1 //===--- TokenAnnotator.cpp - Format C++ code -----------------------------===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is distributed under the University of Illinois Open Source 6 // License. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 /// 10 /// \file 11 /// \brief This file implements a token annotator, i.e. creates 12 /// \c AnnotatedTokens out of \c FormatTokens with required extra information. 13 /// 14 //===----------------------------------------------------------------------===// 15 16 #include "TokenAnnotator.h" 17 #include "clang/Basic/SourceManager.h" 18 #include "llvm/Support/Debug.h" 19 20 #define DEBUG_TYPE "format-token-annotator" 21 22 namespace clang { 23 namespace format { 24 25 namespace { 26 27 /// \brief A parser that gathers additional information about tokens. 28 /// 29 /// The \c TokenAnnotator tries to match parenthesis and square brakets and 30 /// store a parenthesis levels. It also tries to resolve matching "<" and ">" 31 /// into template parameter lists. 32 class AnnotatingParser { 33 public: 34 AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line, 35 IdentifierInfo &Ident_in) 36 : Style(Style), Line(Line), CurrentToken(Line.First), 37 KeywordVirtualFound(false), AutoFound(false), Ident_in(Ident_in) { 38 Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false)); 39 resetTokenMetadata(CurrentToken); 40 } 41 42 private: 43 bool parseAngle() { 44 if (!CurrentToken) 45 return false; 46 ScopedContextCreator ContextCreator(*this, tok::less, 10); 47 FormatToken *Left = CurrentToken->Previous; 48 Contexts.back().IsExpression = false; 49 // If there's a template keyword before the opening angle bracket, this is a 50 // template parameter, not an argument. 51 Contexts.back().InTemplateArgument = 52 Left->Previous && Left->Previous->Tok.isNot(tok::kw_template); 53 54 while (CurrentToken) { 55 if (CurrentToken->is(tok::greater)) { 56 Left->MatchingParen = CurrentToken; 57 CurrentToken->MatchingParen = Left; 58 CurrentToken->Type = TT_TemplateCloser; 59 next(); 60 return true; 61 } 62 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace, 63 tok::question, tok::colon)) 64 return false; 65 // If a && or || is found and interpreted as a binary operator, this set 66 // of angles is likely part of something like "a < b && c > d". If the 67 // angles are inside an expression, the ||/&& might also be a binary 68 // operator that was misinterpreted because we are parsing template 69 // parameters. 70 // FIXME: This is getting out of hand, write a decent parser. 71 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) && 72 CurrentToken->Previous->Type == TT_BinaryOperator && 73 Contexts[Contexts.size() - 2].IsExpression && 74 Line.First->isNot(tok::kw_template)) 75 return false; 76 updateParameterCount(Left, CurrentToken); 77 if (!consumeToken()) 78 return false; 79 } 80 return false; 81 } 82 83 bool parseParens(bool LookForDecls = false) { 84 if (!CurrentToken) 85 return false; 86 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1); 87 88 // FIXME: This is a bit of a hack. Do better. 89 Contexts.back().ColonIsForRangeExpr = 90 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr; 91 92 bool StartsObjCMethodExpr = false; 93 FormatToken *Left = CurrentToken->Previous; 94 if (CurrentToken->is(tok::caret)) { 95 // (^ can start a block type. 96 Left->Type = TT_ObjCBlockLParen; 97 } else if (FormatToken *MaybeSel = Left->Previous) { 98 // @selector( starts a selector. 99 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous && 100 MaybeSel->Previous->is(tok::at)) { 101 StartsObjCMethodExpr = true; 102 } 103 } 104 105 if (Left->Previous && 106 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_if, 107 tok::kw_while, tok::l_paren, tok::comma) || 108 Left->Previous->Type == TT_BinaryOperator)) { 109 // static_assert, if and while usually contain expressions. 110 Contexts.back().IsExpression = true; 111 } else if (Line.InPPDirective && 112 (!Left->Previous || 113 (Left->Previous->isNot(tok::identifier) && 114 Left->Previous->Type != TT_OverloadedOperator))) { 115 Contexts.back().IsExpression = true; 116 } else if (Left->Previous && Left->Previous->is(tok::r_square) && 117 Left->Previous->MatchingParen && 118 Left->Previous->MatchingParen->Type == TT_LambdaLSquare) { 119 // This is a parameter list of a lambda expression. 120 Contexts.back().IsExpression = false; 121 } else if (Contexts[Contexts.size() - 2].CaretFound) { 122 // This is the parameter list of an ObjC block. 123 Contexts.back().IsExpression = false; 124 } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) { 125 Left->Type = TT_AttributeParen; 126 } else if (Left->Previous && Left->Previous->IsForEachMacro) { 127 // The first argument to a foreach macro is a declaration. 128 Contexts.back().IsForEachMacro = true; 129 Contexts.back().IsExpression = false; 130 } 131 132 if (StartsObjCMethodExpr) { 133 Contexts.back().ColonIsObjCMethodExpr = true; 134 Left->Type = TT_ObjCMethodExpr; 135 } 136 137 bool MightBeFunctionType = CurrentToken->is(tok::star); 138 bool HasMultipleLines = false; 139 bool HasMultipleParametersOnALine = false; 140 while (CurrentToken) { 141 // LookForDecls is set when "if (" has been seen. Check for 142 // 'identifier' '*' 'identifier' followed by not '=' -- this 143 // '*' has to be a binary operator but determineStarAmpUsage() will 144 // categorize it as an unary operator, so set the right type here. 145 if (LookForDecls && CurrentToken->Next) { 146 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 147 if (Prev) { 148 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 149 FormatToken *Next = CurrentToken->Next; 150 if (PrevPrev && PrevPrev->is(tok::identifier) && 151 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 152 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 153 Prev->Type = TT_BinaryOperator; 154 LookForDecls = false; 155 } 156 } 157 } 158 159 if (CurrentToken->Previous->Type == TT_PointerOrReference && 160 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 161 tok::coloncolon)) 162 MightBeFunctionType = true; 163 if (CurrentToken->Previous->Type == TT_BinaryOperator) 164 Contexts.back().IsExpression = true; 165 if (CurrentToken->is(tok::r_paren)) { 166 if (MightBeFunctionType && CurrentToken->Next && 167 (CurrentToken->Next->is(tok::l_paren) || 168 (CurrentToken->Next->is(tok::l_square) && 169 !Contexts.back().IsExpression))) 170 Left->Type = TT_FunctionTypeLParen; 171 Left->MatchingParen = CurrentToken; 172 CurrentToken->MatchingParen = Left; 173 174 if (StartsObjCMethodExpr) { 175 CurrentToken->Type = TT_ObjCMethodExpr; 176 if (Contexts.back().FirstObjCSelectorName) { 177 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 178 Contexts.back().LongestObjCSelectorName; 179 } 180 } 181 182 if (Left->Type == TT_AttributeParen) 183 CurrentToken->Type = TT_AttributeParen; 184 185 if (!HasMultipleLines) 186 Left->PackingKind = PPK_Inconclusive; 187 else if (HasMultipleParametersOnALine) 188 Left->PackingKind = PPK_BinPacked; 189 else 190 Left->PackingKind = PPK_OnePerLine; 191 192 next(); 193 return true; 194 } 195 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 196 return false; 197 else if (CurrentToken->is(tok::l_brace)) 198 Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen 199 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 200 !CurrentToken->Next->HasUnescapedNewline && 201 !CurrentToken->Next->isTrailingComment()) 202 HasMultipleParametersOnALine = true; 203 if (CurrentToken->isOneOf(tok::kw_const, tok::kw_auto) || 204 CurrentToken->isSimpleTypeSpecifier()) 205 Contexts.back().IsExpression = false; 206 FormatToken *Tok = CurrentToken; 207 if (!consumeToken()) 208 return false; 209 updateParameterCount(Left, Tok); 210 if (CurrentToken && CurrentToken->HasUnescapedNewline) 211 HasMultipleLines = true; 212 } 213 return false; 214 } 215 216 bool parseSquare() { 217 if (!CurrentToken) 218 return false; 219 220 // A '[' could be an index subscript (after an identifier or after 221 // ')' or ']'), it could be the start of an Objective-C method 222 // expression, or it could the the start of an Objective-C array literal. 223 FormatToken *Left = CurrentToken->Previous; 224 FormatToken *Parent = Left->getPreviousNonComment(); 225 bool StartsObjCMethodExpr = 226 Contexts.back().CanBeExpression && Left->Type != TT_LambdaLSquare && 227 CurrentToken->isNot(tok::l_brace) && 228 (!Parent || Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 229 tok::kw_return, tok::kw_throw) || 230 Parent->isUnaryOperator() || Parent->Type == TT_ObjCForIn || 231 Parent->Type == TT_CastRParen || 232 getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown); 233 ScopedContextCreator ContextCreator(*this, tok::l_square, 10); 234 Contexts.back().IsExpression = true; 235 bool ColonFound = false; 236 237 if (StartsObjCMethodExpr) { 238 Contexts.back().ColonIsObjCMethodExpr = true; 239 Left->Type = TT_ObjCMethodExpr; 240 } else if (Parent && Parent->is(tok::at)) { 241 Left->Type = TT_ArrayInitializerLSquare; 242 } else if (Left->Type == TT_Unknown) { 243 Left->Type = TT_ArraySubscriptLSquare; 244 } 245 246 while (CurrentToken) { 247 if (CurrentToken->is(tok::r_square)) { 248 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_paren) && 249 Left->Type == TT_ObjCMethodExpr) { 250 // An ObjC method call is rarely followed by an open parenthesis. 251 // FIXME: Do we incorrectly label ":" with this? 252 StartsObjCMethodExpr = false; 253 Left->Type = TT_Unknown; 254 } 255 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) { 256 CurrentToken->Type = TT_ObjCMethodExpr; 257 // determineStarAmpUsage() thinks that '*' '[' is allocating an 258 // array of pointers, but if '[' starts a selector then '*' is a 259 // binary operator. 260 if (Parent && Parent->Type == TT_PointerOrReference) 261 Parent->Type = TT_BinaryOperator; 262 } 263 Left->MatchingParen = CurrentToken; 264 CurrentToken->MatchingParen = Left; 265 if (Contexts.back().FirstObjCSelectorName) { 266 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 267 Contexts.back().LongestObjCSelectorName; 268 if (Left->BlockParameterCount > 1) 269 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0; 270 } 271 next(); 272 return true; 273 } 274 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace)) 275 return false; 276 if (CurrentToken->is(tok::colon)) 277 ColonFound = true; 278 if (CurrentToken->is(tok::comma) && 279 Style.Language != FormatStyle::LK_Proto && 280 (Left->Type == TT_ArraySubscriptLSquare || 281 (Left->Type == TT_ObjCMethodExpr && !ColonFound))) 282 Left->Type = TT_ArrayInitializerLSquare; 283 FormatToken* Tok = CurrentToken; 284 if (!consumeToken()) 285 return false; 286 updateParameterCount(Left, Tok); 287 } 288 return false; 289 } 290 291 bool parseBrace() { 292 if (CurrentToken) { 293 FormatToken *Left = CurrentToken->Previous; 294 295 if (Contexts.back().CaretFound) 296 Left->Type = TT_ObjCBlockLBrace; 297 Contexts.back().CaretFound = false; 298 299 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1); 300 Contexts.back().ColonIsDictLiteral = true; 301 if (Left->BlockKind == BK_BracedInit) 302 Contexts.back().IsExpression = true; 303 304 while (CurrentToken) { 305 if (CurrentToken->is(tok::r_brace)) { 306 Left->MatchingParen = CurrentToken; 307 CurrentToken->MatchingParen = Left; 308 next(); 309 return true; 310 } 311 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square)) 312 return false; 313 updateParameterCount(Left, CurrentToken); 314 if (CurrentToken->is(tok::colon)) { 315 if (CurrentToken->getPreviousNonComment()->is(tok::identifier)) 316 CurrentToken->getPreviousNonComment()->Type = TT_SelectorName; 317 Left->Type = TT_DictLiteral; 318 } 319 if (!consumeToken()) 320 return false; 321 } 322 } 323 return true; 324 } 325 326 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 327 if (Current->Type == TT_LambdaLSquare || 328 (Current->is(tok::caret) && Current->Type == TT_UnaryOperator) || 329 (Style.Language == FormatStyle::LK_JavaScript && 330 Current->TokenText == "function")) { 331 ++Left->BlockParameterCount; 332 } 333 if (Current->is(tok::comma)) { 334 ++Left->ParameterCount; 335 if (!Left->Role) 336 Left->Role.reset(new CommaSeparatedList(Style)); 337 Left->Role->CommaFound(Current); 338 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 339 Left->ParameterCount = 1; 340 } 341 } 342 343 bool parseConditional() { 344 while (CurrentToken) { 345 if (CurrentToken->is(tok::colon)) { 346 CurrentToken->Type = TT_ConditionalExpr; 347 next(); 348 return true; 349 } 350 if (!consumeToken()) 351 return false; 352 } 353 return false; 354 } 355 356 bool parseTemplateDeclaration() { 357 if (CurrentToken && CurrentToken->is(tok::less)) { 358 CurrentToken->Type = TT_TemplateOpener; 359 next(); 360 if (!parseAngle()) 361 return false; 362 if (CurrentToken) 363 CurrentToken->Previous->ClosesTemplateDeclaration = true; 364 return true; 365 } 366 return false; 367 } 368 369 bool consumeToken() { 370 FormatToken *Tok = CurrentToken; 371 next(); 372 switch (Tok->Tok.getKind()) { 373 case tok::plus: 374 case tok::minus: 375 if (!Tok->Previous && Line.MustBeDeclaration) 376 Tok->Type = TT_ObjCMethodSpecifier; 377 break; 378 case tok::colon: 379 if (!Tok->Previous) 380 return false; 381 // Colons from ?: are handled in parseConditional(). 382 if (Tok->Previous->is(tok::r_paren) && Contexts.size() == 1 && 383 Line.First->isNot(tok::kw_case)) { 384 Tok->Type = TT_CtorInitializerColon; 385 } else if (Contexts.back().ColonIsDictLiteral) { 386 Tok->Type = TT_DictLiteral; 387 } else if (Contexts.back().ColonIsObjCMethodExpr || 388 Line.First->Type == TT_ObjCMethodSpecifier) { 389 Tok->Type = TT_ObjCMethodExpr; 390 Tok->Previous->Type = TT_SelectorName; 391 if (Tok->Previous->ColumnWidth > 392 Contexts.back().LongestObjCSelectorName) { 393 Contexts.back().LongestObjCSelectorName = Tok->Previous->ColumnWidth; 394 } 395 if (!Contexts.back().FirstObjCSelectorName) 396 Contexts.back().FirstObjCSelectorName = Tok->Previous; 397 } else if (Contexts.back().ColonIsForRangeExpr) { 398 Tok->Type = TT_RangeBasedForLoopColon; 399 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 400 Tok->Type = TT_BitFieldColon; 401 } else if (Contexts.size() == 1 && 402 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) { 403 Tok->Type = TT_InheritanceColon; 404 } else if (Contexts.back().ContextKind == tok::l_paren) { 405 Tok->Type = TT_InlineASMColon; 406 } 407 break; 408 case tok::kw_if: 409 case tok::kw_while: 410 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 411 next(); 412 if (!parseParens(/*LookForDecls=*/true)) 413 return false; 414 } 415 break; 416 case tok::kw_for: 417 Contexts.back().ColonIsForRangeExpr = true; 418 next(); 419 if (!parseParens()) 420 return false; 421 break; 422 case tok::l_paren: 423 if (!parseParens()) 424 return false; 425 if (Line.MustBeDeclaration && Contexts.size() == 1 && 426 !Contexts.back().IsExpression && 427 Line.First->Type != TT_ObjCProperty && 428 (!Tok->Previous || Tok->Previous->isNot(tok::kw_decltype))) 429 Line.MightBeFunctionDecl = true; 430 break; 431 case tok::l_square: 432 if (!parseSquare()) 433 return false; 434 break; 435 case tok::l_brace: 436 if (!parseBrace()) 437 return false; 438 break; 439 case tok::less: 440 if (Tok->Previous && !Tok->Previous->Tok.isLiteral() && parseAngle()) 441 Tok->Type = TT_TemplateOpener; 442 else { 443 Tok->Type = TT_BinaryOperator; 444 CurrentToken = Tok; 445 next(); 446 } 447 break; 448 case tok::r_paren: 449 case tok::r_square: 450 return false; 451 case tok::r_brace: 452 // Lines can start with '}'. 453 if (Tok->Previous) 454 return false; 455 break; 456 case tok::greater: 457 Tok->Type = TT_BinaryOperator; 458 break; 459 case tok::kw_operator: 460 while (CurrentToken && 461 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 462 if (CurrentToken->isOneOf(tok::star, tok::amp)) 463 CurrentToken->Type = TT_PointerOrReference; 464 consumeToken(); 465 if (CurrentToken && CurrentToken->Previous->Type == TT_BinaryOperator) 466 CurrentToken->Previous->Type = TT_OverloadedOperator; 467 } 468 if (CurrentToken) { 469 CurrentToken->Type = TT_OverloadedOperatorLParen; 470 if (CurrentToken->Previous->Type == TT_BinaryOperator) 471 CurrentToken->Previous->Type = TT_OverloadedOperator; 472 } 473 break; 474 case tok::question: 475 parseConditional(); 476 break; 477 case tok::kw_template: 478 parseTemplateDeclaration(); 479 break; 480 case tok::identifier: 481 if (Line.First->is(tok::kw_for) && 482 Tok->Tok.getIdentifierInfo() == &Ident_in) 483 Tok->Type = TT_ObjCForIn; 484 break; 485 case tok::comma: 486 if (Contexts.back().FirstStartOfName) 487 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 488 if (Contexts.back().InCtorInitializer) 489 Tok->Type = TT_CtorInitializerComma; 490 if (Contexts.back().IsForEachMacro) 491 Contexts.back().IsExpression = true; 492 break; 493 default: 494 break; 495 } 496 return true; 497 } 498 499 void parseIncludeDirective() { 500 next(); 501 if (CurrentToken && CurrentToken->is(tok::less)) { 502 next(); 503 while (CurrentToken) { 504 if (CurrentToken->isNot(tok::comment) || CurrentToken->Next) 505 CurrentToken->Type = TT_ImplicitStringLiteral; 506 next(); 507 } 508 } else { 509 while (CurrentToken) { 510 if (CurrentToken->is(tok::string_literal)) 511 // Mark these string literals as "implicit" literals, too, so that 512 // they are not split or line-wrapped. 513 CurrentToken->Type = TT_ImplicitStringLiteral; 514 next(); 515 } 516 } 517 } 518 519 void parseWarningOrError() { 520 next(); 521 // We still want to format the whitespace left of the first token of the 522 // warning or error. 523 next(); 524 while (CurrentToken) { 525 CurrentToken->Type = TT_ImplicitStringLiteral; 526 next(); 527 } 528 } 529 530 void parsePragma() { 531 next(); // Consume "pragma". 532 if (CurrentToken && CurrentToken->TokenText == "mark") { 533 next(); // Consume "mark". 534 next(); // Consume first token (so we fix leading whitespace). 535 while (CurrentToken) { 536 CurrentToken->Type = TT_ImplicitStringLiteral; 537 next(); 538 } 539 } 540 } 541 542 void parsePreprocessorDirective() { 543 next(); 544 if (!CurrentToken) 545 return; 546 if (CurrentToken->Tok.is(tok::numeric_constant)) { 547 CurrentToken->SpacesRequiredBefore = 1; 548 return; 549 } 550 // Hashes in the middle of a line can lead to any strange token 551 // sequence. 552 if (!CurrentToken->Tok.getIdentifierInfo()) 553 return; 554 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 555 case tok::pp_include: 556 case tok::pp_import: 557 parseIncludeDirective(); 558 break; 559 case tok::pp_error: 560 case tok::pp_warning: 561 parseWarningOrError(); 562 break; 563 case tok::pp_pragma: 564 parsePragma(); 565 break; 566 case tok::pp_if: 567 case tok::pp_elif: 568 Contexts.back().IsExpression = true; 569 parseLine(); 570 break; 571 default: 572 break; 573 } 574 while (CurrentToken) 575 next(); 576 } 577 578 public: 579 LineType parseLine() { 580 if (CurrentToken->is(tok::hash)) { 581 parsePreprocessorDirective(); 582 return LT_PreprocessorDirective; 583 } 584 585 // Directly allow to 'import <string-literal>' to support protocol buffer 586 // definitions (code.google.com/p/protobuf) or missing "#" (either way we 587 // should not break the line). 588 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 589 if (Info && Info->getPPKeywordID() == tok::pp_import && 590 CurrentToken->Next && CurrentToken->Next->is(tok::string_literal)) 591 parseIncludeDirective(); 592 593 while (CurrentToken) { 594 if (CurrentToken->is(tok::kw_virtual)) 595 KeywordVirtualFound = true; 596 if (!consumeToken()) 597 return LT_Invalid; 598 } 599 if (KeywordVirtualFound) 600 return LT_VirtualFunctionDecl; 601 602 if (Line.First->Type == TT_ObjCMethodSpecifier) { 603 if (Contexts.back().FirstObjCSelectorName) 604 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 605 Contexts.back().LongestObjCSelectorName; 606 return LT_ObjCMethodDecl; 607 } 608 609 return LT_Other; 610 } 611 612 private: 613 void resetTokenMetadata(FormatToken *Token) { 614 if (!Token) 615 return; 616 617 // Reset token type in case we have already looked at it and then 618 // recovered from an error (e.g. failure to find the matching >). 619 if (CurrentToken->Type != TT_LambdaLSquare && 620 CurrentToken->Type != TT_FunctionLBrace && 621 CurrentToken->Type != TT_ImplicitStringLiteral && 622 CurrentToken->Type != TT_RegexLiteral && 623 CurrentToken->Type != TT_TrailingReturnArrow) 624 CurrentToken->Type = TT_Unknown; 625 CurrentToken->Role.reset(); 626 CurrentToken->FakeLParens.clear(); 627 CurrentToken->FakeRParens = 0; 628 } 629 630 void next() { 631 if (CurrentToken) { 632 determineTokenType(*CurrentToken); 633 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 634 CurrentToken->NestingLevel = Contexts.size() - 1; 635 CurrentToken = CurrentToken->Next; 636 } 637 638 resetTokenMetadata(CurrentToken); 639 } 640 641 /// \brief A struct to hold information valid in a specific context, e.g. 642 /// a pair of parenthesis. 643 struct Context { 644 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 645 bool IsExpression) 646 : ContextKind(ContextKind), BindingStrength(BindingStrength), 647 LongestObjCSelectorName(0), ColonIsForRangeExpr(false), 648 ColonIsDictLiteral(false), ColonIsObjCMethodExpr(false), 649 FirstObjCSelectorName(nullptr), FirstStartOfName(nullptr), 650 IsExpression(IsExpression), CanBeExpression(true), 651 InTemplateArgument(false), InCtorInitializer(false), 652 CaretFound(false), IsForEachMacro(false) {} 653 654 tok::TokenKind ContextKind; 655 unsigned BindingStrength; 656 unsigned LongestObjCSelectorName; 657 bool ColonIsForRangeExpr; 658 bool ColonIsDictLiteral; 659 bool ColonIsObjCMethodExpr; 660 FormatToken *FirstObjCSelectorName; 661 FormatToken *FirstStartOfName; 662 bool IsExpression; 663 bool CanBeExpression; 664 bool InTemplateArgument; 665 bool InCtorInitializer; 666 bool CaretFound; 667 bool IsForEachMacro; 668 }; 669 670 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime 671 /// of each instance. 672 struct ScopedContextCreator { 673 AnnotatingParser &P; 674 675 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 676 unsigned Increase) 677 : P(P) { 678 P.Contexts.push_back(Context(ContextKind, 679 P.Contexts.back().BindingStrength + Increase, 680 P.Contexts.back().IsExpression)); 681 } 682 683 ~ScopedContextCreator() { P.Contexts.pop_back(); } 684 }; 685 686 void determineTokenType(FormatToken &Current) { 687 if (Current.getPrecedence() == prec::Assignment && 688 !Line.First->isOneOf(tok::kw_template, tok::kw_using) && 689 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 690 Contexts.back().IsExpression = true; 691 for (FormatToken *Previous = Current.Previous; 692 Previous && !Previous->isOneOf(tok::comma, tok::semi); 693 Previous = Previous->Previous) { 694 if (Previous->isOneOf(tok::r_square, tok::r_paren)) 695 Previous = Previous->MatchingParen; 696 if ((Previous->Type == TT_BinaryOperator || 697 Previous->Type == TT_UnaryOperator) && 698 Previous->isOneOf(tok::star, tok::amp)) { 699 Previous->Type = TT_PointerOrReference; 700 } 701 } 702 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 703 Contexts.back().IsExpression = true; 704 } else if (Current.is(tok::l_paren) && !Line.MustBeDeclaration && 705 !Line.InPPDirective && 706 (!Current.Previous || 707 Current.Previous->isNot(tok::kw_decltype))) { 708 bool ParametersOfFunctionType = 709 Current.Previous && Current.Previous->is(tok::r_paren) && 710 Current.Previous->MatchingParen && 711 Current.Previous->MatchingParen->Type == TT_FunctionTypeLParen; 712 bool IsForOrCatch = Current.Previous && 713 Current.Previous->isOneOf(tok::kw_for, tok::kw_catch); 714 Contexts.back().IsExpression = !ParametersOfFunctionType && !IsForOrCatch; 715 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 716 for (FormatToken *Previous = Current.Previous; 717 Previous && Previous->isOneOf(tok::star, tok::amp); 718 Previous = Previous->Previous) 719 Previous->Type = TT_PointerOrReference; 720 } else if (Current.Previous && 721 Current.Previous->Type == TT_CtorInitializerColon) { 722 Contexts.back().IsExpression = true; 723 Contexts.back().InCtorInitializer = true; 724 } else if (Current.is(tok::kw_new)) { 725 Contexts.back().CanBeExpression = false; 726 } else if (Current.is(tok::semi) || Current.is(tok::exclaim)) { 727 // This should be the condition or increment in a for-loop. 728 Contexts.back().IsExpression = true; 729 } 730 731 if (Current.Type == TT_Unknown) { 732 // Line.MightBeFunctionDecl can only be true after the parentheses of a 733 // function declaration have been found. In this case, 'Current' is a 734 // trailing token of this declaration and thus cannot be a name. 735 if (isStartOfName(Current) && !Line.MightBeFunctionDecl) { 736 Contexts.back().FirstStartOfName = &Current; 737 Current.Type = TT_StartOfName; 738 } else if (Current.is(tok::kw_auto)) { 739 AutoFound = true; 740 } else if (Current.is(tok::arrow) && AutoFound && 741 Line.MustBeDeclaration) { 742 Current.Type = TT_TrailingReturnArrow; 743 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 744 Current.Type = 745 determineStarAmpUsage(Current, Contexts.back().CanBeExpression && 746 Contexts.back().IsExpression, 747 Contexts.back().InTemplateArgument); 748 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 749 Current.Type = determinePlusMinusCaretUsage(Current); 750 if (Current.Type == TT_UnaryOperator && Current.is(tok::caret)) 751 Contexts.back().CaretFound = true; 752 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 753 Current.Type = determineIncrementUsage(Current); 754 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 755 Current.Type = TT_UnaryOperator; 756 } else if (Current.is(tok::question)) { 757 Current.Type = TT_ConditionalExpr; 758 } else if (Current.isBinaryOperator() && 759 (!Current.Previous || 760 Current.Previous->isNot(tok::l_square))) { 761 Current.Type = TT_BinaryOperator; 762 } else if (Current.is(tok::comment)) { 763 if (Current.TokenText.startswith("//")) 764 Current.Type = TT_LineComment; 765 else 766 Current.Type = TT_BlockComment; 767 } else if (Current.is(tok::r_paren)) { 768 if (rParenEndsCast(Current)) 769 Current.Type = TT_CastRParen; 770 } else if (Current.is(tok::at) && Current.Next) { 771 switch (Current.Next->Tok.getObjCKeywordID()) { 772 case tok::objc_interface: 773 case tok::objc_implementation: 774 case tok::objc_protocol: 775 Current.Type = TT_ObjCDecl; 776 break; 777 case tok::objc_property: 778 Current.Type = TT_ObjCProperty; 779 break; 780 default: 781 break; 782 } 783 } else if (Current.is(tok::period)) { 784 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 785 if (PreviousNoComment && 786 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 787 Current.Type = TT_DesignatedInitializerPeriod; 788 } else if (Current.isOneOf(tok::identifier, tok::kw_const) && 789 Current.Previous && Current.Previous->isNot(tok::equal) && 790 Line.MightBeFunctionDecl && Contexts.size() == 1) { 791 // Line.MightBeFunctionDecl can only be true after the parentheses of a 792 // function declaration have been found. 793 Current.Type = TT_TrailingAnnotation; 794 } 795 } 796 } 797 798 /// \brief Take a guess at whether \p Tok starts a name of a function or 799 /// variable declaration. 800 /// 801 /// This is a heuristic based on whether \p Tok is an identifier following 802 /// something that is likely a type. 803 bool isStartOfName(const FormatToken &Tok) { 804 if (Tok.isNot(tok::identifier) || !Tok.Previous) 805 return false; 806 807 // Skip "const" as it does not have an influence on whether this is a name. 808 FormatToken *PreviousNotConst = Tok.Previous; 809 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 810 PreviousNotConst = PreviousNotConst->Previous; 811 812 if (!PreviousNotConst) 813 return false; 814 815 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 816 PreviousNotConst->Previous && 817 PreviousNotConst->Previous->is(tok::hash); 818 819 if (PreviousNotConst->Type == TT_TemplateCloser) 820 return PreviousNotConst && PreviousNotConst->MatchingParen && 821 PreviousNotConst->MatchingParen->Previous && 822 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 823 824 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen && 825 PreviousNotConst->MatchingParen->Previous && 826 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype)) 827 return true; 828 829 return (!IsPPKeyword && PreviousNotConst->is(tok::identifier)) || 830 PreviousNotConst->Type == TT_PointerOrReference || 831 PreviousNotConst->isSimpleTypeSpecifier(); 832 } 833 834 /// \brief Determine whether ')' is ending a cast. 835 bool rParenEndsCast(const FormatToken &Tok) { 836 FormatToken *LeftOfParens = nullptr; 837 if (Tok.MatchingParen) 838 LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 839 if (LeftOfParens && LeftOfParens->is(tok::r_paren)) 840 return false; 841 if (LeftOfParens && LeftOfParens->is(tok::r_square) && 842 LeftOfParens->MatchingParen && 843 LeftOfParens->MatchingParen->Type == TT_LambdaLSquare) 844 return false; 845 bool IsCast = false; 846 bool ParensAreEmpty = Tok.Previous == Tok.MatchingParen; 847 bool ParensAreType = !Tok.Previous || 848 Tok.Previous->Type == TT_PointerOrReference || 849 Tok.Previous->Type == TT_TemplateCloser || 850 Tok.Previous->isSimpleTypeSpecifier(); 851 bool ParensCouldEndDecl = 852 Tok.Next && Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace); 853 bool IsSizeOfOrAlignOf = 854 LeftOfParens && LeftOfParens->isOneOf(tok::kw_sizeof, tok::kw_alignof); 855 if (ParensAreType && !ParensCouldEndDecl && !IsSizeOfOrAlignOf && 856 ((Contexts.size() > 1 && Contexts[Contexts.size() - 2].IsExpression) || 857 (Tok.Next && Tok.Next->isBinaryOperator()))) 858 IsCast = true; 859 else if (Tok.Next && Tok.Next->isNot(tok::string_literal) && 860 (Tok.Next->Tok.isLiteral() || 861 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 862 IsCast = true; 863 // If there is an identifier after the (), it is likely a cast, unless 864 // there is also an identifier before the (). 865 else if (LeftOfParens && 866 (LeftOfParens->Tok.getIdentifierInfo() == nullptr || 867 LeftOfParens->is(tok::kw_return)) && 868 LeftOfParens->Type != TT_OverloadedOperator && 869 LeftOfParens->isNot(tok::at) && 870 LeftOfParens->Type != TT_TemplateCloser && Tok.Next) { 871 if (Tok.Next->isOneOf(tok::identifier, tok::numeric_constant)) { 872 IsCast = true; 873 } else { 874 // Use heuristics to recognize c style casting. 875 FormatToken *Prev = Tok.Previous; 876 if (Prev && Prev->isOneOf(tok::amp, tok::star)) 877 Prev = Prev->Previous; 878 879 if (Prev && Tok.Next && Tok.Next->Next) { 880 bool NextIsUnary = Tok.Next->isUnaryOperator() || 881 Tok.Next->isOneOf(tok::amp, tok::star); 882 IsCast = NextIsUnary && Tok.Next->Next->isOneOf( 883 tok::identifier, tok::numeric_constant); 884 } 885 886 for (; Prev != Tok.MatchingParen; Prev = Prev->Previous) { 887 if (!Prev || !Prev->isOneOf(tok::kw_const, tok::identifier)) { 888 IsCast = false; 889 break; 890 } 891 } 892 } 893 } 894 return IsCast && !ParensAreEmpty; 895 } 896 897 /// \brief Return the type of the given token assuming it is * or &. 898 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 899 bool InTemplateArgument) { 900 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 901 if (!PrevToken) 902 return TT_UnaryOperator; 903 904 const FormatToken *NextToken = Tok.getNextNonComment(); 905 if (!NextToken || NextToken->is(tok::l_brace)) 906 return TT_Unknown; 907 908 if (PrevToken->is(tok::coloncolon) || 909 (PrevToken->is(tok::l_paren) && !IsExpression)) 910 return TT_PointerOrReference; 911 912 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 913 tok::comma, tok::semi, tok::kw_return, tok::colon, 914 tok::equal, tok::kw_delete, tok::kw_sizeof) || 915 PrevToken->Type == TT_BinaryOperator || 916 PrevToken->Type == TT_ConditionalExpr || 917 PrevToken->Type == TT_UnaryOperator || PrevToken->Type == TT_CastRParen) 918 return TT_UnaryOperator; 919 920 if (NextToken->is(tok::l_square) && NextToken->Type != TT_LambdaLSquare) 921 return TT_PointerOrReference; 922 if (NextToken->isOneOf(tok::kw_operator, tok::comma)) 923 return TT_PointerOrReference; 924 925 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen && 926 PrevToken->MatchingParen->Previous && 927 PrevToken->MatchingParen->Previous->isOneOf(tok::kw_typeof, 928 tok::kw_decltype)) 929 return TT_PointerOrReference; 930 931 if (PrevToken->Tok.isLiteral() || 932 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 933 tok::kw_false) || 934 NextToken->Tok.isLiteral() || 935 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 936 NextToken->isUnaryOperator() || 937 // If we know we're in a template argument, there are no named 938 // declarations. Thus, having an identifier on the right-hand side 939 // indicates a binary operator. 940 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 941 return TT_BinaryOperator; 942 943 // This catches some cases where evaluation order is used as control flow: 944 // aaa && aaa->f(); 945 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 946 if (NextNextToken && NextNextToken->is(tok::arrow)) 947 return TT_BinaryOperator; 948 949 // It is very unlikely that we are going to find a pointer or reference type 950 // definition on the RHS of an assignment. 951 if (IsExpression) 952 return TT_BinaryOperator; 953 954 return TT_PointerOrReference; 955 } 956 957 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 958 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 959 if (!PrevToken || PrevToken->Type == TT_CastRParen) 960 return TT_UnaryOperator; 961 962 // Use heuristics to recognize unary operators. 963 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 964 tok::question, tok::colon, tok::kw_return, 965 tok::kw_case, tok::at, tok::l_brace)) 966 return TT_UnaryOperator; 967 968 // There can't be two consecutive binary operators. 969 if (PrevToken->Type == TT_BinaryOperator) 970 return TT_UnaryOperator; 971 972 // Fall back to marking the token as binary operator. 973 return TT_BinaryOperator; 974 } 975 976 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements. 977 TokenType determineIncrementUsage(const FormatToken &Tok) { 978 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 979 if (!PrevToken || PrevToken->Type == TT_CastRParen) 980 return TT_UnaryOperator; 981 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 982 return TT_TrailingUnaryOperator; 983 984 return TT_UnaryOperator; 985 } 986 987 SmallVector<Context, 8> Contexts; 988 989 const FormatStyle &Style; 990 AnnotatedLine &Line; 991 FormatToken *CurrentToken; 992 bool KeywordVirtualFound; 993 bool AutoFound; 994 IdentifierInfo &Ident_in; 995 }; 996 997 static int PrecedenceUnaryOperator = prec::PointerToMember + 1; 998 static int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 999 1000 /// \brief Parses binary expressions by inserting fake parenthesis based on 1001 /// operator precedence. 1002 class ExpressionParser { 1003 public: 1004 ExpressionParser(AnnotatedLine &Line) : Current(Line.First) { 1005 // Skip leading "}", e.g. in "} else if (...) {". 1006 if (Current->is(tok::r_brace)) 1007 next(); 1008 } 1009 1010 /// \brief Parse expressions with the given operatore precedence. 1011 void parse(int Precedence = 0) { 1012 // Skip 'return' and ObjC selector colons as they are not part of a binary 1013 // expression. 1014 while (Current && 1015 (Current->is(tok::kw_return) || 1016 (Current->is(tok::colon) && (Current->Type == TT_ObjCMethodExpr || 1017 Current->Type == TT_DictLiteral)))) 1018 next(); 1019 1020 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1021 return; 1022 1023 // Conditional expressions need to be parsed separately for proper nesting. 1024 if (Precedence == prec::Conditional) { 1025 parseConditionalExpr(); 1026 return; 1027 } 1028 1029 // Parse unary operators, which all have a higher precedence than binary 1030 // operators. 1031 if (Precedence == PrecedenceUnaryOperator) { 1032 parseUnaryOperator(); 1033 return; 1034 } 1035 1036 FormatToken *Start = Current; 1037 FormatToken *LatestOperator = nullptr; 1038 unsigned OperatorIndex = 0; 1039 1040 while (Current) { 1041 // Consume operators with higher precedence. 1042 parse(Precedence + 1); 1043 1044 int CurrentPrecedence = getCurrentPrecedence(); 1045 1046 if (Current && Current->Type == TT_SelectorName && 1047 Precedence == CurrentPrecedence) { 1048 if (LatestOperator) 1049 addFakeParenthesis(Start, prec::Level(Precedence)); 1050 Start = Current; 1051 } 1052 1053 // At the end of the line or when an operator with higher precedence is 1054 // found, insert fake parenthesis and return. 1055 if (!Current || Current->closesScope() || 1056 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence)) { 1057 if (LatestOperator) { 1058 LatestOperator->LastOperator = true; 1059 if (Precedence == PrecedenceArrowAndPeriod) { 1060 // Call expressions don't have a binary operator precedence. 1061 addFakeParenthesis(Start, prec::Unknown); 1062 } else { 1063 addFakeParenthesis(Start, prec::Level(Precedence)); 1064 } 1065 } 1066 return; 1067 } 1068 1069 // Consume scopes: (), [], <> and {} 1070 if (Current->opensScope()) { 1071 while (Current && !Current->closesScope()) { 1072 next(); 1073 parse(); 1074 } 1075 next(); 1076 } else { 1077 // Operator found. 1078 if (CurrentPrecedence == Precedence) { 1079 LatestOperator = Current; 1080 Current->OperatorIndex = OperatorIndex; 1081 ++OperatorIndex; 1082 } 1083 1084 next(); 1085 } 1086 } 1087 } 1088 1089 private: 1090 /// \brief Gets the precedence (+1) of the given token for binary operators 1091 /// and other tokens that we treat like binary operators. 1092 int getCurrentPrecedence() { 1093 if (Current) { 1094 if (Current->Type == TT_ConditionalExpr) 1095 return prec::Conditional; 1096 else if (Current->is(tok::semi) || Current->Type == TT_InlineASMColon || 1097 Current->Type == TT_SelectorName) 1098 return 0; 1099 else if (Current->Type == TT_RangeBasedForLoopColon) 1100 return prec::Comma; 1101 else if (Current->Type == TT_BinaryOperator || Current->is(tok::comma)) 1102 return Current->getPrecedence(); 1103 else if (Current->isOneOf(tok::period, tok::arrow)) 1104 return PrecedenceArrowAndPeriod; 1105 } 1106 return -1; 1107 } 1108 1109 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1110 Start->FakeLParens.push_back(Precedence); 1111 if (Precedence > prec::Unknown) 1112 Start->StartsBinaryExpression = true; 1113 if (Current) { 1114 ++Current->Previous->FakeRParens; 1115 if (Precedence > prec::Unknown) 1116 Current->Previous->EndsBinaryExpression = true; 1117 } 1118 } 1119 1120 /// \brief Parse unary operator expressions and surround them with fake 1121 /// parentheses if appropriate. 1122 void parseUnaryOperator() { 1123 if (!Current || Current->Type != TT_UnaryOperator) { 1124 parse(PrecedenceArrowAndPeriod); 1125 return; 1126 } 1127 1128 FormatToken *Start = Current; 1129 next(); 1130 parseUnaryOperator(); 1131 1132 // The actual precedence doesn't matter. 1133 addFakeParenthesis(Start, prec::Unknown); 1134 } 1135 1136 void parseConditionalExpr() { 1137 FormatToken *Start = Current; 1138 parse(prec::LogicalOr); 1139 if (!Current || !Current->is(tok::question)) 1140 return; 1141 next(); 1142 parseConditionalExpr(); 1143 if (!Current || Current->Type != TT_ConditionalExpr) 1144 return; 1145 next(); 1146 parseConditionalExpr(); 1147 addFakeParenthesis(Start, prec::Conditional); 1148 } 1149 1150 void next() { 1151 if (Current) 1152 Current = Current->Next; 1153 while (Current && Current->isTrailingComment()) 1154 Current = Current->Next; 1155 } 1156 1157 FormatToken *Current; 1158 }; 1159 1160 } // end anonymous namespace 1161 1162 void 1163 TokenAnnotator::setCommentLineLevels(SmallVectorImpl<AnnotatedLine *> &Lines) { 1164 const AnnotatedLine *NextNonCommentLine = nullptr; 1165 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 1166 E = Lines.rend(); 1167 I != E; ++I) { 1168 if (NextNonCommentLine && (*I)->First->is(tok::comment) && 1169 (*I)->First->Next == nullptr) 1170 (*I)->Level = NextNonCommentLine->Level; 1171 else 1172 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 1173 1174 setCommentLineLevels((*I)->Children); 1175 } 1176 } 1177 1178 void TokenAnnotator::annotate(AnnotatedLine &Line) { 1179 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1180 E = Line.Children.end(); 1181 I != E; ++I) { 1182 annotate(**I); 1183 } 1184 AnnotatingParser Parser(Style, Line, Ident_in); 1185 Line.Type = Parser.parseLine(); 1186 if (Line.Type == LT_Invalid) 1187 return; 1188 1189 ExpressionParser ExprParser(Line); 1190 ExprParser.parse(); 1191 1192 if (Line.First->Type == TT_ObjCMethodSpecifier) 1193 Line.Type = LT_ObjCMethodDecl; 1194 else if (Line.First->Type == TT_ObjCDecl) 1195 Line.Type = LT_ObjCDecl; 1196 else if (Line.First->Type == TT_ObjCProperty) 1197 Line.Type = LT_ObjCProperty; 1198 1199 Line.First->SpacesRequiredBefore = 1; 1200 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 1201 } 1202 1203 // This function heuristically determines whether 'Current' starts the name of a 1204 // function declaration. 1205 static bool isFunctionDeclarationName(const FormatToken &Current) { 1206 if (Current.Type != TT_StartOfName || 1207 Current.NestingLevel != 0 || 1208 Current.Previous->Type == TT_StartOfName) 1209 return false; 1210 const FormatToken *Next = Current.Next; 1211 for (; Next; Next = Next->Next) { 1212 if (Next->Type == TT_TemplateOpener) { 1213 Next = Next->MatchingParen; 1214 } else if (Next->is(tok::coloncolon)) { 1215 Next = Next->Next; 1216 if (!Next || !Next->is(tok::identifier)) 1217 return false; 1218 } else if (Next->is(tok::l_paren)) { 1219 break; 1220 } else { 1221 return false; 1222 } 1223 } 1224 if (!Next) 1225 return false; 1226 assert(Next->is(tok::l_paren)); 1227 if (Next->Next == Next->MatchingParen) 1228 return true; 1229 for (const FormatToken *Tok = Next->Next; Tok != Next->MatchingParen; 1230 Tok = Tok->Next) { 1231 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 1232 Tok->Type == TT_PointerOrReference || Tok->Type == TT_StartOfName) 1233 return true; 1234 if (Tok->isOneOf(tok::l_brace, tok::string_literal) || Tok->Tok.isLiteral()) 1235 return false; 1236 } 1237 return false; 1238 } 1239 1240 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 1241 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1242 E = Line.Children.end(); 1243 I != E; ++I) { 1244 calculateFormattingInformation(**I); 1245 } 1246 1247 Line.First->TotalLength = 1248 Line.First->IsMultiline ? Style.ColumnLimit : Line.First->ColumnWidth; 1249 if (!Line.First->Next) 1250 return; 1251 FormatToken *Current = Line.First->Next; 1252 bool InFunctionDecl = Line.MightBeFunctionDecl; 1253 while (Current) { 1254 if (isFunctionDeclarationName(*Current)) 1255 Current->Type = TT_FunctionDeclarationName; 1256 if (Current->Type == TT_LineComment) { 1257 if (Current->Previous->BlockKind == BK_BracedInit && 1258 Current->Previous->opensScope()) 1259 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1; 1260 else 1261 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 1262 1263 // If we find a trailing comment, iterate backwards to determine whether 1264 // it seems to relate to a specific parameter. If so, break before that 1265 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 1266 // to the previous line in: 1267 // SomeFunction(a, 1268 // b, // comment 1269 // c); 1270 if (!Current->HasUnescapedNewline) { 1271 for (FormatToken *Parameter = Current->Previous; Parameter; 1272 Parameter = Parameter->Previous) { 1273 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 1274 break; 1275 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 1276 if (Parameter->Previous->Type != TT_CtorInitializerComma && 1277 Parameter->HasUnescapedNewline) 1278 Parameter->MustBreakBefore = true; 1279 break; 1280 } 1281 } 1282 } 1283 } else if (Current->SpacesRequiredBefore == 0 && 1284 spaceRequiredBefore(Line, *Current)) { 1285 Current->SpacesRequiredBefore = 1; 1286 } 1287 1288 Current->MustBreakBefore = 1289 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 1290 1291 if (Style.AlwaysBreakAfterDefinitionReturnType && 1292 InFunctionDecl && Current->Type == TT_FunctionDeclarationName && 1293 Line.Last->is(tok::l_brace)) // Only for definitions. 1294 Current->MustBreakBefore = true; 1295 1296 Current->CanBreakBefore = 1297 Current->MustBreakBefore || canBreakBefore(Line, *Current); 1298 unsigned ChildSize = 0; 1299 if (Current->Previous->Children.size() == 1) { 1300 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 1301 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 1302 : LastOfChild.TotalLength + 1; 1303 } 1304 if (Current->MustBreakBefore || Current->Previous->Children.size() > 1 || 1305 Current->IsMultiline) 1306 Current->TotalLength = Current->Previous->TotalLength + Style.ColumnLimit; 1307 else 1308 Current->TotalLength = Current->Previous->TotalLength + 1309 Current->ColumnWidth + ChildSize + 1310 Current->SpacesRequiredBefore; 1311 1312 if (Current->Type == TT_CtorInitializerColon) 1313 InFunctionDecl = false; 1314 1315 // FIXME: Only calculate this if CanBreakBefore is true once static 1316 // initializers etc. are sorted out. 1317 // FIXME: Move magic numbers to a better place. 1318 Current->SplitPenalty = 20 * Current->BindingStrength + 1319 splitPenalty(Line, *Current, InFunctionDecl); 1320 1321 Current = Current->Next; 1322 } 1323 1324 calculateUnbreakableTailLengths(Line); 1325 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 1326 if (Current->Role) 1327 Current->Role->precomputeFormattingInfos(Current); 1328 } 1329 1330 DEBUG({ printDebugInfo(Line); }); 1331 } 1332 1333 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 1334 unsigned UnbreakableTailLength = 0; 1335 FormatToken *Current = Line.Last; 1336 while (Current) { 1337 Current->UnbreakableTailLength = UnbreakableTailLength; 1338 if (Current->CanBreakBefore || 1339 Current->isOneOf(tok::comment, tok::string_literal)) { 1340 UnbreakableTailLength = 0; 1341 } else { 1342 UnbreakableTailLength += 1343 Current->ColumnWidth + Current->SpacesRequiredBefore; 1344 } 1345 Current = Current->Previous; 1346 } 1347 } 1348 1349 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 1350 const FormatToken &Tok, 1351 bool InFunctionDecl) { 1352 const FormatToken &Left = *Tok.Previous; 1353 const FormatToken &Right = Tok; 1354 1355 if (Left.is(tok::semi)) 1356 return 0; 1357 if (Left.is(tok::comma) || (Right.is(tok::identifier) && Right.Next && 1358 Right.Next->Type == TT_DictLiteral)) 1359 return 1; 1360 if (Right.is(tok::l_square)) { 1361 if (Style.Language == FormatStyle::LK_Proto) 1362 return 1; 1363 if (Right.Type != TT_ObjCMethodExpr && Right.Type != TT_LambdaLSquare) 1364 return 500; 1365 } 1366 if (Right.Type == TT_StartOfName || 1367 Right.Type == TT_FunctionDeclarationName || Right.is(tok::kw_operator)) { 1368 if (Line.First->is(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 1369 return 3; 1370 if (Left.Type == TT_StartOfName) 1371 return 20; 1372 if (InFunctionDecl && Right.NestingLevel == 0) 1373 return Style.PenaltyReturnTypeOnItsOwnLine; 1374 return 200; 1375 } 1376 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 1377 return 150; 1378 if (Left.Type == TT_CastRParen) 1379 return 100; 1380 if (Left.is(tok::coloncolon) || 1381 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 1382 return 500; 1383 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 1384 return 5000; 1385 1386 if (Left.Type == TT_RangeBasedForLoopColon || 1387 Left.Type == TT_InheritanceColon) 1388 return 2; 1389 1390 if (Right.isMemberAccess()) { 1391 if (Left.is(tok::r_paren) && Left.MatchingParen && 1392 Left.MatchingParen->ParameterCount > 0) 1393 return 20; // Should be smaller than breaking at a nested comma. 1394 return 150; 1395 } 1396 1397 if (Right.Type == TT_TrailingAnnotation && 1398 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 1399 // Generally, breaking before a trailing annotation is bad unless it is 1400 // function-like. It seems to be especially preferable to keep standard 1401 // annotations (i.e. "const", "final" and "override") on the same line. 1402 // Use a slightly higher penalty after ")" so that annotations like 1403 // "const override" are kept together. 1404 bool is_short_annotation = Right.TokenText.size() < 10; 1405 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 1406 } 1407 1408 // In for-loops, prefer breaking at ',' and ';'. 1409 if (Line.First->is(tok::kw_for) && Left.is(tok::equal)) 1410 return 4; 1411 1412 // In Objective-C method expressions, prefer breaking before "param:" over 1413 // breaking after it. 1414 if (Right.Type == TT_SelectorName) 1415 return 0; 1416 if (Left.is(tok::colon) && Left.Type == TT_ObjCMethodExpr) 1417 return Line.MightBeFunctionDecl ? 50 : 500; 1418 1419 if (Left.is(tok::l_paren) && InFunctionDecl) 1420 return 100; 1421 if (Left.is(tok::equal) && InFunctionDecl) 1422 return 110; 1423 if (Right.is(tok::r_brace)) 1424 return 1; 1425 if (Left.opensScope()) 1426 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 1427 : 19; 1428 1429 if (Right.is(tok::lessless)) { 1430 if (Left.is(tok::string_literal)) { 1431 StringRef Content = Left.TokenText; 1432 if (Content.startswith("\"")) 1433 Content = Content.drop_front(1); 1434 if (Content.endswith("\"")) 1435 Content = Content.drop_back(1); 1436 Content = Content.trim(); 1437 if (Content.size() > 1 && 1438 (Content.back() == ':' || Content.back() == '=')) 1439 return 25; 1440 } 1441 return 1; // Breaking at a << is really cheap. 1442 } 1443 if (Left.Type == TT_ConditionalExpr) 1444 return prec::Conditional; 1445 prec::Level Level = Left.getPrecedence(); 1446 1447 if (Level != prec::Unknown) 1448 return Level; 1449 1450 return 3; 1451 } 1452 1453 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 1454 const FormatToken &Left, 1455 const FormatToken &Right) { 1456 if (Style.Language == FormatStyle::LK_Proto) { 1457 if (Right.is(tok::period) && 1458 (Left.TokenText == "optional" || Left.TokenText == "required" || 1459 Left.TokenText == "repeated")) 1460 return true; 1461 if (Right.is(tok::l_paren) && 1462 (Left.TokenText == "returns" || Left.TokenText == "option")) 1463 return true; 1464 } else if (Style.Language == FormatStyle::LK_JavaScript) { 1465 if (Left.TokenText == "var") 1466 return true; 1467 } 1468 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 1469 return true; 1470 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 1471 Left.Tok.getObjCKeywordID() == tok::objc_property) 1472 return true; 1473 if (Right.is(tok::hashhash)) 1474 return Left.is(tok::hash); 1475 if (Left.isOneOf(tok::hashhash, tok::hash)) 1476 return Right.is(tok::hash); 1477 if (Left.is(tok::l_paren) && Right.is(tok::r_paren)) 1478 return Style.SpaceInEmptyParentheses; 1479 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 1480 return (Right.Type == TT_CastRParen || 1481 (Left.MatchingParen && Left.MatchingParen->Type == TT_CastRParen)) 1482 ? Style.SpacesInCStyleCastParentheses 1483 : Style.SpacesInParentheses; 1484 if (Style.SpacesInAngles && 1485 ((Left.Type == TT_TemplateOpener) != (Right.Type == TT_TemplateCloser))) 1486 return true; 1487 if (Right.isOneOf(tok::semi, tok::comma)) 1488 return false; 1489 if (Right.is(tok::less) && 1490 (Left.isOneOf(tok::kw_template, tok::r_paren) || 1491 (Line.Type == LT_ObjCDecl && Style.ObjCSpaceBeforeProtocolList))) 1492 return true; 1493 if (Left.is(tok::arrow) || Right.is(tok::arrow)) 1494 return false; 1495 if (Left.isOneOf(tok::exclaim, tok::tilde)) 1496 return false; 1497 if (Left.is(tok::at) && 1498 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 1499 tok::numeric_constant, tok::l_paren, tok::l_brace, 1500 tok::kw_true, tok::kw_false)) 1501 return false; 1502 if (Left.is(tok::coloncolon)) 1503 return false; 1504 if (Right.is(tok::coloncolon) && Left.isNot(tok::l_brace)) 1505 return (Left.is(tok::less) && Style.Standard == FormatStyle::LS_Cpp03) || 1506 !Left.isOneOf(tok::identifier, tok::greater, tok::l_paren, 1507 tok::r_paren, tok::less); 1508 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) 1509 return false; 1510 if (Right.is(tok::ellipsis)) 1511 return Left.Tok.isLiteral(); 1512 if (Left.is(tok::l_square) && Right.is(tok::amp)) 1513 return false; 1514 if (Right.Type == TT_PointerOrReference) 1515 return Left.Tok.isLiteral() || 1516 ((Left.Type != TT_PointerOrReference) && Left.isNot(tok::l_paren) && 1517 Style.PointerAlignment != FormatStyle::PAS_Left); 1518 if (Right.Type == TT_FunctionTypeLParen && Left.isNot(tok::l_paren) && 1519 (Left.Type != TT_PointerOrReference || 1520 Style.PointerAlignment != FormatStyle::PAS_Right)) 1521 return true; 1522 if (Left.Type == TT_PointerOrReference) 1523 return Right.Tok.isLiteral() || Right.Type == TT_BlockComment || 1524 ((Right.Type != TT_PointerOrReference) && 1525 Right.isNot(tok::l_paren) && 1526 Style.PointerAlignment != FormatStyle::PAS_Right && Left.Previous && 1527 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon)); 1528 if (Right.is(tok::star) && Left.is(tok::l_paren)) 1529 return false; 1530 if (Left.is(tok::l_square)) 1531 return Left.Type == TT_ArrayInitializerLSquare && 1532 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square); 1533 if (Right.is(tok::r_square)) 1534 return Right.MatchingParen && Style.SpacesInContainerLiterals && 1535 Right.MatchingParen->Type == TT_ArrayInitializerLSquare; 1536 if (Right.is(tok::l_square) && Right.Type != TT_ObjCMethodExpr && 1537 Right.Type != TT_LambdaLSquare && Left.isNot(tok::numeric_constant) && 1538 Left.Type != TT_DictLiteral) 1539 return false; 1540 if (Left.is(tok::colon)) 1541 return Left.Type != TT_ObjCMethodExpr; 1542 if (Left.Type == TT_BlockComment) 1543 return !Left.TokenText.endswith("=*/"); 1544 if (Right.is(tok::l_paren)) { 1545 if (Left.is(tok::r_paren) && Left.Type == TT_AttributeParen) 1546 return true; 1547 return Line.Type == LT_ObjCDecl || 1548 Left.isOneOf(tok::kw_new, tok::kw_delete, tok::semi) || 1549 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 1550 (Left.isOneOf(tok::kw_if, tok::kw_for, tok::kw_while, 1551 tok::kw_switch, tok::kw_catch, tok::kw_case) || 1552 Left.IsForEachMacro)) || 1553 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always && 1554 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword()) && 1555 Line.Type != LT_PreprocessorDirective); 1556 } 1557 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 1558 return false; 1559 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 1560 return !Left.Children.empty(); // No spaces in "{}". 1561 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 1562 (Right.is(tok::r_brace) && Right.MatchingParen && 1563 Right.MatchingParen->BlockKind != BK_Block)) 1564 return !Style.Cpp11BracedListStyle; 1565 if (Right.Type == TT_UnaryOperator) 1566 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 1567 (Left.isNot(tok::colon) || Left.Type != TT_ObjCMethodExpr); 1568 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 1569 tok::r_paren) || 1570 Left.isSimpleTypeSpecifier()) && 1571 Right.is(tok::l_brace) && Right.getNextNonComment() && 1572 Right.BlockKind != BK_Block) 1573 return false; 1574 if (Left.is(tok::period) || Right.is(tok::period)) 1575 return false; 1576 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 1577 return false; 1578 return true; 1579 } 1580 1581 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 1582 const FormatToken &Tok) { 1583 if (Tok.Tok.getIdentifierInfo() && Tok.Previous->Tok.getIdentifierInfo()) 1584 return true; // Never ever merge two identifiers. 1585 if (Tok.Previous->Type == TT_ImplicitStringLiteral) 1586 return Tok.WhitespaceRange.getBegin() != Tok.WhitespaceRange.getEnd(); 1587 if (Line.Type == LT_ObjCMethodDecl) { 1588 if (Tok.Previous->Type == TT_ObjCMethodSpecifier) 1589 return true; 1590 if (Tok.Previous->is(tok::r_paren) && Tok.is(tok::identifier)) 1591 // Don't space between ')' and <id> 1592 return false; 1593 } 1594 if (Line.Type == LT_ObjCProperty && 1595 (Tok.is(tok::equal) || Tok.Previous->is(tok::equal))) 1596 return false; 1597 1598 if (Tok.Type == TT_TrailingReturnArrow || 1599 Tok.Previous->Type == TT_TrailingReturnArrow) 1600 return true; 1601 if (Tok.Previous->is(tok::comma)) 1602 return true; 1603 if (Tok.is(tok::comma)) 1604 return false; 1605 if (Tok.Type == TT_CtorInitializerColon || Tok.Type == TT_ObjCBlockLParen) 1606 return true; 1607 if (Tok.Previous->Tok.is(tok::kw_operator)) 1608 return Tok.is(tok::coloncolon); 1609 if (Tok.Type == TT_OverloadedOperatorLParen) 1610 return false; 1611 if (Tok.is(tok::colon)) 1612 return !Line.First->isOneOf(tok::kw_case, tok::kw_default) && 1613 Tok.getNextNonComment() && Tok.Type != TT_ObjCMethodExpr && 1614 !Tok.Previous->is(tok::question) && 1615 (Tok.Type != TT_DictLiteral || Style.SpacesInContainerLiterals); 1616 if (Tok.Previous->Type == TT_UnaryOperator || 1617 Tok.Previous->Type == TT_CastRParen) 1618 return Tok.Type == TT_BinaryOperator; 1619 if (Tok.Previous->is(tok::greater) && Tok.is(tok::greater)) { 1620 return Tok.Type == TT_TemplateCloser && 1621 Tok.Previous->Type == TT_TemplateCloser && 1622 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 1623 } 1624 if (Tok.isOneOf(tok::arrowstar, tok::periodstar) || 1625 Tok.Previous->isOneOf(tok::arrowstar, tok::periodstar)) 1626 return false; 1627 if (!Style.SpaceBeforeAssignmentOperators && 1628 Tok.getPrecedence() == prec::Assignment) 1629 return false; 1630 if ((Tok.Type == TT_BinaryOperator && !Tok.Previous->is(tok::l_paren)) || 1631 Tok.Previous->Type == TT_BinaryOperator || 1632 Tok.Previous->Type == TT_ConditionalExpr) 1633 return true; 1634 if (Tok.Previous->Type == TT_TemplateCloser && Tok.is(tok::l_paren)) 1635 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 1636 if (Tok.is(tok::less) && Tok.Previous->isNot(tok::l_paren) && 1637 Line.First->is(tok::hash)) 1638 return true; 1639 if (Tok.Type == TT_TrailingUnaryOperator) 1640 return false; 1641 if (Tok.Previous->Type == TT_RegexLiteral) 1642 return false; 1643 return spaceRequiredBetween(Line, *Tok.Previous, Tok); 1644 } 1645 1646 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 1647 static bool isAllmanBrace(const FormatToken &Tok) { 1648 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 1649 Tok.Type != TT_ObjCBlockLBrace && Tok.Type != TT_DictLiteral; 1650 } 1651 1652 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 1653 const FormatToken &Right) { 1654 const FormatToken &Left = *Right.Previous; 1655 if (Right.NewlinesBefore > 1) 1656 return true; 1657 if (Right.is(tok::comment)) { 1658 return Right.Previous->BlockKind != BK_BracedInit && 1659 Right.Previous->Type != TT_CtorInitializerColon && 1660 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 1661 } else if (Right.Previous->isTrailingComment() || 1662 (Right.isStringLiteral() && Right.Previous->isStringLiteral())) { 1663 return true; 1664 } else if (Right.Previous->IsUnterminatedLiteral) { 1665 return true; 1666 } else if (Right.is(tok::lessless) && Right.Next && 1667 Right.Previous->is(tok::string_literal) && 1668 Right.Next->is(tok::string_literal)) { 1669 return true; 1670 } else if (Right.Previous->ClosesTemplateDeclaration && 1671 Right.Previous->MatchingParen && 1672 Right.Previous->MatchingParen->NestingLevel == 0 && 1673 Style.AlwaysBreakTemplateDeclarations) { 1674 return true; 1675 } else if ((Right.Type == TT_CtorInitializerComma || 1676 Right.Type == TT_CtorInitializerColon) && 1677 Style.BreakConstructorInitializersBeforeComma && 1678 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) { 1679 return true; 1680 } else if (Right.is(tok::string_literal) && 1681 Right.TokenText.startswith("R\"")) { 1682 // Raw string literals are special wrt. line breaks. The author has made a 1683 // deliberate choice and might have aligned the contents of the string 1684 // literal accordingly. Thus, we try keep existing line breaks. 1685 return Right.NewlinesBefore > 0; 1686 } else if (Right.Previous->is(tok::l_brace) && Right.NestingLevel == 1 && 1687 Style.Language == FormatStyle::LK_Proto) { 1688 // Don't enums onto single lines in protocol buffers. 1689 return true; 1690 } else if (isAllmanBrace(Left) || isAllmanBrace(Right)) { 1691 return Style.BreakBeforeBraces == FormatStyle::BS_Allman || 1692 Style.BreakBeforeBraces == FormatStyle::BS_GNU; 1693 } else if (Style.Language == FormatStyle::LK_Proto && 1694 Left.isNot(tok::l_brace) && Right.Type == TT_SelectorName) { 1695 return true; 1696 } 1697 1698 // If the last token before a '}' is a comma or a comment, the intention is to 1699 // insert a line break after it in order to make shuffling around entries 1700 // easier. 1701 const FormatToken *BeforeClosingBrace = nullptr; 1702 if (Left.is(tok::l_brace) && Left.MatchingParen) 1703 BeforeClosingBrace = Left.MatchingParen->Previous; 1704 else if (Right.is(tok::r_brace)) 1705 BeforeClosingBrace = Right.Previous; 1706 if (BeforeClosingBrace && 1707 BeforeClosingBrace->isOneOf(tok::comma, tok::comment)) 1708 return true; 1709 1710 if (Style.Language == FormatStyle::LK_JavaScript) { 1711 // FIXME: This might apply to other languages and token kinds. 1712 if (Right.is(tok::char_constant) && Left.is(tok::plus) && Left.Previous && 1713 Left.Previous->is(tok::char_constant)) 1714 return true; 1715 } 1716 1717 return false; 1718 } 1719 1720 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 1721 const FormatToken &Right) { 1722 const FormatToken &Left = *Right.Previous; 1723 if (Left.is(tok::at)) 1724 return false; 1725 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 1726 return false; 1727 if (Right.Type == TT_StartOfName || 1728 Right.Type == TT_FunctionDeclarationName || Right.is(tok::kw_operator)) 1729 return true; 1730 if (Right.isTrailingComment()) 1731 // We rely on MustBreakBefore being set correctly here as we should not 1732 // change the "binding" behavior of a comment. 1733 // The first comment in a braced lists is always interpreted as belonging to 1734 // the first list element. Otherwise, it should be placed outside of the 1735 // list. 1736 return Left.BlockKind == BK_BracedInit; 1737 if (Left.is(tok::question) && Right.is(tok::colon)) 1738 return false; 1739 if (Right.Type == TT_ConditionalExpr || Right.is(tok::question)) 1740 return Style.BreakBeforeTernaryOperators; 1741 if (Left.Type == TT_ConditionalExpr || Left.is(tok::question)) 1742 return !Style.BreakBeforeTernaryOperators; 1743 if (Right.Type == TT_InheritanceColon) 1744 return true; 1745 if (Right.is(tok::colon) && (Right.Type != TT_CtorInitializerColon && 1746 Right.Type != TT_InlineASMColon)) 1747 return false; 1748 if (Left.is(tok::colon) && 1749 (Left.Type == TT_DictLiteral || Left.Type == TT_ObjCMethodExpr)) 1750 return true; 1751 if (Right.Type == TT_SelectorName) 1752 return true; 1753 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 1754 return true; 1755 if (Left.ClosesTemplateDeclaration) 1756 return true; 1757 if (Right.Type == TT_RangeBasedForLoopColon || 1758 Right.Type == TT_OverloadedOperatorLParen || 1759 Right.Type == TT_OverloadedOperator) 1760 return false; 1761 if (Left.Type == TT_RangeBasedForLoopColon) 1762 return true; 1763 if (Right.Type == TT_RangeBasedForLoopColon) 1764 return false; 1765 if (Left.Type == TT_PointerOrReference || Left.Type == TT_TemplateCloser || 1766 Left.Type == TT_UnaryOperator || Left.is(tok::kw_operator)) 1767 return false; 1768 if (Left.is(tok::equal) && Line.Type == LT_VirtualFunctionDecl) 1769 return false; 1770 if (Left.is(tok::l_paren) && Left.Type == TT_AttributeParen) 1771 return false; 1772 if (Left.is(tok::l_paren) && Left.Previous && 1773 (Left.Previous->Type == TT_BinaryOperator || 1774 Left.Previous->Type == TT_CastRParen || Left.Previous->is(tok::kw_if))) 1775 return false; 1776 if (Right.Type == TT_ImplicitStringLiteral) 1777 return false; 1778 1779 if (Right.is(tok::r_paren) || Right.Type == TT_TemplateCloser) 1780 return false; 1781 1782 // We only break before r_brace if there was a corresponding break before 1783 // the l_brace, which is tracked by BreakBeforeClosingBrace. 1784 if (Right.is(tok::r_brace)) 1785 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 1786 1787 // Allow breaking after a trailing annotation, e.g. after a method 1788 // declaration. 1789 if (Left.Type == TT_TrailingAnnotation) 1790 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 1791 tok::less, tok::coloncolon); 1792 1793 if (Right.is(tok::kw___attribute)) 1794 return true; 1795 1796 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 1797 return true; 1798 1799 if (Right.is(tok::identifier) && Right.Next && 1800 Right.Next->Type == TT_DictLiteral) 1801 return true; 1802 1803 if (Left.Type == TT_CtorInitializerComma && 1804 Style.BreakConstructorInitializersBeforeComma) 1805 return false; 1806 if (Right.Type == TT_CtorInitializerComma && 1807 Style.BreakConstructorInitializersBeforeComma) 1808 return true; 1809 if (Left.is(tok::greater) && Right.is(tok::greater) && 1810 Left.Type != TT_TemplateCloser) 1811 return false; 1812 if (Right.Type == TT_BinaryOperator && Style.BreakBeforeBinaryOperators) 1813 return true; 1814 if (Left.Type == TT_ArrayInitializerLSquare) 1815 return true; 1816 return (Left.isBinaryOperator() && 1817 !Left.isOneOf(tok::arrowstar, tok::lessless) && 1818 !Style.BreakBeforeBinaryOperators) || 1819 Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 1820 tok::kw_class, tok::kw_struct) || 1821 Right.isMemberAccess() || 1822 Right.isOneOf(tok::lessless, tok::colon, tok::l_square, tok::at, 1823 tok::kw_typename) || 1824 (Left.is(tok::r_paren) && 1825 Right.isOneOf(tok::identifier, tok::kw_const)) || 1826 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)); 1827 } 1828 1829 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 1830 llvm::errs() << "AnnotatedTokens:\n"; 1831 const FormatToken *Tok = Line.First; 1832 while (Tok) { 1833 llvm::errs() << " M=" << Tok->MustBreakBefore 1834 << " C=" << Tok->CanBreakBefore << " T=" << Tok->Type 1835 << " S=" << Tok->SpacesRequiredBefore 1836 << " B=" << Tok->BlockParameterCount 1837 << " P=" << Tok->SplitPenalty << " Name=" << Tok->Tok.getName() 1838 << " L=" << Tok->TotalLength << " PPK=" << Tok->PackingKind 1839 << " FakeLParens="; 1840 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 1841 llvm::errs() << Tok->FakeLParens[i] << "/"; 1842 llvm::errs() << " FakeRParens=" << Tok->FakeRParens << "\n"; 1843 if (!Tok->Next) 1844 assert(Tok == Line.Last); 1845 Tok = Tok->Next; 1846 } 1847 llvm::errs() << "----\n"; 1848 } 1849 1850 } // namespace format 1851 } // namespace clang 1852