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 bool ParensCouldEndDecl = 872 Tok.Next && Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace); 873 bool IsSizeOfOrAlignOf = 874 LeftOfParens && LeftOfParens->isOneOf(tok::kw_sizeof, tok::kw_alignof); 875 if (ParensAreType && !ParensCouldEndDecl && !IsSizeOfOrAlignOf && 876 ((Contexts.size() > 1 && Contexts[Contexts.size() - 2].IsExpression) || 877 (Tok.Next && Tok.Next->isBinaryOperator()))) 878 IsCast = true; 879 else if (Tok.Next && Tok.Next->isNot(tok::string_literal) && 880 (Tok.Next->Tok.isLiteral() || 881 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 882 IsCast = true; 883 // If there is an identifier after the (), it is likely a cast, unless 884 // there is also an identifier before the (). 885 else if (LeftOfParens && 886 (LeftOfParens->Tok.getIdentifierInfo() == nullptr || 887 LeftOfParens->is(tok::kw_return)) && 888 LeftOfParens->Type != TT_OverloadedOperator && 889 LeftOfParens->isNot(tok::at) && 890 LeftOfParens->Type != TT_TemplateCloser && Tok.Next) { 891 if (Tok.Next->isOneOf(tok::identifier, tok::numeric_constant)) { 892 IsCast = true; 893 } else { 894 // Use heuristics to recognize c style casting. 895 FormatToken *Prev = Tok.Previous; 896 if (Prev && Prev->isOneOf(tok::amp, tok::star)) 897 Prev = Prev->Previous; 898 899 if (Prev && Tok.Next && Tok.Next->Next) { 900 bool NextIsUnary = Tok.Next->isUnaryOperator() || 901 Tok.Next->isOneOf(tok::amp, tok::star); 902 IsCast = NextIsUnary && Tok.Next->Next->isOneOf( 903 tok::identifier, tok::numeric_constant); 904 } 905 906 for (; Prev != Tok.MatchingParen; Prev = Prev->Previous) { 907 if (!Prev || !Prev->isOneOf(tok::kw_const, tok::identifier)) { 908 IsCast = false; 909 break; 910 } 911 } 912 } 913 } 914 return IsCast && !ParensAreEmpty; 915 } 916 917 /// \brief Return the type of the given token assuming it is * or &. 918 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 919 bool InTemplateArgument) { 920 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 921 if (!PrevToken) 922 return TT_UnaryOperator; 923 924 const FormatToken *NextToken = Tok.getNextNonComment(); 925 if (!NextToken || NextToken->is(tok::l_brace)) 926 return TT_Unknown; 927 928 if (PrevToken->is(tok::coloncolon) || 929 (PrevToken->is(tok::l_paren) && !IsExpression)) 930 return TT_PointerOrReference; 931 932 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 933 tok::comma, tok::semi, tok::kw_return, tok::colon, 934 tok::equal, tok::kw_delete, tok::kw_sizeof) || 935 PrevToken->Type == TT_BinaryOperator || 936 PrevToken->Type == TT_ConditionalExpr || 937 PrevToken->Type == TT_UnaryOperator || PrevToken->Type == TT_CastRParen) 938 return TT_UnaryOperator; 939 940 if (NextToken->is(tok::l_square) && NextToken->Type != TT_LambdaLSquare) 941 return TT_PointerOrReference; 942 if (NextToken->isOneOf(tok::kw_operator, tok::comma)) 943 return TT_PointerOrReference; 944 945 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen && 946 PrevToken->MatchingParen->Previous && 947 PrevToken->MatchingParen->Previous->isOneOf(tok::kw_typeof, 948 tok::kw_decltype)) 949 return TT_PointerOrReference; 950 951 if (PrevToken->Tok.isLiteral() || 952 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 953 tok::kw_false) || 954 NextToken->Tok.isLiteral() || 955 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 956 NextToken->isUnaryOperator() || 957 // If we know we're in a template argument, there are no named 958 // declarations. Thus, having an identifier on the right-hand side 959 // indicates a binary operator. 960 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 961 return TT_BinaryOperator; 962 963 // This catches some cases where evaluation order is used as control flow: 964 // aaa && aaa->f(); 965 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 966 if (NextNextToken && NextNextToken->is(tok::arrow)) 967 return TT_BinaryOperator; 968 969 // It is very unlikely that we are going to find a pointer or reference type 970 // definition on the RHS of an assignment. 971 if (IsExpression) 972 return TT_BinaryOperator; 973 974 return TT_PointerOrReference; 975 } 976 977 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 978 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 979 if (!PrevToken || PrevToken->Type == TT_CastRParen) 980 return TT_UnaryOperator; 981 982 // Use heuristics to recognize unary operators. 983 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 984 tok::question, tok::colon, tok::kw_return, 985 tok::kw_case, tok::at, tok::l_brace)) 986 return TT_UnaryOperator; 987 988 // There can't be two consecutive binary operators. 989 if (PrevToken->Type == TT_BinaryOperator) 990 return TT_UnaryOperator; 991 992 // Fall back to marking the token as binary operator. 993 return TT_BinaryOperator; 994 } 995 996 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements. 997 TokenType determineIncrementUsage(const FormatToken &Tok) { 998 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 999 if (!PrevToken || PrevToken->Type == TT_CastRParen) 1000 return TT_UnaryOperator; 1001 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 1002 return TT_TrailingUnaryOperator; 1003 1004 return TT_UnaryOperator; 1005 } 1006 1007 SmallVector<Context, 8> Contexts; 1008 1009 const FormatStyle &Style; 1010 AnnotatedLine &Line; 1011 FormatToken *CurrentToken; 1012 bool KeywordVirtualFound; 1013 bool AutoFound; 1014 IdentifierInfo &Ident_in; 1015 }; 1016 1017 static int PrecedenceUnaryOperator = prec::PointerToMember + 1; 1018 static int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 1019 1020 /// \brief Parses binary expressions by inserting fake parenthesis based on 1021 /// operator precedence. 1022 class ExpressionParser { 1023 public: 1024 ExpressionParser(AnnotatedLine &Line) : Current(Line.First) { 1025 // Skip leading "}", e.g. in "} else if (...) {". 1026 if (Current->is(tok::r_brace)) 1027 next(); 1028 } 1029 1030 /// \brief Parse expressions with the given operatore precedence. 1031 void parse(int Precedence = 0) { 1032 // Skip 'return' and ObjC selector colons as they are not part of a binary 1033 // expression. 1034 while (Current && 1035 (Current->is(tok::kw_return) || 1036 (Current->is(tok::colon) && (Current->Type == TT_ObjCMethodExpr || 1037 Current->Type == TT_DictLiteral)))) 1038 next(); 1039 1040 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1041 return; 1042 1043 // Conditional expressions need to be parsed separately for proper nesting. 1044 if (Precedence == prec::Conditional) { 1045 parseConditionalExpr(); 1046 return; 1047 } 1048 1049 // Parse unary operators, which all have a higher precedence than binary 1050 // operators. 1051 if (Precedence == PrecedenceUnaryOperator) { 1052 parseUnaryOperator(); 1053 return; 1054 } 1055 1056 FormatToken *Start = Current; 1057 FormatToken *LatestOperator = nullptr; 1058 unsigned OperatorIndex = 0; 1059 1060 while (Current) { 1061 // Consume operators with higher precedence. 1062 parse(Precedence + 1); 1063 1064 int CurrentPrecedence = getCurrentPrecedence(); 1065 1066 if (Current && Current->Type == TT_SelectorName && 1067 Precedence == CurrentPrecedence) { 1068 if (LatestOperator) 1069 addFakeParenthesis(Start, prec::Level(Precedence)); 1070 Start = Current; 1071 } 1072 1073 // At the end of the line or when an operator with higher precedence is 1074 // found, insert fake parenthesis and return. 1075 if (!Current || Current->closesScope() || 1076 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence)) { 1077 if (LatestOperator) { 1078 LatestOperator->LastOperator = true; 1079 if (Precedence == PrecedenceArrowAndPeriod) { 1080 // Call expressions don't have a binary operator precedence. 1081 addFakeParenthesis(Start, prec::Unknown); 1082 } else { 1083 addFakeParenthesis(Start, prec::Level(Precedence)); 1084 } 1085 } 1086 return; 1087 } 1088 1089 // Consume scopes: (), [], <> and {} 1090 if (Current->opensScope()) { 1091 while (Current && !Current->closesScope()) { 1092 next(); 1093 parse(); 1094 } 1095 next(); 1096 } else { 1097 // Operator found. 1098 if (CurrentPrecedence == Precedence) { 1099 LatestOperator = Current; 1100 Current->OperatorIndex = OperatorIndex; 1101 ++OperatorIndex; 1102 } 1103 1104 next(); 1105 } 1106 } 1107 } 1108 1109 private: 1110 /// \brief Gets the precedence (+1) of the given token for binary operators 1111 /// and other tokens that we treat like binary operators. 1112 int getCurrentPrecedence() { 1113 if (Current) { 1114 if (Current->Type == TT_ConditionalExpr) 1115 return prec::Conditional; 1116 else if (Current->is(tok::semi) || Current->Type == TT_InlineASMColon || 1117 Current->Type == TT_SelectorName) 1118 return 0; 1119 else if (Current->Type == TT_RangeBasedForLoopColon) 1120 return prec::Comma; 1121 else if (Current->Type == TT_BinaryOperator || Current->is(tok::comma)) 1122 return Current->getPrecedence(); 1123 else if (Current->isOneOf(tok::period, tok::arrow)) 1124 return PrecedenceArrowAndPeriod; 1125 } 1126 return -1; 1127 } 1128 1129 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1130 Start->FakeLParens.push_back(Precedence); 1131 if (Precedence > prec::Unknown) 1132 Start->StartsBinaryExpression = true; 1133 if (Current) { 1134 ++Current->Previous->FakeRParens; 1135 if (Precedence > prec::Unknown) 1136 Current->Previous->EndsBinaryExpression = true; 1137 } 1138 } 1139 1140 /// \brief Parse unary operator expressions and surround them with fake 1141 /// parentheses if appropriate. 1142 void parseUnaryOperator() { 1143 if (!Current || Current->Type != TT_UnaryOperator) { 1144 parse(PrecedenceArrowAndPeriod); 1145 return; 1146 } 1147 1148 FormatToken *Start = Current; 1149 next(); 1150 parseUnaryOperator(); 1151 1152 // The actual precedence doesn't matter. 1153 addFakeParenthesis(Start, prec::Unknown); 1154 } 1155 1156 void parseConditionalExpr() { 1157 FormatToken *Start = Current; 1158 parse(prec::LogicalOr); 1159 if (!Current || !Current->is(tok::question)) 1160 return; 1161 next(); 1162 parseConditionalExpr(); 1163 if (!Current || Current->Type != TT_ConditionalExpr) 1164 return; 1165 next(); 1166 parseConditionalExpr(); 1167 addFakeParenthesis(Start, prec::Conditional); 1168 } 1169 1170 void next() { 1171 if (Current) 1172 Current = Current->Next; 1173 while (Current && Current->isTrailingComment()) 1174 Current = Current->Next; 1175 } 1176 1177 FormatToken *Current; 1178 }; 1179 1180 } // end anonymous namespace 1181 1182 void 1183 TokenAnnotator::setCommentLineLevels(SmallVectorImpl<AnnotatedLine *> &Lines) { 1184 const AnnotatedLine *NextNonCommentLine = nullptr; 1185 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 1186 E = Lines.rend(); 1187 I != E; ++I) { 1188 if (NextNonCommentLine && (*I)->First->is(tok::comment) && 1189 (*I)->First->Next == nullptr) 1190 (*I)->Level = NextNonCommentLine->Level; 1191 else 1192 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 1193 1194 setCommentLineLevels((*I)->Children); 1195 } 1196 } 1197 1198 void TokenAnnotator::annotate(AnnotatedLine &Line) { 1199 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1200 E = Line.Children.end(); 1201 I != E; ++I) { 1202 annotate(**I); 1203 } 1204 AnnotatingParser Parser(Style, Line, Ident_in); 1205 Line.Type = Parser.parseLine(); 1206 if (Line.Type == LT_Invalid) 1207 return; 1208 1209 ExpressionParser ExprParser(Line); 1210 ExprParser.parse(); 1211 1212 if (Line.First->Type == TT_ObjCMethodSpecifier) 1213 Line.Type = LT_ObjCMethodDecl; 1214 else if (Line.First->Type == TT_ObjCDecl) 1215 Line.Type = LT_ObjCDecl; 1216 else if (Line.First->Type == TT_ObjCProperty) 1217 Line.Type = LT_ObjCProperty; 1218 1219 Line.First->SpacesRequiredBefore = 1; 1220 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 1221 } 1222 1223 // This function heuristically determines whether 'Current' starts the name of a 1224 // function declaration. 1225 static bool isFunctionDeclarationName(const FormatToken &Current) { 1226 if (Current.Type != TT_StartOfName || 1227 Current.NestingLevel != 0 || 1228 Current.Previous->Type == TT_StartOfName) 1229 return false; 1230 const FormatToken *Next = Current.Next; 1231 for (; Next; Next = Next->Next) { 1232 if (Next->Type == TT_TemplateOpener) { 1233 Next = Next->MatchingParen; 1234 } else if (Next->is(tok::coloncolon)) { 1235 Next = Next->Next; 1236 if (!Next || !Next->is(tok::identifier)) 1237 return false; 1238 } else if (Next->is(tok::l_paren)) { 1239 break; 1240 } else { 1241 return false; 1242 } 1243 } 1244 if (!Next) 1245 return false; 1246 assert(Next->is(tok::l_paren)); 1247 if (Next->Next == Next->MatchingParen) 1248 return true; 1249 for (const FormatToken *Tok = Next->Next; Tok != Next->MatchingParen; 1250 Tok = Tok->Next) { 1251 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 1252 Tok->Type == TT_PointerOrReference || Tok->Type == TT_StartOfName) 1253 return true; 1254 if (Tok->isOneOf(tok::l_brace, tok::string_literal) || Tok->Tok.isLiteral()) 1255 return false; 1256 } 1257 return false; 1258 } 1259 1260 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 1261 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1262 E = Line.Children.end(); 1263 I != E; ++I) { 1264 calculateFormattingInformation(**I); 1265 } 1266 1267 Line.First->TotalLength = 1268 Line.First->IsMultiline ? Style.ColumnLimit : Line.First->ColumnWidth; 1269 if (!Line.First->Next) 1270 return; 1271 FormatToken *Current = Line.First->Next; 1272 bool InFunctionDecl = Line.MightBeFunctionDecl; 1273 while (Current) { 1274 if (isFunctionDeclarationName(*Current)) 1275 Current->Type = TT_FunctionDeclarationName; 1276 if (Current->Type == TT_LineComment) { 1277 if (Current->Previous->BlockKind == BK_BracedInit && 1278 Current->Previous->opensScope()) 1279 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1; 1280 else 1281 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 1282 1283 // If we find a trailing comment, iterate backwards to determine whether 1284 // it seems to relate to a specific parameter. If so, break before that 1285 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 1286 // to the previous line in: 1287 // SomeFunction(a, 1288 // b, // comment 1289 // c); 1290 if (!Current->HasUnescapedNewline) { 1291 for (FormatToken *Parameter = Current->Previous; Parameter; 1292 Parameter = Parameter->Previous) { 1293 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 1294 break; 1295 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 1296 if (Parameter->Previous->Type != TT_CtorInitializerComma && 1297 Parameter->HasUnescapedNewline) 1298 Parameter->MustBreakBefore = true; 1299 break; 1300 } 1301 } 1302 } 1303 } else if (Current->SpacesRequiredBefore == 0 && 1304 spaceRequiredBefore(Line, *Current)) { 1305 Current->SpacesRequiredBefore = 1; 1306 } 1307 1308 Current->MustBreakBefore = 1309 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 1310 1311 if (Style.AlwaysBreakAfterDefinitionReturnType && 1312 InFunctionDecl && Current->Type == TT_FunctionDeclarationName && 1313 !Line.Last->isOneOf(tok::semi, tok::comment)) // Only for definitions. 1314 // FIXME: Line.Last points to other characters than tok::semi 1315 // and tok::lbrace. 1316 Current->MustBreakBefore = true; 1317 1318 Current->CanBreakBefore = 1319 Current->MustBreakBefore || canBreakBefore(Line, *Current); 1320 unsigned ChildSize = 0; 1321 if (Current->Previous->Children.size() == 1) { 1322 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 1323 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 1324 : LastOfChild.TotalLength + 1; 1325 } 1326 if (Current->MustBreakBefore || Current->Previous->Children.size() > 1 || 1327 Current->IsMultiline) 1328 Current->TotalLength = Current->Previous->TotalLength + Style.ColumnLimit; 1329 else 1330 Current->TotalLength = Current->Previous->TotalLength + 1331 Current->ColumnWidth + ChildSize + 1332 Current->SpacesRequiredBefore; 1333 1334 if (Current->Type == TT_CtorInitializerColon) 1335 InFunctionDecl = false; 1336 1337 // FIXME: Only calculate this if CanBreakBefore is true once static 1338 // initializers etc. are sorted out. 1339 // FIXME: Move magic numbers to a better place. 1340 Current->SplitPenalty = 20 * Current->BindingStrength + 1341 splitPenalty(Line, *Current, InFunctionDecl); 1342 1343 Current = Current->Next; 1344 } 1345 1346 calculateUnbreakableTailLengths(Line); 1347 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 1348 if (Current->Role) 1349 Current->Role->precomputeFormattingInfos(Current); 1350 } 1351 1352 DEBUG({ printDebugInfo(Line); }); 1353 } 1354 1355 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 1356 unsigned UnbreakableTailLength = 0; 1357 FormatToken *Current = Line.Last; 1358 while (Current) { 1359 Current->UnbreakableTailLength = UnbreakableTailLength; 1360 if (Current->CanBreakBefore || 1361 Current->isOneOf(tok::comment, tok::string_literal)) { 1362 UnbreakableTailLength = 0; 1363 } else { 1364 UnbreakableTailLength += 1365 Current->ColumnWidth + Current->SpacesRequiredBefore; 1366 } 1367 Current = Current->Previous; 1368 } 1369 } 1370 1371 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 1372 const FormatToken &Tok, 1373 bool InFunctionDecl) { 1374 const FormatToken &Left = *Tok.Previous; 1375 const FormatToken &Right = Tok; 1376 1377 if (Left.is(tok::semi)) 1378 return 0; 1379 if (Left.is(tok::comma) || (Right.is(tok::identifier) && Right.Next && 1380 Right.Next->Type == TT_DictLiteral)) 1381 return 1; 1382 if (Right.is(tok::l_square)) { 1383 if (Style.Language == FormatStyle::LK_Proto) 1384 return 1; 1385 if (Right.Type != TT_ObjCMethodExpr && Right.Type != TT_LambdaLSquare) 1386 return 500; 1387 } 1388 if (Right.Type == TT_StartOfName || 1389 Right.Type == TT_FunctionDeclarationName || Right.is(tok::kw_operator)) { 1390 if (Line.First->is(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 1391 return 3; 1392 if (Left.Type == TT_StartOfName) 1393 return 20; 1394 if (InFunctionDecl && Right.NestingLevel == 0) 1395 return Style.PenaltyReturnTypeOnItsOwnLine; 1396 return 200; 1397 } 1398 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 1399 return 150; 1400 if (Left.Type == TT_CastRParen) 1401 return 100; 1402 if (Left.is(tok::coloncolon) || 1403 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 1404 return 500; 1405 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 1406 return 5000; 1407 1408 if (Left.Type == TT_RangeBasedForLoopColon || 1409 Left.Type == TT_InheritanceColon) 1410 return 2; 1411 1412 if (Right.isMemberAccess()) { 1413 if (Left.is(tok::r_paren) && Left.MatchingParen && 1414 Left.MatchingParen->ParameterCount > 0) 1415 return 20; // Should be smaller than breaking at a nested comma. 1416 return 150; 1417 } 1418 1419 if (Right.Type == TT_TrailingAnnotation && 1420 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 1421 // Generally, breaking before a trailing annotation is bad unless it is 1422 // function-like. It seems to be especially preferable to keep standard 1423 // annotations (i.e. "const", "final" and "override") on the same line. 1424 // Use a slightly higher penalty after ")" so that annotations like 1425 // "const override" are kept together. 1426 bool is_short_annotation = Right.TokenText.size() < 10; 1427 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 1428 } 1429 1430 // In for-loops, prefer breaking at ',' and ';'. 1431 if (Line.First->is(tok::kw_for) && Left.is(tok::equal)) 1432 return 4; 1433 1434 // In Objective-C method expressions, prefer breaking before "param:" over 1435 // breaking after it. 1436 if (Right.Type == TT_SelectorName) 1437 return 0; 1438 if (Left.is(tok::colon) && Left.Type == TT_ObjCMethodExpr) 1439 return Line.MightBeFunctionDecl ? 50 : 500; 1440 1441 if (Left.is(tok::l_paren) && InFunctionDecl) 1442 return 100; 1443 if (Left.is(tok::equal) && InFunctionDecl) 1444 return 110; 1445 if (Right.is(tok::r_brace)) 1446 return 1; 1447 if (Left.Type == TT_TemplateOpener) 1448 return 100; 1449 if (Left.opensScope()) 1450 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 1451 : 19; 1452 1453 if (Right.is(tok::lessless)) { 1454 if (Left.is(tok::string_literal)) { 1455 StringRef Content = Left.TokenText; 1456 if (Content.startswith("\"")) 1457 Content = Content.drop_front(1); 1458 if (Content.endswith("\"")) 1459 Content = Content.drop_back(1); 1460 Content = Content.trim(); 1461 if (Content.size() > 1 && 1462 (Content.back() == ':' || Content.back() == '=')) 1463 return 25; 1464 } 1465 return 1; // Breaking at a << is really cheap. 1466 } 1467 if (Left.Type == TT_ConditionalExpr) 1468 return prec::Conditional; 1469 prec::Level Level = Left.getPrecedence(); 1470 1471 if (Level != prec::Unknown) 1472 return Level; 1473 1474 return 3; 1475 } 1476 1477 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 1478 const FormatToken &Left, 1479 const FormatToken &Right) { 1480 if (Style.Language == FormatStyle::LK_Proto) { 1481 if (Right.is(tok::period) && 1482 (Left.TokenText == "optional" || Left.TokenText == "required" || 1483 Left.TokenText == "repeated")) 1484 return true; 1485 if (Right.is(tok::l_paren) && 1486 (Left.TokenText == "returns" || Left.TokenText == "option")) 1487 return true; 1488 } else if (Style.Language == FormatStyle::LK_JavaScript) { 1489 if (Left.TokenText == "var") 1490 return true; 1491 } 1492 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 1493 return true; 1494 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 1495 Left.Tok.getObjCKeywordID() == tok::objc_property) 1496 return true; 1497 if (Right.is(tok::hashhash)) 1498 return Left.is(tok::hash); 1499 if (Left.isOneOf(tok::hashhash, tok::hash)) 1500 return Right.is(tok::hash); 1501 if (Left.is(tok::l_paren) && Right.is(tok::r_paren)) 1502 return Style.SpaceInEmptyParentheses; 1503 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 1504 return (Right.Type == TT_CastRParen || 1505 (Left.MatchingParen && Left.MatchingParen->Type == TT_CastRParen)) 1506 ? Style.SpacesInCStyleCastParentheses 1507 : Style.SpacesInParentheses; 1508 if (Style.SpacesInAngles && 1509 ((Left.Type == TT_TemplateOpener) != (Right.Type == TT_TemplateCloser))) 1510 return true; 1511 if (Right.isOneOf(tok::semi, tok::comma)) 1512 return false; 1513 if (Right.is(tok::less) && 1514 (Left.isOneOf(tok::kw_template, tok::r_paren) || 1515 (Line.Type == LT_ObjCDecl && Style.ObjCSpaceBeforeProtocolList))) 1516 return true; 1517 if (Left.is(tok::arrow) || Right.is(tok::arrow)) 1518 return false; 1519 if (Left.isOneOf(tok::exclaim, tok::tilde)) 1520 return false; 1521 if (Left.is(tok::at) && 1522 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 1523 tok::numeric_constant, tok::l_paren, tok::l_brace, 1524 tok::kw_true, tok::kw_false)) 1525 return false; 1526 if (Left.is(tok::coloncolon)) 1527 return false; 1528 if (Right.is(tok::coloncolon) && Left.isNot(tok::l_brace)) 1529 return (Left.is(tok::less) && Style.Standard == FormatStyle::LS_Cpp03) || 1530 !Left.isOneOf(tok::identifier, tok::greater, tok::l_paren, 1531 tok::r_paren, tok::less); 1532 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) 1533 return false; 1534 if (Right.is(tok::ellipsis)) 1535 return Left.Tok.isLiteral(); 1536 if (Left.is(tok::l_square) && Right.is(tok::amp)) 1537 return false; 1538 if (Right.Type == TT_PointerOrReference) 1539 return Left.Tok.isLiteral() || 1540 ((Left.Type != TT_PointerOrReference) && Left.isNot(tok::l_paren) && 1541 Style.PointerAlignment != FormatStyle::PAS_Left); 1542 if (Right.Type == TT_FunctionTypeLParen && Left.isNot(tok::l_paren) && 1543 (Left.Type != TT_PointerOrReference || 1544 Style.PointerAlignment != FormatStyle::PAS_Right)) 1545 return true; 1546 if (Left.Type == TT_PointerOrReference) 1547 return Right.Tok.isLiteral() || Right.Type == TT_BlockComment || 1548 ((Right.Type != TT_PointerOrReference) && 1549 Right.isNot(tok::l_paren) && 1550 Style.PointerAlignment != FormatStyle::PAS_Right && Left.Previous && 1551 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon)); 1552 if (Right.is(tok::star) && Left.is(tok::l_paren)) 1553 return false; 1554 if (Left.is(tok::l_square)) 1555 return (Left.Type == TT_ArrayInitializerLSquare && 1556 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square)) || 1557 (Left.Type == TT_ArraySubscriptLSquare && 1558 Style.SpacesInSquareBrackets && Right.isNot(tok::r_square)); 1559 if (Right.is(tok::r_square)) 1560 return Right.MatchingParen && 1561 ((Style.SpacesInContainerLiterals && 1562 Right.MatchingParen->Type == TT_ArrayInitializerLSquare) || 1563 (Style.SpacesInSquareBrackets && 1564 Right.MatchingParen->Type == TT_ArraySubscriptLSquare)); 1565 if (Right.is(tok::l_square) && Right.Type != TT_ObjCMethodExpr && 1566 Right.Type != TT_LambdaLSquare && Left.isNot(tok::numeric_constant) && 1567 Left.Type != TT_DictLiteral) 1568 return false; 1569 if (Left.is(tok::colon)) 1570 return Left.Type != TT_ObjCMethodExpr; 1571 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 1572 return !Left.Children.empty(); // No spaces in "{}". 1573 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 1574 (Right.is(tok::r_brace) && Right.MatchingParen && 1575 Right.MatchingParen->BlockKind != BK_Block)) 1576 return !Style.Cpp11BracedListStyle; 1577 if (Left.Type == TT_BlockComment) 1578 return !Left.TokenText.endswith("=*/"); 1579 if (Right.is(tok::l_paren)) { 1580 if (Left.is(tok::r_paren) && Left.Type == TT_AttributeParen) 1581 return true; 1582 return Line.Type == LT_ObjCDecl || 1583 Left.isOneOf(tok::kw_new, tok::kw_delete, tok::semi) || 1584 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 1585 (Left.isOneOf(tok::kw_if, tok::kw_for, tok::kw_while, 1586 tok::kw_switch, tok::kw_catch, tok::kw_case) || 1587 Left.IsForEachMacro)) || 1588 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always && 1589 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword()) && 1590 Line.Type != LT_PreprocessorDirective); 1591 } 1592 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 1593 return false; 1594 if (Right.Type == TT_UnaryOperator) 1595 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 1596 (Left.isNot(tok::colon) || Left.Type != TT_ObjCMethodExpr); 1597 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 1598 tok::r_paren) || 1599 Left.isSimpleTypeSpecifier()) && 1600 Right.is(tok::l_brace) && Right.getNextNonComment() && 1601 Right.BlockKind != BK_Block) 1602 return false; 1603 if (Left.is(tok::period) || Right.is(tok::period)) 1604 return false; 1605 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 1606 return false; 1607 return true; 1608 } 1609 1610 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 1611 const FormatToken &Tok) { 1612 if (Tok.Tok.getIdentifierInfo() && Tok.Previous->Tok.getIdentifierInfo()) 1613 return true; // Never ever merge two identifiers. 1614 if (Tok.Previous->Type == TT_ImplicitStringLiteral) 1615 return Tok.WhitespaceRange.getBegin() != Tok.WhitespaceRange.getEnd(); 1616 if (Line.Type == LT_ObjCMethodDecl) { 1617 if (Tok.Previous->Type == TT_ObjCMethodSpecifier) 1618 return true; 1619 if (Tok.Previous->is(tok::r_paren) && Tok.is(tok::identifier)) 1620 // Don't space between ')' and <id> 1621 return false; 1622 } 1623 if (Line.Type == LT_ObjCProperty && 1624 (Tok.is(tok::equal) || Tok.Previous->is(tok::equal))) 1625 return false; 1626 1627 if (Tok.Type == TT_TrailingReturnArrow || 1628 Tok.Previous->Type == TT_TrailingReturnArrow) 1629 return true; 1630 if (Tok.Previous->is(tok::comma)) 1631 return true; 1632 if (Tok.is(tok::comma)) 1633 return false; 1634 if (Tok.Type == TT_CtorInitializerColon || Tok.Type == TT_ObjCBlockLParen) 1635 return true; 1636 if (Tok.Previous->Tok.is(tok::kw_operator)) 1637 return Tok.is(tok::coloncolon); 1638 if (Tok.Type == TT_OverloadedOperatorLParen) 1639 return false; 1640 if (Tok.is(tok::colon)) 1641 return !Line.First->isOneOf(tok::kw_case, tok::kw_default) && 1642 Tok.getNextNonComment() && Tok.Type != TT_ObjCMethodExpr && 1643 !Tok.Previous->is(tok::question) && 1644 (Tok.Type != TT_DictLiteral || Style.SpacesInContainerLiterals); 1645 if (Tok.Previous->Type == TT_UnaryOperator || 1646 Tok.Previous->Type == TT_CastRParen) 1647 return Tok.Type == TT_BinaryOperator; 1648 if (Tok.Previous->is(tok::greater) && Tok.is(tok::greater)) { 1649 return Tok.Type == TT_TemplateCloser && 1650 Tok.Previous->Type == TT_TemplateCloser && 1651 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 1652 } 1653 if (Tok.isOneOf(tok::arrowstar, tok::periodstar) || 1654 Tok.Previous->isOneOf(tok::arrowstar, tok::periodstar)) 1655 return false; 1656 if (!Style.SpaceBeforeAssignmentOperators && 1657 Tok.getPrecedence() == prec::Assignment) 1658 return false; 1659 if ((Tok.Type == TT_BinaryOperator && !Tok.Previous->is(tok::l_paren)) || 1660 Tok.Previous->Type == TT_BinaryOperator || 1661 Tok.Previous->Type == TT_ConditionalExpr) 1662 return true; 1663 if (Tok.Previous->Type == TT_TemplateCloser && Tok.is(tok::l_paren)) 1664 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 1665 if (Tok.is(tok::less) && Tok.Previous->isNot(tok::l_paren) && 1666 Line.First->is(tok::hash)) 1667 return true; 1668 if (Tok.Type == TT_TrailingUnaryOperator) 1669 return false; 1670 if (Tok.Previous->Type == TT_RegexLiteral) 1671 return false; 1672 return spaceRequiredBetween(Line, *Tok.Previous, Tok); 1673 } 1674 1675 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 1676 static bool isAllmanBrace(const FormatToken &Tok) { 1677 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 1678 Tok.Type != TT_ObjCBlockLBrace && Tok.Type != TT_DictLiteral; 1679 } 1680 1681 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 1682 const FormatToken &Right) { 1683 const FormatToken &Left = *Right.Previous; 1684 if (Right.NewlinesBefore > 1) 1685 return true; 1686 if (Right.is(tok::comment)) { 1687 return Right.Previous->BlockKind != BK_BracedInit && 1688 Right.Previous->Type != TT_CtorInitializerColon && 1689 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 1690 } else if (Right.Previous->isTrailingComment() || 1691 (Right.isStringLiteral() && Right.Previous->isStringLiteral())) { 1692 return true; 1693 } else if (Right.Previous->IsUnterminatedLiteral) { 1694 return true; 1695 } else if (Right.is(tok::lessless) && Right.Next && 1696 Right.Previous->is(tok::string_literal) && 1697 Right.Next->is(tok::string_literal)) { 1698 return true; 1699 } else if (Right.Previous->ClosesTemplateDeclaration && 1700 Right.Previous->MatchingParen && 1701 Right.Previous->MatchingParen->NestingLevel == 0 && 1702 Style.AlwaysBreakTemplateDeclarations) { 1703 return true; 1704 } else if ((Right.Type == TT_CtorInitializerComma || 1705 Right.Type == TT_CtorInitializerColon) && 1706 Style.BreakConstructorInitializersBeforeComma && 1707 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) { 1708 return true; 1709 } else if (Right.is(tok::string_literal) && 1710 Right.TokenText.startswith("R\"")) { 1711 // Raw string literals are special wrt. line breaks. The author has made a 1712 // deliberate choice and might have aligned the contents of the string 1713 // literal accordingly. Thus, we try keep existing line breaks. 1714 return Right.NewlinesBefore > 0; 1715 } else if (Right.Previous->is(tok::l_brace) && Right.NestingLevel == 1 && 1716 Style.Language == FormatStyle::LK_Proto) { 1717 // Don't enums onto single lines in protocol buffers. 1718 return true; 1719 } else if (isAllmanBrace(Left) || isAllmanBrace(Right)) { 1720 return Style.BreakBeforeBraces == FormatStyle::BS_Allman || 1721 Style.BreakBeforeBraces == FormatStyle::BS_GNU; 1722 } else if (Style.Language == FormatStyle::LK_Proto && 1723 Left.isNot(tok::l_brace) && Right.Type == TT_SelectorName) { 1724 return true; 1725 } 1726 1727 // If the last token before a '}' is a comma or a trailing comment, the 1728 // intention is to insert a line break after it in order to make shuffling 1729 // around entries easier. 1730 const FormatToken *BeforeClosingBrace = nullptr; 1731 if (Left.is(tok::l_brace) && Left.MatchingParen) 1732 BeforeClosingBrace = Left.MatchingParen->Previous; 1733 else if (Right.is(tok::r_brace)) 1734 BeforeClosingBrace = Right.Previous; 1735 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 1736 BeforeClosingBrace->isTrailingComment())) 1737 return true; 1738 1739 if (Style.Language == FormatStyle::LK_JavaScript) { 1740 // FIXME: This might apply to other languages and token kinds. 1741 if (Right.is(tok::char_constant) && Left.is(tok::plus) && Left.Previous && 1742 Left.Previous->is(tok::char_constant)) 1743 return true; 1744 } 1745 1746 return false; 1747 } 1748 1749 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 1750 const FormatToken &Right) { 1751 const FormatToken &Left = *Right.Previous; 1752 if (Left.is(tok::at)) 1753 return false; 1754 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 1755 return false; 1756 if (Right.Type == TT_StartOfName || 1757 Right.Type == TT_FunctionDeclarationName || Right.is(tok::kw_operator)) 1758 return true; 1759 if (Right.isTrailingComment()) 1760 // We rely on MustBreakBefore being set correctly here as we should not 1761 // change the "binding" behavior of a comment. 1762 // The first comment in a braced lists is always interpreted as belonging to 1763 // the first list element. Otherwise, it should be placed outside of the 1764 // list. 1765 return Left.BlockKind == BK_BracedInit; 1766 if (Left.is(tok::question) && Right.is(tok::colon)) 1767 return false; 1768 if (Right.Type == TT_ConditionalExpr || Right.is(tok::question)) 1769 return Style.BreakBeforeTernaryOperators; 1770 if (Left.Type == TT_ConditionalExpr || Left.is(tok::question)) 1771 return !Style.BreakBeforeTernaryOperators; 1772 if (Right.Type == TT_InheritanceColon) 1773 return true; 1774 if (Right.is(tok::colon) && (Right.Type != TT_CtorInitializerColon && 1775 Right.Type != TT_InlineASMColon)) 1776 return false; 1777 if (Left.is(tok::colon) && 1778 (Left.Type == TT_DictLiteral || Left.Type == TT_ObjCMethodExpr)) 1779 return true; 1780 if (Right.Type == TT_SelectorName) 1781 return true; 1782 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 1783 return true; 1784 if (Left.ClosesTemplateDeclaration) 1785 return true; 1786 if (Right.Type == TT_RangeBasedForLoopColon || 1787 Right.Type == TT_OverloadedOperatorLParen || 1788 Right.Type == TT_OverloadedOperator) 1789 return false; 1790 if (Left.Type == TT_RangeBasedForLoopColon) 1791 return true; 1792 if (Right.Type == TT_RangeBasedForLoopColon) 1793 return false; 1794 if (Left.Type == TT_PointerOrReference || Left.Type == TT_TemplateCloser || 1795 Left.Type == TT_UnaryOperator || Left.is(tok::kw_operator)) 1796 return false; 1797 if (Left.is(tok::equal) && Line.Type == LT_VirtualFunctionDecl) 1798 return false; 1799 if (Left.is(tok::l_paren) && Left.Type == TT_AttributeParen) 1800 return false; 1801 if (Left.is(tok::l_paren) && Left.Previous && 1802 (Left.Previous->Type == TT_BinaryOperator || 1803 Left.Previous->Type == TT_CastRParen || Left.Previous->is(tok::kw_if))) 1804 return false; 1805 if (Right.Type == TT_ImplicitStringLiteral) 1806 return false; 1807 1808 if (Right.is(tok::r_paren) || Right.Type == TT_TemplateCloser) 1809 return false; 1810 1811 // We only break before r_brace if there was a corresponding break before 1812 // the l_brace, which is tracked by BreakBeforeClosingBrace. 1813 if (Right.is(tok::r_brace)) 1814 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 1815 1816 // Allow breaking after a trailing annotation, e.g. after a method 1817 // declaration. 1818 if (Left.Type == TT_TrailingAnnotation) 1819 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 1820 tok::less, tok::coloncolon); 1821 1822 if (Right.is(tok::kw___attribute)) 1823 return true; 1824 1825 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 1826 return true; 1827 1828 if (Right.is(tok::identifier) && Right.Next && 1829 Right.Next->Type == TT_DictLiteral) 1830 return true; 1831 1832 if (Left.Type == TT_CtorInitializerComma && 1833 Style.BreakConstructorInitializersBeforeComma) 1834 return false; 1835 if (Right.Type == TT_CtorInitializerComma && 1836 Style.BreakConstructorInitializersBeforeComma) 1837 return true; 1838 if (Left.is(tok::greater) && Right.is(tok::greater) && 1839 Left.Type != TT_TemplateCloser) 1840 return false; 1841 if (Right.Type == TT_BinaryOperator && Style.BreakBeforeBinaryOperators) 1842 return true; 1843 if (Left.Type == TT_ArrayInitializerLSquare) 1844 return true; 1845 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 1846 return true; 1847 return (Left.isBinaryOperator() && 1848 !Left.isOneOf(tok::arrowstar, tok::lessless) && 1849 !Style.BreakBeforeBinaryOperators) || 1850 Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 1851 tok::kw_class, tok::kw_struct) || 1852 Right.isMemberAccess() || 1853 Right.isOneOf(tok::lessless, tok::colon, tok::l_square, tok::at) || 1854 (Left.is(tok::r_paren) && 1855 Right.isOneOf(tok::identifier, tok::kw_const)) || 1856 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)); 1857 } 1858 1859 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 1860 llvm::errs() << "AnnotatedTokens:\n"; 1861 const FormatToken *Tok = Line.First; 1862 while (Tok) { 1863 llvm::errs() << " M=" << Tok->MustBreakBefore 1864 << " C=" << Tok->CanBreakBefore << " T=" << Tok->Type 1865 << " S=" << Tok->SpacesRequiredBefore 1866 << " B=" << Tok->BlockParameterCount 1867 << " P=" << Tok->SplitPenalty << " Name=" << Tok->Tok.getName() 1868 << " L=" << Tok->TotalLength << " PPK=" << Tok->PackingKind 1869 << " FakeLParens="; 1870 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 1871 llvm::errs() << Tok->FakeLParens[i] << "/"; 1872 llvm::errs() << " FakeRParens=" << Tok->FakeRParens << "\n"; 1873 if (!Tok->Next) 1874 assert(Tok == Line.Last); 1875 Tok = Tok->Next; 1876 } 1877 llvm::errs() << "----\n"; 1878 } 1879 1880 } // namespace format 1881 } // namespace clang 1882