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/ADT/SmallPtrSet.h" 19 #include "llvm/Support/Debug.h" 20 21 #define DEBUG_TYPE "format-token-annotator" 22 23 namespace clang { 24 namespace format { 25 26 namespace { 27 28 /// \brief A parser that gathers additional information about tokens. 29 /// 30 /// The \c TokenAnnotator tries to match parenthesis and square brakets and 31 /// store a parenthesis levels. It also tries to resolve matching "<" and ">" 32 /// into template parameter lists. 33 class AnnotatingParser { 34 public: 35 AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line, 36 const AdditionalKeywords &Keywords) 37 : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false), 38 Keywords(Keywords) { 39 Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false)); 40 resetTokenMetadata(CurrentToken); 41 } 42 43 private: 44 bool parseAngle() { 45 if (!CurrentToken || !CurrentToken->Previous) 46 return false; 47 if (NonTemplateLess.count(CurrentToken->Previous)) 48 return false; 49 50 const FormatToken& Previous = *CurrentToken->Previous; 51 if (Previous.Previous) { 52 if (Previous.Previous->Tok.isLiteral()) 53 return false; 54 if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 && 55 (!Previous.Previous->MatchingParen || 56 !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen))) 57 return false; 58 } 59 60 FormatToken *Left = CurrentToken->Previous; 61 Left->ParentBracket = Contexts.back().ContextKind; 62 ScopedContextCreator ContextCreator(*this, tok::less, 12); 63 64 // If this angle is in the context of an expression, we need to be more 65 // hesitant to detect it as opening template parameters. 66 bool InExprContext = Contexts.back().IsExpression; 67 68 Contexts.back().IsExpression = false; 69 // If there's a template keyword before the opening angle bracket, this is a 70 // template parameter, not an argument. 71 Contexts.back().InTemplateArgument = 72 Left->Previous && Left->Previous->Tok.isNot(tok::kw_template); 73 74 if (Style.Language == FormatStyle::LK_Java && 75 CurrentToken->is(tok::question)) 76 next(); 77 78 while (CurrentToken) { 79 if (CurrentToken->is(tok::greater)) { 80 Left->MatchingParen = CurrentToken; 81 CurrentToken->MatchingParen = Left; 82 CurrentToken->Type = TT_TemplateCloser; 83 next(); 84 return true; 85 } 86 if (CurrentToken->is(tok::question) && 87 Style.Language == FormatStyle::LK_Java) { 88 next(); 89 continue; 90 } 91 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) || 92 (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext)) 93 return false; 94 // If a && or || is found and interpreted as a binary operator, this set 95 // of angles is likely part of something like "a < b && c > d". If the 96 // angles are inside an expression, the ||/&& might also be a binary 97 // operator that was misinterpreted because we are parsing template 98 // parameters. 99 // FIXME: This is getting out of hand, write a decent parser. 100 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) && 101 CurrentToken->Previous->is(TT_BinaryOperator) && 102 Contexts[Contexts.size() - 2].IsExpression && 103 !Line.startsWith(tok::kw_template)) 104 return false; 105 updateParameterCount(Left, CurrentToken); 106 if (!consumeToken()) 107 return false; 108 } 109 return false; 110 } 111 112 bool parseParens(bool LookForDecls = false) { 113 if (!CurrentToken) 114 return false; 115 FormatToken *Left = CurrentToken->Previous; 116 Left->ParentBracket = Contexts.back().ContextKind; 117 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1); 118 119 // FIXME: This is a bit of a hack. Do better. 120 Contexts.back().ColonIsForRangeExpr = 121 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr; 122 123 bool StartsObjCMethodExpr = false; 124 if (CurrentToken->is(tok::caret)) { 125 // (^ can start a block type. 126 Left->Type = TT_ObjCBlockLParen; 127 } else if (FormatToken *MaybeSel = Left->Previous) { 128 // @selector( starts a selector. 129 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous && 130 MaybeSel->Previous->is(tok::at)) { 131 StartsObjCMethodExpr = true; 132 } 133 } 134 135 if (Left->is(TT_OverloadedOperatorLParen)) { 136 Contexts.back().IsExpression = false; 137 } else if (Left->Previous && 138 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype, 139 tok::kw_if, tok::kw_while, tok::l_paren, 140 tok::comma) || 141 Left->Previous->is(TT_BinaryOperator))) { 142 // static_assert, if and while usually contain expressions. 143 Contexts.back().IsExpression = true; 144 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 145 (Left->Previous->is(Keywords.kw_function) || 146 (Left->Previous->endsSequence(tok::identifier, 147 Keywords.kw_function)))) { 148 // function(...) or function f(...) 149 Contexts.back().IsExpression = false; 150 } else if (Left->Previous && Left->Previous->is(tok::r_square) && 151 Left->Previous->MatchingParen && 152 Left->Previous->MatchingParen->is(TT_LambdaLSquare)) { 153 // This is a parameter list of a lambda expression. 154 Contexts.back().IsExpression = false; 155 } else if (Line.InPPDirective && 156 (!Left->Previous || !Left->Previous->is(tok::identifier))) { 157 Contexts.back().IsExpression = true; 158 } else if (Contexts[Contexts.size() - 2].CaretFound) { 159 // This is the parameter list of an ObjC block. 160 Contexts.back().IsExpression = false; 161 } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) { 162 Left->Type = TT_AttributeParen; 163 } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) { 164 // The first argument to a foreach macro is a declaration. 165 Contexts.back().IsForEachMacro = true; 166 Contexts.back().IsExpression = false; 167 } else if (Left->Previous && Left->Previous->MatchingParen && 168 Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) { 169 Contexts.back().IsExpression = false; 170 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) { 171 bool IsForOrCatch = 172 Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch); 173 Contexts.back().IsExpression = !IsForOrCatch; 174 } 175 176 if (StartsObjCMethodExpr) { 177 Contexts.back().ColonIsObjCMethodExpr = true; 178 Left->Type = TT_ObjCMethodExpr; 179 } 180 181 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression; 182 bool ProbablyFunctionType = CurrentToken->isOneOf(tok::star, tok::amp); 183 bool HasMultipleLines = false; 184 bool HasMultipleParametersOnALine = false; 185 bool MightBeObjCForRangeLoop = 186 Left->Previous && Left->Previous->is(tok::kw_for); 187 while (CurrentToken) { 188 // LookForDecls is set when "if (" has been seen. Check for 189 // 'identifier' '*' 'identifier' followed by not '=' -- this 190 // '*' has to be a binary operator but determineStarAmpUsage() will 191 // categorize it as an unary operator, so set the right type here. 192 if (LookForDecls && CurrentToken->Next) { 193 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 194 if (Prev) { 195 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 196 FormatToken *Next = CurrentToken->Next; 197 if (PrevPrev && PrevPrev->is(tok::identifier) && 198 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 199 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 200 Prev->Type = TT_BinaryOperator; 201 LookForDecls = false; 202 } 203 } 204 } 205 206 if (CurrentToken->Previous->is(TT_PointerOrReference) && 207 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 208 tok::coloncolon)) 209 ProbablyFunctionType = true; 210 if (CurrentToken->is(tok::comma)) 211 MightBeFunctionType = false; 212 if (CurrentToken->Previous->is(TT_BinaryOperator)) 213 Contexts.back().IsExpression = true; 214 if (CurrentToken->is(tok::r_paren)) { 215 if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next && 216 (CurrentToken->Next->is(tok::l_paren) || 217 (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration))) 218 Left->Type = TT_FunctionTypeLParen; 219 Left->MatchingParen = CurrentToken; 220 CurrentToken->MatchingParen = Left; 221 222 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) && 223 Left->Previous && Left->Previous->is(tok::l_paren)) { 224 // Detect the case where macros are used to generate lambdas or 225 // function bodies, e.g.: 226 // auto my_lambda = MARCO((Type *type, int i) { .. body .. }); 227 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) { 228 if (Tok->is(TT_BinaryOperator) && 229 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) 230 Tok->Type = TT_PointerOrReference; 231 } 232 } 233 234 if (StartsObjCMethodExpr) { 235 CurrentToken->Type = TT_ObjCMethodExpr; 236 if (Contexts.back().FirstObjCSelectorName) { 237 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 238 Contexts.back().LongestObjCSelectorName; 239 } 240 } 241 242 if (Left->is(TT_AttributeParen)) 243 CurrentToken->Type = TT_AttributeParen; 244 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation)) 245 CurrentToken->Type = TT_JavaAnnotation; 246 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation)) 247 CurrentToken->Type = TT_LeadingJavaAnnotation; 248 249 if (!HasMultipleLines) 250 Left->PackingKind = PPK_Inconclusive; 251 else if (HasMultipleParametersOnALine) 252 Left->PackingKind = PPK_BinPacked; 253 else 254 Left->PackingKind = PPK_OnePerLine; 255 256 next(); 257 return true; 258 } 259 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 260 return false; 261 262 if (CurrentToken->is(tok::l_brace)) 263 Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen 264 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 265 !CurrentToken->Next->HasUnescapedNewline && 266 !CurrentToken->Next->isTrailingComment()) 267 HasMultipleParametersOnALine = true; 268 if (CurrentToken->isOneOf(tok::kw_const, tok::kw_auto) || 269 CurrentToken->isSimpleTypeSpecifier()) 270 Contexts.back().IsExpression = false; 271 if (CurrentToken->isOneOf(tok::semi, tok::colon)) 272 MightBeObjCForRangeLoop = false; 273 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) 274 CurrentToken->Type = TT_ObjCForIn; 275 // When we discover a 'new', we set CanBeExpression to 'false' in order to 276 // parse the type correctly. Reset that after a comma. 277 if (CurrentToken->is(tok::comma)) 278 Contexts.back().CanBeExpression = true; 279 280 FormatToken *Tok = CurrentToken; 281 if (!consumeToken()) 282 return false; 283 updateParameterCount(Left, Tok); 284 if (CurrentToken && CurrentToken->HasUnescapedNewline) 285 HasMultipleLines = true; 286 } 287 return false; 288 } 289 290 bool parseSquare() { 291 if (!CurrentToken) 292 return false; 293 294 // A '[' could be an index subscript (after an identifier or after 295 // ')' or ']'), it could be the start of an Objective-C method 296 // expression, or it could the start of an Objective-C array literal. 297 FormatToken *Left = CurrentToken->Previous; 298 Left->ParentBracket = Contexts.back().ContextKind; 299 FormatToken *Parent = Left->getPreviousNonComment(); 300 bool StartsObjCMethodExpr = 301 Style.Language == FormatStyle::LK_Cpp && 302 Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) && 303 CurrentToken->isNot(tok::l_brace) && 304 (!Parent || 305 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 306 tok::kw_return, tok::kw_throw) || 307 Parent->isUnaryOperator() || 308 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) || 309 getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown); 310 bool ColonFound = false; 311 312 unsigned BindingIncrease = 1; 313 if (Left->is(TT_Unknown)) { 314 if (StartsObjCMethodExpr) { 315 Left->Type = TT_ObjCMethodExpr; 316 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent && 317 Contexts.back().ContextKind == tok::l_brace && 318 Parent->isOneOf(tok::l_brace, tok::comma)) { 319 Left->Type = TT_JsComputedPropertyName; 320 } else if (Style.Language == FormatStyle::LK_Proto || 321 (Parent && 322 Parent->isOneOf(TT_BinaryOperator, tok::at, tok::comma, 323 tok::l_paren, tok::l_square, tok::question, 324 tok::colon, tok::kw_return, 325 // Should only be relevant to JavaScript: 326 tok::kw_default))) { 327 Left->Type = TT_ArrayInitializerLSquare; 328 } else { 329 BindingIncrease = 10; 330 Left->Type = TT_ArraySubscriptLSquare; 331 } 332 } 333 334 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease); 335 Contexts.back().IsExpression = true; 336 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr; 337 338 while (CurrentToken) { 339 if (CurrentToken->is(tok::r_square)) { 340 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_paren) && 341 Left->is(TT_ObjCMethodExpr)) { 342 // An ObjC method call is rarely followed by an open parenthesis. 343 // FIXME: Do we incorrectly label ":" with this? 344 StartsObjCMethodExpr = false; 345 Left->Type = TT_Unknown; 346 } 347 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) { 348 CurrentToken->Type = TT_ObjCMethodExpr; 349 // determineStarAmpUsage() thinks that '*' '[' is allocating an 350 // array of pointers, but if '[' starts a selector then '*' is a 351 // binary operator. 352 if (Parent && Parent->is(TT_PointerOrReference)) 353 Parent->Type = TT_BinaryOperator; 354 } 355 Left->MatchingParen = CurrentToken; 356 CurrentToken->MatchingParen = Left; 357 if (Contexts.back().FirstObjCSelectorName) { 358 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 359 Contexts.back().LongestObjCSelectorName; 360 if (Left->BlockParameterCount > 1) 361 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0; 362 } 363 next(); 364 return true; 365 } 366 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace)) 367 return false; 368 if (CurrentToken->is(tok::colon)) { 369 if (Left->is(TT_ArraySubscriptLSquare)) { 370 Left->Type = TT_ObjCMethodExpr; 371 StartsObjCMethodExpr = true; 372 Contexts.back().ColonIsObjCMethodExpr = true; 373 if (Parent && Parent->is(tok::r_paren)) 374 Parent->Type = TT_CastRParen; 375 } 376 ColonFound = true; 377 } 378 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) && 379 !ColonFound) 380 Left->Type = TT_ArrayInitializerLSquare; 381 FormatToken *Tok = CurrentToken; 382 if (!consumeToken()) 383 return false; 384 updateParameterCount(Left, Tok); 385 } 386 return false; 387 } 388 389 bool parseBrace() { 390 if (CurrentToken) { 391 FormatToken *Left = CurrentToken->Previous; 392 Left->ParentBracket = Contexts.back().ContextKind; 393 394 if (Contexts.back().CaretFound) 395 Left->Type = TT_ObjCBlockLBrace; 396 Contexts.back().CaretFound = false; 397 398 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1); 399 Contexts.back().ColonIsDictLiteral = true; 400 if (Left->BlockKind == BK_BracedInit) 401 Contexts.back().IsExpression = true; 402 403 while (CurrentToken) { 404 if (CurrentToken->is(tok::r_brace)) { 405 Left->MatchingParen = CurrentToken; 406 CurrentToken->MatchingParen = Left; 407 next(); 408 return true; 409 } 410 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square)) 411 return false; 412 updateParameterCount(Left, CurrentToken); 413 if (CurrentToken->isOneOf(tok::colon, tok::l_brace)) { 414 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 415 if (((CurrentToken->is(tok::colon) && 416 (!Contexts.back().ColonIsDictLiteral || 417 Style.Language != FormatStyle::LK_Cpp)) || 418 Style.Language == FormatStyle::LK_Proto) && 419 (Previous->Tok.getIdentifierInfo() || 420 Previous->is(tok::string_literal))) 421 Previous->Type = TT_SelectorName; 422 if (CurrentToken->is(tok::colon) || 423 Style.Language == FormatStyle::LK_JavaScript) 424 Left->Type = TT_DictLiteral; 425 } 426 if (!consumeToken()) 427 return false; 428 } 429 } 430 return true; 431 } 432 433 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 434 if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block) 435 ++Left->BlockParameterCount; 436 if (Current->is(tok::comma)) { 437 ++Left->ParameterCount; 438 if (!Left->Role) 439 Left->Role.reset(new CommaSeparatedList(Style)); 440 Left->Role->CommaFound(Current); 441 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 442 Left->ParameterCount = 1; 443 } 444 } 445 446 bool parseConditional() { 447 while (CurrentToken) { 448 if (CurrentToken->is(tok::colon)) { 449 CurrentToken->Type = TT_ConditionalExpr; 450 next(); 451 return true; 452 } 453 if (!consumeToken()) 454 return false; 455 } 456 return false; 457 } 458 459 bool parseTemplateDeclaration() { 460 if (CurrentToken && CurrentToken->is(tok::less)) { 461 CurrentToken->Type = TT_TemplateOpener; 462 next(); 463 if (!parseAngle()) 464 return false; 465 if (CurrentToken) 466 CurrentToken->Previous->ClosesTemplateDeclaration = true; 467 return true; 468 } 469 return false; 470 } 471 472 bool consumeToken() { 473 FormatToken *Tok = CurrentToken; 474 next(); 475 switch (Tok->Tok.getKind()) { 476 case tok::plus: 477 case tok::minus: 478 if (!Tok->Previous && Line.MustBeDeclaration) 479 Tok->Type = TT_ObjCMethodSpecifier; 480 break; 481 case tok::colon: 482 if (!Tok->Previous) 483 return false; 484 // Colons from ?: are handled in parseConditional(). 485 if (Style.Language == FormatStyle::LK_JavaScript) { 486 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop 487 (Contexts.size() == 1 && // switch/case labels 488 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) || 489 Contexts.back().ContextKind == tok::l_paren || // function params 490 Contexts.back().ContextKind == tok::l_square || // array type 491 (Contexts.size() == 1 && 492 Line.MustBeDeclaration)) { // method/property declaration 493 Tok->Type = TT_JsTypeColon; 494 break; 495 } 496 } 497 if (Contexts.back().ColonIsDictLiteral || 498 Style.Language == FormatStyle::LK_Proto) { 499 Tok->Type = TT_DictLiteral; 500 } else if (Contexts.back().ColonIsObjCMethodExpr || 501 Line.startsWith(TT_ObjCMethodSpecifier)) { 502 Tok->Type = TT_ObjCMethodExpr; 503 Tok->Previous->Type = TT_SelectorName; 504 if (Tok->Previous->ColumnWidth > 505 Contexts.back().LongestObjCSelectorName) 506 Contexts.back().LongestObjCSelectorName = Tok->Previous->ColumnWidth; 507 if (!Contexts.back().FirstObjCSelectorName) 508 Contexts.back().FirstObjCSelectorName = Tok->Previous; 509 } else if (Contexts.back().ColonIsForRangeExpr) { 510 Tok->Type = TT_RangeBasedForLoopColon; 511 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 512 Tok->Type = TT_BitFieldColon; 513 } else if (Contexts.size() == 1 && 514 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) { 515 if (Tok->Previous->isOneOf(tok::r_paren, tok::kw_noexcept)) 516 Tok->Type = TT_CtorInitializerColon; 517 else 518 Tok->Type = TT_InheritanceColon; 519 } else if (Tok->Previous->is(tok::identifier) && Tok->Next && 520 Tok->Next->isOneOf(tok::r_paren, tok::comma)) { 521 // This handles a special macro in ObjC code where selectors including 522 // the colon are passed as macro arguments. 523 Tok->Type = TT_ObjCMethodExpr; 524 } else if (Contexts.back().ContextKind == tok::l_paren) { 525 Tok->Type = TT_InlineASMColon; 526 } 527 break; 528 case tok::pipe: 529 case tok::amp: 530 // | and & in declarations/type expressions represent union and 531 // intersection types, respectively. 532 if (Style.Language == FormatStyle::LK_JavaScript && 533 !Contexts.back().IsExpression) 534 Tok->Type = TT_JsTypeOperator; 535 break; 536 case tok::kw_if: 537 case tok::kw_while: 538 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 539 next(); 540 if (!parseParens(/*LookForDecls=*/true)) 541 return false; 542 } 543 break; 544 case tok::kw_for: 545 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Previous && 546 Tok->Previous->is(tok::period)) 547 break; 548 Contexts.back().ColonIsForRangeExpr = true; 549 next(); 550 if (!parseParens()) 551 return false; 552 break; 553 case tok::l_paren: 554 // When faced with 'operator()()', the kw_operator handler incorrectly 555 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make 556 // the first two parens OverloadedOperators and the second l_paren an 557 // OverloadedOperatorLParen. 558 if (Tok->Previous && 559 Tok->Previous->is(tok::r_paren) && 560 Tok->Previous->MatchingParen && 561 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) { 562 Tok->Previous->Type = TT_OverloadedOperator; 563 Tok->Previous->MatchingParen->Type = TT_OverloadedOperator; 564 Tok->Type = TT_OverloadedOperatorLParen; 565 } 566 567 if (!parseParens()) 568 return false; 569 if (Line.MustBeDeclaration && Contexts.size() == 1 && 570 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) && 571 (!Tok->Previous || 572 !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute, 573 TT_LeadingJavaAnnotation))) 574 Line.MightBeFunctionDecl = true; 575 break; 576 case tok::l_square: 577 if (!parseSquare()) 578 return false; 579 break; 580 case tok::l_brace: 581 if (!parseBrace()) 582 return false; 583 break; 584 case tok::less: 585 if (parseAngle()) { 586 Tok->Type = TT_TemplateOpener; 587 } else { 588 Tok->Type = TT_BinaryOperator; 589 NonTemplateLess.insert(Tok); 590 CurrentToken = Tok; 591 next(); 592 } 593 break; 594 case tok::r_paren: 595 case tok::r_square: 596 return false; 597 case tok::r_brace: 598 // Lines can start with '}'. 599 if (Tok->Previous) 600 return false; 601 break; 602 case tok::greater: 603 Tok->Type = TT_BinaryOperator; 604 break; 605 case tok::kw_operator: 606 while (CurrentToken && 607 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 608 if (CurrentToken->isOneOf(tok::star, tok::amp)) 609 CurrentToken->Type = TT_PointerOrReference; 610 consumeToken(); 611 if (CurrentToken && 612 CurrentToken->Previous->isOneOf(TT_BinaryOperator, tok::comma)) 613 CurrentToken->Previous->Type = TT_OverloadedOperator; 614 } 615 if (CurrentToken) { 616 CurrentToken->Type = TT_OverloadedOperatorLParen; 617 if (CurrentToken->Previous->is(TT_BinaryOperator)) 618 CurrentToken->Previous->Type = TT_OverloadedOperator; 619 } 620 break; 621 case tok::question: 622 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next && 623 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren, 624 tok::r_brace)) { 625 // Question marks before semicolons, colons, etc. indicate optional 626 // types (fields, parameters), e.g. 627 // function(x?: string, y?) {...} 628 // class X { y?; } 629 Tok->Type = TT_JsTypeOptionalQuestion; 630 break; 631 } 632 // Declarations cannot be conditional expressions, this can only be part 633 // of a type declaration. 634 if (Line.MustBeDeclaration && 635 Style.Language == FormatStyle::LK_JavaScript) 636 break; 637 parseConditional(); 638 break; 639 case tok::kw_template: 640 parseTemplateDeclaration(); 641 break; 642 case tok::comma: 643 if (Contexts.back().InCtorInitializer) 644 Tok->Type = TT_CtorInitializerComma; 645 else if (Contexts.back().FirstStartOfName && 646 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) { 647 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 648 Line.IsMultiVariableDeclStmt = true; 649 } 650 if (Contexts.back().IsForEachMacro) 651 Contexts.back().IsExpression = true; 652 break; 653 default: 654 break; 655 } 656 return true; 657 } 658 659 void parseIncludeDirective() { 660 if (CurrentToken && CurrentToken->is(tok::less)) { 661 next(); 662 while (CurrentToken) { 663 if (CurrentToken->isNot(tok::comment) || CurrentToken->Next) 664 CurrentToken->Type = TT_ImplicitStringLiteral; 665 next(); 666 } 667 } 668 } 669 670 void parseWarningOrError() { 671 next(); 672 // We still want to format the whitespace left of the first token of the 673 // warning or error. 674 next(); 675 while (CurrentToken) { 676 CurrentToken->Type = TT_ImplicitStringLiteral; 677 next(); 678 } 679 } 680 681 void parsePragma() { 682 next(); // Consume "pragma". 683 if (CurrentToken && 684 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) { 685 bool IsMark = CurrentToken->is(Keywords.kw_mark); 686 next(); // Consume "mark". 687 next(); // Consume first token (so we fix leading whitespace). 688 while (CurrentToken) { 689 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator)) 690 CurrentToken->Type = TT_ImplicitStringLiteral; 691 next(); 692 } 693 } 694 } 695 696 LineType parsePreprocessorDirective() { 697 bool IsFirstToken = CurrentToken->IsFirst; 698 LineType Type = LT_PreprocessorDirective; 699 next(); 700 if (!CurrentToken) 701 return Type; 702 703 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) { 704 // JavaScript files can contain shebang lines of the form: 705 // #!/usr/bin/env node 706 // Treat these like C++ #include directives. 707 while (CurrentToken) { 708 // Tokens cannot be comments here. 709 CurrentToken->Type = TT_ImplicitStringLiteral; 710 next(); 711 } 712 return LT_ImportStatement; 713 } 714 715 if (CurrentToken->Tok.is(tok::numeric_constant)) { 716 CurrentToken->SpacesRequiredBefore = 1; 717 return Type; 718 } 719 // Hashes in the middle of a line can lead to any strange token 720 // sequence. 721 if (!CurrentToken->Tok.getIdentifierInfo()) 722 return Type; 723 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 724 case tok::pp_include: 725 case tok::pp_include_next: 726 case tok::pp_import: 727 next(); 728 parseIncludeDirective(); 729 Type = LT_ImportStatement; 730 break; 731 case tok::pp_error: 732 case tok::pp_warning: 733 parseWarningOrError(); 734 break; 735 case tok::pp_pragma: 736 parsePragma(); 737 break; 738 case tok::pp_if: 739 case tok::pp_elif: 740 Contexts.back().IsExpression = true; 741 parseLine(); 742 break; 743 default: 744 break; 745 } 746 while (CurrentToken) 747 next(); 748 return Type; 749 } 750 751 public: 752 LineType parseLine() { 753 NonTemplateLess.clear(); 754 if (CurrentToken->is(tok::hash)) 755 return parsePreprocessorDirective(); 756 757 // Directly allow to 'import <string-literal>' to support protocol buffer 758 // definitions (code.google.com/p/protobuf) or missing "#" (either way we 759 // should not break the line). 760 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 761 if ((Style.Language == FormatStyle::LK_Java && 762 CurrentToken->is(Keywords.kw_package)) || 763 (Info && Info->getPPKeywordID() == tok::pp_import && 764 CurrentToken->Next && 765 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier, 766 tok::kw_static))) { 767 next(); 768 parseIncludeDirective(); 769 return LT_ImportStatement; 770 } 771 772 // If this line starts and ends in '<' and '>', respectively, it is likely 773 // part of "#define <a/b.h>". 774 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 775 parseIncludeDirective(); 776 return LT_ImportStatement; 777 } 778 779 // In .proto files, top-level options are very similar to import statements 780 // and should not be line-wrapped. 781 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 && 782 CurrentToken->is(Keywords.kw_option)) { 783 next(); 784 if (CurrentToken && CurrentToken->is(tok::identifier)) 785 return LT_ImportStatement; 786 } 787 788 bool KeywordVirtualFound = false; 789 bool ImportStatement = false; 790 791 // import {...} from '...'; 792 if (Style.Language == FormatStyle::LK_JavaScript && 793 CurrentToken->is(Keywords.kw_import)) 794 ImportStatement = true; 795 796 while (CurrentToken) { 797 if (CurrentToken->is(tok::kw_virtual)) 798 KeywordVirtualFound = true; 799 if (Style.Language == FormatStyle::LK_JavaScript) { 800 // export {...} from '...'; 801 // An export followed by "from 'some string';" is a re-export from 802 // another module identified by a URI and is treated as a 803 // LT_ImportStatement (i.e. prevent wraps on it for long URIs). 804 // Just "export {...};" or "export class ..." should not be treated as 805 // an import in this sense. 806 if (Line.First->is(tok::kw_export) && 807 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next && 808 CurrentToken->Next->isStringLiteral()) 809 ImportStatement = true; 810 if (isClosureImportStatement(*CurrentToken)) 811 ImportStatement = true; 812 } 813 if (!consumeToken()) 814 return LT_Invalid; 815 } 816 if (KeywordVirtualFound) 817 return LT_VirtualFunctionDecl; 818 if (ImportStatement) 819 return LT_ImportStatement; 820 821 if (Line.startsWith(TT_ObjCMethodSpecifier)) { 822 if (Contexts.back().FirstObjCSelectorName) 823 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 824 Contexts.back().LongestObjCSelectorName; 825 return LT_ObjCMethodDecl; 826 } 827 828 return LT_Other; 829 } 830 831 private: 832 bool isClosureImportStatement(const FormatToken &Tok) { 833 // FIXME: Closure-library specific stuff should not be hard-coded but be 834 // configurable. 835 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) && 836 Tok.Next->Next && (Tok.Next->Next->TokenText == "module" || 837 Tok.Next->Next->TokenText == "provide" || 838 Tok.Next->Next->TokenText == "require" || 839 Tok.Next->Next->TokenText == "setTestOnly" || 840 Tok.Next->Next->TokenText == "forwardDeclare") && 841 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren); 842 } 843 844 void resetTokenMetadata(FormatToken *Token) { 845 if (!Token) 846 return; 847 848 // Reset token type in case we have already looked at it and then 849 // recovered from an error (e.g. failure to find the matching >). 850 if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro, 851 TT_FunctionLBrace, TT_ImplicitStringLiteral, 852 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow, 853 TT_RegexLiteral)) 854 CurrentToken->Type = TT_Unknown; 855 CurrentToken->Role.reset(); 856 CurrentToken->MatchingParen = nullptr; 857 CurrentToken->FakeLParens.clear(); 858 CurrentToken->FakeRParens = 0; 859 } 860 861 void next() { 862 if (CurrentToken) { 863 CurrentToken->NestingLevel = Contexts.size() - 1; 864 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 865 modifyContext(*CurrentToken); 866 determineTokenType(*CurrentToken); 867 CurrentToken = CurrentToken->Next; 868 } 869 870 resetTokenMetadata(CurrentToken); 871 } 872 873 /// \brief A struct to hold information valid in a specific context, e.g. 874 /// a pair of parenthesis. 875 struct Context { 876 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 877 bool IsExpression) 878 : ContextKind(ContextKind), BindingStrength(BindingStrength), 879 IsExpression(IsExpression) {} 880 881 tok::TokenKind ContextKind; 882 unsigned BindingStrength; 883 bool IsExpression; 884 unsigned LongestObjCSelectorName = 0; 885 bool ColonIsForRangeExpr = false; 886 bool ColonIsDictLiteral = false; 887 bool ColonIsObjCMethodExpr = false; 888 FormatToken *FirstObjCSelectorName = nullptr; 889 FormatToken *FirstStartOfName = nullptr; 890 bool CanBeExpression = true; 891 bool InTemplateArgument = false; 892 bool InCtorInitializer = false; 893 bool CaretFound = false; 894 bool IsForEachMacro = false; 895 }; 896 897 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime 898 /// of each instance. 899 struct ScopedContextCreator { 900 AnnotatingParser &P; 901 902 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 903 unsigned Increase) 904 : P(P) { 905 P.Contexts.push_back(Context(ContextKind, 906 P.Contexts.back().BindingStrength + Increase, 907 P.Contexts.back().IsExpression)); 908 } 909 910 ~ScopedContextCreator() { P.Contexts.pop_back(); } 911 }; 912 913 void modifyContext(const FormatToken &Current) { 914 if (Current.getPrecedence() == prec::Assignment && 915 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) && 916 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 917 Contexts.back().IsExpression = true; 918 if (!Line.startsWith(TT_UnaryOperator)) { 919 for (FormatToken *Previous = Current.Previous; 920 Previous && Previous->Previous && 921 !Previous->Previous->isOneOf(tok::comma, tok::semi); 922 Previous = Previous->Previous) { 923 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 924 Previous = Previous->MatchingParen; 925 if (!Previous) 926 break; 927 } 928 if (Previous->opensScope()) 929 break; 930 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) && 931 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) && 932 Previous->Previous && Previous->Previous->isNot(tok::equal)) 933 Previous->Type = TT_PointerOrReference; 934 } 935 } 936 } else if (Current.is(tok::lessless) && 937 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) { 938 Contexts.back().IsExpression = true; 939 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 940 Contexts.back().IsExpression = true; 941 } else if (Current.is(TT_TrailingReturnArrow)) { 942 Contexts.back().IsExpression = false; 943 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) { 944 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java; 945 } else if (Current.Previous && 946 Current.Previous->is(TT_CtorInitializerColon)) { 947 Contexts.back().IsExpression = true; 948 Contexts.back().InCtorInitializer = true; 949 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 950 for (FormatToken *Previous = Current.Previous; 951 Previous && Previous->isOneOf(tok::star, tok::amp); 952 Previous = Previous->Previous) 953 Previous->Type = TT_PointerOrReference; 954 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer) 955 Contexts.back().IsExpression = false; 956 } else if (Current.is(tok::kw_new)) { 957 Contexts.back().CanBeExpression = false; 958 } else if (Current.isOneOf(tok::semi, tok::exclaim)) { 959 // This should be the condition or increment in a for-loop. 960 Contexts.back().IsExpression = true; 961 } 962 } 963 964 void determineTokenType(FormatToken &Current) { 965 if (!Current.is(TT_Unknown)) 966 // The token type is already known. 967 return; 968 969 // Line.MightBeFunctionDecl can only be true after the parentheses of a 970 // function declaration have been found. In this case, 'Current' is a 971 // trailing token of this declaration and thus cannot be a name. 972 if (Current.is(Keywords.kw_instanceof)) { 973 Current.Type = TT_BinaryOperator; 974 } else if (isStartOfName(Current) && 975 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 976 Contexts.back().FirstStartOfName = &Current; 977 Current.Type = TT_StartOfName; 978 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) { 979 AutoFound = true; 980 } else if (Current.is(tok::arrow) && 981 Style.Language == FormatStyle::LK_Java) { 982 Current.Type = TT_LambdaArrow; 983 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration && 984 Current.NestingLevel == 0) { 985 Current.Type = TT_TrailingReturnArrow; 986 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 987 Current.Type = 988 determineStarAmpUsage(Current, Contexts.back().CanBeExpression && 989 Contexts.back().IsExpression, 990 Contexts.back().InTemplateArgument); 991 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 992 Current.Type = determinePlusMinusCaretUsage(Current); 993 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret)) 994 Contexts.back().CaretFound = true; 995 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 996 Current.Type = determineIncrementUsage(Current); 997 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 998 Current.Type = TT_UnaryOperator; 999 } else if (Current.is(tok::question)) { 1000 if (Style.Language == FormatStyle::LK_JavaScript && 1001 Line.MustBeDeclaration) { 1002 // In JavaScript, `interface X { foo?(): bar; }` is an optional method 1003 // on the interface, not a ternary expression. 1004 Current.Type = TT_JsTypeOptionalQuestion; 1005 } else { 1006 Current.Type = TT_ConditionalExpr; 1007 } 1008 } else if (Current.isBinaryOperator() && 1009 (!Current.Previous || Current.Previous->isNot(tok::l_square))) { 1010 Current.Type = TT_BinaryOperator; 1011 } else if (Current.is(tok::comment)) { 1012 if (Current.TokenText.startswith("/*")) { 1013 if (Current.TokenText.endswith("*/")) 1014 Current.Type = TT_BlockComment; 1015 else 1016 // The lexer has for some reason determined a comment here. But we 1017 // cannot really handle it, if it isn't properly terminated. 1018 Current.Tok.setKind(tok::unknown); 1019 } else { 1020 Current.Type = TT_LineComment; 1021 } 1022 } else if (Current.is(tok::r_paren)) { 1023 if (rParenEndsCast(Current)) 1024 Current.Type = TT_CastRParen; 1025 if (Current.MatchingParen && Current.Next && 1026 !Current.Next->isBinaryOperator() && 1027 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace, 1028 tok::period, tok::arrow, tok::coloncolon)) 1029 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) 1030 if (BeforeParen->is(tok::identifier) && 1031 BeforeParen->TokenText == BeforeParen->TokenText.upper() && 1032 (!BeforeParen->Previous || 1033 BeforeParen->Previous->ClosesTemplateDeclaration)) 1034 Current.Type = TT_FunctionAnnotationRParen; 1035 } else if (Current.is(tok::at) && Current.Next) { 1036 if (Current.Next->isStringLiteral()) { 1037 Current.Type = TT_ObjCStringLiteral; 1038 } else { 1039 switch (Current.Next->Tok.getObjCKeywordID()) { 1040 case tok::objc_interface: 1041 case tok::objc_implementation: 1042 case tok::objc_protocol: 1043 Current.Type = TT_ObjCDecl; 1044 break; 1045 case tok::objc_property: 1046 Current.Type = TT_ObjCProperty; 1047 break; 1048 default: 1049 break; 1050 } 1051 } 1052 } else if (Current.is(tok::period)) { 1053 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 1054 if (PreviousNoComment && 1055 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 1056 Current.Type = TT_DesignatedInitializerPeriod; 1057 else if (Style.Language == FormatStyle::LK_Java && Current.Previous && 1058 Current.Previous->isOneOf(TT_JavaAnnotation, 1059 TT_LeadingJavaAnnotation)) { 1060 Current.Type = Current.Previous->Type; 1061 } 1062 } else if (Current.isOneOf(tok::identifier, tok::kw_const) && 1063 Current.Previous && 1064 !Current.Previous->isOneOf(tok::equal, tok::at) && 1065 Line.MightBeFunctionDecl && Contexts.size() == 1) { 1066 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1067 // function declaration have been found. 1068 Current.Type = TT_TrailingAnnotation; 1069 } else if ((Style.Language == FormatStyle::LK_Java || 1070 Style.Language == FormatStyle::LK_JavaScript) && 1071 Current.Previous) { 1072 if (Current.Previous->is(tok::at) && 1073 Current.isNot(Keywords.kw_interface)) { 1074 const FormatToken &AtToken = *Current.Previous; 1075 const FormatToken *Previous = AtToken.getPreviousNonComment(); 1076 if (!Previous || Previous->is(TT_LeadingJavaAnnotation)) 1077 Current.Type = TT_LeadingJavaAnnotation; 1078 else 1079 Current.Type = TT_JavaAnnotation; 1080 } else if (Current.Previous->is(tok::period) && 1081 Current.Previous->isOneOf(TT_JavaAnnotation, 1082 TT_LeadingJavaAnnotation)) { 1083 Current.Type = Current.Previous->Type; 1084 } 1085 } 1086 } 1087 1088 /// \brief Take a guess at whether \p Tok starts a name of a function or 1089 /// variable declaration. 1090 /// 1091 /// This is a heuristic based on whether \p Tok is an identifier following 1092 /// something that is likely a type. 1093 bool isStartOfName(const FormatToken &Tok) { 1094 if (Tok.isNot(tok::identifier) || !Tok.Previous) 1095 return false; 1096 1097 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof)) 1098 return false; 1099 if (Style.Language == FormatStyle::LK_JavaScript && 1100 Tok.Previous->is(Keywords.kw_in)) 1101 return false; 1102 1103 // Skip "const" as it does not have an influence on whether this is a name. 1104 FormatToken *PreviousNotConst = Tok.Previous; 1105 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 1106 PreviousNotConst = PreviousNotConst->Previous; 1107 1108 if (!PreviousNotConst) 1109 return false; 1110 1111 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 1112 PreviousNotConst->Previous && 1113 PreviousNotConst->Previous->is(tok::hash); 1114 1115 if (PreviousNotConst->is(TT_TemplateCloser)) 1116 return PreviousNotConst && PreviousNotConst->MatchingParen && 1117 PreviousNotConst->MatchingParen->Previous && 1118 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) && 1119 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 1120 1121 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen && 1122 PreviousNotConst->MatchingParen->Previous && 1123 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype)) 1124 return true; 1125 1126 return (!IsPPKeyword && 1127 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) || 1128 PreviousNotConst->is(TT_PointerOrReference) || 1129 PreviousNotConst->isSimpleTypeSpecifier(); 1130 } 1131 1132 /// \brief Determine whether ')' is ending a cast. 1133 bool rParenEndsCast(const FormatToken &Tok) { 1134 // C-style casts are only used in C++ and Java. 1135 if (Style.Language != FormatStyle::LK_Cpp && 1136 Style.Language != FormatStyle::LK_Java) 1137 return false; 1138 1139 // Empty parens aren't casts and there are no casts at the end of the line. 1140 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen) 1141 return false; 1142 1143 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 1144 if (LeftOfParens) { 1145 // If there is a closing parenthesis left of the current parentheses, 1146 // look past it as these might be chained casts. 1147 if (LeftOfParens->is(tok::r_paren)) { 1148 if (!LeftOfParens->MatchingParen || 1149 !LeftOfParens->MatchingParen->Previous) 1150 return false; 1151 LeftOfParens = LeftOfParens->MatchingParen->Previous; 1152 } 1153 1154 // If there is an identifier (or with a few exceptions a keyword) right 1155 // before the parentheses, this is unlikely to be a cast. 1156 if (LeftOfParens->Tok.getIdentifierInfo() && 1157 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case, 1158 tok::kw_delete)) 1159 return false; 1160 1161 // Certain other tokens right before the parentheses are also signals that 1162 // this cannot be a cast. 1163 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator, 1164 TT_TemplateCloser, tok::ellipsis)) 1165 return false; 1166 } 1167 1168 if (Tok.Next->is(tok::question)) 1169 return false; 1170 1171 // As Java has no function types, a "(" after the ")" likely means that this 1172 // is a cast. 1173 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren)) 1174 return true; 1175 1176 // If a (non-string) literal follows, this is likely a cast. 1177 if (Tok.Next->isNot(tok::string_literal) && 1178 (Tok.Next->Tok.isLiteral() || 1179 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 1180 return true; 1181 1182 // Heuristically try to determine whether the parentheses contain a type. 1183 bool ParensAreType = 1184 !Tok.Previous || 1185 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) || 1186 Tok.Previous->isSimpleTypeSpecifier(); 1187 bool ParensCouldEndDecl = 1188 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater); 1189 if (ParensAreType && !ParensCouldEndDecl) 1190 return true; 1191 1192 // At this point, we heuristically assume that there are no casts at the 1193 // start of the line. We assume that we have found most cases where there 1194 // are by the logic above, e.g. "(void)x;". 1195 if (!LeftOfParens) 1196 return false; 1197 1198 // If the following token is an identifier or 'this', this is a cast. All 1199 // cases where this can be something else are handled above. 1200 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this)) 1201 return true; 1202 1203 if (!Tok.Next->Next) 1204 return false; 1205 1206 // If the next token after the parenthesis is a unary operator, assume 1207 // that this is cast, unless there are unexpected tokens inside the 1208 // parenthesis. 1209 bool NextIsUnary = 1210 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star); 1211 if (!NextIsUnary || Tok.Next->is(tok::plus) || 1212 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) 1213 return false; 1214 // Search for unexpected tokens. 1215 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen; 1216 Prev = Prev->Previous) { 1217 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon)) 1218 return false; 1219 } 1220 return true; 1221 } 1222 1223 /// \brief Return the type of the given token assuming it is * or &. 1224 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 1225 bool InTemplateArgument) { 1226 if (Style.Language == FormatStyle::LK_JavaScript) 1227 return TT_BinaryOperator; 1228 1229 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1230 if (!PrevToken) 1231 return TT_UnaryOperator; 1232 1233 const FormatToken *NextToken = Tok.getNextNonComment(); 1234 if (!NextToken || 1235 NextToken->isOneOf(tok::arrow, Keywords.kw_final, 1236 Keywords.kw_override) || 1237 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) 1238 return TT_PointerOrReference; 1239 1240 if (PrevToken->is(tok::coloncolon)) 1241 return TT_PointerOrReference; 1242 1243 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 1244 tok::comma, tok::semi, tok::kw_return, tok::colon, 1245 tok::equal, tok::kw_delete, tok::kw_sizeof) || 1246 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr, 1247 TT_UnaryOperator, TT_CastRParen)) 1248 return TT_UnaryOperator; 1249 1250 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare)) 1251 return TT_PointerOrReference; 1252 if (NextToken->is(tok::kw_operator) && !IsExpression) 1253 return TT_PointerOrReference; 1254 if (NextToken->isOneOf(tok::comma, tok::semi)) 1255 return TT_PointerOrReference; 1256 1257 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen && 1258 PrevToken->MatchingParen->Previous && 1259 PrevToken->MatchingParen->Previous->isOneOf(tok::kw_typeof, 1260 tok::kw_decltype)) 1261 return TT_PointerOrReference; 1262 1263 if (PrevToken->Tok.isLiteral() || 1264 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 1265 tok::kw_false, tok::r_brace) || 1266 NextToken->Tok.isLiteral() || 1267 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 1268 NextToken->isUnaryOperator() || 1269 // If we know we're in a template argument, there are no named 1270 // declarations. Thus, having an identifier on the right-hand side 1271 // indicates a binary operator. 1272 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 1273 return TT_BinaryOperator; 1274 1275 // "&&(" is quite unlikely to be two successive unary "&". 1276 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren)) 1277 return TT_BinaryOperator; 1278 1279 // This catches some cases where evaluation order is used as control flow: 1280 // aaa && aaa->f(); 1281 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 1282 if (NextNextToken && NextNextToken->is(tok::arrow)) 1283 return TT_BinaryOperator; 1284 1285 // It is very unlikely that we are going to find a pointer or reference type 1286 // definition on the RHS of an assignment. 1287 if (IsExpression && !Contexts.back().CaretFound) 1288 return TT_BinaryOperator; 1289 1290 return TT_PointerOrReference; 1291 } 1292 1293 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 1294 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1295 if (!PrevToken || PrevToken->is(TT_CastRParen)) 1296 return TT_UnaryOperator; 1297 1298 // Use heuristics to recognize unary operators. 1299 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 1300 tok::question, tok::colon, tok::kw_return, 1301 tok::kw_case, tok::at, tok::l_brace)) 1302 return TT_UnaryOperator; 1303 1304 // There can't be two consecutive binary operators. 1305 if (PrevToken->is(TT_BinaryOperator)) 1306 return TT_UnaryOperator; 1307 1308 // Fall back to marking the token as binary operator. 1309 return TT_BinaryOperator; 1310 } 1311 1312 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements. 1313 TokenType determineIncrementUsage(const FormatToken &Tok) { 1314 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1315 if (!PrevToken || PrevToken->is(TT_CastRParen)) 1316 return TT_UnaryOperator; 1317 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 1318 return TT_TrailingUnaryOperator; 1319 1320 return TT_UnaryOperator; 1321 } 1322 1323 SmallVector<Context, 8> Contexts; 1324 1325 const FormatStyle &Style; 1326 AnnotatedLine &Line; 1327 FormatToken *CurrentToken; 1328 bool AutoFound; 1329 const AdditionalKeywords &Keywords; 1330 1331 // Set of "<" tokens that do not open a template parameter list. If parseAngle 1332 // determines that a specific token can't be a template opener, it will make 1333 // same decision irrespective of the decisions for tokens leading up to it. 1334 // Store this information to prevent this from causing exponential runtime. 1335 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess; 1336 }; 1337 1338 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1; 1339 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 1340 1341 /// \brief Parses binary expressions by inserting fake parenthesis based on 1342 /// operator precedence. 1343 class ExpressionParser { 1344 public: 1345 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords, 1346 AnnotatedLine &Line) 1347 : Style(Style), Keywords(Keywords), Current(Line.First) {} 1348 1349 /// \brief Parse expressions with the given operatore precedence. 1350 void parse(int Precedence = 0) { 1351 // Skip 'return' and ObjC selector colons as they are not part of a binary 1352 // expression. 1353 while (Current && (Current->is(tok::kw_return) || 1354 (Current->is(tok::colon) && 1355 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) 1356 next(); 1357 1358 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1359 return; 1360 1361 // Conditional expressions need to be parsed separately for proper nesting. 1362 if (Precedence == prec::Conditional) { 1363 parseConditionalExpr(); 1364 return; 1365 } 1366 1367 // Parse unary operators, which all have a higher precedence than binary 1368 // operators. 1369 if (Precedence == PrecedenceUnaryOperator) { 1370 parseUnaryOperator(); 1371 return; 1372 } 1373 1374 FormatToken *Start = Current; 1375 FormatToken *LatestOperator = nullptr; 1376 unsigned OperatorIndex = 0; 1377 1378 while (Current) { 1379 // Consume operators with higher precedence. 1380 parse(Precedence + 1); 1381 1382 int CurrentPrecedence = getCurrentPrecedence(); 1383 1384 if (Current && Current->is(TT_SelectorName) && 1385 Precedence == CurrentPrecedence) { 1386 if (LatestOperator) 1387 addFakeParenthesis(Start, prec::Level(Precedence)); 1388 Start = Current; 1389 } 1390 1391 // At the end of the line or when an operator with higher precedence is 1392 // found, insert fake parenthesis and return. 1393 if (!Current || (Current->closesScope() && Current->MatchingParen) || 1394 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) || 1395 (CurrentPrecedence == prec::Conditional && 1396 Precedence == prec::Assignment && Current->is(tok::colon))) { 1397 break; 1398 } 1399 1400 // Consume scopes: (), [], <> and {} 1401 if (Current->opensScope()) { 1402 while (Current && !Current->closesScope()) { 1403 next(); 1404 parse(); 1405 } 1406 next(); 1407 } else { 1408 // Operator found. 1409 if (CurrentPrecedence == Precedence) { 1410 if (LatestOperator) 1411 LatestOperator->NextOperator = Current; 1412 LatestOperator = Current; 1413 Current->OperatorIndex = OperatorIndex; 1414 ++OperatorIndex; 1415 } 1416 next(/*SkipPastLeadingComments=*/Precedence > 0); 1417 } 1418 } 1419 1420 if (LatestOperator && (Current || Precedence > 0)) { 1421 // LatestOperator->LastOperator = true; 1422 if (Precedence == PrecedenceArrowAndPeriod) { 1423 // Call expressions don't have a binary operator precedence. 1424 addFakeParenthesis(Start, prec::Unknown); 1425 } else { 1426 addFakeParenthesis(Start, prec::Level(Precedence)); 1427 } 1428 } 1429 } 1430 1431 private: 1432 /// \brief Gets the precedence (+1) of the given token for binary operators 1433 /// and other tokens that we treat like binary operators. 1434 int getCurrentPrecedence() { 1435 if (Current) { 1436 const FormatToken *NextNonComment = Current->getNextNonComment(); 1437 if (Current->is(TT_ConditionalExpr)) 1438 return prec::Conditional; 1439 if (NextNonComment && NextNonComment->is(tok::colon) && 1440 NextNonComment->is(TT_DictLiteral)) 1441 return prec::Comma; 1442 if (Current->is(TT_LambdaArrow)) 1443 return prec::Comma; 1444 if (Current->is(TT_JsFatArrow)) 1445 return prec::Assignment; 1446 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName, 1447 TT_JsComputedPropertyName) || 1448 (Current->is(tok::comment) && NextNonComment && 1449 NextNonComment->is(TT_SelectorName))) 1450 return 0; 1451 if (Current->is(TT_RangeBasedForLoopColon)) 1452 return prec::Comma; 1453 if ((Style.Language == FormatStyle::LK_Java || 1454 Style.Language == FormatStyle::LK_JavaScript) && 1455 Current->is(Keywords.kw_instanceof)) 1456 return prec::Relational; 1457 if (Style.Language == FormatStyle::LK_JavaScript && 1458 Current->is(Keywords.kw_in)) 1459 return prec::Relational; 1460 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma)) 1461 return Current->getPrecedence(); 1462 if (Current->isOneOf(tok::period, tok::arrow)) 1463 return PrecedenceArrowAndPeriod; 1464 if (Style.Language == FormatStyle::LK_Java && 1465 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements, 1466 Keywords.kw_throws)) 1467 return 0; 1468 } 1469 return -1; 1470 } 1471 1472 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1473 Start->FakeLParens.push_back(Precedence); 1474 if (Precedence > prec::Unknown) 1475 Start->StartsBinaryExpression = true; 1476 if (Current) { 1477 FormatToken *Previous = Current->Previous; 1478 while (Previous->is(tok::comment) && Previous->Previous) 1479 Previous = Previous->Previous; 1480 ++Previous->FakeRParens; 1481 if (Precedence > prec::Unknown) 1482 Previous->EndsBinaryExpression = true; 1483 } 1484 } 1485 1486 /// \brief Parse unary operator expressions and surround them with fake 1487 /// parentheses if appropriate. 1488 void parseUnaryOperator() { 1489 if (!Current || Current->isNot(TT_UnaryOperator)) { 1490 parse(PrecedenceArrowAndPeriod); 1491 return; 1492 } 1493 1494 FormatToken *Start = Current; 1495 next(); 1496 parseUnaryOperator(); 1497 1498 // The actual precedence doesn't matter. 1499 addFakeParenthesis(Start, prec::Unknown); 1500 } 1501 1502 void parseConditionalExpr() { 1503 while (Current && Current->isTrailingComment()) { 1504 next(); 1505 } 1506 FormatToken *Start = Current; 1507 parse(prec::LogicalOr); 1508 if (!Current || !Current->is(tok::question)) 1509 return; 1510 next(); 1511 parse(prec::Assignment); 1512 if (!Current || Current->isNot(TT_ConditionalExpr)) 1513 return; 1514 next(); 1515 parse(prec::Assignment); 1516 addFakeParenthesis(Start, prec::Conditional); 1517 } 1518 1519 void next(bool SkipPastLeadingComments = true) { 1520 if (Current) 1521 Current = Current->Next; 1522 while (Current && 1523 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 1524 Current->isTrailingComment()) 1525 Current = Current->Next; 1526 } 1527 1528 const FormatStyle &Style; 1529 const AdditionalKeywords &Keywords; 1530 FormatToken *Current; 1531 }; 1532 1533 } // end anonymous namespace 1534 1535 void TokenAnnotator::setCommentLineLevels( 1536 SmallVectorImpl<AnnotatedLine *> &Lines) { 1537 const AnnotatedLine *NextNonCommentLine = nullptr; 1538 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 1539 E = Lines.rend(); 1540 I != E; ++I) { 1541 if (NextNonCommentLine && (*I)->First->is(tok::comment) && 1542 (*I)->First->Next == nullptr) 1543 (*I)->Level = NextNonCommentLine->Level; 1544 else 1545 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 1546 1547 setCommentLineLevels((*I)->Children); 1548 } 1549 } 1550 1551 void TokenAnnotator::annotate(AnnotatedLine &Line) { 1552 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1553 E = Line.Children.end(); 1554 I != E; ++I) { 1555 annotate(**I); 1556 } 1557 AnnotatingParser Parser(Style, Line, Keywords); 1558 Line.Type = Parser.parseLine(); 1559 if (Line.Type == LT_Invalid) 1560 return; 1561 1562 ExpressionParser ExprParser(Style, Keywords, Line); 1563 ExprParser.parse(); 1564 1565 if (Line.startsWith(TT_ObjCMethodSpecifier)) 1566 Line.Type = LT_ObjCMethodDecl; 1567 else if (Line.startsWith(TT_ObjCDecl)) 1568 Line.Type = LT_ObjCDecl; 1569 else if (Line.startsWith(TT_ObjCProperty)) 1570 Line.Type = LT_ObjCProperty; 1571 1572 Line.First->SpacesRequiredBefore = 1; 1573 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 1574 } 1575 1576 // This function heuristically determines whether 'Current' starts the name of a 1577 // function declaration. 1578 static bool isFunctionDeclarationName(const FormatToken &Current, 1579 const AnnotatedLine &Line) { 1580 auto skipOperatorName = [](const FormatToken* Next) -> const FormatToken* { 1581 for (; Next; Next = Next->Next) { 1582 if (Next->is(TT_OverloadedOperatorLParen)) 1583 return Next; 1584 if (Next->is(TT_OverloadedOperator)) 1585 continue; 1586 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) { 1587 // For 'new[]' and 'delete[]'. 1588 if (Next->Next && Next->Next->is(tok::l_square) && 1589 Next->Next->Next && Next->Next->Next->is(tok::r_square)) 1590 Next = Next->Next->Next; 1591 continue; 1592 } 1593 1594 break; 1595 } 1596 return nullptr; 1597 }; 1598 1599 // Find parentheses of parameter list. 1600 const FormatToken *Next = Current.Next; 1601 if (Current.is(tok::kw_operator)) { 1602 if (Current.Previous && Current.Previous->is(tok::coloncolon)) 1603 return false; 1604 Next = skipOperatorName(Next); 1605 } else { 1606 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0) 1607 return false; 1608 for (; Next; Next = Next->Next) { 1609 if (Next->is(TT_TemplateOpener)) { 1610 Next = Next->MatchingParen; 1611 } else if (Next->is(tok::coloncolon)) { 1612 Next = Next->Next; 1613 if (!Next) 1614 return false; 1615 if (Next->is(tok::kw_operator)) { 1616 Next = skipOperatorName(Next->Next); 1617 break; 1618 } 1619 if (!Next->is(tok::identifier)) 1620 return false; 1621 } else if (Next->is(tok::l_paren)) { 1622 break; 1623 } else { 1624 return false; 1625 } 1626 } 1627 } 1628 1629 // Check whether parameter list can be long to a function declaration. 1630 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen) 1631 return false; 1632 // If the lines ends with "{", this is likely an function definition. 1633 if (Line.Last->is(tok::l_brace)) 1634 return true; 1635 if (Next->Next == Next->MatchingParen) 1636 return true; // Empty parentheses. 1637 // If there is an &/&& after the r_paren, this is likely a function. 1638 if (Next->MatchingParen->Next && 1639 Next->MatchingParen->Next->is(TT_PointerOrReference)) 1640 return true; 1641 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen; 1642 Tok = Tok->Next) { 1643 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 1644 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis)) 1645 return true; 1646 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) || 1647 Tok->Tok.isLiteral()) 1648 return false; 1649 } 1650 return false; 1651 } 1652 1653 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const { 1654 assert(Line.MightBeFunctionDecl); 1655 1656 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel || 1657 Style.AlwaysBreakAfterReturnType == 1658 FormatStyle::RTBS_TopLevelDefinitions) && 1659 Line.Level > 0) 1660 return false; 1661 1662 switch (Style.AlwaysBreakAfterReturnType) { 1663 case FormatStyle::RTBS_None: 1664 return false; 1665 case FormatStyle::RTBS_All: 1666 case FormatStyle::RTBS_TopLevel: 1667 return true; 1668 case FormatStyle::RTBS_AllDefinitions: 1669 case FormatStyle::RTBS_TopLevelDefinitions: 1670 return Line.mightBeFunctionDefinition(); 1671 } 1672 1673 return false; 1674 } 1675 1676 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 1677 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1678 E = Line.Children.end(); 1679 I != E; ++I) { 1680 calculateFormattingInformation(**I); 1681 } 1682 1683 Line.First->TotalLength = 1684 Line.First->IsMultiline ? Style.ColumnLimit : Line.First->ColumnWidth; 1685 if (!Line.First->Next) 1686 return; 1687 FormatToken *Current = Line.First->Next; 1688 bool InFunctionDecl = Line.MightBeFunctionDecl; 1689 while (Current) { 1690 if (isFunctionDeclarationName(*Current, Line)) 1691 Current->Type = TT_FunctionDeclarationName; 1692 if (Current->is(TT_LineComment)) { 1693 if (Current->Previous->BlockKind == BK_BracedInit && 1694 Current->Previous->opensScope()) 1695 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1; 1696 else 1697 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 1698 1699 // If we find a trailing comment, iterate backwards to determine whether 1700 // it seems to relate to a specific parameter. If so, break before that 1701 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 1702 // to the previous line in: 1703 // SomeFunction(a, 1704 // b, // comment 1705 // c); 1706 if (!Current->HasUnescapedNewline) { 1707 for (FormatToken *Parameter = Current->Previous; Parameter; 1708 Parameter = Parameter->Previous) { 1709 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 1710 break; 1711 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 1712 if (!Parameter->Previous->is(TT_CtorInitializerComma) && 1713 Parameter->HasUnescapedNewline) 1714 Parameter->MustBreakBefore = true; 1715 break; 1716 } 1717 } 1718 } 1719 } else if (Current->SpacesRequiredBefore == 0 && 1720 spaceRequiredBefore(Line, *Current)) { 1721 Current->SpacesRequiredBefore = 1; 1722 } 1723 1724 Current->MustBreakBefore = 1725 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 1726 1727 if (!Current->MustBreakBefore && InFunctionDecl && 1728 Current->is(TT_FunctionDeclarationName)) 1729 Current->MustBreakBefore = mustBreakForReturnType(Line); 1730 1731 Current->CanBreakBefore = 1732 Current->MustBreakBefore || canBreakBefore(Line, *Current); 1733 unsigned ChildSize = 0; 1734 if (Current->Previous->Children.size() == 1) { 1735 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 1736 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 1737 : LastOfChild.TotalLength + 1; 1738 } 1739 const FormatToken *Prev = Current->Previous; 1740 if (Current->MustBreakBefore || Prev->Children.size() > 1 || 1741 (Prev->Children.size() == 1 && 1742 Prev->Children[0]->First->MustBreakBefore) || 1743 Current->IsMultiline) 1744 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit; 1745 else 1746 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth + 1747 ChildSize + Current->SpacesRequiredBefore; 1748 1749 if (Current->is(TT_CtorInitializerColon)) 1750 InFunctionDecl = false; 1751 1752 // FIXME: Only calculate this if CanBreakBefore is true once static 1753 // initializers etc. are sorted out. 1754 // FIXME: Move magic numbers to a better place. 1755 Current->SplitPenalty = 20 * Current->BindingStrength + 1756 splitPenalty(Line, *Current, InFunctionDecl); 1757 1758 Current = Current->Next; 1759 } 1760 1761 calculateUnbreakableTailLengths(Line); 1762 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 1763 if (Current->Role) 1764 Current->Role->precomputeFormattingInfos(Current); 1765 } 1766 1767 DEBUG({ printDebugInfo(Line); }); 1768 } 1769 1770 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 1771 unsigned UnbreakableTailLength = 0; 1772 FormatToken *Current = Line.Last; 1773 while (Current) { 1774 Current->UnbreakableTailLength = UnbreakableTailLength; 1775 if (Current->CanBreakBefore || 1776 Current->isOneOf(tok::comment, tok::string_literal)) { 1777 UnbreakableTailLength = 0; 1778 } else { 1779 UnbreakableTailLength += 1780 Current->ColumnWidth + Current->SpacesRequiredBefore; 1781 } 1782 Current = Current->Previous; 1783 } 1784 } 1785 1786 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 1787 const FormatToken &Tok, 1788 bool InFunctionDecl) { 1789 const FormatToken &Left = *Tok.Previous; 1790 const FormatToken &Right = Tok; 1791 1792 if (Left.is(tok::semi)) 1793 return 0; 1794 1795 if (Style.Language == FormatStyle::LK_Java) { 1796 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws)) 1797 return 1; 1798 if (Right.is(Keywords.kw_implements)) 1799 return 2; 1800 if (Left.is(tok::comma) && Left.NestingLevel == 0) 1801 return 3; 1802 } else if (Style.Language == FormatStyle::LK_JavaScript) { 1803 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma)) 1804 return 100; 1805 if (Left.is(TT_JsTypeColon)) 1806 return 35; 1807 } 1808 1809 if (Left.is(tok::comma) || (Right.is(tok::identifier) && Right.Next && 1810 Right.Next->is(TT_DictLiteral))) 1811 return 1; 1812 if (Right.is(tok::l_square)) { 1813 if (Style.Language == FormatStyle::LK_Proto) 1814 return 1; 1815 if (Left.is(tok::r_square)) 1816 return 200; 1817 // Slightly prefer formatting local lambda definitions like functions. 1818 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal)) 1819 return 35; 1820 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 1821 TT_ArrayInitializerLSquare)) 1822 return 500; 1823 } 1824 1825 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 1826 Right.is(tok::kw_operator)) { 1827 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 1828 return 3; 1829 if (Left.is(TT_StartOfName)) 1830 return 110; 1831 if (InFunctionDecl && Right.NestingLevel == 0) 1832 return Style.PenaltyReturnTypeOnItsOwnLine; 1833 return 200; 1834 } 1835 if (Right.is(TT_PointerOrReference)) 1836 return 190; 1837 if (Right.is(TT_LambdaArrow)) 1838 return 110; 1839 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 1840 return 150; 1841 if (Left.is(TT_CastRParen)) 1842 return 100; 1843 if (Left.is(tok::coloncolon) || 1844 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 1845 return 500; 1846 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 1847 return 5000; 1848 if (Left.is(tok::comment)) 1849 return 1000; 1850 1851 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon)) 1852 return 2; 1853 1854 if (Right.isMemberAccess()) { 1855 // Breaking before the "./->" of a chained call/member access is reasonably 1856 // cheap, as formatting those with one call per line is generally 1857 // desirable. In particular, it should be cheaper to break before the call 1858 // than it is to break inside a call's parameters, which could lead to weird 1859 // "hanging" indents. The exception is the very last "./->" to support this 1860 // frequent pattern: 1861 // 1862 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc( 1863 // dddddddd); 1864 // 1865 // which might otherwise be blown up onto many lines. Here, clang-format 1866 // won't produce "hanging" indents anyway as there is no other trailing 1867 // call. 1868 // 1869 // Also apply higher penalty is not a call as that might lead to a wrapping 1870 // like: 1871 // 1872 // aaaaaaa 1873 // .aaaaaaaaa.bbbbbbbb(cccccccc); 1874 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope() 1875 ? 150 1876 : 35; 1877 } 1878 1879 if (Right.is(TT_TrailingAnnotation) && 1880 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 1881 // Moving trailing annotations to the next line is fine for ObjC method 1882 // declarations. 1883 if (Line.startsWith(TT_ObjCMethodSpecifier)) 1884 return 10; 1885 // Generally, breaking before a trailing annotation is bad unless it is 1886 // function-like. It seems to be especially preferable to keep standard 1887 // annotations (i.e. "const", "final" and "override") on the same line. 1888 // Use a slightly higher penalty after ")" so that annotations like 1889 // "const override" are kept together. 1890 bool is_short_annotation = Right.TokenText.size() < 10; 1891 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 1892 } 1893 1894 // In for-loops, prefer breaking at ',' and ';'. 1895 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal)) 1896 return 4; 1897 1898 // In Objective-C method expressions, prefer breaking before "param:" over 1899 // breaking after it. 1900 if (Right.is(TT_SelectorName)) 1901 return 0; 1902 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr)) 1903 return Line.MightBeFunctionDecl ? 50 : 500; 1904 1905 if (Left.is(tok::l_paren) && InFunctionDecl && 1906 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) 1907 return 100; 1908 if (Left.is(tok::l_paren) && Left.Previous && 1909 Left.Previous->isOneOf(tok::kw_if, tok::kw_for)) 1910 return 1000; 1911 if (Left.is(tok::equal) && InFunctionDecl) 1912 return 110; 1913 if (Right.is(tok::r_brace)) 1914 return 1; 1915 if (Left.is(TT_TemplateOpener)) 1916 return 100; 1917 if (Left.opensScope()) { 1918 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign) 1919 return 0; 1920 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 1921 : 19; 1922 } 1923 if (Left.is(TT_JavaAnnotation)) 1924 return 50; 1925 1926 if (Right.is(tok::lessless)) { 1927 if (Left.is(tok::string_literal) && 1928 (Right.NextOperator || Right.OperatorIndex != 1)) { 1929 StringRef Content = Left.TokenText; 1930 if (Content.startswith("\"")) 1931 Content = Content.drop_front(1); 1932 if (Content.endswith("\"")) 1933 Content = Content.drop_back(1); 1934 Content = Content.trim(); 1935 if (Content.size() > 1 && 1936 (Content.back() == ':' || Content.back() == '=')) 1937 return 25; 1938 } 1939 return 1; // Breaking at a << is really cheap. 1940 } 1941 if (Left.is(TT_ConditionalExpr)) 1942 return prec::Conditional; 1943 prec::Level Level = Left.getPrecedence(); 1944 if (Level != prec::Unknown) 1945 return Level; 1946 Level = Right.getPrecedence(); 1947 if (Level != prec::Unknown) 1948 return Level; 1949 1950 return 3; 1951 } 1952 1953 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 1954 const FormatToken &Left, 1955 const FormatToken &Right) { 1956 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 1957 return true; 1958 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 1959 Left.Tok.getObjCKeywordID() == tok::objc_property) 1960 return true; 1961 if (Right.is(tok::hashhash)) 1962 return Left.is(tok::hash); 1963 if (Left.isOneOf(tok::hashhash, tok::hash)) 1964 return Right.is(tok::hash); 1965 if (Left.is(tok::l_paren) && Right.is(tok::r_paren)) 1966 return Style.SpaceInEmptyParentheses; 1967 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 1968 return (Right.is(TT_CastRParen) || 1969 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen))) 1970 ? Style.SpacesInCStyleCastParentheses 1971 : Style.SpacesInParentheses; 1972 if (Right.isOneOf(tok::semi, tok::comma)) 1973 return false; 1974 if (Right.is(tok::less) && 1975 (Left.is(tok::kw_template) || 1976 (Line.Type == LT_ObjCDecl && Style.ObjCSpaceBeforeProtocolList))) 1977 return true; 1978 if (Left.isOneOf(tok::exclaim, tok::tilde)) 1979 return false; 1980 if (Left.is(tok::at) && 1981 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 1982 tok::numeric_constant, tok::l_paren, tok::l_brace, 1983 tok::kw_true, tok::kw_false)) 1984 return false; 1985 if (Left.is(tok::colon)) 1986 return !Left.is(TT_ObjCMethodExpr); 1987 if (Left.is(tok::coloncolon)) 1988 return false; 1989 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) 1990 return false; 1991 if (Right.is(tok::ellipsis)) 1992 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous && 1993 Left.Previous->is(tok::kw_case)); 1994 if (Left.is(tok::l_square) && Right.is(tok::amp)) 1995 return false; 1996 if (Right.is(TT_PointerOrReference)) 1997 return (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) || 1998 (Left.Tok.isLiteral() || (Left.is(tok::kw_const) && Left.Previous && 1999 Left.Previous->is(tok::r_paren)) || 2000 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) && 2001 (Style.PointerAlignment != FormatStyle::PAS_Left || 2002 Line.IsMultiVariableDeclStmt))); 2003 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) && 2004 (!Left.is(TT_PointerOrReference) || 2005 (Style.PointerAlignment != FormatStyle::PAS_Right && 2006 !Line.IsMultiVariableDeclStmt))) 2007 return true; 2008 if (Left.is(TT_PointerOrReference)) 2009 return Right.Tok.isLiteral() || 2010 Right.isOneOf(TT_BlockComment, Keywords.kw_final, 2011 Keywords.kw_override) || 2012 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) || 2013 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare, 2014 tok::l_paren) && 2015 (Style.PointerAlignment != FormatStyle::PAS_Right && 2016 !Line.IsMultiVariableDeclStmt) && 2017 Left.Previous && 2018 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon)); 2019 if (Right.is(tok::star) && Left.is(tok::l_paren)) 2020 return false; 2021 if (Left.is(tok::l_square)) 2022 return (Left.is(TT_ArrayInitializerLSquare) && 2023 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square)) || 2024 (Left.is(TT_ArraySubscriptLSquare) && Style.SpacesInSquareBrackets && 2025 Right.isNot(tok::r_square)); 2026 if (Right.is(tok::r_square)) 2027 return Right.MatchingParen && 2028 ((Style.SpacesInContainerLiterals && 2029 Right.MatchingParen->is(TT_ArrayInitializerLSquare)) || 2030 (Style.SpacesInSquareBrackets && 2031 Right.MatchingParen->is(TT_ArraySubscriptLSquare))); 2032 if (Right.is(tok::l_square) && 2033 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare) && 2034 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral)) 2035 return false; 2036 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 2037 return !Left.Children.empty(); // No spaces in "{}". 2038 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 2039 (Right.is(tok::r_brace) && Right.MatchingParen && 2040 Right.MatchingParen->BlockKind != BK_Block)) 2041 return !Style.Cpp11BracedListStyle; 2042 if (Left.is(TT_BlockComment)) 2043 return !Left.TokenText.endswith("=*/"); 2044 if (Right.is(tok::l_paren)) { 2045 if (Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) 2046 return true; 2047 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) || 2048 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 2049 (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while, 2050 tok::kw_switch, tok::kw_case, TT_ForEachMacro, 2051 TT_ObjCForIn) || 2052 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch, 2053 tok::kw_new, tok::kw_delete) && 2054 (!Left.Previous || Left.Previous->isNot(tok::period))))) || 2055 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always && 2056 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() || 2057 Left.is(tok::r_paren)) && 2058 Line.Type != LT_PreprocessorDirective); 2059 } 2060 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 2061 return false; 2062 if (Right.is(TT_UnaryOperator)) 2063 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 2064 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr)); 2065 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 2066 tok::r_paren) || 2067 Left.isSimpleTypeSpecifier()) && 2068 Right.is(tok::l_brace) && Right.getNextNonComment() && 2069 Right.BlockKind != BK_Block) 2070 return false; 2071 if (Left.is(tok::period) || Right.is(tok::period)) 2072 return false; 2073 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 2074 return false; 2075 if (Left.is(TT_TemplateCloser) && Left.MatchingParen && 2076 Left.MatchingParen->Previous && 2077 Left.MatchingParen->Previous->is(tok::period)) 2078 // A.<B>DoSomething(); 2079 return false; 2080 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square)) 2081 return false; 2082 return true; 2083 } 2084 2085 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 2086 const FormatToken &Right) { 2087 const FormatToken &Left = *Right.Previous; 2088 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo()) 2089 return true; // Never ever merge two identifiers. 2090 if (Style.Language == FormatStyle::LK_Cpp) { 2091 if (Left.is(tok::kw_operator)) 2092 return Right.is(tok::coloncolon); 2093 } else if (Style.Language == FormatStyle::LK_Proto) { 2094 if (Right.is(tok::period) && 2095 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required, 2096 Keywords.kw_repeated, Keywords.kw_extend)) 2097 return true; 2098 if (Right.is(tok::l_paren) && 2099 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) 2100 return true; 2101 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2102 if (Left.is(TT_JsFatArrow)) 2103 return true; 2104 if (Right.is(tok::star) && 2105 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) 2106 return false; 2107 if (Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in, 2108 Keywords.kw_of, tok::kw_const) && 2109 (!Left.Previous || !Left.Previous->is(tok::period))) 2110 return true; 2111 if (Left.is(tok::kw_default) && Left.Previous && 2112 Left.Previous->is(tok::kw_export)) 2113 return true; 2114 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace)) 2115 return true; 2116 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion)) 2117 return false; 2118 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator)) 2119 return false; 2120 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) && 2121 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) 2122 return false; 2123 if (Left.is(tok::ellipsis)) 2124 return false; 2125 if (Left.is(TT_TemplateCloser) && 2126 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square, 2127 Keywords.kw_implements, Keywords.kw_extends)) 2128 // Type assertions ('<type>expr') are not followed by whitespace. Other 2129 // locations that should have whitespace following are identified by the 2130 // above set of follower tokens. 2131 return false; 2132 // Postfix non-null assertion operator, as in `foo!.bar()`. 2133 if (Right.is(tok::exclaim) && (Left.isOneOf(tok::identifier, tok::r_paren, 2134 tok::r_square, tok::r_brace) || 2135 Left.Tok.isLiteral())) 2136 return false; 2137 } else if (Style.Language == FormatStyle::LK_Java) { 2138 if (Left.is(tok::r_square) && Right.is(tok::l_brace)) 2139 return true; 2140 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) 2141 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never; 2142 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private, 2143 tok::kw_protected) || 2144 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract, 2145 Keywords.kw_native)) && 2146 Right.is(TT_TemplateOpener)) 2147 return true; 2148 } 2149 if (Left.is(TT_ImplicitStringLiteral)) 2150 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2151 if (Line.Type == LT_ObjCMethodDecl) { 2152 if (Left.is(TT_ObjCMethodSpecifier)) 2153 return true; 2154 if (Left.is(tok::r_paren) && Right.is(tok::identifier)) 2155 // Don't space between ')' and <id> 2156 return false; 2157 } 2158 if (Line.Type == LT_ObjCProperty && 2159 (Right.is(tok::equal) || Left.is(tok::equal))) 2160 return false; 2161 2162 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) || 2163 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) 2164 return true; 2165 if (Right.is(TT_OverloadedOperatorLParen)) 2166 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2167 if (Left.is(tok::comma)) 2168 return true; 2169 if (Right.is(tok::comma)) 2170 return false; 2171 if (Right.isOneOf(TT_CtorInitializerColon, TT_ObjCBlockLParen)) 2172 return true; 2173 if (Right.is(tok::colon)) { 2174 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) || 2175 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi)) 2176 return false; 2177 if (Right.is(TT_ObjCMethodExpr)) 2178 return false; 2179 if (Left.is(tok::question)) 2180 return false; 2181 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon)) 2182 return false; 2183 if (Right.is(TT_DictLiteral)) 2184 return Style.SpacesInContainerLiterals; 2185 return true; 2186 } 2187 if (Left.is(TT_UnaryOperator)) 2188 return Right.is(TT_BinaryOperator); 2189 2190 // If the next token is a binary operator or a selector name, we have 2191 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly. 2192 if (Left.is(TT_CastRParen)) 2193 return Style.SpaceAfterCStyleCast || 2194 Right.isOneOf(TT_BinaryOperator, TT_SelectorName); 2195 2196 if (Left.is(tok::greater) && Right.is(tok::greater)) 2197 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) && 2198 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 2199 if (Right.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) || 2200 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar)) 2201 return false; 2202 if (!Style.SpaceBeforeAssignmentOperators && 2203 Right.getPrecedence() == prec::Assignment) 2204 return false; 2205 if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment)) 2206 return (Left.is(TT_TemplateOpener) && 2207 Style.Standard == FormatStyle::LS_Cpp03) || 2208 !(Left.isOneOf(tok::identifier, tok::l_paren, tok::r_paren, 2209 tok::l_square) || 2210 Left.isOneOf(TT_TemplateCloser, TT_TemplateOpener)); 2211 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser))) 2212 return Style.SpacesInAngles; 2213 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) || 2214 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && 2215 !Right.is(tok::r_paren))) 2216 return true; 2217 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) && 2218 Right.isNot(TT_FunctionTypeLParen)) 2219 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2220 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) && 2221 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen)) 2222 return false; 2223 if (Right.is(tok::less) && Left.isNot(tok::l_paren) && 2224 Line.startsWith(tok::hash)) 2225 return true; 2226 if (Right.is(TT_TrailingUnaryOperator)) 2227 return false; 2228 if (Left.is(TT_RegexLiteral)) 2229 return false; 2230 return spaceRequiredBetween(Line, Left, Right); 2231 } 2232 2233 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 2234 static bool isAllmanBrace(const FormatToken &Tok) { 2235 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 2236 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral); 2237 } 2238 2239 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 2240 const FormatToken &Right) { 2241 const FormatToken &Left = *Right.Previous; 2242 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0) 2243 return true; 2244 2245 if (Style.Language == FormatStyle::LK_JavaScript) { 2246 // FIXME: This might apply to other languages and token kinds. 2247 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous && 2248 Left.Previous->is(tok::string_literal)) 2249 return true; 2250 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 && 2251 Left.Previous && Left.Previous->is(tok::equal) && 2252 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export, 2253 tok::kw_const) && 2254 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match 2255 // above. 2256 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) 2257 // Object literals on the top level of a file are treated as "enum-style". 2258 // Each key/value pair is put on a separate line, instead of bin-packing. 2259 return true; 2260 if (Left.is(tok::l_brace) && Line.Level == 0 && 2261 (Line.startsWith(tok::kw_enum) || 2262 Line.startsWith(tok::kw_export, tok::kw_enum))) 2263 // JavaScript top-level enum key/value pairs are put on separate lines 2264 // instead of bin-packing. 2265 return true; 2266 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && 2267 !Left.Children.empty()) 2268 // Support AllowShortFunctionsOnASingleLine for JavaScript. 2269 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 2270 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty || 2271 (Left.NestingLevel == 0 && Line.Level == 0 && 2272 Style.AllowShortFunctionsOnASingleLine == 2273 FormatStyle::SFS_Inline); 2274 } else if (Style.Language == FormatStyle::LK_Java) { 2275 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next && 2276 Right.Next->is(tok::string_literal)) 2277 return true; 2278 } 2279 2280 // If the last token before a '}' is a comma or a trailing comment, the 2281 // intention is to insert a line break after it in order to make shuffling 2282 // around entries easier. 2283 const FormatToken *BeforeClosingBrace = nullptr; 2284 if (Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) && 2285 Left.BlockKind != BK_Block && Left.MatchingParen) 2286 BeforeClosingBrace = Left.MatchingParen->Previous; 2287 else if (Right.MatchingParen && 2288 Right.MatchingParen->isOneOf(tok::l_brace, 2289 TT_ArrayInitializerLSquare)) 2290 BeforeClosingBrace = &Left; 2291 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 2292 BeforeClosingBrace->isTrailingComment())) 2293 return true; 2294 2295 if (Right.is(tok::comment)) 2296 return Left.BlockKind != BK_BracedInit && 2297 Left.isNot(TT_CtorInitializerColon) && 2298 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 2299 if (Left.isTrailingComment()) 2300 return true; 2301 if (Left.isStringLiteral() && 2302 (Right.isStringLiteral() || Right.is(TT_ObjCStringLiteral))) 2303 return true; 2304 if (Right.Previous->IsUnterminatedLiteral) 2305 return true; 2306 if (Right.is(tok::lessless) && Right.Next && 2307 Right.Previous->is(tok::string_literal) && 2308 Right.Next->is(tok::string_literal)) 2309 return true; 2310 if (Right.Previous->ClosesTemplateDeclaration && 2311 Right.Previous->MatchingParen && 2312 Right.Previous->MatchingParen->NestingLevel == 0 && 2313 Style.AlwaysBreakTemplateDeclarations) 2314 return true; 2315 if ((Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon)) && 2316 Style.BreakConstructorInitializersBeforeComma && 2317 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 2318 return true; 2319 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) 2320 // Raw string literals are special wrt. line breaks. The author has made a 2321 // deliberate choice and might have aligned the contents of the string 2322 // literal accordingly. Thus, we try keep existing line breaks. 2323 return Right.NewlinesBefore > 0; 2324 if (Right.Previous->is(tok::l_brace) && Right.NestingLevel == 1 && 2325 Style.Language == FormatStyle::LK_Proto) 2326 // Don't put enums onto single lines in protocol buffers. 2327 return true; 2328 if (Right.is(TT_InlineASMBrace)) 2329 return Right.HasUnescapedNewline; 2330 if (isAllmanBrace(Left) || isAllmanBrace(Right)) 2331 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) || 2332 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) || 2333 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct); 2334 if (Style.Language == FormatStyle::LK_Proto && Left.isNot(tok::l_brace) && 2335 Right.is(TT_SelectorName)) 2336 return true; 2337 if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine) 2338 return true; 2339 2340 if ((Style.Language == FormatStyle::LK_Java || 2341 Style.Language == FormatStyle::LK_JavaScript) && 2342 Left.is(TT_LeadingJavaAnnotation) && 2343 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) && 2344 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) 2345 return true; 2346 2347 return false; 2348 } 2349 2350 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 2351 const FormatToken &Right) { 2352 const FormatToken &Left = *Right.Previous; 2353 2354 // Language-specific stuff. 2355 if (Style.Language == FormatStyle::LK_Java) { 2356 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2357 Keywords.kw_implements)) 2358 return false; 2359 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2360 Keywords.kw_implements)) 2361 return true; 2362 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2363 if (Left.is(tok::kw_return)) 2364 return false; // Otherwise a semicolon is inserted. 2365 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace)) 2366 return false; 2367 if (Left.is(TT_JsTypeColon)) 2368 return true; 2369 if (Right.NestingLevel == 0 && Right.is(Keywords.kw_is)) 2370 return false; 2371 if (Left.is(Keywords.kw_in)) 2372 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None; 2373 if (Right.is(Keywords.kw_in)) 2374 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None; 2375 } 2376 2377 if (Left.is(tok::at)) 2378 return false; 2379 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 2380 return false; 2381 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation)) 2382 return !Right.is(tok::l_paren); 2383 if (Right.is(TT_PointerOrReference)) 2384 return Line.IsMultiVariableDeclStmt || 2385 (Style.PointerAlignment == FormatStyle::PAS_Right && 2386 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName))); 2387 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2388 Right.is(tok::kw_operator)) 2389 return true; 2390 if (Left.is(TT_PointerOrReference)) 2391 return false; 2392 if (Right.isTrailingComment()) 2393 // We rely on MustBreakBefore being set correctly here as we should not 2394 // change the "binding" behavior of a comment. 2395 // The first comment in a braced lists is always interpreted as belonging to 2396 // the first list element. Otherwise, it should be placed outside of the 2397 // list. 2398 return Left.BlockKind == BK_BracedInit; 2399 if (Left.is(tok::question) && Right.is(tok::colon)) 2400 return false; 2401 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question)) 2402 return Style.BreakBeforeTernaryOperators; 2403 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question)) 2404 return !Style.BreakBeforeTernaryOperators; 2405 if (Right.is(TT_InheritanceColon)) 2406 return true; 2407 if (Right.is(tok::colon) && 2408 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) 2409 return false; 2410 if (Left.is(tok::colon) && (Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr))) 2411 return true; 2412 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next && 2413 Right.Next->is(TT_ObjCMethodExpr))) 2414 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls. 2415 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 2416 return true; 2417 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen)) 2418 return true; 2419 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen, 2420 TT_OverloadedOperator)) 2421 return false; 2422 if (Left.is(TT_RangeBasedForLoopColon)) 2423 return true; 2424 if (Right.is(TT_RangeBasedForLoopColon)) 2425 return false; 2426 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) || 2427 Left.is(tok::kw_operator)) 2428 return false; 2429 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) && 2430 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) 2431 return false; 2432 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen)) 2433 return false; 2434 if (Left.is(tok::l_paren) && Left.Previous && 2435 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) 2436 return false; 2437 if (Right.is(TT_ImplicitStringLiteral)) 2438 return false; 2439 2440 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser)) 2441 return false; 2442 if (Right.is(tok::r_square) && Right.MatchingParen && 2443 Right.MatchingParen->is(TT_LambdaLSquare)) 2444 return false; 2445 2446 // We only break before r_brace if there was a corresponding break before 2447 // the l_brace, which is tracked by BreakBeforeClosingBrace. 2448 if (Right.is(tok::r_brace)) 2449 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 2450 2451 // Allow breaking after a trailing annotation, e.g. after a method 2452 // declaration. 2453 if (Left.is(TT_TrailingAnnotation)) 2454 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 2455 tok::less, tok::coloncolon); 2456 2457 if (Right.is(tok::kw___attribute)) 2458 return true; 2459 2460 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 2461 return true; 2462 2463 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2464 return true; 2465 2466 if (Left.is(TT_CtorInitializerComma) && 2467 Style.BreakConstructorInitializersBeforeComma) 2468 return false; 2469 if (Right.is(TT_CtorInitializerComma) && 2470 Style.BreakConstructorInitializersBeforeComma) 2471 return true; 2472 if ((Left.is(tok::greater) && Right.is(tok::greater)) || 2473 (Left.is(tok::less) && Right.is(tok::less))) 2474 return false; 2475 if (Right.is(TT_BinaryOperator) && 2476 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 2477 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 2478 Right.getPrecedence() != prec::Assignment)) 2479 return true; 2480 if (Left.is(TT_ArrayInitializerLSquare)) 2481 return true; 2482 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 2483 return true; 2484 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) && 2485 !Left.isOneOf(tok::arrowstar, tok::lessless) && 2486 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 2487 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 2488 Left.getPrecedence() == prec::Assignment)) 2489 return true; 2490 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 2491 tok::kw_class, tok::kw_struct, tok::comment) || 2492 Right.isMemberAccess() || 2493 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless, 2494 tok::colon, tok::l_square, tok::at) || 2495 (Left.is(tok::r_paren) && 2496 Right.isOneOf(tok::identifier, tok::kw_const)) || 2497 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)); 2498 } 2499 2500 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 2501 llvm::errs() << "AnnotatedTokens:\n"; 2502 const FormatToken *Tok = Line.First; 2503 while (Tok) { 2504 llvm::errs() << " M=" << Tok->MustBreakBefore 2505 << " C=" << Tok->CanBreakBefore 2506 << " T=" << getTokenTypeName(Tok->Type) 2507 << " S=" << Tok->SpacesRequiredBefore 2508 << " B=" << Tok->BlockParameterCount 2509 << " P=" << Tok->SplitPenalty << " Name=" << Tok->Tok.getName() 2510 << " L=" << Tok->TotalLength << " PPK=" << Tok->PackingKind 2511 << " FakeLParens="; 2512 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 2513 llvm::errs() << Tok->FakeLParens[i] << "/"; 2514 llvm::errs() << " FakeRParens=" << Tok->FakeRParens << "\n"; 2515 if (!Tok->Next) 2516 assert(Tok == Line.Last); 2517 Tok = Tok->Next; 2518 } 2519 llvm::errs() << "----\n"; 2520 } 2521 2522 } // namespace format 2523 } // namespace clang 2524