1 //===--- UnwrappedLineParser.cpp - Format C++ code ------------------------===// 2 // 3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 4 // See https://llvm.org/LICENSE.txt for license information. 5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 6 // 7 //===----------------------------------------------------------------------===// 8 /// 9 /// \file 10 /// This file contains the implementation of the UnwrappedLineParser, 11 /// which turns a stream of tokens into UnwrappedLines. 12 /// 13 //===----------------------------------------------------------------------===// 14 15 #include "UnwrappedLineParser.h" 16 #include "FormatToken.h" 17 #include "llvm/ADT/STLExtras.h" 18 #include "llvm/Support/Debug.h" 19 #include "llvm/Support/raw_ostream.h" 20 21 #include <algorithm> 22 23 #define DEBUG_TYPE "format-parser" 24 25 namespace clang { 26 namespace format { 27 28 class FormatTokenSource { 29 public: 30 virtual ~FormatTokenSource() {} 31 virtual FormatToken *getNextToken() = 0; 32 33 virtual unsigned getPosition() = 0; 34 virtual FormatToken *setPosition(unsigned Position) = 0; 35 }; 36 37 namespace { 38 39 class ScopedDeclarationState { 40 public: 41 ScopedDeclarationState(UnwrappedLine &Line, std::vector<bool> &Stack, 42 bool MustBeDeclaration) 43 : Line(Line), Stack(Stack) { 44 Line.MustBeDeclaration = MustBeDeclaration; 45 Stack.push_back(MustBeDeclaration); 46 } 47 ~ScopedDeclarationState() { 48 Stack.pop_back(); 49 if (!Stack.empty()) 50 Line.MustBeDeclaration = Stack.back(); 51 else 52 Line.MustBeDeclaration = true; 53 } 54 55 private: 56 UnwrappedLine &Line; 57 std::vector<bool> &Stack; 58 }; 59 60 static bool isLineComment(const FormatToken &FormatTok) { 61 return FormatTok.is(tok::comment) && !FormatTok.TokenText.startswith("/*"); 62 } 63 64 // Checks if \p FormatTok is a line comment that continues the line comment 65 // \p Previous. The original column of \p MinColumnToken is used to determine 66 // whether \p FormatTok is indented enough to the right to continue \p Previous. 67 static bool continuesLineComment(const FormatToken &FormatTok, 68 const FormatToken *Previous, 69 const FormatToken *MinColumnToken) { 70 if (!Previous || !MinColumnToken) 71 return false; 72 unsigned MinContinueColumn = 73 MinColumnToken->OriginalColumn + (isLineComment(*MinColumnToken) ? 0 : 1); 74 return isLineComment(FormatTok) && FormatTok.NewlinesBefore == 1 && 75 isLineComment(*Previous) && 76 FormatTok.OriginalColumn >= MinContinueColumn; 77 } 78 79 class ScopedMacroState : public FormatTokenSource { 80 public: 81 ScopedMacroState(UnwrappedLine &Line, FormatTokenSource *&TokenSource, 82 FormatToken *&ResetToken) 83 : Line(Line), TokenSource(TokenSource), ResetToken(ResetToken), 84 PreviousLineLevel(Line.Level), PreviousTokenSource(TokenSource), 85 Token(nullptr), PreviousToken(nullptr) { 86 FakeEOF.Tok.startToken(); 87 FakeEOF.Tok.setKind(tok::eof); 88 TokenSource = this; 89 Line.Level = 0; 90 Line.InPPDirective = true; 91 } 92 93 ~ScopedMacroState() override { 94 TokenSource = PreviousTokenSource; 95 ResetToken = Token; 96 Line.InPPDirective = false; 97 Line.Level = PreviousLineLevel; 98 } 99 100 FormatToken *getNextToken() override { 101 // The \c UnwrappedLineParser guards against this by never calling 102 // \c getNextToken() after it has encountered the first eof token. 103 assert(!eof()); 104 PreviousToken = Token; 105 Token = PreviousTokenSource->getNextToken(); 106 if (eof()) 107 return &FakeEOF; 108 return Token; 109 } 110 111 unsigned getPosition() override { return PreviousTokenSource->getPosition(); } 112 113 FormatToken *setPosition(unsigned Position) override { 114 PreviousToken = nullptr; 115 Token = PreviousTokenSource->setPosition(Position); 116 return Token; 117 } 118 119 private: 120 bool eof() { 121 return Token && Token->HasUnescapedNewline && 122 !continuesLineComment(*Token, PreviousToken, 123 /*MinColumnToken=*/PreviousToken); 124 } 125 126 FormatToken FakeEOF; 127 UnwrappedLine &Line; 128 FormatTokenSource *&TokenSource; 129 FormatToken *&ResetToken; 130 unsigned PreviousLineLevel; 131 FormatTokenSource *PreviousTokenSource; 132 133 FormatToken *Token; 134 FormatToken *PreviousToken; 135 }; 136 137 } // end anonymous namespace 138 139 class ScopedLineState { 140 public: 141 ScopedLineState(UnwrappedLineParser &Parser, 142 bool SwitchToPreprocessorLines = false) 143 : Parser(Parser), OriginalLines(Parser.CurrentLines) { 144 if (SwitchToPreprocessorLines) 145 Parser.CurrentLines = &Parser.PreprocessorDirectives; 146 else if (!Parser.Line->Tokens.empty()) 147 Parser.CurrentLines = &Parser.Line->Tokens.back().Children; 148 PreBlockLine = std::move(Parser.Line); 149 Parser.Line = std::make_unique<UnwrappedLine>(); 150 Parser.Line->Level = PreBlockLine->Level; 151 Parser.Line->InPPDirective = PreBlockLine->InPPDirective; 152 } 153 154 ~ScopedLineState() { 155 if (!Parser.Line->Tokens.empty()) { 156 Parser.addUnwrappedLine(); 157 } 158 assert(Parser.Line->Tokens.empty()); 159 Parser.Line = std::move(PreBlockLine); 160 if (Parser.CurrentLines == &Parser.PreprocessorDirectives) 161 Parser.MustBreakBeforeNextToken = true; 162 Parser.CurrentLines = OriginalLines; 163 } 164 165 private: 166 UnwrappedLineParser &Parser; 167 168 std::unique_ptr<UnwrappedLine> PreBlockLine; 169 SmallVectorImpl<UnwrappedLine> *OriginalLines; 170 }; 171 172 class CompoundStatementIndenter { 173 public: 174 CompoundStatementIndenter(UnwrappedLineParser *Parser, 175 const FormatStyle &Style, unsigned &LineLevel) 176 : CompoundStatementIndenter(Parser, LineLevel, 177 Style.BraceWrapping.AfterControlStatement, 178 Style.BraceWrapping.IndentBraces) {} 179 CompoundStatementIndenter(UnwrappedLineParser *Parser, unsigned &LineLevel, 180 bool WrapBrace, bool IndentBrace) 181 : LineLevel(LineLevel), OldLineLevel(LineLevel) { 182 if (WrapBrace) 183 Parser->addUnwrappedLine(); 184 if (IndentBrace) 185 ++LineLevel; 186 } 187 ~CompoundStatementIndenter() { LineLevel = OldLineLevel; } 188 189 private: 190 unsigned &LineLevel; 191 unsigned OldLineLevel; 192 }; 193 194 namespace { 195 196 class IndexedTokenSource : public FormatTokenSource { 197 public: 198 IndexedTokenSource(ArrayRef<FormatToken *> Tokens) 199 : Tokens(Tokens), Position(-1) {} 200 201 FormatToken *getNextToken() override { 202 ++Position; 203 return Tokens[Position]; 204 } 205 206 unsigned getPosition() override { 207 assert(Position >= 0); 208 return Position; 209 } 210 211 FormatToken *setPosition(unsigned P) override { 212 Position = P; 213 return Tokens[Position]; 214 } 215 216 void reset() { Position = -1; } 217 218 private: 219 ArrayRef<FormatToken *> Tokens; 220 int Position; 221 }; 222 223 } // end anonymous namespace 224 225 UnwrappedLineParser::UnwrappedLineParser(const FormatStyle &Style, 226 const AdditionalKeywords &Keywords, 227 unsigned FirstStartColumn, 228 ArrayRef<FormatToken *> Tokens, 229 UnwrappedLineConsumer &Callback) 230 : Line(new UnwrappedLine), MustBreakBeforeNextToken(false), 231 CurrentLines(&Lines), Style(Style), Keywords(Keywords), 232 CommentPragmasRegex(Style.CommentPragmas), Tokens(nullptr), 233 Callback(Callback), AllTokens(Tokens), PPBranchLevel(-1), 234 IncludeGuard(Style.IndentPPDirectives == FormatStyle::PPDIS_None 235 ? IG_Rejected 236 : IG_Inited), 237 IncludeGuardToken(nullptr), FirstStartColumn(FirstStartColumn) {} 238 239 void UnwrappedLineParser::reset() { 240 PPBranchLevel = -1; 241 IncludeGuard = Style.IndentPPDirectives == FormatStyle::PPDIS_None 242 ? IG_Rejected 243 : IG_Inited; 244 IncludeGuardToken = nullptr; 245 Line.reset(new UnwrappedLine); 246 CommentsBeforeNextToken.clear(); 247 FormatTok = nullptr; 248 MustBreakBeforeNextToken = false; 249 PreprocessorDirectives.clear(); 250 CurrentLines = &Lines; 251 DeclarationScopeStack.clear(); 252 PPStack.clear(); 253 Line->FirstStartColumn = FirstStartColumn; 254 } 255 256 void UnwrappedLineParser::parse() { 257 IndexedTokenSource TokenSource(AllTokens); 258 Line->FirstStartColumn = FirstStartColumn; 259 do { 260 LLVM_DEBUG(llvm::dbgs() << "----\n"); 261 reset(); 262 Tokens = &TokenSource; 263 TokenSource.reset(); 264 265 readToken(); 266 parseFile(); 267 268 // If we found an include guard then all preprocessor directives (other than 269 // the guard) are over-indented by one. 270 if (IncludeGuard == IG_Found) 271 for (auto &Line : Lines) 272 if (Line.InPPDirective && Line.Level > 0) 273 --Line.Level; 274 275 // Create line with eof token. 276 pushToken(FormatTok); 277 addUnwrappedLine(); 278 279 for (SmallVectorImpl<UnwrappedLine>::iterator I = Lines.begin(), 280 E = Lines.end(); 281 I != E; ++I) { 282 Callback.consumeUnwrappedLine(*I); 283 } 284 Callback.finishRun(); 285 Lines.clear(); 286 while (!PPLevelBranchIndex.empty() && 287 PPLevelBranchIndex.back() + 1 >= PPLevelBranchCount.back()) { 288 PPLevelBranchIndex.resize(PPLevelBranchIndex.size() - 1); 289 PPLevelBranchCount.resize(PPLevelBranchCount.size() - 1); 290 } 291 if (!PPLevelBranchIndex.empty()) { 292 ++PPLevelBranchIndex.back(); 293 assert(PPLevelBranchIndex.size() == PPLevelBranchCount.size()); 294 assert(PPLevelBranchIndex.back() <= PPLevelBranchCount.back()); 295 } 296 } while (!PPLevelBranchIndex.empty()); 297 } 298 299 void UnwrappedLineParser::parseFile() { 300 // The top-level context in a file always has declarations, except for pre- 301 // processor directives and JavaScript files. 302 bool MustBeDeclaration = 303 !Line->InPPDirective && Style.Language != FormatStyle::LK_JavaScript; 304 ScopedDeclarationState DeclarationState(*Line, DeclarationScopeStack, 305 MustBeDeclaration); 306 if (Style.Language == FormatStyle::LK_TextProto) 307 parseBracedList(); 308 else 309 parseLevel(/*HasOpeningBrace=*/false); 310 // Make sure to format the remaining tokens. 311 // 312 // LK_TextProto is special since its top-level is parsed as the body of a 313 // braced list, which does not necessarily have natural line separators such 314 // as a semicolon. Comments after the last entry that have been determined to 315 // not belong to that line, as in: 316 // key: value 317 // // endfile comment 318 // do not have a chance to be put on a line of their own until this point. 319 // Here we add this newline before end-of-file comments. 320 if (Style.Language == FormatStyle::LK_TextProto && 321 !CommentsBeforeNextToken.empty()) 322 addUnwrappedLine(); 323 flushComments(true); 324 addUnwrappedLine(); 325 } 326 327 void UnwrappedLineParser::parseCSharpGenericTypeConstraint() { 328 do { 329 switch (FormatTok->Tok.getKind()) { 330 case tok::l_brace: 331 return; 332 default: 333 if (FormatTok->is(Keywords.kw_where)) { 334 addUnwrappedLine(); 335 nextToken(); 336 parseCSharpGenericTypeConstraint(); 337 break; 338 } 339 nextToken(); 340 break; 341 } 342 } while (!eof()); 343 } 344 345 void UnwrappedLineParser::parseCSharpAttribute() { 346 int UnpairedSquareBrackets = 1; 347 do { 348 switch (FormatTok->Tok.getKind()) { 349 case tok::r_square: 350 nextToken(); 351 --UnpairedSquareBrackets; 352 if (UnpairedSquareBrackets == 0) { 353 addUnwrappedLine(); 354 return; 355 } 356 break; 357 case tok::l_square: 358 ++UnpairedSquareBrackets; 359 nextToken(); 360 break; 361 default: 362 nextToken(); 363 break; 364 } 365 } while (!eof()); 366 } 367 368 void UnwrappedLineParser::parseLevel(bool HasOpeningBrace) { 369 bool SwitchLabelEncountered = false; 370 do { 371 tok::TokenKind kind = FormatTok->Tok.getKind(); 372 if (FormatTok->getType() == TT_MacroBlockBegin) { 373 kind = tok::l_brace; 374 } else if (FormatTok->getType() == TT_MacroBlockEnd) { 375 kind = tok::r_brace; 376 } 377 378 switch (kind) { 379 case tok::comment: 380 nextToken(); 381 addUnwrappedLine(); 382 break; 383 case tok::l_brace: 384 // FIXME: Add parameter whether this can happen - if this happens, we must 385 // be in a non-declaration context. 386 if (!FormatTok->is(TT_MacroBlockBegin) && tryToParseBracedList()) 387 continue; 388 parseBlock(/*MustBeDeclaration=*/false); 389 addUnwrappedLine(); 390 break; 391 case tok::r_brace: 392 if (HasOpeningBrace) 393 return; 394 nextToken(); 395 addUnwrappedLine(); 396 break; 397 case tok::kw_default: { 398 unsigned StoredPosition = Tokens->getPosition(); 399 FormatToken *Next; 400 do { 401 Next = Tokens->getNextToken(); 402 } while (Next && Next->is(tok::comment)); 403 FormatTok = Tokens->setPosition(StoredPosition); 404 if (Next && Next->isNot(tok::colon)) { 405 // default not followed by ':' is not a case label; treat it like 406 // an identifier. 407 parseStructuralElement(); 408 break; 409 } 410 // Else, if it is 'default:', fall through to the case handling. 411 LLVM_FALLTHROUGH; 412 } 413 case tok::kw_case: 414 if (Style.Language == FormatStyle::LK_JavaScript && 415 Line->MustBeDeclaration) { 416 // A 'case: string' style field declaration. 417 parseStructuralElement(); 418 break; 419 } 420 if (!SwitchLabelEncountered && 421 (Style.IndentCaseLabels || (Line->InPPDirective && Line->Level == 1))) 422 ++Line->Level; 423 SwitchLabelEncountered = true; 424 parseStructuralElement(); 425 break; 426 case tok::l_square: 427 if (Style.isCSharp()) { 428 nextToken(); 429 parseCSharpAttribute(); 430 break; 431 } 432 LLVM_FALLTHROUGH; 433 default: 434 parseStructuralElement(/*IsTopLevel=*/true); 435 break; 436 } 437 } while (!eof()); 438 } 439 440 void UnwrappedLineParser::calculateBraceTypes(bool ExpectClassBody) { 441 // We'll parse forward through the tokens until we hit 442 // a closing brace or eof - note that getNextToken() will 443 // parse macros, so this will magically work inside macro 444 // definitions, too. 445 unsigned StoredPosition = Tokens->getPosition(); 446 FormatToken *Tok = FormatTok; 447 const FormatToken *PrevTok = Tok->Previous; 448 // Keep a stack of positions of lbrace tokens. We will 449 // update information about whether an lbrace starts a 450 // braced init list or a different block during the loop. 451 SmallVector<FormatToken *, 8> LBraceStack; 452 assert(Tok->Tok.is(tok::l_brace)); 453 do { 454 // Get next non-comment token. 455 FormatToken *NextTok; 456 unsigned ReadTokens = 0; 457 do { 458 NextTok = Tokens->getNextToken(); 459 ++ReadTokens; 460 } while (NextTok->is(tok::comment)); 461 462 switch (Tok->Tok.getKind()) { 463 case tok::l_brace: 464 if (Style.Language == FormatStyle::LK_JavaScript && PrevTok) { 465 if (PrevTok->isOneOf(tok::colon, tok::less)) 466 // A ':' indicates this code is in a type, or a braced list 467 // following a label in an object literal ({a: {b: 1}}). 468 // A '<' could be an object used in a comparison, but that is nonsense 469 // code (can never return true), so more likely it is a generic type 470 // argument (`X<{a: string; b: number}>`). 471 // The code below could be confused by semicolons between the 472 // individual members in a type member list, which would normally 473 // trigger BK_Block. In both cases, this must be parsed as an inline 474 // braced init. 475 Tok->setBlockKind(BK_BracedInit); 476 else if (PrevTok->is(tok::r_paren)) 477 // `) { }` can only occur in function or method declarations in JS. 478 Tok->setBlockKind(BK_Block); 479 } else { 480 Tok->setBlockKind(BK_Unknown); 481 } 482 LBraceStack.push_back(Tok); 483 break; 484 case tok::r_brace: 485 if (LBraceStack.empty()) 486 break; 487 if (LBraceStack.back()->is(BK_Unknown)) { 488 bool ProbablyBracedList = false; 489 if (Style.Language == FormatStyle::LK_Proto) { 490 ProbablyBracedList = NextTok->isOneOf(tok::comma, tok::r_square); 491 } else { 492 // Using OriginalColumn to distinguish between ObjC methods and 493 // binary operators is a bit hacky. 494 bool NextIsObjCMethod = NextTok->isOneOf(tok::plus, tok::minus) && 495 NextTok->OriginalColumn == 0; 496 497 // If there is a comma, semicolon or right paren after the closing 498 // brace, we assume this is a braced initializer list. Note that 499 // regardless how we mark inner braces here, we will overwrite the 500 // BlockKind later if we parse a braced list (where all blocks 501 // inside are by default braced lists), or when we explicitly detect 502 // blocks (for example while parsing lambdas). 503 // FIXME: Some of these do not apply to JS, e.g. "} {" can never be a 504 // braced list in JS. 505 ProbablyBracedList = 506 (Style.Language == FormatStyle::LK_JavaScript && 507 NextTok->isOneOf(Keywords.kw_of, Keywords.kw_in, 508 Keywords.kw_as)) || 509 (Style.isCpp() && NextTok->is(tok::l_paren)) || 510 NextTok->isOneOf(tok::comma, tok::period, tok::colon, 511 tok::r_paren, tok::r_square, tok::l_brace, 512 tok::ellipsis) || 513 (NextTok->is(tok::identifier) && 514 !PrevTok->isOneOf(tok::semi, tok::r_brace, tok::l_brace)) || 515 (NextTok->is(tok::semi) && 516 (!ExpectClassBody || LBraceStack.size() != 1)) || 517 (NextTok->isBinaryOperator() && !NextIsObjCMethod); 518 if (!Style.isCSharp() && NextTok->is(tok::l_square)) { 519 // We can have an array subscript after a braced init 520 // list, but C++11 attributes are expected after blocks. 521 NextTok = Tokens->getNextToken(); 522 ++ReadTokens; 523 ProbablyBracedList = NextTok->isNot(tok::l_square); 524 } 525 } 526 if (ProbablyBracedList) { 527 Tok->setBlockKind(BK_BracedInit); 528 LBraceStack.back()->setBlockKind(BK_BracedInit); 529 } else { 530 Tok->setBlockKind(BK_Block); 531 LBraceStack.back()->setBlockKind(BK_Block); 532 } 533 } 534 LBraceStack.pop_back(); 535 break; 536 case tok::identifier: 537 if (!Tok->is(TT_StatementMacro)) 538 break; 539 LLVM_FALLTHROUGH; 540 case tok::at: 541 case tok::semi: 542 case tok::kw_if: 543 case tok::kw_while: 544 case tok::kw_for: 545 case tok::kw_switch: 546 case tok::kw_try: 547 case tok::kw___try: 548 if (!LBraceStack.empty() && LBraceStack.back()->is(BK_Unknown)) 549 LBraceStack.back()->setBlockKind(BK_Block); 550 break; 551 default: 552 break; 553 } 554 PrevTok = Tok; 555 Tok = NextTok; 556 } while (Tok->Tok.isNot(tok::eof) && !LBraceStack.empty()); 557 558 // Assume other blocks for all unclosed opening braces. 559 for (unsigned i = 0, e = LBraceStack.size(); i != e; ++i) { 560 if (LBraceStack[i]->is(BK_Unknown)) 561 LBraceStack[i]->setBlockKind(BK_Block); 562 } 563 564 FormatTok = Tokens->setPosition(StoredPosition); 565 } 566 567 template <class T> 568 static inline void hash_combine(std::size_t &seed, const T &v) { 569 std::hash<T> hasher; 570 seed ^= hasher(v) + 0x9e3779b9 + (seed << 6) + (seed >> 2); 571 } 572 573 size_t UnwrappedLineParser::computePPHash() const { 574 size_t h = 0; 575 for (const auto &i : PPStack) { 576 hash_combine(h, size_t(i.Kind)); 577 hash_combine(h, i.Line); 578 } 579 return h; 580 } 581 582 void UnwrappedLineParser::parseBlock(bool MustBeDeclaration, unsigned AddLevels, 583 bool MunchSemi, 584 bool UnindentWhitesmithsBraces) { 585 assert(FormatTok->isOneOf(tok::l_brace, TT_MacroBlockBegin) && 586 "'{' or macro block token expected"); 587 const bool MacroBlock = FormatTok->is(TT_MacroBlockBegin); 588 FormatTok->setBlockKind(BK_Block); 589 590 // For Whitesmiths mode, jump to the next level prior to skipping over the 591 // braces. 592 if (AddLevels > 0 && Style.BreakBeforeBraces == FormatStyle::BS_Whitesmiths) 593 ++Line->Level; 594 595 size_t PPStartHash = computePPHash(); 596 597 unsigned InitialLevel = Line->Level; 598 nextToken(/*LevelDifference=*/AddLevels); 599 600 if (MacroBlock && FormatTok->is(tok::l_paren)) 601 parseParens(); 602 603 size_t NbPreprocessorDirectives = 604 CurrentLines == &Lines ? PreprocessorDirectives.size() : 0; 605 addUnwrappedLine(); 606 size_t OpeningLineIndex = 607 CurrentLines->empty() 608 ? (UnwrappedLine::kInvalidIndex) 609 : (CurrentLines->size() - 1 - NbPreprocessorDirectives); 610 611 // Whitesmiths is weird here. The brace needs to be indented for the namespace 612 // block, but the block itself may not be indented depending on the style 613 // settings. This allows the format to back up one level in those cases. 614 if (UnindentWhitesmithsBraces) 615 --Line->Level; 616 617 ScopedDeclarationState DeclarationState(*Line, DeclarationScopeStack, 618 MustBeDeclaration); 619 if (AddLevels > 0u && Style.BreakBeforeBraces != FormatStyle::BS_Whitesmiths) 620 Line->Level += AddLevels; 621 parseLevel(/*HasOpeningBrace=*/true); 622 623 if (eof()) 624 return; 625 626 if (MacroBlock ? !FormatTok->is(TT_MacroBlockEnd) 627 : !FormatTok->is(tok::r_brace)) { 628 Line->Level = InitialLevel; 629 FormatTok->setBlockKind(BK_Block); 630 return; 631 } 632 633 size_t PPEndHash = computePPHash(); 634 635 // Munch the closing brace. 636 nextToken(/*LevelDifference=*/-AddLevels); 637 638 if (MacroBlock && FormatTok->is(tok::l_paren)) 639 parseParens(); 640 641 if (FormatTok->is(tok::arrow)) { 642 // Following the } we can find a trailing return type arrow 643 // as part of an implicit conversion constraint. 644 nextToken(); 645 parseStructuralElement(); 646 } 647 648 if (MunchSemi && FormatTok->Tok.is(tok::semi)) 649 nextToken(); 650 651 Line->Level = InitialLevel; 652 FormatTok->setBlockKind(BK_Block); 653 654 if (PPStartHash == PPEndHash) { 655 Line->MatchingOpeningBlockLineIndex = OpeningLineIndex; 656 if (OpeningLineIndex != UnwrappedLine::kInvalidIndex) { 657 // Update the opening line to add the forward reference as well 658 (*CurrentLines)[OpeningLineIndex].MatchingClosingBlockLineIndex = 659 CurrentLines->size() - 1; 660 } 661 } 662 } 663 664 static bool isGoogScope(const UnwrappedLine &Line) { 665 // FIXME: Closure-library specific stuff should not be hard-coded but be 666 // configurable. 667 if (Line.Tokens.size() < 4) 668 return false; 669 auto I = Line.Tokens.begin(); 670 if (I->Tok->TokenText != "goog") 671 return false; 672 ++I; 673 if (I->Tok->isNot(tok::period)) 674 return false; 675 ++I; 676 if (I->Tok->TokenText != "scope") 677 return false; 678 ++I; 679 return I->Tok->is(tok::l_paren); 680 } 681 682 static bool isIIFE(const UnwrappedLine &Line, 683 const AdditionalKeywords &Keywords) { 684 // Look for the start of an immediately invoked anonymous function. 685 // https://en.wikipedia.org/wiki/Immediately-invoked_function_expression 686 // This is commonly done in JavaScript to create a new, anonymous scope. 687 // Example: (function() { ... })() 688 if (Line.Tokens.size() < 3) 689 return false; 690 auto I = Line.Tokens.begin(); 691 if (I->Tok->isNot(tok::l_paren)) 692 return false; 693 ++I; 694 if (I->Tok->isNot(Keywords.kw_function)) 695 return false; 696 ++I; 697 return I->Tok->is(tok::l_paren); 698 } 699 700 static bool ShouldBreakBeforeBrace(const FormatStyle &Style, 701 const FormatToken &InitialToken) { 702 if (InitialToken.isOneOf(tok::kw_namespace, TT_NamespaceMacro)) 703 return Style.BraceWrapping.AfterNamespace; 704 if (InitialToken.is(tok::kw_class)) 705 return Style.BraceWrapping.AfterClass; 706 if (InitialToken.is(tok::kw_union)) 707 return Style.BraceWrapping.AfterUnion; 708 if (InitialToken.is(tok::kw_struct)) 709 return Style.BraceWrapping.AfterStruct; 710 return false; 711 } 712 713 void UnwrappedLineParser::parseChildBlock() { 714 FormatTok->setBlockKind(BK_Block); 715 nextToken(); 716 { 717 bool SkipIndent = (Style.Language == FormatStyle::LK_JavaScript && 718 (isGoogScope(*Line) || isIIFE(*Line, Keywords))); 719 ScopedLineState LineState(*this); 720 ScopedDeclarationState DeclarationState(*Line, DeclarationScopeStack, 721 /*MustBeDeclaration=*/false); 722 Line->Level += SkipIndent ? 0 : 1; 723 parseLevel(/*HasOpeningBrace=*/true); 724 flushComments(isOnNewLine(*FormatTok)); 725 Line->Level -= SkipIndent ? 0 : 1; 726 } 727 nextToken(); 728 } 729 730 void UnwrappedLineParser::parsePPDirective() { 731 assert(FormatTok->Tok.is(tok::hash) && "'#' expected"); 732 ScopedMacroState MacroState(*Line, Tokens, FormatTok); 733 734 nextToken(); 735 736 if (!FormatTok->Tok.getIdentifierInfo()) { 737 parsePPUnknown(); 738 return; 739 } 740 741 switch (FormatTok->Tok.getIdentifierInfo()->getPPKeywordID()) { 742 case tok::pp_define: 743 parsePPDefine(); 744 return; 745 case tok::pp_if: 746 parsePPIf(/*IfDef=*/false); 747 break; 748 case tok::pp_ifdef: 749 case tok::pp_ifndef: 750 parsePPIf(/*IfDef=*/true); 751 break; 752 case tok::pp_else: 753 parsePPElse(); 754 break; 755 case tok::pp_elifdef: 756 case tok::pp_elifndef: 757 case tok::pp_elif: 758 parsePPElIf(); 759 break; 760 case tok::pp_endif: 761 parsePPEndIf(); 762 break; 763 default: 764 parsePPUnknown(); 765 break; 766 } 767 } 768 769 void UnwrappedLineParser::conditionalCompilationCondition(bool Unreachable) { 770 size_t Line = CurrentLines->size(); 771 if (CurrentLines == &PreprocessorDirectives) 772 Line += Lines.size(); 773 774 if (Unreachable || 775 (!PPStack.empty() && PPStack.back().Kind == PP_Unreachable)) 776 PPStack.push_back({PP_Unreachable, Line}); 777 else 778 PPStack.push_back({PP_Conditional, Line}); 779 } 780 781 void UnwrappedLineParser::conditionalCompilationStart(bool Unreachable) { 782 ++PPBranchLevel; 783 assert(PPBranchLevel >= 0 && PPBranchLevel <= (int)PPLevelBranchIndex.size()); 784 if (PPBranchLevel == (int)PPLevelBranchIndex.size()) { 785 PPLevelBranchIndex.push_back(0); 786 PPLevelBranchCount.push_back(0); 787 } 788 PPChainBranchIndex.push(0); 789 bool Skip = PPLevelBranchIndex[PPBranchLevel] > 0; 790 conditionalCompilationCondition(Unreachable || Skip); 791 } 792 793 void UnwrappedLineParser::conditionalCompilationAlternative() { 794 if (!PPStack.empty()) 795 PPStack.pop_back(); 796 assert(PPBranchLevel < (int)PPLevelBranchIndex.size()); 797 if (!PPChainBranchIndex.empty()) 798 ++PPChainBranchIndex.top(); 799 conditionalCompilationCondition( 800 PPBranchLevel >= 0 && !PPChainBranchIndex.empty() && 801 PPLevelBranchIndex[PPBranchLevel] != PPChainBranchIndex.top()); 802 } 803 804 void UnwrappedLineParser::conditionalCompilationEnd() { 805 assert(PPBranchLevel < (int)PPLevelBranchIndex.size()); 806 if (PPBranchLevel >= 0 && !PPChainBranchIndex.empty()) { 807 if (PPChainBranchIndex.top() + 1 > PPLevelBranchCount[PPBranchLevel]) { 808 PPLevelBranchCount[PPBranchLevel] = PPChainBranchIndex.top() + 1; 809 } 810 } 811 // Guard against #endif's without #if. 812 if (PPBranchLevel > -1) 813 --PPBranchLevel; 814 if (!PPChainBranchIndex.empty()) 815 PPChainBranchIndex.pop(); 816 if (!PPStack.empty()) 817 PPStack.pop_back(); 818 } 819 820 void UnwrappedLineParser::parsePPIf(bool IfDef) { 821 bool IfNDef = FormatTok->is(tok::pp_ifndef); 822 nextToken(); 823 bool Unreachable = false; 824 if (!IfDef && (FormatTok->is(tok::kw_false) || FormatTok->TokenText == "0")) 825 Unreachable = true; 826 if (IfDef && !IfNDef && FormatTok->TokenText == "SWIG") 827 Unreachable = true; 828 conditionalCompilationStart(Unreachable); 829 FormatToken *IfCondition = FormatTok; 830 // If there's a #ifndef on the first line, and the only lines before it are 831 // comments, it could be an include guard. 832 bool MaybeIncludeGuard = IfNDef; 833 if (IncludeGuard == IG_Inited && MaybeIncludeGuard) 834 for (auto &Line : Lines) { 835 if (!Line.Tokens.front().Tok->is(tok::comment)) { 836 MaybeIncludeGuard = false; 837 IncludeGuard = IG_Rejected; 838 break; 839 } 840 } 841 --PPBranchLevel; 842 parsePPUnknown(); 843 ++PPBranchLevel; 844 if (IncludeGuard == IG_Inited && MaybeIncludeGuard) { 845 IncludeGuard = IG_IfNdefed; 846 IncludeGuardToken = IfCondition; 847 } 848 } 849 850 void UnwrappedLineParser::parsePPElse() { 851 // If a potential include guard has an #else, it's not an include guard. 852 if (IncludeGuard == IG_Defined && PPBranchLevel == 0) 853 IncludeGuard = IG_Rejected; 854 conditionalCompilationAlternative(); 855 if (PPBranchLevel > -1) 856 --PPBranchLevel; 857 parsePPUnknown(); 858 ++PPBranchLevel; 859 } 860 861 void UnwrappedLineParser::parsePPElIf() { parsePPElse(); } 862 863 void UnwrappedLineParser::parsePPEndIf() { 864 conditionalCompilationEnd(); 865 parsePPUnknown(); 866 // If the #endif of a potential include guard is the last thing in the file, 867 // then we found an include guard. 868 unsigned TokenPosition = Tokens->getPosition(); 869 FormatToken *PeekNext = AllTokens[TokenPosition]; 870 if (IncludeGuard == IG_Defined && PPBranchLevel == -1 && 871 PeekNext->is(tok::eof) && 872 Style.IndentPPDirectives != FormatStyle::PPDIS_None) 873 IncludeGuard = IG_Found; 874 } 875 876 void UnwrappedLineParser::parsePPDefine() { 877 nextToken(); 878 879 if (!FormatTok->Tok.getIdentifierInfo()) { 880 IncludeGuard = IG_Rejected; 881 IncludeGuardToken = nullptr; 882 parsePPUnknown(); 883 return; 884 } 885 886 if (IncludeGuard == IG_IfNdefed && 887 IncludeGuardToken->TokenText == FormatTok->TokenText) { 888 IncludeGuard = IG_Defined; 889 IncludeGuardToken = nullptr; 890 for (auto &Line : Lines) { 891 if (!Line.Tokens.front().Tok->isOneOf(tok::comment, tok::hash)) { 892 IncludeGuard = IG_Rejected; 893 break; 894 } 895 } 896 } 897 898 nextToken(); 899 if (FormatTok->Tok.getKind() == tok::l_paren && 900 FormatTok->WhitespaceRange.getBegin() == 901 FormatTok->WhitespaceRange.getEnd()) { 902 parseParens(); 903 } 904 if (Style.IndentPPDirectives != FormatStyle::PPDIS_None) 905 Line->Level += PPBranchLevel + 1; 906 addUnwrappedLine(); 907 ++Line->Level; 908 909 // Errors during a preprocessor directive can only affect the layout of the 910 // preprocessor directive, and thus we ignore them. An alternative approach 911 // would be to use the same approach we use on the file level (no 912 // re-indentation if there was a structural error) within the macro 913 // definition. 914 parseFile(); 915 } 916 917 void UnwrappedLineParser::parsePPUnknown() { 918 do { 919 nextToken(); 920 } while (!eof()); 921 if (Style.IndentPPDirectives != FormatStyle::PPDIS_None) 922 Line->Level += PPBranchLevel + 1; 923 addUnwrappedLine(); 924 } 925 926 // Here we exclude certain tokens that are not usually the first token in an 927 // unwrapped line. This is used in attempt to distinguish macro calls without 928 // trailing semicolons from other constructs split to several lines. 929 static bool tokenCanStartNewLine(const FormatToken &Tok) { 930 // Semicolon can be a null-statement, l_square can be a start of a macro or 931 // a C++11 attribute, but this doesn't seem to be common. 932 return Tok.isNot(tok::semi) && Tok.isNot(tok::l_brace) && 933 Tok.isNot(TT_AttributeSquare) && 934 // Tokens that can only be used as binary operators and a part of 935 // overloaded operator names. 936 Tok.isNot(tok::period) && Tok.isNot(tok::periodstar) && 937 Tok.isNot(tok::arrow) && Tok.isNot(tok::arrowstar) && 938 Tok.isNot(tok::less) && Tok.isNot(tok::greater) && 939 Tok.isNot(tok::slash) && Tok.isNot(tok::percent) && 940 Tok.isNot(tok::lessless) && Tok.isNot(tok::greatergreater) && 941 Tok.isNot(tok::equal) && Tok.isNot(tok::plusequal) && 942 Tok.isNot(tok::minusequal) && Tok.isNot(tok::starequal) && 943 Tok.isNot(tok::slashequal) && Tok.isNot(tok::percentequal) && 944 Tok.isNot(tok::ampequal) && Tok.isNot(tok::pipeequal) && 945 Tok.isNot(tok::caretequal) && Tok.isNot(tok::greatergreaterequal) && 946 Tok.isNot(tok::lesslessequal) && 947 // Colon is used in labels, base class lists, initializer lists, 948 // range-based for loops, ternary operator, but should never be the 949 // first token in an unwrapped line. 950 Tok.isNot(tok::colon) && 951 // 'noexcept' is a trailing annotation. 952 Tok.isNot(tok::kw_noexcept); 953 } 954 955 static bool mustBeJSIdent(const AdditionalKeywords &Keywords, 956 const FormatToken *FormatTok) { 957 // FIXME: This returns true for C/C++ keywords like 'struct'. 958 return FormatTok->is(tok::identifier) && 959 (FormatTok->Tok.getIdentifierInfo() == nullptr || 960 !FormatTok->isOneOf( 961 Keywords.kw_in, Keywords.kw_of, Keywords.kw_as, Keywords.kw_async, 962 Keywords.kw_await, Keywords.kw_yield, Keywords.kw_finally, 963 Keywords.kw_function, Keywords.kw_import, Keywords.kw_is, 964 Keywords.kw_let, Keywords.kw_var, tok::kw_const, 965 Keywords.kw_abstract, Keywords.kw_extends, Keywords.kw_implements, 966 Keywords.kw_instanceof, Keywords.kw_interface, Keywords.kw_throws, 967 Keywords.kw_from)); 968 } 969 970 static bool mustBeJSIdentOrValue(const AdditionalKeywords &Keywords, 971 const FormatToken *FormatTok) { 972 return FormatTok->Tok.isLiteral() || 973 FormatTok->isOneOf(tok::kw_true, tok::kw_false) || 974 mustBeJSIdent(Keywords, FormatTok); 975 } 976 977 // isJSDeclOrStmt returns true if |FormatTok| starts a declaration or statement 978 // when encountered after a value (see mustBeJSIdentOrValue). 979 static bool isJSDeclOrStmt(const AdditionalKeywords &Keywords, 980 const FormatToken *FormatTok) { 981 return FormatTok->isOneOf( 982 tok::kw_return, Keywords.kw_yield, 983 // conditionals 984 tok::kw_if, tok::kw_else, 985 // loops 986 tok::kw_for, tok::kw_while, tok::kw_do, tok::kw_continue, tok::kw_break, 987 // switch/case 988 tok::kw_switch, tok::kw_case, 989 // exceptions 990 tok::kw_throw, tok::kw_try, tok::kw_catch, Keywords.kw_finally, 991 // declaration 992 tok::kw_const, tok::kw_class, Keywords.kw_var, Keywords.kw_let, 993 Keywords.kw_async, Keywords.kw_function, 994 // import/export 995 Keywords.kw_import, tok::kw_export); 996 } 997 998 // Checks whether a token is a type in K&R C (aka C78). 999 static bool isC78Type(const FormatToken &Tok) { 1000 return Tok.isOneOf(tok::kw_char, tok::kw_short, tok::kw_int, tok::kw_long, 1001 tok::kw_unsigned, tok::kw_float, tok::kw_double, 1002 tok::identifier); 1003 } 1004 1005 // This function checks whether a token starts the first parameter declaration 1006 // in a K&R C (aka C78) function definition, e.g.: 1007 // int f(a, b) 1008 // short a, b; 1009 // { 1010 // return a + b; 1011 // } 1012 static bool isC78ParameterDecl(const FormatToken *Tok, const FormatToken *Next, 1013 const FormatToken *FuncName) { 1014 assert(Tok); 1015 assert(Next); 1016 assert(FuncName); 1017 1018 if (FuncName->isNot(tok::identifier)) 1019 return false; 1020 1021 const FormatToken *Prev = FuncName->Previous; 1022 if (!Prev || (Prev->isNot(tok::star) && !isC78Type(*Prev))) 1023 return false; 1024 1025 if (!isC78Type(*Tok) && 1026 !Tok->isOneOf(tok::kw_register, tok::kw_struct, tok::kw_union)) 1027 return false; 1028 1029 if (Next->isNot(tok::star) && !Next->Tok.getIdentifierInfo()) 1030 return false; 1031 1032 Tok = Tok->Previous; 1033 if (!Tok || Tok->isNot(tok::r_paren)) 1034 return false; 1035 1036 Tok = Tok->Previous; 1037 if (!Tok || Tok->isNot(tok::identifier)) 1038 return false; 1039 1040 return Tok->Previous && Tok->Previous->isOneOf(tok::l_paren, tok::comma); 1041 } 1042 1043 // readTokenWithJavaScriptASI reads the next token and terminates the current 1044 // line if JavaScript Automatic Semicolon Insertion must 1045 // happen between the current token and the next token. 1046 // 1047 // This method is conservative - it cannot cover all edge cases of JavaScript, 1048 // but only aims to correctly handle certain well known cases. It *must not* 1049 // return true in speculative cases. 1050 void UnwrappedLineParser::readTokenWithJavaScriptASI() { 1051 FormatToken *Previous = FormatTok; 1052 readToken(); 1053 FormatToken *Next = FormatTok; 1054 1055 bool IsOnSameLine = 1056 CommentsBeforeNextToken.empty() 1057 ? Next->NewlinesBefore == 0 1058 : CommentsBeforeNextToken.front()->NewlinesBefore == 0; 1059 if (IsOnSameLine) 1060 return; 1061 1062 bool PreviousMustBeValue = mustBeJSIdentOrValue(Keywords, Previous); 1063 bool PreviousStartsTemplateExpr = 1064 Previous->is(TT_TemplateString) && Previous->TokenText.endswith("${"); 1065 if (PreviousMustBeValue || Previous->is(tok::r_paren)) { 1066 // If the line contains an '@' sign, the previous token might be an 1067 // annotation, which can precede another identifier/value. 1068 bool HasAt = std::find_if(Line->Tokens.begin(), Line->Tokens.end(), 1069 [](UnwrappedLineNode &LineNode) { 1070 return LineNode.Tok->is(tok::at); 1071 }) != Line->Tokens.end(); 1072 if (HasAt) 1073 return; 1074 } 1075 if (Next->is(tok::exclaim) && PreviousMustBeValue) 1076 return addUnwrappedLine(); 1077 bool NextMustBeValue = mustBeJSIdentOrValue(Keywords, Next); 1078 bool NextEndsTemplateExpr = 1079 Next->is(TT_TemplateString) && Next->TokenText.startswith("}"); 1080 if (NextMustBeValue && !NextEndsTemplateExpr && !PreviousStartsTemplateExpr && 1081 (PreviousMustBeValue || 1082 Previous->isOneOf(tok::r_square, tok::r_paren, tok::plusplus, 1083 tok::minusminus))) 1084 return addUnwrappedLine(); 1085 if ((PreviousMustBeValue || Previous->is(tok::r_paren)) && 1086 isJSDeclOrStmt(Keywords, Next)) 1087 return addUnwrappedLine(); 1088 } 1089 1090 void UnwrappedLineParser::parseStructuralElement(bool IsTopLevel) { 1091 assert(!FormatTok->is(tok::l_brace)); 1092 if (Style.Language == FormatStyle::LK_TableGen && 1093 FormatTok->is(tok::pp_include)) { 1094 nextToken(); 1095 if (FormatTok->is(tok::string_literal)) 1096 nextToken(); 1097 addUnwrappedLine(); 1098 return; 1099 } 1100 switch (FormatTok->Tok.getKind()) { 1101 case tok::kw_asm: 1102 nextToken(); 1103 if (FormatTok->is(tok::l_brace)) { 1104 FormatTok->setType(TT_InlineASMBrace); 1105 nextToken(); 1106 while (FormatTok && FormatTok->isNot(tok::eof)) { 1107 if (FormatTok->is(tok::r_brace)) { 1108 FormatTok->setType(TT_InlineASMBrace); 1109 nextToken(); 1110 addUnwrappedLine(); 1111 break; 1112 } 1113 FormatTok->Finalized = true; 1114 nextToken(); 1115 } 1116 } 1117 break; 1118 case tok::kw_namespace: 1119 parseNamespace(); 1120 return; 1121 case tok::kw_public: 1122 case tok::kw_protected: 1123 case tok::kw_private: 1124 if (Style.Language == FormatStyle::LK_Java || 1125 Style.Language == FormatStyle::LK_JavaScript || Style.isCSharp()) 1126 nextToken(); 1127 else 1128 parseAccessSpecifier(); 1129 return; 1130 case tok::kw_if: 1131 if (Style.Language == FormatStyle::LK_JavaScript && Line->MustBeDeclaration) 1132 // field/method declaration. 1133 break; 1134 parseIfThenElse(); 1135 return; 1136 case tok::kw_for: 1137 case tok::kw_while: 1138 if (Style.Language == FormatStyle::LK_JavaScript && Line->MustBeDeclaration) 1139 // field/method declaration. 1140 break; 1141 parseForOrWhileLoop(); 1142 return; 1143 case tok::kw_do: 1144 if (Style.Language == FormatStyle::LK_JavaScript && Line->MustBeDeclaration) 1145 // field/method declaration. 1146 break; 1147 parseDoWhile(); 1148 return; 1149 case tok::kw_switch: 1150 if (Style.Language == FormatStyle::LK_JavaScript && Line->MustBeDeclaration) 1151 // 'switch: string' field declaration. 1152 break; 1153 parseSwitch(); 1154 return; 1155 case tok::kw_default: 1156 if (Style.Language == FormatStyle::LK_JavaScript && Line->MustBeDeclaration) 1157 // 'default: string' field declaration. 1158 break; 1159 nextToken(); 1160 if (FormatTok->is(tok::colon)) { 1161 parseLabel(); 1162 return; 1163 } 1164 // e.g. "default void f() {}" in a Java interface. 1165 break; 1166 case tok::kw_case: 1167 if (Style.Language == FormatStyle::LK_JavaScript && Line->MustBeDeclaration) 1168 // 'case: string' field declaration. 1169 break; 1170 parseCaseLabel(); 1171 return; 1172 case tok::kw_try: 1173 case tok::kw___try: 1174 if (Style.Language == FormatStyle::LK_JavaScript && Line->MustBeDeclaration) 1175 // field/method declaration. 1176 break; 1177 parseTryCatch(); 1178 return; 1179 case tok::kw_extern: 1180 nextToken(); 1181 if (FormatTok->Tok.is(tok::string_literal)) { 1182 nextToken(); 1183 if (FormatTok->Tok.is(tok::l_brace)) { 1184 if (!Style.IndentExternBlock) { 1185 if (Style.BraceWrapping.AfterExternBlock) { 1186 addUnwrappedLine(); 1187 } 1188 unsigned AddLevels = Style.BraceWrapping.AfterExternBlock ? 1u : 0u; 1189 parseBlock(/*MustBeDeclaration=*/true, AddLevels); 1190 } else { 1191 unsigned AddLevels = 1192 Style.IndentExternBlock == FormatStyle::IEBS_Indent ? 1u : 0u; 1193 parseBlock(/*MustBeDeclaration=*/true, AddLevels); 1194 } 1195 addUnwrappedLine(); 1196 return; 1197 } 1198 } 1199 break; 1200 case tok::kw_export: 1201 if (Style.Language == FormatStyle::LK_JavaScript) { 1202 parseJavaScriptEs6ImportExport(); 1203 return; 1204 } 1205 if (!Style.isCpp()) 1206 break; 1207 // Handle C++ "(inline|export) namespace". 1208 LLVM_FALLTHROUGH; 1209 case tok::kw_inline: 1210 nextToken(); 1211 if (FormatTok->Tok.is(tok::kw_namespace)) { 1212 parseNamespace(); 1213 return; 1214 } 1215 break; 1216 case tok::identifier: 1217 if (FormatTok->is(TT_ForEachMacro)) { 1218 parseForOrWhileLoop(); 1219 return; 1220 } 1221 if (FormatTok->is(TT_MacroBlockBegin)) { 1222 parseBlock(/*MustBeDeclaration=*/false, /*AddLevels=*/1u, 1223 /*MunchSemi=*/false); 1224 return; 1225 } 1226 if (FormatTok->is(Keywords.kw_import)) { 1227 if (Style.Language == FormatStyle::LK_JavaScript) { 1228 parseJavaScriptEs6ImportExport(); 1229 return; 1230 } 1231 if (Style.Language == FormatStyle::LK_Proto) { 1232 nextToken(); 1233 if (FormatTok->is(tok::kw_public)) 1234 nextToken(); 1235 if (!FormatTok->is(tok::string_literal)) 1236 return; 1237 nextToken(); 1238 if (FormatTok->is(tok::semi)) 1239 nextToken(); 1240 addUnwrappedLine(); 1241 return; 1242 } 1243 } 1244 if (Style.isCpp() && 1245 FormatTok->isOneOf(Keywords.kw_signals, Keywords.kw_qsignals, 1246 Keywords.kw_slots, Keywords.kw_qslots)) { 1247 nextToken(); 1248 if (FormatTok->is(tok::colon)) { 1249 nextToken(); 1250 addUnwrappedLine(); 1251 return; 1252 } 1253 } 1254 if (Style.isCpp() && FormatTok->is(TT_StatementMacro)) { 1255 parseStatementMacro(); 1256 return; 1257 } 1258 if (Style.isCpp() && FormatTok->is(TT_NamespaceMacro)) { 1259 parseNamespace(); 1260 return; 1261 } 1262 // In all other cases, parse the declaration. 1263 break; 1264 default: 1265 break; 1266 } 1267 do { 1268 const FormatToken *Previous = FormatTok->Previous; 1269 switch (FormatTok->Tok.getKind()) { 1270 case tok::at: 1271 nextToken(); 1272 if (FormatTok->Tok.is(tok::l_brace)) { 1273 nextToken(); 1274 parseBracedList(); 1275 break; 1276 } else if (Style.Language == FormatStyle::LK_Java && 1277 FormatTok->is(Keywords.kw_interface)) { 1278 nextToken(); 1279 break; 1280 } 1281 switch (FormatTok->Tok.getObjCKeywordID()) { 1282 case tok::objc_public: 1283 case tok::objc_protected: 1284 case tok::objc_package: 1285 case tok::objc_private: 1286 return parseAccessSpecifier(); 1287 case tok::objc_interface: 1288 case tok::objc_implementation: 1289 return parseObjCInterfaceOrImplementation(); 1290 case tok::objc_protocol: 1291 if (parseObjCProtocol()) 1292 return; 1293 break; 1294 case tok::objc_end: 1295 return; // Handled by the caller. 1296 case tok::objc_optional: 1297 case tok::objc_required: 1298 nextToken(); 1299 addUnwrappedLine(); 1300 return; 1301 case tok::objc_autoreleasepool: 1302 nextToken(); 1303 if (FormatTok->Tok.is(tok::l_brace)) { 1304 if (Style.BraceWrapping.AfterControlStatement == 1305 FormatStyle::BWACS_Always) 1306 addUnwrappedLine(); 1307 parseBlock(/*MustBeDeclaration=*/false); 1308 } 1309 addUnwrappedLine(); 1310 return; 1311 case tok::objc_synchronized: 1312 nextToken(); 1313 if (FormatTok->Tok.is(tok::l_paren)) 1314 // Skip synchronization object 1315 parseParens(); 1316 if (FormatTok->Tok.is(tok::l_brace)) { 1317 if (Style.BraceWrapping.AfterControlStatement == 1318 FormatStyle::BWACS_Always) 1319 addUnwrappedLine(); 1320 parseBlock(/*MustBeDeclaration=*/false); 1321 } 1322 addUnwrappedLine(); 1323 return; 1324 case tok::objc_try: 1325 // This branch isn't strictly necessary (the kw_try case below would 1326 // do this too after the tok::at is parsed above). But be explicit. 1327 parseTryCatch(); 1328 return; 1329 default: 1330 break; 1331 } 1332 break; 1333 case tok::kw_concept: 1334 parseConcept(); 1335 break; 1336 case tok::kw_requires: 1337 parseRequires(); 1338 break; 1339 case tok::kw_enum: 1340 // Ignore if this is part of "template <enum ...". 1341 if (Previous && Previous->is(tok::less)) { 1342 nextToken(); 1343 break; 1344 } 1345 1346 // parseEnum falls through and does not yet add an unwrapped line as an 1347 // enum definition can start a structural element. 1348 if (!parseEnum()) 1349 break; 1350 // This only applies for C++. 1351 if (!Style.isCpp()) { 1352 addUnwrappedLine(); 1353 return; 1354 } 1355 break; 1356 case tok::kw_typedef: 1357 nextToken(); 1358 if (FormatTok->isOneOf(Keywords.kw_NS_ENUM, Keywords.kw_NS_OPTIONS, 1359 Keywords.kw_CF_ENUM, Keywords.kw_CF_OPTIONS, 1360 Keywords.kw_CF_CLOSED_ENUM, 1361 Keywords.kw_NS_CLOSED_ENUM)) 1362 parseEnum(); 1363 break; 1364 case tok::kw_struct: 1365 case tok::kw_union: 1366 case tok::kw_class: 1367 if (parseStructLike()) { 1368 return; 1369 } 1370 break; 1371 case tok::period: 1372 nextToken(); 1373 // In Java, classes have an implicit static member "class". 1374 if (Style.Language == FormatStyle::LK_Java && FormatTok && 1375 FormatTok->is(tok::kw_class)) 1376 nextToken(); 1377 if (Style.Language == FormatStyle::LK_JavaScript && FormatTok && 1378 FormatTok->Tok.getIdentifierInfo()) 1379 // JavaScript only has pseudo keywords, all keywords are allowed to 1380 // appear in "IdentifierName" positions. See http://es5.github.io/#x7.6 1381 nextToken(); 1382 break; 1383 case tok::semi: 1384 nextToken(); 1385 addUnwrappedLine(); 1386 return; 1387 case tok::r_brace: 1388 addUnwrappedLine(); 1389 return; 1390 case tok::l_paren: { 1391 parseParens(); 1392 // Break the unwrapped line if a K&R C function definition has a parameter 1393 // declaration. 1394 if (!IsTopLevel || !Style.isCpp() || !Previous || FormatTok->is(tok::eof)) 1395 break; 1396 const unsigned Position = Tokens->getPosition() + 1; 1397 assert(Position < AllTokens.size()); 1398 if (isC78ParameterDecl(FormatTok, AllTokens[Position], Previous)) { 1399 addUnwrappedLine(); 1400 return; 1401 } 1402 break; 1403 } 1404 case tok::kw_operator: 1405 nextToken(); 1406 if (FormatTok->isBinaryOperator()) 1407 nextToken(); 1408 break; 1409 case tok::caret: 1410 nextToken(); 1411 if (FormatTok->Tok.isAnyIdentifier() || 1412 FormatTok->isSimpleTypeSpecifier()) 1413 nextToken(); 1414 if (FormatTok->is(tok::l_paren)) 1415 parseParens(); 1416 if (FormatTok->is(tok::l_brace)) 1417 parseChildBlock(); 1418 break; 1419 case tok::l_brace: 1420 if (!tryToParsePropertyAccessor() && !tryToParseBracedList()) { 1421 // A block outside of parentheses must be the last part of a 1422 // structural element. 1423 // FIXME: Figure out cases where this is not true, and add projections 1424 // for them (the one we know is missing are lambdas). 1425 if (Style.BraceWrapping.AfterFunction) 1426 addUnwrappedLine(); 1427 FormatTok->setType(TT_FunctionLBrace); 1428 parseBlock(/*MustBeDeclaration=*/false); 1429 addUnwrappedLine(); 1430 return; 1431 } 1432 // Otherwise this was a braced init list, and the structural 1433 // element continues. 1434 break; 1435 case tok::kw_try: 1436 if (Style.Language == FormatStyle::LK_JavaScript && 1437 Line->MustBeDeclaration) { 1438 // field/method declaration. 1439 nextToken(); 1440 break; 1441 } 1442 // We arrive here when parsing function-try blocks. 1443 if (Style.BraceWrapping.AfterFunction) 1444 addUnwrappedLine(); 1445 parseTryCatch(); 1446 return; 1447 case tok::identifier: { 1448 if (Style.isCSharp() && FormatTok->is(Keywords.kw_where) && 1449 Line->MustBeDeclaration) { 1450 addUnwrappedLine(); 1451 parseCSharpGenericTypeConstraint(); 1452 break; 1453 } 1454 if (FormatTok->is(TT_MacroBlockEnd)) { 1455 addUnwrappedLine(); 1456 return; 1457 } 1458 1459 // Function declarations (as opposed to function expressions) are parsed 1460 // on their own unwrapped line by continuing this loop. Function 1461 // expressions (functions that are not on their own line) must not create 1462 // a new unwrapped line, so they are special cased below. 1463 size_t TokenCount = Line->Tokens.size(); 1464 if (Style.Language == FormatStyle::LK_JavaScript && 1465 FormatTok->is(Keywords.kw_function) && 1466 (TokenCount > 1 || (TokenCount == 1 && !Line->Tokens.front().Tok->is( 1467 Keywords.kw_async)))) { 1468 tryToParseJSFunction(); 1469 break; 1470 } 1471 if ((Style.Language == FormatStyle::LK_JavaScript || 1472 Style.Language == FormatStyle::LK_Java) && 1473 FormatTok->is(Keywords.kw_interface)) { 1474 if (Style.Language == FormatStyle::LK_JavaScript) { 1475 // In JavaScript/TypeScript, "interface" can be used as a standalone 1476 // identifier, e.g. in `var interface = 1;`. If "interface" is 1477 // followed by another identifier, it is very like to be an actual 1478 // interface declaration. 1479 unsigned StoredPosition = Tokens->getPosition(); 1480 FormatToken *Next = Tokens->getNextToken(); 1481 FormatTok = Tokens->setPosition(StoredPosition); 1482 if (Next && !mustBeJSIdent(Keywords, Next)) { 1483 nextToken(); 1484 break; 1485 } 1486 } 1487 parseRecord(); 1488 addUnwrappedLine(); 1489 return; 1490 } 1491 1492 if (FormatTok->is(Keywords.kw_interface)) { 1493 if (parseStructLike()) { 1494 return; 1495 } 1496 break; 1497 } 1498 1499 if (Style.isCpp() && FormatTok->is(TT_StatementMacro)) { 1500 parseStatementMacro(); 1501 return; 1502 } 1503 1504 // See if the following token should start a new unwrapped line. 1505 StringRef Text = FormatTok->TokenText; 1506 nextToken(); 1507 1508 // JS doesn't have macros, and within classes colons indicate fields, not 1509 // labels. 1510 if (Style.Language == FormatStyle::LK_JavaScript) 1511 break; 1512 1513 TokenCount = Line->Tokens.size(); 1514 if (TokenCount == 1 || 1515 (TokenCount == 2 && Line->Tokens.front().Tok->is(tok::comment))) { 1516 if (FormatTok->Tok.is(tok::colon) && !Line->MustBeDeclaration) { 1517 Line->Tokens.begin()->Tok->MustBreakBefore = true; 1518 parseLabel(!Style.IndentGotoLabels); 1519 return; 1520 } 1521 // Recognize function-like macro usages without trailing semicolon as 1522 // well as free-standing macros like Q_OBJECT. 1523 bool FunctionLike = FormatTok->is(tok::l_paren); 1524 if (FunctionLike) 1525 parseParens(); 1526 1527 bool FollowedByNewline = 1528 CommentsBeforeNextToken.empty() 1529 ? FormatTok->NewlinesBefore > 0 1530 : CommentsBeforeNextToken.front()->NewlinesBefore > 0; 1531 1532 if (FollowedByNewline && (Text.size() >= 5 || FunctionLike) && 1533 tokenCanStartNewLine(*FormatTok) && Text == Text.upper()) { 1534 addUnwrappedLine(); 1535 return; 1536 } 1537 } 1538 break; 1539 } 1540 case tok::equal: 1541 // Fat arrows (=>) have tok::TokenKind tok::equal but TokenType 1542 // TT_FatArrow. They always start an expression or a child block if 1543 // followed by a curly brace. 1544 if (FormatTok->is(TT_FatArrow)) { 1545 nextToken(); 1546 if (FormatTok->is(tok::l_brace)) { 1547 // C# may break after => if the next character is a newline. 1548 if (Style.isCSharp() && Style.BraceWrapping.AfterFunction == true) { 1549 // calling `addUnwrappedLine()` here causes odd parsing errors. 1550 FormatTok->MustBreakBefore = true; 1551 } 1552 parseChildBlock(); 1553 } 1554 break; 1555 } 1556 1557 nextToken(); 1558 if (FormatTok->Tok.is(tok::l_brace)) { 1559 // Block kind should probably be set to BK_BracedInit for any language. 1560 // C# needs this change to ensure that array initialisers and object 1561 // initialisers are indented the same way. 1562 if (Style.isCSharp()) 1563 FormatTok->setBlockKind(BK_BracedInit); 1564 nextToken(); 1565 parseBracedList(); 1566 } else if (Style.Language == FormatStyle::LK_Proto && 1567 FormatTok->Tok.is(tok::less)) { 1568 nextToken(); 1569 parseBracedList(/*ContinueOnSemicolons=*/false, /*IsEnum=*/false, 1570 /*ClosingBraceKind=*/tok::greater); 1571 } 1572 break; 1573 case tok::l_square: 1574 parseSquare(); 1575 break; 1576 case tok::kw_new: 1577 parseNew(); 1578 break; 1579 default: 1580 nextToken(); 1581 break; 1582 } 1583 } while (!eof()); 1584 } 1585 1586 bool UnwrappedLineParser::tryToParsePropertyAccessor() { 1587 assert(FormatTok->is(tok::l_brace)); 1588 if (!Style.isCSharp()) 1589 return false; 1590 // See if it's a property accessor. 1591 if (FormatTok->Previous->isNot(tok::identifier)) 1592 return false; 1593 1594 // See if we are inside a property accessor. 1595 // 1596 // Record the current tokenPosition so that we can advance and 1597 // reset the current token. `Next` is not set yet so we need 1598 // another way to advance along the token stream. 1599 unsigned int StoredPosition = Tokens->getPosition(); 1600 FormatToken *Tok = Tokens->getNextToken(); 1601 1602 // A trivial property accessor is of the form: 1603 // { [ACCESS_SPECIFIER] [get]; [ACCESS_SPECIFIER] [set] } 1604 // Track these as they do not require line breaks to be introduced. 1605 bool HasGetOrSet = false; 1606 bool IsTrivialPropertyAccessor = true; 1607 while (!eof()) { 1608 if (Tok->isOneOf(tok::semi, tok::kw_public, tok::kw_private, 1609 tok::kw_protected, Keywords.kw_internal, Keywords.kw_get, 1610 Keywords.kw_set)) { 1611 if (Tok->isOneOf(Keywords.kw_get, Keywords.kw_set)) 1612 HasGetOrSet = true; 1613 Tok = Tokens->getNextToken(); 1614 continue; 1615 } 1616 if (Tok->isNot(tok::r_brace)) 1617 IsTrivialPropertyAccessor = false; 1618 break; 1619 } 1620 1621 if (!HasGetOrSet) { 1622 Tokens->setPosition(StoredPosition); 1623 return false; 1624 } 1625 1626 // Try to parse the property accessor: 1627 // https://docs.microsoft.com/en-us/dotnet/csharp/programming-guide/classes-and-structs/properties 1628 Tokens->setPosition(StoredPosition); 1629 if (!IsTrivialPropertyAccessor && Style.BraceWrapping.AfterFunction == true) 1630 addUnwrappedLine(); 1631 nextToken(); 1632 do { 1633 switch (FormatTok->Tok.getKind()) { 1634 case tok::r_brace: 1635 nextToken(); 1636 if (FormatTok->is(tok::equal)) { 1637 while (!eof() && FormatTok->isNot(tok::semi)) 1638 nextToken(); 1639 nextToken(); 1640 } 1641 addUnwrappedLine(); 1642 return true; 1643 case tok::l_brace: 1644 ++Line->Level; 1645 parseBlock(/*MustBeDeclaration=*/true); 1646 addUnwrappedLine(); 1647 --Line->Level; 1648 break; 1649 case tok::equal: 1650 if (FormatTok->is(TT_FatArrow)) { 1651 ++Line->Level; 1652 do { 1653 nextToken(); 1654 } while (!eof() && FormatTok->isNot(tok::semi)); 1655 nextToken(); 1656 addUnwrappedLine(); 1657 --Line->Level; 1658 break; 1659 } 1660 nextToken(); 1661 break; 1662 default: 1663 if (FormatTok->isOneOf(Keywords.kw_get, Keywords.kw_set) && 1664 !IsTrivialPropertyAccessor) { 1665 // Non-trivial get/set needs to be on its own line. 1666 addUnwrappedLine(); 1667 } 1668 nextToken(); 1669 } 1670 } while (!eof()); 1671 1672 // Unreachable for well-formed code (paired '{' and '}'). 1673 return true; 1674 } 1675 1676 bool UnwrappedLineParser::tryToParseLambda() { 1677 if (!Style.isCpp()) { 1678 nextToken(); 1679 return false; 1680 } 1681 assert(FormatTok->is(tok::l_square)); 1682 FormatToken &LSquare = *FormatTok; 1683 if (!tryToParseLambdaIntroducer()) 1684 return false; 1685 1686 bool SeenArrow = false; 1687 1688 while (FormatTok->isNot(tok::l_brace)) { 1689 if (FormatTok->isSimpleTypeSpecifier()) { 1690 nextToken(); 1691 continue; 1692 } 1693 switch (FormatTok->Tok.getKind()) { 1694 case tok::l_brace: 1695 break; 1696 case tok::l_paren: 1697 parseParens(); 1698 break; 1699 case tok::amp: 1700 case tok::star: 1701 case tok::kw_const: 1702 case tok::comma: 1703 case tok::less: 1704 case tok::greater: 1705 case tok::identifier: 1706 case tok::numeric_constant: 1707 case tok::coloncolon: 1708 case tok::kw_class: 1709 case tok::kw_mutable: 1710 case tok::kw_noexcept: 1711 case tok::kw_template: 1712 case tok::kw_typename: 1713 nextToken(); 1714 break; 1715 // Specialization of a template with an integer parameter can contain 1716 // arithmetic, logical, comparison and ternary operators. 1717 // 1718 // FIXME: This also accepts sequences of operators that are not in the scope 1719 // of a template argument list. 1720 // 1721 // In a C++ lambda a template type can only occur after an arrow. We use 1722 // this as an heuristic to distinguish between Objective-C expressions 1723 // followed by an `a->b` expression, such as: 1724 // ([obj func:arg] + a->b) 1725 // Otherwise the code below would parse as a lambda. 1726 // 1727 // FIXME: This heuristic is incorrect for C++20 generic lambdas with 1728 // explicit template lists: []<bool b = true && false>(U &&u){} 1729 case tok::plus: 1730 case tok::minus: 1731 case tok::exclaim: 1732 case tok::tilde: 1733 case tok::slash: 1734 case tok::percent: 1735 case tok::lessless: 1736 case tok::pipe: 1737 case tok::pipepipe: 1738 case tok::ampamp: 1739 case tok::caret: 1740 case tok::equalequal: 1741 case tok::exclaimequal: 1742 case tok::greaterequal: 1743 case tok::lessequal: 1744 case tok::question: 1745 case tok::colon: 1746 case tok::ellipsis: 1747 case tok::kw_true: 1748 case tok::kw_false: 1749 if (SeenArrow) { 1750 nextToken(); 1751 break; 1752 } 1753 return true; 1754 case tok::arrow: 1755 // This might or might not actually be a lambda arrow (this could be an 1756 // ObjC method invocation followed by a dereferencing arrow). We might 1757 // reset this back to TT_Unknown in TokenAnnotator. 1758 FormatTok->setType(TT_LambdaArrow); 1759 SeenArrow = true; 1760 nextToken(); 1761 break; 1762 default: 1763 return true; 1764 } 1765 } 1766 FormatTok->setType(TT_LambdaLBrace); 1767 LSquare.setType(TT_LambdaLSquare); 1768 parseChildBlock(); 1769 return true; 1770 } 1771 1772 bool UnwrappedLineParser::tryToParseLambdaIntroducer() { 1773 const FormatToken *Previous = FormatTok->Previous; 1774 if (Previous && 1775 (Previous->isOneOf(tok::identifier, tok::kw_operator, tok::kw_new, 1776 tok::kw_delete, tok::l_square) || 1777 FormatTok->isCppStructuredBinding(Style) || Previous->closesScope() || 1778 Previous->isSimpleTypeSpecifier())) { 1779 nextToken(); 1780 return false; 1781 } 1782 nextToken(); 1783 if (FormatTok->is(tok::l_square)) { 1784 return false; 1785 } 1786 parseSquare(/*LambdaIntroducer=*/true); 1787 return true; 1788 } 1789 1790 void UnwrappedLineParser::tryToParseJSFunction() { 1791 assert(FormatTok->is(Keywords.kw_function) || 1792 FormatTok->startsSequence(Keywords.kw_async, Keywords.kw_function)); 1793 if (FormatTok->is(Keywords.kw_async)) 1794 nextToken(); 1795 // Consume "function". 1796 nextToken(); 1797 1798 // Consume * (generator function). Treat it like C++'s overloaded operators. 1799 if (FormatTok->is(tok::star)) { 1800 FormatTok->setType(TT_OverloadedOperator); 1801 nextToken(); 1802 } 1803 1804 // Consume function name. 1805 if (FormatTok->is(tok::identifier)) 1806 nextToken(); 1807 1808 if (FormatTok->isNot(tok::l_paren)) 1809 return; 1810 1811 // Parse formal parameter list. 1812 parseParens(); 1813 1814 if (FormatTok->is(tok::colon)) { 1815 // Parse a type definition. 1816 nextToken(); 1817 1818 // Eat the type declaration. For braced inline object types, balance braces, 1819 // otherwise just parse until finding an l_brace for the function body. 1820 if (FormatTok->is(tok::l_brace)) 1821 tryToParseBracedList(); 1822 else 1823 while (!FormatTok->isOneOf(tok::l_brace, tok::semi) && !eof()) 1824 nextToken(); 1825 } 1826 1827 if (FormatTok->is(tok::semi)) 1828 return; 1829 1830 parseChildBlock(); 1831 } 1832 1833 bool UnwrappedLineParser::tryToParseBracedList() { 1834 if (FormatTok->is(BK_Unknown)) 1835 calculateBraceTypes(); 1836 assert(FormatTok->isNot(BK_Unknown)); 1837 if (FormatTok->is(BK_Block)) 1838 return false; 1839 nextToken(); 1840 parseBracedList(); 1841 return true; 1842 } 1843 1844 bool UnwrappedLineParser::parseBracedList(bool ContinueOnSemicolons, 1845 bool IsEnum, 1846 tok::TokenKind ClosingBraceKind) { 1847 bool HasError = false; 1848 1849 // FIXME: Once we have an expression parser in the UnwrappedLineParser, 1850 // replace this by using parseAssignmentExpression() inside. 1851 do { 1852 if (Style.isCSharp()) { 1853 // Fat arrows (=>) have tok::TokenKind tok::equal but TokenType 1854 // TT_FatArrow. They always start an expression or a child block if 1855 // followed by a curly brace. 1856 if (FormatTok->is(TT_FatArrow)) { 1857 nextToken(); 1858 if (FormatTok->is(tok::l_brace)) { 1859 // C# may break after => if the next character is a newline. 1860 if (Style.isCSharp() && Style.BraceWrapping.AfterFunction == true) { 1861 // calling `addUnwrappedLine()` here causes odd parsing errors. 1862 FormatTok->MustBreakBefore = true; 1863 } 1864 parseChildBlock(); 1865 continue; 1866 } 1867 } 1868 } 1869 if (Style.Language == FormatStyle::LK_JavaScript) { 1870 if (FormatTok->is(Keywords.kw_function) || 1871 FormatTok->startsSequence(Keywords.kw_async, Keywords.kw_function)) { 1872 tryToParseJSFunction(); 1873 continue; 1874 } 1875 if (FormatTok->is(TT_FatArrow)) { 1876 nextToken(); 1877 // Fat arrows can be followed by simple expressions or by child blocks 1878 // in curly braces. 1879 if (FormatTok->is(tok::l_brace)) { 1880 parseChildBlock(); 1881 continue; 1882 } 1883 } 1884 if (FormatTok->is(tok::l_brace)) { 1885 // Could be a method inside of a braced list `{a() { return 1; }}`. 1886 if (tryToParseBracedList()) 1887 continue; 1888 parseChildBlock(); 1889 } 1890 } 1891 if (FormatTok->Tok.getKind() == ClosingBraceKind) { 1892 if (IsEnum && !Style.AllowShortEnumsOnASingleLine) 1893 addUnwrappedLine(); 1894 nextToken(); 1895 return !HasError; 1896 } 1897 switch (FormatTok->Tok.getKind()) { 1898 case tok::caret: 1899 nextToken(); 1900 if (FormatTok->is(tok::l_brace)) { 1901 parseChildBlock(); 1902 } 1903 break; 1904 case tok::l_square: 1905 if (Style.isCSharp()) 1906 parseSquare(); 1907 else 1908 tryToParseLambda(); 1909 break; 1910 case tok::l_paren: 1911 parseParens(); 1912 // JavaScript can just have free standing methods and getters/setters in 1913 // object literals. Detect them by a "{" following ")". 1914 if (Style.Language == FormatStyle::LK_JavaScript) { 1915 if (FormatTok->is(tok::l_brace)) 1916 parseChildBlock(); 1917 break; 1918 } 1919 break; 1920 case tok::l_brace: 1921 // Assume there are no blocks inside a braced init list apart 1922 // from the ones we explicitly parse out (like lambdas). 1923 FormatTok->setBlockKind(BK_BracedInit); 1924 nextToken(); 1925 parseBracedList(); 1926 break; 1927 case tok::less: 1928 if (Style.Language == FormatStyle::LK_Proto) { 1929 nextToken(); 1930 parseBracedList(/*ContinueOnSemicolons=*/false, /*IsEnum=*/false, 1931 /*ClosingBraceKind=*/tok::greater); 1932 } else { 1933 nextToken(); 1934 } 1935 break; 1936 case tok::semi: 1937 // JavaScript (or more precisely TypeScript) can have semicolons in braced 1938 // lists (in so-called TypeMemberLists). Thus, the semicolon cannot be 1939 // used for error recovery if we have otherwise determined that this is 1940 // a braced list. 1941 if (Style.Language == FormatStyle::LK_JavaScript) { 1942 nextToken(); 1943 break; 1944 } 1945 HasError = true; 1946 if (!ContinueOnSemicolons) 1947 return !HasError; 1948 nextToken(); 1949 break; 1950 case tok::comma: 1951 nextToken(); 1952 if (IsEnum && !Style.AllowShortEnumsOnASingleLine) 1953 addUnwrappedLine(); 1954 break; 1955 default: 1956 nextToken(); 1957 break; 1958 } 1959 } while (!eof()); 1960 return false; 1961 } 1962 1963 void UnwrappedLineParser::parseParens() { 1964 assert(FormatTok->Tok.is(tok::l_paren) && "'(' expected."); 1965 nextToken(); 1966 do { 1967 switch (FormatTok->Tok.getKind()) { 1968 case tok::l_paren: 1969 parseParens(); 1970 if (Style.Language == FormatStyle::LK_Java && FormatTok->is(tok::l_brace)) 1971 parseChildBlock(); 1972 break; 1973 case tok::r_paren: 1974 nextToken(); 1975 return; 1976 case tok::r_brace: 1977 // A "}" inside parenthesis is an error if there wasn't a matching "{". 1978 return; 1979 case tok::l_square: 1980 tryToParseLambda(); 1981 break; 1982 case tok::l_brace: 1983 if (!tryToParseBracedList()) 1984 parseChildBlock(); 1985 break; 1986 case tok::at: 1987 nextToken(); 1988 if (FormatTok->Tok.is(tok::l_brace)) { 1989 nextToken(); 1990 parseBracedList(); 1991 } 1992 break; 1993 case tok::equal: 1994 if (Style.isCSharp() && FormatTok->is(TT_FatArrow)) 1995 parseStructuralElement(); 1996 else 1997 nextToken(); 1998 break; 1999 case tok::kw_class: 2000 if (Style.Language == FormatStyle::LK_JavaScript) 2001 parseRecord(/*ParseAsExpr=*/true); 2002 else 2003 nextToken(); 2004 break; 2005 case tok::identifier: 2006 if (Style.Language == FormatStyle::LK_JavaScript && 2007 (FormatTok->is(Keywords.kw_function) || 2008 FormatTok->startsSequence(Keywords.kw_async, Keywords.kw_function))) 2009 tryToParseJSFunction(); 2010 else 2011 nextToken(); 2012 break; 2013 default: 2014 nextToken(); 2015 break; 2016 } 2017 } while (!eof()); 2018 } 2019 2020 void UnwrappedLineParser::parseSquare(bool LambdaIntroducer) { 2021 if (!LambdaIntroducer) { 2022 assert(FormatTok->Tok.is(tok::l_square) && "'[' expected."); 2023 if (tryToParseLambda()) 2024 return; 2025 } 2026 do { 2027 switch (FormatTok->Tok.getKind()) { 2028 case tok::l_paren: 2029 parseParens(); 2030 break; 2031 case tok::r_square: 2032 nextToken(); 2033 return; 2034 case tok::r_brace: 2035 // A "}" inside parenthesis is an error if there wasn't a matching "{". 2036 return; 2037 case tok::l_square: 2038 parseSquare(); 2039 break; 2040 case tok::l_brace: { 2041 if (!tryToParseBracedList()) 2042 parseChildBlock(); 2043 break; 2044 } 2045 case tok::at: 2046 nextToken(); 2047 if (FormatTok->Tok.is(tok::l_brace)) { 2048 nextToken(); 2049 parseBracedList(); 2050 } 2051 break; 2052 default: 2053 nextToken(); 2054 break; 2055 } 2056 } while (!eof()); 2057 } 2058 2059 void UnwrappedLineParser::parseIfThenElse() { 2060 assert(FormatTok->Tok.is(tok::kw_if) && "'if' expected"); 2061 nextToken(); 2062 if (FormatTok->Tok.isOneOf(tok::kw_constexpr, tok::identifier)) 2063 nextToken(); 2064 if (FormatTok->Tok.is(tok::l_paren)) 2065 parseParens(); 2066 // handle [[likely]] / [[unlikely]] 2067 if (FormatTok->is(tok::l_square) && tryToParseSimpleAttribute()) 2068 parseSquare(); 2069 bool NeedsUnwrappedLine = false; 2070 if (FormatTok->Tok.is(tok::l_brace)) { 2071 CompoundStatementIndenter Indenter(this, Style, Line->Level); 2072 parseBlock(/*MustBeDeclaration=*/false); 2073 if (Style.BraceWrapping.BeforeElse) 2074 addUnwrappedLine(); 2075 else 2076 NeedsUnwrappedLine = true; 2077 } else { 2078 addUnwrappedLine(); 2079 ++Line->Level; 2080 parseStructuralElement(); 2081 --Line->Level; 2082 } 2083 if (FormatTok->Tok.is(tok::kw_else)) { 2084 nextToken(); 2085 // handle [[likely]] / [[unlikely]] 2086 if (FormatTok->Tok.is(tok::l_square) && tryToParseSimpleAttribute()) 2087 parseSquare(); 2088 if (FormatTok->Tok.is(tok::l_brace)) { 2089 CompoundStatementIndenter Indenter(this, Style, Line->Level); 2090 parseBlock(/*MustBeDeclaration=*/false); 2091 addUnwrappedLine(); 2092 } else if (FormatTok->Tok.is(tok::kw_if)) { 2093 FormatToken *Previous = AllTokens[Tokens->getPosition() - 1]; 2094 bool PrecededByComment = Previous->is(tok::comment); 2095 if (PrecededByComment) { 2096 addUnwrappedLine(); 2097 ++Line->Level; 2098 } 2099 parseIfThenElse(); 2100 if (PrecededByComment) 2101 --Line->Level; 2102 } else { 2103 addUnwrappedLine(); 2104 ++Line->Level; 2105 parseStructuralElement(); 2106 if (FormatTok->is(tok::eof)) 2107 addUnwrappedLine(); 2108 --Line->Level; 2109 } 2110 } else if (NeedsUnwrappedLine) { 2111 addUnwrappedLine(); 2112 } 2113 } 2114 2115 void UnwrappedLineParser::parseTryCatch() { 2116 assert(FormatTok->isOneOf(tok::kw_try, tok::kw___try) && "'try' expected"); 2117 nextToken(); 2118 bool NeedsUnwrappedLine = false; 2119 if (FormatTok->is(tok::colon)) { 2120 // We are in a function try block, what comes is an initializer list. 2121 nextToken(); 2122 2123 // In case identifiers were removed by clang-tidy, what might follow is 2124 // multiple commas in sequence - before the first identifier. 2125 while (FormatTok->is(tok::comma)) 2126 nextToken(); 2127 2128 while (FormatTok->is(tok::identifier)) { 2129 nextToken(); 2130 if (FormatTok->is(tok::l_paren)) 2131 parseParens(); 2132 if (FormatTok->Previous && FormatTok->Previous->is(tok::identifier) && 2133 FormatTok->is(tok::l_brace)) { 2134 do { 2135 nextToken(); 2136 } while (!FormatTok->is(tok::r_brace)); 2137 nextToken(); 2138 } 2139 2140 // In case identifiers were removed by clang-tidy, what might follow is 2141 // multiple commas in sequence - after the first identifier. 2142 while (FormatTok->is(tok::comma)) 2143 nextToken(); 2144 } 2145 } 2146 // Parse try with resource. 2147 if (Style.Language == FormatStyle::LK_Java && FormatTok->is(tok::l_paren)) { 2148 parseParens(); 2149 } 2150 if (FormatTok->is(tok::l_brace)) { 2151 CompoundStatementIndenter Indenter(this, Style, Line->Level); 2152 parseBlock(/*MustBeDeclaration=*/false); 2153 if (Style.BraceWrapping.BeforeCatch) { 2154 addUnwrappedLine(); 2155 } else { 2156 NeedsUnwrappedLine = true; 2157 } 2158 } else if (!FormatTok->is(tok::kw_catch)) { 2159 // The C++ standard requires a compound-statement after a try. 2160 // If there's none, we try to assume there's a structuralElement 2161 // and try to continue. 2162 addUnwrappedLine(); 2163 ++Line->Level; 2164 parseStructuralElement(); 2165 --Line->Level; 2166 } 2167 while (1) { 2168 if (FormatTok->is(tok::at)) 2169 nextToken(); 2170 if (!(FormatTok->isOneOf(tok::kw_catch, Keywords.kw___except, 2171 tok::kw___finally) || 2172 ((Style.Language == FormatStyle::LK_Java || 2173 Style.Language == FormatStyle::LK_JavaScript) && 2174 FormatTok->is(Keywords.kw_finally)) || 2175 (FormatTok->Tok.isObjCAtKeyword(tok::objc_catch) || 2176 FormatTok->Tok.isObjCAtKeyword(tok::objc_finally)))) 2177 break; 2178 nextToken(); 2179 while (FormatTok->isNot(tok::l_brace)) { 2180 if (FormatTok->is(tok::l_paren)) { 2181 parseParens(); 2182 continue; 2183 } 2184 if (FormatTok->isOneOf(tok::semi, tok::r_brace, tok::eof)) 2185 return; 2186 nextToken(); 2187 } 2188 NeedsUnwrappedLine = false; 2189 CompoundStatementIndenter Indenter(this, Style, Line->Level); 2190 parseBlock(/*MustBeDeclaration=*/false); 2191 if (Style.BraceWrapping.BeforeCatch) 2192 addUnwrappedLine(); 2193 else 2194 NeedsUnwrappedLine = true; 2195 } 2196 if (NeedsUnwrappedLine) 2197 addUnwrappedLine(); 2198 } 2199 2200 void UnwrappedLineParser::parseNamespace() { 2201 assert(FormatTok->isOneOf(tok::kw_namespace, TT_NamespaceMacro) && 2202 "'namespace' expected"); 2203 2204 const FormatToken &InitialToken = *FormatTok; 2205 nextToken(); 2206 if (InitialToken.is(TT_NamespaceMacro)) { 2207 parseParens(); 2208 } else { 2209 while (FormatTok->isOneOf(tok::identifier, tok::coloncolon, tok::kw_inline, 2210 tok::l_square)) { 2211 if (FormatTok->is(tok::l_square)) 2212 parseSquare(); 2213 else 2214 nextToken(); 2215 } 2216 } 2217 if (FormatTok->Tok.is(tok::l_brace)) { 2218 if (ShouldBreakBeforeBrace(Style, InitialToken)) 2219 addUnwrappedLine(); 2220 2221 unsigned AddLevels = 2222 Style.NamespaceIndentation == FormatStyle::NI_All || 2223 (Style.NamespaceIndentation == FormatStyle::NI_Inner && 2224 DeclarationScopeStack.size() > 1) 2225 ? 1u 2226 : 0u; 2227 bool ManageWhitesmithsBraces = 2228 AddLevels == 0u && 2229 Style.BreakBeforeBraces == FormatStyle::BS_Whitesmiths; 2230 2231 // If we're in Whitesmiths mode, indent the brace if we're not indenting 2232 // the whole block. 2233 if (ManageWhitesmithsBraces) 2234 ++Line->Level; 2235 2236 parseBlock(/*MustBeDeclaration=*/true, AddLevels, 2237 /*MunchSemi=*/true, 2238 /*UnindentWhitesmithsBraces=*/ManageWhitesmithsBraces); 2239 2240 // Munch the semicolon after a namespace. This is more common than one would 2241 // think. Putting the semicolon into its own line is very ugly. 2242 if (FormatTok->Tok.is(tok::semi)) 2243 nextToken(); 2244 2245 addUnwrappedLine(AddLevels > 0 ? LineLevel::Remove : LineLevel::Keep); 2246 2247 if (ManageWhitesmithsBraces) 2248 --Line->Level; 2249 } 2250 // FIXME: Add error handling. 2251 } 2252 2253 void UnwrappedLineParser::parseNew() { 2254 assert(FormatTok->is(tok::kw_new) && "'new' expected"); 2255 nextToken(); 2256 2257 if (Style.isCSharp()) { 2258 do { 2259 if (FormatTok->is(tok::l_brace)) 2260 parseBracedList(); 2261 2262 if (FormatTok->isOneOf(tok::semi, tok::comma)) 2263 return; 2264 2265 nextToken(); 2266 } while (!eof()); 2267 } 2268 2269 if (Style.Language != FormatStyle::LK_Java) 2270 return; 2271 2272 // In Java, we can parse everything up to the parens, which aren't optional. 2273 do { 2274 // There should not be a ;, { or } before the new's open paren. 2275 if (FormatTok->isOneOf(tok::semi, tok::l_brace, tok::r_brace)) 2276 return; 2277 2278 // Consume the parens. 2279 if (FormatTok->is(tok::l_paren)) { 2280 parseParens(); 2281 2282 // If there is a class body of an anonymous class, consume that as child. 2283 if (FormatTok->is(tok::l_brace)) 2284 parseChildBlock(); 2285 return; 2286 } 2287 nextToken(); 2288 } while (!eof()); 2289 } 2290 2291 void UnwrappedLineParser::parseForOrWhileLoop() { 2292 assert(FormatTok->isOneOf(tok::kw_for, tok::kw_while, TT_ForEachMacro) && 2293 "'for', 'while' or foreach macro expected"); 2294 nextToken(); 2295 // JS' for await ( ... 2296 if (Style.Language == FormatStyle::LK_JavaScript && 2297 FormatTok->is(Keywords.kw_await)) 2298 nextToken(); 2299 if (FormatTok->Tok.is(tok::l_paren)) 2300 parseParens(); 2301 if (FormatTok->Tok.is(tok::l_brace)) { 2302 CompoundStatementIndenter Indenter(this, Style, Line->Level); 2303 parseBlock(/*MustBeDeclaration=*/false); 2304 addUnwrappedLine(); 2305 } else { 2306 addUnwrappedLine(); 2307 ++Line->Level; 2308 parseStructuralElement(); 2309 --Line->Level; 2310 } 2311 } 2312 2313 void UnwrappedLineParser::parseDoWhile() { 2314 assert(FormatTok->Tok.is(tok::kw_do) && "'do' expected"); 2315 nextToken(); 2316 if (FormatTok->Tok.is(tok::l_brace)) { 2317 CompoundStatementIndenter Indenter(this, Style, Line->Level); 2318 parseBlock(/*MustBeDeclaration=*/false); 2319 if (Style.BraceWrapping.BeforeWhile) 2320 addUnwrappedLine(); 2321 } else { 2322 addUnwrappedLine(); 2323 ++Line->Level; 2324 parseStructuralElement(); 2325 --Line->Level; 2326 } 2327 2328 // FIXME: Add error handling. 2329 if (!FormatTok->Tok.is(tok::kw_while)) { 2330 addUnwrappedLine(); 2331 return; 2332 } 2333 2334 // If in Whitesmiths mode, the line with the while() needs to be indented 2335 // to the same level as the block. 2336 if (Style.BreakBeforeBraces == FormatStyle::BS_Whitesmiths) 2337 ++Line->Level; 2338 2339 nextToken(); 2340 parseStructuralElement(); 2341 } 2342 2343 void UnwrappedLineParser::parseLabel(bool LeftAlignLabel) { 2344 nextToken(); 2345 unsigned OldLineLevel = Line->Level; 2346 if (Line->Level > 1 || (!Line->InPPDirective && Line->Level > 0)) 2347 --Line->Level; 2348 if (LeftAlignLabel) 2349 Line->Level = 0; 2350 2351 if (!Style.IndentCaseBlocks && CommentsBeforeNextToken.empty() && 2352 FormatTok->Tok.is(tok::l_brace)) { 2353 2354 CompoundStatementIndenter Indenter(this, Line->Level, 2355 Style.BraceWrapping.AfterCaseLabel, 2356 Style.BraceWrapping.IndentBraces); 2357 parseBlock(/*MustBeDeclaration=*/false); 2358 if (FormatTok->Tok.is(tok::kw_break)) { 2359 if (Style.BraceWrapping.AfterControlStatement == 2360 FormatStyle::BWACS_Always) { 2361 addUnwrappedLine(); 2362 if (!Style.IndentCaseBlocks && 2363 Style.BreakBeforeBraces == FormatStyle::BS_Whitesmiths) { 2364 Line->Level++; 2365 } 2366 } 2367 parseStructuralElement(); 2368 } 2369 addUnwrappedLine(); 2370 } else { 2371 if (FormatTok->is(tok::semi)) 2372 nextToken(); 2373 addUnwrappedLine(); 2374 } 2375 Line->Level = OldLineLevel; 2376 if (FormatTok->isNot(tok::l_brace)) { 2377 parseStructuralElement(); 2378 addUnwrappedLine(); 2379 } 2380 } 2381 2382 void UnwrappedLineParser::parseCaseLabel() { 2383 assert(FormatTok->Tok.is(tok::kw_case) && "'case' expected"); 2384 2385 // FIXME: fix handling of complex expressions here. 2386 do { 2387 nextToken(); 2388 } while (!eof() && !FormatTok->Tok.is(tok::colon)); 2389 parseLabel(); 2390 } 2391 2392 void UnwrappedLineParser::parseSwitch() { 2393 assert(FormatTok->Tok.is(tok::kw_switch) && "'switch' expected"); 2394 nextToken(); 2395 if (FormatTok->Tok.is(tok::l_paren)) 2396 parseParens(); 2397 if (FormatTok->Tok.is(tok::l_brace)) { 2398 CompoundStatementIndenter Indenter(this, Style, Line->Level); 2399 parseBlock(/*MustBeDeclaration=*/false); 2400 addUnwrappedLine(); 2401 } else { 2402 addUnwrappedLine(); 2403 ++Line->Level; 2404 parseStructuralElement(); 2405 --Line->Level; 2406 } 2407 } 2408 2409 void UnwrappedLineParser::parseAccessSpecifier() { 2410 nextToken(); 2411 // Understand Qt's slots. 2412 if (FormatTok->isOneOf(Keywords.kw_slots, Keywords.kw_qslots)) 2413 nextToken(); 2414 // Otherwise, we don't know what it is, and we'd better keep the next token. 2415 if (FormatTok->Tok.is(tok::colon)) 2416 nextToken(); 2417 addUnwrappedLine(); 2418 } 2419 2420 void UnwrappedLineParser::parseConcept() { 2421 assert(FormatTok->Tok.is(tok::kw_concept) && "'concept' expected"); 2422 nextToken(); 2423 if (!FormatTok->Tok.is(tok::identifier)) 2424 return; 2425 nextToken(); 2426 if (!FormatTok->Tok.is(tok::equal)) 2427 return; 2428 nextToken(); 2429 if (FormatTok->Tok.is(tok::kw_requires)) { 2430 nextToken(); 2431 parseRequiresExpression(Line->Level); 2432 } else { 2433 parseConstraintExpression(Line->Level); 2434 } 2435 } 2436 2437 void UnwrappedLineParser::parseRequiresExpression(unsigned int OriginalLevel) { 2438 // requires (R range) 2439 if (FormatTok->Tok.is(tok::l_paren)) { 2440 parseParens(); 2441 if (Style.IndentRequires && OriginalLevel != Line->Level) { 2442 addUnwrappedLine(); 2443 --Line->Level; 2444 } 2445 } 2446 2447 if (FormatTok->Tok.is(tok::l_brace)) { 2448 if (Style.BraceWrapping.AfterFunction) 2449 addUnwrappedLine(); 2450 FormatTok->setType(TT_FunctionLBrace); 2451 parseBlock(/*MustBeDeclaration=*/false); 2452 addUnwrappedLine(); 2453 } else { 2454 parseConstraintExpression(OriginalLevel); 2455 } 2456 } 2457 2458 void UnwrappedLineParser::parseConstraintExpression( 2459 unsigned int OriginalLevel) { 2460 // requires Id<T> && Id<T> || Id<T> 2461 while ( 2462 FormatTok->isOneOf(tok::identifier, tok::kw_requires, tok::coloncolon)) { 2463 nextToken(); 2464 while (FormatTok->isOneOf(tok::identifier, tok::coloncolon, tok::less, 2465 tok::greater, tok::comma, tok::ellipsis)) { 2466 if (FormatTok->Tok.is(tok::less)) { 2467 parseBracedList(/*ContinueOnSemicolons=*/false, /*IsEnum=*/false, 2468 /*ClosingBraceKind=*/tok::greater); 2469 continue; 2470 } 2471 nextToken(); 2472 } 2473 if (FormatTok->Tok.is(tok::kw_requires)) { 2474 parseRequiresExpression(OriginalLevel); 2475 } 2476 if (FormatTok->Tok.is(tok::less)) { 2477 parseBracedList(/*ContinueOnSemicolons=*/false, /*IsEnum=*/false, 2478 /*ClosingBraceKind=*/tok::greater); 2479 } 2480 2481 if (FormatTok->Tok.is(tok::l_paren)) { 2482 parseParens(); 2483 } 2484 if (FormatTok->Tok.is(tok::l_brace)) { 2485 if (Style.BraceWrapping.AfterFunction) 2486 addUnwrappedLine(); 2487 FormatTok->setType(TT_FunctionLBrace); 2488 parseBlock(/*MustBeDeclaration=*/false); 2489 } 2490 if (FormatTok->Tok.is(tok::semi)) { 2491 // Eat any trailing semi. 2492 nextToken(); 2493 addUnwrappedLine(); 2494 } 2495 if (FormatTok->Tok.is(tok::colon)) { 2496 return; 2497 } 2498 if (!FormatTok->Tok.isOneOf(tok::ampamp, tok::pipepipe)) { 2499 if (FormatTok->Previous && 2500 !FormatTok->Previous->isOneOf(tok::identifier, tok::kw_requires, 2501 tok::coloncolon)) { 2502 addUnwrappedLine(); 2503 } 2504 if (Style.IndentRequires && OriginalLevel != Line->Level) { 2505 --Line->Level; 2506 } 2507 break; 2508 } else { 2509 FormatTok->setType(TT_ConstraintJunctions); 2510 } 2511 2512 nextToken(); 2513 } 2514 } 2515 2516 void UnwrappedLineParser::parseRequires() { 2517 assert(FormatTok->Tok.is(tok::kw_requires) && "'requires' expected"); 2518 2519 unsigned OriginalLevel = Line->Level; 2520 if (FormatTok->Previous && FormatTok->Previous->is(tok::greater)) { 2521 addUnwrappedLine(); 2522 if (Style.IndentRequires) { 2523 Line->Level++; 2524 } 2525 } 2526 nextToken(); 2527 2528 parseRequiresExpression(OriginalLevel); 2529 } 2530 2531 bool UnwrappedLineParser::parseEnum() { 2532 // Won't be 'enum' for NS_ENUMs. 2533 if (FormatTok->Tok.is(tok::kw_enum)) 2534 nextToken(); 2535 2536 // In TypeScript, "enum" can also be used as property name, e.g. in interface 2537 // declarations. An "enum" keyword followed by a colon would be a syntax 2538 // error and thus assume it is just an identifier. 2539 if (Style.Language == FormatStyle::LK_JavaScript && 2540 FormatTok->isOneOf(tok::colon, tok::question)) 2541 return false; 2542 2543 // In protobuf, "enum" can be used as a field name. 2544 if (Style.Language == FormatStyle::LK_Proto && FormatTok->is(tok::equal)) 2545 return false; 2546 2547 // Eat up enum class ... 2548 if (FormatTok->Tok.is(tok::kw_class) || FormatTok->Tok.is(tok::kw_struct)) 2549 nextToken(); 2550 2551 while (FormatTok->Tok.getIdentifierInfo() || 2552 FormatTok->isOneOf(tok::colon, tok::coloncolon, tok::less, 2553 tok::greater, tok::comma, tok::question)) { 2554 nextToken(); 2555 // We can have macros or attributes in between 'enum' and the enum name. 2556 if (FormatTok->is(tok::l_paren)) 2557 parseParens(); 2558 if (FormatTok->is(tok::identifier)) { 2559 nextToken(); 2560 // If there are two identifiers in a row, this is likely an elaborate 2561 // return type. In Java, this can be "implements", etc. 2562 if (Style.isCpp() && FormatTok->is(tok::identifier)) 2563 return false; 2564 } 2565 } 2566 2567 // Just a declaration or something is wrong. 2568 if (FormatTok->isNot(tok::l_brace)) 2569 return true; 2570 FormatTok->setBlockKind(BK_Block); 2571 2572 if (Style.Language == FormatStyle::LK_Java) { 2573 // Java enums are different. 2574 parseJavaEnumBody(); 2575 return true; 2576 } 2577 if (Style.Language == FormatStyle::LK_Proto) { 2578 parseBlock(/*MustBeDeclaration=*/true); 2579 return true; 2580 } 2581 2582 if (!Style.AllowShortEnumsOnASingleLine) 2583 addUnwrappedLine(); 2584 // Parse enum body. 2585 nextToken(); 2586 if (!Style.AllowShortEnumsOnASingleLine) { 2587 addUnwrappedLine(); 2588 Line->Level += 1; 2589 } 2590 bool HasError = !parseBracedList(/*ContinueOnSemicolons=*/true, 2591 /*IsEnum=*/true); 2592 if (!Style.AllowShortEnumsOnASingleLine) 2593 Line->Level -= 1; 2594 if (HasError) { 2595 if (FormatTok->is(tok::semi)) 2596 nextToken(); 2597 addUnwrappedLine(); 2598 } 2599 return true; 2600 2601 // There is no addUnwrappedLine() here so that we fall through to parsing a 2602 // structural element afterwards. Thus, in "enum A {} n, m;", 2603 // "} n, m;" will end up in one unwrapped line. 2604 } 2605 2606 bool UnwrappedLineParser::parseStructLike() { 2607 // parseRecord falls through and does not yet add an unwrapped line as a 2608 // record declaration or definition can start a structural element. 2609 parseRecord(); 2610 // This does not apply to Java, JavaScript and C#. 2611 if (Style.Language == FormatStyle::LK_Java || 2612 Style.Language == FormatStyle::LK_JavaScript || Style.isCSharp()) { 2613 if (FormatTok->is(tok::semi)) 2614 nextToken(); 2615 addUnwrappedLine(); 2616 return true; 2617 } 2618 return false; 2619 } 2620 2621 namespace { 2622 // A class used to set and restore the Token position when peeking 2623 // ahead in the token source. 2624 class ScopedTokenPosition { 2625 unsigned StoredPosition; 2626 FormatTokenSource *Tokens; 2627 2628 public: 2629 ScopedTokenPosition(FormatTokenSource *Tokens) : Tokens(Tokens) { 2630 assert(Tokens && "Tokens expected to not be null"); 2631 StoredPosition = Tokens->getPosition(); 2632 } 2633 2634 ~ScopedTokenPosition() { Tokens->setPosition(StoredPosition); } 2635 }; 2636 } // namespace 2637 2638 // Look to see if we have [[ by looking ahead, if 2639 // its not then rewind to the original position. 2640 bool UnwrappedLineParser::tryToParseSimpleAttribute() { 2641 ScopedTokenPosition AutoPosition(Tokens); 2642 FormatToken *Tok = Tokens->getNextToken(); 2643 // We already read the first [ check for the second. 2644 if (Tok && !Tok->is(tok::l_square)) { 2645 return false; 2646 } 2647 // Double check that the attribute is just something 2648 // fairly simple. 2649 while (Tok) { 2650 if (Tok->is(tok::r_square)) { 2651 break; 2652 } 2653 Tok = Tokens->getNextToken(); 2654 } 2655 Tok = Tokens->getNextToken(); 2656 if (Tok && !Tok->is(tok::r_square)) { 2657 return false; 2658 } 2659 Tok = Tokens->getNextToken(); 2660 if (Tok && Tok->is(tok::semi)) { 2661 return false; 2662 } 2663 return true; 2664 } 2665 2666 void UnwrappedLineParser::parseJavaEnumBody() { 2667 // Determine whether the enum is simple, i.e. does not have a semicolon or 2668 // constants with class bodies. Simple enums can be formatted like braced 2669 // lists, contracted to a single line, etc. 2670 unsigned StoredPosition = Tokens->getPosition(); 2671 bool IsSimple = true; 2672 FormatToken *Tok = Tokens->getNextToken(); 2673 while (Tok) { 2674 if (Tok->is(tok::r_brace)) 2675 break; 2676 if (Tok->isOneOf(tok::l_brace, tok::semi)) { 2677 IsSimple = false; 2678 break; 2679 } 2680 // FIXME: This will also mark enums with braces in the arguments to enum 2681 // constants as "not simple". This is probably fine in practice, though. 2682 Tok = Tokens->getNextToken(); 2683 } 2684 FormatTok = Tokens->setPosition(StoredPosition); 2685 2686 if (IsSimple) { 2687 nextToken(); 2688 parseBracedList(); 2689 addUnwrappedLine(); 2690 return; 2691 } 2692 2693 // Parse the body of a more complex enum. 2694 // First add a line for everything up to the "{". 2695 nextToken(); 2696 addUnwrappedLine(); 2697 ++Line->Level; 2698 2699 // Parse the enum constants. 2700 while (FormatTok) { 2701 if (FormatTok->is(tok::l_brace)) { 2702 // Parse the constant's class body. 2703 parseBlock(/*MustBeDeclaration=*/true, /*AddLevels=*/1u, 2704 /*MunchSemi=*/false); 2705 } else if (FormatTok->is(tok::l_paren)) { 2706 parseParens(); 2707 } else if (FormatTok->is(tok::comma)) { 2708 nextToken(); 2709 addUnwrappedLine(); 2710 } else if (FormatTok->is(tok::semi)) { 2711 nextToken(); 2712 addUnwrappedLine(); 2713 break; 2714 } else if (FormatTok->is(tok::r_brace)) { 2715 addUnwrappedLine(); 2716 break; 2717 } else { 2718 nextToken(); 2719 } 2720 } 2721 2722 // Parse the class body after the enum's ";" if any. 2723 parseLevel(/*HasOpeningBrace=*/true); 2724 nextToken(); 2725 --Line->Level; 2726 addUnwrappedLine(); 2727 } 2728 2729 void UnwrappedLineParser::parseRecord(bool ParseAsExpr) { 2730 const FormatToken &InitialToken = *FormatTok; 2731 nextToken(); 2732 2733 // The actual identifier can be a nested name specifier, and in macros 2734 // it is often token-pasted. 2735 // An [[attribute]] can be before the identifier. 2736 while (FormatTok->isOneOf(tok::identifier, tok::coloncolon, tok::hashhash, 2737 tok::kw___attribute, tok::kw___declspec, 2738 tok::kw_alignas, tok::l_square, tok::r_square) || 2739 ((Style.Language == FormatStyle::LK_Java || 2740 Style.Language == FormatStyle::LK_JavaScript) && 2741 FormatTok->isOneOf(tok::period, tok::comma))) { 2742 if (Style.Language == FormatStyle::LK_JavaScript && 2743 FormatTok->isOneOf(Keywords.kw_extends, Keywords.kw_implements)) { 2744 // JavaScript/TypeScript supports inline object types in 2745 // extends/implements positions: 2746 // class Foo implements {bar: number} { } 2747 nextToken(); 2748 if (FormatTok->is(tok::l_brace)) { 2749 tryToParseBracedList(); 2750 continue; 2751 } 2752 } 2753 bool IsNonMacroIdentifier = 2754 FormatTok->is(tok::identifier) && 2755 FormatTok->TokenText != FormatTok->TokenText.upper(); 2756 nextToken(); 2757 // We can have macros or attributes in between 'class' and the class name. 2758 if (!IsNonMacroIdentifier) { 2759 if (FormatTok->Tok.is(tok::l_paren)) { 2760 parseParens(); 2761 } else if (FormatTok->is(TT_AttributeSquare)) { 2762 parseSquare(); 2763 // Consume the closing TT_AttributeSquare. 2764 if (FormatTok->Next && FormatTok->is(TT_AttributeSquare)) 2765 nextToken(); 2766 } 2767 } 2768 } 2769 2770 // Note that parsing away template declarations here leads to incorrectly 2771 // accepting function declarations as record declarations. 2772 // In general, we cannot solve this problem. Consider: 2773 // class A<int> B() {} 2774 // which can be a function definition or a class definition when B() is a 2775 // macro. If we find enough real-world cases where this is a problem, we 2776 // can parse for the 'template' keyword in the beginning of the statement, 2777 // and thus rule out the record production in case there is no template 2778 // (this would still leave us with an ambiguity between template function 2779 // and class declarations). 2780 if (FormatTok->isOneOf(tok::colon, tok::less)) { 2781 while (!eof()) { 2782 if (FormatTok->is(tok::l_brace)) { 2783 calculateBraceTypes(/*ExpectClassBody=*/true); 2784 if (!tryToParseBracedList()) 2785 break; 2786 } 2787 if (FormatTok->Tok.is(tok::semi)) 2788 return; 2789 if (Style.isCSharp() && FormatTok->is(Keywords.kw_where)) { 2790 addUnwrappedLine(); 2791 nextToken(); 2792 parseCSharpGenericTypeConstraint(); 2793 break; 2794 } 2795 nextToken(); 2796 } 2797 } 2798 if (FormatTok->Tok.is(tok::l_brace)) { 2799 if (ParseAsExpr) { 2800 parseChildBlock(); 2801 } else { 2802 if (ShouldBreakBeforeBrace(Style, InitialToken)) 2803 addUnwrappedLine(); 2804 2805 unsigned AddLevels = Style.IndentAccessModifiers ? 2u : 1u; 2806 parseBlock(/*MustBeDeclaration=*/true, AddLevels, /*MunchSemi=*/false); 2807 } 2808 } 2809 // There is no addUnwrappedLine() here so that we fall through to parsing a 2810 // structural element afterwards. Thus, in "class A {} n, m;", 2811 // "} n, m;" will end up in one unwrapped line. 2812 } 2813 2814 void UnwrappedLineParser::parseObjCMethod() { 2815 assert(FormatTok->Tok.isOneOf(tok::l_paren, tok::identifier) && 2816 "'(' or identifier expected."); 2817 do { 2818 if (FormatTok->Tok.is(tok::semi)) { 2819 nextToken(); 2820 addUnwrappedLine(); 2821 return; 2822 } else if (FormatTok->Tok.is(tok::l_brace)) { 2823 if (Style.BraceWrapping.AfterFunction) 2824 addUnwrappedLine(); 2825 parseBlock(/*MustBeDeclaration=*/false); 2826 addUnwrappedLine(); 2827 return; 2828 } else { 2829 nextToken(); 2830 } 2831 } while (!eof()); 2832 } 2833 2834 void UnwrappedLineParser::parseObjCProtocolList() { 2835 assert(FormatTok->Tok.is(tok::less) && "'<' expected."); 2836 do { 2837 nextToken(); 2838 // Early exit in case someone forgot a close angle. 2839 if (FormatTok->isOneOf(tok::semi, tok::l_brace) || 2840 FormatTok->Tok.isObjCAtKeyword(tok::objc_end)) 2841 return; 2842 } while (!eof() && FormatTok->Tok.isNot(tok::greater)); 2843 nextToken(); // Skip '>'. 2844 } 2845 2846 void UnwrappedLineParser::parseObjCUntilAtEnd() { 2847 do { 2848 if (FormatTok->Tok.isObjCAtKeyword(tok::objc_end)) { 2849 nextToken(); 2850 addUnwrappedLine(); 2851 break; 2852 } 2853 if (FormatTok->is(tok::l_brace)) { 2854 parseBlock(/*MustBeDeclaration=*/false); 2855 // In ObjC interfaces, nothing should be following the "}". 2856 addUnwrappedLine(); 2857 } else if (FormatTok->is(tok::r_brace)) { 2858 // Ignore stray "}". parseStructuralElement doesn't consume them. 2859 nextToken(); 2860 addUnwrappedLine(); 2861 } else if (FormatTok->isOneOf(tok::minus, tok::plus)) { 2862 nextToken(); 2863 parseObjCMethod(); 2864 } else { 2865 parseStructuralElement(); 2866 } 2867 } while (!eof()); 2868 } 2869 2870 void UnwrappedLineParser::parseObjCInterfaceOrImplementation() { 2871 assert(FormatTok->Tok.getObjCKeywordID() == tok::objc_interface || 2872 FormatTok->Tok.getObjCKeywordID() == tok::objc_implementation); 2873 nextToken(); 2874 nextToken(); // interface name 2875 2876 // @interface can be followed by a lightweight generic 2877 // specialization list, then either a base class or a category. 2878 if (FormatTok->Tok.is(tok::less)) { 2879 parseObjCLightweightGenerics(); 2880 } 2881 if (FormatTok->Tok.is(tok::colon)) { 2882 nextToken(); 2883 nextToken(); // base class name 2884 // The base class can also have lightweight generics applied to it. 2885 if (FormatTok->Tok.is(tok::less)) { 2886 parseObjCLightweightGenerics(); 2887 } 2888 } else if (FormatTok->Tok.is(tok::l_paren)) 2889 // Skip category, if present. 2890 parseParens(); 2891 2892 if (FormatTok->Tok.is(tok::less)) 2893 parseObjCProtocolList(); 2894 2895 if (FormatTok->Tok.is(tok::l_brace)) { 2896 if (Style.BraceWrapping.AfterObjCDeclaration) 2897 addUnwrappedLine(); 2898 parseBlock(/*MustBeDeclaration=*/true); 2899 } 2900 2901 // With instance variables, this puts '}' on its own line. Without instance 2902 // variables, this ends the @interface line. 2903 addUnwrappedLine(); 2904 2905 parseObjCUntilAtEnd(); 2906 } 2907 2908 void UnwrappedLineParser::parseObjCLightweightGenerics() { 2909 assert(FormatTok->Tok.is(tok::less)); 2910 // Unlike protocol lists, generic parameterizations support 2911 // nested angles: 2912 // 2913 // @interface Foo<ValueType : id <NSCopying, NSSecureCoding>> : 2914 // NSObject <NSCopying, NSSecureCoding> 2915 // 2916 // so we need to count how many open angles we have left. 2917 unsigned NumOpenAngles = 1; 2918 do { 2919 nextToken(); 2920 // Early exit in case someone forgot a close angle. 2921 if (FormatTok->isOneOf(tok::semi, tok::l_brace) || 2922 FormatTok->Tok.isObjCAtKeyword(tok::objc_end)) 2923 break; 2924 if (FormatTok->Tok.is(tok::less)) 2925 ++NumOpenAngles; 2926 else if (FormatTok->Tok.is(tok::greater)) { 2927 assert(NumOpenAngles > 0 && "'>' makes NumOpenAngles negative"); 2928 --NumOpenAngles; 2929 } 2930 } while (!eof() && NumOpenAngles != 0); 2931 nextToken(); // Skip '>'. 2932 } 2933 2934 // Returns true for the declaration/definition form of @protocol, 2935 // false for the expression form. 2936 bool UnwrappedLineParser::parseObjCProtocol() { 2937 assert(FormatTok->Tok.getObjCKeywordID() == tok::objc_protocol); 2938 nextToken(); 2939 2940 if (FormatTok->is(tok::l_paren)) 2941 // The expression form of @protocol, e.g. "Protocol* p = @protocol(foo);". 2942 return false; 2943 2944 // The definition/declaration form, 2945 // @protocol Foo 2946 // - (int)someMethod; 2947 // @end 2948 2949 nextToken(); // protocol name 2950 2951 if (FormatTok->Tok.is(tok::less)) 2952 parseObjCProtocolList(); 2953 2954 // Check for protocol declaration. 2955 if (FormatTok->Tok.is(tok::semi)) { 2956 nextToken(); 2957 addUnwrappedLine(); 2958 return true; 2959 } 2960 2961 addUnwrappedLine(); 2962 parseObjCUntilAtEnd(); 2963 return true; 2964 } 2965 2966 void UnwrappedLineParser::parseJavaScriptEs6ImportExport() { 2967 bool IsImport = FormatTok->is(Keywords.kw_import); 2968 assert(IsImport || FormatTok->is(tok::kw_export)); 2969 nextToken(); 2970 2971 // Consume the "default" in "export default class/function". 2972 if (FormatTok->is(tok::kw_default)) 2973 nextToken(); 2974 2975 // Consume "async function", "function" and "default function", so that these 2976 // get parsed as free-standing JS functions, i.e. do not require a trailing 2977 // semicolon. 2978 if (FormatTok->is(Keywords.kw_async)) 2979 nextToken(); 2980 if (FormatTok->is(Keywords.kw_function)) { 2981 nextToken(); 2982 return; 2983 } 2984 2985 // For imports, `export *`, `export {...}`, consume the rest of the line up 2986 // to the terminating `;`. For everything else, just return and continue 2987 // parsing the structural element, i.e. the declaration or expression for 2988 // `export default`. 2989 if (!IsImport && !FormatTok->isOneOf(tok::l_brace, tok::star) && 2990 !FormatTok->isStringLiteral()) 2991 return; 2992 2993 while (!eof()) { 2994 if (FormatTok->is(tok::semi)) 2995 return; 2996 if (Line->Tokens.empty()) { 2997 // Common issue: Automatic Semicolon Insertion wrapped the line, so the 2998 // import statement should terminate. 2999 return; 3000 } 3001 if (FormatTok->is(tok::l_brace)) { 3002 FormatTok->setBlockKind(BK_Block); 3003 nextToken(); 3004 parseBracedList(); 3005 } else { 3006 nextToken(); 3007 } 3008 } 3009 } 3010 3011 void UnwrappedLineParser::parseStatementMacro() { 3012 nextToken(); 3013 if (FormatTok->is(tok::l_paren)) 3014 parseParens(); 3015 if (FormatTok->is(tok::semi)) 3016 nextToken(); 3017 addUnwrappedLine(); 3018 } 3019 3020 LLVM_ATTRIBUTE_UNUSED static void printDebugInfo(const UnwrappedLine &Line, 3021 StringRef Prefix = "") { 3022 llvm::dbgs() << Prefix << "Line(" << Line.Level 3023 << ", FSC=" << Line.FirstStartColumn << ")" 3024 << (Line.InPPDirective ? " MACRO" : "") << ": "; 3025 for (std::list<UnwrappedLineNode>::const_iterator I = Line.Tokens.begin(), 3026 E = Line.Tokens.end(); 3027 I != E; ++I) { 3028 llvm::dbgs() << I->Tok->Tok.getName() << "[" 3029 << "T=" << (unsigned)I->Tok->getType() 3030 << ", OC=" << I->Tok->OriginalColumn << "] "; 3031 } 3032 for (std::list<UnwrappedLineNode>::const_iterator I = Line.Tokens.begin(), 3033 E = Line.Tokens.end(); 3034 I != E; ++I) { 3035 const UnwrappedLineNode &Node = *I; 3036 for (SmallVectorImpl<UnwrappedLine>::const_iterator 3037 I = Node.Children.begin(), 3038 E = Node.Children.end(); 3039 I != E; ++I) { 3040 printDebugInfo(*I, "\nChild: "); 3041 } 3042 } 3043 llvm::dbgs() << "\n"; 3044 } 3045 3046 void UnwrappedLineParser::addUnwrappedLine(LineLevel AdjustLevel) { 3047 if (Line->Tokens.empty()) 3048 return; 3049 LLVM_DEBUG({ 3050 if (CurrentLines == &Lines) 3051 printDebugInfo(*Line); 3052 }); 3053 3054 // If this line closes a block when in Whitesmiths mode, remember that 3055 // information so that the level can be decreased after the line is added. 3056 // This has to happen after the addition of the line since the line itself 3057 // needs to be indented. 3058 bool ClosesWhitesmithsBlock = 3059 Line->MatchingOpeningBlockLineIndex != UnwrappedLine::kInvalidIndex && 3060 Style.BreakBeforeBraces == FormatStyle::BS_Whitesmiths; 3061 3062 CurrentLines->push_back(std::move(*Line)); 3063 Line->Tokens.clear(); 3064 Line->MatchingOpeningBlockLineIndex = UnwrappedLine::kInvalidIndex; 3065 Line->FirstStartColumn = 0; 3066 3067 if (ClosesWhitesmithsBlock && AdjustLevel == LineLevel::Remove) 3068 --Line->Level; 3069 if (CurrentLines == &Lines && !PreprocessorDirectives.empty()) { 3070 CurrentLines->append( 3071 std::make_move_iterator(PreprocessorDirectives.begin()), 3072 std::make_move_iterator(PreprocessorDirectives.end())); 3073 PreprocessorDirectives.clear(); 3074 } 3075 // Disconnect the current token from the last token on the previous line. 3076 FormatTok->Previous = nullptr; 3077 } 3078 3079 bool UnwrappedLineParser::eof() const { return FormatTok->Tok.is(tok::eof); } 3080 3081 bool UnwrappedLineParser::isOnNewLine(const FormatToken &FormatTok) { 3082 return (Line->InPPDirective || FormatTok.HasUnescapedNewline) && 3083 FormatTok.NewlinesBefore > 0; 3084 } 3085 3086 // Checks if \p FormatTok is a line comment that continues the line comment 3087 // section on \p Line. 3088 static bool 3089 continuesLineCommentSection(const FormatToken &FormatTok, 3090 const UnwrappedLine &Line, 3091 const llvm::Regex &CommentPragmasRegex) { 3092 if (Line.Tokens.empty()) 3093 return false; 3094 3095 StringRef IndentContent = FormatTok.TokenText; 3096 if (FormatTok.TokenText.startswith("//") || 3097 FormatTok.TokenText.startswith("/*")) 3098 IndentContent = FormatTok.TokenText.substr(2); 3099 if (CommentPragmasRegex.match(IndentContent)) 3100 return false; 3101 3102 // If Line starts with a line comment, then FormatTok continues the comment 3103 // section if its original column is greater or equal to the original start 3104 // column of the line. 3105 // 3106 // Define the min column token of a line as follows: if a line ends in '{' or 3107 // contains a '{' followed by a line comment, then the min column token is 3108 // that '{'. Otherwise, the min column token of the line is the first token of 3109 // the line. 3110 // 3111 // If Line starts with a token other than a line comment, then FormatTok 3112 // continues the comment section if its original column is greater than the 3113 // original start column of the min column token of the line. 3114 // 3115 // For example, the second line comment continues the first in these cases: 3116 // 3117 // // first line 3118 // // second line 3119 // 3120 // and: 3121 // 3122 // // first line 3123 // // second line 3124 // 3125 // and: 3126 // 3127 // int i; // first line 3128 // // second line 3129 // 3130 // and: 3131 // 3132 // do { // first line 3133 // // second line 3134 // int i; 3135 // } while (true); 3136 // 3137 // and: 3138 // 3139 // enum { 3140 // a, // first line 3141 // // second line 3142 // b 3143 // }; 3144 // 3145 // The second line comment doesn't continue the first in these cases: 3146 // 3147 // // first line 3148 // // second line 3149 // 3150 // and: 3151 // 3152 // int i; // first line 3153 // // second line 3154 // 3155 // and: 3156 // 3157 // do { // first line 3158 // // second line 3159 // int i; 3160 // } while (true); 3161 // 3162 // and: 3163 // 3164 // enum { 3165 // a, // first line 3166 // // second line 3167 // }; 3168 const FormatToken *MinColumnToken = Line.Tokens.front().Tok; 3169 3170 // Scan for '{//'. If found, use the column of '{' as a min column for line 3171 // comment section continuation. 3172 const FormatToken *PreviousToken = nullptr; 3173 for (const UnwrappedLineNode &Node : Line.Tokens) { 3174 if (PreviousToken && PreviousToken->is(tok::l_brace) && 3175 isLineComment(*Node.Tok)) { 3176 MinColumnToken = PreviousToken; 3177 break; 3178 } 3179 PreviousToken = Node.Tok; 3180 3181 // Grab the last newline preceding a token in this unwrapped line. 3182 if (Node.Tok->NewlinesBefore > 0) { 3183 MinColumnToken = Node.Tok; 3184 } 3185 } 3186 if (PreviousToken && PreviousToken->is(tok::l_brace)) { 3187 MinColumnToken = PreviousToken; 3188 } 3189 3190 return continuesLineComment(FormatTok, /*Previous=*/Line.Tokens.back().Tok, 3191 MinColumnToken); 3192 } 3193 3194 void UnwrappedLineParser::flushComments(bool NewlineBeforeNext) { 3195 bool JustComments = Line->Tokens.empty(); 3196 for (SmallVectorImpl<FormatToken *>::const_iterator 3197 I = CommentsBeforeNextToken.begin(), 3198 E = CommentsBeforeNextToken.end(); 3199 I != E; ++I) { 3200 // Line comments that belong to the same line comment section are put on the 3201 // same line since later we might want to reflow content between them. 3202 // Additional fine-grained breaking of line comment sections is controlled 3203 // by the class BreakableLineCommentSection in case it is desirable to keep 3204 // several line comment sections in the same unwrapped line. 3205 // 3206 // FIXME: Consider putting separate line comment sections as children to the 3207 // unwrapped line instead. 3208 (*I)->ContinuesLineCommentSection = 3209 continuesLineCommentSection(**I, *Line, CommentPragmasRegex); 3210 if (isOnNewLine(**I) && JustComments && !(*I)->ContinuesLineCommentSection) 3211 addUnwrappedLine(); 3212 pushToken(*I); 3213 } 3214 if (NewlineBeforeNext && JustComments) 3215 addUnwrappedLine(); 3216 CommentsBeforeNextToken.clear(); 3217 } 3218 3219 void UnwrappedLineParser::nextToken(int LevelDifference) { 3220 if (eof()) 3221 return; 3222 flushComments(isOnNewLine(*FormatTok)); 3223 pushToken(FormatTok); 3224 FormatToken *Previous = FormatTok; 3225 if (Style.Language != FormatStyle::LK_JavaScript) 3226 readToken(LevelDifference); 3227 else 3228 readTokenWithJavaScriptASI(); 3229 FormatTok->Previous = Previous; 3230 } 3231 3232 void UnwrappedLineParser::distributeComments( 3233 const SmallVectorImpl<FormatToken *> &Comments, 3234 const FormatToken *NextTok) { 3235 // Whether or not a line comment token continues a line is controlled by 3236 // the method continuesLineCommentSection, with the following caveat: 3237 // 3238 // Define a trail of Comments to be a nonempty proper postfix of Comments such 3239 // that each comment line from the trail is aligned with the next token, if 3240 // the next token exists. If a trail exists, the beginning of the maximal 3241 // trail is marked as a start of a new comment section. 3242 // 3243 // For example in this code: 3244 // 3245 // int a; // line about a 3246 // // line 1 about b 3247 // // line 2 about b 3248 // int b; 3249 // 3250 // the two lines about b form a maximal trail, so there are two sections, the 3251 // first one consisting of the single comment "// line about a" and the 3252 // second one consisting of the next two comments. 3253 if (Comments.empty()) 3254 return; 3255 bool ShouldPushCommentsInCurrentLine = true; 3256 bool HasTrailAlignedWithNextToken = false; 3257 unsigned StartOfTrailAlignedWithNextToken = 0; 3258 if (NextTok) { 3259 // We are skipping the first element intentionally. 3260 for (unsigned i = Comments.size() - 1; i > 0; --i) { 3261 if (Comments[i]->OriginalColumn == NextTok->OriginalColumn) { 3262 HasTrailAlignedWithNextToken = true; 3263 StartOfTrailAlignedWithNextToken = i; 3264 } 3265 } 3266 } 3267 for (unsigned i = 0, e = Comments.size(); i < e; ++i) { 3268 FormatToken *FormatTok = Comments[i]; 3269 if (HasTrailAlignedWithNextToken && i == StartOfTrailAlignedWithNextToken) { 3270 FormatTok->ContinuesLineCommentSection = false; 3271 } else { 3272 FormatTok->ContinuesLineCommentSection = 3273 continuesLineCommentSection(*FormatTok, *Line, CommentPragmasRegex); 3274 } 3275 if (!FormatTok->ContinuesLineCommentSection && 3276 (isOnNewLine(*FormatTok) || FormatTok->IsFirst)) { 3277 ShouldPushCommentsInCurrentLine = false; 3278 } 3279 if (ShouldPushCommentsInCurrentLine) { 3280 pushToken(FormatTok); 3281 } else { 3282 CommentsBeforeNextToken.push_back(FormatTok); 3283 } 3284 } 3285 } 3286 3287 void UnwrappedLineParser::readToken(int LevelDifference) { 3288 SmallVector<FormatToken *, 1> Comments; 3289 do { 3290 FormatTok = Tokens->getNextToken(); 3291 assert(FormatTok); 3292 while (!Line->InPPDirective && FormatTok->Tok.is(tok::hash) && 3293 (FormatTok->HasUnescapedNewline || FormatTok->IsFirst)) { 3294 distributeComments(Comments, FormatTok); 3295 Comments.clear(); 3296 // If there is an unfinished unwrapped line, we flush the preprocessor 3297 // directives only after that unwrapped line was finished later. 3298 bool SwitchToPreprocessorLines = !Line->Tokens.empty(); 3299 ScopedLineState BlockState(*this, SwitchToPreprocessorLines); 3300 assert((LevelDifference >= 0 || 3301 static_cast<unsigned>(-LevelDifference) <= Line->Level) && 3302 "LevelDifference makes Line->Level negative"); 3303 Line->Level += LevelDifference; 3304 // Comments stored before the preprocessor directive need to be output 3305 // before the preprocessor directive, at the same level as the 3306 // preprocessor directive, as we consider them to apply to the directive. 3307 if (Style.IndentPPDirectives == FormatStyle::PPDIS_BeforeHash && 3308 PPBranchLevel > 0) 3309 Line->Level += PPBranchLevel; 3310 flushComments(isOnNewLine(*FormatTok)); 3311 parsePPDirective(); 3312 } 3313 while (FormatTok->getType() == TT_ConflictStart || 3314 FormatTok->getType() == TT_ConflictEnd || 3315 FormatTok->getType() == TT_ConflictAlternative) { 3316 if (FormatTok->getType() == TT_ConflictStart) { 3317 conditionalCompilationStart(/*Unreachable=*/false); 3318 } else if (FormatTok->getType() == TT_ConflictAlternative) { 3319 conditionalCompilationAlternative(); 3320 } else if (FormatTok->getType() == TT_ConflictEnd) { 3321 conditionalCompilationEnd(); 3322 } 3323 FormatTok = Tokens->getNextToken(); 3324 FormatTok->MustBreakBefore = true; 3325 } 3326 3327 if (!PPStack.empty() && (PPStack.back().Kind == PP_Unreachable) && 3328 !Line->InPPDirective) { 3329 continue; 3330 } 3331 3332 if (!FormatTok->Tok.is(tok::comment)) { 3333 distributeComments(Comments, FormatTok); 3334 Comments.clear(); 3335 return; 3336 } 3337 3338 Comments.push_back(FormatTok); 3339 } while (!eof()); 3340 3341 distributeComments(Comments, nullptr); 3342 Comments.clear(); 3343 } 3344 3345 void UnwrappedLineParser::pushToken(FormatToken *Tok) { 3346 Line->Tokens.push_back(UnwrappedLineNode(Tok)); 3347 if (MustBreakBeforeNextToken) { 3348 Line->Tokens.back().Tok->MustBreakBefore = true; 3349 MustBreakBeforeNextToken = false; 3350 } 3351 } 3352 3353 } // end namespace format 3354 } // end namespace clang 3355