| File: | root/firefox-clang/obj-x86_64-pc-linux-gnu/extensions/spellcheck/hunspell/src/./../../../../../extensions/spellcheck/hunspell/src/affixmgr.cxx |
| Warning: | line 4475, column 13 Value stored to 'nummap' is never read |
Press '?' to see keyboard shortcuts
Keyboard shortcuts:
| 1 | /* ***** BEGIN LICENSE BLOCK ***** |
| 2 | * Version: MPL 1.1/GPL 2.0/LGPL 2.1 |
| 3 | * |
| 4 | * Copyright (C) 2002-2022 Németh László |
| 5 | * |
| 6 | * The contents of this file are subject to the Mozilla Public License Version |
| 7 | * 1.1 (the "License"); you may not use this file except in compliance with |
| 8 | * the License. You may obtain a copy of the License at |
| 9 | * http://www.mozilla.org/MPL/ |
| 10 | * |
| 11 | * Software distributed under the License is distributed on an "AS IS" basis, |
| 12 | * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License |
| 13 | * for the specific language governing rights and limitations under the |
| 14 | * License. |
| 15 | * |
| 16 | * Hunspell is based on MySpell which is Copyright (C) 2002 Kevin Hendricks. |
| 17 | * |
| 18 | * Contributor(s): David Einstein, Davide Prina, Giuseppe Modugno, |
| 19 | * Gianluca Turconi, Simon Brouwer, Noll János, Bíró Árpád, |
| 20 | * Goldman Eleonóra, Sarlós Tamás, Bencsáth Boldizsár, Halácsy Péter, |
| 21 | * Dvornik László, Gefferth András, Nagy Viktor, Varga Dániel, Chris Halls, |
| 22 | * Rene Engelhard, Bram Moolenaar, Dafydd Jones, Harri Pitkänen |
| 23 | * |
| 24 | * Alternatively, the contents of this file may be used under the terms of |
| 25 | * either the GNU General Public License Version 2 or later (the "GPL"), or |
| 26 | * the GNU Lesser General Public License Version 2.1 or later (the "LGPL"), |
| 27 | * in which case the provisions of the GPL or the LGPL are applicable instead |
| 28 | * of those above. If you wish to allow use of your version of this file only |
| 29 | * under the terms of either the GPL or the LGPL, and not to allow others to |
| 30 | * use your version of this file under the terms of the MPL, indicate your |
| 31 | * decision by deleting the provisions above and replace them with the notice |
| 32 | * and other provisions required by the GPL or the LGPL. If you do not delete |
| 33 | * the provisions above, a recipient may use your version of this file under |
| 34 | * the terms of any one of the MPL, the GPL or the LGPL. |
| 35 | * |
| 36 | * ***** END LICENSE BLOCK ***** */ |
| 37 | /* |
| 38 | * Copyright 2002 Kevin B. Hendricks, Stratford, Ontario, Canada |
| 39 | * And Contributors. All rights reserved. |
| 40 | * |
| 41 | * Redistribution and use in source and binary forms, with or without |
| 42 | * modification, are permitted provided that the following conditions |
| 43 | * are met: |
| 44 | * |
| 45 | * 1. Redistributions of source code must retain the above copyright |
| 46 | * notice, this list of conditions and the following disclaimer. |
| 47 | * |
| 48 | * 2. Redistributions in binary form must reproduce the above copyright |
| 49 | * notice, this list of conditions and the following disclaimer in the |
| 50 | * documentation and/or other materials provided with the distribution. |
| 51 | * |
| 52 | * 3. All modifications to the source code must be clearly marked as |
| 53 | * such. Binary redistributions based on modified source code |
| 54 | * must be clearly marked as modified versions in the documentation |
| 55 | * and/or other materials provided with the distribution. |
| 56 | * |
| 57 | * THIS SOFTWARE IS PROVIDED BY KEVIN B. HENDRICKS AND CONTRIBUTORS |
| 58 | * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
| 59 | * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS |
| 60 | * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL |
| 61 | * KEVIN B. HENDRICKS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, |
| 62 | * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, |
| 63 | * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; |
| 64 | * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
| 65 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
| 66 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
| 67 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
| 68 | * SUCH DAMAGE. |
| 69 | */ |
| 70 | |
| 71 | #include <cstdlib> |
| 72 | #include <cstring> |
| 73 | #include <cstdio> |
| 74 | #include <cctype> |
| 75 | #include <ctime> |
| 76 | |
| 77 | #include <algorithm> |
| 78 | #include <chrono> |
| 79 | #include <memory> |
| 80 | #include <limits> |
| 81 | #include <string> |
| 82 | #include <unordered_set> |
| 83 | #include <vector> |
| 84 | |
| 85 | #include "affixmgr.hxx" |
| 86 | #include "affentry.hxx" |
| 87 | #include "hunspelltrace.hxx" |
| 88 | #include "langnum.hxx" |
| 89 | |
| 90 | #include "csutil.hxx" |
| 91 | |
| 92 | namespace { |
| 93 | // the distinct records of one analysis, in the order they were first added. a record is a run of |
| 94 | // text between two breakchar separators |
| 95 | class DistinctRecords { |
| 96 | public: |
| 97 | explicit DistinctRecords(char breakchar) : sep(breakchar) {} |
| 98 | |
| 99 | // add each record of add that is not held already |
| 100 | void append(const std::string& add) { |
| 101 | size_t pos = 0; |
| 102 | while (pos < add.size()) { |
| 103 | size_t end = add.find(sep, pos); |
| 104 | size_t len = (end == std::string::npos ? add.size() : end) - pos; |
| 105 | if (len) { |
| 106 | std::string record(add, pos, len); |
| 107 | if (seen.insert(record).second) |
| 108 | records.push_back(std::move(record)); |
| 109 | } |
| 110 | if (end == std::string::npos) |
| 111 | break; |
| 112 | pos = end + 1; |
| 113 | } |
| 114 | } |
| 115 | |
| 116 | // the records joined back into one string, each one terminated by breakchar |
| 117 | std::string join() const { |
| 118 | std::string text; |
| 119 | for (const std::string& record : records) { |
| 120 | text.append(record); |
| 121 | text.push_back(sep); |
| 122 | } |
| 123 | return text; |
| 124 | } |
| 125 | |
| 126 | private: |
| 127 | std::vector<std::string> records; |
| 128 | // the same records again, to answer "have I seen this one" without a walk over records |
| 129 | std::unordered_set<std::string> seen; |
| 130 | char sep; |
| 131 | }; |
| 132 | } |
| 133 | |
| 134 | AffixMgr::AffixMgr(const char* affpath, |
| 135 | const std::vector<std::unique_ptr<HashMgr>>& ptr, |
| 136 | const char* key) |
| 137 | : alldic(ptr) |
| 138 | , pHMgr(ptr[0].get()) { |
| 139 | |
| 140 | // register hash manager and load affix data from aff file |
| 141 | csconv = nullptr; |
| 142 | utf8 = 0; |
| 143 | complexprefixes = 0; |
| 144 | parsedmaptable = false; |
| 145 | parsedbreaktable = false; |
| 146 | iconvtable = nullptr; |
| 147 | oconvtable = nullptr; |
| 148 | // allow simplified compound forms (see 3rd field of CHECKCOMPOUNDPATTERN) |
| 149 | simplifiedcpd = 0; |
| 150 | parsedcheckcpd = false; |
| 151 | parseddefcpd = false; |
| 152 | phone = nullptr; |
| 153 | compoundflag = FLAG_NULL0x00; // permits word in compound forms |
| 154 | compoundbegin = FLAG_NULL0x00; // may be first word in compound forms |
| 155 | compoundmiddle = FLAG_NULL0x00; // may be middle word in compound forms |
| 156 | compoundend = FLAG_NULL0x00; // may be last word in compound forms |
| 157 | compoundroot = FLAG_NULL0x00; // compound word signing flag |
| 158 | compoundpermitflag = FLAG_NULL0x00; // compound permitting flag for suffixed word |
| 159 | compoundforbidflag = FLAG_NULL0x00; // compound fordidden flag for suffixed word |
| 160 | compoundmoresuffixes = 0; // allow more suffixes within compound words |
| 161 | checkcompounddup = 0; // forbid double words in compounds |
| 162 | checkcompoundrep = 0; // forbid bad compounds (may be non-compound word with |
| 163 | // a REP substitution) |
| 164 | checkcompoundcase = |
| 165 | 0; // forbid upper and lowercase combinations at word bounds |
| 166 | checkcompoundtriple = 0; // forbid compounds with triple letters |
| 167 | simplifiedtriple = 0; // allow simplified triple letters in compounds |
| 168 | // (Schiff+fahrt -> Schiffahrt) |
| 169 | forbiddenword = FORBIDDENWORD65510; // forbidden word signing flag |
| 170 | nosuggest = FLAG_NULL0x00; // don't suggest words signed with NOSUGGEST flag |
| 171 | nongramsuggest = FLAG_NULL0x00; |
| 172 | langnum = 0; // language code (see http://l10n.openoffice.org/languages.html) |
| 173 | needaffix = FLAG_NULL0x00; // forbidden root, allowed only with suffixes |
| 174 | cpdwordmax = -1; // default: unlimited wordcount in compound words |
| 175 | cpdmin = -1; // undefined |
| 176 | cpdmaxsyllable = 0; // default: unlimited syllablecount in compound words |
| 177 | pfxappnd = nullptr; // previous prefix for counting syllables of the prefix BUG |
| 178 | sfxappnd = nullptr; // previous suffix for counting syllables of the suffix BUG |
| 179 | sfxextra = 0; // modifier for syllable count of sfxappnd BUG |
| 180 | checknum = 0; // checking numbers, and word with numbers |
| 181 | havecontclass = 0; // flags of possible continuing classes (double affix) |
| 182 | // LEMMA_PRESENT: not put root into the morphological output. Lemma presents |
| 183 | // in morhological description in dictionary file. It's often combined with |
| 184 | // PSEUDOROOT. |
| 185 | lemma_present = FLAG_NULL0x00; |
| 186 | circumfix = FLAG_NULL0x00; |
| 187 | onlyincompound = FLAG_NULL0x00; |
| 188 | maxngramsugs = -1; // undefined |
| 189 | maxdiff = -1; // undefined |
| 190 | onlymaxdiff = 0; |
| 191 | maxcpdsugs = -1; // undefined |
| 192 | nosplitsugs = 0; |
| 193 | sugswithdots = 0; |
| 194 | keepcase = 0; |
| 195 | forceucase = 0; |
| 196 | warn = 0; |
| 197 | forbidwarn = 0; |
| 198 | checksharps = 0; |
| 199 | substandard = FLAG_NULL0x00; |
| 200 | fullstrip = 0; |
| 201 | |
| 202 | sfx = nullptr; |
| 203 | pfx = nullptr; |
| 204 | |
| 205 | for (int i = 0; i < SETSIZE256; i++) { |
| 206 | pStart[i] = nullptr; |
| 207 | sStart[i] = nullptr; |
| 208 | pFlag[i] = nullptr; |
| 209 | sFlag[i] = nullptr; |
| 210 | } |
| 211 | |
| 212 | memset(contclasses, 0, CONTSIZE65536 * sizeof(char)); |
| 213 | |
| 214 | if (parse_file(affpath, key)) { |
| 215 | fprintf(stderrstderr, "Failure loading aff file %s\n", affpath); |
| 216 | } |
| 217 | |
| 218 | /* get encoding for CHECKCOMPOUNDCASE */ |
| 219 | if (!utf8) { |
| 220 | csconv = get_current_cs(get_encoding()); |
| 221 | for (int i = 0; i <= 255; i++) { |
| 222 | if ((csconv[i].cupper != csconv[i].clower) && |
| 223 | (wordchars.find((char)i) == std::string::npos)) { |
| 224 | wordchars.push_back((char)i); |
| 225 | } |
| 226 | } |
| 227 | } |
| 228 | |
| 229 | // default BREAK definition |
| 230 | if (!parsedbreaktable) { |
| 231 | breaktable.emplace_back("-"); |
| 232 | breaktable.emplace_back("^-"); |
| 233 | breaktable.emplace_back("-$"); |
| 234 | parsedbreaktable = true; |
| 235 | } |
| 236 | |
| 237 | #if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) |
| 238 | // not entirely sure this is invalid, so only for fuzzing for now |
| 239 | if (iconvtable && !iconvtable->check_against_breaktable(breaktable)) { |
| 240 | delete iconvtable; |
| 241 | iconvtable = nullptr; |
| 242 | } |
| 243 | #endif |
| 244 | |
| 245 | if (cpdmin == -1) |
| 246 | cpdmin = MINCPDLEN3; |
| 247 | } |
| 248 | |
| 249 | AffixMgr::~AffixMgr() { |
| 250 | // pass through linked prefix entries and clean up |
| 251 | for (int i = 0; i < SETSIZE256; i++) { |
| 252 | pFlag[i] = nullptr; |
| 253 | PfxEntry* ptr = pStart[i]; |
| 254 | PfxEntry* nptr = nullptr; |
| 255 | while (ptr) { |
| 256 | nptr = ptr->getNext(); |
| 257 | delete (ptr); |
| 258 | ptr = nptr; |
| 259 | } |
| 260 | } |
| 261 | |
| 262 | // pass through linked suffix entries and clean up |
| 263 | for (int j = 0; j < SETSIZE256; j++) { |
| 264 | sFlag[j] = nullptr; |
| 265 | SfxEntry* ptr = sStart[j]; |
| 266 | SfxEntry* nptr = nullptr; |
| 267 | while (ptr) { |
| 268 | nptr = ptr->getNext(); |
| 269 | delete (ptr); |
| 270 | ptr = nptr; |
| 271 | } |
| 272 | sStart[j] = nullptr; |
| 273 | } |
| 274 | |
| 275 | delete iconvtable; |
| 276 | delete oconvtable; |
| 277 | delete phone; |
| 278 | |
| 279 | FREE_FLAG(compoundflag)compoundflag = 0; |
| 280 | FREE_FLAG(compoundbegin)compoundbegin = 0; |
| 281 | FREE_FLAG(compoundmiddle)compoundmiddle = 0; |
| 282 | FREE_FLAG(compoundend)compoundend = 0; |
| 283 | FREE_FLAG(compoundpermitflag)compoundpermitflag = 0; |
| 284 | FREE_FLAG(compoundforbidflag)compoundforbidflag = 0; |
| 285 | FREE_FLAG(compoundroot)compoundroot = 0; |
| 286 | FREE_FLAG(forbiddenword)forbiddenword = 0; |
| 287 | FREE_FLAG(nosuggest)nosuggest = 0; |
| 288 | FREE_FLAG(nongramsuggest)nongramsuggest = 0; |
| 289 | FREE_FLAG(needaffix)needaffix = 0; |
| 290 | FREE_FLAG(lemma_present)lemma_present = 0; |
| 291 | FREE_FLAG(circumfix)circumfix = 0; |
| 292 | FREE_FLAG(onlyincompound)onlyincompound = 0; |
| 293 | |
| 294 | cpdwordmax = 0; |
| 295 | pHMgr = nullptr; |
| 296 | cpdmin = 0; |
| 297 | cpdmaxsyllable = 0; |
| 298 | checknum = 0; |
| 299 | #ifdef MOZILLA_CLIENT1 |
| 300 | delete[] csconv; |
| 301 | #endif |
| 302 | } |
| 303 | |
| 304 | void AffixMgr::finishFileMgr(FileMgr* afflst) { |
| 305 | delete afflst; |
| 306 | |
| 307 | // convert affix trees to sorted list |
| 308 | process_pfx_tree_to_list(); |
| 309 | process_sfx_tree_to_list(); |
| 310 | } |
| 311 | |
| 312 | // read in aff file and build up prefix and suffix entry objects |
| 313 | int AffixMgr::parse_file(const char* affpath, const char* key) { |
| 314 | |
| 315 | // checking flag duplication |
| 316 | char dupflags[CONTSIZE65536]; |
| 317 | char dupflags_ini = 1; |
| 318 | |
| 319 | // first line indicator for removing byte order mark |
| 320 | int firstline = 1; |
| 321 | |
| 322 | // open the affix file |
| 323 | FileMgr* afflst = new FileMgr(affpath, key); |
| 324 | |
| 325 | // step one is to parse the affix file building up the internal |
| 326 | // affix data structures |
| 327 | |
| 328 | // read in each line ignoring any that do not |
| 329 | // start with a known line type indicator |
| 330 | std::string line; |
| 331 | while (afflst->getline(line)) { |
| 332 | mychomp(line); |
| 333 | |
| 334 | /* remove byte order mark */ |
| 335 | if (firstline) { |
| 336 | firstline = 0; |
| 337 | // Affix file begins with byte order mark: possible incompatibility with |
| 338 | // old Hunspell versions |
| 339 | if (line.compare(0, 3, "\xEF\xBB\xBF", 3) == 0) { |
| 340 | line.erase(0, 3); |
| 341 | } |
| 342 | } |
| 343 | |
| 344 | /* parse in the keyboard string */ |
| 345 | if (line.compare(0, 3, "KEY", 3) == 0) { |
| 346 | if (!parse_string(line, keystring, afflst->getlinenum())) { |
| 347 | finishFileMgr(afflst); |
| 348 | return 1; |
| 349 | } |
| 350 | } |
| 351 | |
| 352 | /* parse in the try string */ |
| 353 | if (line.compare(0, 3, "TRY", 3) == 0) { |
| 354 | if (!parse_string(line, trystring, afflst->getlinenum())) { |
| 355 | finishFileMgr(afflst); |
| 356 | return 1; |
| 357 | } |
| 358 | } |
| 359 | |
| 360 | /* parse in the name of the character set used by the .dict and .aff */ |
| 361 | if (line.compare(0, 3, "SET", 3) == 0) { |
| 362 | if (!parse_string(line, encoding, afflst->getlinenum())) { |
| 363 | finishFileMgr(afflst); |
| 364 | return 1; |
| 365 | } |
| 366 | if (encoding == "UTF-8") { |
| 367 | utf8 = 1; |
| 368 | } |
| 369 | } |
| 370 | |
| 371 | /* parse COMPLEXPREFIXES for agglutinative languages with right-to-left |
| 372 | * writing system */ |
| 373 | if (line.compare(0, 15, "COMPLEXPREFIXES", 15) == 0) |
| 374 | complexprefixes = 1; |
| 375 | |
| 376 | /* parse in the flag used by the controlled compound words */ |
| 377 | if (line.compare(0, 12, "COMPOUNDFLAG", 12) == 0) { |
| 378 | if (!parse_flag(line, &compoundflag, afflst)) { |
| 379 | finishFileMgr(afflst); |
| 380 | return 1; |
| 381 | } |
| 382 | } |
| 383 | |
| 384 | /* parse in the flag used by compound words */ |
| 385 | if (line.compare(0, 13, "COMPOUNDBEGIN", 13) == 0) { |
| 386 | if (complexprefixes) { |
| 387 | if (!parse_flag(line, &compoundend, afflst)) { |
| 388 | finishFileMgr(afflst); |
| 389 | return 1; |
| 390 | } |
| 391 | } else { |
| 392 | if (!parse_flag(line, &compoundbegin, afflst)) { |
| 393 | finishFileMgr(afflst); |
| 394 | return 1; |
| 395 | } |
| 396 | } |
| 397 | } |
| 398 | |
| 399 | /* parse in the flag used by compound words */ |
| 400 | if (line.compare(0, 14, "COMPOUNDMIDDLE", 14) == 0) { |
| 401 | if (!parse_flag(line, &compoundmiddle, afflst)) { |
| 402 | finishFileMgr(afflst); |
| 403 | return 1; |
| 404 | } |
| 405 | } |
| 406 | |
| 407 | /* parse in the flag used by compound words */ |
| 408 | if (line.compare(0, 11, "COMPOUNDEND", 11) == 0) { |
| 409 | if (complexprefixes) { |
| 410 | if (!parse_flag(line, &compoundbegin, afflst)) { |
| 411 | finishFileMgr(afflst); |
| 412 | return 1; |
| 413 | } |
| 414 | } else { |
| 415 | if (!parse_flag(line, &compoundend, afflst)) { |
| 416 | finishFileMgr(afflst); |
| 417 | return 1; |
| 418 | } |
| 419 | } |
| 420 | } |
| 421 | |
| 422 | /* parse in the data used by compound_check() method */ |
| 423 | if (line.compare(0, 15, "COMPOUNDWORDMAX", 15) == 0) { |
| 424 | if (!parse_num(line, &cpdwordmax, afflst)) { |
| 425 | finishFileMgr(afflst); |
| 426 | return 1; |
| 427 | } |
| 428 | } |
| 429 | |
| 430 | /* parse in the flag sign compounds in dictionary */ |
| 431 | if (line.compare(0, 12, "COMPOUNDROOT", 12) == 0) { |
| 432 | if (!parse_flag(line, &compoundroot, afflst)) { |
| 433 | finishFileMgr(afflst); |
| 434 | return 1; |
| 435 | } |
| 436 | } |
| 437 | |
| 438 | /* parse in the flag used by compound_check() method */ |
| 439 | if (line.compare(0, 18, "COMPOUNDPERMITFLAG", 18) == 0) { |
| 440 | if (!parse_flag(line, &compoundpermitflag, afflst)) { |
| 441 | finishFileMgr(afflst); |
| 442 | return 1; |
| 443 | } |
| 444 | } |
| 445 | |
| 446 | /* parse in the flag used by compound_check() method */ |
| 447 | if (line.compare(0, 18, "COMPOUNDFORBIDFLAG", 18) == 0) { |
| 448 | if (!parse_flag(line, &compoundforbidflag, afflst)) { |
| 449 | finishFileMgr(afflst); |
| 450 | return 1; |
| 451 | } |
| 452 | } |
| 453 | |
| 454 | if (line.compare(0, 20, "COMPOUNDMORESUFFIXES", 20) == 0) { |
| 455 | compoundmoresuffixes = 1; |
| 456 | } |
| 457 | |
| 458 | if (line.compare(0, 16, "CHECKCOMPOUNDDUP", 16) == 0) { |
| 459 | checkcompounddup = 1; |
| 460 | } |
| 461 | |
| 462 | if (line.compare(0, 16, "CHECKCOMPOUNDREP", 16) == 0) { |
| 463 | checkcompoundrep = 1; |
| 464 | } |
| 465 | |
| 466 | if (line.compare(0, 19, "CHECKCOMPOUNDTRIPLE", 19) == 0) { |
| 467 | checkcompoundtriple = 1; |
| 468 | } |
| 469 | |
| 470 | if (line.compare(0, 16, "SIMPLIFIEDTRIPLE", 16) == 0) { |
| 471 | simplifiedtriple = 1; |
| 472 | } |
| 473 | |
| 474 | if (line.compare(0, 17, "CHECKCOMPOUNDCASE", 17) == 0) { |
| 475 | checkcompoundcase = 1; |
| 476 | } |
| 477 | |
| 478 | if (line.compare(0, 9, "NOSUGGEST", 9) == 0) { |
| 479 | if (!parse_flag(line, &nosuggest, afflst)) { |
| 480 | finishFileMgr(afflst); |
| 481 | return 1; |
| 482 | } |
| 483 | } |
| 484 | |
| 485 | if (line.compare(0, 14, "NONGRAMSUGGEST", 14) == 0) { |
| 486 | if (!parse_flag(line, &nongramsuggest, afflst)) { |
| 487 | finishFileMgr(afflst); |
| 488 | return 1; |
| 489 | } |
| 490 | } |
| 491 | |
| 492 | /* parse in the flag used by forbidden words */ |
| 493 | if (line.compare(0, 13, "FORBIDDENWORD", 13) == 0) { |
| 494 | if (!parse_flag(line, &forbiddenword, afflst)) { |
| 495 | finishFileMgr(afflst); |
| 496 | return 1; |
| 497 | } |
| 498 | } |
| 499 | |
| 500 | /* parse in the flag used by forbidden words (is deprecated) */ |
| 501 | if (line.compare(0, 13, "LEMMA_PRESENT", 13) == 0) { |
| 502 | if (!parse_flag(line, &lemma_present, afflst)) { |
| 503 | finishFileMgr(afflst); |
| 504 | return 1; |
| 505 | } |
| 506 | } |
| 507 | |
| 508 | /* parse in the flag used by circumfixes */ |
| 509 | if (line.compare(0, 9, "CIRCUMFIX", 9) == 0) { |
| 510 | if (!parse_flag(line, &circumfix, afflst)) { |
| 511 | finishFileMgr(afflst); |
| 512 | return 1; |
| 513 | } |
| 514 | } |
| 515 | |
| 516 | /* parse in the flag used by fogemorphemes */ |
| 517 | if (line.compare(0, 14, "ONLYINCOMPOUND", 14) == 0) { |
| 518 | if (!parse_flag(line, &onlyincompound, afflst)) { |
| 519 | finishFileMgr(afflst); |
| 520 | return 1; |
| 521 | } |
| 522 | } |
| 523 | |
| 524 | /* parse in the flag used by `needaffixs' (is deprecated) */ |
| 525 | if (line.compare(0, 10, "PSEUDOROOT", 10) == 0) { |
| 526 | if (!parse_flag(line, &needaffix, afflst)) { |
| 527 | finishFileMgr(afflst); |
| 528 | return 1; |
| 529 | } |
| 530 | } |
| 531 | |
| 532 | /* parse in the flag used by `needaffixs' */ |
| 533 | if (line.compare(0, 9, "NEEDAFFIX", 9) == 0) { |
| 534 | if (!parse_flag(line, &needaffix, afflst)) { |
| 535 | finishFileMgr(afflst); |
| 536 | return 1; |
| 537 | } |
| 538 | } |
| 539 | |
| 540 | /* parse in the minimal length for words in compounds */ |
| 541 | if (line.compare(0, 11, "COMPOUNDMIN", 11) == 0) { |
| 542 | if (!parse_num(line, &cpdmin, afflst)) { |
| 543 | finishFileMgr(afflst); |
| 544 | return 1; |
| 545 | } |
| 546 | if (cpdmin < 1) |
| 547 | cpdmin = 1; |
| 548 | } |
| 549 | |
| 550 | /* parse in the max. words and syllables in compounds */ |
| 551 | if (line.compare(0, 16, "COMPOUNDSYLLABLE", 16) == 0) { |
| 552 | if (!parse_cpdsyllable(line, afflst)) { |
| 553 | finishFileMgr(afflst); |
| 554 | return 1; |
| 555 | } |
| 556 | } |
| 557 | |
| 558 | /* parse in the flag used by compound_check() method */ |
| 559 | if (line.compare(0, 11, "SYLLABLENUM", 11) == 0) { |
| 560 | if (!parse_string(line, cpdsyllablenum, afflst->getlinenum())) { |
| 561 | finishFileMgr(afflst); |
| 562 | return 1; |
| 563 | } |
| 564 | } |
| 565 | |
| 566 | /* parse in the flag used by the controlled compound words */ |
| 567 | if (line.compare(0, 8, "CHECKNUM", 8) == 0) { |
| 568 | checknum = 1; |
| 569 | } |
| 570 | |
| 571 | /* parse in the extra word characters */ |
| 572 | if (line.compare(0, 9, "WORDCHARS", 9) == 0) { |
| 573 | if (!parse_array(line, wordchars, wordchars_utf16, |
| 574 | utf8, afflst->getlinenum())) { |
| 575 | finishFileMgr(afflst); |
| 576 | return 1; |
| 577 | } |
| 578 | } |
| 579 | |
| 580 | /* parse in the ignored characters (for example, Arabic optional diacretics |
| 581 | * charachters */ |
| 582 | if (line.compare(0, 6, "IGNORE", 6) == 0) { |
| 583 | if (!parse_array(line, ignorechars, ignorechars_utf16, |
| 584 | utf8, afflst->getlinenum())) { |
| 585 | finishFileMgr(afflst); |
| 586 | return 1; |
| 587 | } |
| 588 | } |
| 589 | |
| 590 | /* parse in the input conversion table */ |
| 591 | if (line.compare(0, 5, "ICONV", 5) == 0) { |
| 592 | if (!parse_convtable(line, afflst, &iconvtable, "ICONV")) { |
| 593 | finishFileMgr(afflst); |
| 594 | return 1; |
| 595 | } |
| 596 | } |
| 597 | |
| 598 | /* parse in the output conversion table */ |
| 599 | if (line.compare(0, 5, "OCONV", 5) == 0) { |
| 600 | if (!parse_convtable(line, afflst, &oconvtable, "OCONV")) { |
| 601 | finishFileMgr(afflst); |
| 602 | return 1; |
| 603 | } |
| 604 | } |
| 605 | |
| 606 | /* parse in the phonetic translation table */ |
| 607 | if (line.compare(0, 5, "PHONE", 5) == 0) { |
| 608 | if (!parse_phonetable(line, afflst)) { |
| 609 | finishFileMgr(afflst); |
| 610 | return 1; |
| 611 | } |
| 612 | } |
| 613 | |
| 614 | /* parse in the checkcompoundpattern table */ |
| 615 | if (line.compare(0, 20, "CHECKCOMPOUNDPATTERN", 20) == 0) { |
| 616 | if (!parse_checkcpdtable(line, afflst)) { |
| 617 | finishFileMgr(afflst); |
| 618 | return 1; |
| 619 | } |
| 620 | } |
| 621 | |
| 622 | /* parse in the defcompound table */ |
| 623 | if (line.compare(0, 12, "COMPOUNDRULE", 12) == 0) { |
| 624 | if (!parse_defcpdtable(line, afflst)) { |
| 625 | finishFileMgr(afflst); |
| 626 | return 1; |
| 627 | } |
| 628 | } |
| 629 | |
| 630 | /* parse in the related character map table */ |
| 631 | if (line.compare(0, 3, "MAP", 3) == 0) { |
| 632 | if (!parse_maptable(line, afflst)) { |
| 633 | finishFileMgr(afflst); |
| 634 | return 1; |
| 635 | } |
| 636 | } |
| 637 | |
| 638 | /* parse in the word breakpoints table */ |
| 639 | if (line.compare(0, 5, "BREAK", 5) == 0) { |
| 640 | if (!parse_breaktable(line, afflst)) { |
| 641 | finishFileMgr(afflst); |
| 642 | return 1; |
| 643 | } |
| 644 | } |
| 645 | |
| 646 | /* parse in the language for language specific codes */ |
| 647 | if (line.compare(0, 4, "LANG", 4) == 0) { |
| 648 | if (!parse_string(line, lang, afflst->getlinenum())) { |
| 649 | finishFileMgr(afflst); |
| 650 | return 1; |
| 651 | } |
| 652 | langnum = get_lang_num(lang); |
| 653 | } |
| 654 | |
| 655 | if (line.compare(0, 7, "VERSION", 7) == 0) { |
| 656 | size_t startpos = line.find_first_not_of(" \t", 7); |
| 657 | if (startpos != std::string::npos) { |
| 658 | version = line.substr(startpos); |
| 659 | } |
| 660 | } |
| 661 | |
| 662 | if (line.compare(0, 12, "MAXNGRAMSUGS", 12) == 0) { |
| 663 | if (!parse_num(line, &maxngramsugs, afflst)) { |
| 664 | finishFileMgr(afflst); |
| 665 | return 1; |
| 666 | } |
| 667 | } |
| 668 | |
| 669 | if (line.compare(0, 11, "ONLYMAXDIFF", 11) == 0) |
| 670 | onlymaxdiff = 1; |
| 671 | |
| 672 | if (line.compare(0, 7, "MAXDIFF", 7) == 0) { |
| 673 | if (!parse_num(line, &maxdiff, afflst)) { |
| 674 | finishFileMgr(afflst); |
| 675 | return 1; |
| 676 | } |
| 677 | } |
| 678 | |
| 679 | if (line.compare(0, 10, "MAXCPDSUGS", 10) == 0) { |
| 680 | if (!parse_num(line, &maxcpdsugs, afflst)) { |
| 681 | finishFileMgr(afflst); |
| 682 | return 1; |
| 683 | } |
| 684 | } |
| 685 | |
| 686 | if (line.compare(0, 11, "NOSPLITSUGS", 11) == 0) { |
| 687 | nosplitsugs = 1; |
| 688 | } |
| 689 | |
| 690 | if (line.compare(0, 9, "FULLSTRIP", 9) == 0) { |
| 691 | fullstrip = 1; |
| 692 | } |
| 693 | |
| 694 | if (line.compare(0, 12, "SUGSWITHDOTS", 12) == 0) { |
| 695 | sugswithdots = 1; |
| 696 | } |
| 697 | |
| 698 | /* parse in the flag used by forbidden words */ |
| 699 | if (line.compare(0, 8, "KEEPCASE", 8) == 0) { |
| 700 | if (!parse_flag(line, &keepcase, afflst)) { |
| 701 | finishFileMgr(afflst); |
| 702 | return 1; |
| 703 | } |
| 704 | } |
| 705 | |
| 706 | /* parse in the flag used by `forceucase' */ |
| 707 | if (line.compare(0, 10, "FORCEUCASE", 10) == 0) { |
| 708 | if (!parse_flag(line, &forceucase, afflst)) { |
| 709 | finishFileMgr(afflst); |
| 710 | return 1; |
| 711 | } |
| 712 | } |
| 713 | |
| 714 | /* parse in the flag used by `warn' */ |
| 715 | if (line.compare(0, 4, "WARN", 4) == 0) { |
| 716 | if (!parse_flag(line, &warn, afflst)) { |
| 717 | finishFileMgr(afflst); |
| 718 | return 1; |
| 719 | } |
| 720 | } |
| 721 | |
| 722 | if (line.compare(0, 10, "FORBIDWARN", 10) == 0) { |
| 723 | forbidwarn = 1; |
| 724 | } |
| 725 | |
| 726 | /* parse in the flag used by the affix generator */ |
| 727 | if (line.compare(0, 11, "SUBSTANDARD", 11) == 0) { |
| 728 | if (!parse_flag(line, &substandard, afflst)) { |
| 729 | finishFileMgr(afflst); |
| 730 | return 1; |
| 731 | } |
| 732 | } |
| 733 | |
| 734 | if (line.compare(0, 11, "CHECKSHARPS", 11) == 0) { |
| 735 | checksharps = 1; |
| 736 | } |
| 737 | |
| 738 | /* parse this affix: P - prefix, S - suffix */ |
| 739 | // affix type |
| 740 | char ft = ' '; |
| 741 | if (line.compare(0, 3, "PFX", 3) == 0) |
| 742 | ft = complexprefixes ? 'S' : 'P'; |
| 743 | if (line.compare(0, 3, "SFX", 3) == 0) |
| 744 | ft = complexprefixes ? 'P' : 'S'; |
| 745 | if (ft != ' ') { |
| 746 | if (dupflags_ini) { |
| 747 | memset(dupflags, 0, sizeof(dupflags)); |
| 748 | dupflags_ini = 0; |
| 749 | } |
| 750 | if (!parse_affix(line, ft, afflst, dupflags)) { |
| 751 | finishFileMgr(afflst); |
| 752 | return 1; |
| 753 | } |
| 754 | } |
| 755 | } |
| 756 | |
| 757 | finishFileMgr(afflst); |
| 758 | // affix trees are sorted now |
| 759 | |
| 760 | // now we can speed up performance greatly taking advantage of the |
| 761 | // relationship between the affixes and the idea of "subsets". |
| 762 | |
| 763 | // View each prefix as a potential leading subset of another and view |
| 764 | // each suffix (reversed) as a potential trailing subset of another. |
| 765 | |
| 766 | // To illustrate this relationship if we know the prefix "ab" is found in the |
| 767 | // word to examine, only prefixes that "ab" is a leading subset of need be |
| 768 | // examined. |
| 769 | // Furthermore is "ab" is not present then none of the prefixes that "ab" is |
| 770 | // is a subset need be examined. |
| 771 | // The same argument goes for suffix string that are reversed. |
| 772 | |
| 773 | // Then to top this off why not examine the first char of the word to quickly |
| 774 | // limit the set of prefixes to examine (i.e. the prefixes to examine must |
| 775 | // be leading supersets of the first character of the word (if they exist) |
| 776 | |
| 777 | // To take advantage of this "subset" relationship, we need to add two links |
| 778 | // from entry. One to take next if the current prefix is found (call it |
| 779 | // nexteq) |
| 780 | // and one to take next if the current prefix is not found (call it nextne). |
| 781 | |
| 782 | // Since we have built ordered lists, all that remains is to properly |
| 783 | // initialize |
| 784 | // the nextne and nexteq pointers that relate them |
| 785 | |
| 786 | process_pfx_order(); |
| 787 | process_sfx_order(); |
| 788 | |
| 789 | return 0; |
| 790 | } |
| 791 | |
| 792 | // we want to be able to quickly access prefix information |
| 793 | // both by prefix flag, and sorted by prefix string itself |
| 794 | // so we need to set up two indexes |
| 795 | |
| 796 | int AffixMgr::build_pfxtree(PfxEntry* pfxptr) { |
| 797 | PfxEntry* ptr; |
| 798 | PfxEntry* pptr; |
| 799 | PfxEntry* ep = pfxptr; |
| 800 | |
| 801 | // get the right starting points |
| 802 | const char* key = ep->getKey(); |
| 803 | const auto flg = (unsigned char)(ep->getFlag() & 0x00FF); |
| 804 | |
| 805 | // first index by flag which must exist |
| 806 | ptr = pFlag[flg]; |
| 807 | ep->setFlgNxt(ptr); |
| 808 | pFlag[flg] = ep; |
| 809 | |
| 810 | // handle the special case of null affix string |
| 811 | if (*key == '\0') { |
| 812 | // always inset them at head of list at element 0 |
| 813 | ptr = pStart[0]; |
| 814 | ep->setNext(ptr); |
| 815 | pStart[0] = ep; |
| 816 | return 0; |
| 817 | } |
| 818 | |
| 819 | // now handle the normal case |
| 820 | ep->setNextEQ(nullptr); |
| 821 | ep->setNextNE(nullptr); |
| 822 | |
| 823 | unsigned char sp = *((const unsigned char*)key); |
| 824 | ptr = pStart[sp]; |
| 825 | |
| 826 | // handle the first insert |
| 827 | if (!ptr) { |
| 828 | pStart[sp] = ep; |
| 829 | return 0; |
| 830 | } |
| 831 | |
| 832 | // otherwise use binary tree insertion so that a sorted |
| 833 | // list can easily be generated later |
| 834 | pptr = nullptr; |
| 835 | for (;;) { |
| 836 | pptr = ptr; |
| 837 | if (strcmp(ep->getKey(), ptr->getKey()) <= 0) { |
| 838 | ptr = ptr->getNextEQ(); |
| 839 | if (!ptr) { |
| 840 | pptr->setNextEQ(ep); |
| 841 | break; |
| 842 | } |
| 843 | } else { |
| 844 | ptr = ptr->getNextNE(); |
| 845 | if (!ptr) { |
| 846 | pptr->setNextNE(ep); |
| 847 | break; |
| 848 | } |
| 849 | } |
| 850 | } |
| 851 | return 0; |
| 852 | } |
| 853 | |
| 854 | // we want to be able to quickly access suffix information |
| 855 | // both by suffix flag, and sorted by the reverse of the |
| 856 | // suffix string itself; so we need to set up two indexes |
| 857 | int AffixMgr::build_sfxtree(SfxEntry* sfxptr) { |
| 858 | |
| 859 | sfxptr->initReverseWord(); |
| 860 | |
| 861 | SfxEntry* ptr; |
| 862 | SfxEntry* pptr; |
| 863 | SfxEntry* ep = sfxptr; |
| 864 | |
| 865 | /* get the right starting point */ |
| 866 | const char* key = ep->getKey(); |
| 867 | const auto flg = (unsigned char)(ep->getFlag() & 0x00FF); |
| 868 | |
| 869 | // first index by flag which must exist |
| 870 | ptr = sFlag[flg]; |
| 871 | ep->setFlgNxt(ptr); |
| 872 | sFlag[flg] = ep; |
| 873 | |
| 874 | // next index by affix string |
| 875 | |
| 876 | // handle the special case of null affix string |
| 877 | if (*key == '\0') { |
| 878 | // always inset them at head of list at element 0 |
| 879 | ptr = sStart[0]; |
| 880 | ep->setNext(ptr); |
| 881 | sStart[0] = ep; |
| 882 | return 0; |
| 883 | } |
| 884 | |
| 885 | // now handle the normal case |
| 886 | ep->setNextEQ(nullptr); |
| 887 | ep->setNextNE(nullptr); |
| 888 | |
| 889 | unsigned char sp = *((const unsigned char*)key); |
| 890 | ptr = sStart[sp]; |
| 891 | |
| 892 | // handle the first insert |
| 893 | if (!ptr) { |
| 894 | sStart[sp] = ep; |
| 895 | return 0; |
| 896 | } |
| 897 | |
| 898 | // otherwise use binary tree insertion so that a sorted |
| 899 | // list can easily be generated later |
| 900 | pptr = nullptr; |
| 901 | for (;;) { |
| 902 | pptr = ptr; |
| 903 | if (strcmp(ep->getKey(), ptr->getKey()) <= 0) { |
| 904 | ptr = ptr->getNextEQ(); |
| 905 | if (!ptr) { |
| 906 | pptr->setNextEQ(ep); |
| 907 | break; |
| 908 | } |
| 909 | } else { |
| 910 | ptr = ptr->getNextNE(); |
| 911 | if (!ptr) { |
| 912 | pptr->setNextNE(ep); |
| 913 | break; |
| 914 | } |
| 915 | } |
| 916 | } |
| 917 | return 0; |
| 918 | } |
| 919 | |
| 920 | // convert from binary tree to sorted list |
| 921 | int AffixMgr::process_pfx_tree_to_list() { |
| 922 | for (int i = 1; i < SETSIZE256; i++) { |
| 923 | pStart[i] = process_pfx_in_order(pStart[i], nullptr); |
| 924 | } |
| 925 | return 0; |
| 926 | } |
| 927 | |
| 928 | PfxEntry* AffixMgr::process_pfx_in_order(PfxEntry* ptr, PfxEntry* nptr) { |
| 929 | if (ptr) { |
| 930 | nptr = process_pfx_in_order(ptr->getNextNE(), nptr); |
| 931 | ptr->setNext(nptr); |
| 932 | nptr = process_pfx_in_order(ptr->getNextEQ(), ptr); |
| 933 | } |
| 934 | return nptr; |
| 935 | } |
| 936 | |
| 937 | // convert from binary tree to sorted list |
| 938 | int AffixMgr::process_sfx_tree_to_list() { |
| 939 | for (int i = 1; i < SETSIZE256; i++) { |
| 940 | sStart[i] = process_sfx_in_order(sStart[i], nullptr); |
| 941 | } |
| 942 | return 0; |
| 943 | } |
| 944 | |
| 945 | SfxEntry* AffixMgr::process_sfx_in_order(SfxEntry* ptr, SfxEntry* nptr) { |
| 946 | if (ptr) { |
| 947 | nptr = process_sfx_in_order(ptr->getNextNE(), nptr); |
| 948 | ptr->setNext(nptr); |
| 949 | nptr = process_sfx_in_order(ptr->getNextEQ(), ptr); |
| 950 | } |
| 951 | return nptr; |
| 952 | } |
| 953 | |
| 954 | // reinitialize the PfxEntry links NextEQ and NextNE to speed searching |
| 955 | // using the idea of leading subsets this time |
| 956 | int AffixMgr::process_pfx_order() { |
| 957 | PfxEntry* ptr; |
| 958 | |
| 959 | // loop through each prefix list starting point |
| 960 | for (int i = 1; i < SETSIZE256; i++) { |
| 961 | ptr = pStart[i]; |
| 962 | |
| 963 | // look through the remainder of the list |
| 964 | // and find next entry with affix that |
| 965 | // the current one is not a subset of |
| 966 | // mark that as destination for NextNE |
| 967 | // use next in list that you are a subset |
| 968 | // of as NextEQ |
| 969 | |
| 970 | for (; ptr != nullptr; ptr = ptr->getNext()) { |
| 971 | PfxEntry* nptr = ptr->getNext(); |
| 972 | for (; nptr != nullptr; nptr = nptr->getNext()) { |
| 973 | if (!isSubset(ptr->getKey(), nptr->getKey())) |
| 974 | break; |
| 975 | } |
| 976 | ptr->setNextNE(nptr); |
| 977 | ptr->setNextEQ(nullptr); |
| 978 | if ((ptr->getNext()) && |
| 979 | isSubset(ptr->getKey(), (ptr->getNext())->getKey())) |
| 980 | ptr->setNextEQ(ptr->getNext()); |
| 981 | } |
| 982 | |
| 983 | // now clean up by adding smart search termination strings: |
| 984 | // if you are already a superset of the previous prefix |
| 985 | // but not a subset of the next, search can end here |
| 986 | // so set NextNE properly |
| 987 | |
| 988 | ptr = pStart[i]; |
| 989 | for (; ptr != nullptr; ptr = ptr->getNext()) { |
| 990 | PfxEntry* nptr = ptr->getNext(); |
| 991 | PfxEntry* mptr = nullptr; |
| 992 | for (; nptr != nullptr; nptr = nptr->getNext()) { |
| 993 | if (!isSubset(ptr->getKey(), nptr->getKey())) |
| 994 | break; |
| 995 | mptr = nptr; |
| 996 | } |
| 997 | if (mptr) |
| 998 | mptr->setNextNE(nullptr); |
| 999 | } |
| 1000 | } |
| 1001 | return 0; |
| 1002 | } |
| 1003 | |
| 1004 | // initialize the SfxEntry links NextEQ and NextNE to speed searching |
| 1005 | // using the idea of leading subsets this time |
| 1006 | int AffixMgr::process_sfx_order() { |
| 1007 | SfxEntry* ptr; |
| 1008 | |
| 1009 | // loop through each prefix list starting point |
| 1010 | for (int i = 1; i < SETSIZE256; i++) { |
| 1011 | ptr = sStart[i]; |
| 1012 | |
| 1013 | // look through the remainder of the list |
| 1014 | // and find next entry with affix that |
| 1015 | // the current one is not a subset of |
| 1016 | // mark that as destination for NextNE |
| 1017 | // use next in list that you are a subset |
| 1018 | // of as NextEQ |
| 1019 | |
| 1020 | for (; ptr != nullptr; ptr = ptr->getNext()) { |
| 1021 | SfxEntry* nptr = ptr->getNext(); |
| 1022 | for (; nptr != nullptr; nptr = nptr->getNext()) { |
| 1023 | if (!isSubset(ptr->getKey(), nptr->getKey())) |
| 1024 | break; |
| 1025 | } |
| 1026 | ptr->setNextNE(nptr); |
| 1027 | ptr->setNextEQ(nullptr); |
| 1028 | if ((ptr->getNext()) && |
| 1029 | isSubset(ptr->getKey(), (ptr->getNext())->getKey())) |
| 1030 | ptr->setNextEQ(ptr->getNext()); |
| 1031 | } |
| 1032 | |
| 1033 | // now clean up by adding smart search termination strings: |
| 1034 | // if you are already a superset of the previous suffix |
| 1035 | // but not a subset of the next, search can end here |
| 1036 | // so set NextNE properly |
| 1037 | |
| 1038 | ptr = sStart[i]; |
| 1039 | for (; ptr != nullptr; ptr = ptr->getNext()) { |
| 1040 | SfxEntry* nptr = ptr->getNext(); |
| 1041 | SfxEntry* mptr = nullptr; |
| 1042 | for (; nptr != nullptr; nptr = nptr->getNext()) { |
| 1043 | if (!isSubset(ptr->getKey(), nptr->getKey())) |
| 1044 | break; |
| 1045 | mptr = nptr; |
| 1046 | } |
| 1047 | if (mptr) |
| 1048 | mptr->setNextNE(nullptr); |
| 1049 | } |
| 1050 | } |
| 1051 | return 0; |
| 1052 | } |
| 1053 | |
| 1054 | // add flags to the result for dictionary debugging |
| 1055 | std::string& AffixMgr::debugflag(std::string& result, unsigned short flag) { |
| 1056 | std::string st = encode_flag(flag); |
| 1057 | result.push_back(MSEP_FLD' '); |
| 1058 | result.append(MORPH_FLAG"fl:"); |
| 1059 | result.append(st); |
| 1060 | return result; |
| 1061 | } |
| 1062 | |
| 1063 | // Count match positions in an .aff condition pattern: each [...] group |
| 1064 | // counts as one, each non-grouped codepoint counts as one. |
| 1065 | int AffixMgr::condlen(const std::string& s) { |
| 1066 | int l = 0; |
| 1067 | bool group = false; |
| 1068 | size_t i = 0; |
| 1069 | while (i < s.size()) { |
| 1070 | if (s[i] == '[') { |
| 1071 | group = true; |
| 1072 | ++l; |
| 1073 | ++i; |
| 1074 | } else if (s[i] == ']') { |
| 1075 | group = false; |
| 1076 | ++i; |
| 1077 | } else if (group) { |
| 1078 | ++i; |
| 1079 | } else { |
| 1080 | ++l; |
| 1081 | i = utf8 ? utf8_next(s, i) : i + 1; |
| 1082 | } |
| 1083 | } |
| 1084 | return l; |
| 1085 | } |
| 1086 | |
| 1087 | int AffixMgr::encodeit(AffEntry& entry, const std::string& cs) { |
| 1088 | if (cs.compare(".") != 0) { |
| 1089 | int n = condlen(cs); |
| 1090 | if (n > std::numeric_limits<unsigned char>::max()) { |
| 1091 | HUNSPELL_WARNING(stderrstderr, "error: condition length %d is over max limit\n", n); |
| 1092 | return 1; |
| 1093 | } |
| 1094 | entry.numconds = (unsigned char)n; |
| 1095 | const size_t cslen = cs.size(); |
| 1096 | const size_t short_part = std::min<size_t>(MAXCONDLEN20, cslen); |
| 1097 | memcpy(entry.c.conds, cs.data(), short_part); |
| 1098 | if (short_part < MAXCONDLEN20) { |
| 1099 | //blank out the remaining space |
| 1100 | memset(entry.c.conds + short_part, 0, MAXCONDLEN20 - short_part); |
| 1101 | } else if (cs[MAXCONDLEN20]) { |
| 1102 | //there is more conditions than fit in fixed space, so its |
| 1103 | //a long condition |
| 1104 | entry.opts |= aeLONGCOND(1 << 4); |
| 1105 | size_t remaining = cs.size() - MAXCONDLEN_1(20 - sizeof(char*)); |
| 1106 | entry.c.l.conds2 = new char[1 + remaining]; |
| 1107 | memcpy(entry.c.l.conds2, cs.data() + MAXCONDLEN_1(20 - sizeof(char*)), remaining); |
| 1108 | entry.c.l.conds2[remaining] = 0; |
| 1109 | } |
| 1110 | } else { |
| 1111 | entry.numconds = 0; |
| 1112 | entry.c.conds[0] = '\0'; |
| 1113 | } |
| 1114 | return 0; |
| 1115 | } |
| 1116 | |
| 1117 | // return 1 if s1 is a leading subset of s2 (dots are for infixes) |
| 1118 | inline int AffixMgr::isSubset(const char* s1, const char* s2) { |
| 1119 | while (((*s1 == *s2) || (*s1 == '.')) && (*s1 != '\0') && (*s2 != '\0')) { |
| 1120 | s1++; |
| 1121 | s2++; |
| 1122 | } |
| 1123 | return (*s1 == '\0'); |
| 1124 | } |
| 1125 | |
| 1126 | // check word for prefixes |
| 1127 | struct hentry* AffixMgr::prefix_check(const std::string& word, |
| 1128 | int start, |
| 1129 | int len, |
| 1130 | char in_compound, |
| 1131 | AffixScratch& scratch, |
| 1132 | const FLAGunsigned short needflag, |
| 1133 | const FLAGunsigned short avoidflag) { |
| 1134 | struct hentry* rv = nullptr; |
| 1135 | |
| 1136 | pfx = nullptr; |
| 1137 | pfxappnd = nullptr; |
| 1138 | sfxappnd = nullptr; |
| 1139 | sfxextra = 0; |
| 1140 | |
| 1141 | TraceCtx* t = trace_on(scratch.trace); |
| 1142 | int candidates = 0; |
| 1143 | |
| 1144 | // first handle the special case of 0 length prefixes |
| 1145 | PfxEntry* pe = pStart[0]; |
| 1146 | while (pe) { |
| 1147 | if ( |
| 1148 | // fogemorpheme |
| 1149 | ((in_compound != IN_CPD_NOT0) || |
| 1150 | !(pe->getCont() && |
| 1151 | (TESTAFF(pe->getCont(), onlyincompound, pe->getContLen())(std::binary_search(pe->getCont(), pe->getCont() + pe-> getContLen(), onlyincompound))))) && |
| 1152 | // permit prefixes in compounds |
| 1153 | ((in_compound != IN_CPD_END2) || |
| 1154 | (pe->getCont() && |
| 1155 | (TESTAFF(pe->getCont(), compoundpermitflag, pe->getContLen())(std::binary_search(pe->getCont(), pe->getCont() + pe-> getContLen(), compoundpermitflag)))))) { |
| 1156 | // check prefix |
| 1157 | ++candidates; |
| 1158 | rv = pe->checkword(word, start, len, in_compound, needflag, scratch); |
| 1159 | // Skip a stem with the avoid flag and keep scanning the other prefixes. |
| 1160 | if (rv && avoidflag != FLAG_NULL0x00 && TESTAFF(rv->astr, avoidflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, avoidflag ))) { |
| 1161 | trace_avoidflag(t, avoidflag, rv); |
| 1162 | rv = nullptr; |
| 1163 | } |
| 1164 | if (rv) { |
| 1165 | pfx = pe; // BUG: pfx not stateless |
| 1166 | return rv; |
| 1167 | } |
| 1168 | } |
| 1169 | pe = pe->getNext(); |
| 1170 | } |
| 1171 | |
| 1172 | // now handle the general case |
| 1173 | unsigned char sp = word[start]; |
| 1174 | PfxEntry* pptr = pStart[sp]; |
| 1175 | |
| 1176 | while (pptr) { |
| 1177 | if (isSubset(pptr->getKey(), word.c_str() + start)) { |
| 1178 | if ( |
| 1179 | // fogemorpheme |
| 1180 | ((in_compound != IN_CPD_NOT0) || |
| 1181 | !(pptr->getCont() && |
| 1182 | (TESTAFF(pptr->getCont(), onlyincompound, pptr->getContLen())(std::binary_search(pptr->getCont(), pptr->getCont() + pptr ->getContLen(), onlyincompound))))) && |
| 1183 | // permit prefixes in compounds |
| 1184 | ((in_compound != IN_CPD_END2) || |
| 1185 | (pptr->getCont() && (TESTAFF(pptr->getCont(), compoundpermitflag,(std::binary_search(pptr->getCont(), pptr->getCont() + pptr ->getContLen(), compoundpermitflag)) |
| 1186 | pptr->getContLen())(std::binary_search(pptr->getCont(), pptr->getCont() + pptr ->getContLen(), compoundpermitflag)))))) { |
| 1187 | // check prefix |
| 1188 | ++candidates; |
| 1189 | rv = pptr->checkword(word, start, len, in_compound, needflag, scratch); |
| 1190 | if (rv && avoidflag != FLAG_NULL0x00 && TESTAFF(rv->astr, avoidflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, avoidflag ))) { |
| 1191 | trace_avoidflag(t, avoidflag, rv); |
| 1192 | rv = nullptr; |
| 1193 | } |
| 1194 | if (rv) { |
| 1195 | pfx = pptr; // BUG: pfx not stateless |
| 1196 | return rv; |
| 1197 | } |
| 1198 | } |
| 1199 | pptr = pptr->getNextEQ(); |
| 1200 | } else { |
| 1201 | pptr = pptr->getNextNE(); |
| 1202 | } |
| 1203 | } |
| 1204 | |
| 1205 | // a pass that had nothing to try says so, which is different from a pass that |
| 1206 | // did not run |
| 1207 | if (t && candidates == 0) |
| 1208 | trace(*t, "pfx \"%s\" candidates=0", word.substr(start, len).c_str()); |
| 1209 | |
| 1210 | return nullptr; |
| 1211 | } |
| 1212 | |
| 1213 | // check word for prefixes and two-level suffixes |
| 1214 | struct hentry* AffixMgr::prefix_check_twosfx(const std::string& word, |
| 1215 | int start, |
| 1216 | int len, |
| 1217 | char in_compound, |
| 1218 | AffixScratch& scratch, |
| 1219 | const FLAGunsigned short needflag) { |
| 1220 | struct hentry* rv = nullptr; |
| 1221 | |
| 1222 | pfx = nullptr; |
| 1223 | sfxappnd = nullptr; |
| 1224 | sfxextra = 0; |
| 1225 | |
| 1226 | // first handle the special case of 0 length prefixes |
| 1227 | PfxEntry* pe = pStart[0]; |
| 1228 | |
| 1229 | while (pe) { |
| 1230 | rv = pe->check_twosfx(word, start, len, in_compound, needflag, scratch); |
| 1231 | if (rv) |
| 1232 | return rv; |
| 1233 | pe = pe->getNext(); |
| 1234 | } |
| 1235 | |
| 1236 | // now handle the general case |
| 1237 | unsigned char sp = word[start]; |
| 1238 | PfxEntry* pptr = pStart[sp]; |
| 1239 | |
| 1240 | while (pptr) { |
| 1241 | if (isSubset(pptr->getKey(), word.c_str() + start)) { |
| 1242 | rv = pptr->check_twosfx(word, start, len, in_compound, needflag, scratch); |
| 1243 | if (rv) { |
| 1244 | pfx = pptr; |
| 1245 | return rv; |
| 1246 | } |
| 1247 | pptr = pptr->getNextEQ(); |
| 1248 | } else { |
| 1249 | pptr = pptr->getNextNE(); |
| 1250 | } |
| 1251 | } |
| 1252 | |
| 1253 | return nullptr; |
| 1254 | } |
| 1255 | |
| 1256 | // check word for prefixes and morph |
| 1257 | std::string AffixMgr::prefix_check_morph(const std::string& word, |
| 1258 | int start, |
| 1259 | int len, |
| 1260 | char in_compound, |
| 1261 | AffixScratch& scratch, |
| 1262 | const FLAGunsigned short needflag) { |
| 1263 | |
| 1264 | // many prefixes can analyse a word the same way |
| 1265 | DistinctRecords result(MSEP_REC'\n'); |
| 1266 | |
| 1267 | pfx = nullptr; |
| 1268 | sfxappnd = nullptr; |
| 1269 | sfxextra = 0; |
| 1270 | |
| 1271 | // first handle the special case of 0 length prefixes |
| 1272 | PfxEntry* pe = pStart[0]; |
| 1273 | while (pe) { |
| 1274 | std::string st = pe->check_morph(word, start, len, in_compound, needflag, scratch); |
| 1275 | if (!st.empty()) { |
| 1276 | result.append(st); |
| 1277 | } |
| 1278 | pe = pe->getNext(); |
| 1279 | } |
| 1280 | |
| 1281 | // now handle the general case |
| 1282 | unsigned char sp = word[start]; |
| 1283 | PfxEntry* pptr = pStart[sp]; |
| 1284 | |
| 1285 | while (pptr) { |
| 1286 | if (isSubset(pptr->getKey(), word.c_str() + start)) { |
| 1287 | std::string st = pptr->check_morph(word, start, len, in_compound, needflag, scratch); |
| 1288 | if (!st.empty()) { |
| 1289 | // fogemorpheme |
| 1290 | if ((in_compound != IN_CPD_NOT0) || |
| 1291 | !((pptr->getCont() && (TESTAFF(pptr->getCont(), onlyincompound,(std::binary_search(pptr->getCont(), pptr->getCont() + pptr ->getContLen(), onlyincompound)) |
| 1292 | pptr->getContLen())(std::binary_search(pptr->getCont(), pptr->getCont() + pptr ->getContLen(), onlyincompound)))))) { |
| 1293 | result.append(st); |
| 1294 | pfx = pptr; |
| 1295 | } |
| 1296 | } |
| 1297 | pptr = pptr->getNextEQ(); |
| 1298 | } else { |
| 1299 | pptr = pptr->getNextNE(); |
| 1300 | } |
| 1301 | } |
| 1302 | |
| 1303 | return result.join(); |
| 1304 | } |
| 1305 | |
| 1306 | // check word for prefixes and morph and two-level suffixes |
| 1307 | std::string AffixMgr::prefix_check_twosfx_morph(const std::string& word, |
| 1308 | int start, |
| 1309 | int len, |
| 1310 | char in_compound, |
| 1311 | AffixScratch& scratch, |
| 1312 | const FLAGunsigned short needflag) { |
| 1313 | // every prefix crosses with every pair of suffixes |
| 1314 | DistinctRecords result(MSEP_REC'\n'); |
| 1315 | |
| 1316 | pfx = nullptr; |
| 1317 | sfxappnd = nullptr; |
| 1318 | sfxextra = 0; |
| 1319 | |
| 1320 | // first handle the special case of 0 length prefixes |
| 1321 | PfxEntry* pe = pStart[0]; |
| 1322 | while (pe) { |
| 1323 | std::string st = pe->check_twosfx_morph(word, start, len, in_compound, needflag, scratch); |
| 1324 | if (!st.empty()) { |
| 1325 | result.append(st); |
| 1326 | } |
| 1327 | pe = pe->getNext(); |
| 1328 | } |
| 1329 | |
| 1330 | // now handle the general case |
| 1331 | unsigned char sp = word[start]; |
| 1332 | PfxEntry* pptr = pStart[sp]; |
| 1333 | |
| 1334 | while (pptr) { |
| 1335 | if (isSubset(pptr->getKey(), word.c_str() + start)) { |
| 1336 | std::string st = pptr->check_twosfx_morph(word, start, len, in_compound, needflag, scratch); |
| 1337 | if (!st.empty()) { |
| 1338 | result.append(st); |
| 1339 | pfx = pptr; |
| 1340 | } |
| 1341 | pptr = pptr->getNextEQ(); |
| 1342 | } else { |
| 1343 | pptr = pptr->getNextNE(); |
| 1344 | } |
| 1345 | } |
| 1346 | |
| 1347 | return result.join(); |
| 1348 | } |
| 1349 | |
| 1350 | // Is word a non-compound with a REP substitution (see checkcompoundrep)? |
| 1351 | int AffixMgr::cpdrep_check(const std::string& in_word, |
| 1352 | int wl, |
| 1353 | AffixScratch& scratch, |
| 1354 | bool& timelimit_exceeded, |
| 1355 | std::chrono::steady_clock::time_point clock_time_start) { |
| 1356 | if ((wl < 2) || get_reptable().empty()) |
| 1357 | return 0; |
| 1358 | |
| 1359 | std::string word(in_word, 0, wl); |
| 1360 | |
| 1361 | for (const auto& i : get_reptable()) { |
| 1362 | if (timelimit_exceeded || std::chrono::steady_clock::now() - clock_time_start > TIMELIMIT_MSstd::chrono::milliseconds(50)) { |
| 1363 | timelimit_exceeded = true; |
| 1364 | return 0; |
| 1365 | } |
| 1366 | // use only available mid patterns |
| 1367 | if (!i.outstrings[0].empty()) { |
| 1368 | size_t r = 0; |
| 1369 | const size_t lenp = i.pattern.size(); |
| 1370 | // search every occurence of the pattern in the word |
| 1371 | while ((r = word.find(i.pattern, r)) != std::string::npos) { |
| 1372 | std::string candidate(word); |
| 1373 | candidate.replace(r, lenp, i.outstrings[0]); |
| 1374 | if (candidate_check(candidate, scratch)) |
| 1375 | return 1; |
| 1376 | ++r; // search for the next letter |
| 1377 | } |
| 1378 | } |
| 1379 | } |
| 1380 | |
| 1381 | return 0; |
| 1382 | } |
| 1383 | |
| 1384 | // forbid compound words, if they are in the dictionary as a |
| 1385 | // word pair separated by space |
| 1386 | int AffixMgr::cpdwordpair_check(const std::string& word, |
| 1387 | int wl, |
| 1388 | AffixScratch& scratch, |
| 1389 | bool& timelimit_exceeded, |
| 1390 | std::chrono::steady_clock::time_point clock_time_start) { |
| 1391 | TraceCtx* t = trace_on(scratch.trace); |
| 1392 | int pair_found = 0; |
| 1393 | std::string pair; |
| 1394 | |
| 1395 | { |
| 1396 | // this check puts a space at every position in turn, so the trace reports the verdict alone |
| 1397 | TraceSuppress no_trace(scratch.trace); |
| 1398 | |
| 1399 | if (wl > 2) { |
| 1400 | std::string candidate(word, 0, wl); |
| 1401 | for (size_t i = 1; i < candidate.size(); i++) { |
| 1402 | if (timelimit_exceeded || std::chrono::steady_clock::now() - clock_time_start > TIMELIMIT_MSstd::chrono::milliseconds(50)) { |
| 1403 | timelimit_exceeded = true; |
| 1404 | break; |
| 1405 | } |
| 1406 | // go to end of the UTF-8 character |
| 1407 | if (utf8 && is_utf8_cont(candidate[i])) |
| 1408 | continue; |
| 1409 | candidate.insert(i, 1, ' '); |
| 1410 | if (candidate_check(candidate, scratch)) { |
| 1411 | pair_found = 1; |
| 1412 | pair = candidate; |
| 1413 | break; |
| 1414 | } |
| 1415 | candidate.erase(i, 1); |
| 1416 | } |
| 1417 | } |
| 1418 | } |
| 1419 | |
| 1420 | if (t) { |
| 1421 | if (pair_found) |
| 1422 | trace(*t, "test wordpair pair=\"%s\" -> fail, the parts are a known word pair", pair.c_str()); |
| 1423 | else if (!timelimit_exceeded) |
| 1424 | trace(*t, "test wordpair -> pass, no space splits this into a known word pair"); |
| 1425 | } |
| 1426 | |
| 1427 | return pair_found; |
| 1428 | } |
| 1429 | |
| 1430 | // a CHECKCOMPOUNDPATTERN half may leave its flag out, and then any word is allowed to stand there |
| 1431 | static std::string trace_cond_flag(const AffixMgr* pAMgr, FLAGunsigned short cond) { |
| 1432 | if (cond == FLAG_NULL0x00) |
| 1433 | return "(any)"; |
| 1434 | return trace_flag(pAMgr, cond); |
| 1435 | } |
| 1436 | |
| 1437 | // one side of a compound join, as the entry it was found in and the flags that entry carries |
| 1438 | static std::string trace_join_word(const AffixMgr* pAMgr, const hentry* entry) { |
| 1439 | if (!entry) |
| 1440 | return "(none)"; |
| 1441 | return "\"" + std::string(entry->word, entry->blen) + "\"/" + |
| 1442 | trace_flags(pAMgr, entry->astr, entry->alen); |
| 1443 | } |
| 1444 | |
| 1445 | // one side of a compound join carries a flag when the dictionary entry has it, or when a prefix or |
| 1446 | // suffix that built that side passes it on through its continuation classes |
| 1447 | static bool join_side_has_flag(const hentry* entry, FLAGunsigned short cond, PfxEntry* p, SfxEntry* s) { |
| 1448 | return (entry->astr && TESTAFF(entry->astr, cond, entry->alen)(std::binary_search(entry->astr, entry->astr + entry-> alen, cond))) || |
| 1449 | (p && p->getCont() && TESTAFF(p->getCont(), cond, p->getContLen())(std::binary_search(p->getCont(), p->getCont() + p-> getContLen(), cond))) || |
| 1450 | (s && s->getCont() && TESTAFF(s->getCont(), cond, s->getContLen())(std::binary_search(s->getCont(), s->getCont() + s-> getContLen(), cond))); |
| 1451 | } |
| 1452 | |
| 1453 | // forbid compoundings when there are special patterns at word bound |
| 1454 | int AffixMgr::cpdpat_check(const std::string& word, |
| 1455 | size_t pos, |
| 1456 | hentry* r1, |
| 1457 | hentry* r2, |
| 1458 | const char /*affixed*/, |
| 1459 | const TraceCtx* t, |
| 1460 | PfxEntry* p1, |
| 1461 | SfxEntry* s1, |
| 1462 | PfxEntry* p2, |
| 1463 | SfxEntry* s2) { |
| 1464 | for (auto& i : checkcpdtable) { |
| 1465 | size_t len; |
| 1466 | bool right_text_ok = isSubset(i.pattern2.c_str(), word.c_str() + pos); |
| 1467 | bool left_flag_ok = right_text_ok && |
| 1468 | (!r1 || !i.cond || join_side_has_flag(r1, i.cond, p1, s1)); |
| 1469 | bool right_flag_ok = left_flag_ok && |
| 1470 | (!r2 || !i.cond2 || join_side_has_flag(r2, i.cond2, p2, s2)); |
| 1471 | // zero length pattern => only TESTAFF |
| 1472 | // zero pattern (0/flag) => unmodified stem (zero affixes allowed) |
| 1473 | bool left_text_ok = right_flag_ok && |
| 1474 | (i.pattern.empty() || |
| 1475 | ((i.pattern[0] == '0' && r1->blen <= pos && |
| 1476 | strncmp(word.c_str() + pos - r1->blen, r1->word, r1->blen) == 0) || |
| 1477 | (i.pattern[0] != '0' && |
| 1478 | ((len = i.pattern.size()) != 0) && len <= pos && |
| 1479 | strncmp(word.c_str() + pos - len, i.pattern.c_str(), len) == 0))); |
| 1480 | |
| 1481 | if (t) { |
| 1482 | std::string fields = "left=\"" + i.pattern + "\"/" + trace_cond_flag(this, i.cond) + |
| 1483 | " right=\"" + i.pattern2 + "\"/" + trace_cond_flag(this, i.cond2) + |
| 1484 | " first=" + trace_join_word(this, r1) + |
| 1485 | " second=" + trace_join_word(this, r2); |
| 1486 | if (left_text_ok) |
| 1487 | trace(*t, "test cpdpattern %s -> fail, this pair is forbidden at the join", fields.c_str()); |
| 1488 | else if (!right_text_ok) |
| 1489 | trace(*t, "test cpdpattern %s -> pass, the text after the join does not start with \"%s\"", |
| 1490 | fields.c_str(), i.pattern2.c_str()); |
| 1491 | else if (!left_flag_ok) |
| 1492 | trace(*t, "test cpdpattern %s -> pass, \"%s\" has no %s", fields.c_str(), |
| 1493 | std::string(r1->word, r1->blen).c_str(), trace_flag(this, i.cond).c_str()); |
| 1494 | else if (!right_flag_ok) |
| 1495 | trace(*t, "test cpdpattern %s -> pass, \"%s\" has no %s", fields.c_str(), |
| 1496 | std::string(r2->word, r2->blen).c_str(), trace_flag(this, i.cond2).c_str()); |
| 1497 | else |
| 1498 | trace(*t, "test cpdpattern %s -> pass, the text before the join does not end with \"%s\"", |
| 1499 | fields.c_str(), i.pattern.c_str()); |
| 1500 | } |
| 1501 | |
| 1502 | if (left_text_ok) |
| 1503 | return 1; |
| 1504 | } |
| 1505 | return 0; |
| 1506 | } |
| 1507 | |
| 1508 | // forbid compounding with neighbouring upper and lower case characters at word |
| 1509 | // bounds |
| 1510 | int AffixMgr::cpdcase_check(const std::string& word, int pos) { |
| 1511 | if (utf8) { |
| 1512 | const char* p; |
| 1513 | const char* wordp = word.c_str(); |
| 1514 | for (p = wordp + pos - 1; p > wordp && is_utf8_cont(*p); p--) |
| 1515 | ; |
| 1516 | std::string pair(p); |
| 1517 | std::vector<w_char> pair_u; |
| 1518 | u8_u16(pair_u, pair); |
| 1519 | unsigned short a = pair_u.size() > 1 ? (unsigned short)pair_u[1] : 0, |
| 1520 | b = !pair_u.empty() ? (unsigned short)pair_u[0] : 0; |
| 1521 | if (((unicodetoupper(a, langnum) == a && unicodetolower(a, langnum) != a) || |
| 1522 | (unicodetoupper(b, langnum) == b && unicodetolower(b, langnum) != b)) && |
| 1523 | (a != '-') && (b != '-')) |
| 1524 | return 1; |
| 1525 | } else { |
| 1526 | const unsigned char a = word[pos - 1], b = word[pos]; |
| 1527 | if ((csconv[a].ccase || csconv[b].ccase) && (a != '-') && (b != '-')) |
| 1528 | return 1; |
| 1529 | } |
| 1530 | return 0; |
| 1531 | } |
| 1532 | |
| 1533 | struct metachar_data { |
| 1534 | size_t btpp; // metacharacter (*, ?) position for backtracking |
| 1535 | signed short btwp; // word position for metacharacters |
| 1536 | int btnum; // number of matched characters in metacharacter |
| 1537 | }; |
| 1538 | |
| 1539 | // check compound patterns |
| 1540 | int AffixMgr::defcpd_check(hentry*** words, |
| 1541 | short wnum, |
| 1542 | short maxwordnum, |
| 1543 | hentry* rv, |
| 1544 | hentry** def, |
| 1545 | char all) { |
| 1546 | int w = 0; |
| 1547 | |
| 1548 | if (!*words) { |
| 1549 | w = 1; |
| 1550 | *words = def; |
| 1551 | } |
| 1552 | |
| 1553 | if (!*words) { |
| 1554 | return 0; |
| 1555 | } |
| 1556 | |
| 1557 | if (wnum >= maxwordnum) { |
| 1558 | if (w) |
| 1559 | *words = nullptr; |
| 1560 | return 0; |
| 1561 | } |
| 1562 | |
| 1563 | std::vector<metachar_data> btinfo(1); |
| 1564 | |
| 1565 | short bt = 0; |
| 1566 | |
| 1567 | (*words)[wnum] = rv; |
| 1568 | |
| 1569 | // has the last word COMPOUNDRULE flag? |
| 1570 | if (rv->alen == 0) { |
| 1571 | (*words)[wnum] = nullptr; |
| 1572 | if (w) |
| 1573 | *words = nullptr; |
| 1574 | return 0; |
| 1575 | } |
| 1576 | int ok = 0; |
| 1577 | for (auto& i : defcpdtable) { |
| 1578 | for (auto& j : i) { |
| 1579 | if (j != '*' && j != '?' && |
| 1580 | TESTAFF(rv->astr, j, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, j ))) { |
| 1581 | ok = 1; |
| 1582 | break; |
| 1583 | } |
| 1584 | } |
| 1585 | } |
| 1586 | if (ok == 0) { |
| 1587 | (*words)[wnum] = nullptr; |
| 1588 | if (w) |
| 1589 | *words = nullptr; |
| 1590 | return 0; |
| 1591 | } |
| 1592 | |
| 1593 | for (auto& i : defcpdtable) { |
| 1594 | size_t pp = 0; // pattern position |
| 1595 | signed short wp = 0; // "words" position |
| 1596 | int ok2 = 1; |
| 1597 | ok = 1; |
| 1598 | do { |
| 1599 | while ((pp < i.size()) && (wp <= wnum)) { |
| 1600 | if (((pp + 1) < i.size()) && |
| 1601 | ((i[pp + 1] == '*') || |
| 1602 | (i[pp + 1] == '?'))) { |
| 1603 | int wend = (i[pp + 1] == '?') ? wp : wnum; |
| 1604 | ok2 = 1; |
| 1605 | pp += 2; |
| 1606 | btinfo[bt].btpp = pp; |
| 1607 | btinfo[bt].btwp = wp; |
| 1608 | while (wp <= wend) { |
| 1609 | if (!(*words)[wp] || |
| 1610 | !(*words)[wp]->alen || |
| 1611 | !TESTAFF((*words)[wp]->astr, i[pp - 2],(std::binary_search((*words)[wp]->astr, (*words)[wp]->astr + (*words)[wp]->alen, i[pp - 2])) |
| 1612 | (*words)[wp]->alen)(std::binary_search((*words)[wp]->astr, (*words)[wp]->astr + (*words)[wp]->alen, i[pp - 2]))) { |
| 1613 | ok2 = 0; |
| 1614 | break; |
| 1615 | } |
| 1616 | wp++; |
| 1617 | } |
| 1618 | if (wp <= wnum) |
| 1619 | ok2 = 0; |
| 1620 | btinfo[bt].btnum = wp - btinfo[bt].btwp; |
| 1621 | if (btinfo[bt].btnum > 0) { |
| 1622 | ++bt; |
| 1623 | btinfo.resize(bt+1); |
| 1624 | } |
| 1625 | if (ok2) |
| 1626 | break; |
| 1627 | } else { |
| 1628 | ok2 = 1; |
| 1629 | if (!(*words)[wp] || !(*words)[wp]->alen || |
| 1630 | !TESTAFF((*words)[wp]->astr, i[pp],(std::binary_search((*words)[wp]->astr, (*words)[wp]->astr + (*words)[wp]->alen, i[pp])) |
| 1631 | (*words)[wp]->alen)(std::binary_search((*words)[wp]->astr, (*words)[wp]->astr + (*words)[wp]->alen, i[pp]))) { |
| 1632 | ok = 0; |
| 1633 | break; |
| 1634 | } |
| 1635 | pp++; |
| 1636 | wp++; |
| 1637 | if ((i.size() == pp) && !(wp > wnum)) |
| 1638 | ok = 0; |
| 1639 | } |
| 1640 | } |
| 1641 | if (ok && ok2) { |
| 1642 | size_t r = pp; |
| 1643 | while ((i.size() > r) && ((r + 1) < i.size()) && |
| 1644 | ((i[r + 1] == '*') || |
| 1645 | (i[r + 1] == '?'))) |
| 1646 | r += 2; |
| 1647 | if (i.size() <= r) |
| 1648 | return 1; |
| 1649 | } |
| 1650 | // backtrack |
| 1651 | if (bt) |
| 1652 | do { |
| 1653 | ok = 1; |
| 1654 | btinfo[bt - 1].btnum--; |
| 1655 | pp = btinfo[bt - 1].btpp; |
| 1656 | wp = btinfo[bt - 1].btwp + (signed short)btinfo[bt - 1].btnum; |
| 1657 | } while ((btinfo[bt - 1].btnum < 0) && --bt); |
| 1658 | } while (bt); |
| 1659 | |
| 1660 | if (ok && ok2 && (!all || (i.size() <= pp))) |
| 1661 | return 1; |
| 1662 | |
| 1663 | // check zero ending |
| 1664 | while (ok && ok2 && (i.size() > pp) && |
| 1665 | ((pp + 1) < i.size()) && |
| 1666 | ((i[pp + 1] == '*') || |
| 1667 | (i[pp + 1] == '?'))) |
| 1668 | pp += 2; |
| 1669 | if (ok && ok2 && (i.size() <= pp)) |
| 1670 | return 1; |
| 1671 | } |
| 1672 | (*words)[wnum] = nullptr; |
| 1673 | if (w) |
| 1674 | *words = nullptr; |
| 1675 | return 0; |
| 1676 | } |
| 1677 | |
| 1678 | inline int AffixMgr::candidate_check(const std::string& word, AffixScratch& scratch) { |
| 1679 | |
| 1680 | struct hentry* rv = lookup(word.c_str(), word.size()); |
| 1681 | if (rv) |
| 1682 | return 1; |
| 1683 | |
| 1684 | // rv = prefix_check(word,0,len,1); |
| 1685 | // if (rv) return 1; |
| 1686 | |
| 1687 | rv = affix_check(word, 0, word.size(), scratch); |
| 1688 | if (rv) |
| 1689 | return 1; |
| 1690 | return 0; |
| 1691 | } |
| 1692 | |
| 1693 | // calculate number of syllable for compound-checking |
| 1694 | short AffixMgr::get_syllable(const std::string& word) { |
| 1695 | if (cpdmaxsyllable == 0) |
| 1696 | return 0; |
| 1697 | |
| 1698 | short num = 0; |
| 1699 | |
| 1700 | if (!utf8) { |
| 1701 | num = (short)std::count_if(word.begin(), word.end(), |
| 1702 | [&](char c) { |
| 1703 | return std::binary_search(cpdvowels.begin(), cpdvowels.end(), c); |
| 1704 | }); |
| 1705 | } else if (!cpdvowels_utf16.empty()) { |
| 1706 | std::vector<w_char> w; |
| 1707 | u8_u16(w, word); |
| 1708 | num = (short)std::count_if(w.begin(), w.end(), |
| 1709 | [&](w_char wc) { |
| 1710 | return std::binary_search(cpdvowels_utf16.begin(), cpdvowels_utf16.end(), wc); |
| 1711 | }); |
| 1712 | } |
| 1713 | |
| 1714 | return num; |
| 1715 | } |
| 1716 | |
| 1717 | void AffixMgr::setcminmax(size_t* cmin, size_t* cmax, const char* word, size_t len) { |
| 1718 | if (utf8) { |
| 1719 | int i; |
| 1720 | for (*cmin = 0, i = 0; (i < cpdmin) && *cmin < len; i++) { |
| 1721 | for ((*cmin)++; *cmin < len && is_utf8_cont(word[*cmin]); (*cmin)++) |
| 1722 | ; |
| 1723 | } |
| 1724 | for (*cmax = len, i = 0; (i < (cpdmin - 1)) && *cmax > 0; i++) { |
| 1725 | for ((*cmax)--; *cmax > 0 && is_utf8_cont(word[*cmax]); (*cmax)--) |
| 1726 | ; |
| 1727 | } |
| 1728 | } else { |
| 1729 | *cmin = cpdmin; |
| 1730 | *cmax = len - cpdmin + 1; |
| 1731 | } |
| 1732 | } |
| 1733 | |
| 1734 | // check if compound word is correctly spelled |
| 1735 | // hu_mov_rule = spec. Hungarian rule (XXX) |
| 1736 | struct hentry* AffixMgr::compound_check(const std::string& word, |
| 1737 | short wordnum, |
| 1738 | short numsyllable, |
| 1739 | short maxwordnum, |
| 1740 | short wnum, |
| 1741 | hentry** words, |
| 1742 | hentry** rwords, |
| 1743 | char hu_mov_rule, |
| 1744 | char is_sug, |
| 1745 | int* info, |
| 1746 | AffixScratch& scratch) { |
| 1747 | short oldnumsyllable, oldnumsyllable2, oldwordnum, oldwordnum2; |
| 1748 | hentry *rv = nullptr, *rv_first; |
| 1749 | std::string st; |
| 1750 | char ch = '\0', affixed; |
| 1751 | size_t cmin, cmax; |
| 1752 | int striple = 0, soldi = 0, oldcmin = 0, oldcmax = 0, oldlen = 0, checkedstriple = 0; |
| 1753 | hentry** oldwords = words; |
| 1754 | size_t scpd = 0, len = word.size(); |
| 1755 | |
| 1756 | TraceCtx* t = trace_on(scratch.trace); |
| 1757 | |
| 1758 | // protect subsequent words[wnum + 1] reads and any recursion |
| 1759 | if (wnum + 1 >= maxwordnum) |
| 1760 | return nullptr; |
| 1761 | |
| 1762 | int checked_prefix; |
| 1763 | |
| 1764 | // add a time limit to handle possible |
| 1765 | // combinatorical explosion of the overlapping words |
| 1766 | |
| 1767 | HUNSPELL_THREAD_LOCALthread_local std::chrono::steady_clock::time_point clock_time_start; |
| 1768 | HUNSPELL_THREAD_LOCALthread_local bool timelimit_exceeded; |
| 1769 | |
| 1770 | // get the current time |
| 1771 | std::chrono::steady_clock::time_point clock_now = std::chrono::steady_clock::now(); |
| 1772 | |
| 1773 | if (wnum == 0) { |
| 1774 | // set the start time |
| 1775 | clock_time_start = clock_now; |
| 1776 | timelimit_exceeded = false; |
| 1777 | } |
| 1778 | else if (clock_now - clock_time_start > TIMELIMIT_MSstd::chrono::milliseconds(50)) |
| 1779 | timelimit_exceeded = true; |
| 1780 | |
| 1781 | setcminmax(&cmin, &cmax, word.c_str(), len); |
| 1782 | |
| 1783 | // with SIMPLIFIEDTRIPLE a second part that starts at a doubled letter is one letter longer |
| 1784 | // than the surface shows, so the split can go one letter past cmax |
| 1785 | size_t cmaxtriple = simplifiedtriple ? (utf8 ? utf8_next(word, cmax) : cmax + 1) : cmax; |
| 1786 | |
| 1787 | st.assign(word); |
| 1788 | |
| 1789 | for (size_t i = cmin; i < cmaxtriple; ++i) { |
| 1790 | // go to end of the UTF-8 character |
| 1791 | if (utf8) { |
| 1792 | for (; is_utf8_cont(st[i]); i++) |
| 1793 | ; |
| 1794 | if (i >= cmaxtriple) |
| 1795 | return nullptr; |
| 1796 | } |
| 1797 | if (i >= cmax && !(i > 2 && word[i - 1] == word[i - 2])) |
| 1798 | break; |
| 1799 | |
| 1800 | words = oldwords; |
| 1801 | int onlycpdrule = (words) ? 1 : 0; |
| 1802 | |
| 1803 | do { // onlycpdrule loop |
| 1804 | |
| 1805 | oldnumsyllable = numsyllable; |
| 1806 | oldwordnum = wordnum; |
| 1807 | checked_prefix = 0; |
| 1808 | |
| 1809 | do { // simplified checkcompoundpattern loop |
| 1810 | |
| 1811 | if (timelimit_exceeded || |
| 1812 | std::chrono::steady_clock::now() - clock_time_start > TIMELIMIT_MSstd::chrono::milliseconds(50)) { |
| 1813 | if (t && !timelimit_exceeded) |
| 1814 | trace(*t, "test timelimit -> fail, the compound search stops here and gives up"); |
| 1815 | timelimit_exceeded = true; |
| 1816 | return nullptr; |
| 1817 | } |
| 1818 | |
| 1819 | if (scpd > 0) { |
| 1820 | for (; scpd <= checkcpdtable.size() && |
| 1821 | (checkcpdtable[scpd - 1].pattern3.empty() || |
| 1822 | i > word.size() || |
| 1823 | word.compare(i, checkcpdtable[scpd - 1].pattern3.size(), checkcpdtable[scpd - 1].pattern3) != 0); |
| 1824 | scpd++) |
| 1825 | ; |
| 1826 | |
| 1827 | if (scpd > checkcpdtable.size()) |
| 1828 | break; // break simplified checkcompoundpattern loop |
| 1829 | st.replace(i, std::string::npos, checkcpdtable[scpd - 1].pattern); |
| 1830 | soldi = i; |
| 1831 | i += checkcpdtable[scpd - 1].pattern.size(); |
| 1832 | st.replace(i, std::string::npos, checkcpdtable[scpd - 1].pattern2); |
| 1833 | st.replace(i + checkcpdtable[scpd - 1].pattern2.size(), std::string::npos, |
| 1834 | word.substr(soldi + checkcpdtable[scpd - 1].pattern3.size())); |
| 1835 | |
| 1836 | oldlen = len; |
| 1837 | len += checkcpdtable[scpd - 1].pattern.size() + |
| 1838 | checkcpdtable[scpd - 1].pattern2.size() - |
| 1839 | checkcpdtable[scpd - 1].pattern3.size(); |
| 1840 | oldcmin = cmin; |
| 1841 | oldcmax = cmax; |
| 1842 | setcminmax(&cmin, &cmax, st.c_str(), len); |
| 1843 | |
| 1844 | cmax = len - cpdmin + 1; |
| 1845 | } |
| 1846 | |
| 1847 | if (i >= st.size()) |
| 1848 | return nullptr; |
| 1849 | |
| 1850 | ch = st[i]; |
| 1851 | if (t) { |
| 1852 | // a non-zero scpd means this split is the retry that one CHECKCOMPOUNDPATTERN entry asks |
| 1853 | // for, with the text at the join replaced by the two sides of that entry |
| 1854 | if (scpd != 0) |
| 1855 | trace(*t, "split at=%d left=\"%s\" right=\"%s\" cpdpattern=%d", (int)i, |
| 1856 | st.substr(0, i).c_str(), st.substr(i).c_str(), (int)scpd); |
| 1857 | else |
| 1858 | trace(*t, "split at=%d left=\"%s\" right=\"%s\"", (int)i, |
| 1859 | st.substr(0, i).c_str(), st.substr(i).c_str()); |
| 1860 | } |
| 1861 | // everything this split point tries belongs to the split |
| 1862 | TraceScope split_depth(t); |
| 1863 | st[i] = '\0'; |
| 1864 | |
| 1865 | sfx = nullptr; |
| 1866 | pfx = nullptr; |
| 1867 | |
| 1868 | // FIRST WORD |
| 1869 | |
| 1870 | affixed = 1; |
| 1871 | rv = lookup(st.c_str(), i); // perhaps without prefix |
| 1872 | if (t) { |
| 1873 | if (rv) |
| 1874 | trace(*t, "first \"%s\" -> entry \"%s\" flags=%s", st.c_str(), |
| 1875 | rv->word, trace_flags(this, rv->astr, rv->alen).c_str()); |
| 1876 | else |
| 1877 | trace(*t, "first \"%s\" -> miss", st.c_str()); |
| 1878 | } |
| 1879 | |
| 1880 | // forbid dictionary stems with COMPOUNDFORBIDFLAG in |
| 1881 | // compound words, overriding the effect of COMPOUNDPERMITFLAG |
| 1882 | if ((rv) && compoundforbidflag && |
| 1883 | TESTAFF(rv->astr, compoundforbidflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundforbidflag )) && !hu_mov_rule) { |
| 1884 | bool would_continue = !onlycpdrule && simplifiedcpd; |
| 1885 | if (!scpd && would_continue) { |
| 1886 | // given the while conditions that continue jumps to, this situation |
| 1887 | // never ends |
| 1888 | HUNSPELL_WARNING(stderrstderr, "break infinite loop\n"); |
| 1889 | break; |
| 1890 | } |
| 1891 | |
| 1892 | if (scpd > 0 && would_continue) { |
| 1893 | // under these conditions we loop again, but the assumption above |
| 1894 | // appears to be that cmin and cmax are the original values they |
| 1895 | // had in the outside loop |
| 1896 | cmin = oldcmin; |
| 1897 | cmax = oldcmax; |
| 1898 | } |
| 1899 | continue; |
| 1900 | } |
| 1901 | |
| 1902 | // search homonym with compound flag |
| 1903 | while ((rv) && !hu_mov_rule && |
| 1904 | ((needaffix && TESTAFF(rv->astr, needaffix, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, needaffix ))) || |
| 1905 | !((compoundflag && !words && !onlycpdrule && |
| 1906 | TESTAFF(rv->astr, compoundflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundflag ))) || |
| 1907 | (compoundbegin && !wordnum && !onlycpdrule && |
| 1908 | TESTAFF(rv->astr, compoundbegin, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundbegin ))) || |
| 1909 | (compoundmiddle && wordnum && !words && !onlycpdrule && |
| 1910 | TESTAFF(rv->astr, compoundmiddle, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundmiddle ))) || |
| 1911 | (!defcpdtable.empty() && onlycpdrule && |
| 1912 | ((!words && !wordnum && |
| 1913 | defcpd_check(&words, wnum, maxwordnum, rv, rwords, 0)) || |
| 1914 | (words && |
| 1915 | defcpd_check(&words, wnum, maxwordnum, rv, rwords, 0))))) || |
| 1916 | (scpd != 0 && checkcpdtable[scpd - 1].cond != FLAG_NULL0x00 && |
| 1917 | !TESTAFF(rv->astr, checkcpdtable[scpd - 1].cond, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, checkcpdtable [scpd - 1].cond))))) { |
| 1918 | rv = rv->next_homonym; |
| 1919 | } |
| 1920 | |
| 1921 | if (rv) |
| 1922 | affixed = 0; |
| 1923 | |
| 1924 | if (!rv) { |
| 1925 | if (onlycpdrule) |
| 1926 | break; |
| 1927 | if (compoundflag && |
| 1928 | !(rv = prefix_check(st, 0, i, |
| 1929 | hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1, |
| 1930 | scratch, compoundflag))) { |
| 1931 | if (((rv = suffix_check(st, 0, i, 0, nullptr, scratch, FLAG_NULL0x00, compoundflag, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1)) || |
| 1932 | (compoundmoresuffixes && (rv = suffix_check_twosfx(st, 0, i, 0, nullptr, scratch, compoundflag)))) && |
| 1933 | !hu_mov_rule && sfx->getCont() && |
| 1934 | ((compoundforbidflag && TESTAFF(sfx->getCont(), compoundforbidflag, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundforbidflag))) || |
| 1935 | (compoundend && TESTAFF(sfx->getCont(), compoundend, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundend))))) { |
| 1936 | rv = nullptr; |
| 1937 | // the suffix is dropped with the word it built |
| 1938 | sfx = nullptr; |
| 1939 | } |
| 1940 | } |
| 1941 | |
| 1942 | if (rv || |
| 1943 | (((wordnum == 0) && compoundbegin && |
| 1944 | ((rv = suffix_check(st, 0, i, 0, nullptr, scratch, FLAG_NULL0x00, compoundbegin, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1)) || |
| 1945 | (compoundmoresuffixes && (rv = suffix_check_twosfx(st, 0, i, 0, nullptr, scratch, |
| 1946 | compoundbegin))) || // twofold suffixes + compound |
| 1947 | (rv = prefix_check(st, 0, i, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1, scratch, compoundbegin)))) || |
| 1948 | ((wordnum > 0) && compoundmiddle && |
| 1949 | ((rv = suffix_check(st, 0, i, 0, nullptr, scratch, FLAG_NULL0x00, compoundmiddle, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1)) || |
| 1950 | (compoundmoresuffixes && (rv = suffix_check_twosfx(st, 0, i, 0, nullptr, scratch, |
| 1951 | compoundmiddle))) || // twofold suffixes + compound |
| 1952 | (rv = prefix_check(st, 0, i, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1, scratch, compoundmiddle)))))) |
| 1953 | checked_prefix = 1; |
| 1954 | // else check forbiddenwords and needaffix |
| 1955 | } else if (rv->astr && (TESTAFF(rv->astr, forbiddenword, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forbiddenword )) || |
| 1956 | TESTAFF(rv->astr, needaffix, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, needaffix )) || |
| 1957 | TESTAFF(rv->astr, ONLYUPCASEFLAG, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 65511 )) || |
| 1958 | (is_sug && nosuggest && |
| 1959 | TESTAFF(rv->astr, nosuggest, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, nosuggest ))))) { |
| 1960 | st[i] = ch; |
| 1961 | // continue; |
| 1962 | break; |
| 1963 | } |
| 1964 | |
| 1965 | // check non_compound flag in suffix and prefix |
| 1966 | if ((rv) && !hu_mov_rule && |
| 1967 | ((pfx && pfx->getCont() && |
| 1968 | TESTAFF(pfx->getCont(), compoundforbidflag, pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), compoundforbidflag))) || |
| 1969 | (sfx && sfx->getCont() && |
| 1970 | TESTAFF(sfx->getCont(), compoundforbidflag,(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundforbidflag)) |
| 1971 | sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundforbidflag))))) { |
| 1972 | rv = nullptr; |
| 1973 | } |
| 1974 | |
| 1975 | // check compoundend flag in suffix and prefix |
| 1976 | if ((rv) && !checked_prefix && compoundend && !hu_mov_rule && |
| 1977 | ((pfx && pfx->getCont() && |
| 1978 | TESTAFF(pfx->getCont(), compoundend, pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), compoundend))) || |
| 1979 | (sfx && sfx->getCont() && |
| 1980 | TESTAFF(sfx->getCont(), compoundend, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundend))))) { |
| 1981 | rv = nullptr; |
| 1982 | } |
| 1983 | |
| 1984 | // check compoundmiddle flag in suffix and prefix |
| 1985 | if ((rv) && !checked_prefix && (wordnum == 0) && compoundmiddle && |
| 1986 | !hu_mov_rule && |
| 1987 | ((pfx && pfx->getCont() && |
| 1988 | TESTAFF(pfx->getCont(), compoundmiddle, pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), compoundmiddle))) || |
| 1989 | (sfx && sfx->getCont() && |
| 1990 | TESTAFF(sfx->getCont(), compoundmiddle, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundmiddle))))) { |
| 1991 | rv = nullptr; |
| 1992 | } |
| 1993 | |
| 1994 | // check forbiddenwords |
| 1995 | if ((rv) && (rv->astr) && |
| 1996 | (TESTAFF(rv->astr, forbiddenword, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forbiddenword )) || |
| 1997 | TESTAFF(rv->astr, ONLYUPCASEFLAG, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 65511 )) || |
| 1998 | (is_sug && nosuggest && TESTAFF(rv->astr, nosuggest, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, nosuggest ))))) { |
| 1999 | return nullptr; |
| 2000 | } |
| 2001 | |
| 2002 | // increment word number, if the second root has a compoundroot flag |
| 2003 | if ((rv) && compoundroot && |
| 2004 | (TESTAFF(rv->astr, compoundroot, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundroot )))) { |
| 2005 | wordnum++; |
| 2006 | } |
| 2007 | |
| 2008 | // first word is acceptable in compound words? |
| 2009 | if (((rv) && |
| 2010 | (checked_prefix || (words && words[wnum]) || |
| 2011 | (compoundflag && TESTAFF(rv->astr, compoundflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundflag ))) || |
| 2012 | ((oldwordnum == 0) && compoundbegin && |
| 2013 | TESTAFF(rv->astr, compoundbegin, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundbegin ))) || |
| 2014 | ((oldwordnum > 0) && compoundmiddle && |
| 2015 | TESTAFF(rv->astr, compoundmiddle, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundmiddle ))) |
| 2016 | |
| 2017 | // LANG_hu section: spec. Hungarian rule |
| 2018 | || ((langnum == LANG_hu) && hu_mov_rule && |
| 2019 | (TESTAFF((std::binary_search(rv->astr, rv->astr + rv->alen, 'F' )) |
| 2020 | rv->astr, 'F',(std::binary_search(rv->astr, rv->astr + rv->alen, 'F' )) |
| 2021 | rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'F' )) || // XXX hardwired Hungarian dictionary codes |
| 2022 | TESTAFF(rv->astr, 'G', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'G' )) || |
| 2023 | TESTAFF(rv->astr, 'H', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'H' )))) |
| 2024 | // END of LANG_hu section |
| 2025 | ) && |
| 2026 | ( |
| 2027 | // test CHECKCOMPOUNDPATTERN conditions |
| 2028 | scpd == 0 || checkcpdtable[scpd - 1].cond == FLAG_NULL0x00 || |
| 2029 | join_side_has_flag(rv, checkcpdtable[scpd - 1].cond, pfx, sfx)) && |
| 2030 | !((checkcompoundtriple && scpd == 0 && |
| 2031 | !words && i < word.size() && // test triple letters |
| 2032 | (word[i - 1] == word[i]) && |
| 2033 | (((i > 1) && (word[i - 1] == word[i - 2])) || |
| 2034 | ((word[i - 1] == word[i + 1])) // may be word[i+1] == '\0' |
| 2035 | )) || |
| 2036 | (checkcompoundcase && scpd == 0 && !words && i < word.size() && |
| 2037 | cpdcase_check(word, i)))) |
| 2038 | // LANG_hu section: spec. Hungarian rule |
| 2039 | || ((!rv) && (langnum == LANG_hu) && hu_mov_rule && |
| 2040 | (rv = affix_check(st, 0, i, scratch)) && |
| 2041 | (sfx && sfx->getCont() && |
| 2042 | ( // XXX hardwired Hungarian dic. codes |
| 2043 | TESTAFF(sfx->getCont(), (unsigned short)'x',(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), (unsigned short)'x')) |
| 2044 | sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), (unsigned short)'x')) || |
| 2045 | TESTAFF((std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), (unsigned short)'%')) |
| 2046 | sfx->getCont(), (unsigned short)'%',(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), (unsigned short)'%')) |
| 2047 | sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), (unsigned short)'%')))))) { // first word is ok condition |
| 2048 | |
| 2049 | // LANG_hu section: spec. Hungarian rule |
| 2050 | if (langnum == LANG_hu) { |
| 2051 | // calculate syllable number of the word |
| 2052 | numsyllable += get_syllable(st.substr(0, i)); |
| 2053 | // + 1 word, if syllable number of the prefix > 1 (hungarian |
| 2054 | // convention) |
| 2055 | if (pfx && (get_syllable(pfx->getKey()) > 1)) |
| 2056 | wordnum++; |
| 2057 | } |
| 2058 | // END of LANG_hu section |
| 2059 | |
| 2060 | // NEXT WORD(S) |
| 2061 | rv_first = rv; |
| 2062 | // the affixes that built the first word, so its continuation classes stay reachable once |
| 2063 | // the second word has overwritten pfx and sfx |
| 2064 | PfxEntry* rv_first_pfx = pfx; |
| 2065 | SfxEntry* rv_first_sfx = sfx; |
| 2066 | st[i] = ch; |
| 2067 | |
| 2068 | do { // striple loop |
| 2069 | |
| 2070 | // check simplifiedtriple |
| 2071 | if (simplifiedtriple) { |
| 2072 | if (striple) { |
| 2073 | checkedstriple = 1; |
| 2074 | i--; // check "fahrt" instead of "ahrt" in "Schiffahrt" |
| 2075 | } else if (i > 2 && i <= word.size() && word[i - 1] == word[i - 2]) { |
| 2076 | striple = 1; |
| 2077 | // past cmax the surface split is too short, so only the form with the letter |
| 2078 | // put back is tried |
| 2079 | if (i >= cmax) { |
| 2080 | checkedstriple = 1; |
| 2081 | i--; |
| 2082 | } |
| 2083 | } |
| 2084 | } |
| 2085 | |
| 2086 | rv = lookup(st.c_str() + i, st.size() - i); // perhaps without prefix |
| 2087 | if (t) { |
| 2088 | if (rv) |
| 2089 | trace(*t, "second \"%s\" -> entry \"%s\" flags=%s", st.c_str() + i, |
| 2090 | rv->word, trace_flags(this, rv->astr, rv->alen).c_str()); |
| 2091 | else |
| 2092 | trace(*t, "second \"%s\" -> miss", st.c_str() + i); |
| 2093 | } |
| 2094 | |
| 2095 | // search homonym with compound flag |
| 2096 | while ((rv) && ((needaffix && TESTAFF(rv->astr, needaffix, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, needaffix ))) || |
| 2097 | !((compoundflag && !words && TESTAFF(rv->astr, compoundflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundflag ))) || |
| 2098 | (compoundend && !words && TESTAFF(rv->astr, compoundend, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundend ))) || |
| 2099 | (!defcpdtable.empty() && words && defcpd_check(&words, wnum + 1, maxwordnum, rv, nullptr, 1))) || |
| 2100 | (scpd != 0 && checkcpdtable[scpd - 1].cond2 != FLAG_NULL0x00 && |
| 2101 | !TESTAFF(rv->astr, checkcpdtable[scpd - 1].cond2, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, checkcpdtable [scpd - 1].cond2))))) { |
| 2102 | rv = rv->next_homonym; |
| 2103 | } |
| 2104 | |
| 2105 | // check FORCEUCASE |
| 2106 | if (rv && forceucase && |
| 2107 | (TESTAFF(rv->astr, forceucase, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forceucase ))) && |
| 2108 | !(info && *info & SPELL_ORIGCAP(1 << 5))) |
| 2109 | rv = nullptr; |
| 2110 | |
| 2111 | if (rv && words && words[wnum + 1]) |
| 2112 | return rv_first; |
| 2113 | |
| 2114 | oldnumsyllable2 = numsyllable; |
| 2115 | oldwordnum2 = wordnum; |
| 2116 | |
| 2117 | // LANG_hu section: spec. Hungarian rule, XXX hardwired dictionary |
| 2118 | // code |
| 2119 | if ((rv) && (langnum == LANG_hu) && |
| 2120 | (TESTAFF(rv->astr, 'I', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'I' ))) && |
| 2121 | !(TESTAFF(rv->astr, 'J', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'J' )))) { |
| 2122 | numsyllable--; |
| 2123 | } |
| 2124 | // END of LANG_hu section |
| 2125 | |
| 2126 | // increment word number, if the second root has a compoundroot flag |
| 2127 | if ((rv) && (compoundroot) && |
| 2128 | (TESTAFF(rv->astr, compoundroot, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundroot )))) { |
| 2129 | wordnum++; |
| 2130 | } |
| 2131 | |
| 2132 | // check forbiddenwords |
| 2133 | if ((rv) && (rv->astr) && |
| 2134 | (TESTAFF(rv->astr, forbiddenword, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forbiddenword )) || |
| 2135 | TESTAFF(rv->astr, ONLYUPCASEFLAG, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 65511 )) || |
| 2136 | (is_sug && nosuggest && |
| 2137 | TESTAFF(rv->astr, nosuggest, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, nosuggest ))))) |
| 2138 | return nullptr; |
| 2139 | |
| 2140 | // second word is acceptable, as a root? |
| 2141 | // hungarian conventions: compounding is acceptable, |
| 2142 | // when compound forms consist of 2 words, or if more, |
| 2143 | // then the syllable number of root words must be 6, or lesser. |
| 2144 | |
| 2145 | if ((rv) && |
| 2146 | ((compoundflag && TESTAFF(rv->astr, compoundflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundflag ))) || |
| 2147 | (compoundend && TESTAFF(rv->astr, compoundend, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundend )))) && |
| 2148 | (((cpdwordmax == -1) || (wordnum + 1 < cpdwordmax)) || |
| 2149 | ((cpdmaxsyllable != 0) && |
| 2150 | (numsyllable + get_syllable(std::string(HENTRY_WORD(rv)&(rv->word[0]), rv->blen)) <= |
| 2151 | cpdmaxsyllable))) && |
| 2152 | ( |
| 2153 | // test CHECKCOMPOUNDPATTERN |
| 2154 | checkcpdtable.empty() || scpd != 0 || |
| 2155 | (i < word.size() && !cpdpat_check(word, i, rv_first, rv, 0, t, rv_first_pfx, rv_first_sfx, nullptr, nullptr))) && |
| 2156 | ((!checkcompounddup || (rv != rv_first))) |
| 2157 | // test CHECKCOMPOUNDPATTERN conditions |
| 2158 | && |
| 2159 | (scpd == 0 || checkcpdtable[scpd - 1].cond2 == FLAG_NULL0x00 || |
| 2160 | TESTAFF(rv->astr, checkcpdtable[scpd - 1].cond2, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, checkcpdtable [scpd - 1].cond2)))) { |
| 2161 | // forbid compound word, if it is a non-compound word with typical |
| 2162 | // fault |
| 2163 | if ((checkcompoundrep && cpdrep_check(word, len, scratch, timelimit_exceeded, clock_time_start)) || |
| 2164 | cpdwordpair_check(word, len, scratch, timelimit_exceeded, clock_time_start)) |
| 2165 | return nullptr; |
| 2166 | return rv_first; |
| 2167 | } |
| 2168 | |
| 2169 | numsyllable = oldnumsyllable2; |
| 2170 | wordnum = oldwordnum2; |
| 2171 | |
| 2172 | // perhaps second word has prefix or/and suffix |
| 2173 | sfx = nullptr; |
| 2174 | sfxflag = FLAG_NULL0x00; |
| 2175 | // the affixes that built the second word, kept because affix_check clears the members |
| 2176 | PfxEntry* rv_second_pfx = nullptr; |
| 2177 | SfxEntry* rv_second_sfx = nullptr; |
| 2178 | rv = (compoundflag && !onlycpdrule && i < word.size()) ? affix_check(word, i, word.size() - i, scratch, compoundflag, IN_CPD_END2, FLAG_NULL0x00, &rv_second_pfx, &rv_second_sfx) |
| 2179 | : nullptr; |
| 2180 | if (!rv && compoundend && !onlycpdrule) { |
| 2181 | sfx = nullptr; |
| 2182 | pfx = nullptr; |
| 2183 | if (i < word.size()) |
| 2184 | rv = affix_check(word, i, word.size() - i, scratch, compoundend, IN_CPD_END2, FLAG_NULL0x00, &rv_second_pfx, &rv_second_sfx); |
| 2185 | } |
| 2186 | |
| 2187 | if (!rv && !defcpdtable.empty() && words) { |
| 2188 | if (i < word.size()) |
| 2189 | rv = affix_check(word, i, word.size() - i, scratch, 0, IN_CPD_END2); |
| 2190 | if (rv && defcpd_check(&words, wnum + 1, maxwordnum, rv, nullptr, 1)) |
| 2191 | return rv_first; |
| 2192 | rv = nullptr; |
| 2193 | } |
| 2194 | |
| 2195 | // test CHECKCOMPOUNDPATTERN conditions (allowed forms) |
| 2196 | if (rv && |
| 2197 | !(scpd == 0 || checkcpdtable[scpd - 1].cond2 == FLAG_NULL0x00 || |
| 2198 | join_side_has_flag(rv, checkcpdtable[scpd - 1].cond2, rv_second_pfx, |
| 2199 | rv_second_sfx))) |
| 2200 | rv = nullptr; |
| 2201 | |
| 2202 | // test CHECKCOMPOUNDPATTERN conditions (forbidden compounds) |
| 2203 | if (rv && !checkcpdtable.empty() && scpd == 0 && |
| 2204 | cpdpat_check(word, i, rv_first, rv, affixed, t, rv_first_pfx, rv_first_sfx, rv_second_pfx, rv_second_sfx)) |
| 2205 | rv = nullptr; |
| 2206 | |
| 2207 | // check non_compound flag in suffix and prefix |
| 2208 | if ((rv) && ((pfx && pfx->getCont() && |
| 2209 | TESTAFF(pfx->getCont(), compoundforbidflag,(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), compoundforbidflag)) |
| 2210 | pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), compoundforbidflag))) || |
| 2211 | (sfx && sfx->getCont() && |
| 2212 | TESTAFF(sfx->getCont(), compoundforbidflag,(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundforbidflag)) |
| 2213 | sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundforbidflag))))) { |
| 2214 | rv = nullptr; |
| 2215 | } |
| 2216 | |
| 2217 | // check FORCEUCASE |
| 2218 | if (rv && forceucase && |
| 2219 | (TESTAFF(rv->astr, forceucase, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forceucase ))) && |
| 2220 | !(info && *info & SPELL_ORIGCAP(1 << 5))) |
| 2221 | rv = nullptr; |
| 2222 | |
| 2223 | // check forbiddenwords |
| 2224 | if ((rv) && (rv->astr) && |
| 2225 | (TESTAFF(rv->astr, forbiddenword, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forbiddenword )) || |
| 2226 | TESTAFF(rv->astr, ONLYUPCASEFLAG, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 65511 )) || |
| 2227 | (is_sug && nosuggest && |
| 2228 | TESTAFF(rv->astr, nosuggest, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, nosuggest ))))) |
| 2229 | return nullptr; |
| 2230 | |
| 2231 | // pfxappnd = prefix of word+i, or NULL |
| 2232 | // calculate syllable number of prefix. |
| 2233 | // hungarian convention: when syllable number of prefix is more, |
| 2234 | // than 1, the prefix+word counts as two words. |
| 2235 | |
| 2236 | if (langnum == LANG_hu) { |
| 2237 | if (i < word.size()) { |
| 2238 | // calculate syllable number of the word |
| 2239 | numsyllable += get_syllable(word.substr(i)); |
| 2240 | } |
| 2241 | |
| 2242 | // - affix syllable num. |
| 2243 | // XXX only second suffix (inflections, not derivations) |
| 2244 | if (sfxappnd) { |
| 2245 | std::string tmp(sfxappnd); |
| 2246 | reverseword(tmp); |
| 2247 | numsyllable -= short(get_syllable(tmp) + sfxextra); |
| 2248 | } else { |
| 2249 | numsyllable -= short(sfxextra); |
| 2250 | } |
| 2251 | |
| 2252 | // + 1 word, if syllable number of the prefix > 1 (hungarian |
| 2253 | // convention) |
| 2254 | if (pfx && (get_syllable(pfx->getKey()) > 1)) |
| 2255 | wordnum++; |
| 2256 | |
| 2257 | // increment syllable num, if last word has a SYLLABLENUM flag |
| 2258 | // and the suffix is beginning `s' |
| 2259 | |
| 2260 | if (!cpdsyllablenum.empty()) { |
| 2261 | switch (sfxflag) { |
| 2262 | case 'c': { |
| 2263 | numsyllable += 2; |
| 2264 | break; |
| 2265 | } |
| 2266 | case 'J': { |
| 2267 | numsyllable += 1; |
| 2268 | break; |
| 2269 | } |
| 2270 | case 'I': { |
| 2271 | if (rv && TESTAFF(rv->astr, 'J', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'J' ))) |
| 2272 | numsyllable += 1; |
| 2273 | break; |
| 2274 | } |
| 2275 | } |
| 2276 | } |
| 2277 | } |
| 2278 | |
| 2279 | // increment word number, if the second word has a compoundroot flag |
| 2280 | if ((rv) && (compoundroot) && |
| 2281 | (TESTAFF(rv->astr, compoundroot, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundroot )))) { |
| 2282 | wordnum++; |
| 2283 | } |
| 2284 | // second word is acceptable, as a word with prefix or/and suffix? |
| 2285 | // hungarian conventions: compounding is acceptable, |
| 2286 | // when compound forms consist 2 word, otherwise |
| 2287 | // the syllable number of root words is 6, or lesser. |
| 2288 | if ((rv) && |
| 2289 | (((cpdwordmax == -1) || (wordnum + 1 < cpdwordmax)) || |
| 2290 | ((cpdmaxsyllable != 0) && (numsyllable <= cpdmaxsyllable))) && |
| 2291 | ((!checkcompounddup || (rv != rv_first)))) { |
| 2292 | // forbid compound word, if it is a non-compound word with typical |
| 2293 | // fault |
| 2294 | if ((checkcompoundrep && cpdrep_check(word, len, scratch, timelimit_exceeded, clock_time_start)) || |
| 2295 | cpdwordpair_check(word, len, scratch, timelimit_exceeded, clock_time_start)) |
| 2296 | return nullptr; |
| 2297 | return rv_first; |
| 2298 | } |
| 2299 | |
| 2300 | numsyllable = oldnumsyllable2; |
| 2301 | wordnum = oldwordnum2; |
| 2302 | |
| 2303 | // perhaps second word is a compound word (recursive call) |
| 2304 | // (only if SPELL_COMPOUND_2 is not set and maxwordnum is not exceeded) |
| 2305 | if ((!info || !(*info & SPELL_COMPOUND_2(1 << 7))) && wordnum + 2 < maxwordnum && wnum + 1 < maxwordnum) { |
| 2306 | rv = compound_check(st.substr(i), wordnum + 1, |
| 2307 | numsyllable, maxwordnum, wnum + 1, words, rwords, 0, |
| 2308 | is_sug, info, scratch); |
| 2309 | |
| 2310 | if (rv && !checkcpdtable.empty() && i < word.size() && |
| 2311 | ((scpd == 0 && |
| 2312 | cpdpat_check(word, i, rv_first, rv, affixed, t, rv_first_pfx, rv_first_sfx, nullptr, nullptr)) || |
| 2313 | (scpd != 0 && |
| 2314 | !cpdpat_check(word, i, rv_first, rv, affixed, t, rv_first_pfx, rv_first_sfx, nullptr, nullptr)))) |
| 2315 | rv = nullptr; |
| 2316 | } else { |
| 2317 | rv = nullptr; |
| 2318 | } |
| 2319 | if (rv) { |
| 2320 | // forbid compound word, if it is a non-compound word with typical |
| 2321 | // fault, or a dictionary word pair |
| 2322 | |
| 2323 | if (cpdwordpair_check(word, len, scratch, timelimit_exceeded, clock_time_start)) |
| 2324 | return nullptr; |
| 2325 | |
| 2326 | if (checkcompoundrep || forbiddenword) { |
| 2327 | if (checkcompoundrep && cpdrep_check(word, len, scratch, timelimit_exceeded, clock_time_start)) |
| 2328 | return nullptr; |
| 2329 | |
| 2330 | // check first part |
| 2331 | if (i < word.size() && word.compare(i, rv->blen, rv->word, rv->blen) == 0) { |
| 2332 | char r = st[i + rv->blen]; |
| 2333 | st[i + rv->blen] = '\0'; |
| 2334 | |
| 2335 | if ((checkcompoundrep && cpdrep_check(st, i + rv->blen, scratch, timelimit_exceeded, clock_time_start)) || |
| 2336 | cpdwordpair_check(st, i + rv->blen, scratch, timelimit_exceeded, clock_time_start)) { |
| 2337 | st[ + i + rv->blen] = r; |
| 2338 | continue; |
| 2339 | } |
| 2340 | |
| 2341 | if (forbiddenword) { |
| 2342 | struct hentry* rv2 = lookup(word.c_str(), word.size()); |
| 2343 | if (!rv2 && len <= word.size()) |
| 2344 | rv2 = affix_check(word, 0, len, scratch); |
| 2345 | if (rv2 && rv2->astr && |
| 2346 | TESTAFF(rv2->astr, forbiddenword, rv2->alen)(std::binary_search(rv2->astr, rv2->astr + rv2->alen , forbiddenword)) && |
| 2347 | (strncmp(rv2->word, st.c_str(), i + rv->blen) == 0)) { |
| 2348 | return nullptr; |
| 2349 | } |
| 2350 | } |
| 2351 | st[i + rv->blen] = r; |
| 2352 | } |
| 2353 | } |
| 2354 | return rv_first; |
| 2355 | } |
| 2356 | } while (striple && !checkedstriple); // end of striple loop |
| 2357 | |
| 2358 | if (checkedstriple) { |
| 2359 | i++; |
| 2360 | checkedstriple = 0; |
| 2361 | striple = 0; |
| 2362 | } |
| 2363 | |
| 2364 | } // first word is ok condition |
| 2365 | |
| 2366 | if (soldi != 0) { |
| 2367 | i = soldi; |
| 2368 | len = oldlen; |
| 2369 | cmin = oldcmin; |
| 2370 | cmax = oldcmax; |
| 2371 | } |
| 2372 | scpd++; |
| 2373 | |
| 2374 | } while (!onlycpdrule && simplifiedcpd && |
| 2375 | scpd <= checkcpdtable.size()); // end of simplifiedcpd loop |
| 2376 | |
| 2377 | scpd = 0; |
| 2378 | wordnum = oldwordnum; |
| 2379 | numsyllable = oldnumsyllable; |
| 2380 | |
| 2381 | if (soldi != 0) { |
| 2382 | i = soldi; |
| 2383 | st.assign(word); // XXX add more optim. |
| 2384 | soldi = 0; |
| 2385 | len = oldlen; |
| 2386 | cmin = oldcmin; |
| 2387 | cmax = oldcmax; |
| 2388 | } else |
| 2389 | st[i] = ch; |
| 2390 | |
| 2391 | } while (!defcpdtable.empty() && oldwordnum == 0 && |
| 2392 | onlycpdrule++ < 1); // end of onlycpd loop |
| 2393 | } |
| 2394 | |
| 2395 | return nullptr; |
| 2396 | } |
| 2397 | |
| 2398 | // check if compound word is correctly spelled |
| 2399 | // hu_mov_rule = spec. Hungarian rule (XXX) |
| 2400 | int AffixMgr::compound_check_morph(const std::string& word, |
| 2401 | short wordnum, |
| 2402 | short numsyllable, |
| 2403 | short maxwordnum, |
| 2404 | short wnum, |
| 2405 | hentry** words, |
| 2406 | hentry** rwords, |
| 2407 | char hu_mov_rule, |
| 2408 | std::string& result, |
| 2409 | const std::string* partresult, |
| 2410 | AffixScratch& scratch) { |
| 2411 | short oldnumsyllable, oldnumsyllable2, oldwordnum, oldwordnum2; |
| 2412 | hentry *rv = nullptr, *rv_first; |
| 2413 | std::string st, presult; |
| 2414 | char ch, affixed = 0; |
| 2415 | int checked_prefix, ok = 0; |
| 2416 | size_t cmin, cmax; |
| 2417 | hentry** oldwords = words; |
| 2418 | size_t len = word.size(); |
| 2419 | |
| 2420 | // protect subsequent words[wnum + 1] reads and any recursion |
| 2421 | if (wnum + 1 >= maxwordnum) |
| 2422 | return 0; |
| 2423 | |
| 2424 | // add a time limit to handle possible |
| 2425 | // combinatorical explosion of the overlapping words |
| 2426 | |
| 2427 | HUNSPELL_THREAD_LOCALthread_local std::chrono::steady_clock::time_point clock_time_start; |
| 2428 | HUNSPELL_THREAD_LOCALthread_local bool timelimit_exceeded; |
| 2429 | |
| 2430 | // get the current time |
| 2431 | std::chrono::steady_clock::time_point clock_now = std::chrono::steady_clock::now(); |
| 2432 | |
| 2433 | if (wnum == 0) { |
| 2434 | // set the start time |
| 2435 | clock_time_start = clock_now; |
| 2436 | timelimit_exceeded = false; |
| 2437 | } |
| 2438 | else if (clock_now - clock_time_start > TIMELIMIT_MSstd::chrono::milliseconds(50)) |
| 2439 | timelimit_exceeded = true; |
| 2440 | |
| 2441 | setcminmax(&cmin, &cmax, word.c_str(), len); |
| 2442 | |
| 2443 | st.assign(word); |
| 2444 | |
| 2445 | for (size_t i = cmin; i < cmax; ++i) { |
| 2446 | // go to end of the UTF-8 character |
| 2447 | if (utf8) { |
| 2448 | for (; is_utf8_cont(st[i]); i++) |
| 2449 | ; |
| 2450 | if (i >= cmax) |
| 2451 | return 0; |
| 2452 | } |
| 2453 | |
| 2454 | words = oldwords; |
| 2455 | int onlycpdrule = (words) ? 1 : 0; |
| 2456 | |
| 2457 | do { // onlycpdrule loop |
| 2458 | |
| 2459 | if (timelimit_exceeded || |
| 2460 | std::chrono::steady_clock::now() - clock_time_start > TIMELIMIT_MSstd::chrono::milliseconds(50)) { |
| 2461 | timelimit_exceeded = true; |
| 2462 | return 0; |
| 2463 | } |
| 2464 | |
| 2465 | if (result.size() > MAXMORPHRESULT(4 * 1024 * 1024)) |
| 2466 | return 0; |
| 2467 | |
| 2468 | oldnumsyllable = numsyllable; |
| 2469 | oldwordnum = wordnum; |
| 2470 | checked_prefix = 0; |
| 2471 | |
| 2472 | if (i >= st.size()) |
| 2473 | return 0; |
| 2474 | |
| 2475 | ch = st[i]; |
| 2476 | st[i] = '\0'; |
| 2477 | sfx = nullptr; |
| 2478 | |
| 2479 | // FIRST WORD |
| 2480 | |
| 2481 | affixed = 1; |
| 2482 | |
| 2483 | presult.clear(); |
| 2484 | if (partresult) |
| 2485 | presult.append(*partresult); |
| 2486 | |
| 2487 | rv = lookup(st.c_str(), i); // perhaps without prefix |
| 2488 | |
| 2489 | // forbid dictionary stems with COMPOUNDFORBIDFLAG in |
| 2490 | // compound words, overriding the effect of COMPOUNDPERMITFLAG |
| 2491 | if ((rv) && compoundforbidflag && |
| 2492 | TESTAFF(rv->astr, compoundforbidflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundforbidflag )) && !hu_mov_rule) |
| 2493 | continue; |
| 2494 | |
| 2495 | // search homonym with compound flag |
| 2496 | while ((rv) && !hu_mov_rule && |
| 2497 | ((needaffix && TESTAFF(rv->astr, needaffix, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, needaffix ))) || |
| 2498 | !((compoundflag && !words && !onlycpdrule && |
| 2499 | TESTAFF(rv->astr, compoundflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundflag ))) || |
| 2500 | (compoundbegin && !wordnum && !onlycpdrule && |
| 2501 | TESTAFF(rv->astr, compoundbegin, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundbegin ))) || |
| 2502 | (compoundmiddle && wordnum && !words && !onlycpdrule && |
| 2503 | TESTAFF(rv->astr, compoundmiddle, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundmiddle ))) || |
| 2504 | (!defcpdtable.empty() && onlycpdrule && |
| 2505 | ((!words && !wordnum && |
| 2506 | defcpd_check(&words, wnum, maxwordnum, rv, rwords, 0)) || |
| 2507 | (words && |
| 2508 | defcpd_check(&words, wnum, maxwordnum, rv, rwords, 0))))))) { |
| 2509 | rv = rv->next_homonym; |
| 2510 | } |
| 2511 | |
| 2512 | |
| 2513 | if (rv) |
| 2514 | affixed = 0; |
| 2515 | |
| 2516 | if (rv) { |
| 2517 | presult.push_back(MSEP_FLD' '); |
| 2518 | presult.append(MORPH_PART"pa:"); |
| 2519 | presult.append(st, 0, i); |
| 2520 | if (!HENTRY_FIND(rv, MORPH_STEM"st:")) { |
| 2521 | presult.push_back(MSEP_FLD' '); |
| 2522 | presult.append(MORPH_STEM"st:"); |
| 2523 | presult.append(st, 0, i); |
| 2524 | } |
| 2525 | if (HENTRY_DATA(rv)) { |
| 2526 | presult.push_back(MSEP_FLD' '); |
| 2527 | presult.append(HENTRY_DATA2(rv)); |
| 2528 | } |
| 2529 | } |
| 2530 | |
| 2531 | if (!rv) { |
| 2532 | if (compoundflag && |
| 2533 | !(rv = |
| 2534 | prefix_check(st, 0, i, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1, |
| 2535 | scratch, compoundflag))) { |
| 2536 | if (((rv = suffix_check(st, 0, i, 0, nullptr, scratch, FLAG_NULL0x00, compoundflag, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1)) || |
| 2537 | (compoundmoresuffixes && (rv = suffix_check_twosfx(st, 0, i, 0, nullptr, scratch, compoundflag)))) && |
| 2538 | !hu_mov_rule && sfx->getCont() && |
| 2539 | ((compoundforbidflag && TESTAFF(sfx->getCont(), compoundforbidflag, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundforbidflag))) || |
| 2540 | (compoundend && TESTAFF(sfx->getCont(), compoundend, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundend))))) { |
| 2541 | rv = nullptr; |
| 2542 | } |
| 2543 | } |
| 2544 | |
| 2545 | if (rv || |
| 2546 | (((wordnum == 0) && compoundbegin && |
| 2547 | ((rv = suffix_check(st, 0, i, 0, nullptr, scratch, FLAG_NULL0x00, compoundbegin, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1)) || |
| 2548 | (compoundmoresuffixes && (rv = suffix_check_twosfx(st, 0, i, 0, nullptr, scratch, |
| 2549 | compoundbegin))) || // twofold suffix+compound |
| 2550 | (rv = prefix_check(st, 0, i, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1, scratch, compoundbegin)))) || |
| 2551 | ((wordnum > 0) && compoundmiddle && |
| 2552 | ((rv = suffix_check(st, 0, i, 0, nullptr, scratch, FLAG_NULL0x00, compoundmiddle, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1)) || |
| 2553 | (compoundmoresuffixes && (rv = suffix_check_twosfx(st, 0, i, 0, nullptr, scratch, |
| 2554 | compoundmiddle))) || // twofold suffix+compound |
| 2555 | (rv = prefix_check(st, 0, i, hu_mov_rule ? IN_CPD_OTHER3 : IN_CPD_BEGIN1, scratch, compoundmiddle)))))) { |
| 2556 | std::string p; |
| 2557 | if (compoundflag) |
| 2558 | p = affix_check_morph(st, 0, i, scratch, compoundflag); |
| 2559 | if (p.empty()) { |
| 2560 | if ((wordnum == 0) && compoundbegin) { |
| 2561 | p = affix_check_morph(st, 0, i, scratch, compoundbegin); |
| 2562 | } else if ((wordnum > 0) && compoundmiddle) { |
| 2563 | p = affix_check_morph(st, 0, i, scratch, compoundmiddle); |
| 2564 | } |
| 2565 | } |
| 2566 | presult.push_back(MSEP_FLD' '); |
| 2567 | presult.append(MORPH_PART"pa:"); |
| 2568 | presult.append(st, 0, i); |
| 2569 | if (!p.empty()) { |
| 2570 | line_uniq_app(p, MSEP_REC'\n'); |
| 2571 | if (!p.empty() && p[0] != MSEP_FLD' ') |
| 2572 | presult.push_back(MSEP_FLD' '); |
| 2573 | presult.append(p); |
| 2574 | } |
| 2575 | checked_prefix = 1; |
| 2576 | } |
| 2577 | // else check forbiddenwords |
| 2578 | } else if (rv->astr && (TESTAFF(rv->astr, forbiddenword, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forbiddenword )) || |
| 2579 | TESTAFF(rv->astr, ONLYUPCASEFLAG, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 65511 )) || |
| 2580 | TESTAFF(rv->astr, needaffix, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, needaffix )))) { |
| 2581 | st[i] = ch; |
| 2582 | continue; |
| 2583 | } |
| 2584 | |
| 2585 | // check non_compound flag in suffix and prefix |
| 2586 | if ((rv) && !hu_mov_rule && |
| 2587 | ((pfx && pfx->getCont() && |
| 2588 | TESTAFF(pfx->getCont(), compoundforbidflag, pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), compoundforbidflag))) || |
| 2589 | (sfx && sfx->getCont() && |
| 2590 | TESTAFF(sfx->getCont(), compoundforbidflag, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundforbidflag))))) { |
| 2591 | continue; |
| 2592 | } |
| 2593 | |
| 2594 | // check compoundend flag in suffix and prefix |
| 2595 | if ((rv) && !checked_prefix && compoundend && !hu_mov_rule && |
| 2596 | ((pfx && pfx->getCont() && |
| 2597 | TESTAFF(pfx->getCont(), compoundend, pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), compoundend))) || |
| 2598 | (sfx && sfx->getCont() && |
| 2599 | TESTAFF(sfx->getCont(), compoundend, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundend))))) { |
| 2600 | continue; |
| 2601 | } |
| 2602 | |
| 2603 | // check compoundmiddle flag in suffix and prefix |
| 2604 | if ((rv) && !checked_prefix && (wordnum == 0) && compoundmiddle && |
| 2605 | !hu_mov_rule && |
| 2606 | ((pfx && pfx->getCont() && |
| 2607 | TESTAFF(pfx->getCont(), compoundmiddle, pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), compoundmiddle))) || |
| 2608 | (sfx && sfx->getCont() && |
| 2609 | TESTAFF(sfx->getCont(), compoundmiddle, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundmiddle))))) { |
| 2610 | rv = nullptr; |
| 2611 | } |
| 2612 | |
| 2613 | // check forbiddenwords |
| 2614 | if ((rv) && (rv->astr) && (TESTAFF(rv->astr, forbiddenword, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forbiddenword )) || |
| 2615 | TESTAFF(rv->astr, ONLYUPCASEFLAG, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 65511 )))) |
| 2616 | continue; |
| 2617 | |
| 2618 | // increment word number, if the second root has a compoundroot flag |
| 2619 | if ((rv) && (compoundroot) && |
| 2620 | (TESTAFF(rv->astr, compoundroot, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundroot )))) { |
| 2621 | wordnum++; |
| 2622 | } |
| 2623 | |
| 2624 | // first word is acceptable in compound words? |
| 2625 | if (((rv) && |
| 2626 | (checked_prefix || (words && words[wnum]) || (compoundflag && TESTAFF(rv->astr, compoundflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundflag ))) || |
| 2627 | ((oldwordnum == 0) && compoundbegin && TESTAFF(rv->astr, compoundbegin, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundbegin ))) || |
| 2628 | ((oldwordnum > 0) && compoundmiddle && TESTAFF(rv->astr, compoundmiddle, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundmiddle ))) |
| 2629 | // LANG_hu section: spec. Hungarian rule |
| 2630 | || ((langnum == LANG_hu) && // hu_mov_rule |
| 2631 | hu_mov_rule && |
| 2632 | (TESTAFF(rv->astr, 'F', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'F' )) || TESTAFF(rv->astr, 'G', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'G' )) || TESTAFF(rv->astr, 'H', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'H' )))) |
| 2633 | // END of LANG_hu section |
| 2634 | ) && |
| 2635 | !((checkcompoundtriple && !words && // test triple letters |
| 2636 | (word[i - 1] == word[i]) && |
| 2637 | (((i > 1) && (word[i - 1] == word[i - 2])) || ((word[i - 1] == word[i + 1])) // may be word[i+1] == '\0' |
| 2638 | )) || |
| 2639 | ( |
| 2640 | // test CHECKCOMPOUNDPATTERN |
| 2641 | !checkcpdtable.empty() && !words && cpdpat_check(word, i, rv, nullptr, affixed, nullptr, nullptr, nullptr, nullptr, nullptr)) || |
| 2642 | (checkcompoundcase && !words && cpdcase_check(word, i)))) |
| 2643 | // LANG_hu section: spec. Hungarian rule |
| 2644 | || ((!rv) && (langnum == LANG_hu) && hu_mov_rule && (rv = affix_check(st, 0, i, scratch)) && |
| 2645 | (sfx && sfx->getCont() && |
| 2646 | (TESTAFF(sfx->getCont(), (unsigned short)'x', sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), (unsigned short)'x')) || |
| 2647 | TESTAFF(sfx->getCont(), (unsigned short)'%', sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), (unsigned short)'%'))))) |
| 2648 | // END of LANG_hu section |
| 2649 | ) { |
| 2650 | // LANG_hu section: spec. Hungarian rule |
| 2651 | if (langnum == LANG_hu) { |
| 2652 | // calculate syllable number of the word |
| 2653 | numsyllable += get_syllable(st.substr(0, i)); |
| 2654 | |
| 2655 | // + 1 word, if syllable number of the prefix > 1 (hungarian |
| 2656 | // convention) |
| 2657 | if (pfx && (get_syllable(pfx->getKey()) > 1)) |
| 2658 | wordnum++; |
| 2659 | } |
| 2660 | // END of LANG_hu section |
| 2661 | |
| 2662 | // NEXT WORD(S) |
| 2663 | rv_first = rv; |
| 2664 | rv = lookup(word.c_str() + i, word.size() - i); // perhaps without prefix |
| 2665 | |
| 2666 | // search homonym with compound flag |
| 2667 | while ((rv) && ((needaffix && TESTAFF(rv->astr, needaffix, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, needaffix ))) || |
| 2668 | !((compoundflag && !words && TESTAFF(rv->astr, compoundflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundflag ))) || |
| 2669 | (compoundend && !words && TESTAFF(rv->astr, compoundend, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundend ))) || |
| 2670 | (!defcpdtable.empty() && words && defcpd_check(&words, wnum + 1, maxwordnum, rv, nullptr, 1))))) { |
| 2671 | rv = rv->next_homonym; |
| 2672 | } |
| 2673 | |
| 2674 | if (rv && words && words[wnum + 1]) { |
| 2675 | result.append(presult); |
| 2676 | result.push_back(MSEP_FLD' '); |
| 2677 | result.append(MORPH_PART"pa:"); |
| 2678 | result.append(word, i, word.size()); |
| 2679 | if (complexprefixes && HENTRY_DATA(rv)) |
| 2680 | result.append(HENTRY_DATA2(rv)); |
| 2681 | if (!HENTRY_FIND(rv, MORPH_STEM"st:")) { |
| 2682 | result.push_back(MSEP_FLD' '); |
| 2683 | result.append(MORPH_STEM"st:"); |
| 2684 | result.append(HENTRY_WORD(rv)&(rv->word[0])); |
| 2685 | } |
| 2686 | // store the pointer of the hash entry |
| 2687 | if (!complexprefixes && HENTRY_DATA(rv)) { |
| 2688 | result.push_back(MSEP_FLD' '); |
| 2689 | result.append(HENTRY_DATA2(rv)); |
| 2690 | } |
| 2691 | result.push_back(MSEP_REC'\n'); |
| 2692 | return 0; |
| 2693 | } |
| 2694 | |
| 2695 | oldnumsyllable2 = numsyllable; |
| 2696 | oldwordnum2 = wordnum; |
| 2697 | |
| 2698 | // LANG_hu section: spec. Hungarian rule |
| 2699 | if ((rv) && (langnum == LANG_hu) && |
| 2700 | (TESTAFF(rv->astr, 'I', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'I' ))) && |
| 2701 | !(TESTAFF(rv->astr, 'J', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'J' )))) { |
| 2702 | numsyllable--; |
| 2703 | } |
| 2704 | // END of LANG_hu section |
| 2705 | // increment word number, if the second root has a compoundroot flag |
| 2706 | if ((rv) && (compoundroot) && |
| 2707 | (TESTAFF(rv->astr, compoundroot, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundroot )))) { |
| 2708 | wordnum++; |
| 2709 | } |
| 2710 | |
| 2711 | // check forbiddenwords |
| 2712 | if ((rv) && (rv->astr) && |
| 2713 | (TESTAFF(rv->astr, forbiddenword, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forbiddenword )) || |
| 2714 | TESTAFF(rv->astr, ONLYUPCASEFLAG, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 65511 )))) { |
| 2715 | st[i] = ch; |
| 2716 | continue; |
| 2717 | } |
| 2718 | |
| 2719 | // second word is acceptable, as a root? |
| 2720 | // hungarian conventions: compounding is acceptable, |
| 2721 | // when compound forms consist of 2 words, or if more, |
| 2722 | // then the syllable number of root words must be 6, or lesser. |
| 2723 | if ((rv) && |
| 2724 | ((compoundflag && TESTAFF(rv->astr, compoundflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundflag ))) || |
| 2725 | (compoundend && TESTAFF(rv->astr, compoundend, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundend )))) && |
| 2726 | (((cpdwordmax == -1) || (wordnum + 1 < cpdwordmax)) || |
| 2727 | ((cpdmaxsyllable != 0) && |
| 2728 | (numsyllable + get_syllable(std::string(HENTRY_WORD(rv)&(rv->word[0]), rv->blen)) <= |
| 2729 | cpdmaxsyllable))) && |
| 2730 | ((!checkcompounddup || (rv != rv_first)))) { |
| 2731 | // bad compound word |
| 2732 | result.append(presult); |
| 2733 | result.push_back(MSEP_FLD' '); |
| 2734 | result.append(MORPH_PART"pa:"); |
| 2735 | result.append(word, i, word.size()); |
| 2736 | |
| 2737 | if (HENTRY_DATA(rv)) { |
| 2738 | if (complexprefixes) |
| 2739 | result.append(HENTRY_DATA2(rv)); |
| 2740 | if (!HENTRY_FIND(rv, MORPH_STEM"st:")) { |
| 2741 | result.push_back(MSEP_FLD' '); |
| 2742 | result.append(MORPH_STEM"st:"); |
| 2743 | result.append(HENTRY_WORD(rv)&(rv->word[0])); |
| 2744 | } |
| 2745 | // store the pointer of the hash entry |
| 2746 | if (!complexprefixes) { |
| 2747 | result.push_back(MSEP_FLD' '); |
| 2748 | result.append(HENTRY_DATA2(rv)); |
| 2749 | } |
| 2750 | } |
| 2751 | result.push_back(MSEP_REC'\n'); |
| 2752 | ok = 1; |
| 2753 | } |
| 2754 | |
| 2755 | numsyllable = oldnumsyllable2; |
| 2756 | wordnum = oldwordnum2; |
| 2757 | |
| 2758 | // perhaps second word has prefix or/and suffix |
| 2759 | sfx = nullptr; |
| 2760 | sfxflag = FLAG_NULL0x00; |
| 2761 | |
| 2762 | if (compoundflag && !onlycpdrule) |
| 2763 | rv = affix_check(word, i, word.size() - i, scratch, compoundflag); |
| 2764 | else |
| 2765 | rv = nullptr; |
| 2766 | |
| 2767 | if (!rv && compoundend && !onlycpdrule) { |
| 2768 | sfx = nullptr; |
| 2769 | pfx = nullptr; |
| 2770 | rv = affix_check(word, i, word.size() - i, scratch, compoundend, IN_CPD_END2); |
| 2771 | } |
| 2772 | |
| 2773 | if (!rv && !defcpdtable.empty() && words) { |
| 2774 | rv = affix_check(word, i, word.size() - i, scratch, 0, IN_CPD_END2); |
| 2775 | if (rv && words && defcpd_check(&words, wnum + 1, maxwordnum, rv, nullptr, 1)) { |
| 2776 | std::string m; |
| 2777 | if (compoundflag) |
| 2778 | m = affix_check_morph(word, i, word.size() - i, scratch, compoundflag); |
| 2779 | if (m.empty() && compoundend) { |
| 2780 | m = affix_check_morph(word, i, word.size() - i, scratch, compoundend); |
| 2781 | } |
| 2782 | result.append(presult); |
| 2783 | if (!m.empty()) { |
| 2784 | result.push_back(MSEP_FLD' '); |
| 2785 | result.append(MORPH_PART"pa:"); |
| 2786 | result.append(word, i, word.size()); |
| 2787 | line_uniq_app(m, MSEP_REC'\n'); |
| 2788 | result.append(m); |
| 2789 | } |
| 2790 | result.push_back(MSEP_REC'\n'); |
| 2791 | ok = 1; |
| 2792 | } |
| 2793 | } |
| 2794 | |
| 2795 | // check non_compound flag in suffix and prefix |
| 2796 | if ((rv) && |
| 2797 | ((pfx && pfx->getCont() && |
| 2798 | TESTAFF(pfx->getCont(), compoundforbidflag, pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), compoundforbidflag))) || |
| 2799 | (sfx && sfx->getCont() && |
| 2800 | TESTAFF(sfx->getCont(), compoundforbidflag,(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundforbidflag)) |
| 2801 | sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundforbidflag))))) { |
| 2802 | rv = nullptr; |
| 2803 | } |
| 2804 | |
| 2805 | // check forbiddenwords |
| 2806 | if ((rv) && (rv->astr) && |
| 2807 | (TESTAFF(rv->astr, forbiddenword, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, forbiddenword )) || |
| 2808 | TESTAFF(rv->astr, ONLYUPCASEFLAG, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 65511 ))) && |
| 2809 | (!TESTAFF(rv->astr, needaffix, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, needaffix )))) { |
| 2810 | st[i] = ch; |
| 2811 | continue; |
| 2812 | } |
| 2813 | |
| 2814 | if (langnum == LANG_hu) { |
| 2815 | // calculate syllable number of the word |
| 2816 | numsyllable += get_syllable(word.c_str() + i); |
| 2817 | |
| 2818 | // - affix syllable num. |
| 2819 | // XXX only second suffix (inflections, not derivations) |
| 2820 | if (sfxappnd) { |
| 2821 | std::string tmp(sfxappnd); |
| 2822 | reverseword(tmp); |
| 2823 | numsyllable -= short(get_syllable(tmp) + sfxextra); |
| 2824 | } else { |
| 2825 | numsyllable -= short(sfxextra); |
| 2826 | } |
| 2827 | |
| 2828 | // + 1 word, if syllable number of the prefix > 1 (hungarian |
| 2829 | // convention) |
| 2830 | if (pfx && (get_syllable(pfx->getKey()) > 1)) |
| 2831 | wordnum++; |
| 2832 | |
| 2833 | // increment syllable num, if last word has a SYLLABLENUM flag |
| 2834 | // and the suffix is beginning `s' |
| 2835 | |
| 2836 | if (!cpdsyllablenum.empty()) { |
| 2837 | switch (sfxflag) { |
| 2838 | case 'c': { |
| 2839 | numsyllable += 2; |
| 2840 | break; |
| 2841 | } |
| 2842 | case 'J': { |
| 2843 | numsyllable += 1; |
| 2844 | break; |
| 2845 | } |
| 2846 | case 'I': { |
| 2847 | if (rv && TESTAFF(rv->astr, 'J', rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, 'J' ))) |
| 2848 | numsyllable += 1; |
| 2849 | break; |
| 2850 | } |
| 2851 | } |
| 2852 | } |
| 2853 | } |
| 2854 | |
| 2855 | // increment word number, if the second word has a compoundroot flag |
| 2856 | if ((rv) && (compoundroot) && |
| 2857 | (TESTAFF(rv->astr, compoundroot, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, compoundroot )))) { |
| 2858 | wordnum++; |
| 2859 | } |
| 2860 | // second word is acceptable, as a word with prefix or/and suffix? |
| 2861 | // hungarian conventions: compounding is acceptable, |
| 2862 | // when compound forms consist 2 word, otherwise |
| 2863 | // the syllable number of root words is 6, or lesser. |
| 2864 | if ((rv) && |
| 2865 | (((cpdwordmax == -1) || (wordnum + 1 < cpdwordmax)) || |
| 2866 | ((cpdmaxsyllable != 0) && (numsyllable <= cpdmaxsyllable))) && |
| 2867 | ((!checkcompounddup || (rv != rv_first)))) { |
| 2868 | std::string m; |
| 2869 | if (compoundflag) |
| 2870 | m = affix_check_morph(word, i, word.size() - i, scratch, compoundflag, IN_CPD_END2); |
| 2871 | if (m.empty() && compoundend) { |
| 2872 | m = affix_check_morph(word, i, word.size() - i, scratch, compoundend, IN_CPD_END2); |
| 2873 | } |
| 2874 | result.append(presult); |
| 2875 | if (!m.empty()) { |
| 2876 | result.push_back(MSEP_FLD' '); |
| 2877 | result.append(MORPH_PART"pa:"); |
| 2878 | result.append(word, i, word.size()); |
| 2879 | line_uniq_app(m, MSEP_REC'\n'); |
| 2880 | result.push_back(MSEP_FLD' '); |
| 2881 | result.append(m); |
| 2882 | } |
| 2883 | result.push_back(MSEP_REC'\n'); |
| 2884 | ok = 1; |
| 2885 | } |
| 2886 | |
| 2887 | numsyllable = oldnumsyllable2; |
| 2888 | wordnum = oldwordnum2; |
| 2889 | |
| 2890 | // perhaps second word is a compound word (recursive call) |
| 2891 | if ((wordnum + 2 < maxwordnum) && (wnum + 1 < maxwordnum) && (ok == 0)) { |
| 2892 | compound_check_morph(word.substr(i), wordnum + 1, |
| 2893 | numsyllable, maxwordnum, wnum + 1, words, rwords, 0, |
| 2894 | result, &presult, scratch); |
| 2895 | } else { |
| 2896 | rv = nullptr; |
| 2897 | } |
| 2898 | } |
| 2899 | st[i] = ch; |
| 2900 | wordnum = oldwordnum; |
| 2901 | numsyllable = oldnumsyllable; |
| 2902 | |
| 2903 | } while (!defcpdtable.empty() && oldwordnum == 0 && |
| 2904 | onlycpdrule++ < 1); // end of onlycpd loop |
| 2905 | } |
| 2906 | return 0; |
| 2907 | } |
| 2908 | |
| 2909 | |
| 2910 | inline int AffixMgr::isRevSubset(const char* s1, |
| 2911 | const char* end_of_s2, |
| 2912 | int len) { |
| 2913 | while ((len > 0) && (*s1 != '\0') && ((*s1 == *end_of_s2) || (*s1 == '.'))) { |
| 2914 | s1++; |
| 2915 | end_of_s2--; |
| 2916 | len--; |
| 2917 | } |
| 2918 | return (*s1 == '\0'); |
| 2919 | } |
| 2920 | |
| 2921 | // a circumfix is one affix split in two halves, so the flag has to be on both the prefix and the |
| 2922 | // suffix, or on neither of them |
| 2923 | bool AffixMgr::circumfix_ok(PfxEntry* pfx, SfxEntry* sfx, const TraceCtx* t) const { |
| 2924 | if (!circumfix) |
| 2925 | return true; |
| 2926 | bool in_prefix = pfx && pfx->getCont() && |
| 2927 | TESTAFF(pfx->getCont(), circumfix, pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), circumfix)); |
| 2928 | bool in_suffix = sfx->getCont() && |
| 2929 | TESTAFF(sfx->getCont(), circumfix, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), circumfix)); |
| 2930 | if (t) |
| 2931 | trace_circumfix(*t, this, circumfix, pfx, sfx, in_prefix, in_suffix); |
| 2932 | return in_prefix == in_suffix; |
| 2933 | } |
| 2934 | |
| 2935 | // decide whether a suffix entry may be applied at all, before the suffix itself is matched |
| 2936 | // against the word |
| 2937 | // the stem was found and then dropped again for carrying the flag the caller asked to keep away |
| 2938 | // from |
| 2939 | void AffixMgr::trace_avoidflag(TraceCtx* t, |
| 2940 | const FLAGunsigned short avoidflag, |
| 2941 | const struct hentry* stem) const { |
| 2942 | if (!t) |
| 2943 | return; |
| 2944 | TraceScope trace_depth(t); |
| 2945 | trace_test(*t, "avoidflag", this, avoidflag, "dic", stem->astr, stem->alen, |
| 2946 | "fail, the caller is skipping stems with this flag"); |
| 2947 | } |
| 2948 | |
| 2949 | bool AffixMgr::suffix_applicable(PfxEntry* pfx, |
| 2950 | SfxEntry* sfx, |
| 2951 | const FLAGunsigned short cclass, |
| 2952 | char in_compound, |
| 2953 | const TraceCtx* t) const { |
| 2954 | // suffixes are only allowed at the beginning of a compound when they are signed with the |
| 2955 | // compoundpermitflag flag |
| 2956 | if (in_compound == IN_CPD_BEGIN1 && |
| 2957 | !(sfx->getCont() && compoundpermitflag && |
| 2958 | TESTAFF(sfx->getCont(), compoundpermitflag, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), compoundpermitflag)))) { |
| 2959 | if (t) |
| 2960 | trace_test(*t, "compoundpermit", this, compoundpermitflag, "sfx-cont", |
| 2961 | sfx->getCont(), sfx->getContLen(), |
| 2962 | "fail, a suffix at the start of a compound needs this flag"); |
| 2963 | return false; |
| 2964 | } |
| 2965 | |
| 2966 | if (!circumfix_ok(pfx, sfx, t)) |
| 2967 | return false; |
| 2968 | |
| 2969 | // a fogemorpheme is only allowed inside a compound |
| 2970 | if (!in_compound && sfx->getCont() && |
| 2971 | TESTAFF(sfx->getCont(), onlyincompound, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), onlyincompound))) { |
| 2972 | if (t) |
| 2973 | trace_test(*t, "onlyincompound", this, onlyincompound, "sfx-cont", |
| 2974 | sfx->getCont(), sfx->getContLen(), |
| 2975 | "fail, this suffix is only allowed inside a compound"); |
| 2976 | return false; |
| 2977 | } |
| 2978 | |
| 2979 | // a needaffix suffix needs a further affix, so it is allowed either as a second suffix or |
| 2980 | // when a prefix is present that does not itself need one |
| 2981 | if (!cclass && sfx->getCont() && |
| 2982 | TESTAFF(sfx->getCont(), needaffix, sfx->getContLen())(std::binary_search(sfx->getCont(), sfx->getCont() + sfx ->getContLen(), needaffix)) && |
| 2983 | !(pfx && !(pfx->getCont() && |
| 2984 | TESTAFF(pfx->getCont(), needaffix, pfx->getContLen())(std::binary_search(pfx->getCont(), pfx->getCont() + pfx ->getContLen(), needaffix))))) { |
| 2985 | if (t) |
| 2986 | trace_test(*t, "needaffix", this, needaffix, "sfx-cont", sfx->getCont(), |
| 2987 | sfx->getContLen(), |
| 2988 | "fail, this suffix needs a further affix and has none"); |
| 2989 | return false; |
| 2990 | } |
| 2991 | |
| 2992 | return true; |
| 2993 | } |
| 2994 | |
| 2995 | // check word for suffixes |
| 2996 | struct hentry* AffixMgr::suffix_check(const std::string& word, |
| 2997 | int start, |
| 2998 | int len, |
| 2999 | int sfxopts, |
| 3000 | PfxEntry* ppfx, |
| 3001 | AffixScratch& scratch, |
| 3002 | const FLAGunsigned short cclass, |
| 3003 | const FLAGunsigned short needflag, |
| 3004 | char in_compound, |
| 3005 | const FLAGunsigned short avoidflag) { |
| 3006 | struct hentry* rv = nullptr; |
| 3007 | |
| 3008 | TraceCtx* t = trace_on(scratch.trace); |
| 3009 | int candidates = 0; |
| 3010 | // a pass that had nothing to try says so, which is different from a pass that |
| 3011 | // did not run |
| 3012 | auto report_empty_pass = [t, &word, start, len, sfxopts, &candidates]() { |
| 3013 | if (t && candidates == 0) |
| 3014 | trace(*t, "sfx \"%s\" candidates=0%s", word.substr(start, len).c_str(), |
| 3015 | (sfxopts & aeXPRODUCT(1 << 0)) != 0 ? " xprod=Y" : ""); |
| 3016 | }; |
| 3017 | // a rule the suffix conditions turned away never reaches checkword, so name |
| 3018 | // the rule here and run the conditions again to say which of them refused it |
| 3019 | auto report_refused = [this, t, ppfx, cclass, in_compound, |
| 3020 | &candidates](SfxEntry* se) { |
| 3021 | if (!t) |
| 3022 | return; |
| 3023 | ++candidates; |
| 3024 | trace_affix(*t, "sfx", this, *se); |
| 3025 | TraceScope trace_depth(t); |
| 3026 | suffix_applicable(ppfx, se, cclass, in_compound, t); |
| 3027 | }; |
| 3028 | |
| 3029 | // first handle the special case of 0 length suffixes |
| 3030 | SfxEntry* se = sStart[0]; |
| 3031 | |
| 3032 | while (se) { |
| 3033 | if (!cclass || se->getCont()) { |
| 3034 | if (suffix_applicable(ppfx, se, cclass, in_compound, nullptr)) { |
| 3035 | ++candidates; |
| 3036 | rv = se->checkword(word, start, len, sfxopts, ppfx, |
| 3037 | (FLAGunsigned short)cclass, needflag, |
| 3038 | (in_compound ? 0 : onlyincompound), |
| 3039 | scratch); |
| 3040 | // Skip a stem with the avoid flag and keep scanning the other suffixes. |
| 3041 | if (rv && avoidflag != FLAG_NULL0x00 && TESTAFF(rv->astr, avoidflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, avoidflag ))) { |
| 3042 | trace_avoidflag(t, avoidflag, rv); |
| 3043 | rv = nullptr; |
| 3044 | } |
| 3045 | if (rv) { |
| 3046 | sfx = se; // BUG: sfx not stateless |
| 3047 | return rv; |
| 3048 | } |
| 3049 | } else { |
| 3050 | report_refused(se); |
| 3051 | } |
| 3052 | } |
| 3053 | se = se->getNext(); |
| 3054 | } |
| 3055 | |
| 3056 | // now handle the general case |
| 3057 | if (len == 0) { |
| 3058 | report_empty_pass(); |
| 3059 | return nullptr; // FULLSTRIP |
| 3060 | } |
| 3061 | unsigned char sp = word[start + len - 1]; |
| 3062 | SfxEntry* sptr = sStart[sp]; |
| 3063 | |
| 3064 | while (sptr) { |
| 3065 | if (isRevSubset(sptr->getKey(), word.c_str() + start + len - 1, len)) { |
| 3066 | if (!suffix_applicable(ppfx, sptr, cclass, in_compound, nullptr)) |
| 3067 | report_refused(sptr); |
| 3068 | else if (in_compound != IN_CPD_END2 || ppfx || |
| 3069 | !(sptr->getCont() && |
| 3070 | TESTAFF(sptr->getCont(), onlyincompound, sptr->getContLen())(std::binary_search(sptr->getCont(), sptr->getCont() + sptr ->getContLen(), onlyincompound)))) { |
| 3071 | ++candidates; |
| 3072 | rv = sptr->checkword(word, start, len, sfxopts, ppfx, |
| 3073 | cclass, needflag, |
| 3074 | (in_compound ? 0 : onlyincompound), |
| 3075 | scratch); |
| 3076 | if (rv && avoidflag != FLAG_NULL0x00 && TESTAFF(rv->astr, avoidflag, rv->alen)(std::binary_search(rv->astr, rv->astr + rv->alen, avoidflag ))) { |
| 3077 | trace_avoidflag(t, avoidflag, rv); |
| 3078 | rv = nullptr; |
| 3079 | } |
| 3080 | if (rv) { |
| 3081 | sfx = sptr; // BUG: sfx not stateless |
| 3082 | sfxflag = sptr->getFlag(); // BUG: sfxflag not stateless |
| 3083 | if (!sptr->getCont()) |
| 3084 | sfxappnd = sptr->getKey(); // BUG: sfxappnd not stateless |
| 3085 | // LANG_hu section: spec. Hungarian rule |
| 3086 | else if (langnum == LANG_hu && sptr->getKeyLen() && |
| 3087 | sptr->getKey()[0] == 'i' && sptr->getKey()[1] != 'y' && |
| 3088 | sptr->getKey()[1] != 't') { |
| 3089 | sfxextra = 1; |
| 3090 | } |
| 3091 | // END of LANG_hu section |
| 3092 | return rv; |
| 3093 | } |
| 3094 | } |
| 3095 | sptr = sptr->getNextEQ(); |
| 3096 | } else { |
| 3097 | sptr = sptr->getNextNE(); |
| 3098 | } |
| 3099 | } |
| 3100 | |
| 3101 | report_empty_pass(); |
| 3102 | |
| 3103 | return nullptr; |
| 3104 | } |
| 3105 | |
| 3106 | // check word for two-level suffixes |
| 3107 | struct hentry* AffixMgr::suffix_check_twosfx(const std::string& word, |
| 3108 | int start, |
| 3109 | int len, |
| 3110 | int sfxopts, |
| 3111 | PfxEntry* ppfx, |
| 3112 | AffixScratch& scratch, |
| 3113 | const FLAGunsigned short needflag) { |
| 3114 | struct hentry* rv = nullptr; |
| 3115 | |
| 3116 | // first handle the special case of 0 length suffixes |
| 3117 | SfxEntry* se = sStart[0]; |
| 3118 | while (se) { |
| 3119 | if (contclasses[se->getFlag()]) { |
| 3120 | rv = se->check_twosfx(word, start, len, sfxopts, ppfx, needflag, scratch); |
| 3121 | if (rv) |
| 3122 | return rv; |
| 3123 | } |
| 3124 | se = se->getNext(); |
| 3125 | } |
| 3126 | |
| 3127 | // now handle the general case |
| 3128 | if (len == 0) |
| 3129 | return nullptr; // FULLSTRIP |
| 3130 | unsigned char sp = word[start + len - 1]; |
| 3131 | SfxEntry* sptr = sStart[sp]; |
| 3132 | |
| 3133 | while (sptr) { |
| 3134 | if (isRevSubset(sptr->getKey(), word.c_str() + start + len - 1, len)) { |
| 3135 | if (contclasses[sptr->getFlag()]) { |
| 3136 | rv = sptr->check_twosfx(word, start, len, sfxopts, ppfx, needflag, scratch); |
| 3137 | if (rv) { |
| 3138 | sfxflag = sptr->getFlag(); // BUG: sfxflag not stateless |
| 3139 | if (!sptr->getCont()) |
| 3140 | sfxappnd = sptr->getKey(); // BUG: sfxappnd not stateless |
| 3141 | return rv; |
| 3142 | } |
| 3143 | } |
| 3144 | sptr = sptr->getNextEQ(); |
| 3145 | } else { |
| 3146 | sptr = sptr->getNextNE(); |
| 3147 | } |
| 3148 | } |
| 3149 | |
| 3150 | return nullptr; |
| 3151 | } |
| 3152 | |
| 3153 | // check word for two-level suffixes and morph |
| 3154 | std::string AffixMgr::suffix_check_twosfx_morph(const std::string& word, |
| 3155 | int start, |
| 3156 | int len, |
| 3157 | int sfxopts, |
| 3158 | PfxEntry* ppfx, |
| 3159 | AffixScratch& scratch, |
| 3160 | const FLAGunsigned short needflag) { |
| 3161 | std::string result2; |
| 3162 | std::string result3; |
| 3163 | // the second suffix runs the whole suffix table again for every first suffix |
| 3164 | DistinctRecords result(MSEP_REC'\n'); |
| 3165 | |
| 3166 | // first handle the special case of 0 length suffixes |
| 3167 | SfxEntry* se = sStart[0]; |
| 3168 | while (se) { |
| 3169 | if (contclasses[se->getFlag()]) { |
| 3170 | std::string st = se->check_twosfx_morph(word, start, len, sfxopts, ppfx, needflag, scratch); |
| 3171 | if (!st.empty()) { |
| 3172 | std::string analysis; |
| 3173 | if (ppfx) { |
| 3174 | if (ppfx->getMorph()) { |
| 3175 | analysis.append(ppfx->getMorph()); |
| 3176 | analysis.push_back(MSEP_FLD' '); |
| 3177 | } else |
| 3178 | debugflag(analysis, ppfx->getFlag()); |
| 3179 | } |
| 3180 | analysis.append(st); |
| 3181 | if (se->getMorph()) { |
| 3182 | analysis.push_back(MSEP_FLD' '); |
| 3183 | analysis.append(se->getMorph()); |
| 3184 | } else |
| 3185 | debugflag(analysis, se->getFlag()); |
| 3186 | result.append(analysis); |
| 3187 | } |
| 3188 | } |
| 3189 | se = se->getNext(); |
| 3190 | } |
| 3191 | |
| 3192 | // now handle the general case |
| 3193 | if (len == 0) |
| 3194 | return { }; // FULLSTRIP |
| 3195 | unsigned char sp = word[start + len - 1]; |
| 3196 | SfxEntry* sptr = sStart[sp]; |
| 3197 | |
| 3198 | while (sptr) { |
| 3199 | if (isRevSubset(sptr->getKey(), word.c_str() + start + len - 1, len)) { |
| 3200 | if (contclasses[sptr->getFlag()]) { |
| 3201 | std::string st = sptr->check_twosfx_morph(word, start, len, sfxopts, ppfx, needflag, scratch); |
| 3202 | if (!st.empty()) { |
| 3203 | sfxflag = sptr->getFlag(); // BUG: sfxflag not stateless |
| 3204 | if (!sptr->getCont()) |
| 3205 | sfxappnd = sptr->getKey(); // BUG: sfxappnd not stateless |
| 3206 | result2.assign(st); |
| 3207 | |
| 3208 | result3.clear(); |
| 3209 | |
| 3210 | if (sptr->getMorph()) { |
| 3211 | result3.push_back(MSEP_FLD' '); |
| 3212 | result3.append(sptr->getMorph()); |
| 3213 | } else |
| 3214 | debugflag(result3, sptr->getFlag()); |
| 3215 | strlinecat(result2, result3); |
| 3216 | result.append(result2); |
| 3217 | } |
| 3218 | } |
| 3219 | sptr = sptr->getNextEQ(); |
| 3220 | } else { |
| 3221 | sptr = sptr->getNextNE(); |
| 3222 | } |
| 3223 | } |
| 3224 | |
| 3225 | return result.join(); |
| 3226 | } |
| 3227 | |
| 3228 | std::string AffixMgr::suffix_check_morph(const std::string& word, |
| 3229 | int start, |
| 3230 | int len, |
| 3231 | int sfxopts, |
| 3232 | PfxEntry* ppfx, |
| 3233 | AffixScratch& scratch, |
| 3234 | const FLAGunsigned short cclass, |
| 3235 | const FLAGunsigned short needflag, |
| 3236 | char in_compound) { |
| 3237 | std::string result; |
| 3238 | |
| 3239 | struct hentry* rv = nullptr; |
| 3240 | |
| 3241 | // first handle the special case of 0 length suffixes |
| 3242 | SfxEntry* se = sStart[0]; |
| 3243 | while (se) { |
| 3244 | if (!cclass || se->getCont()) { |
| 3245 | if (suffix_applicable(ppfx, se, cclass, in_compound, nullptr)) |
| 3246 | rv = se->checkword(word, start, len, sfxopts, ppfx, cclass, |
| 3247 | needflag, FLAG_NULL0x00, scratch); |
| 3248 | while (rv) { |
| 3249 | if (ppfx) { |
| 3250 | if (ppfx->getMorph()) { |
| 3251 | result.append(ppfx->getMorph()); |
| 3252 | result.push_back(MSEP_FLD' '); |
| 3253 | } else |
| 3254 | debugflag(result, ppfx->getFlag()); |
| 3255 | } |
| 3256 | if (complexprefixes && HENTRY_DATA(rv)) |
| 3257 | result.append(HENTRY_DATA2(rv)); |
| 3258 | if (!HENTRY_FIND(rv, MORPH_STEM"st:")) { |
| 3259 | result.push_back(MSEP_FLD' '); |
| 3260 | result.append(MORPH_STEM"st:"); |
| 3261 | result.append(HENTRY_WORD(rv)&(rv->word[0])); |
| 3262 | } |
| 3263 | |
| 3264 | if (!complexprefixes && HENTRY_DATA(rv)) { |
| 3265 | result.push_back(MSEP_FLD' '); |
| 3266 | result.append(HENTRY_DATA2(rv)); |
| 3267 | } |
| 3268 | if (se->getMorph()) { |
| 3269 | result.push_back(MSEP_FLD' '); |
| 3270 | result.append(se->getMorph()); |
| 3271 | } else |
| 3272 | debugflag(result, se->getFlag()); |
| 3273 | result.push_back(MSEP_REC'\n'); |
| 3274 | rv = se->get_next_homonym(rv, sfxopts, ppfx, cclass, needflag); |
| 3275 | } |
| 3276 | } |
| 3277 | se = se->getNext(); |
| 3278 | } |
| 3279 | |
| 3280 | // now handle the general case |
| 3281 | if (len == 0) |
| 3282 | return { }; // FULLSTRIP |
| 3283 | unsigned char sp = word[start + len - 1]; |
| 3284 | SfxEntry* sptr = sStart[sp]; |
| 3285 | |
| 3286 | while (sptr) { |
| 3287 | if (isRevSubset(sptr->getKey(), word.c_str() + start + len - 1, len)) { |
| 3288 | if (suffix_applicable(ppfx, sptr, cclass, in_compound, nullptr)) |
| 3289 | rv = sptr->checkword(word, start, len, sfxopts, ppfx, cclass, |
| 3290 | needflag, FLAG_NULL0x00, scratch); |
| 3291 | while (rv) { |
| 3292 | if (ppfx) { |
| 3293 | if (ppfx->getMorph()) { |
| 3294 | result.append(ppfx->getMorph()); |
| 3295 | result.push_back(MSEP_FLD' '); |
| 3296 | } else |
| 3297 | debugflag(result, ppfx->getFlag()); |
| 3298 | } |
| 3299 | if (complexprefixes && HENTRY_DATA(rv)) |
| 3300 | result.append(HENTRY_DATA2(rv)); |
| 3301 | if (!HENTRY_FIND(rv, MORPH_STEM"st:")) { |
| 3302 | result.push_back(MSEP_FLD' '); |
| 3303 | result.append(MORPH_STEM"st:"); |
| 3304 | result.append(HENTRY_WORD(rv)&(rv->word[0])); |
| 3305 | } |
| 3306 | |
| 3307 | if (!complexprefixes && HENTRY_DATA(rv)) { |
| 3308 | result.push_back(MSEP_FLD' '); |
| 3309 | result.append(HENTRY_DATA2(rv)); |
| 3310 | } |
| 3311 | |
| 3312 | if (sptr->getMorph()) { |
| 3313 | result.push_back(MSEP_FLD' '); |
| 3314 | result.append(sptr->getMorph()); |
| 3315 | } else |
| 3316 | debugflag(result, sptr->getFlag()); |
| 3317 | result.push_back(MSEP_REC'\n'); |
| 3318 | rv = sptr->get_next_homonym(rv, sfxopts, ppfx, cclass, needflag); |
| 3319 | } |
| 3320 | sptr = sptr->getNextEQ(); |
| 3321 | } else { |
| 3322 | sptr = sptr->getNextNE(); |
| 3323 | } |
| 3324 | } |
| 3325 | |
| 3326 | return result; |
| 3327 | } |
| 3328 | |
| 3329 | // check if word with affixes is correctly spelled |
| 3330 | struct hentry* AffixMgr::affix_check(const std::string& word, |
| 3331 | int start, |
| 3332 | int len, |
| 3333 | AffixScratch& scratch, |
| 3334 | const FLAGunsigned short needflag, |
| 3335 | char in_compound, |
| 3336 | const FLAGunsigned short avoidflag, |
| 3337 | PfxEntry** found_pfx, |
| 3338 | SfxEntry** found_sfx) { |
| 3339 | |
| 3340 | TraceCtx* t = trace_on(scratch.trace); |
| 3341 | // the affixes that built the word are reported before the members holding them are cleared |
| 3342 | auto report_form = [this, t, &word, start, len, found_pfx, found_sfx](const struct hentry* rv) { |
| 3343 | if (!rv) |
| 3344 | return; |
| 3345 | if (t) |
| 3346 | trace_form(*t, this, word.substr(start, len), rv->word, pfx, sfx); |
| 3347 | if (found_pfx) |
| 3348 | *found_pfx = pfx; |
| 3349 | if (found_sfx) |
| 3350 | *found_sfx = sfx; |
| 3351 | }; |
| 3352 | |
| 3353 | // check all prefixes (also crossed with suffixes if allowed) |
| 3354 | struct hentry* rv = prefix_check(word, start, len, in_compound, scratch, needflag, avoidflag); |
| 3355 | if (rv) { |
| 3356 | report_form(rv); |
| 3357 | return rv; |
| 3358 | } |
| 3359 | |
| 3360 | // if still not found check all suffixes |
| 3361 | rv = suffix_check(word, start, len, 0, nullptr, scratch, FLAG_NULL0x00, needflag, in_compound, avoidflag); |
| 3362 | |
| 3363 | if (havecontclass) { |
| 3364 | if (rv) |
| 3365 | report_form(rv); |
| 3366 | |
| 3367 | sfx = nullptr; |
| 3368 | pfx = nullptr; |
| 3369 | |
| 3370 | if (rv) |
| 3371 | return rv; |
| 3372 | // if still not found check all two-level suffixes |
| 3373 | rv = suffix_check_twosfx(word, start, len, 0, nullptr, scratch, needflag); |
| 3374 | |
| 3375 | if (rv) { |
| 3376 | report_form(rv); |
| 3377 | return rv; |
| 3378 | } |
| 3379 | // if still not found check all two-level suffixes |
| 3380 | rv = prefix_check_twosfx(word, start, len, IN_CPD_NOT0, scratch, needflag); |
| 3381 | } |
| 3382 | |
| 3383 | report_form(rv); |
| 3384 | |
| 3385 | return rv; |
| 3386 | } |
| 3387 | |
| 3388 | // check if word with affixes is correctly spelled |
| 3389 | std::string AffixMgr::affix_check_morph(const std::string& word, |
| 3390 | int start, |
| 3391 | int len, |
| 3392 | AffixScratch& scratch, |
| 3393 | const FLAGunsigned short needflag, |
| 3394 | char in_compound) { |
| 3395 | // the four checks below overlap |
| 3396 | DistinctRecords result(MSEP_REC'\n'); |
| 3397 | |
| 3398 | // check all prefixes (also crossed with suffixes if allowed) |
| 3399 | std::string st = prefix_check_morph(word, start, len, in_compound, scratch); |
| 3400 | if (!st.empty()) { |
| 3401 | result.append(st); |
| 3402 | } |
| 3403 | |
| 3404 | // if still not found check all suffixes |
| 3405 | st = suffix_check_morph(word, start, len, 0, nullptr, scratch, '\0', needflag, in_compound); |
| 3406 | if (!st.empty()) { |
| 3407 | result.append(st); |
| 3408 | } |
| 3409 | |
| 3410 | if (havecontclass) { |
| 3411 | sfx = nullptr; |
| 3412 | pfx = nullptr; |
| 3413 | // if still not found check all two-level suffixes |
| 3414 | st = suffix_check_twosfx_morph(word, start, len, 0, nullptr, scratch, needflag); |
| 3415 | if (!st.empty()) { |
| 3416 | result.append(st); |
| 3417 | } |
| 3418 | |
| 3419 | // if still not found check all two-level suffixes |
| 3420 | st = prefix_check_twosfx_morph(word, start, len, IN_CPD_NOT0, scratch, needflag); |
| 3421 | if (!st.empty()) { |
| 3422 | result.append(st); |
| 3423 | } |
| 3424 | } |
| 3425 | |
| 3426 | return result.join(); |
| 3427 | } |
| 3428 | |
| 3429 | // morphcmp(): compare MORPH_DERI_SFX, MORPH_INFL_SFX and MORPH_TERM_SFX fields |
| 3430 | // in the first line of the inputs |
| 3431 | // return 0, if inputs equal |
| 3432 | // return 1, if inputs may equal with a secondary suffix |
| 3433 | // otherwise return -1 |
| 3434 | static int morphcmp(const char* s, const char* t) { |
| 3435 | bool se = false, te = false; |
| 3436 | const char* sl; |
| 3437 | const char* tl; |
| 3438 | const char* olds; |
| 3439 | const char* oldt; |
| 3440 | if (!s || !t) |
| 3441 | return 1; |
| 3442 | olds = s; |
| 3443 | sl = strchr(s, '\n'); |
| 3444 | s = strstr(s, MORPH_DERI_SFX"ds:"); |
| 3445 | if (!s || (sl && sl < s)) |
| 3446 | s = strstr(olds, MORPH_INFL_SFX"is:"); |
| 3447 | if (!s || (sl && sl < s)) { |
| 3448 | s = strstr(olds, MORPH_TERM_SFX"ts:"); |
| 3449 | olds = nullptr; |
| 3450 | } |
| 3451 | oldt = t; |
| 3452 | tl = strchr(t, '\n'); |
| 3453 | t = strstr(t, MORPH_DERI_SFX"ds:"); |
| 3454 | if (!t || (tl && tl < t)) |
| 3455 | t = strstr(oldt, MORPH_INFL_SFX"is:"); |
| 3456 | if (!t || (tl && tl < t)) |
| 3457 | t = strstr(oldt, MORPH_TERM_SFX"ts:"); |
| 3458 | while (s && t && (!sl || sl > s) && (!tl || tl > t)) { |
| 3459 | s += MORPH_TAG_LENstrlen("st:"); |
| 3460 | t += MORPH_TAG_LENstrlen("st:"); |
| 3461 | se = false; |
| 3462 | te = false; |
| 3463 | // both values empty: matched at the shared field terminator. |
| 3464 | // skip the inner loop, which would step past the '\0'. |
| 3465 | if (*s == '\0' && *t == '\0') { |
| 3466 | se = true; |
| 3467 | te = true; |
| 3468 | } |
| 3469 | while ((*s == *t) && !se && !te) { |
| 3470 | s++; |
| 3471 | t++; |
| 3472 | switch (*s) { |
| 3473 | case ' ': |
| 3474 | case '\n': |
| 3475 | case '\t': |
| 3476 | case '\0': |
| 3477 | se = true; |
| 3478 | } |
| 3479 | switch (*t) { |
| 3480 | case ' ': |
| 3481 | case '\n': |
| 3482 | case '\t': |
| 3483 | case '\0': |
| 3484 | te = true; |
| 3485 | } |
| 3486 | } |
| 3487 | if (!se || !te) { |
| 3488 | // not terminal suffix difference |
| 3489 | if (olds) |
| 3490 | return -1; |
| 3491 | return 1; |
| 3492 | } |
| 3493 | olds = s; |
| 3494 | s = strstr(s, MORPH_DERI_SFX"ds:"); |
| 3495 | if (!s || (sl && sl < s)) |
| 3496 | s = strstr(olds, MORPH_INFL_SFX"is:"); |
| 3497 | if (!s || (sl && sl < s)) { |
| 3498 | s = strstr(olds, MORPH_TERM_SFX"ts:"); |
| 3499 | olds = nullptr; |
| 3500 | } |
| 3501 | oldt = t; |
| 3502 | t = strstr(t, MORPH_DERI_SFX"ds:"); |
| 3503 | if (!t || (tl && tl < t)) |
| 3504 | t = strstr(oldt, MORPH_INFL_SFX"is:"); |
| 3505 | if (!t || (tl && tl < t)) |
| 3506 | t = strstr(oldt, MORPH_TERM_SFX"ts:"); |
| 3507 | } |
| 3508 | if (!s && !t && se && te) |
| 3509 | return 0; |
| 3510 | return 1; |
| 3511 | } |
| 3512 | |
| 3513 | std::string AffixMgr::morphgen(const char* ts, |
| 3514 | int wl, |
| 3515 | const unsigned short* ap, |
| 3516 | unsigned short al, |
| 3517 | const char* morph, |
| 3518 | const char* targetmorph, |
| 3519 | int level, |
| 3520 | const FLAGunsigned short avoidflag) { |
| 3521 | // handle suffixes |
| 3522 | if (!morph) |
| 3523 | return {}; |
| 3524 | |
| 3525 | // check substandard flag |
| 3526 | if (TESTAFF(ap, substandard, al)(std::binary_search(ap, ap + al, substandard))) |
| 3527 | return {}; |
| 3528 | |
| 3529 | if (morphcmp(morph, targetmorph) == 0) |
| 3530 | return ts; |
| 3531 | |
| 3532 | size_t stemmorphcatpos; |
| 3533 | std::string mymorph; |
| 3534 | |
| 3535 | // use input suffix fields, if exist |
| 3536 | if (strstr(morph, MORPH_INFL_SFX"is:") || strstr(morph, MORPH_DERI_SFX"ds:")) { |
| 3537 | mymorph.assign(morph); |
| 3538 | mymorph.push_back(MSEP_FLD' '); |
| 3539 | stemmorphcatpos = mymorph.size(); |
| 3540 | } else { |
| 3541 | stemmorphcatpos = std::string::npos; |
| 3542 | } |
| 3543 | |
| 3544 | for (int i = 0; i < al; i++) { |
| 3545 | // A derivational suffix whose own continuation class re-enables the same |
| 3546 | // suffix would otherwise be applied twice here, doubling the ending (for |
| 3547 | // example the Hungarian -szeru adjective suffix producing peldaszeruszeru). |
| 3548 | // Skip the flag that was just applied one level up. |
| 3549 | if (avoidflag != FLAG_NULL0x00 && ap[i] == avoidflag) |
| 3550 | continue; |
| 3551 | const auto c = (unsigned char)(ap[i] & 0x00FF); |
| 3552 | SfxEntry* sptr = sFlag[c]; |
| 3553 | while (sptr) { |
| 3554 | if (sptr->getFlag() == ap[i] && sptr->getMorph() && |
| 3555 | ((sptr->getContLen() == 0) || |
| 3556 | // don't generate forms with substandard affixes |
| 3557 | !TESTAFF(sptr->getCont(), substandard, sptr->getContLen())(std::binary_search(sptr->getCont(), sptr->getCont() + sptr ->getContLen(), substandard)))) { |
| 3558 | const char* stemmorph; |
| 3559 | if (stemmorphcatpos != std::string::npos) { |
| 3560 | mymorph.replace(stemmorphcatpos, std::string::npos, sptr->getMorph()); |
| 3561 | stemmorph = mymorph.c_str(); |
| 3562 | } else { |
| 3563 | stemmorph = sptr->getMorph(); |
| 3564 | } |
| 3565 | |
| 3566 | int cmp = morphcmp(stemmorph, targetmorph); |
| 3567 | |
| 3568 | if (cmp == 0) { |
| 3569 | std::string newword = sptr->add(ts, wl); |
| 3570 | if (!newword.empty()) { |
| 3571 | hentry* check = pHMgr->lookup(newword.c_str(), newword.size()); // XXX extra dic |
| 3572 | if (!check || !check->astr || |
| 3573 | !(TESTAFF(check->astr, forbiddenword, check->alen)(std::binary_search(check->astr, check->astr + check-> alen, forbiddenword)) || |
| 3574 | TESTAFF(check->astr, ONLYUPCASEFLAG, check->alen)(std::binary_search(check->astr, check->astr + check-> alen, 65511)))) { |
| 3575 | return newword; |
| 3576 | } |
| 3577 | } |
| 3578 | } |
| 3579 | |
| 3580 | // recursive call for secondary suffixes |
| 3581 | if ((level == 0) && (cmp == 1) && (sptr->getContLen() > 0) && |
| 3582 | !TESTAFF(sptr->getCont(), substandard, sptr->getContLen())(std::binary_search(sptr->getCont(), sptr->getCont() + sptr ->getContLen(), substandard))) { |
| 3583 | std::string newword = sptr->add(ts, wl); |
| 3584 | if (!newword.empty()) { |
| 3585 | std::string newword2 = |
| 3586 | morphgen(newword.c_str(), newword.size(), sptr->getCont(), |
| 3587 | sptr->getContLen(), stemmorph, targetmorph, 1, |
| 3588 | sptr->getFlag()); |
| 3589 | |
| 3590 | if (!newword2.empty()) { |
| 3591 | return newword2; |
| 3592 | } |
| 3593 | } |
| 3594 | } |
| 3595 | } |
| 3596 | sptr = sptr->getFlgNxt(); |
| 3597 | } |
| 3598 | } |
| 3599 | return { }; |
| 3600 | } |
| 3601 | |
| 3602 | namespace { |
| 3603 | // replaces strdup with ansi version |
| 3604 | char* mystrdup(const char* s) { |
| 3605 | char* d = nullptr; |
| 3606 | if (s) { |
| 3607 | size_t sl = strlen(s) + 1; |
| 3608 | d = new char[sl]; |
| 3609 | memcpy(d, s, sl); |
| 3610 | } |
| 3611 | return d; |
| 3612 | } |
| 3613 | } |
| 3614 | |
| 3615 | int AffixMgr::expand_rootword(struct guessword* wlst, |
| 3616 | int maxn, |
| 3617 | const char* ts, |
| 3618 | int wl, |
| 3619 | const unsigned short* ap, |
| 3620 | unsigned short al, |
| 3621 | const char* bad, |
| 3622 | int badl, |
| 3623 | const char* phon) { |
| 3624 | int nh = 0; |
| 3625 | // first add root word to list |
| 3626 | if ((nh < maxn) && |
| 3627 | !(al && ((needaffix && TESTAFF(ap, needaffix, al)(std::binary_search(ap, ap + al, needaffix))) || |
| 3628 | (onlyincompound && TESTAFF(ap, onlyincompound, al)(std::binary_search(ap, ap + al, onlyincompound)))))) { |
| 3629 | wlst[nh].word = mystrdup(ts); |
| 3630 | wlst[nh].allow = false; |
| 3631 | wlst[nh].orig = nullptr; |
| 3632 | nh++; |
| 3633 | // add special phonetic version |
| 3634 | if (phon && (nh < maxn)) { |
| 3635 | wlst[nh].word = mystrdup(phon); |
| 3636 | wlst[nh].allow = false; |
| 3637 | wlst[nh].orig = mystrdup(ts); |
| 3638 | nh++; |
| 3639 | } |
| 3640 | } |
| 3641 | |
| 3642 | // handle suffixes |
| 3643 | for (int i = 0; i < al; i++) { |
| 3644 | const auto c = (unsigned char)(ap[i] & 0x00FF); |
| 3645 | SfxEntry* sptr = sFlag[c]; |
| 3646 | while (sptr) { |
| 3647 | if ((sptr->getFlag() == ap[i]) && |
| 3648 | (!sptr->getKeyLen() || |
| 3649 | ((badl > sptr->getKeyLen()) && |
| 3650 | (strcmp(sptr->getAffix(), bad + badl - sptr->getKeyLen()) == 0))) && |
| 3651 | // check needaffix flag |
| 3652 | !(sptr->getCont() && |
| 3653 | ((needaffix && |
| 3654 | TESTAFF(sptr->getCont(), needaffix, sptr->getContLen())(std::binary_search(sptr->getCont(), sptr->getCont() + sptr ->getContLen(), needaffix))) || |
| 3655 | (circumfix && |
| 3656 | TESTAFF(sptr->getCont(), circumfix, sptr->getContLen())(std::binary_search(sptr->getCont(), sptr->getCont() + sptr ->getContLen(), circumfix))) || |
| 3657 | (onlyincompound && |
| 3658 | TESTAFF(sptr->getCont(), onlyincompound, sptr->getContLen())(std::binary_search(sptr->getCont(), sptr->getCont() + sptr ->getContLen(), onlyincompound)))))) { |
| 3659 | std::string newword = sptr->add(ts, wl); |
| 3660 | if (!newword.empty()) { |
| 3661 | if (nh < maxn) { |
| 3662 | wlst[nh].word = mystrdup(newword.c_str()); |
| 3663 | wlst[nh].allow = sptr->allowCross(); |
| 3664 | wlst[nh].orig = nullptr; |
| 3665 | nh++; |
| 3666 | // add special phonetic version |
| 3667 | if (phon && (nh < maxn)) { |
| 3668 | std::string prefix(phon); |
| 3669 | std::string key(sptr->getKey()); |
| 3670 | reverseword(key); |
| 3671 | prefix.append(key); |
| 3672 | wlst[nh].word = mystrdup(prefix.c_str()); |
| 3673 | wlst[nh].allow = false; |
| 3674 | wlst[nh].orig = mystrdup(newword.c_str()); |
| 3675 | nh++; |
| 3676 | } |
| 3677 | } |
| 3678 | } |
| 3679 | } |
| 3680 | sptr = sptr->getFlgNxt(); |
| 3681 | } |
| 3682 | } |
| 3683 | |
| 3684 | int n = nh; |
| 3685 | |
| 3686 | // handle cross products of prefixes and suffixes |
| 3687 | for (int j = 1; j < n; j++) |
| 3688 | if (wlst[j].allow) { |
| 3689 | for (int k = 0; k < al; k++) { |
| 3690 | const auto c = (unsigned char)(ap[k] & 0x00FF); |
| 3691 | PfxEntry* cptr = pFlag[c]; |
| 3692 | while (cptr) { |
| 3693 | if ((cptr->getFlag() == ap[k]) && cptr->allowCross() && |
| 3694 | (!cptr->getKeyLen() || |
| 3695 | ((badl > cptr->getKeyLen()) && |
| 3696 | (strncmp(cptr->getKey(), bad, cptr->getKeyLen()) == 0)))) { |
| 3697 | int l1 = strlen(wlst[j].word); |
| 3698 | std::string newword = cptr->add(wlst[j].word, l1); |
| 3699 | if (!newword.empty()) { |
| 3700 | if (nh < maxn) { |
| 3701 | wlst[nh].word = mystrdup(newword.c_str()); |
| 3702 | wlst[nh].allow = cptr->allowCross(); |
| 3703 | wlst[nh].orig = nullptr; |
| 3704 | nh++; |
| 3705 | } |
| 3706 | } |
| 3707 | } |
| 3708 | cptr = cptr->getFlgNxt(); |
| 3709 | } |
| 3710 | } |
| 3711 | } |
| 3712 | |
| 3713 | // now handle pure prefixes |
| 3714 | for (int m = 0; m < al; m++) { |
| 3715 | const auto c = (unsigned char)(ap[m] & 0x00FF); |
| 3716 | PfxEntry* ptr = pFlag[c]; |
| 3717 | while (ptr) { |
| 3718 | if ((ptr->getFlag() == ap[m]) && |
| 3719 | (!ptr->getKeyLen() || |
| 3720 | ((badl > ptr->getKeyLen()) && |
| 3721 | (strncmp(ptr->getKey(), bad, ptr->getKeyLen()) == 0))) && |
| 3722 | // check needaffix flag |
| 3723 | !(ptr->getCont() && |
| 3724 | ((needaffix && |
| 3725 | TESTAFF(ptr->getCont(), needaffix, ptr->getContLen())(std::binary_search(ptr->getCont(), ptr->getCont() + ptr ->getContLen(), needaffix))) || |
| 3726 | (circumfix && |
| 3727 | TESTAFF(ptr->getCont(), circumfix, ptr->getContLen())(std::binary_search(ptr->getCont(), ptr->getCont() + ptr ->getContLen(), circumfix))) || |
| 3728 | (onlyincompound && |
| 3729 | TESTAFF(ptr->getCont(), onlyincompound, ptr->getContLen())(std::binary_search(ptr->getCont(), ptr->getCont() + ptr ->getContLen(), onlyincompound)))))) { |
| 3730 | std::string newword = ptr->add(ts, wl); |
| 3731 | if (!newword.empty()) { |
| 3732 | if (nh < maxn) { |
| 3733 | wlst[nh].word = mystrdup(newword.c_str()); |
| 3734 | wlst[nh].allow = ptr->allowCross(); |
| 3735 | wlst[nh].orig = nullptr; |
| 3736 | nh++; |
| 3737 | } |
| 3738 | } |
| 3739 | } |
| 3740 | ptr = ptr->getFlgNxt(); |
| 3741 | } |
| 3742 | } |
| 3743 | |
| 3744 | return nh; |
| 3745 | } |
| 3746 | |
| 3747 | // return replacing table |
| 3748 | const std::vector<replentry>& AffixMgr::get_reptable() const { |
| 3749 | return pHMgr->get_reptable(); |
| 3750 | } |
| 3751 | |
| 3752 | // return iconv table |
| 3753 | RepList* AffixMgr::get_iconvtable() const { |
| 3754 | if (!iconvtable) |
| 3755 | return nullptr; |
| 3756 | return iconvtable; |
| 3757 | } |
| 3758 | |
| 3759 | // return oconv table |
| 3760 | RepList* AffixMgr::get_oconvtable() const { |
| 3761 | if (!oconvtable) |
| 3762 | return nullptr; |
| 3763 | return oconvtable; |
| 3764 | } |
| 3765 | |
| 3766 | // return replacing table |
| 3767 | struct phonetable* AffixMgr::get_phonetable() const { |
| 3768 | if (!phone) |
| 3769 | return nullptr; |
| 3770 | return phone; |
| 3771 | } |
| 3772 | |
| 3773 | // return character map table |
| 3774 | const std::vector<mapentry>& AffixMgr::get_maptable() const { |
| 3775 | return maptable; |
| 3776 | } |
| 3777 | |
| 3778 | // return character map table |
| 3779 | const std::vector<std::string>& AffixMgr::get_breaktable() const { |
| 3780 | return breaktable; |
| 3781 | } |
| 3782 | |
| 3783 | // return text encoding of dictionary |
| 3784 | const std::string& AffixMgr::get_encoding() { |
| 3785 | if (encoding.empty()) |
| 3786 | encoding = SPELL_ENCODING"ISO8859-1"; |
| 3787 | return encoding; |
| 3788 | } |
| 3789 | |
| 3790 | // return text encoding of dictionary |
| 3791 | int AffixMgr::get_langnum() const { |
| 3792 | return langnum; |
| 3793 | } |
| 3794 | |
| 3795 | // return double prefix option |
| 3796 | int AffixMgr::get_complexprefixes() const { |
| 3797 | return complexprefixes; |
| 3798 | } |
| 3799 | |
| 3800 | // return FULLSTRIP option |
| 3801 | int AffixMgr::get_fullstrip() const { |
| 3802 | return fullstrip; |
| 3803 | } |
| 3804 | |
| 3805 | FLAGunsigned short AffixMgr::get_keepcase() const { |
| 3806 | return keepcase; |
| 3807 | } |
| 3808 | |
| 3809 | FLAGunsigned short AffixMgr::get_forceucase() const { |
| 3810 | return forceucase; |
| 3811 | } |
| 3812 | |
| 3813 | FLAGunsigned short AffixMgr::get_warn() const { |
| 3814 | return warn; |
| 3815 | } |
| 3816 | |
| 3817 | int AffixMgr::get_forbidwarn() const { |
| 3818 | return forbidwarn; |
| 3819 | } |
| 3820 | |
| 3821 | int AffixMgr::get_checksharps() const { |
| 3822 | return checksharps; |
| 3823 | } |
| 3824 | |
| 3825 | std::string AffixMgr::encode_flag(unsigned short aflag) const { |
| 3826 | return pHMgr->encode_flag(aflag); |
| 3827 | } |
| 3828 | |
| 3829 | // return the preferred ignore string for suggestions |
| 3830 | const char* AffixMgr::get_ignore() const { |
| 3831 | if (ignorechars.empty()) |
| 3832 | return nullptr; |
| 3833 | return ignorechars.c_str(); |
| 3834 | } |
| 3835 | |
| 3836 | // return the preferred ignore string for suggestions |
| 3837 | const std::vector<w_char>& AffixMgr::get_ignore_utf16() const { |
| 3838 | return ignorechars_utf16; |
| 3839 | } |
| 3840 | |
| 3841 | // return the keyboard string for suggestions |
| 3842 | const std::string& AffixMgr::get_key_string() { |
| 3843 | if (keystring.empty()) |
| 3844 | keystring = SPELL_KEYSTRING"qwertyuiop|asdfghjkl|zxcvbnm"; |
| 3845 | return keystring; |
| 3846 | } |
| 3847 | |
| 3848 | // return the preferred try string for suggestions |
| 3849 | const std::string& AffixMgr::get_try_string() const { |
| 3850 | return trystring; |
| 3851 | } |
| 3852 | |
| 3853 | // return the preferred try string for suggestions |
| 3854 | const std::string& AffixMgr::get_wordchars() const { |
| 3855 | return wordchars; |
| 3856 | } |
| 3857 | |
| 3858 | const std::vector<w_char>& AffixMgr::get_wordchars_utf16() const { |
| 3859 | return wordchars_utf16; |
| 3860 | } |
| 3861 | |
| 3862 | // is there compounding? |
| 3863 | int AffixMgr::get_compound() const { |
| 3864 | return compoundflag || compoundbegin || !defcpdtable.empty(); |
| 3865 | } |
| 3866 | |
| 3867 | // return the compound words control flag |
| 3868 | FLAGunsigned short AffixMgr::get_compoundflag() const { |
| 3869 | return compoundflag; |
| 3870 | } |
| 3871 | |
| 3872 | // return the forbidden words control flag |
| 3873 | FLAGunsigned short AffixMgr::get_forbiddenword() const { |
| 3874 | return forbiddenword; |
| 3875 | } |
| 3876 | |
| 3877 | // return the forbidden words control flag |
| 3878 | FLAGunsigned short AffixMgr::get_nosuggest() const { |
| 3879 | return nosuggest; |
| 3880 | } |
| 3881 | |
| 3882 | // return the forbidden words control flag |
| 3883 | FLAGunsigned short AffixMgr::get_nongramsuggest() const { |
| 3884 | return nongramsuggest; |
| 3885 | } |
| 3886 | |
| 3887 | // return the substandard root/affix control flag |
| 3888 | FLAGunsigned short AffixMgr::get_substandard() const { |
| 3889 | return substandard; |
| 3890 | } |
| 3891 | |
| 3892 | // return the forbidden words flag modify flag |
| 3893 | FLAGunsigned short AffixMgr::get_needaffix() const { |
| 3894 | return needaffix; |
| 3895 | } |
| 3896 | |
| 3897 | // return the circumfix flag |
| 3898 | FLAGunsigned short AffixMgr::get_circumfix() const { |
| 3899 | return circumfix; |
| 3900 | } |
| 3901 | |
| 3902 | // return the onlyincompound flag |
| 3903 | FLAGunsigned short AffixMgr::get_onlyincompound() const { |
| 3904 | return onlyincompound; |
| 3905 | } |
| 3906 | |
| 3907 | // return the value of suffix |
| 3908 | const std::string& AffixMgr::get_version() const { |
| 3909 | return version; |
| 3910 | } |
| 3911 | |
| 3912 | // utility method to look up root words in hash table |
| 3913 | struct hentry* AffixMgr::lookup(const char* word, size_t len) { |
| 3914 | struct hentry* he = nullptr; |
| 3915 | for (size_t i = 0; i < alldic.size() && !he; ++i) { |
| 3916 | he = alldic[i]->lookup(word, len); |
| 3917 | } |
| 3918 | return he; |
| 3919 | } |
| 3920 | |
| 3921 | // return the value of suffix |
| 3922 | int AffixMgr::have_contclass() const { |
| 3923 | return havecontclass; |
| 3924 | } |
| 3925 | |
| 3926 | // return utf8 |
| 3927 | int AffixMgr::get_utf8() const { |
| 3928 | return utf8; |
| 3929 | } |
| 3930 | |
| 3931 | int AffixMgr::get_maxngramsugs() const { |
| 3932 | return maxngramsugs; |
| 3933 | } |
| 3934 | |
| 3935 | int AffixMgr::get_maxcpdsugs() const { |
| 3936 | return maxcpdsugs; |
| 3937 | } |
| 3938 | |
| 3939 | int AffixMgr::get_maxdiff() const { |
| 3940 | return maxdiff; |
| 3941 | } |
| 3942 | |
| 3943 | int AffixMgr::get_onlymaxdiff() const { |
| 3944 | return onlymaxdiff; |
| 3945 | } |
| 3946 | |
| 3947 | // return nosplitsugs |
| 3948 | int AffixMgr::get_nosplitsugs() const { |
| 3949 | return nosplitsugs; |
| 3950 | } |
| 3951 | |
| 3952 | // return sugswithdots |
| 3953 | int AffixMgr::get_sugswithdots() const { |
| 3954 | return sugswithdots; |
| 3955 | } |
| 3956 | |
| 3957 | /* parse flag */ |
| 3958 | bool AffixMgr::parse_flag(const std::string& line, unsigned short* out, FileMgr* af) { |
| 3959 | if (*out != FLAG_NULL0x00 && !(*out >= DEFAULTFLAGS65510)) { |
| 3960 | HUNSPELL_WARNING( |
| 3961 | stderrstderr, |
| 3962 | "error: line %d: multiple definitions of an affix file parameter\n", |
| 3963 | af->getlinenum()); |
| 3964 | return false; |
| 3965 | } |
| 3966 | std::string s; |
| 3967 | if (!parse_string(line, s, af->getlinenum())) |
| 3968 | return false; |
| 3969 | *out = pHMgr->decode_flag(s); |
| 3970 | return true; |
| 3971 | } |
| 3972 | |
| 3973 | /* parse num */ |
| 3974 | bool AffixMgr::parse_num(const std::string& line, int* out, FileMgr* af) { |
| 3975 | if (*out != -1) { |
| 3976 | HUNSPELL_WARNING( |
| 3977 | stderrstderr, |
| 3978 | "error: line %d: multiple definitions of an affix file parameter\n", |
| 3979 | af->getlinenum()); |
| 3980 | return false; |
| 3981 | } |
| 3982 | std::string s; |
| 3983 | if (!parse_string(line, s, af->getlinenum())) |
| 3984 | return false; |
| 3985 | *out = atoi(s.c_str()); |
| 3986 | return true; |
| 3987 | } |
| 3988 | |
| 3989 | /* parse in the max syllablecount of compound words and */ |
| 3990 | bool AffixMgr::parse_cpdsyllable(const std::string& line, FileMgr* af) { |
| 3991 | int i = 0; |
| 3992 | int np = 0; |
| 3993 | auto iter = line.begin(), start_piece = mystrsep(line, iter); |
| 3994 | while (start_piece != line.end()) { |
| 3995 | switch (i) { |
| 3996 | case 0: { |
| 3997 | np++; |
| 3998 | break; |
| 3999 | } |
| 4000 | case 1: { |
| 4001 | cpdmaxsyllable = atoi(std::string(start_piece, iter).c_str()); |
| 4002 | np++; |
| 4003 | break; |
| 4004 | } |
| 4005 | case 2: { |
| 4006 | if (!utf8) { |
| 4007 | cpdvowels.assign(start_piece, iter); |
| 4008 | std::sort(cpdvowels.begin(), cpdvowels.end()); |
| 4009 | } else { |
| 4010 | std::string piece(start_piece, iter); |
| 4011 | u8_u16(cpdvowels_utf16, piece); |
| 4012 | std::sort(cpdvowels_utf16.begin(), cpdvowels_utf16.end()); |
| 4013 | } |
| 4014 | np++; |
| 4015 | break; |
| 4016 | } |
| 4017 | default: |
| 4018 | break; |
| 4019 | } |
| 4020 | ++i; |
| 4021 | start_piece = mystrsep(line, iter); |
| 4022 | } |
| 4023 | if (np < 2) { |
| 4024 | HUNSPELL_WARNING(stderrstderr, |
| 4025 | "error: line %d: missing compoundsyllable information\n", |
| 4026 | af->getlinenum()); |
| 4027 | return false; |
| 4028 | } |
| 4029 | if (np == 2) |
| 4030 | cpdvowels = "AEIOUaeiou"; |
| 4031 | return true; |
| 4032 | } |
| 4033 | |
| 4034 | bool AffixMgr::parse_convtable(const std::string& line, |
| 4035 | FileMgr* af, |
| 4036 | RepList** rl, |
| 4037 | const std::string& keyword) { |
| 4038 | if (*rl) { |
| 4039 | HUNSPELL_WARNING(stderrstderr, "error: line %d: multiple table definitions\n", |
| 4040 | af->getlinenum()); |
| 4041 | return false; |
| 4042 | } |
| 4043 | int i = 0; |
| 4044 | int np = 0; |
| 4045 | int numrl = 0; |
| 4046 | auto iter = line.begin(), start_piece = mystrsep(line, iter); |
| 4047 | while (start_piece != line.end()) { |
| 4048 | switch (i) { |
| 4049 | case 0: { |
| 4050 | np++; |
| 4051 | break; |
| 4052 | } |
| 4053 | case 1: { |
| 4054 | numrl = atoi(std::string(start_piece, iter).c_str()); |
| 4055 | if (numrl < 1) { |
| 4056 | HUNSPELL_WARNING(stderrstderr, "error: line %d: incorrect entry number\n", |
| 4057 | af->getlinenum()); |
| 4058 | return false; |
| 4059 | } |
| 4060 | *rl = new RepList(numrl); |
| 4061 | if (!*rl) |
| 4062 | return false; |
| 4063 | np++; |
| 4064 | break; |
| 4065 | } |
| 4066 | default: |
| 4067 | break; |
| 4068 | } |
| 4069 | ++i; |
| 4070 | start_piece = mystrsep(line, iter); |
| 4071 | } |
| 4072 | if (np != 2) { |
| 4073 | HUNSPELL_WARNING(stderrstderr, "error: line %d: missing data\n", |
| 4074 | af->getlinenum()); |
| 4075 | return false; |
| 4076 | } |
| 4077 | |
| 4078 | /* now parse the num lines to read in the remainder of the table */ |
| 4079 | for (int j = 0; j < numrl; j++) { |
| 4080 | std::string nl; |
| 4081 | if (!af->getline(nl)) |
| 4082 | return false; |
| 4083 | mychomp(nl); |
| 4084 | i = 0; |
| 4085 | std::string pattern; |
| 4086 | std::string pattern2; |
| 4087 | iter = nl.begin(); |
| 4088 | start_piece = mystrsep(nl, iter); |
| 4089 | while (start_piece != nl.end()) { |
| 4090 | { |
| 4091 | switch (i) { |
| 4092 | case 0: { |
| 4093 | if (nl.compare(start_piece - nl.begin(), keyword.size(), keyword, 0, keyword.size()) != 0) { |
| 4094 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4095 | af->getlinenum()); |
| 4096 | delete *rl; |
| 4097 | *rl = nullptr; |
| 4098 | return false; |
| 4099 | } |
| 4100 | break; |
| 4101 | } |
| 4102 | case 1: { |
| 4103 | pattern.assign(start_piece, iter); |
| 4104 | break; |
| 4105 | } |
| 4106 | case 2: { |
| 4107 | pattern2.assign(start_piece, iter); |
| 4108 | break; |
| 4109 | } |
| 4110 | default: |
| 4111 | break; |
| 4112 | } |
| 4113 | ++i; |
| 4114 | } |
| 4115 | start_piece = mystrsep(nl, iter); |
| 4116 | } |
| 4117 | if (pattern.empty() || pattern2.empty()) { |
| 4118 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4119 | af->getlinenum()); |
| 4120 | return false; |
| 4121 | } |
| 4122 | |
| 4123 | (*rl)->add(pattern, pattern2); |
| 4124 | } |
| 4125 | return true; |
| 4126 | } |
| 4127 | |
| 4128 | /* parse in the typical fault correcting table */ |
| 4129 | bool AffixMgr::parse_phonetable(const std::string& line, FileMgr* af) { |
| 4130 | if (phone) { |
| 4131 | HUNSPELL_WARNING(stderrstderr, "error: line %d: multiple table definitions\n", |
| 4132 | af->getlinenum()); |
| 4133 | return false; |
| 4134 | } |
| 4135 | std::unique_ptr<phonetable> new_phone; |
| 4136 | int num = -1; |
| 4137 | int i = 0; |
| 4138 | int np = 0; |
| 4139 | auto iter = line.begin(), start_piece = mystrsep(line, iter); |
| 4140 | while (start_piece != line.end()) { |
| 4141 | switch (i) { |
| 4142 | case 0: { |
| 4143 | np++; |
| 4144 | break; |
| 4145 | } |
| 4146 | case 1: { |
| 4147 | num = atoi(std::string(start_piece, iter).c_str()); |
| 4148 | if (num < 1) { |
| 4149 | HUNSPELL_WARNING(stderrstderr, "error: line %d: bad entry number\n", |
| 4150 | af->getlinenum()); |
| 4151 | return false; |
| 4152 | } |
| 4153 | new_phone = std::make_unique<phonetable>(); |
| 4154 | new_phone->utf8 = (char)utf8; |
| 4155 | np++; |
| 4156 | break; |
| 4157 | } |
| 4158 | default: |
| 4159 | break; |
| 4160 | } |
| 4161 | ++i; |
| 4162 | start_piece = mystrsep(line, iter); |
| 4163 | } |
| 4164 | if (np != 2) { |
| 4165 | HUNSPELL_WARNING(stderrstderr, "error: line %d: missing data\n", |
| 4166 | af->getlinenum()); |
| 4167 | return false; |
| 4168 | } |
| 4169 | |
| 4170 | /* now parse the phone->num lines to read in the remainder of the table */ |
| 4171 | for (int j = 0; j < num; ++j) { |
| 4172 | std::string nl; |
| 4173 | if (!af->getline(nl)) |
| 4174 | return false; |
| 4175 | mychomp(nl); |
| 4176 | i = 0; |
| 4177 | const size_t old_size = new_phone->rules.size(); |
| 4178 | iter = nl.begin(); |
| 4179 | start_piece = mystrsep(nl, iter); |
| 4180 | while (start_piece != nl.end()) { |
| 4181 | { |
| 4182 | switch (i) { |
| 4183 | case 0: { |
| 4184 | if (nl.compare(start_piece - nl.begin(), 5, "PHONE", 5) != 0) { |
| 4185 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4186 | af->getlinenum()); |
| 4187 | return false; |
| 4188 | } |
| 4189 | break; |
| 4190 | } |
| 4191 | case 1: { |
| 4192 | new_phone->rules.emplace_back(start_piece, iter); |
| 4193 | break; |
| 4194 | } |
| 4195 | case 2: { |
| 4196 | new_phone->rules.emplace_back(start_piece, iter); |
| 4197 | mystrrep(new_phone->rules.back(), "_", ""); |
| 4198 | break; |
| 4199 | } |
| 4200 | default: |
| 4201 | break; |
| 4202 | } |
| 4203 | ++i; |
| 4204 | } |
| 4205 | start_piece = mystrsep(nl, iter); |
| 4206 | } |
| 4207 | if (new_phone->rules.size() != old_size + 2) { |
| 4208 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4209 | af->getlinenum()); |
| 4210 | return false; |
| 4211 | } |
| 4212 | } |
| 4213 | new_phone->rules.emplace_back(""); |
| 4214 | new_phone->rules.emplace_back(""); |
| 4215 | init_phonet_hash(*new_phone); |
| 4216 | phone = new_phone.release(); |
| 4217 | return true; |
| 4218 | } |
| 4219 | |
| 4220 | /* parse in the checkcompoundpattern table */ |
| 4221 | bool AffixMgr::parse_checkcpdtable(const std::string& line, FileMgr* af) { |
| 4222 | if (parsedcheckcpd) { |
| 4223 | HUNSPELL_WARNING(stderrstderr, "error: line %d: multiple table definitions\n", |
| 4224 | af->getlinenum()); |
| 4225 | return false; |
| 4226 | } |
| 4227 | parsedcheckcpd = true; |
| 4228 | int numcheckcpd = -1; |
| 4229 | int i = 0; |
| 4230 | int np = 0; |
| 4231 | auto iter = line.begin(), start_piece = mystrsep(line, iter); |
| 4232 | while (start_piece != line.end()) { |
| 4233 | switch (i) { |
| 4234 | case 0: { |
| 4235 | np++; |
| 4236 | break; |
| 4237 | } |
| 4238 | case 1: { |
| 4239 | numcheckcpd = atoi(std::string(start_piece, iter).c_str()); |
| 4240 | if (numcheckcpd < 1) { |
| 4241 | HUNSPELL_WARNING(stderrstderr, "error: line %d: bad entry number\n", |
| 4242 | af->getlinenum()); |
| 4243 | return false; |
| 4244 | } |
| 4245 | checkcpdtable.reserve(std::min(numcheckcpd, 16384)); |
| 4246 | np++; |
| 4247 | break; |
| 4248 | } |
| 4249 | default: |
| 4250 | break; |
| 4251 | } |
| 4252 | ++i; |
| 4253 | start_piece = mystrsep(line, iter); |
| 4254 | } |
| 4255 | if (np != 2) { |
| 4256 | HUNSPELL_WARNING(stderrstderr, "error: line %d: missing data\n", |
| 4257 | af->getlinenum()); |
| 4258 | return false; |
| 4259 | } |
| 4260 | |
| 4261 | /* now parse the numcheckcpd lines to read in the remainder of the table */ |
| 4262 | for (int j = 0; j < numcheckcpd; ++j) { |
| 4263 | std::string nl; |
| 4264 | if (!af->getline(nl)) |
| 4265 | return false; |
| 4266 | mychomp(nl); |
| 4267 | i = 0; |
| 4268 | checkcpdtable.emplace_back(); |
| 4269 | iter = nl.begin(); |
| 4270 | start_piece = mystrsep(nl, iter); |
| 4271 | while (start_piece != nl.end()) { |
| 4272 | switch (i) { |
| 4273 | case 0: { |
| 4274 | if (nl.compare(start_piece - nl.begin(), 20, "CHECKCOMPOUNDPATTERN", 20) != 0) { |
| 4275 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4276 | af->getlinenum()); |
| 4277 | checkcpdtable.clear(); |
| 4278 | return false; |
| 4279 | } |
| 4280 | break; |
| 4281 | } |
| 4282 | case 1: { |
| 4283 | checkcpdtable.back().pattern.assign(start_piece, iter); |
| 4284 | size_t slash_pos = checkcpdtable.back().pattern.find('/'); |
| 4285 | if (slash_pos != std::string::npos) { |
| 4286 | std::string chunk(checkcpdtable.back().pattern, slash_pos + 1); |
| 4287 | checkcpdtable.back().pattern.resize(slash_pos); |
| 4288 | checkcpdtable.back().cond = pHMgr->decode_flag(chunk); |
| 4289 | } |
| 4290 | break; |
| 4291 | } |
| 4292 | case 2: { |
| 4293 | checkcpdtable.back().pattern2.assign(start_piece, iter); |
| 4294 | size_t slash_pos = checkcpdtable.back().pattern2.find('/'); |
| 4295 | if (slash_pos != std::string::npos) { |
| 4296 | std::string chunk(checkcpdtable.back().pattern2, slash_pos + 1); |
| 4297 | checkcpdtable.back().pattern2.resize(slash_pos); |
| 4298 | checkcpdtable.back().cond2 = pHMgr->decode_flag(chunk); |
| 4299 | } |
| 4300 | break; |
| 4301 | } |
| 4302 | case 3: { |
| 4303 | checkcpdtable.back().pattern3.assign(start_piece, iter); |
| 4304 | simplifiedcpd = 1; |
| 4305 | break; |
| 4306 | } |
| 4307 | default: |
| 4308 | break; |
| 4309 | } |
| 4310 | i++; |
| 4311 | start_piece = mystrsep(nl, iter); |
| 4312 | } |
| 4313 | } |
| 4314 | return true; |
| 4315 | } |
| 4316 | |
| 4317 | /* parse in the compound rule table */ |
| 4318 | bool AffixMgr::parse_defcpdtable(const std::string& line, FileMgr* af) { |
| 4319 | if (parseddefcpd) { |
| 4320 | HUNSPELL_WARNING(stderrstderr, "error: line %d: multiple table definitions\n", |
| 4321 | af->getlinenum()); |
| 4322 | return false; |
| 4323 | } |
| 4324 | parseddefcpd = true; |
| 4325 | int numdefcpd = -1; |
| 4326 | int i = 0; |
| 4327 | int np = 0; |
| 4328 | auto iter = line.begin(), start_piece = mystrsep(line, iter); |
| 4329 | while (start_piece != line.end()) { |
| 4330 | switch (i) { |
| 4331 | case 0: { |
| 4332 | np++; |
| 4333 | break; |
| 4334 | } |
| 4335 | case 1: { |
| 4336 | numdefcpd = atoi(std::string(start_piece, iter).c_str()); |
| 4337 | if (numdefcpd < 1) { |
| 4338 | HUNSPELL_WARNING(stderrstderr, "error: line %d: bad entry number\n", |
| 4339 | af->getlinenum()); |
| 4340 | return false; |
| 4341 | } |
| 4342 | defcpdtable.reserve(std::min(numdefcpd, 16384)); |
| 4343 | np++; |
| 4344 | break; |
| 4345 | } |
| 4346 | default: |
| 4347 | break; |
| 4348 | } |
| 4349 | ++i; |
| 4350 | start_piece = mystrsep(line, iter); |
| 4351 | } |
| 4352 | if (np != 2) { |
| 4353 | HUNSPELL_WARNING(stderrstderr, "error: line %d: missing data\n", |
| 4354 | af->getlinenum()); |
| 4355 | return false; |
| 4356 | } |
| 4357 | |
| 4358 | /* now parse the numdefcpd lines to read in the remainder of the table */ |
| 4359 | for (int j = 0; j < numdefcpd; ++j) { |
| 4360 | std::string nl; |
| 4361 | if (!af->getline(nl)) |
| 4362 | return false; |
| 4363 | mychomp(nl); |
| 4364 | i = 0; |
| 4365 | defcpdtable.emplace_back(); |
| 4366 | iter = nl.begin(); |
| 4367 | start_piece = mystrsep(nl, iter); |
| 4368 | while (start_piece != nl.end()) { |
| 4369 | switch (i) { |
| 4370 | case 0: { |
| 4371 | if (nl.compare(start_piece - nl.begin(), 12, "COMPOUNDRULE", 12) != 0) { |
| 4372 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4373 | af->getlinenum()); |
| 4374 | numdefcpd = 0; |
| 4375 | return false; |
| 4376 | } |
| 4377 | break; |
| 4378 | } |
| 4379 | case 1: { // handle parenthesized flags |
| 4380 | if (std::find(start_piece, iter, '(') != iter) { |
| 4381 | for (auto k = start_piece; k != iter; ++k) { |
| 4382 | auto chb = k, che = k + 1; |
| 4383 | if (*k == '(') { |
| 4384 | auto parpos = std::find(k, iter, ')'); |
| 4385 | if (parpos != iter) { |
| 4386 | chb = k + 1; |
| 4387 | che = parpos; |
| 4388 | k = parpos; |
| 4389 | } |
| 4390 | } |
| 4391 | |
| 4392 | if (*chb == '*' || *chb == '?') { |
| 4393 | defcpdtable.back().push_back((FLAGunsigned short)*chb); |
| 4394 | } else { |
| 4395 | pHMgr->decode_flags(defcpdtable.back(), std::string(chb, che), af); |
| 4396 | } |
| 4397 | } |
| 4398 | } else { |
| 4399 | pHMgr->decode_flags(defcpdtable.back(), std::string(start_piece, iter), af); |
| 4400 | } |
| 4401 | break; |
| 4402 | } |
| 4403 | default: |
| 4404 | break; |
| 4405 | } |
| 4406 | ++i; |
| 4407 | start_piece = mystrsep(nl, iter); |
| 4408 | } |
| 4409 | if (defcpdtable.back().empty()) { |
| 4410 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4411 | af->getlinenum()); |
| 4412 | return false; |
| 4413 | } |
| 4414 | } |
| 4415 | return true; |
| 4416 | } |
| 4417 | |
| 4418 | /* parse in the character map table */ |
| 4419 | bool AffixMgr::parse_maptable(const std::string& line, FileMgr* af) { |
| 4420 | if (parsedmaptable) { |
| 4421 | HUNSPELL_WARNING(stderrstderr, "error: line %d: multiple table definitions\n", |
| 4422 | af->getlinenum()); |
| 4423 | return false; |
| 4424 | } |
| 4425 | parsedmaptable = true; |
| 4426 | int nummap = -1; |
| 4427 | int i = 0; |
| 4428 | int np = 0; |
| 4429 | auto iter = line.begin(), start_piece = mystrsep(line, iter); |
| 4430 | while (start_piece != line.end()) { |
| 4431 | switch (i) { |
| 4432 | case 0: { |
| 4433 | np++; |
| 4434 | break; |
| 4435 | } |
| 4436 | case 1: { |
| 4437 | nummap = atoi(std::string(start_piece, iter).c_str()); |
| 4438 | if (nummap < 1) { |
| 4439 | HUNSPELL_WARNING(stderrstderr, "error: line %d: bad entry number\n", |
| 4440 | af->getlinenum()); |
| 4441 | return false; |
| 4442 | } |
| 4443 | maptable.reserve(std::min(nummap, 16384)); |
| 4444 | np++; |
| 4445 | break; |
| 4446 | } |
| 4447 | default: |
| 4448 | break; |
| 4449 | } |
| 4450 | ++i; |
| 4451 | start_piece = mystrsep(line, iter); |
| 4452 | } |
| 4453 | if (np != 2) { |
| 4454 | HUNSPELL_WARNING(stderrstderr, "error: line %d: missing data\n", |
| 4455 | af->getlinenum()); |
| 4456 | return false; |
| 4457 | } |
| 4458 | |
| 4459 | /* now parse the nummap lines to read in the remainder of the table */ |
| 4460 | for (int j = 0; j < nummap; ++j) { |
| 4461 | std::string nl; |
| 4462 | if (!af->getline(nl)) |
| 4463 | return false; |
| 4464 | mychomp(nl); |
| 4465 | i = 0; |
| 4466 | maptable.emplace_back(); |
| 4467 | iter = nl.begin(); |
| 4468 | start_piece = mystrsep(nl, iter); |
| 4469 | while (start_piece != nl.end()) { |
| 4470 | switch (i) { |
| 4471 | case 0: { |
| 4472 | if (nl.compare(start_piece - nl.begin(), 3, "MAP", 3) != 0) { |
| 4473 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4474 | af->getlinenum()); |
| 4475 | nummap = 0; |
Value stored to 'nummap' is never read | |
| 4476 | return false; |
| 4477 | } |
| 4478 | break; |
| 4479 | } |
| 4480 | case 1: { |
| 4481 | for (auto k = start_piece; k != iter; ++k) { |
| 4482 | auto chb = k, che = k + 1; |
| 4483 | if (*k == '(') { |
| 4484 | auto parpos = std::find(k, iter, ')'); |
| 4485 | if (parpos != iter) { |
| 4486 | chb = k + 1; |
| 4487 | che = parpos; |
| 4488 | k = parpos; |
| 4489 | } |
| 4490 | } else { |
| 4491 | if (utf8 && (*k & 0xc0) == 0xc0) { |
| 4492 | ++k; |
| 4493 | while (k != iter && is_utf8_cont(*k)) |
| 4494 | ++k; |
| 4495 | che = k; |
| 4496 | --k; |
| 4497 | } |
| 4498 | } |
| 4499 | if (chb == che) { |
| 4500 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4501 | af->getlinenum()); |
| 4502 | } |
| 4503 | |
| 4504 | maptable.back().emplace_back(chb, che); |
| 4505 | } |
| 4506 | break; |
| 4507 | } |
| 4508 | default: |
| 4509 | break; |
| 4510 | } |
| 4511 | ++i; |
| 4512 | start_piece = mystrsep(nl, iter); |
| 4513 | } |
| 4514 | if (maptable.back().empty()) { |
| 4515 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4516 | af->getlinenum()); |
| 4517 | return false; |
| 4518 | } |
| 4519 | } |
| 4520 | return true; |
| 4521 | } |
| 4522 | |
| 4523 | /* parse in the word breakpoint table */ |
| 4524 | bool AffixMgr::parse_breaktable(const std::string& line, FileMgr* af) { |
| 4525 | if (parsedbreaktable) { |
| 4526 | HUNSPELL_WARNING(stderrstderr, "error: line %d: multiple table definitions\n", |
| 4527 | af->getlinenum()); |
| 4528 | return false; |
| 4529 | } |
| 4530 | parsedbreaktable = true; |
| 4531 | int numbreak = -1; |
| 4532 | int i = 0; |
| 4533 | int np = 0; |
| 4534 | auto iter = line.begin(), start_piece = mystrsep(line, iter); |
| 4535 | while (start_piece != line.end()) { |
| 4536 | switch (i) { |
| 4537 | case 0: { |
| 4538 | np++; |
| 4539 | break; |
| 4540 | } |
| 4541 | case 1: { |
| 4542 | numbreak = atoi(std::string(start_piece, iter).c_str()); |
| 4543 | if (numbreak < 0) { |
| 4544 | HUNSPELL_WARNING(stderrstderr, "error: line %d: bad entry number\n", |
| 4545 | af->getlinenum()); |
| 4546 | return false; |
| 4547 | } |
| 4548 | if (numbreak == 0) |
| 4549 | return true; |
| 4550 | breaktable.reserve(std::min(numbreak, 16384)); |
| 4551 | np++; |
| 4552 | break; |
| 4553 | } |
| 4554 | default: |
| 4555 | break; |
| 4556 | } |
| 4557 | ++i; |
| 4558 | start_piece = mystrsep(line, iter); |
| 4559 | } |
| 4560 | if (np != 2) { |
| 4561 | HUNSPELL_WARNING(stderrstderr, "error: line %d: missing data\n", |
| 4562 | af->getlinenum()); |
| 4563 | return false; |
| 4564 | } |
| 4565 | |
| 4566 | /* now parse the numbreak lines to read in the remainder of the table */ |
| 4567 | for (int j = 0; j < numbreak; ++j) { |
| 4568 | std::string nl; |
| 4569 | if (!af->getline(nl)) |
| 4570 | return false; |
| 4571 | mychomp(nl); |
| 4572 | i = 0; |
| 4573 | iter = nl.begin(); |
| 4574 | start_piece = mystrsep(nl, iter); |
| 4575 | while (start_piece != nl.end()) { |
| 4576 | switch (i) { |
| 4577 | case 0: { |
| 4578 | if (nl.compare(start_piece - nl.begin(), 5, "BREAK", 5) != 0) { |
| 4579 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4580 | af->getlinenum()); |
| 4581 | numbreak = 0; |
| 4582 | return false; |
| 4583 | } |
| 4584 | break; |
| 4585 | } |
| 4586 | case 1: { |
| 4587 | breaktable.emplace_back(start_piece, iter); |
| 4588 | break; |
| 4589 | } |
| 4590 | default: |
| 4591 | break; |
| 4592 | } |
| 4593 | ++i; |
| 4594 | start_piece = mystrsep(nl, iter); |
| 4595 | } |
| 4596 | } |
| 4597 | |
| 4598 | if (breaktable.size() != static_cast<size_t>(numbreak)) { |
| 4599 | HUNSPELL_WARNING(stderrstderr, "error: line %d: table is corrupt\n", |
| 4600 | af->getlinenum()); |
| 4601 | return false; |
| 4602 | } |
| 4603 | |
| 4604 | return true; |
| 4605 | } |
| 4606 | |
| 4607 | void AffixMgr::reverse_condition(std::string& piece) { |
| 4608 | if (piece.empty()) |
| 4609 | return; |
| 4610 | |
| 4611 | int neg = 0; |
| 4612 | // iterate backwards; k wraps to npos (SIZE_MAX) when decremented past 0 |
| 4613 | for (size_t k = piece.size() - 1; k != std::string::npos; --k) { |
| 4614 | switch (piece[k]) { |
| 4615 | case '[': { |
| 4616 | if (neg) |
| 4617 | piece[k + 1] = '['; |
| 4618 | else |
| 4619 | piece[k] = ']'; |
| 4620 | break; |
| 4621 | } |
| 4622 | case ']': { |
| 4623 | piece[k] = '['; |
| 4624 | if (neg) |
| 4625 | piece[k + 1] = '^'; |
| 4626 | neg = 0; |
| 4627 | break; |
| 4628 | } |
| 4629 | case '^': { |
| 4630 | if (piece[k + 1] == ']') |
| 4631 | neg = 1; |
| 4632 | else if (neg) |
| 4633 | piece[k + 1] = piece[k]; |
| 4634 | break; |
| 4635 | } |
| 4636 | default: { |
| 4637 | if (neg) |
| 4638 | piece[k + 1] = piece[k]; |
| 4639 | } |
| 4640 | } |
| 4641 | } |
| 4642 | } |
| 4643 | |
| 4644 | class entries_container { |
| 4645 | std::vector<AffEntry*> entries; |
| 4646 | AffixMgr* m_mgr; |
| 4647 | char m_at; |
| 4648 | public: |
| 4649 | entries_container(char at, AffixMgr* mgr) |
| 4650 | : m_mgr(mgr) |
| 4651 | , m_at(at) { |
| 4652 | } |
| 4653 | void release() { |
| 4654 | entries.clear(); |
| 4655 | } |
| 4656 | void initialize(int numents, |
| 4657 | char opts, unsigned short aflag) { |
| 4658 | entries.reserve(std::min(numents, 16384)); |
| 4659 | |
| 4660 | if (m_at == 'P') { |
| 4661 | entries.push_back(new PfxEntry(m_mgr)); |
| 4662 | } else { |
| 4663 | entries.push_back(new SfxEntry(m_mgr)); |
| 4664 | } |
| 4665 | |
| 4666 | entries.back()->opts = opts; |
| 4667 | entries.back()->aflag = aflag; |
| 4668 | } |
| 4669 | |
| 4670 | AffEntry* add_entry(char opts) { |
| 4671 | if (m_at == 'P') { |
| 4672 | entries.push_back(new PfxEntry(m_mgr)); |
| 4673 | } else { |
| 4674 | entries.push_back(new SfxEntry(m_mgr)); |
| 4675 | } |
| 4676 | AffEntry* ret = entries.back(); |
| 4677 | ret->opts = entries[0]->opts & opts; |
| 4678 | return ret; |
| 4679 | } |
| 4680 | |
| 4681 | AffEntry* first_entry() { return entries.empty() ? nullptr : entries[0]; } |
| 4682 | |
| 4683 | ~entries_container() { |
| 4684 | for (auto& entry : entries) { |
| 4685 | delete entry; |
| 4686 | } |
| 4687 | } |
| 4688 | |
| 4689 | std::vector<AffEntry*>::iterator begin() { return entries.begin(); } |
| 4690 | std::vector<AffEntry*>::iterator end() { return entries.end(); } |
| 4691 | }; |
| 4692 | |
| 4693 | bool AffixMgr::parse_affix(const std::string& line, |
| 4694 | const char at, |
| 4695 | FileMgr* af, |
| 4696 | char* dupflags) { |
| 4697 | int numents = 0; // number of AffEntry structures to parse |
| 4698 | |
| 4699 | unsigned short aflag = 0; // affix char identifier |
| 4700 | |
| 4701 | char ff = 0; |
| 4702 | char xprod = 0; |
| 4703 | int headerline = af->getlinenum(); |
| 4704 | entries_container affentries(at, this); |
| 4705 | |
| 4706 | int i = 0; |
| 4707 | |
| 4708 | // checking lines with bad syntax |
| 4709 | #ifdef DEBUG1 |
| 4710 | int basefieldnum = 0; |
| 4711 | #endif |
| 4712 | |
| 4713 | // split affix header line into pieces |
| 4714 | |
| 4715 | int np = 0; |
| 4716 | auto iter = line.begin(), start_piece = mystrsep(line, iter); |
| 4717 | while (start_piece != line.end()) { |
| 4718 | switch (i) { |
| 4719 | // piece 1 - is type of affix |
| 4720 | case 0: { |
| 4721 | np++; |
| 4722 | break; |
| 4723 | } |
| 4724 | |
| 4725 | // piece 2 - is affix char |
| 4726 | case 1: { |
| 4727 | np++; |
| 4728 | aflag = pHMgr->decode_flag(std::string(start_piece, iter)); |
| 4729 | if (((at == 'S') && (dupflags[aflag] & dupSFX(1 << 0))) || |
| 4730 | ((at == 'P') && (dupflags[aflag] & dupPFX(1 << 1)))) { |
| 4731 | HUNSPELL_WARNING( |
| 4732 | stderrstderr, |
| 4733 | "error: line %d: multiple definitions of an affix flag\n", |
| 4734 | af->getlinenum()); |
| 4735 | } |
| 4736 | dupflags[aflag] += (char)((at == 'S') ? dupSFX(1 << 0) : dupPFX(1 << 1)); |
| 4737 | break; |
| 4738 | } |
| 4739 | // piece 3 - is cross product indicator |
| 4740 | case 2: { |
| 4741 | np++; |
| 4742 | xprod = *start_piece; |
| 4743 | if (xprod == 'Y') |
| 4744 | ff = aeXPRODUCT(1 << 0); |
| 4745 | break; |
| 4746 | } |
| 4747 | |
| 4748 | // piece 4 - is number of affentries |
| 4749 | case 3: { |
| 4750 | np++; |
| 4751 | numents = atoi(std::string(start_piece, iter).c_str()); |
| 4752 | if ((numents <= 0) || ((std::numeric_limits<size_t>::max() / |
| 4753 | sizeof(AffEntry)) < static_cast<size_t>(numents))) { |
| 4754 | std::string err = pHMgr->encode_flag(aflag); |
| 4755 | HUNSPELL_WARNING(stderrstderr, "error: line %d: affix %s: bad entry number\n", |
| 4756 | af->getlinenum(), err.c_str()); |
| 4757 | return false; |
| 4758 | } |
| 4759 | |
| 4760 | char opts = ff; |
| 4761 | if (utf8) |
| 4762 | opts |= aeUTF8(1 << 1); |
| 4763 | if (pHMgr->is_aliasf()) |
| 4764 | opts |= aeALIASF(1 << 2); |
| 4765 | if (pHMgr->is_aliasm()) |
| 4766 | opts |= aeALIASM(1 << 3); |
| 4767 | affentries.initialize(numents, opts, aflag); |
| 4768 | } |
| 4769 | |
| 4770 | default: |
| 4771 | break; |
| 4772 | } |
| 4773 | ++i; |
| 4774 | start_piece = mystrsep(line, iter); |
| 4775 | } |
| 4776 | // check to make sure we parsed enough pieces |
| 4777 | if (np != 4) { |
| 4778 | std::string err = pHMgr->encode_flag(aflag); |
| 4779 | HUNSPELL_WARNING(stderrstderr, "error: line %d: affix %s: missing data\n", |
| 4780 | af->getlinenum(), err.c_str()); |
| 4781 | return false; |
| 4782 | } |
| 4783 | |
| 4784 | // now parse numents affentries for this affix |
| 4785 | AffEntry* entry = affentries.first_entry(); |
| 4786 | for (int ent = 0; ent < numents; ++ent) { |
| 4787 | std::string nl; |
| 4788 | if (!af->getline(nl)) |
| 4789 | return false; |
| 4790 | mychomp(nl); |
| 4791 | int ruleline = af->getlinenum(); |
| 4792 | |
| 4793 | iter = nl.begin(); |
| 4794 | i = 0; |
| 4795 | np = 0; |
| 4796 | |
| 4797 | // split line into pieces |
| 4798 | start_piece = mystrsep(nl, iter); |
| 4799 | while (start_piece != nl.end()) { |
| 4800 | switch (i) { |
| 4801 | // piece 1 - is type |
| 4802 | case 0: { |
| 4803 | np++; |
| 4804 | if (ent != 0) |
| 4805 | entry = affentries.add_entry((char)(aeXPRODUCT(1 << 0) | aeUTF8(1 << 1) | aeALIASF(1 << 2) | aeALIASM(1 << 3))); |
| 4806 | break; |
| 4807 | } |
| 4808 | |
| 4809 | // piece 2 - is affix char |
| 4810 | case 1: { |
| 4811 | np++; |
| 4812 | std::string chunk(start_piece, iter); |
| 4813 | if (pHMgr->decode_flag(chunk) != aflag) { |
| 4814 | std::string err = pHMgr->encode_flag(aflag); |
| 4815 | HUNSPELL_WARNING(stderrstderr, |
| 4816 | "error: line %d: affix %s is corrupt\n", |
| 4817 | af->getlinenum(), err.c_str()); |
| 4818 | return false; |
| 4819 | } |
| 4820 | |
| 4821 | if (ent != 0) { |
| 4822 | AffEntry* start_entry = affentries.first_entry(); |
| 4823 | entry->aflag = start_entry->aflag; |
| 4824 | } |
| 4825 | break; |
| 4826 | } |
| 4827 | |
| 4828 | // piece 3 - is string to strip or 0 for null |
| 4829 | case 2: { |
| 4830 | np++; |
| 4831 | entry->strip = std::string(start_piece, iter); |
| 4832 | if (complexprefixes) { |
| 4833 | if (utf8) |
| 4834 | reverseword_utf(entry->strip); |
| 4835 | else |
| 4836 | reverseword(entry->strip); |
| 4837 | } |
| 4838 | if (entry->strip.compare("0") == 0) { |
| 4839 | entry->strip.clear(); |
| 4840 | } |
| 4841 | break; |
| 4842 | } |
| 4843 | |
| 4844 | // piece 4 - is affix string or 0 for null |
| 4845 | case 3: { |
| 4846 | entry->morphcode = nullptr; |
| 4847 | entry->contclass = nullptr; |
| 4848 | entry->contclasslen = 0; |
| 4849 | np++; |
| 4850 | std::string::const_iterator dash = std::find(start_piece, iter, '/'); |
| 4851 | if (dash != iter) { |
| 4852 | entry->appnd = std::string(start_piece, dash); |
| 4853 | std::string dash_str(dash + 1, iter); |
| 4854 | |
| 4855 | if (!ignorechars.empty() && !has_no_ignored_chars(entry->appnd, ignorechars)) { |
| 4856 | if (utf8) { |
| 4857 | remove_ignored_chars_utf(entry->appnd, ignorechars_utf16); |
| 4858 | } else { |
| 4859 | remove_ignored_chars(entry->appnd, ignorechars); |
| 4860 | } |
| 4861 | } |
| 4862 | |
| 4863 | if (complexprefixes) { |
| 4864 | if (utf8) |
| 4865 | reverseword_utf(entry->appnd); |
| 4866 | else |
| 4867 | reverseword(entry->appnd); |
| 4868 | } |
| 4869 | |
| 4870 | if (pHMgr->is_aliasf()) { |
| 4871 | int index = atoi(dash_str.c_str()); |
| 4872 | entry->contclasslen = (unsigned short)pHMgr->get_aliasf( |
| 4873 | index, &(entry->contclass), af); |
| 4874 | if (!entry->contclasslen) |
| 4875 | HUNSPELL_WARNING(stderrstderr, |
| 4876 | "error: bad affix flag alias: \"%s\"\n", |
| 4877 | dash_str.c_str()); |
| 4878 | } else { |
| 4879 | entry->contclasslen = (unsigned short)pHMgr->decode_flags( |
| 4880 | &(entry->contclass), dash_str, af); |
| 4881 | std::sort(entry->contclass, entry->contclass + entry->contclasslen); |
| 4882 | } |
| 4883 | |
| 4884 | havecontclass = 1; |
| 4885 | for (unsigned short _i = 0; _i < entry->contclasslen; _i++) { |
| 4886 | contclasses[(entry->contclass)[_i]] = 1; |
| 4887 | } |
| 4888 | } else { |
| 4889 | entry->appnd = std::string(start_piece, iter); |
| 4890 | |
| 4891 | if (!ignorechars.empty() && !has_no_ignored_chars(entry->appnd, ignorechars)) { |
| 4892 | if (utf8) { |
| 4893 | remove_ignored_chars_utf(entry->appnd, ignorechars_utf16); |
| 4894 | } else { |
| 4895 | remove_ignored_chars(entry->appnd, ignorechars); |
| 4896 | } |
| 4897 | } |
| 4898 | |
| 4899 | if (complexprefixes) { |
| 4900 | if (utf8) |
| 4901 | reverseword_utf(entry->appnd); |
| 4902 | else |
| 4903 | reverseword(entry->appnd); |
| 4904 | } |
| 4905 | } |
| 4906 | |
| 4907 | if (entry->appnd.compare("0") == 0) { |
| 4908 | entry->appnd.clear(); |
| 4909 | } |
| 4910 | break; |
| 4911 | } |
| 4912 | |
| 4913 | // piece 5 - is the conditions descriptions |
| 4914 | case 4: { |
| 4915 | std::string chunk(start_piece, iter); |
| 4916 | np++; |
| 4917 | if (complexprefixes) { |
| 4918 | if (utf8) |
| 4919 | reverseword_utf(chunk); |
| 4920 | else |
| 4921 | reverseword(chunk); |
| 4922 | reverse_condition(chunk); |
| 4923 | } |
| 4924 | if (!entry->strip.empty() && chunk != "." && |
| 4925 | redundant_condition(at, entry->strip, chunk, |
| 4926 | af->getlinenum())) { |
| 4927 | chunk = "."; |
| 4928 | entry->opts |= aeREDUNDANTCOND(1 << 5); |
| 4929 | } |
| 4930 | if (at == 'S') { |
| 4931 | reverseword(chunk); |
| 4932 | reverse_condition(chunk); |
| 4933 | } |
| 4934 | if (encodeit(*entry, chunk)) |
| 4935 | return false; |
| 4936 | break; |
| 4937 | } |
| 4938 | |
| 4939 | case 5: { |
| 4940 | std::string chunk(start_piece, iter); |
| 4941 | np++; |
| 4942 | if (pHMgr->is_aliasm()) { |
| 4943 | int index = atoi(chunk.c_str()); |
| 4944 | entry->morphcode = pHMgr->get_aliasm(index); |
| 4945 | } else { |
| 4946 | if (complexprefixes) { // XXX - fix me for morph. gen. |
| 4947 | if (utf8) |
| 4948 | reverseword_utf(chunk); |
| 4949 | else |
| 4950 | reverseword(chunk); |
| 4951 | } |
| 4952 | // add the remaining of the line |
| 4953 | std::string::const_iterator end = nl.end(); |
| 4954 | if (iter != end) { |
| 4955 | chunk.append(iter, end); |
| 4956 | } |
| 4957 | entry->morphcode = mystrdup(chunk.c_str()); |
| 4958 | } |
| 4959 | break; |
| 4960 | } |
| 4961 | default: |
| 4962 | break; |
| 4963 | } |
| 4964 | i++; |
| 4965 | start_piece = mystrsep(nl, iter); |
| 4966 | } |
| 4967 | // check to make sure we parsed enough pieces |
| 4968 | if (np < 4) { |
| 4969 | std::string err = pHMgr->encode_flag(aflag); |
| 4970 | HUNSPELL_WARNING(stderrstderr, "error: line %d: affix %s is corrupt\n", |
| 4971 | af->getlinenum(), err.c_str()); |
| 4972 | return false; |
| 4973 | } |
| 4974 | |
| 4975 | entry->line = ruleline; |
| 4976 | entry->headerline = headerline; |
| 4977 | entry->xprod = xprod; |
| 4978 | |
| 4979 | #ifdef DEBUG1 |
| 4980 | // detect unnecessary fields, excepting comments |
| 4981 | if (basefieldnum) { |
| 4982 | int fieldnum = |
| 4983 | !(entry->morphcode) ? 5 : ((*(entry->morphcode) == '#') ? 5 : 6); |
| 4984 | if (fieldnum != basefieldnum) |
| 4985 | HUNSPELL_WARNING(stderrstderr, "warning: line %d: bad field number\n", |
| 4986 | af->getlinenum()); |
| 4987 | } else { |
| 4988 | basefieldnum = |
| 4989 | !(entry->morphcode) ? 5 : ((*(entry->morphcode) == '#') ? 5 : 6); |
| 4990 | } |
| 4991 | #endif |
| 4992 | } |
| 4993 | |
| 4994 | // now create SfxEntry or PfxEntry objects and use links to |
| 4995 | // build an ordered (sorted by affix string) list |
| 4996 | auto start = affentries.begin(), end = affentries.end(); |
| 4997 | for (auto affentry = start; affentry != end; ++affentry) { |
| 4998 | if (at == 'P') { |
| 4999 | build_pfxtree(static_cast<PfxEntry*>(*affentry)); |
| 5000 | } else { |
| 5001 | build_sfxtree(static_cast<SfxEntry*>(*affentry)); |
| 5002 | } |
| 5003 | } |
| 5004 | |
| 5005 | //contents belong to AffixMgr now |
| 5006 | affentries.release(); |
| 5007 | |
| 5008 | return true; |
| 5009 | } |
| 5010 | |
| 5011 | int AffixMgr::redundant_condition(char ft, |
| 5012 | const std::string& strip, |
| 5013 | const std::string& cond, |
| 5014 | int linenum) { |
| 5015 | int stripl = strip.size(), condl = cond.size(), i, j, neg, in; |
| 5016 | if (ft == 'P') { // prefix |
| 5017 | if (strip.compare(0, condl, cond) == 0) |
| 5018 | return 1; |
| 5019 | if (utf8) { |
| 5020 | } else { |
| 5021 | for (i = 0, j = 0; (i < stripl) && (j < condl); i++, j++) { |
| 5022 | if (cond[j] != '[') { |
| 5023 | if (cond[j] != strip[i]) { |
| 5024 | HUNSPELL_WARNING(stderrstderr, |
| 5025 | "warning: line %d: incompatible stripping " |
| 5026 | "characters and condition\n", |
| 5027 | linenum); |
| 5028 | return 0; |
| 5029 | } |
| 5030 | } else { |
| 5031 | neg = (cond[j + 1] == '^') ? 1 : 0; |
| 5032 | in = 0; |
| 5033 | do { |
| 5034 | j++; |
| 5035 | if (strip[i] == cond[j]) |
| 5036 | in = 1; |
| 5037 | } while ((j < (condl - 1)) && (cond[j] != ']')); |
| 5038 | if (j == (condl - 1) && (cond[j] != ']')) { |
| 5039 | HUNSPELL_WARNING(stderrstderr, |
| 5040 | "error: line %d: missing ] in condition:\n%s\n", |
| 5041 | linenum, cond.c_str()); |
| 5042 | return 0; |
| 5043 | } |
| 5044 | if ((!neg && !in) || (neg && in)) { |
| 5045 | HUNSPELL_WARNING(stderrstderr, |
| 5046 | "warning: line %d: incompatible stripping " |
| 5047 | "characters and condition\n", |
| 5048 | linenum); |
| 5049 | return 0; |
| 5050 | } |
| 5051 | } |
| 5052 | } |
| 5053 | if (j >= condl) |
| 5054 | return 1; |
| 5055 | } |
| 5056 | } else { // suffix |
| 5057 | if ((stripl >= condl) && strip.compare(stripl - condl, std::string::npos, cond) == 0) |
| 5058 | return 1; |
| 5059 | if (utf8) { |
| 5060 | } else { |
| 5061 | for (i = stripl - 1, j = condl - 1; (i >= 0) && (j >= 0); i--, j--) { |
| 5062 | if (cond[j] != ']') { |
| 5063 | if (cond[j] != strip[i]) { |
| 5064 | HUNSPELL_WARNING(stderrstderr, |
| 5065 | "warning: line %d: incompatible stripping " |
| 5066 | "characters and condition\n", |
| 5067 | linenum); |
| 5068 | return 0; |
| 5069 | } |
| 5070 | } else if (j > 0) { |
| 5071 | in = 0; |
| 5072 | do { |
| 5073 | j--; |
| 5074 | if (strip[i] == cond[j]) |
| 5075 | in = 1; |
| 5076 | } while ((j > 0) && (cond[j] != '[')); |
| 5077 | if ((j == 0) && (cond[j] != '[')) { |
| 5078 | HUNSPELL_WARNING(stderrstderr, |
| 5079 | "error: line: %d: missing ] in condition:\n%s\n", |
| 5080 | linenum, cond.c_str()); |
| 5081 | return 0; |
| 5082 | } |
| 5083 | neg = (cond[j + 1] == '^') ? 1 : 0; |
| 5084 | if ((!neg && !in) || (neg && in)) { |
| 5085 | HUNSPELL_WARNING(stderrstderr, |
| 5086 | "warning: line %d: incompatible stripping " |
| 5087 | "characters and condition\n", |
| 5088 | linenum); |
| 5089 | return 0; |
| 5090 | } |
| 5091 | } |
| 5092 | } |
| 5093 | if (j < 0) |
| 5094 | return 1; |
| 5095 | } |
| 5096 | } |
| 5097 | return 0; |
| 5098 | } |
| 5099 | |
| 5100 | std::vector<std::string> AffixMgr::get_suffix_words(short unsigned* suff, |
| 5101 | int len, |
| 5102 | const std::string& root_word) { |
| 5103 | std::vector<std::string> slst; |
| 5104 | AffixScratch scratch; |
| 5105 | short unsigned* start_ptr = suff; |
| 5106 | for (auto ptr : sStart) { |
| 5107 | while (ptr) { |
| 5108 | suff = start_ptr; |
| 5109 | for (int i = 0; i < len; i++) { |
| 5110 | if ((*suff) == ptr->getFlag()) { |
| 5111 | std::string nw(root_word); |
| 5112 | nw.append(ptr->getAffix()); |
| 5113 | hentry* ht = ptr->checkword(nw, 0, nw.size(), 0, nullptr, 0, 0, 0, scratch); |
| 5114 | if (ht) { |
| 5115 | slst.push_back(std::move(nw)); |
| 5116 | } |
| 5117 | } |
| 5118 | suff++; |
| 5119 | } |
| 5120 | ptr = ptr->getNext(); |
| 5121 | } |
| 5122 | } |
| 5123 | return slst; |
| 5124 | } |