| 498 | RBBIDebugPrintf( "\nRanges grouped by Unicode Set Membership...\n"); | 
|---|
| 499 | for (RangeDescriptor *rlRange = fRangeList; rlRange!=nullptr; rlRange=rlRange->fNext) { | 
|---|
| 500 | if (rlRange->fFirstInGroup) { | 
|---|
| 501 | int groupNum = rlRange->fNum; | 
|---|
| 502 | RBBIDebugPrintf( "%2i  ", groupNum); | 
|---|
| 503 |  | 
|---|
| 504 | if (groupNum >= fDictCategoriesStart) { RBBIDebugPrintf( " <DICT> ");} | 
|---|
| 505 |  | 
|---|
| 506 | for (i=0; i<rlRange->fIncludesSets->size(); i++) { | 
|---|
| 507 | RBBINode       *usetNode    = (RBBINode *)rlRange->fIncludesSets->elementAt(i); | 
|---|
| 508 | UnicodeString   setName = UNICODE_STRING( "anon", 4); | 
|---|
| 509 | RBBINode       *setRef = usetNode->fParent; | 
|---|
| 510 | if (setRef != nullptr) { | 
|---|
| 511 | RBBINode *varRef = setRef->fParent; | 
|---|
| 512 | if (varRef != nullptr  &&  varRef->fType == RBBINode::varRef) { | 
|---|
| 513 | setName = varRef->fText; | 
|---|
| 514 | } | 
|---|
| 515 | } | 
|---|
| 516 | RBBI_DEBUG_printUnicodeString(setName); RBBIDebugPrintf( " "); | 
|---|
| 517 | } | 
|---|
| 518 |  | 
|---|
| 519 | i = 0; | 
|---|
| 520 | for (RangeDescriptor *tRange = rlRange; tRange != nullptr; tRange = tRange->fNext) { | 
|---|
| 521 | if (tRange->fNum == rlRange->fNum) { | 
|---|
| 522 | if (i++ % 5 == 0) { | 
|---|
| 523 | RBBIDebugPrintf( "\n    "); | 
|---|
| 524 | } | 
|---|
| 525 | RBBIDebugPrintf( "  %05x-%05x", tRange->fStartChar, tRange->fEndChar); | 
|---|
| 526 | } | 
|---|
| 527 | } | 
|---|
| 528 | RBBIDebugPrintf( "\n"); | 
|---|
| 529 | } | 
|---|
| 530 | } | 
|---|
| 531 | RBBIDebugPrintf( "\n"); | 
|---|
| 532 | } | 
|---|
| 533 | #endif | 
|---|
| 534 |  | 
|---|
| 535 |  | 
|---|
| 536 | //------------------------------------------------------------------------ | 
|---|
| 537 | // | 
|---|
| 538 | //   printSets          A debugging function. | 
|---|
| 539 | //                      dump out all of the set definitions. | 
|---|
| 540 | // | 
|---|
| 541 | //------------------------------------------------------------------------ | 
|---|
| 542 | #ifdef RBBI_DEBUG | 
|---|
| 543 | void RBBISetBuilder::printSets() { | 
|---|
| 544 | int                   i; | 
|---|
| 545 |  | 
|---|
| 546 | RBBIDebugPrintf( "\n\nUnicode Sets List\n------------------\n"); | 
|---|
| 547 | for (i=0; ; i++) { | 
|---|
| 548 | RBBINode        *usetNode; | 
|---|
| 549 | RBBINode        *setRef; | 
|---|
| 550 | RBBINode        *varRef; | 
|---|
| 551 | UnicodeString    setName; | 
|---|
| 552 |  | 
|---|
| 553 | usetNode = (RBBINode *)fRB->fUSetNodes->elementAt(i); | 
|---|
| 554 | if (usetNode == nullptr) { | 
|---|
| 555 | break; | 
|---|
| 556 | } | 
|---|
| 557 |  | 
|---|
| 558 | RBBIDebugPrintf( "%3d    ", i); | 
|---|
| 559 | setName = UNICODE_STRING( "anonymous", 9); | 
|---|
| 560 | setRef = usetNode->fParent; | 
|---|
| 561 | if (setRef != nullptr) { | 
|---|
| 562 | varRef = setRef->fParent; | 
|---|
| 563 | if (varRef != nullptr  &&  varRef->fType == RBBINode::varRef) { | 
|---|
| 564 | setName = varRef->fText; | 
|---|
| 565 | } | 
|---|
| 566 | } | 
|---|
| 567 | RBBI_DEBUG_printUnicodeString(setName); | 
|---|
| 568 | RBBIDebugPrintf( "   "); | 
|---|
| 569 | RBBI_DEBUG_printUnicodeString(usetNode->fText); | 
|---|
| 570 | RBBIDebugPrintf( "\n"); | 
|---|
| 571 | if (usetNode->fLeftChild != nullptr) { | 
|---|
| 572 | RBBINode::printTree(usetNode->fLeftChild, true); | 
|---|
| 573 | } | 
|---|
| 574 | } | 
|---|
| 575 | RBBIDebugPrintf( "\n"); | 
|---|
| 576 | } | 
|---|
| 577 | #endif | 
|---|
| 578 |  | 
|---|
| 579 |  | 
|---|
| 580 |  | 
|---|
| 581 | //------------------------------------------------------------------------------------- | 
|---|
| 582 | // | 
|---|
| 583 | //  RangeDescriptor copy constructor | 
|---|
| 584 | // | 
|---|
| 585 | //------------------------------------------------------------------------------------- | 
|---|
| 586 |  | 
|---|
| 587 | RangeDescriptor::RangeDescriptor(const RangeDescriptor &other, UErrorCode &status) : | 
|---|
| 588 | fStartChar(other.fStartChar), fEndChar {other.fEndChar}, fNum {other.fNum}, | 
|---|
| 589 | fIncludesDict{other.fIncludesDict}, fFirstInGroup{other.fFirstInGroup} { | 
|---|
| 590 |  | 
|---|
| 591 | if (U_FAILURE(status)) { | 
|---|
| 592 | return; | 
|---|
| 593 | } | 
|---|
| 594 | fIncludesSets = new UVector(status); | 
|---|
| 595 | if (this->fIncludesSets == nullptr) { | 
|---|
| 596 | status = U_MEMORY_ALLOCATION_ERROR; | 
|---|
| 597 | } | 
|---|
| 598 | if (U_FAILURE(status)) { | 
|---|
| 599 | return; | 
|---|
| 600 | } | 
|---|
| 601 |  | 
|---|
| 602 | for (int32_t i=0; i<other.fIncludesSets->size(); i++) { | 
|---|
| 603 | this->fIncludesSets->addElement(other.fIncludesSets->elementAt(i), status); | 
|---|
| 604 | } | 
|---|
| 605 | } | 
|---|
| 606 |  | 
|---|
| 607 |  | 
|---|
| 608 | //------------------------------------------------------------------------------------- | 
|---|
| 609 | // | 
|---|
| 610 | //  RangeDesriptor default constructor | 
|---|
| 611 | // | 
|---|
| 612 | //------------------------------------------------------------------------------------- | 
|---|
| 613 | RangeDescriptor::RangeDescriptor(UErrorCode &status) { | 
|---|
| 614 | if (U_FAILURE(status)) { | 
|---|
| 615 | return; | 
|---|
| 616 | } | 
|---|
| 617 | fIncludesSets = new UVector(status); | 
|---|
| 618 | if (fIncludesSets == nullptr) { | 
|---|
| 619 | status = U_MEMORY_ALLOCATION_ERROR; | 
|---|
| 620 | } | 
|---|
| 621 | } | 
|---|
| 622 |  | 
|---|
| 623 |  | 
|---|
| 624 | //------------------------------------------------------------------------------------- | 
|---|
| 625 | // | 
|---|
| 626 | //  RangeDesriptor Destructor | 
|---|
| 627 | // | 
|---|
| 628 | //------------------------------------------------------------------------------------- | 
|---|
| 629 | RangeDescriptor::~RangeDescriptor() { | 
|---|
| 630 | delete  fIncludesSets; | 
|---|
| 631 | fIncludesSets = nullptr; | 
|---|
| 632 | } | 
|---|
| 633 |  | 
|---|
| 634 | //------------------------------------------------------------------------------------- | 
|---|
| 635 | // | 
|---|
| 636 | //  RangeDesriptor::split() | 
|---|
| 637 | // | 
|---|
| 638 | //------------------------------------------------------------------------------------- | 
|---|
| 639 | void RangeDescriptor::split(UChar32 where, UErrorCode &status) { | 
|---|
| 640 | U_ASSERT(where>fStartChar && where<=fEndChar); | 
|---|
| 641 | RangeDescriptor *nr = new RangeDescriptor(*this, status); | 
|---|
| 642 | if(nr == nullptr) { | 
|---|
| 643 | status = U_MEMORY_ALLOCATION_ERROR; | 
|---|
| 644 | return; | 
|---|
| 645 | } | 
|---|
| 646 | if (U_FAILURE(status)) { | 
|---|
| 647 | delete nr; | 
|---|
| 648 | return; | 
|---|
| 649 | } | 
|---|
| 650 | //  RangeDescriptor copy constructor copies all fields. | 
|---|
| 651 | //  Only need to update those that are different after the split. | 
|---|
| 652 | nr->fStartChar = where; | 
|---|
| 653 | this->fEndChar = where-1; | 
|---|
| 654 | nr->fNext      = this->fNext; | 
|---|
| 655 | this->fNext    = nr; | 
|---|
| 656 | } | 
|---|
| 657 |  | 
|---|
| 658 |  | 
|---|
| 659 | //------------------------------------------------------------------------------------- | 
|---|
| 660 | // | 
|---|
| 661 | //   RangeDescriptor::isDictionaryRange | 
|---|
| 662 | // | 
|---|
| 663 | //            Test whether this range includes characters from | 
|---|
| 664 | //            the original Unicode Set named "dictionary". | 
|---|
| 665 | // | 
|---|
| 666 | //            This function looks through the Unicode Sets that | 
|---|
| 667 | //            the range includes, checking for one named "dictionary" | 
|---|
| 668 | // | 
|---|
| 669 | //            TODO:  a faster way would be to find the set node for | 
|---|
| 670 | //                   "dictionary" just once, rather than looking it | 
|---|
| 671 | //                   up by name every time. | 
|---|
| 672 | // | 
|---|
| 673 | //------------------------------------------------------------------------------------- | 
|---|
| 674 | bool RangeDescriptor::isDictionaryRange() { | 
|---|
| 675 | static const char16_t *dictionary = u "dictionary"; | 
|---|
| 676 | for (int32_t i=0; i<fIncludesSets->size(); i++) { | 
|---|
| 677 | RBBINode *usetNode  = (RBBINode *)fIncludesSets->elementAt(i); | 
|---|
| 678 | RBBINode *setRef = usetNode->fParent; | 
|---|
| 679 | if (setRef != nullptr) { | 
|---|
| 680 | RBBINode *varRef = setRef->fParent; | 
|---|
| 681 | if (varRef && varRef->fType == RBBINode::varRef) { | 
|---|
| 682 | const UnicodeString *setName = &varRef->fText; | 
|---|
| 683 | if (setName->compare(dictionary, -1) == 0) { | 
|---|
| 684 | return true; | 
|---|
| 685 | } | 
|---|
| 686 | } | 
|---|
| 687 | } | 
|---|
| 688 | } | 
|---|
| 689 | return false; | 
|---|
| 690 | } | 
|---|
| 691 |  | 
|---|
| 692 | U_NAMESPACE_END | 
|---|
| 693 |  | 
|---|
| 694 | #endif /* #if !UCONFIG_NO_BREAK_ITERATION */ | 
|---|
| 695 |  | 
|---|