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->isOneOf(tok::colon, tok::l_brace)) { 315 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 316 if ((CurrentToken->is(tok::colon) || 317 Style.Language == FormatStyle::LK_Proto) && 318 Previous->is(tok::identifier)) 319 Previous->Type = TT_SelectorName; 320 if (CurrentToken->is(tok::colon)) 321 Left->Type = TT_DictLiteral; 322 } 323 if (!consumeToken()) 324 return false; 325 } 326 } 327 return true; 328 } 329 330 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 331 if (Current->Type == TT_LambdaLSquare || 332 (Current->is(tok::caret) && Current->Type == TT_UnaryOperator) || 333 (Style.Language == FormatStyle::LK_JavaScript && 334 Current->TokenText == "function")) { 335 ++Left->BlockParameterCount; 336 } 337 if (Current->is(tok::comma)) { 338 ++Left->ParameterCount; 339 if (!Left->Role) 340 Left->Role.reset(new CommaSeparatedList(Style)); 341 Left->Role->CommaFound(Current); 342 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 343 Left->ParameterCount = 1; 344 } 345 } 346 347 bool parseConditional() { 348 while (CurrentToken) { 349 if (CurrentToken->is(tok::colon)) { 350 CurrentToken->Type = TT_ConditionalExpr; 351 next(); 352 return true; 353 } 354 if (!consumeToken()) 355 return false; 356 } 357 return false; 358 } 359 360 bool parseTemplateDeclaration() { 361 if (CurrentToken && CurrentToken->is(tok::less)) { 362 CurrentToken->Type = TT_TemplateOpener; 363 next(); 364 if (!parseAngle()) 365 return false; 366 if (CurrentToken) 367 CurrentToken->Previous->ClosesTemplateDeclaration = true; 368 return true; 369 } 370 return false; 371 } 372 373 bool consumeToken() { 374 FormatToken *Tok = CurrentToken; 375 next(); 376 switch (Tok->Tok.getKind()) { 377 case tok::plus: 378 case tok::minus: 379 if (!Tok->Previous && Line.MustBeDeclaration) 380 Tok->Type = TT_ObjCMethodSpecifier; 381 break; 382 case tok::colon: 383 if (!Tok->Previous) 384 return false; 385 // Colons from ?: are handled in parseConditional(). 386 if (Tok->Previous->is(tok::r_paren) && Contexts.size() == 1 && 387 Line.First->isNot(tok::kw_case)) { 388 Tok->Type = TT_CtorInitializerColon; 389 } else if (Contexts.back().ColonIsDictLiteral) { 390 Tok->Type = TT_DictLiteral; 391 } else if (Contexts.back().ColonIsObjCMethodExpr || 392 Line.First->Type == TT_ObjCMethodSpecifier) { 393 Tok->Type = TT_ObjCMethodExpr; 394 Tok->Previous->Type = TT_SelectorName; 395 if (Tok->Previous->ColumnWidth > 396 Contexts.back().LongestObjCSelectorName) { 397 Contexts.back().LongestObjCSelectorName = Tok->Previous->ColumnWidth; 398 } 399 if (!Contexts.back().FirstObjCSelectorName) 400 Contexts.back().FirstObjCSelectorName = Tok->Previous; 401 } else if (Contexts.back().ColonIsForRangeExpr) { 402 Tok->Type = TT_RangeBasedForLoopColon; 403 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 404 Tok->Type = TT_BitFieldColon; 405 } else if (Contexts.size() == 1 && 406 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) { 407 Tok->Type = TT_InheritanceColon; 408 } else if (Contexts.back().ContextKind == tok::l_paren) { 409 Tok->Type = TT_InlineASMColon; 410 } 411 break; 412 case tok::kw_if: 413 case tok::kw_while: 414 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 415 next(); 416 if (!parseParens(/*LookForDecls=*/true)) 417 return false; 418 } 419 break; 420 case tok::kw_for: 421 Contexts.back().ColonIsForRangeExpr = true; 422 next(); 423 if (!parseParens()) 424 return false; 425 break; 426 case tok::l_paren: 427 if (!parseParens()) 428 return false; 429 if (Line.MustBeDeclaration && Contexts.size() == 1 && 430 !Contexts.back().IsExpression && 431 Line.First->Type != TT_ObjCProperty && 432 (!Tok->Previous || Tok->Previous->isNot(tok::kw_decltype))) 433 Line.MightBeFunctionDecl = true; 434 break; 435 case tok::l_square: 436 if (!parseSquare()) 437 return false; 438 break; 439 case tok::l_brace: 440 if (!parseBrace()) 441 return false; 442 break; 443 case tok::less: 444 if (Tok->Previous && !Tok->Previous->Tok.isLiteral() && parseAngle()) 445 Tok->Type = TT_TemplateOpener; 446 else { 447 Tok->Type = TT_BinaryOperator; 448 CurrentToken = Tok; 449 next(); 450 } 451 break; 452 case tok::r_paren: 453 case tok::r_square: 454 return false; 455 case tok::r_brace: 456 // Lines can start with '}'. 457 if (Tok->Previous) 458 return false; 459 break; 460 case tok::greater: 461 Tok->Type = TT_BinaryOperator; 462 break; 463 case tok::kw_operator: 464 while (CurrentToken && 465 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 466 if (CurrentToken->isOneOf(tok::star, tok::amp)) 467 CurrentToken->Type = TT_PointerOrReference; 468 consumeToken(); 469 if (CurrentToken && CurrentToken->Previous->Type == TT_BinaryOperator) 470 CurrentToken->Previous->Type = TT_OverloadedOperator; 471 } 472 if (CurrentToken) { 473 CurrentToken->Type = TT_OverloadedOperatorLParen; 474 if (CurrentToken->Previous->Type == TT_BinaryOperator) 475 CurrentToken->Previous->Type = TT_OverloadedOperator; 476 } 477 break; 478 case tok::question: 479 parseConditional(); 480 break; 481 case tok::kw_template: 482 parseTemplateDeclaration(); 483 break; 484 case tok::identifier: 485 if (Line.First->is(tok::kw_for) && 486 Tok->Tok.getIdentifierInfo() == &Ident_in) 487 Tok->Type = TT_ObjCForIn; 488 break; 489 case tok::comma: 490 if (Contexts.back().FirstStartOfName) 491 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 492 if (Contexts.back().InCtorInitializer) 493 Tok->Type = TT_CtorInitializerComma; 494 if (Contexts.back().IsForEachMacro) 495 Contexts.back().IsExpression = true; 496 break; 497 default: 498 break; 499 } 500 return true; 501 } 502 503 void parseIncludeDirective() { 504 if (CurrentToken && CurrentToken->is(tok::less)) { 505 next(); 506 while (CurrentToken) { 507 if (CurrentToken->isNot(tok::comment) || CurrentToken->Next) 508 CurrentToken->Type = TT_ImplicitStringLiteral; 509 next(); 510 } 511 } else { 512 while (CurrentToken) { 513 if (CurrentToken->is(tok::string_literal)) 514 // Mark these string literals as "implicit" literals, too, so that 515 // they are not split or line-wrapped. 516 CurrentToken->Type = TT_ImplicitStringLiteral; 517 next(); 518 } 519 } 520 } 521 522 void parseWarningOrError() { 523 next(); 524 // We still want to format the whitespace left of the first token of the 525 // warning or error. 526 next(); 527 while (CurrentToken) { 528 CurrentToken->Type = TT_ImplicitStringLiteral; 529 next(); 530 } 531 } 532 533 void parsePragma() { 534 next(); // Consume "pragma". 535 if (CurrentToken && CurrentToken->TokenText == "mark") { 536 next(); // Consume "mark". 537 next(); // Consume first token (so we fix leading whitespace). 538 while (CurrentToken) { 539 CurrentToken->Type = TT_ImplicitStringLiteral; 540 next(); 541 } 542 } 543 } 544 545 void parsePreprocessorDirective() { 546 next(); 547 if (!CurrentToken) 548 return; 549 if (CurrentToken->Tok.is(tok::numeric_constant)) { 550 CurrentToken->SpacesRequiredBefore = 1; 551 return; 552 } 553 // Hashes in the middle of a line can lead to any strange token 554 // sequence. 555 if (!CurrentToken->Tok.getIdentifierInfo()) 556 return; 557 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 558 case tok::pp_include: 559 case tok::pp_import: 560 next(); 561 parseIncludeDirective(); 562 break; 563 case tok::pp_error: 564 case tok::pp_warning: 565 parseWarningOrError(); 566 break; 567 case tok::pp_pragma: 568 parsePragma(); 569 break; 570 case tok::pp_if: 571 case tok::pp_elif: 572 Contexts.back().IsExpression = true; 573 parseLine(); 574 break; 575 default: 576 break; 577 } 578 while (CurrentToken) 579 next(); 580 } 581 582 public: 583 LineType parseLine() { 584 if (CurrentToken->is(tok::hash)) { 585 parsePreprocessorDirective(); 586 return LT_PreprocessorDirective; 587 } 588 589 // Directly allow to 'import <string-literal>' to support protocol buffer 590 // definitions (code.google.com/p/protobuf) or missing "#" (either way we 591 // should not break the line). 592 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 593 if (Info && Info->getPPKeywordID() == tok::pp_import && 594 CurrentToken->Next && CurrentToken->Next->is(tok::string_literal)) { 595 next(); 596 parseIncludeDirective(); 597 return LT_Other; 598 } 599 600 // If this line starts and ends in '<' and '>', respectively, it is likely 601 // part of "#define <a/b.h>". 602 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 603 parseIncludeDirective(); 604 return LT_Other; 605 } 606 607 while (CurrentToken) { 608 if (CurrentToken->is(tok::kw_virtual)) 609 KeywordVirtualFound = true; 610 if (!consumeToken()) 611 return LT_Invalid; 612 } 613 if (KeywordVirtualFound) 614 return LT_VirtualFunctionDecl; 615 616 if (Line.First->Type == TT_ObjCMethodSpecifier) { 617 if (Contexts.back().FirstObjCSelectorName) 618 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 619 Contexts.back().LongestObjCSelectorName; 620 return LT_ObjCMethodDecl; 621 } 622 623 return LT_Other; 624 } 625 626 private: 627 void resetTokenMetadata(FormatToken *Token) { 628 if (!Token) 629 return; 630 631 // Reset token type in case we have already looked at it and then 632 // recovered from an error (e.g. failure to find the matching >). 633 if (CurrentToken->Type != TT_LambdaLSquare && 634 CurrentToken->Type != TT_FunctionLBrace && 635 CurrentToken->Type != TT_ImplicitStringLiteral && 636 CurrentToken->Type != TT_RegexLiteral && 637 CurrentToken->Type != TT_TrailingReturnArrow) 638 CurrentToken->Type = TT_Unknown; 639 CurrentToken->Role.reset(); 640 CurrentToken->FakeLParens.clear(); 641 CurrentToken->FakeRParens = 0; 642 } 643 644 void next() { 645 if (CurrentToken) { 646 CurrentToken->NestingLevel = Contexts.size() - 1; 647 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 648 determineTokenType(*CurrentToken); 649 CurrentToken = CurrentToken->Next; 650 } 651 652 resetTokenMetadata(CurrentToken); 653 } 654 655 /// \brief A struct to hold information valid in a specific context, e.g. 656 /// a pair of parenthesis. 657 struct Context { 658 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 659 bool IsExpression) 660 : ContextKind(ContextKind), BindingStrength(BindingStrength), 661 LongestObjCSelectorName(0), ColonIsForRangeExpr(false), 662 ColonIsDictLiteral(false), ColonIsObjCMethodExpr(false), 663 FirstObjCSelectorName(nullptr), FirstStartOfName(nullptr), 664 IsExpression(IsExpression), CanBeExpression(true), 665 InTemplateArgument(false), InCtorInitializer(false), 666 CaretFound(false), IsForEachMacro(false) {} 667 668 tok::TokenKind ContextKind; 669 unsigned BindingStrength; 670 unsigned LongestObjCSelectorName; 671 bool ColonIsForRangeExpr; 672 bool ColonIsDictLiteral; 673 bool ColonIsObjCMethodExpr; 674 FormatToken *FirstObjCSelectorName; 675 FormatToken *FirstStartOfName; 676 bool IsExpression; 677 bool CanBeExpression; 678 bool InTemplateArgument; 679 bool InCtorInitializer; 680 bool CaretFound; 681 bool IsForEachMacro; 682 }; 683 684 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime 685 /// of each instance. 686 struct ScopedContextCreator { 687 AnnotatingParser &P; 688 689 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 690 unsigned Increase) 691 : P(P) { 692 P.Contexts.push_back(Context(ContextKind, 693 P.Contexts.back().BindingStrength + Increase, 694 P.Contexts.back().IsExpression)); 695 } 696 697 ~ScopedContextCreator() { P.Contexts.pop_back(); } 698 }; 699 700 void determineTokenType(FormatToken &Current) { 701 if (Current.getPrecedence() == prec::Assignment && 702 !Line.First->isOneOf(tok::kw_template, tok::kw_using) && 703 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 704 Contexts.back().IsExpression = true; 705 for (FormatToken *Previous = Current.Previous; 706 Previous && !Previous->isOneOf(tok::comma, tok::semi); 707 Previous = Previous->Previous) { 708 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 709 Previous = Previous->MatchingParen; 710 if (!Previous) 711 break; 712 } 713 if ((Previous->Type == TT_BinaryOperator || 714 Previous->Type == TT_UnaryOperator) && 715 Previous->isOneOf(tok::star, tok::amp) && Previous->Previous && 716 Previous->Previous->isNot(tok::equal)) { 717 Previous->Type = TT_PointerOrReference; 718 } 719 } 720 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 721 Contexts.back().IsExpression = true; 722 } else if (Current.is(tok::l_paren) && !Line.MustBeDeclaration && 723 !Line.InPPDirective && 724 (!Current.Previous || 725 Current.Previous->isNot(tok::kw_decltype))) { 726 bool ParametersOfFunctionType = 727 Current.Previous && Current.Previous->is(tok::r_paren) && 728 Current.Previous->MatchingParen && 729 Current.Previous->MatchingParen->Type == TT_FunctionTypeLParen; 730 bool IsForOrCatch = Current.Previous && 731 Current.Previous->isOneOf(tok::kw_for, tok::kw_catch); 732 Contexts.back().IsExpression = !ParametersOfFunctionType && !IsForOrCatch; 733 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 734 for (FormatToken *Previous = Current.Previous; 735 Previous && Previous->isOneOf(tok::star, tok::amp); 736 Previous = Previous->Previous) 737 Previous->Type = TT_PointerOrReference; 738 } else if (Current.Previous && 739 Current.Previous->Type == TT_CtorInitializerColon) { 740 Contexts.back().IsExpression = true; 741 Contexts.back().InCtorInitializer = true; 742 } else if (Current.is(tok::kw_new)) { 743 Contexts.back().CanBeExpression = false; 744 } else if (Current.is(tok::semi) || Current.is(tok::exclaim)) { 745 // This should be the condition or increment in a for-loop. 746 Contexts.back().IsExpression = true; 747 } 748 749 if (Current.Type == TT_Unknown) { 750 // Line.MightBeFunctionDecl can only be true after the parentheses of a 751 // function declaration have been found. In this case, 'Current' is a 752 // trailing token of this declaration and thus cannot be a name. 753 if (isStartOfName(Current) && 754 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 755 Contexts.back().FirstStartOfName = &Current; 756 Current.Type = TT_StartOfName; 757 } else if (Current.is(tok::kw_auto)) { 758 AutoFound = true; 759 } else if (Current.is(tok::arrow) && AutoFound && 760 Line.MustBeDeclaration) { 761 Current.Type = TT_TrailingReturnArrow; 762 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 763 Current.Type = 764 determineStarAmpUsage(Current, Contexts.back().CanBeExpression && 765 Contexts.back().IsExpression, 766 Contexts.back().InTemplateArgument); 767 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 768 Current.Type = determinePlusMinusCaretUsage(Current); 769 if (Current.Type == TT_UnaryOperator && Current.is(tok::caret)) 770 Contexts.back().CaretFound = true; 771 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 772 Current.Type = determineIncrementUsage(Current); 773 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 774 Current.Type = TT_UnaryOperator; 775 } else if (Current.is(tok::question)) { 776 Current.Type = TT_ConditionalExpr; 777 } else if (Current.isBinaryOperator() && 778 (!Current.Previous || 779 Current.Previous->isNot(tok::l_square))) { 780 Current.Type = TT_BinaryOperator; 781 } else if (Current.is(tok::comment)) { 782 if (Current.TokenText.startswith("//")) 783 Current.Type = TT_LineComment; 784 else 785 Current.Type = TT_BlockComment; 786 } else if (Current.is(tok::r_paren)) { 787 if (rParenEndsCast(Current)) 788 Current.Type = TT_CastRParen; 789 } else if (Current.is(tok::at) && Current.Next) { 790 switch (Current.Next->Tok.getObjCKeywordID()) { 791 case tok::objc_interface: 792 case tok::objc_implementation: 793 case tok::objc_protocol: 794 Current.Type = TT_ObjCDecl; 795 break; 796 case tok::objc_property: 797 Current.Type = TT_ObjCProperty; 798 break; 799 default: 800 break; 801 } 802 } else if (Current.is(tok::period)) { 803 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 804 if (PreviousNoComment && 805 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 806 Current.Type = TT_DesignatedInitializerPeriod; 807 } else if (Current.isOneOf(tok::identifier, tok::kw_const) && 808 Current.Previous && Current.Previous->isNot(tok::equal) && 809 Line.MightBeFunctionDecl && Contexts.size() == 1) { 810 // Line.MightBeFunctionDecl can only be true after the parentheses of a 811 // function declaration have been found. 812 Current.Type = TT_TrailingAnnotation; 813 } 814 } 815 } 816 817 /// \brief Take a guess at whether \p Tok starts a name of a function or 818 /// variable declaration. 819 /// 820 /// This is a heuristic based on whether \p Tok is an identifier following 821 /// something that is likely a type. 822 bool isStartOfName(const FormatToken &Tok) { 823 if (Tok.isNot(tok::identifier) || !Tok.Previous) 824 return false; 825 826 // Skip "const" as it does not have an influence on whether this is a name. 827 FormatToken *PreviousNotConst = Tok.Previous; 828 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 829 PreviousNotConst = PreviousNotConst->Previous; 830 831 if (!PreviousNotConst) 832 return false; 833 834 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 835 PreviousNotConst->Previous && 836 PreviousNotConst->Previous->is(tok::hash); 837 838 if (PreviousNotConst->Type == TT_TemplateCloser) 839 return PreviousNotConst && PreviousNotConst->MatchingParen && 840 PreviousNotConst->MatchingParen->Previous && 841 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 842 843 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen && 844 PreviousNotConst->MatchingParen->Previous && 845 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype)) 846 return true; 847 848 return (!IsPPKeyword && PreviousNotConst->is(tok::identifier)) || 849 PreviousNotConst->Type == TT_PointerOrReference || 850 PreviousNotConst->isSimpleTypeSpecifier(); 851 } 852 853 /// \brief Determine whether ')' is ending a cast. 854 bool rParenEndsCast(const FormatToken &Tok) { 855 FormatToken *LeftOfParens = nullptr; 856 if (Tok.MatchingParen) 857 LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 858 if (LeftOfParens && LeftOfParens->is(tok::r_paren) && 859 LeftOfParens->MatchingParen) 860 LeftOfParens = LeftOfParens->MatchingParen->Previous; 861 if (LeftOfParens && LeftOfParens->is(tok::r_square) && 862 LeftOfParens->MatchingParen && 863 LeftOfParens->MatchingParen->Type == TT_LambdaLSquare) 864 return false; 865 bool IsCast = false; 866 bool ParensAreEmpty = Tok.Previous == Tok.MatchingParen; 867 bool ParensAreType = !Tok.Previous || 868 Tok.Previous->Type == TT_PointerOrReference || 869 Tok.Previous->Type == TT_TemplateCloser || 870 Tok.Previous->isSimpleTypeSpecifier(); 871 if (Style.Language == FormatStyle::LK_JavaScript && Tok.Next && 872 Tok.Next->TokenText == "in") 873 return false; 874 bool ParensCouldEndDecl = 875 Tok.Next && Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace); 876 bool IsSizeOfOrAlignOf = 877 LeftOfParens && LeftOfParens->isOneOf(tok::kw_sizeof, tok::kw_alignof); 878 if (ParensAreType && !ParensCouldEndDecl && !IsSizeOfOrAlignOf && 879 ((Contexts.size() > 1 && Contexts[Contexts.size() - 2].IsExpression) || 880 (Tok.Next && Tok.Next->isBinaryOperator()))) 881 IsCast = true; 882 else if (Tok.Next && Tok.Next->isNot(tok::string_literal) && 883 (Tok.Next->Tok.isLiteral() || 884 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 885 IsCast = true; 886 // If there is an identifier after the (), it is likely a cast, unless 887 // there is also an identifier before the (). 888 else if (LeftOfParens && 889 (LeftOfParens->Tok.getIdentifierInfo() == nullptr || 890 LeftOfParens->is(tok::kw_return)) && 891 LeftOfParens->Type != TT_OverloadedOperator && 892 LeftOfParens->isNot(tok::at) && 893 LeftOfParens->Type != TT_TemplateCloser && Tok.Next) { 894 if (Tok.Next->isOneOf(tok::identifier, tok::numeric_constant)) { 895 IsCast = true; 896 } else { 897 // Use heuristics to recognize c style casting. 898 FormatToken *Prev = Tok.Previous; 899 if (Prev && Prev->isOneOf(tok::amp, tok::star)) 900 Prev = Prev->Previous; 901 902 if (Prev && Tok.Next && Tok.Next->Next) { 903 bool NextIsUnary = Tok.Next->isUnaryOperator() || 904 Tok.Next->isOneOf(tok::amp, tok::star); 905 IsCast = 906 NextIsUnary && !Tok.Next->is(tok::plus) && 907 Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant); 908 } 909 910 for (; Prev != Tok.MatchingParen; Prev = Prev->Previous) { 911 if (!Prev || !Prev->isOneOf(tok::kw_const, tok::identifier)) { 912 IsCast = false; 913 break; 914 } 915 } 916 } 917 } 918 return IsCast && !ParensAreEmpty; 919 } 920 921 /// \brief Return the type of the given token assuming it is * or &. 922 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 923 bool InTemplateArgument) { 924 if (Style.Language == FormatStyle::LK_JavaScript) 925 return TT_BinaryOperator; 926 927 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 928 if (!PrevToken) 929 return TT_UnaryOperator; 930 931 const FormatToken *NextToken = Tok.getNextNonComment(); 932 if (!NextToken || NextToken->is(tok::l_brace)) 933 return TT_Unknown; 934 935 if (PrevToken->is(tok::coloncolon)) 936 return TT_PointerOrReference; 937 938 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 939 tok::comma, tok::semi, tok::kw_return, tok::colon, 940 tok::equal, tok::kw_delete, tok::kw_sizeof) || 941 PrevToken->Type == TT_BinaryOperator || 942 PrevToken->Type == TT_ConditionalExpr || 943 PrevToken->Type == TT_UnaryOperator || PrevToken->Type == TT_CastRParen) 944 return TT_UnaryOperator; 945 946 if (NextToken->is(tok::l_square) && NextToken->Type != TT_LambdaLSquare) 947 return TT_PointerOrReference; 948 if (NextToken->isOneOf(tok::kw_operator, tok::comma)) 949 return TT_PointerOrReference; 950 951 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen && 952 PrevToken->MatchingParen->Previous && 953 PrevToken->MatchingParen->Previous->isOneOf(tok::kw_typeof, 954 tok::kw_decltype)) 955 return TT_PointerOrReference; 956 957 if (PrevToken->Tok.isLiteral() || 958 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 959 tok::kw_false) || 960 NextToken->Tok.isLiteral() || 961 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 962 NextToken->isUnaryOperator() || 963 // If we know we're in a template argument, there are no named 964 // declarations. Thus, having an identifier on the right-hand side 965 // indicates a binary operator. 966 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 967 return TT_BinaryOperator; 968 969 // This catches some cases where evaluation order is used as control flow: 970 // aaa && aaa->f(); 971 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 972 if (NextNextToken && NextNextToken->is(tok::arrow)) 973 return TT_BinaryOperator; 974 975 // It is very unlikely that we are going to find a pointer or reference type 976 // definition on the RHS of an assignment. 977 if (IsExpression) 978 return TT_BinaryOperator; 979 980 return TT_PointerOrReference; 981 } 982 983 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 984 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 985 if (!PrevToken || PrevToken->Type == TT_CastRParen) 986 return TT_UnaryOperator; 987 988 // Use heuristics to recognize unary operators. 989 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 990 tok::question, tok::colon, tok::kw_return, 991 tok::kw_case, tok::at, tok::l_brace)) 992 return TT_UnaryOperator; 993 994 // There can't be two consecutive binary operators. 995 if (PrevToken->Type == TT_BinaryOperator) 996 return TT_UnaryOperator; 997 998 // Fall back to marking the token as binary operator. 999 return TT_BinaryOperator; 1000 } 1001 1002 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements. 1003 TokenType determineIncrementUsage(const FormatToken &Tok) { 1004 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1005 if (!PrevToken || PrevToken->Type == TT_CastRParen) 1006 return TT_UnaryOperator; 1007 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 1008 return TT_TrailingUnaryOperator; 1009 1010 return TT_UnaryOperator; 1011 } 1012 1013 SmallVector<Context, 8> Contexts; 1014 1015 const FormatStyle &Style; 1016 AnnotatedLine &Line; 1017 FormatToken *CurrentToken; 1018 bool KeywordVirtualFound; 1019 bool AutoFound; 1020 IdentifierInfo &Ident_in; 1021 }; 1022 1023 static int PrecedenceUnaryOperator = prec::PointerToMember + 1; 1024 static int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 1025 1026 /// \brief Parses binary expressions by inserting fake parenthesis based on 1027 /// operator precedence. 1028 class ExpressionParser { 1029 public: 1030 ExpressionParser(AnnotatedLine &Line) : Current(Line.First) { 1031 // Skip leading "}", e.g. in "} else if (...) {". 1032 if (Current->is(tok::r_brace)) 1033 next(); 1034 } 1035 1036 /// \brief Parse expressions with the given operatore precedence. 1037 void parse(int Precedence = 0) { 1038 // Skip 'return' and ObjC selector colons as they are not part of a binary 1039 // expression. 1040 while (Current && 1041 (Current->is(tok::kw_return) || 1042 (Current->is(tok::colon) && (Current->Type == TT_ObjCMethodExpr || 1043 Current->Type == TT_DictLiteral)))) 1044 next(); 1045 1046 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1047 return; 1048 1049 // Conditional expressions need to be parsed separately for proper nesting. 1050 if (Precedence == prec::Conditional) { 1051 parseConditionalExpr(); 1052 return; 1053 } 1054 1055 // Parse unary operators, which all have a higher precedence than binary 1056 // operators. 1057 if (Precedence == PrecedenceUnaryOperator) { 1058 parseUnaryOperator(); 1059 return; 1060 } 1061 1062 FormatToken *Start = Current; 1063 FormatToken *LatestOperator = nullptr; 1064 unsigned OperatorIndex = 0; 1065 1066 while (Current) { 1067 // Consume operators with higher precedence. 1068 parse(Precedence + 1); 1069 1070 int CurrentPrecedence = getCurrentPrecedence(); 1071 1072 if (Current && Current->Type == TT_SelectorName && 1073 Precedence == CurrentPrecedence) { 1074 if (LatestOperator) 1075 addFakeParenthesis(Start, prec::Level(Precedence)); 1076 Start = Current; 1077 } 1078 1079 // At the end of the line or when an operator with higher precedence is 1080 // found, insert fake parenthesis and return. 1081 if (!Current || Current->closesScope() || 1082 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence)) { 1083 if (LatestOperator) { 1084 LatestOperator->LastOperator = true; 1085 if (Precedence == PrecedenceArrowAndPeriod) { 1086 // Call expressions don't have a binary operator precedence. 1087 addFakeParenthesis(Start, prec::Unknown); 1088 } else { 1089 addFakeParenthesis(Start, prec::Level(Precedence)); 1090 } 1091 } 1092 return; 1093 } 1094 1095 // Consume scopes: (), [], <> and {} 1096 if (Current->opensScope()) { 1097 while (Current && !Current->closesScope()) { 1098 next(); 1099 parse(); 1100 } 1101 next(); 1102 } else { 1103 // Operator found. 1104 if (CurrentPrecedence == Precedence) { 1105 LatestOperator = Current; 1106 Current->OperatorIndex = OperatorIndex; 1107 ++OperatorIndex; 1108 } 1109 1110 next(/*SkipPastLeadingComments=*/false); 1111 } 1112 } 1113 } 1114 1115 private: 1116 /// \brief Gets the precedence (+1) of the given token for binary operators 1117 /// and other tokens that we treat like binary operators. 1118 int getCurrentPrecedence() { 1119 if (Current) { 1120 const FormatToken *NextNonComment = Current->getNextNonComment(); 1121 if (Current->Type == TT_ConditionalExpr) 1122 return prec::Conditional; 1123 else if (NextNonComment && NextNonComment->is(tok::colon) && 1124 NextNonComment->Type == TT_DictLiteral) 1125 return prec::Comma; 1126 else if (Current->is(tok::semi) || Current->Type == TT_InlineASMColon || 1127 Current->Type == TT_SelectorName || 1128 (Current->is(tok::comment) && NextNonComment && 1129 NextNonComment->Type == TT_SelectorName)) 1130 return 0; 1131 else if (Current->Type == TT_RangeBasedForLoopColon) 1132 return prec::Comma; 1133 else if (Current->Type == TT_BinaryOperator || Current->is(tok::comma)) 1134 return Current->getPrecedence(); 1135 else if (Current->isOneOf(tok::period, tok::arrow)) 1136 return PrecedenceArrowAndPeriod; 1137 } 1138 return -1; 1139 } 1140 1141 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1142 Start->FakeLParens.push_back(Precedence); 1143 if (Precedence > prec::Unknown) 1144 Start->StartsBinaryExpression = true; 1145 if (Current) { 1146 ++Current->Previous->FakeRParens; 1147 if (Precedence > prec::Unknown) 1148 Current->Previous->EndsBinaryExpression = true; 1149 } 1150 } 1151 1152 /// \brief Parse unary operator expressions and surround them with fake 1153 /// parentheses if appropriate. 1154 void parseUnaryOperator() { 1155 if (!Current || Current->Type != TT_UnaryOperator) { 1156 parse(PrecedenceArrowAndPeriod); 1157 return; 1158 } 1159 1160 FormatToken *Start = Current; 1161 next(); 1162 parseUnaryOperator(); 1163 1164 // The actual precedence doesn't matter. 1165 addFakeParenthesis(Start, prec::Unknown); 1166 } 1167 1168 void parseConditionalExpr() { 1169 FormatToken *Start = Current; 1170 parse(prec::LogicalOr); 1171 if (!Current || !Current->is(tok::question)) 1172 return; 1173 next(); 1174 parseConditionalExpr(); 1175 if (!Current || Current->Type != TT_ConditionalExpr) 1176 return; 1177 next(); 1178 parseConditionalExpr(); 1179 addFakeParenthesis(Start, prec::Conditional); 1180 } 1181 1182 void next(bool SkipPastLeadingComments = true) { 1183 if (Current) 1184 Current = Current->Next; 1185 while (Current && 1186 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 1187 Current->isTrailingComment()) 1188 Current = Current->Next; 1189 } 1190 1191 FormatToken *Current; 1192 }; 1193 1194 } // end anonymous namespace 1195 1196 void 1197 TokenAnnotator::setCommentLineLevels(SmallVectorImpl<AnnotatedLine *> &Lines) { 1198 const AnnotatedLine *NextNonCommentLine = nullptr; 1199 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 1200 E = Lines.rend(); 1201 I != E; ++I) { 1202 if (NextNonCommentLine && (*I)->First->is(tok::comment) && 1203 (*I)->First->Next == nullptr) 1204 (*I)->Level = NextNonCommentLine->Level; 1205 else 1206 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 1207 1208 setCommentLineLevels((*I)->Children); 1209 } 1210 } 1211 1212 void TokenAnnotator::annotate(AnnotatedLine &Line) { 1213 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1214 E = Line.Children.end(); 1215 I != E; ++I) { 1216 annotate(**I); 1217 } 1218 AnnotatingParser Parser(Style, Line, Ident_in); 1219 Line.Type = Parser.parseLine(); 1220 if (Line.Type == LT_Invalid) 1221 return; 1222 1223 ExpressionParser ExprParser(Line); 1224 ExprParser.parse(); 1225 1226 if (Line.First->Type == TT_ObjCMethodSpecifier) 1227 Line.Type = LT_ObjCMethodDecl; 1228 else if (Line.First->Type == TT_ObjCDecl) 1229 Line.Type = LT_ObjCDecl; 1230 else if (Line.First->Type == TT_ObjCProperty) 1231 Line.Type = LT_ObjCProperty; 1232 1233 Line.First->SpacesRequiredBefore = 1; 1234 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 1235 } 1236 1237 // This function heuristically determines whether 'Current' starts the name of a 1238 // function declaration. 1239 static bool isFunctionDeclarationName(const FormatToken &Current) { 1240 if (Current.Type != TT_StartOfName || 1241 Current.NestingLevel != 0 || 1242 Current.Previous->Type == TT_StartOfName) 1243 return false; 1244 const FormatToken *Next = Current.Next; 1245 for (; Next; Next = Next->Next) { 1246 if (Next->Type == TT_TemplateOpener) { 1247 Next = Next->MatchingParen; 1248 } else if (Next->is(tok::coloncolon)) { 1249 Next = Next->Next; 1250 if (!Next || !Next->is(tok::identifier)) 1251 return false; 1252 } else if (Next->is(tok::l_paren)) { 1253 break; 1254 } else { 1255 return false; 1256 } 1257 } 1258 if (!Next) 1259 return false; 1260 assert(Next->is(tok::l_paren)); 1261 if (Next->Next == Next->MatchingParen) 1262 return true; 1263 for (const FormatToken *Tok = Next->Next; Tok != Next->MatchingParen; 1264 Tok = Tok->Next) { 1265 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 1266 Tok->Type == TT_PointerOrReference || Tok->Type == TT_StartOfName) 1267 return true; 1268 if (Tok->isOneOf(tok::l_brace, tok::string_literal) || Tok->Tok.isLiteral()) 1269 return false; 1270 } 1271 return false; 1272 } 1273 1274 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 1275 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1276 E = Line.Children.end(); 1277 I != E; ++I) { 1278 calculateFormattingInformation(**I); 1279 } 1280 1281 Line.First->TotalLength = 1282 Line.First->IsMultiline ? Style.ColumnLimit : Line.First->ColumnWidth; 1283 if (!Line.First->Next) 1284 return; 1285 FormatToken *Current = Line.First->Next; 1286 bool InFunctionDecl = Line.MightBeFunctionDecl; 1287 while (Current) { 1288 if (isFunctionDeclarationName(*Current)) 1289 Current->Type = TT_FunctionDeclarationName; 1290 if (Current->Type == TT_LineComment) { 1291 if (Current->Previous->BlockKind == BK_BracedInit && 1292 Current->Previous->opensScope()) 1293 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1; 1294 else 1295 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 1296 1297 // If we find a trailing comment, iterate backwards to determine whether 1298 // it seems to relate to a specific parameter. If so, break before that 1299 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 1300 // to the previous line in: 1301 // SomeFunction(a, 1302 // b, // comment 1303 // c); 1304 if (!Current->HasUnescapedNewline) { 1305 for (FormatToken *Parameter = Current->Previous; Parameter; 1306 Parameter = Parameter->Previous) { 1307 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 1308 break; 1309 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 1310 if (Parameter->Previous->Type != TT_CtorInitializerComma && 1311 Parameter->HasUnescapedNewline) 1312 Parameter->MustBreakBefore = true; 1313 break; 1314 } 1315 } 1316 } 1317 } else if (Current->SpacesRequiredBefore == 0 && 1318 spaceRequiredBefore(Line, *Current)) { 1319 Current->SpacesRequiredBefore = 1; 1320 } 1321 1322 Current->MustBreakBefore = 1323 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 1324 1325 if (Style.AlwaysBreakAfterDefinitionReturnType && 1326 InFunctionDecl && Current->Type == TT_FunctionDeclarationName && 1327 !Line.Last->isOneOf(tok::semi, tok::comment)) // Only for definitions. 1328 // FIXME: Line.Last points to other characters than tok::semi 1329 // and tok::lbrace. 1330 Current->MustBreakBefore = true; 1331 1332 Current->CanBreakBefore = 1333 Current->MustBreakBefore || canBreakBefore(Line, *Current); 1334 unsigned ChildSize = 0; 1335 if (Current->Previous->Children.size() == 1) { 1336 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 1337 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 1338 : LastOfChild.TotalLength + 1; 1339 } 1340 if (Current->MustBreakBefore || Current->Previous->Children.size() > 1 || 1341 Current->IsMultiline) 1342 Current->TotalLength = Current->Previous->TotalLength + Style.ColumnLimit; 1343 else 1344 Current->TotalLength = Current->Previous->TotalLength + 1345 Current->ColumnWidth + ChildSize + 1346 Current->SpacesRequiredBefore; 1347 1348 if (Current->Type == TT_CtorInitializerColon) 1349 InFunctionDecl = false; 1350 1351 // FIXME: Only calculate this if CanBreakBefore is true once static 1352 // initializers etc. are sorted out. 1353 // FIXME: Move magic numbers to a better place. 1354 Current->SplitPenalty = 20 * Current->BindingStrength + 1355 splitPenalty(Line, *Current, InFunctionDecl); 1356 1357 Current = Current->Next; 1358 } 1359 1360 calculateUnbreakableTailLengths(Line); 1361 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 1362 if (Current->Role) 1363 Current->Role->precomputeFormattingInfos(Current); 1364 } 1365 1366 DEBUG({ printDebugInfo(Line); }); 1367 } 1368 1369 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 1370 unsigned UnbreakableTailLength = 0; 1371 FormatToken *Current = Line.Last; 1372 while (Current) { 1373 Current->UnbreakableTailLength = UnbreakableTailLength; 1374 if (Current->CanBreakBefore || 1375 Current->isOneOf(tok::comment, tok::string_literal)) { 1376 UnbreakableTailLength = 0; 1377 } else { 1378 UnbreakableTailLength += 1379 Current->ColumnWidth + Current->SpacesRequiredBefore; 1380 } 1381 Current = Current->Previous; 1382 } 1383 } 1384 1385 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 1386 const FormatToken &Tok, 1387 bool InFunctionDecl) { 1388 const FormatToken &Left = *Tok.Previous; 1389 const FormatToken &Right = Tok; 1390 1391 if (Left.is(tok::semi)) 1392 return 0; 1393 if (Left.is(tok::comma) || (Right.is(tok::identifier) && Right.Next && 1394 Right.Next->Type == TT_DictLiteral)) 1395 return 1; 1396 if (Right.is(tok::l_square)) { 1397 if (Style.Language == FormatStyle::LK_Proto) 1398 return 1; 1399 if (Right.Type != TT_ObjCMethodExpr && Right.Type != TT_LambdaLSquare) 1400 return 500; 1401 } 1402 if (Right.Type == TT_StartOfName || 1403 Right.Type == TT_FunctionDeclarationName || Right.is(tok::kw_operator)) { 1404 if (Line.First->is(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 1405 return 3; 1406 if (Left.Type == TT_StartOfName) 1407 return 20; 1408 if (InFunctionDecl && Right.NestingLevel == 0) 1409 return Style.PenaltyReturnTypeOnItsOwnLine; 1410 return 200; 1411 } 1412 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 1413 return 150; 1414 if (Left.Type == TT_CastRParen) 1415 return 100; 1416 if (Left.is(tok::coloncolon) || 1417 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 1418 return 500; 1419 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 1420 return 5000; 1421 1422 if (Left.Type == TT_RangeBasedForLoopColon || 1423 Left.Type == TT_InheritanceColon) 1424 return 2; 1425 1426 if (Right.isMemberAccess()) { 1427 if (Left.is(tok::r_paren) && Left.MatchingParen && 1428 Left.MatchingParen->ParameterCount > 0) 1429 return 20; // Should be smaller than breaking at a nested comma. 1430 return 150; 1431 } 1432 1433 if (Right.Type == TT_TrailingAnnotation && 1434 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 1435 // Generally, breaking before a trailing annotation is bad unless it is 1436 // function-like. It seems to be especially preferable to keep standard 1437 // annotations (i.e. "const", "final" and "override") on the same line. 1438 // Use a slightly higher penalty after ")" so that annotations like 1439 // "const override" are kept together. 1440 bool is_short_annotation = Right.TokenText.size() < 10; 1441 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 1442 } 1443 1444 // In for-loops, prefer breaking at ',' and ';'. 1445 if (Line.First->is(tok::kw_for) && Left.is(tok::equal)) 1446 return 4; 1447 1448 // In Objective-C method expressions, prefer breaking before "param:" over 1449 // breaking after it. 1450 if (Right.Type == TT_SelectorName) 1451 return 0; 1452 if (Left.is(tok::colon) && Left.Type == TT_ObjCMethodExpr) 1453 return Line.MightBeFunctionDecl ? 50 : 500; 1454 1455 if (Left.is(tok::l_paren) && InFunctionDecl) 1456 return 100; 1457 if (Left.is(tok::equal) && InFunctionDecl) 1458 return 110; 1459 if (Right.is(tok::r_brace)) 1460 return 1; 1461 if (Left.Type == TT_TemplateOpener) 1462 return 100; 1463 if (Left.opensScope()) 1464 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 1465 : 19; 1466 1467 if (Right.is(tok::lessless)) { 1468 if (Left.is(tok::string_literal)) { 1469 StringRef Content = Left.TokenText; 1470 if (Content.startswith("\"")) 1471 Content = Content.drop_front(1); 1472 if (Content.endswith("\"")) 1473 Content = Content.drop_back(1); 1474 Content = Content.trim(); 1475 if (Content.size() > 1 && 1476 (Content.back() == ':' || Content.back() == '=')) 1477 return 25; 1478 } 1479 return 1; // Breaking at a << is really cheap. 1480 } 1481 if (Left.Type == TT_ConditionalExpr) 1482 return prec::Conditional; 1483 prec::Level Level = Left.getPrecedence(); 1484 1485 if (Level != prec::Unknown) 1486 return Level; 1487 1488 return 3; 1489 } 1490 1491 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 1492 const FormatToken &Left, 1493 const FormatToken &Right) { 1494 if (Style.Language == FormatStyle::LK_Proto) { 1495 if (Right.is(tok::period) && 1496 (Left.TokenText == "optional" || Left.TokenText == "required" || 1497 Left.TokenText == "repeated")) 1498 return true; 1499 if (Right.is(tok::l_paren) && 1500 (Left.TokenText == "returns" || Left.TokenText == "option")) 1501 return true; 1502 } else if (Style.Language == FormatStyle::LK_JavaScript) { 1503 if (Left.TokenText == "var") 1504 return true; 1505 } 1506 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 1507 return true; 1508 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 1509 Left.Tok.getObjCKeywordID() == tok::objc_property) 1510 return true; 1511 if (Right.is(tok::hashhash)) 1512 return Left.is(tok::hash); 1513 if (Left.isOneOf(tok::hashhash, tok::hash)) 1514 return Right.is(tok::hash); 1515 if (Left.is(tok::l_paren) && Right.is(tok::r_paren)) 1516 return Style.SpaceInEmptyParentheses; 1517 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 1518 return (Right.Type == TT_CastRParen || 1519 (Left.MatchingParen && Left.MatchingParen->Type == TT_CastRParen)) 1520 ? Style.SpacesInCStyleCastParentheses 1521 : Style.SpacesInParentheses; 1522 if (Style.SpacesInAngles && 1523 ((Left.Type == TT_TemplateOpener) != (Right.Type == TT_TemplateCloser))) 1524 return true; 1525 if (Right.isOneOf(tok::semi, tok::comma)) 1526 return false; 1527 if (Right.is(tok::less) && 1528 (Left.isOneOf(tok::kw_template, tok::r_paren) || 1529 (Line.Type == LT_ObjCDecl && Style.ObjCSpaceBeforeProtocolList))) 1530 return true; 1531 if (Left.is(tok::arrow) || Right.is(tok::arrow)) 1532 return false; 1533 if (Left.isOneOf(tok::exclaim, tok::tilde)) 1534 return false; 1535 if (Left.is(tok::at) && 1536 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 1537 tok::numeric_constant, tok::l_paren, tok::l_brace, 1538 tok::kw_true, tok::kw_false)) 1539 return false; 1540 if (Left.is(tok::coloncolon)) 1541 return false; 1542 if (Right.is(tok::coloncolon) && Left.isNot(tok::l_brace)) 1543 return (Left.is(tok::less) && Style.Standard == FormatStyle::LS_Cpp03) || 1544 !Left.isOneOf(tok::identifier, tok::greater, tok::l_paren, 1545 tok::r_paren, tok::less); 1546 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) 1547 return false; 1548 if (Right.is(tok::ellipsis)) 1549 return Left.Tok.isLiteral(); 1550 if (Left.is(tok::l_square) && Right.is(tok::amp)) 1551 return false; 1552 if (Right.Type == TT_PointerOrReference) 1553 return Left.Tok.isLiteral() || 1554 ((Left.Type != TT_PointerOrReference) && Left.isNot(tok::l_paren) && 1555 Style.PointerAlignment != FormatStyle::PAS_Left); 1556 if (Right.Type == TT_FunctionTypeLParen && Left.isNot(tok::l_paren) && 1557 (Left.Type != TT_PointerOrReference || 1558 Style.PointerAlignment != FormatStyle::PAS_Right)) 1559 return true; 1560 if (Left.Type == TT_PointerOrReference) 1561 return Right.Tok.isLiteral() || Right.Type == TT_BlockComment || 1562 ((Right.Type != TT_PointerOrReference) && 1563 Right.isNot(tok::l_paren) && 1564 Style.PointerAlignment != FormatStyle::PAS_Right && Left.Previous && 1565 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon)); 1566 if (Right.is(tok::star) && Left.is(tok::l_paren)) 1567 return false; 1568 if (Left.is(tok::l_square)) 1569 return (Left.Type == TT_ArrayInitializerLSquare && 1570 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square)) || 1571 (Left.Type == TT_ArraySubscriptLSquare && 1572 Style.SpacesInSquareBrackets && Right.isNot(tok::r_square)); 1573 if (Right.is(tok::r_square)) 1574 return Right.MatchingParen && 1575 ((Style.SpacesInContainerLiterals && 1576 Right.MatchingParen->Type == TT_ArrayInitializerLSquare) || 1577 (Style.SpacesInSquareBrackets && 1578 Right.MatchingParen->Type == TT_ArraySubscriptLSquare)); 1579 if (Right.is(tok::l_square) && Right.Type != TT_ObjCMethodExpr && 1580 Right.Type != TT_LambdaLSquare && Left.isNot(tok::numeric_constant) && 1581 Left.Type != TT_DictLiteral) 1582 return false; 1583 if (Left.is(tok::colon)) 1584 return Left.Type != TT_ObjCMethodExpr; 1585 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 1586 return !Left.Children.empty(); // No spaces in "{}". 1587 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 1588 (Right.is(tok::r_brace) && Right.MatchingParen && 1589 Right.MatchingParen->BlockKind != BK_Block)) 1590 return !Style.Cpp11BracedListStyle; 1591 if (Left.Type == TT_BlockComment) 1592 return !Left.TokenText.endswith("=*/"); 1593 if (Right.is(tok::l_paren)) { 1594 if (Left.is(tok::r_paren) && Left.Type == TT_AttributeParen) 1595 return true; 1596 return Line.Type == LT_ObjCDecl || 1597 Left.isOneOf(tok::kw_new, tok::kw_delete, tok::semi) || 1598 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 1599 (Left.isOneOf(tok::kw_if, tok::kw_for, tok::kw_while, 1600 tok::kw_switch, tok::kw_case) || 1601 (Left.is(tok::kw_catch) && 1602 (!Left.Previous || Left.Previous->isNot(tok::period))) || 1603 Left.IsForEachMacro)) || 1604 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always && 1605 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword()) && 1606 Line.Type != LT_PreprocessorDirective); 1607 } 1608 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 1609 return false; 1610 if (Right.Type == TT_UnaryOperator) 1611 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 1612 (Left.isNot(tok::colon) || Left.Type != TT_ObjCMethodExpr); 1613 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 1614 tok::r_paren) || 1615 Left.isSimpleTypeSpecifier()) && 1616 Right.is(tok::l_brace) && Right.getNextNonComment() && 1617 Right.BlockKind != BK_Block) 1618 return false; 1619 if (Left.is(tok::period) || Right.is(tok::period)) 1620 return false; 1621 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 1622 return false; 1623 return true; 1624 } 1625 1626 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 1627 const FormatToken &Tok) { 1628 if (Tok.Tok.getIdentifierInfo() && Tok.Previous->Tok.getIdentifierInfo()) 1629 return true; // Never ever merge two identifiers. 1630 if (Tok.Previous->Type == TT_ImplicitStringLiteral) 1631 return Tok.WhitespaceRange.getBegin() != Tok.WhitespaceRange.getEnd(); 1632 if (Line.Type == LT_ObjCMethodDecl) { 1633 if (Tok.Previous->Type == TT_ObjCMethodSpecifier) 1634 return true; 1635 if (Tok.Previous->is(tok::r_paren) && Tok.is(tok::identifier)) 1636 // Don't space between ')' and <id> 1637 return false; 1638 } 1639 if (Line.Type == LT_ObjCProperty && 1640 (Tok.is(tok::equal) || Tok.Previous->is(tok::equal))) 1641 return false; 1642 1643 if (Tok.Type == TT_TrailingReturnArrow || 1644 Tok.Previous->Type == TT_TrailingReturnArrow) 1645 return true; 1646 if (Tok.Previous->is(tok::comma)) 1647 return true; 1648 if (Tok.is(tok::comma)) 1649 return false; 1650 if (Tok.Type == TT_CtorInitializerColon || Tok.Type == TT_ObjCBlockLParen) 1651 return true; 1652 if (Tok.Previous->Tok.is(tok::kw_operator)) 1653 return Tok.is(tok::coloncolon); 1654 if (Tok.Type == TT_OverloadedOperatorLParen) 1655 return false; 1656 if (Tok.is(tok::colon)) 1657 return !Line.First->isOneOf(tok::kw_case, tok::kw_default) && 1658 Tok.getNextNonComment() && Tok.Type != TT_ObjCMethodExpr && 1659 !Tok.Previous->is(tok::question) && 1660 !(Tok.Type == TT_InlineASMColon && 1661 Tok.Previous->is(tok::coloncolon)) && 1662 (Tok.Type != TT_DictLiteral || Style.SpacesInContainerLiterals); 1663 if (Tok.Previous->Type == TT_UnaryOperator) 1664 return Tok.Type == TT_BinaryOperator; 1665 if (Tok.Previous->Type == TT_CastRParen) 1666 return Style.SpaceAfterCStyleCast || Tok.Type == TT_BinaryOperator; 1667 if (Tok.Previous->is(tok::greater) && Tok.is(tok::greater)) { 1668 return Tok.Type == TT_TemplateCloser && 1669 Tok.Previous->Type == TT_TemplateCloser && 1670 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 1671 } 1672 if (Tok.isOneOf(tok::arrowstar, tok::periodstar) || 1673 Tok.Previous->isOneOf(tok::arrowstar, tok::periodstar)) 1674 return false; 1675 if (!Style.SpaceBeforeAssignmentOperators && 1676 Tok.getPrecedence() == prec::Assignment) 1677 return false; 1678 if ((Tok.Type == TT_BinaryOperator && !Tok.Previous->is(tok::l_paren)) || 1679 Tok.Previous->Type == TT_BinaryOperator || 1680 Tok.Previous->Type == TT_ConditionalExpr) 1681 return true; 1682 if (Tok.Previous->Type == TT_TemplateCloser && Tok.is(tok::l_paren)) 1683 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 1684 if (Tok.is(tok::less) && Tok.Previous->isNot(tok::l_paren) && 1685 Line.First->is(tok::hash)) 1686 return true; 1687 if (Tok.Type == TT_TrailingUnaryOperator) 1688 return false; 1689 if (Tok.Previous->Type == TT_RegexLiteral) 1690 return false; 1691 return spaceRequiredBetween(Line, *Tok.Previous, Tok); 1692 } 1693 1694 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 1695 static bool isAllmanBrace(const FormatToken &Tok) { 1696 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 1697 Tok.Type != TT_ObjCBlockLBrace && Tok.Type != TT_DictLiteral; 1698 } 1699 1700 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 1701 const FormatToken &Right) { 1702 const FormatToken &Left = *Right.Previous; 1703 if (Right.NewlinesBefore > 1) 1704 return true; 1705 if (Right.is(tok::comment)) { 1706 return Right.Previous->BlockKind != BK_BracedInit && 1707 Right.Previous->Type != TT_CtorInitializerColon && 1708 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 1709 } else if (Right.Previous->isTrailingComment() || 1710 (Right.isStringLiteral() && Right.Previous->isStringLiteral())) { 1711 return true; 1712 } else if (Right.Previous->IsUnterminatedLiteral) { 1713 return true; 1714 } else if (Right.is(tok::lessless) && Right.Next && 1715 Right.Previous->is(tok::string_literal) && 1716 Right.Next->is(tok::string_literal)) { 1717 return true; 1718 } else if (Right.Previous->ClosesTemplateDeclaration && 1719 Right.Previous->MatchingParen && 1720 Right.Previous->MatchingParen->NestingLevel == 0 && 1721 Style.AlwaysBreakTemplateDeclarations) { 1722 return true; 1723 } else if ((Right.Type == TT_CtorInitializerComma || 1724 Right.Type == TT_CtorInitializerColon) && 1725 Style.BreakConstructorInitializersBeforeComma && 1726 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) { 1727 return true; 1728 } else if (Right.is(tok::string_literal) && 1729 Right.TokenText.startswith("R\"")) { 1730 // Raw string literals are special wrt. line breaks. The author has made a 1731 // deliberate choice and might have aligned the contents of the string 1732 // literal accordingly. Thus, we try keep existing line breaks. 1733 return Right.NewlinesBefore > 0; 1734 } else if (Right.Previous->is(tok::l_brace) && Right.NestingLevel == 1 && 1735 Style.Language == FormatStyle::LK_Proto) { 1736 // Don't enums onto single lines in protocol buffers. 1737 return true; 1738 } else if (Style.Language == FormatStyle::LK_JavaScript && 1739 Right.is(tok::r_brace) && Left.is(tok::l_brace) && 1740 !Left.Children.empty()) { 1741 // Support AllowShortFunctionsOnASingleLine for JavaScript. 1742 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 1743 (Left.NestingLevel == 0 && Line.Level == 0 && 1744 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Inline); 1745 } else if (isAllmanBrace(Left) || isAllmanBrace(Right)) { 1746 return Style.BreakBeforeBraces == FormatStyle::BS_Allman || 1747 Style.BreakBeforeBraces == FormatStyle::BS_GNU; 1748 } else if (Style.Language == FormatStyle::LK_Proto && 1749 Left.isNot(tok::l_brace) && Right.Type == TT_SelectorName) { 1750 return true; 1751 } 1752 1753 // If the last token before a '}' is a comma or a trailing comment, the 1754 // intention is to insert a line break after it in order to make shuffling 1755 // around entries easier. 1756 const FormatToken *BeforeClosingBrace = nullptr; 1757 if (Left.is(tok::l_brace) && Left.MatchingParen) 1758 BeforeClosingBrace = Left.MatchingParen->Previous; 1759 else if (Right.is(tok::r_brace)) 1760 BeforeClosingBrace = Right.Previous; 1761 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 1762 BeforeClosingBrace->isTrailingComment())) 1763 return true; 1764 1765 if (Style.Language == FormatStyle::LK_JavaScript) { 1766 // FIXME: This might apply to other languages and token kinds. 1767 if (Right.is(tok::char_constant) && Left.is(tok::plus) && Left.Previous && 1768 Left.Previous->is(tok::char_constant)) 1769 return true; 1770 } 1771 1772 return false; 1773 } 1774 1775 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 1776 const FormatToken &Right) { 1777 const FormatToken &Left = *Right.Previous; 1778 if (Left.is(tok::at)) 1779 return false; 1780 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 1781 return false; 1782 if (Right.Type == TT_StartOfName || 1783 Right.Type == TT_FunctionDeclarationName || Right.is(tok::kw_operator)) 1784 return true; 1785 if (Right.isTrailingComment()) 1786 // We rely on MustBreakBefore being set correctly here as we should not 1787 // change the "binding" behavior of a comment. 1788 // The first comment in a braced lists is always interpreted as belonging to 1789 // the first list element. Otherwise, it should be placed outside of the 1790 // list. 1791 return Left.BlockKind == BK_BracedInit; 1792 if (Left.is(tok::question) && Right.is(tok::colon)) 1793 return false; 1794 if (Right.Type == TT_ConditionalExpr || Right.is(tok::question)) 1795 return Style.BreakBeforeTernaryOperators; 1796 if (Left.Type == TT_ConditionalExpr || Left.is(tok::question)) 1797 return !Style.BreakBeforeTernaryOperators; 1798 if (Right.Type == TT_InheritanceColon) 1799 return true; 1800 if (Right.is(tok::colon) && (Right.Type != TT_CtorInitializerColon && 1801 Right.Type != TT_InlineASMColon)) 1802 return false; 1803 if (Left.is(tok::colon) && 1804 (Left.Type == TT_DictLiteral || Left.Type == TT_ObjCMethodExpr)) 1805 return true; 1806 if (Right.Type == TT_SelectorName) 1807 return true; 1808 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 1809 return true; 1810 if (Left.ClosesTemplateDeclaration) 1811 return true; 1812 if (Right.Type == TT_RangeBasedForLoopColon || 1813 Right.Type == TT_OverloadedOperatorLParen || 1814 Right.Type == TT_OverloadedOperator) 1815 return false; 1816 if (Left.Type == TT_RangeBasedForLoopColon) 1817 return true; 1818 if (Right.Type == TT_RangeBasedForLoopColon) 1819 return false; 1820 if (Left.Type == TT_PointerOrReference || Left.Type == TT_TemplateCloser || 1821 Left.Type == TT_UnaryOperator || Left.is(tok::kw_operator)) 1822 return false; 1823 if (Left.is(tok::equal) && Line.Type == LT_VirtualFunctionDecl) 1824 return false; 1825 if (Left.is(tok::l_paren) && Left.Type == TT_AttributeParen) 1826 return false; 1827 if (Left.is(tok::l_paren) && Left.Previous && 1828 (Left.Previous->Type == TT_BinaryOperator || 1829 Left.Previous->Type == TT_CastRParen || Left.Previous->is(tok::kw_if))) 1830 return false; 1831 if (Right.Type == TT_ImplicitStringLiteral) 1832 return false; 1833 1834 if (Right.is(tok::r_paren) || Right.Type == TT_TemplateCloser) 1835 return false; 1836 1837 // We only break before r_brace if there was a corresponding break before 1838 // the l_brace, which is tracked by BreakBeforeClosingBrace. 1839 if (Right.is(tok::r_brace)) 1840 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 1841 1842 // Allow breaking after a trailing annotation, e.g. after a method 1843 // declaration. 1844 if (Left.Type == TT_TrailingAnnotation) 1845 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 1846 tok::less, tok::coloncolon); 1847 1848 if (Right.is(tok::kw___attribute)) 1849 return true; 1850 1851 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 1852 return true; 1853 1854 if (Right.is(tok::identifier) && Right.Next && 1855 Right.Next->Type == TT_DictLiteral) 1856 return true; 1857 1858 if (Left.Type == TT_CtorInitializerComma && 1859 Style.BreakConstructorInitializersBeforeComma) 1860 return false; 1861 if (Right.Type == TT_CtorInitializerComma && 1862 Style.BreakConstructorInitializersBeforeComma) 1863 return true; 1864 if (Left.is(tok::greater) && Right.is(tok::greater) && 1865 Left.Type != TT_TemplateCloser) 1866 return false; 1867 if (Right.Type == TT_BinaryOperator && 1868 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 1869 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 1870 Right.getPrecedence() != prec::Assignment)) 1871 return true; 1872 if (Left.Type == TT_ArrayInitializerLSquare) 1873 return true; 1874 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 1875 return true; 1876 if (Left.isBinaryOperator() && !Left.isOneOf(tok::arrowstar, tok::lessless) && 1877 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 1878 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 1879 Left.getPrecedence() == prec::Assignment)) 1880 return true; 1881 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 1882 tok::kw_class, tok::kw_struct) || 1883 Right.isMemberAccess() || 1884 Right.isOneOf(tok::lessless, tok::colon, tok::l_square, tok::at) || 1885 (Left.is(tok::r_paren) && 1886 Right.isOneOf(tok::identifier, tok::kw_const)) || 1887 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)); 1888 } 1889 1890 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 1891 llvm::errs() << "AnnotatedTokens:\n"; 1892 const FormatToken *Tok = Line.First; 1893 while (Tok) { 1894 llvm::errs() << " M=" << Tok->MustBreakBefore 1895 << " C=" << Tok->CanBreakBefore << " T=" << Tok->Type 1896 << " S=" << Tok->SpacesRequiredBefore 1897 << " B=" << Tok->BlockParameterCount 1898 << " P=" << Tok->SplitPenalty << " Name=" << Tok->Tok.getName() 1899 << " L=" << Tok->TotalLength << " PPK=" << Tok->PackingKind 1900 << " FakeLParens="; 1901 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 1902 llvm::errs() << Tok->FakeLParens[i] << "/"; 1903 llvm::errs() << " FakeRParens=" << Tok->FakeRParens << "\n"; 1904 if (!Tok->Next) 1905 assert(Tok == Line.Last); 1906 Tok = Tok->Next; 1907 } 1908 llvm::errs() << "----\n"; 1909 } 1910 1911 } // namespace format 1912 } // namespace clang 1913