yarr/YarrPattern.cpp

   1 /*
   2  * Copyright (C) 2009, 2013 Apple Inc. All rights reserved.
   3  * Copyright (C) 2010 Peter Varga (pvarga@inf.u-szeged.hu), University of Szeged
   4  *
   5  * Redistribution and use in source and binary forms, with or without
   6  * modification, are permitted provided that the following conditions
   7  * are met:
   8  * 1. Redistributions of source code must retain the above copyright
   9  *    notice, this list of conditions and the following disclaimer.
  10  * 2. Redistributions in binary form must reproduce the above copyright
  11  *    notice, this list of conditions and the following disclaimer in the
  12  *    documentation and/or other materials provided with the distribution.
  13  *
  14  * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY
  15  * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  16  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
  17  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL APPLE INC. OR
  18  * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
  19  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
  20  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
  21  * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
  22  * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
  23  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
  24  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  25  */
  26
  27 #include "config.h"
  28 #include "YarrPattern.h"
  29
  30 #include "Yarr.h"
  31 #include "YarrCanonicalizeUCS2.h"
  32 #include "YarrParser.h"
  33 #include <wtf/Vector.h>
  34
  35 using namespace WTF;
  36
  37 namespace JSC { namespace Yarr {
  38
  39 #include "RegExpJitTables.h"
  40
  41 class CharacterClassConstructor {
  42 public:
  43     CharacterClassConstructor(bool isCaseInsensitive = false)
  44         : m_isCaseInsensitive(isCaseInsensitive)
  45     {
  46     }
  47
  48     void reset()
  49     {
  50         m_matches.clear();
  51         m_ranges.clear();
  52         m_matchesUnicode.clear();
  53         m_rangesUnicode.clear();
  54     }
  55
  56     void append(const CharacterClass* other)
  57     {
  58         for (size_t i = 0; i < other->m_matches.size(); ++i)
  59             addSorted(m_matches, other->m_matches[i]);
  60         for (size_t i = 0; i < other->m_ranges.size(); ++i)
  61             addSortedRange(m_ranges, other->m_ranges[i].begin, other->m_ranges[i].end);
  62         for (size_t i = 0; i < other->m_matchesUnicode.size(); ++i)
  63             addSorted(m_matchesUnicode, other->m_matchesUnicode[i]);
  64         for (size_t i = 0; i < other->m_rangesUnicode.size(); ++i)
  65             addSortedRange(m_rangesUnicode, other->m_rangesUnicode[i].begin, other->m_rangesUnicode[i].end);
  66     }
  67
  68     void putChar(UChar ch)
  69     {
  70         // Handle ascii cases.
  71         if (ch <= 0x7f) {
  72             if (m_isCaseInsensitive && isASCIIAlpha(ch)) {
  73                 addSorted(m_matches, toASCIIUpper(ch));
  74                 addSorted(m_matches, toASCIILower(ch));
  75             } else
  76                 addSorted(m_matches, ch);
  77             return;
  78         }
  79
  80         // Simple case, not a case-insensitive match.
  81         if (!m_isCaseInsensitive) {
  82             addSorted(m_matchesUnicode, ch);
  83             return;
  84         }
  85
  86         // Add multiple matches, if necessary.
  87         UCS2CanonicalizationRange* info = rangeInfoFor(ch);
  88         if (info->type == CanonicalizeUnique)
  89             addSorted(m_matchesUnicode, ch);
  90         else
  91             putUnicodeIgnoreCase(ch, info);
  92     }
  93
  94     void putUnicodeIgnoreCase(UChar ch, UCS2CanonicalizationRange* info)
  95     {
  96         ASSERT(m_isCaseInsensitive);
  97         ASSERT(ch > 0x7f);
  98         ASSERT(ch >= info->begin && ch <= info->end);
  99         ASSERT(info->type != CanonicalizeUnique);
 100         if (info->type == CanonicalizeSet) {
 101             for (uint16_t* set = characterSetInfo[info->value]; (ch = *set); ++set)
 102                 addSorted(m_matchesUnicode, ch);
 103         } else {
 104             addSorted(m_matchesUnicode, ch);
 105             addSorted(m_matchesUnicode, getCanonicalPair(info, ch));
 106         }
 107     }
 108
 109     void putRange(UChar lo, UChar hi)
 110     {
 111         if (lo <= 0x7f) {
 112             char asciiLo = lo;
 113             char asciiHi = std::min(hi, (UChar)0x7f);
 114             addSortedRange(m_ranges, lo, asciiHi);
 115
 116             if (m_isCaseInsensitive) {
 117                 if ((asciiLo <= 'Z') && (asciiHi >= 'A'))
 118                     addSortedRange(m_ranges, std::max(asciiLo, 'A')+('a'-'A'), std::min(asciiHi, 'Z')+('a'-'A'));
 119                 if ((asciiLo <= 'z') && (asciiHi >= 'a'))
 120                     addSortedRange(m_ranges, std::max(asciiLo, 'a')+('A'-'a'), std::min(asciiHi, 'z')+('A'-'a'));
 121             }
 122         }
 123         if (hi <= 0x7f)
 124             return;
 125
 126         lo = std::max(lo, (UChar)0x80);
 127         addSortedRange(m_rangesUnicode, lo, hi);
 128
 129         if (!m_isCaseInsensitive)
 130             return;
 131
 132         UCS2CanonicalizationRange* info = rangeInfoFor(lo);
 133         while (true) {
 134             // Handle the range [lo .. end]
 135             UChar end = std::min<UChar>(info->end, hi);
 136
 137             switch (info->type) {
 138             case CanonicalizeUnique:
 139                 // Nothing to do - no canonical equivalents.
 140                 break;
 141             case CanonicalizeSet: {
 142                 UChar ch;
 143                 for (uint16_t* set = characterSetInfo[info->value]; (ch = *set); ++set)
 144                     addSorted(m_matchesUnicode, ch);
 145                 break;
 146             }
 147             case CanonicalizeRangeLo:
 148                 addSortedRange(m_rangesUnicode, lo + info->value, end + info->value);
 149                 break;
 150             case CanonicalizeRangeHi:
 151                 addSortedRange(m_rangesUnicode, lo - info->value, end - info->value);
 152                 break;
 153             case CanonicalizeAlternatingAligned:
 154                 // Use addSortedRange since there is likely an abutting range to combine with.
 155                 if (lo & 1)
 156                     addSortedRange(m_rangesUnicode, lo - 1, lo - 1);
 157                 if (!(end & 1))
 158                     addSortedRange(m_rangesUnicode, end + 1, end + 1);
 159                 break;
 160             case CanonicalizeAlternatingUnaligned:
 161                 // Use addSortedRange since there is likely an abutting range to combine with.
 162                 if (!(lo & 1))
 163                     addSortedRange(m_rangesUnicode, lo - 1, lo - 1);
 164                 if (end & 1)
 165                     addSortedRange(m_rangesUnicode, end + 1, end + 1);
 166                 break;
 167             }
 168
 169             if (hi == end)
 170                 return;
 171
 172             ++info;
 173             lo = info->begin;
 174         };
 175
 176     }
 177
 178     PassOwnPtr<CharacterClass> charClass()
 179     {
 180         OwnPtr<CharacterClass> characterClass = adoptPtr(new CharacterClass);
 181
 182         characterClass->m_matches.swap(m_matches);
 183         characterClass->m_ranges.swap(m_ranges);
 184         characterClass->m_matchesUnicode.swap(m_matchesUnicode);
 185         characterClass->m_rangesUnicode.swap(m_rangesUnicode);
 186
 187         return characterClass.release();
 188     }
 189
 190 private:
 191     void addSorted(Vector<UChar>& matches, UChar ch)
 192     {
 193         unsigned pos = 0;
 194         unsigned range = matches.size();
 195
 196         // binary chop, find position to insert char.
 197         while (range) {
 198             unsigned index = range >> 1;
 199
 200             int val = matches[pos+index] - ch;
 201             if (!val)
 202                 return;
 203             else if (val > 0)
 204                 range = index;
 205             else {
 206                 pos += (index+1);
 207                 range -= (index+1);
 208             }
 209         }
 210
 211         if (pos == matches.size())
 212             matches.append(ch);
 213         else
 214             matches.insert(pos, ch);
 215     }
 216
 217     void addSortedRange(Vector<CharacterRange>& ranges, UChar lo, UChar hi)
 218     {
 219         unsigned end = ranges.size();
 220
 221         // Simple linear scan - I doubt there are that many ranges anyway...
 222         // feel free to fix this with something faster (eg binary chop).
 223         for (unsigned i = 0; i < end; ++i) {
 224             // does the new range fall before the current position in the array
 225             if (hi < ranges[i].begin) {
 226                 // optional optimization: concatenate appending ranges? - may not be worthwhile.
 227                 if (hi == (ranges[i].begin - 1)) {
 228                     ranges[i].begin = lo;
 229                     return;
 230                 }
 231                 ranges.insert(i, CharacterRange(lo, hi));
 232                 return;
 233             }
 234             // Okay, since we didn't hit the last case, the end of the new range is definitely at or after the begining
 235             // If the new range start at or before the end of the last range, then the overlap (if it starts one after the
 236             // end of the last range they concatenate, which is just as good.
 237             if (lo <= (ranges[i].end + 1)) {
 238                 // found an intersect! we'll replace this entry in the array.
 239                 ranges[i].begin = std::min(ranges[i].begin, lo);
 240                 ranges[i].end = std::max(ranges[i].end, hi);
 241
 242                 // now check if the new range can subsume any subsequent ranges.
 243                 unsigned next = i+1;
 244                 // each iteration of the loop we will either remove something from the list, or break the loop.
 245                 while (next < ranges.size()) {
 246                     if (ranges[next].begin <= (ranges[i].end + 1)) {
 247                         // the next entry now overlaps / concatenates this one.
 248                         ranges[i].end = std::max(ranges[i].end, ranges[next].end);
 249                         ranges.remove(next);
 250                     } else
 251                         break;
 252                 }
 253
 254                 return;
 255             }
 256         }
 257
 258         // CharacterRange comes after all existing ranges.
 259         ranges.append(CharacterRange(lo, hi));
 260     }
 261
 262     bool m_isCaseInsensitive;
 263
 264     Vector<UChar> m_matches;
 265     Vector<CharacterRange> m_ranges;
 266     Vector<UChar> m_matchesUnicode;
 267     Vector<CharacterRange> m_rangesUnicode;
 268 };
 269
 270 class YarrPatternConstructor {
 271 public:
 272     YarrPatternConstructor(YarrPattern& pattern)
 273         : m_pattern(pattern)
 274         , m_characterClassConstructor(pattern.m_ignoreCase)
 275         , m_invertParentheticalAssertion(false)
 276     {
 277         OwnPtr<PatternDisjunction> body = adoptPtr(new PatternDisjunction);
 278         m_pattern.m_body = body.get();
 279         m_alternative = body->addNewAlternative();
 280         m_pattern.m_disjunctions.append(body.release());
 281     }
 282
 283     ~YarrPatternConstructor()
 284     {
 285     }
 286
 287     void reset()
 288     {
 289         m_pattern.reset();
 290         m_characterClassConstructor.reset();
 291
 292         OwnPtr<PatternDisjunction> body = adoptPtr(new PatternDisjunction);
 293         m_pattern.m_body = body.get();
 294         m_alternative = body->addNewAlternative();
 295         m_pattern.m_disjunctions.append(body.release());
 296     }
 297
 298     void assertionBOL()
 299     {
 300         if (!m_alternative->m_terms.size() & !m_invertParentheticalAssertion) {
 301             m_alternative->m_startsWithBOL = true;
 302             m_alternative->m_containsBOL = true;
 303             m_pattern.m_containsBOL = true;
 304         }
 305         m_alternative->m_terms.append(PatternTerm::BOL());
 306     }
 307     void assertionEOL()
 308     {
 309         m_alternative->m_terms.append(PatternTerm::EOL());
 310     }
 311     void assertionWordBoundary(bool invert)
 312     {
 313         m_alternative->m_terms.append(PatternTerm::WordBoundary(invert));
 314     }
 315
 316     void atomPatternCharacter(UChar ch)
 317     {
 318         // We handle case-insensitive checking of unicode characters which do have both
 319         // cases by handling them as if they were defined using a CharacterClass.
 320         if (!m_pattern.m_ignoreCase || isASCII(ch)) {
 321             m_alternative->m_terms.append(PatternTerm(ch));
 322             return;
 323         }
 324
 325         UCS2CanonicalizationRange* info = rangeInfoFor(ch);
 326         if (info->type == CanonicalizeUnique) {
 327             m_alternative->m_terms.append(PatternTerm(ch));
 328             return;
 329         }
 330
 331         m_characterClassConstructor.putUnicodeIgnoreCase(ch, info);
 332         OwnPtr<CharacterClass> newCharacterClass = m_characterClassConstructor.charClass();
 333         m_alternative->m_terms.append(PatternTerm(newCharacterClass.get(), false));
 334         m_pattern.m_userCharacterClasses.append(newCharacterClass.release());
 335     }
 336
 337     void atomBuiltInCharacterClass(BuiltInCharacterClassID classID, bool invert)
 338     {
 339         switch (classID) {
 340         case DigitClassID:
 341             m_alternative->m_terms.append(PatternTerm(m_pattern.digitsCharacterClass(), invert));
 342             break;
 343         case SpaceClassID:
 344             m_alternative->m_terms.append(PatternTerm(m_pattern.spacesCharacterClass(), invert));
 345             break;
 346         case WordClassID:
 347             m_alternative->m_terms.append(PatternTerm(m_pattern.wordcharCharacterClass(), invert));
 348             break;
 349         case NewlineClassID:
 350             m_alternative->m_terms.append(PatternTerm(m_pattern.newlineCharacterClass(), invert));
 351             break;
 352         }
 353     }
 354
 355     void atomCharacterClassBegin(bool invert = false)
 356     {
 357         m_invertCharacterClass = invert;
 358     }
 359
 360     void atomCharacterClassAtom(UChar ch)
 361     {
 362         m_characterClassConstructor.putChar(ch);
 363     }
 364
 365     void atomCharacterClassRange(UChar begin, UChar end)
 366     {
 367         m_characterClassConstructor.putRange(begin, end);
 368     }
 369
 370     void atomCharacterClassBuiltIn(BuiltInCharacterClassID classID, bool invert)
 371     {
 372         ASSERT(classID != NewlineClassID);
 373
 374         switch (classID) {
 375         case DigitClassID:
 376             m_characterClassConstructor.append(invert ? m_pattern.nondigitsCharacterClass() : m_pattern.digitsCharacterClass());
 377             break;
 378
 379         case SpaceClassID:
 380             m_characterClassConstructor.append(invert ? m_pattern.nonspacesCharacterClass() : m_pattern.spacesCharacterClass());
 381             break;
 382
 383         case WordClassID:
 384             m_characterClassConstructor.append(invert ? m_pattern.nonwordcharCharacterClass() : m_pattern.wordcharCharacterClass());
 385             break;
 386
 387         default:
 388             RELEASE_ASSERT_NOT_REACHED();
 389         }
 390     }
 391
 392     void atomCharacterClassEnd()
 393     {
 394         OwnPtr<CharacterClass> newCharacterClass = m_characterClassConstructor.charClass();
 395         m_alternative->m_terms.append(PatternTerm(newCharacterClass.get(), m_invertCharacterClass));
 396         m_pattern.m_userCharacterClasses.append(newCharacterClass.release());
 397     }
 398
 399     void atomParenthesesSubpatternBegin(bool capture = true)
 400     {
 401         unsigned subpatternId = m_pattern.m_numSubpatterns + 1;
 402         if (capture)
 403             m_pattern.m_numSubpatterns++;
 404
 405         OwnPtr<PatternDisjunction> parenthesesDisjunction = adoptPtr(new PatternDisjunction(m_alternative));
 406         m_alternative->m_terms.append(PatternTerm(PatternTerm::TypeParenthesesSubpattern, subpatternId, parenthesesDisjunction.get(), capture, false));
 407         m_alternative = parenthesesDisjunction->addNewAlternative();
 408         m_pattern.m_disjunctions.append(parenthesesDisjunction.release());
 409     }
 410
 411     void atomParentheticalAssertionBegin(bool invert = false)
 412     {
 413         OwnPtr<PatternDisjunction> parenthesesDisjunction = adoptPtr(new PatternDisjunction(m_alternative));
 414         m_alternative->m_terms.append(PatternTerm(PatternTerm::TypeParentheticalAssertion, m_pattern.m_numSubpatterns + 1, parenthesesDisjunction.get(), false, invert));
 415         m_alternative = parenthesesDisjunction->addNewAlternative();
 416         m_invertParentheticalAssertion = invert;
 417         m_pattern.m_disjunctions.append(parenthesesDisjunction.release());
 418     }
 419
 420     void atomParenthesesEnd()
 421     {
 422         ASSERT(m_alternative->m_parent);
 423         ASSERT(m_alternative->m_parent->m_parent);
 424
 425         PatternDisjunction* parenthesesDisjunction = m_alternative->m_parent;
 426         m_alternative = m_alternative->m_parent->m_parent;
 427
 428         PatternTerm& lastTerm = m_alternative->lastTerm();
 429
 430         unsigned numParenAlternatives = parenthesesDisjunction->m_alternatives.size();
 431         unsigned numBOLAnchoredAlts = 0;
 432
 433         for (unsigned i = 0; i < numParenAlternatives; i++) {
 434             // Bubble up BOL flags
 435             if (parenthesesDisjunction->m_alternatives[i]->m_startsWithBOL)
 436                 numBOLAnchoredAlts++;
 437         }
 438
 439         if (numBOLAnchoredAlts) {
 440             m_alternative->m_containsBOL = true;
 441             // If all the alternatives in parens start with BOL, then so does this one
 442             if (numBOLAnchoredAlts == numParenAlternatives)
 443                 m_alternative->m_startsWithBOL = true;
 444         }
 445
 446         lastTerm.parentheses.lastSubpatternId = m_pattern.m_numSubpatterns;
 447         m_invertParentheticalAssertion = false;
 448     }
 449
 450     void atomBackReference(unsigned subpatternId)
 451     {
 452         ASSERT(subpatternId);
 453         m_pattern.m_containsBackreferences = true;
 454         m_pattern.m_maxBackReference = std::max(m_pattern.m_maxBackReference, subpatternId);
 455
 456         if (subpatternId > m_pattern.m_numSubpatterns) {
 457             m_alternative->m_terms.append(PatternTerm::ForwardReference());
 458             return;
 459         }
 460
 461         PatternAlternative* currentAlternative = m_alternative;
 462         ASSERT(currentAlternative);
 463
 464         // Note to self: if we waited until the AST was baked, we could also remove forwards refs
 465         while ((currentAlternative = currentAlternative->m_parent->m_parent)) {
 466             PatternTerm& term = currentAlternative->lastTerm();
 467             ASSERT((term.type == PatternTerm::TypeParenthesesSubpattern) || (term.type == PatternTerm::TypeParentheticalAssertion));
 468
 469             if ((term.type == PatternTerm::TypeParenthesesSubpattern) && term.capture() && (subpatternId == term.parentheses.subpatternId)) {
 470                 m_alternative->m_terms.append(PatternTerm::ForwardReference());
 471                 return;
 472             }
 473         }
 474
 475         m_alternative->m_terms.append(PatternTerm(subpatternId));
 476     }
 477
 478     // deep copy the argument disjunction.  If filterStartsWithBOL is true,
 479     // skip alternatives with m_startsWithBOL set true.
 480     PatternDisjunction* copyDisjunction(PatternDisjunction* disjunction, bool filterStartsWithBOL = false)
 481     {
 482         OwnPtr<PatternDisjunction> newDisjunction;
 483         for (unsigned alt = 0; alt < disjunction->m_alternatives.size(); ++alt) {
 484             PatternAlternative* alternative = disjunction->m_alternatives[alt].get();
 485             if (!filterStartsWithBOL || !alternative->m_startsWithBOL) {
 486                 if (!newDisjunction) {
 487                     newDisjunction = adoptPtr(new PatternDisjunction());
 488                     newDisjunction->m_parent = disjunction->m_parent;
 489                 }
 490                 PatternAlternative* newAlternative = newDisjunction->addNewAlternative();
 491                 newAlternative->m_terms.reserveInitialCapacity(alternative->m_terms.size());
 492                 for (unsigned i = 0; i < alternative->m_terms.size(); ++i)
 493                     newAlternative->m_terms.append(copyTerm(alternative->m_terms[i], filterStartsWithBOL));
 494             }
 495         }
 496
 497         if (!newDisjunction)
 498             return 0;
 499
 500         PatternDisjunction* copiedDisjunction = newDisjunction.get();
 501         m_pattern.m_disjunctions.append(newDisjunction.release());
 502         return copiedDisjunction;
 503     }
 504
 505     PatternTerm copyTerm(PatternTerm& term, bool filterStartsWithBOL = false)
 506     {
 507         if ((term.type != PatternTerm::TypeParenthesesSubpattern) && (term.type != PatternTerm::TypeParentheticalAssertion))
 508             return PatternTerm(term);
 509
 510         PatternTerm termCopy = term;
 511         termCopy.parentheses.disjunction = copyDisjunction(termCopy.parentheses.disjunction, filterStartsWithBOL);
 512         return termCopy;
 513     }
 514
 515     void quantifyAtom(unsigned min, unsigned max, bool greedy)
 516     {
 517         ASSERT(min <= max);
 518         ASSERT(m_alternative->m_terms.size());
 519
 520         if (!max) {
 521             m_alternative->removeLastTerm();
 522             return;
 523         }
 524
 525         PatternTerm& term = m_alternative->lastTerm();
 526         ASSERT(term.type > PatternTerm::TypeAssertionWordBoundary);
 527         ASSERT((term.quantityCount == 1) && (term.quantityType == QuantifierFixedCount));
 528
 529         if (term.type == PatternTerm::TypeParentheticalAssertion) {
 530             // If an assertion is quantified with a minimum count of zero, it can simply be removed.
 531             // This arises from the RepeatMatcher behaviour in the spec. Matching an assertion never
 532             // results in any input being consumed, however the continuation passed to the assertion
 533             // (called in steps, 8c and 9 of the RepeatMatcher definition, ES5.1 15.10.2.5) will
 534             // reject all zero length matches (see step 2.1). A match from the continuation of the
 535             // expression will still be accepted regardless (via steps 8a and 11) - the upshot of all
 536             // this is that matches from the assertion are not required, and won't be accepted anyway,
 537             // so no need to ever run it.
 538             if (!min)
 539                 m_alternative->removeLastTerm();
 540             // We never need to run an assertion more than once. Subsequent interations will be run
 541             // with the same start index (since assertions are non-capturing) and the same captures
 542             // (per step 4 of RepeatMatcher in ES5.1 15.10.2.5), and as such will always produce the
 543             // same result and captures. If the first match succeeds then the subsequent (min - 1)
 544             // matches will too. Any additional optional matches will fail (on the same basis as the
 545             // minimum zero quantified assertions, above), but this will still result in a match.
 546             return;
 547         }
 548
 549         if (min == 0)
 550             term.quantify(max, greedy   ? QuantifierGreedy : QuantifierNonGreedy);
 551         else if (min == max)
 552             term.quantify(min, QuantifierFixedCount);
 553         else {
 554             term.quantify(min, QuantifierFixedCount);
 555             m_alternative->m_terms.append(copyTerm(term));
 556             // NOTE: this term is interesting from an analysis perspective, in that it can be ignored.....
 557             m_alternative->lastTerm().quantify((max == quantifyInfinite) ? max : max - min, greedy ? QuantifierGreedy : QuantifierNonGreedy);
 558             if (m_alternative->lastTerm().type == PatternTerm::TypeParenthesesSubpattern)
 559                 m_alternative->lastTerm().parentheses.isCopy = true;
 560         }
 561     }
 562
 563     void disjunction()
 564     {
 565         m_alternative = m_alternative->m_parent->addNewAlternative();
 566     }
 567
 568     unsigned setupAlternativeOffsets(PatternAlternative* alternative, unsigned currentCallFrameSize, unsigned initialInputPosition)
 569     {
 570         alternative->m_hasFixedSize = true;
 571         Checked<unsigned> currentInputPosition = initialInputPosition;
 572
 573         for (unsigned i = 0; i < alternative->m_terms.size(); ++i) {
 574             PatternTerm& term = alternative->m_terms[i];
 575
 576             switch (term.type) {
 577             case PatternTerm::TypeAssertionBOL:
 578             case PatternTerm::TypeAssertionEOL:
 579             case PatternTerm::TypeAssertionWordBoundary:
 580                 term.inputPosition = currentInputPosition.unsafeGet();
 581                 break;
 582
 583             case PatternTerm::TypeBackReference:
 584                 term.inputPosition = currentInputPosition.unsafeGet();
 585                 term.frameLocation = currentCallFrameSize;
 586                 currentCallFrameSize += YarrStackSpaceForBackTrackInfoBackReference;
 587                 alternative->m_hasFixedSize = false;
 588                 break;
 589
 590             case PatternTerm::TypeForwardReference:
 591                 break;
 592
 593             case PatternTerm::TypePatternCharacter:
 594                 term.inputPosition = currentInputPosition.unsafeGet();
 595                 if (term.quantityType != QuantifierFixedCount) {
 596                     term.frameLocation = currentCallFrameSize;
 597                     currentCallFrameSize += YarrStackSpaceForBackTrackInfoPatternCharacter;
 598                     alternative->m_hasFixedSize = false;
 599                 } else
 600                     currentInputPosition += term.quantityCount;
 601                 break;
 602
 603             case PatternTerm::TypeCharacterClass:
 604                 term.inputPosition = currentInputPosition.unsafeGet();
 605                 if (term.quantityType != QuantifierFixedCount) {
 606                     term.frameLocation = currentCallFrameSize;
 607                     currentCallFrameSize += YarrStackSpaceForBackTrackInfoCharacterClass;
 608                     alternative->m_hasFixedSize = false;
 609                 } else
 610                     currentInputPosition += term.quantityCount;
 611                 break;
 612
 613             case PatternTerm::TypeParenthesesSubpattern:
 614                 // Note: for fixed once parentheses we will ensure at least the minimum is available; others are on their own.
 615                 term.frameLocation = currentCallFrameSize;
 616                 if (term.quantityCount == 1 && !term.parentheses.isCopy) {
 617                     if (term.quantityType != QuantifierFixedCount)
 618                         currentCallFrameSize += YarrStackSpaceForBackTrackInfoParenthesesOnce;
 619                     currentCallFrameSize = setupDisjunctionOffsets(term.parentheses.disjunction, currentCallFrameSize, currentInputPosition.unsafeGet());
 620                     // If quantity is fixed, then pre-check its minimum size.
 621                     if (term.quantityType == QuantifierFixedCount)
 622                         currentInputPosition += term.parentheses.disjunction->m_minimumSize;
 623                     term.inputPosition = currentInputPosition.unsafeGet();
 624                 } else if (term.parentheses.isTerminal) {
 625                     currentCallFrameSize += YarrStackSpaceForBackTrackInfoParenthesesTerminal;
 626                     currentCallFrameSize = setupDisjunctionOffsets(term.parentheses.disjunction, currentCallFrameSize, currentInputPosition.unsafeGet());
 627                     term.inputPosition = currentInputPosition.unsafeGet();
 628                 } else {
 629                     term.inputPosition = currentInputPosition.unsafeGet();
 630                     setupDisjunctionOffsets(term.parentheses.disjunction, 0, currentInputPosition.unsafeGet());
 631                     currentCallFrameSize += YarrStackSpaceForBackTrackInfoParentheses;
 632                 }
 633                 // Fixed count of 1 could be accepted, if they have a fixed size *AND* if all alternatives are of the same length.
 634                 alternative->m_hasFixedSize = false;
 635                 break;
 636
 637             case PatternTerm::TypeParentheticalAssertion:
 638                 term.inputPosition = currentInputPosition.unsafeGet();
 639                 term.frameLocation = currentCallFrameSize;
 640                 currentCallFrameSize = setupDisjunctionOffsets(term.parentheses.disjunction, currentCallFrameSize + YarrStackSpaceForBackTrackInfoParentheticalAssertion, currentInputPosition.unsafeGet());
 641                 break;
 642
 643             case PatternTerm::TypeDotStarEnclosure:
 644                 alternative->m_hasFixedSize = false;
 645                 term.inputPosition = initialInputPosition;
 646                 break;
 647             }
 648         }
 649
 650         alternative->m_minimumSize = (currentInputPosition - initialInputPosition).unsafeGet();
 651         return currentCallFrameSize;
 652     }
 653
 654     unsigned setupDisjunctionOffsets(PatternDisjunction* disjunction, unsigned initialCallFrameSize, unsigned initialInputPosition)
 655     {
 656         if ((disjunction != m_pattern.m_body) && (disjunction->m_alternatives.size() > 1))
 657             initialCallFrameSize += YarrStackSpaceForBackTrackInfoAlternative;
 658
 659         unsigned minimumInputSize = UINT_MAX;
 660         unsigned maximumCallFrameSize = 0;
 661         bool hasFixedSize = true;
 662
 663         for (unsigned alt = 0; alt < disjunction->m_alternatives.size(); ++alt) {
 664             PatternAlternative* alternative = disjunction->m_alternatives[alt].get();
 665             unsigned currentAlternativeCallFrameSize = setupAlternativeOffsets(alternative, initialCallFrameSize, initialInputPosition);
 666             minimumInputSize = std::min(minimumInputSize, alternative->m_minimumSize);
 667             maximumCallFrameSize = std::max(maximumCallFrameSize, currentAlternativeCallFrameSize);
 668             hasFixedSize &= alternative->m_hasFixedSize;
 669         }
 670
 671         ASSERT(minimumInputSize != UINT_MAX);
 672         ASSERT(maximumCallFrameSize >= initialCallFrameSize);
 673
 674         disjunction->m_hasFixedSize = hasFixedSize;
 675         disjunction->m_minimumSize = minimumInputSize;
 676         disjunction->m_callFrameSize = maximumCallFrameSize;
 677         return maximumCallFrameSize;
 678     }
 679
 680     void setupOffsets()
 681     {
 682         setupDisjunctionOffsets(m_pattern.m_body, 0, 0);
 683     }
 684
 685     // This optimization identifies sets of parentheses that we will never need to backtrack.
 686     // In these cases we do not need to store state from prior iterations.
 687     // We can presently avoid backtracking for:
 688     //   * where the parens are at the end of the regular expression (last term in any of the
 689     //     alternatives of the main body disjunction).
 690     //   * where the parens are non-capturing, and quantified unbounded greedy (*).
 691     //   * where the parens do not contain any capturing subpatterns.
 692     void checkForTerminalParentheses()
 693     {
 694         // This check is much too crude; should be just checking whether the candidate
 695         // node contains nested capturing subpatterns, not the whole expression!
 696         if (m_pattern.m_numSubpatterns)
 697             return;
 698
 699         Vector<OwnPtr<PatternAlternative> >& alternatives = m_pattern.m_body->m_alternatives;
 700         for (size_t i = 0; i < alternatives.size(); ++i) {
 701             Vector<PatternTerm>& terms = alternatives[i]->m_terms;
 702             if (terms.size()) {
 703                 PatternTerm& term = terms.last();
 704                 if (term.type == PatternTerm::TypeParenthesesSubpattern
 705                     && term.quantityType == QuantifierGreedy
 706                     && term.quantityCount == quantifyInfinite
 707                     && !term.capture())
 708                     term.parentheses.isTerminal = true;
 709             }
 710         }
 711     }
 712
 713     void optimizeBOL()
 714     {
 715         // Look for expressions containing beginning of line (^) anchoring and unroll them.
 716         // e.g. /^a|^b|c/ becomes /^a|^b|c/ which is executed once followed by /c/ which loops
 717         // This code relies on the parsing code tagging alternatives with m_containsBOL and
 718         // m_startsWithBOL and rolling those up to containing alternatives.
 719         // At this point, this is only valid for non-multiline expressions.
 720         PatternDisjunction* disjunction = m_pattern.m_body;
 721
 722         if (!m_pattern.m_containsBOL || m_pattern.m_multiline)
 723             return;
 724
 725         PatternDisjunction* loopDisjunction = copyDisjunction(disjunction, true);
 726
 727         // Set alternatives in disjunction to "onceThrough"
 728         for (unsigned alt = 0; alt < disjunction->m_alternatives.size(); ++alt)
 729             disjunction->m_alternatives[alt]->setOnceThrough();
 730
 731         if (loopDisjunction) {
 732             // Move alternatives from loopDisjunction to disjunction
 733             for (unsigned alt = 0; alt < loopDisjunction->m_alternatives.size(); ++alt)
 734                 disjunction->m_alternatives.append(loopDisjunction->m_alternatives[alt].release());
 735
 736             loopDisjunction->m_alternatives.clear();
 737         }
 738     }
 739
 740     bool containsCapturingTerms(PatternAlternative* alternative, size_t firstTermIndex, size_t lastTermIndex)
 741     {
 742         Vector<PatternTerm>& terms = alternative->m_terms;
 743
 744         for (size_t termIndex = firstTermIndex; termIndex <= lastTermIndex; ++termIndex) {
 745             PatternTerm& term = terms[termIndex];
 746
 747             if (term.m_capture)
 748                 return true;
 749
 750             if (term.type == PatternTerm::TypeParenthesesSubpattern) {
 751                 PatternDisjunction* nestedDisjunction = term.parentheses.disjunction;
 752                 for (unsigned alt = 0; alt < nestedDisjunction->m_alternatives.size(); ++alt) {
 753                     if (containsCapturingTerms(nestedDisjunction->m_alternatives[alt].get(), 0, nestedDisjunction->m_alternatives[alt]->m_terms.size() - 1))
 754                         return true;
 755                 }
 756             }
 757         }
 758
 759         return false;
 760     }
 761
 762     // This optimization identifies alternatives in the form of
 763     // [^].*[?]<expression>.*[$] for expressions that don't have any
 764     // capturing terms. The alternative is changed to <expression>
 765     // followed by processing of the dot stars to find and adjust the
 766     // beginning and the end of the match.
 767     void optimizeDotStarWrappedExpressions()
 768     {
 769         Vector<OwnPtr<PatternAlternative> >& alternatives = m_pattern.m_body->m_alternatives;
 770         if (alternatives.size() != 1)
 771             return;
 772
 773         PatternAlternative* alternative = alternatives[0].get();
 774         Vector<PatternTerm>& terms = alternative->m_terms;
 775         if (terms.size() >= 3) {
 776             bool startsWithBOL = false;
 777             bool endsWithEOL = false;
 778             size_t termIndex, firstExpressionTerm, lastExpressionTerm;
 779
 780             termIndex = 0;
 781             if (terms[termIndex].type == PatternTerm::TypeAssertionBOL) {
 782                 startsWithBOL = true;
 783                 ++termIndex;
 784             }
 785
 786             PatternTerm& firstNonAnchorTerm = terms[termIndex];
 787             if ((firstNonAnchorTerm.type != PatternTerm::TypeCharacterClass) || (firstNonAnchorTerm.characterClass != m_pattern.newlineCharacterClass()) || !((firstNonAnchorTerm.quantityType == QuantifierGreedy) || (firstNonAnchorTerm.quantityType == QuantifierNonGreedy)))
 788                 return;
 789
 790             firstExpressionTerm = termIndex + 1;
 791
 792             termIndex = terms.size() - 1;
 793             if (terms[termIndex].type == PatternTerm::TypeAssertionEOL) {
 794                 endsWithEOL = true;
 795                 --termIndex;
 796             }
 797
 798             PatternTerm& lastNonAnchorTerm = terms[termIndex];
 799             if ((lastNonAnchorTerm.type != PatternTerm::TypeCharacterClass) || (lastNonAnchorTerm.characterClass != m_pattern.newlineCharacterClass()) || (lastNonAnchorTerm.quantityType != QuantifierGreedy))
 800                 return;
 801
 802             lastExpressionTerm = termIndex - 1;
 803
 804             if (firstExpressionTerm > lastExpressionTerm)
 805                 return;
 806
 807             if (!containsCapturingTerms(alternative, firstExpressionTerm, lastExpressionTerm)) {
 808                 for (termIndex = terms.size() - 1; termIndex > lastExpressionTerm; --termIndex)
 809                     terms.remove(termIndex);
 810
 811                 for (termIndex = firstExpressionTerm; termIndex > 0; --termIndex)
 812                     terms.remove(termIndex - 1);
 813
 814                 terms.append(PatternTerm(startsWithBOL, endsWithEOL));
 815
 816                 m_pattern.m_containsBOL = false;
 817             }
 818         }
 819     }
 820
 821 private:
 822     YarrPattern& m_pattern;
 823     PatternAlternative* m_alternative;
 824     CharacterClassConstructor m_characterClassConstructor;
 825     bool m_invertCharacterClass;
 826     bool m_invertParentheticalAssertion;
 827 };
 828
 829 const char* YarrPattern::compile(const String& patternString)
 830 {
 831     YarrPatternConstructor constructor(*this);
 832
 833     if (const char* error = parse(constructor, patternString))
 834         return error;
 835
 836     // If the pattern contains illegal backreferences reset & reparse.
 837     // Quoting Netscape's "What's new in JavaScript 1.2",
 838     //      "Note: if the number of left parentheses is less than the number specified
 839     //       in \#, the \# is taken as an octal escape as described in the next row."
 840     if (containsIllegalBackReference()) {
 841         unsigned numSubpatterns = m_numSubpatterns;
 842
 843         constructor.reset();
 844 #if !ASSERT_DISABLED
 845         const char* error =
 846 #endif
 847             parse(constructor, patternString, numSubpatterns);
 848
 849         ASSERT(!error);
 850         ASSERT(numSubpatterns == m_numSubpatterns);
 851     }
 852
 853     constructor.checkForTerminalParentheses();
 854     constructor.optimizeDotStarWrappedExpressions();
 855     constructor.optimizeBOL();
 856
 857     constructor.setupOffsets();
 858
 859     return 0;
 860 }
 861
 862 YarrPattern::YarrPattern(const String& pattern, bool ignoreCase, bool multiline, const char** error)
 863     : m_ignoreCase(ignoreCase)
 864     , m_multiline(multiline)
 865     , m_containsBackreferences(false)
 866     , m_containsBOL(false)
 867     , m_numSubpatterns(0)
 868     , m_maxBackReference(0)
 869     , newlineCached(0)
 870     , digitsCached(0)
 871     , spacesCached(0)
 872     , wordcharCached(0)
 873     , nondigitsCached(0)
 874     , nonspacesCached(0)
 875     , nonwordcharCached(0)
 876 {
 877     *error = compile(pattern);
 878 }
 879
 880 } }