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 (Style.Language == FormatStyle::LK_JavaScript && 138 Line.startsWith(Keywords.kw_type, tok::identifier)) { 139 // type X = (...); 140 Contexts.back().IsExpression = false; 141 } else if (Left->Previous && 142 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype, 143 tok::kw_if, tok::kw_while, tok::l_paren, 144 tok::comma) || 145 Left->Previous->is(TT_BinaryOperator))) { 146 // static_assert, if and while usually contain expressions. 147 Contexts.back().IsExpression = true; 148 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 149 (Left->Previous->is(Keywords.kw_function) || 150 (Left->Previous->endsSequence(tok::identifier, 151 Keywords.kw_function)))) { 152 // function(...) or function f(...) 153 Contexts.back().IsExpression = false; 154 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 155 Left->Previous->is(TT_JsTypeColon)) { 156 // let x: (SomeType); 157 Contexts.back().IsExpression = false; 158 } else if (Left->Previous && Left->Previous->is(tok::r_square) && 159 Left->Previous->MatchingParen && 160 Left->Previous->MatchingParen->is(TT_LambdaLSquare)) { 161 // This is a parameter list of a lambda expression. 162 Contexts.back().IsExpression = false; 163 } else if (Line.InPPDirective && 164 (!Left->Previous || !Left->Previous->is(tok::identifier))) { 165 Contexts.back().IsExpression = true; 166 } else if (Contexts[Contexts.size() - 2].CaretFound) { 167 // This is the parameter list of an ObjC block. 168 Contexts.back().IsExpression = false; 169 } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) { 170 Left->Type = TT_AttributeParen; 171 } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) { 172 // The first argument to a foreach macro is a declaration. 173 Contexts.back().IsForEachMacro = true; 174 Contexts.back().IsExpression = false; 175 } else if (Left->Previous && Left->Previous->MatchingParen && 176 Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) { 177 Contexts.back().IsExpression = false; 178 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) { 179 bool IsForOrCatch = 180 Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch); 181 Contexts.back().IsExpression = !IsForOrCatch; 182 } 183 184 if (StartsObjCMethodExpr) { 185 Contexts.back().ColonIsObjCMethodExpr = true; 186 Left->Type = TT_ObjCMethodExpr; 187 } 188 189 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression; 190 bool ProbablyFunctionType = CurrentToken->isOneOf(tok::star, tok::amp); 191 bool HasMultipleLines = false; 192 bool HasMultipleParametersOnALine = false; 193 bool MightBeObjCForRangeLoop = 194 Left->Previous && Left->Previous->is(tok::kw_for); 195 while (CurrentToken) { 196 // LookForDecls is set when "if (" has been seen. Check for 197 // 'identifier' '*' 'identifier' followed by not '=' -- this 198 // '*' has to be a binary operator but determineStarAmpUsage() will 199 // categorize it as an unary operator, so set the right type here. 200 if (LookForDecls && CurrentToken->Next) { 201 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 202 if (Prev) { 203 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 204 FormatToken *Next = CurrentToken->Next; 205 if (PrevPrev && PrevPrev->is(tok::identifier) && 206 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 207 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 208 Prev->Type = TT_BinaryOperator; 209 LookForDecls = false; 210 } 211 } 212 } 213 214 if (CurrentToken->Previous->is(TT_PointerOrReference) && 215 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 216 tok::coloncolon)) 217 ProbablyFunctionType = true; 218 if (CurrentToken->is(tok::comma)) 219 MightBeFunctionType = false; 220 if (CurrentToken->Previous->is(TT_BinaryOperator)) 221 Contexts.back().IsExpression = true; 222 if (CurrentToken->is(tok::r_paren)) { 223 if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next && 224 (CurrentToken->Next->is(tok::l_paren) || 225 (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration))) 226 Left->Type = TT_FunctionTypeLParen; 227 Left->MatchingParen = CurrentToken; 228 CurrentToken->MatchingParen = Left; 229 230 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) && 231 Left->Previous && Left->Previous->is(tok::l_paren)) { 232 // Detect the case where macros are used to generate lambdas or 233 // function bodies, e.g.: 234 // auto my_lambda = MARCO((Type *type, int i) { .. body .. }); 235 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) { 236 if (Tok->is(TT_BinaryOperator) && 237 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) 238 Tok->Type = TT_PointerOrReference; 239 } 240 } 241 242 if (StartsObjCMethodExpr) { 243 CurrentToken->Type = TT_ObjCMethodExpr; 244 if (Contexts.back().FirstObjCSelectorName) { 245 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 246 Contexts.back().LongestObjCSelectorName; 247 } 248 } 249 250 if (Left->is(TT_AttributeParen)) 251 CurrentToken->Type = TT_AttributeParen; 252 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation)) 253 CurrentToken->Type = TT_JavaAnnotation; 254 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation)) 255 CurrentToken->Type = TT_LeadingJavaAnnotation; 256 257 if (!HasMultipleLines) 258 Left->PackingKind = PPK_Inconclusive; 259 else if (HasMultipleParametersOnALine) 260 Left->PackingKind = PPK_BinPacked; 261 else 262 Left->PackingKind = PPK_OnePerLine; 263 264 next(); 265 return true; 266 } 267 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 268 return false; 269 270 if (CurrentToken->is(tok::l_brace)) 271 Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen 272 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 273 !CurrentToken->Next->HasUnescapedNewline && 274 !CurrentToken->Next->isTrailingComment()) 275 HasMultipleParametersOnALine = true; 276 if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) || 277 CurrentToken->Previous->isSimpleTypeSpecifier()) && 278 !CurrentToken->is(tok::l_brace)) 279 Contexts.back().IsExpression = false; 280 if (CurrentToken->isOneOf(tok::semi, tok::colon)) 281 MightBeObjCForRangeLoop = false; 282 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) 283 CurrentToken->Type = TT_ObjCForIn; 284 // When we discover a 'new', we set CanBeExpression to 'false' in order to 285 // parse the type correctly. Reset that after a comma. 286 if (CurrentToken->is(tok::comma)) 287 Contexts.back().CanBeExpression = true; 288 289 FormatToken *Tok = CurrentToken; 290 if (!consumeToken()) 291 return false; 292 updateParameterCount(Left, Tok); 293 if (CurrentToken && CurrentToken->HasUnescapedNewline) 294 HasMultipleLines = true; 295 } 296 return false; 297 } 298 299 bool parseSquare() { 300 if (!CurrentToken) 301 return false; 302 303 // A '[' could be an index subscript (after an identifier or after 304 // ')' or ']'), it could be the start of an Objective-C method 305 // expression, or it could the start of an Objective-C array literal. 306 FormatToken *Left = CurrentToken->Previous; 307 Left->ParentBracket = Contexts.back().ContextKind; 308 FormatToken *Parent = Left->getPreviousNonComment(); 309 310 // Cases where '>' is followed by '['. 311 // In C++, this can happen either in array of templates (foo<int>[10]) 312 // or when array is a nested template type (unique_ptr<type1<type2>[]>). 313 bool CppArrayTemplates = 314 Style.Language == FormatStyle::LK_Cpp && Parent && 315 Parent->is(TT_TemplateCloser) && 316 (Contexts.back().CanBeExpression || Contexts.back().IsExpression || 317 Contexts.back().InTemplateArgument); 318 319 bool StartsObjCMethodExpr = 320 !CppArrayTemplates && (Style.Language == FormatStyle::LK_Cpp || 321 Style.Language == FormatStyle::LK_ObjC) && 322 Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) && 323 CurrentToken->isNot(tok::l_brace) && 324 (!Parent || 325 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 326 tok::kw_return, tok::kw_throw) || 327 Parent->isUnaryOperator() || 328 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) || 329 getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown); 330 bool ColonFound = false; 331 332 unsigned BindingIncrease = 1; 333 if (Left->is(TT_Unknown)) { 334 if (StartsObjCMethodExpr) { 335 Left->Type = TT_ObjCMethodExpr; 336 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent && 337 Contexts.back().ContextKind == tok::l_brace && 338 Parent->isOneOf(tok::l_brace, tok::comma)) { 339 Left->Type = TT_JsComputedPropertyName; 340 } else if (Style.Language == FormatStyle::LK_Proto || 341 (!CppArrayTemplates && Parent && 342 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at, 343 tok::comma, tok::l_paren, tok::l_square, 344 tok::question, tok::colon, tok::kw_return, 345 // Should only be relevant to JavaScript: 346 tok::kw_default))) { 347 Left->Type = TT_ArrayInitializerLSquare; 348 } else { 349 BindingIncrease = 10; 350 Left->Type = TT_ArraySubscriptLSquare; 351 } 352 } 353 354 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease); 355 Contexts.back().IsExpression = true; 356 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr; 357 358 while (CurrentToken) { 359 if (CurrentToken->is(tok::r_square)) { 360 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_paren) && 361 Left->is(TT_ObjCMethodExpr)) { 362 // An ObjC method call is rarely followed by an open parenthesis. 363 // FIXME: Do we incorrectly label ":" with this? 364 StartsObjCMethodExpr = false; 365 Left->Type = TT_Unknown; 366 } 367 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) { 368 CurrentToken->Type = TT_ObjCMethodExpr; 369 // determineStarAmpUsage() thinks that '*' '[' is allocating an 370 // array of pointers, but if '[' starts a selector then '*' is a 371 // binary operator. 372 if (Parent && Parent->is(TT_PointerOrReference)) 373 Parent->Type = TT_BinaryOperator; 374 } 375 Left->MatchingParen = CurrentToken; 376 CurrentToken->MatchingParen = Left; 377 if (Contexts.back().FirstObjCSelectorName) { 378 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 379 Contexts.back().LongestObjCSelectorName; 380 if (Left->BlockParameterCount > 1) 381 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0; 382 } 383 next(); 384 return true; 385 } 386 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace)) 387 return false; 388 if (CurrentToken->is(tok::colon)) { 389 if (Left->is(TT_ArraySubscriptLSquare)) { 390 Left->Type = TT_ObjCMethodExpr; 391 StartsObjCMethodExpr = true; 392 Contexts.back().ColonIsObjCMethodExpr = true; 393 if (Parent && Parent->is(tok::r_paren)) 394 Parent->Type = TT_CastRParen; 395 } 396 ColonFound = true; 397 } 398 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) && 399 !ColonFound) 400 Left->Type = TT_ArrayInitializerLSquare; 401 FormatToken *Tok = CurrentToken; 402 if (!consumeToken()) 403 return false; 404 updateParameterCount(Left, Tok); 405 } 406 return false; 407 } 408 409 bool parseBrace() { 410 if (CurrentToken) { 411 FormatToken *Left = CurrentToken->Previous; 412 Left->ParentBracket = Contexts.back().ContextKind; 413 414 if (Contexts.back().CaretFound) 415 Left->Type = TT_ObjCBlockLBrace; 416 Contexts.back().CaretFound = false; 417 418 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1); 419 Contexts.back().ColonIsDictLiteral = true; 420 if (Left->BlockKind == BK_BracedInit) 421 Contexts.back().IsExpression = true; 422 423 while (CurrentToken) { 424 if (CurrentToken->is(tok::r_brace)) { 425 Left->MatchingParen = CurrentToken; 426 CurrentToken->MatchingParen = Left; 427 next(); 428 return true; 429 } 430 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square)) 431 return false; 432 updateParameterCount(Left, CurrentToken); 433 if (CurrentToken->isOneOf(tok::colon, tok::l_brace)) { 434 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 435 if (((CurrentToken->is(tok::colon) && 436 (!Contexts.back().ColonIsDictLiteral || 437 (Style.Language != FormatStyle::LK_Cpp && 438 Style.Language != FormatStyle::LK_ObjC))) || 439 Style.Language == FormatStyle::LK_Proto) && 440 (Previous->Tok.getIdentifierInfo() || 441 Previous->is(tok::string_literal))) 442 Previous->Type = TT_SelectorName; 443 if (CurrentToken->is(tok::colon) || 444 Style.Language == FormatStyle::LK_JavaScript) 445 Left->Type = TT_DictLiteral; 446 } 447 if (CurrentToken->is(tok::comma) && 448 Style.Language == FormatStyle::LK_JavaScript) 449 Left->Type = TT_DictLiteral; 450 if (!consumeToken()) 451 return false; 452 } 453 } 454 return true; 455 } 456 457 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 458 if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block) 459 ++Left->BlockParameterCount; 460 if (Current->is(tok::comma)) { 461 ++Left->ParameterCount; 462 if (!Left->Role) 463 Left->Role.reset(new CommaSeparatedList(Style)); 464 Left->Role->CommaFound(Current); 465 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 466 Left->ParameterCount = 1; 467 } 468 } 469 470 bool parseConditional() { 471 while (CurrentToken) { 472 if (CurrentToken->is(tok::colon)) { 473 CurrentToken->Type = TT_ConditionalExpr; 474 next(); 475 return true; 476 } 477 if (!consumeToken()) 478 return false; 479 } 480 return false; 481 } 482 483 bool parseTemplateDeclaration() { 484 if (CurrentToken && CurrentToken->is(tok::less)) { 485 CurrentToken->Type = TT_TemplateOpener; 486 next(); 487 if (!parseAngle()) 488 return false; 489 if (CurrentToken) 490 CurrentToken->Previous->ClosesTemplateDeclaration = true; 491 return true; 492 } 493 return false; 494 } 495 496 bool consumeToken() { 497 FormatToken *Tok = CurrentToken; 498 next(); 499 switch (Tok->Tok.getKind()) { 500 case tok::plus: 501 case tok::minus: 502 if (!Tok->Previous && Line.MustBeDeclaration) 503 Tok->Type = TT_ObjCMethodSpecifier; 504 break; 505 case tok::colon: 506 if (!Tok->Previous) 507 return false; 508 // Colons from ?: are handled in parseConditional(). 509 if (Style.Language == FormatStyle::LK_JavaScript) { 510 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop 511 (Contexts.size() == 1 && // switch/case labels 512 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) || 513 Contexts.back().ContextKind == tok::l_paren || // function params 514 Contexts.back().ContextKind == tok::l_square || // array type 515 (Contexts.size() == 1 && 516 Line.MustBeDeclaration)) { // method/property declaration 517 Tok->Type = TT_JsTypeColon; 518 break; 519 } 520 } 521 if (Contexts.back().ColonIsDictLiteral || 522 Style.Language == FormatStyle::LK_Proto) { 523 Tok->Type = TT_DictLiteral; 524 } else if (Contexts.back().ColonIsObjCMethodExpr || 525 Line.startsWith(TT_ObjCMethodSpecifier)) { 526 Tok->Type = TT_ObjCMethodExpr; 527 const FormatToken *BeforePrevious = Tok->Previous->Previous; 528 if (!BeforePrevious || 529 !(BeforePrevious->is(TT_CastRParen) || 530 (BeforePrevious->is(TT_ObjCMethodExpr) && 531 BeforePrevious->is(tok::colon))) || 532 BeforePrevious->is(tok::r_square) || 533 Contexts.back().LongestObjCSelectorName == 0) { 534 Tok->Previous->Type = TT_SelectorName; 535 if (Tok->Previous->ColumnWidth > 536 Contexts.back().LongestObjCSelectorName) 537 Contexts.back().LongestObjCSelectorName = 538 Tok->Previous->ColumnWidth; 539 if (!Contexts.back().FirstObjCSelectorName) 540 Contexts.back().FirstObjCSelectorName = Tok->Previous; 541 } 542 } else if (Contexts.back().ColonIsForRangeExpr) { 543 Tok->Type = TT_RangeBasedForLoopColon; 544 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 545 Tok->Type = TT_BitFieldColon; 546 } else if (Contexts.size() == 1 && 547 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) { 548 if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren, 549 tok::kw_noexcept)) 550 Tok->Type = TT_CtorInitializerColon; 551 else 552 Tok->Type = TT_InheritanceColon; 553 } else if (Tok->Previous->is(tok::identifier) && Tok->Next && 554 Tok->Next->isOneOf(tok::r_paren, tok::comma)) { 555 // This handles a special macro in ObjC code where selectors including 556 // the colon are passed as macro arguments. 557 Tok->Type = TT_ObjCMethodExpr; 558 } else if (Contexts.back().ContextKind == tok::l_paren) { 559 Tok->Type = TT_InlineASMColon; 560 } 561 break; 562 case tok::pipe: 563 case tok::amp: 564 // | and & in declarations/type expressions represent union and 565 // intersection types, respectively. 566 if (Style.Language == FormatStyle::LK_JavaScript && 567 !Contexts.back().IsExpression) 568 Tok->Type = TT_JsTypeOperator; 569 break; 570 case tok::kw_if: 571 case tok::kw_while: 572 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 573 next(); 574 if (!parseParens(/*LookForDecls=*/true)) 575 return false; 576 } 577 break; 578 case tok::kw_for: 579 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Previous && 580 Tok->Previous->is(tok::period)) 581 break; 582 Contexts.back().ColonIsForRangeExpr = true; 583 next(); 584 if (!parseParens()) 585 return false; 586 break; 587 case tok::l_paren: 588 // When faced with 'operator()()', the kw_operator handler incorrectly 589 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make 590 // the first two parens OverloadedOperators and the second l_paren an 591 // OverloadedOperatorLParen. 592 if (Tok->Previous && 593 Tok->Previous->is(tok::r_paren) && 594 Tok->Previous->MatchingParen && 595 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) { 596 Tok->Previous->Type = TT_OverloadedOperator; 597 Tok->Previous->MatchingParen->Type = TT_OverloadedOperator; 598 Tok->Type = TT_OverloadedOperatorLParen; 599 } 600 601 if (!parseParens()) 602 return false; 603 if (Line.MustBeDeclaration && Contexts.size() == 1 && 604 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) && 605 (!Tok->Previous || 606 !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute, 607 TT_LeadingJavaAnnotation))) 608 Line.MightBeFunctionDecl = true; 609 break; 610 case tok::l_square: 611 if (!parseSquare()) 612 return false; 613 break; 614 case tok::l_brace: 615 if (!parseBrace()) 616 return false; 617 break; 618 case tok::less: 619 if (parseAngle()) { 620 Tok->Type = TT_TemplateOpener; 621 } else { 622 Tok->Type = TT_BinaryOperator; 623 NonTemplateLess.insert(Tok); 624 CurrentToken = Tok; 625 next(); 626 } 627 break; 628 case tok::r_paren: 629 case tok::r_square: 630 return false; 631 case tok::r_brace: 632 // Lines can start with '}'. 633 if (Tok->Previous) 634 return false; 635 break; 636 case tok::greater: 637 Tok->Type = TT_BinaryOperator; 638 break; 639 case tok::kw_operator: 640 while (CurrentToken && 641 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 642 if (CurrentToken->isOneOf(tok::star, tok::amp)) 643 CurrentToken->Type = TT_PointerOrReference; 644 consumeToken(); 645 if (CurrentToken && 646 CurrentToken->Previous->isOneOf(TT_BinaryOperator, tok::comma)) 647 CurrentToken->Previous->Type = TT_OverloadedOperator; 648 } 649 if (CurrentToken) { 650 CurrentToken->Type = TT_OverloadedOperatorLParen; 651 if (CurrentToken->Previous->is(TT_BinaryOperator)) 652 CurrentToken->Previous->Type = TT_OverloadedOperator; 653 } 654 break; 655 case tok::question: 656 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next && 657 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren, 658 tok::r_brace)) { 659 // Question marks before semicolons, colons, etc. indicate optional 660 // types (fields, parameters), e.g. 661 // function(x?: string, y?) {...} 662 // class X { y?; } 663 Tok->Type = TT_JsTypeOptionalQuestion; 664 break; 665 } 666 // Declarations cannot be conditional expressions, this can only be part 667 // of a type declaration. 668 if (Line.MustBeDeclaration && !Contexts.back().IsExpression && 669 Style.Language == FormatStyle::LK_JavaScript) 670 break; 671 parseConditional(); 672 break; 673 case tok::kw_template: 674 parseTemplateDeclaration(); 675 break; 676 case tok::comma: 677 if (Contexts.back().InCtorInitializer) 678 Tok->Type = TT_CtorInitializerComma; 679 else if (Contexts.back().FirstStartOfName && 680 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) { 681 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 682 Line.IsMultiVariableDeclStmt = true; 683 } 684 if (Contexts.back().IsForEachMacro) 685 Contexts.back().IsExpression = true; 686 break; 687 default: 688 break; 689 } 690 return true; 691 } 692 693 void parseIncludeDirective() { 694 if (CurrentToken && CurrentToken->is(tok::less)) { 695 next(); 696 while (CurrentToken) { 697 if (CurrentToken->isNot(tok::comment) || CurrentToken->Next) 698 CurrentToken->Type = TT_ImplicitStringLiteral; 699 next(); 700 } 701 } 702 } 703 704 void parseWarningOrError() { 705 next(); 706 // We still want to format the whitespace left of the first token of the 707 // warning or error. 708 next(); 709 while (CurrentToken) { 710 CurrentToken->Type = TT_ImplicitStringLiteral; 711 next(); 712 } 713 } 714 715 void parsePragma() { 716 next(); // Consume "pragma". 717 if (CurrentToken && 718 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) { 719 bool IsMark = CurrentToken->is(Keywords.kw_mark); 720 next(); // Consume "mark". 721 next(); // Consume first token (so we fix leading whitespace). 722 while (CurrentToken) { 723 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator)) 724 CurrentToken->Type = TT_ImplicitStringLiteral; 725 next(); 726 } 727 } 728 } 729 730 LineType parsePreprocessorDirective() { 731 bool IsFirstToken = CurrentToken->IsFirst; 732 LineType Type = LT_PreprocessorDirective; 733 next(); 734 if (!CurrentToken) 735 return Type; 736 737 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) { 738 // JavaScript files can contain shebang lines of the form: 739 // #!/usr/bin/env node 740 // Treat these like C++ #include directives. 741 while (CurrentToken) { 742 // Tokens cannot be comments here. 743 CurrentToken->Type = TT_ImplicitStringLiteral; 744 next(); 745 } 746 return LT_ImportStatement; 747 } 748 749 if (CurrentToken->Tok.is(tok::numeric_constant)) { 750 CurrentToken->SpacesRequiredBefore = 1; 751 return Type; 752 } 753 // Hashes in the middle of a line can lead to any strange token 754 // sequence. 755 if (!CurrentToken->Tok.getIdentifierInfo()) 756 return Type; 757 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 758 case tok::pp_include: 759 case tok::pp_include_next: 760 case tok::pp_import: 761 next(); 762 parseIncludeDirective(); 763 Type = LT_ImportStatement; 764 break; 765 case tok::pp_error: 766 case tok::pp_warning: 767 parseWarningOrError(); 768 break; 769 case tok::pp_pragma: 770 parsePragma(); 771 break; 772 case tok::pp_if: 773 case tok::pp_elif: 774 Contexts.back().IsExpression = true; 775 parseLine(); 776 break; 777 default: 778 break; 779 } 780 while (CurrentToken) 781 next(); 782 return Type; 783 } 784 785 public: 786 LineType parseLine() { 787 NonTemplateLess.clear(); 788 if (CurrentToken->is(tok::hash)) 789 return parsePreprocessorDirective(); 790 791 // Directly allow to 'import <string-literal>' to support protocol buffer 792 // definitions (code.google.com/p/protobuf) or missing "#" (either way we 793 // should not break the line). 794 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 795 if ((Style.Language == FormatStyle::LK_Java && 796 CurrentToken->is(Keywords.kw_package)) || 797 (Info && Info->getPPKeywordID() == tok::pp_import && 798 CurrentToken->Next && 799 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier, 800 tok::kw_static))) { 801 next(); 802 parseIncludeDirective(); 803 return LT_ImportStatement; 804 } 805 806 // If this line starts and ends in '<' and '>', respectively, it is likely 807 // part of "#define <a/b.h>". 808 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 809 parseIncludeDirective(); 810 return LT_ImportStatement; 811 } 812 813 // In .proto files, top-level options are very similar to import statements 814 // and should not be line-wrapped. 815 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 && 816 CurrentToken->is(Keywords.kw_option)) { 817 next(); 818 if (CurrentToken && CurrentToken->is(tok::identifier)) 819 return LT_ImportStatement; 820 } 821 822 bool KeywordVirtualFound = false; 823 bool ImportStatement = false; 824 825 // import {...} from '...'; 826 if (Style.Language == FormatStyle::LK_JavaScript && 827 CurrentToken->is(Keywords.kw_import)) 828 ImportStatement = true; 829 830 while (CurrentToken) { 831 if (CurrentToken->is(tok::kw_virtual)) 832 KeywordVirtualFound = true; 833 if (Style.Language == FormatStyle::LK_JavaScript) { 834 // export {...} from '...'; 835 // An export followed by "from 'some string';" is a re-export from 836 // another module identified by a URI and is treated as a 837 // LT_ImportStatement (i.e. prevent wraps on it for long URIs). 838 // Just "export {...};" or "export class ..." should not be treated as 839 // an import in this sense. 840 if (Line.First->is(tok::kw_export) && 841 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next && 842 CurrentToken->Next->isStringLiteral()) 843 ImportStatement = true; 844 if (isClosureImportStatement(*CurrentToken)) 845 ImportStatement = true; 846 } 847 if (!consumeToken()) 848 return LT_Invalid; 849 } 850 if (KeywordVirtualFound) 851 return LT_VirtualFunctionDecl; 852 if (ImportStatement) 853 return LT_ImportStatement; 854 855 if (Line.startsWith(TT_ObjCMethodSpecifier)) { 856 if (Contexts.back().FirstObjCSelectorName) 857 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 858 Contexts.back().LongestObjCSelectorName; 859 return LT_ObjCMethodDecl; 860 } 861 862 return LT_Other; 863 } 864 865 private: 866 bool isClosureImportStatement(const FormatToken &Tok) { 867 // FIXME: Closure-library specific stuff should not be hard-coded but be 868 // configurable. 869 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) && 870 Tok.Next->Next && (Tok.Next->Next->TokenText == "module" || 871 Tok.Next->Next->TokenText == "provide" || 872 Tok.Next->Next->TokenText == "require" || 873 Tok.Next->Next->TokenText == "setTestOnly" || 874 Tok.Next->Next->TokenText == "forwardDeclare") && 875 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren); 876 } 877 878 void resetTokenMetadata(FormatToken *Token) { 879 if (!Token) 880 return; 881 882 // Reset token type in case we have already looked at it and then 883 // recovered from an error (e.g. failure to find the matching >). 884 if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro, 885 TT_FunctionLBrace, TT_ImplicitStringLiteral, 886 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow, 887 TT_OverloadedOperator, TT_RegexLiteral, 888 TT_TemplateString)) 889 CurrentToken->Type = TT_Unknown; 890 CurrentToken->Role.reset(); 891 CurrentToken->MatchingParen = nullptr; 892 CurrentToken->FakeLParens.clear(); 893 CurrentToken->FakeRParens = 0; 894 } 895 896 void next() { 897 if (CurrentToken) { 898 CurrentToken->NestingLevel = Contexts.size() - 1; 899 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 900 modifyContext(*CurrentToken); 901 determineTokenType(*CurrentToken); 902 CurrentToken = CurrentToken->Next; 903 } 904 905 resetTokenMetadata(CurrentToken); 906 } 907 908 /// \brief A struct to hold information valid in a specific context, e.g. 909 /// a pair of parenthesis. 910 struct Context { 911 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 912 bool IsExpression) 913 : ContextKind(ContextKind), BindingStrength(BindingStrength), 914 IsExpression(IsExpression) {} 915 916 tok::TokenKind ContextKind; 917 unsigned BindingStrength; 918 bool IsExpression; 919 unsigned LongestObjCSelectorName = 0; 920 bool ColonIsForRangeExpr = false; 921 bool ColonIsDictLiteral = false; 922 bool ColonIsObjCMethodExpr = false; 923 FormatToken *FirstObjCSelectorName = nullptr; 924 FormatToken *FirstStartOfName = nullptr; 925 bool CanBeExpression = true; 926 bool InTemplateArgument = false; 927 bool InCtorInitializer = false; 928 bool CaretFound = false; 929 bool IsForEachMacro = false; 930 }; 931 932 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime 933 /// of each instance. 934 struct ScopedContextCreator { 935 AnnotatingParser &P; 936 937 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 938 unsigned Increase) 939 : P(P) { 940 P.Contexts.push_back(Context(ContextKind, 941 P.Contexts.back().BindingStrength + Increase, 942 P.Contexts.back().IsExpression)); 943 } 944 945 ~ScopedContextCreator() { P.Contexts.pop_back(); } 946 }; 947 948 void modifyContext(const FormatToken &Current) { 949 if (Current.getPrecedence() == prec::Assignment && 950 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) && 951 // Type aliases use `type X = ...;` in TypeScript. 952 !(Style.Language == FormatStyle::LK_JavaScript && 953 Line.startsWith(Keywords.kw_type, tok::identifier)) && 954 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 955 Contexts.back().IsExpression = true; 956 if (!Line.startsWith(TT_UnaryOperator)) { 957 for (FormatToken *Previous = Current.Previous; 958 Previous && Previous->Previous && 959 !Previous->Previous->isOneOf(tok::comma, tok::semi); 960 Previous = Previous->Previous) { 961 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 962 Previous = Previous->MatchingParen; 963 if (!Previous) 964 break; 965 } 966 if (Previous->opensScope()) 967 break; 968 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) && 969 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) && 970 Previous->Previous && Previous->Previous->isNot(tok::equal)) 971 Previous->Type = TT_PointerOrReference; 972 } 973 } 974 } else if (Current.is(tok::lessless) && 975 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) { 976 Contexts.back().IsExpression = true; 977 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 978 Contexts.back().IsExpression = true; 979 } else if (Current.is(TT_TrailingReturnArrow)) { 980 Contexts.back().IsExpression = false; 981 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) { 982 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java; 983 } else if (Current.Previous && 984 Current.Previous->is(TT_CtorInitializerColon)) { 985 Contexts.back().IsExpression = true; 986 Contexts.back().InCtorInitializer = true; 987 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 988 for (FormatToken *Previous = Current.Previous; 989 Previous && Previous->isOneOf(tok::star, tok::amp); 990 Previous = Previous->Previous) 991 Previous->Type = TT_PointerOrReference; 992 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer) 993 Contexts.back().IsExpression = false; 994 } else if (Current.is(tok::kw_new)) { 995 Contexts.back().CanBeExpression = false; 996 } else if (Current.isOneOf(tok::semi, tok::exclaim)) { 997 // This should be the condition or increment in a for-loop. 998 Contexts.back().IsExpression = true; 999 } 1000 } 1001 1002 void determineTokenType(FormatToken &Current) { 1003 if (!Current.is(TT_Unknown)) 1004 // The token type is already known. 1005 return; 1006 1007 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1008 // function declaration have been found. In this case, 'Current' is a 1009 // trailing token of this declaration and thus cannot be a name. 1010 if (Current.is(Keywords.kw_instanceof)) { 1011 Current.Type = TT_BinaryOperator; 1012 } else if (isStartOfName(Current) && 1013 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 1014 Contexts.back().FirstStartOfName = &Current; 1015 Current.Type = TT_StartOfName; 1016 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) { 1017 AutoFound = true; 1018 } else if (Current.is(tok::arrow) && 1019 Style.Language == FormatStyle::LK_Java) { 1020 Current.Type = TT_LambdaArrow; 1021 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration && 1022 Current.NestingLevel == 0) { 1023 Current.Type = TT_TrailingReturnArrow; 1024 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 1025 Current.Type = 1026 determineStarAmpUsage(Current, Contexts.back().CanBeExpression && 1027 Contexts.back().IsExpression, 1028 Contexts.back().InTemplateArgument); 1029 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 1030 Current.Type = determinePlusMinusCaretUsage(Current); 1031 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret)) 1032 Contexts.back().CaretFound = true; 1033 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 1034 Current.Type = determineIncrementUsage(Current); 1035 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 1036 Current.Type = TT_UnaryOperator; 1037 } else if (Current.is(tok::question)) { 1038 if (Style.Language == FormatStyle::LK_JavaScript && 1039 Line.MustBeDeclaration && !Contexts.back().IsExpression) { 1040 // In JavaScript, `interface X { foo?(): bar; }` is an optional method 1041 // on the interface, not a ternary expression. 1042 Current.Type = TT_JsTypeOptionalQuestion; 1043 } else { 1044 Current.Type = TT_ConditionalExpr; 1045 } 1046 } else if (Current.isBinaryOperator() && 1047 (!Current.Previous || Current.Previous->isNot(tok::l_square))) { 1048 Current.Type = TT_BinaryOperator; 1049 } else if (Current.is(tok::comment)) { 1050 if (Current.TokenText.startswith("/*")) { 1051 if (Current.TokenText.endswith("*/")) 1052 Current.Type = TT_BlockComment; 1053 else 1054 // The lexer has for some reason determined a comment here. But we 1055 // cannot really handle it, if it isn't properly terminated. 1056 Current.Tok.setKind(tok::unknown); 1057 } else { 1058 Current.Type = TT_LineComment; 1059 } 1060 } else if (Current.is(tok::r_paren)) { 1061 if (rParenEndsCast(Current)) 1062 Current.Type = TT_CastRParen; 1063 if (Current.MatchingParen && Current.Next && 1064 !Current.Next->isBinaryOperator() && 1065 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace, 1066 tok::period, tok::arrow, tok::coloncolon)) 1067 if (FormatToken *AfterParen = Current.MatchingParen->Next) { 1068 // Make sure this isn't the return type of an Obj-C block declaration 1069 if (AfterParen->Tok.isNot(tok::caret)) { 1070 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) 1071 if (BeforeParen->is(tok::identifier) && 1072 BeforeParen->TokenText == BeforeParen->TokenText.upper() && 1073 (!BeforeParen->Previous || 1074 BeforeParen->Previous->ClosesTemplateDeclaration)) 1075 Current.Type = TT_FunctionAnnotationRParen; 1076 } 1077 } 1078 } else if (Current.is(tok::at) && Current.Next) { 1079 if (Current.Next->isStringLiteral()) { 1080 Current.Type = TT_ObjCStringLiteral; 1081 } else { 1082 switch (Current.Next->Tok.getObjCKeywordID()) { 1083 case tok::objc_interface: 1084 case tok::objc_implementation: 1085 case tok::objc_protocol: 1086 Current.Type = TT_ObjCDecl; 1087 break; 1088 case tok::objc_property: 1089 Current.Type = TT_ObjCProperty; 1090 break; 1091 default: 1092 break; 1093 } 1094 } 1095 } else if (Current.is(tok::period)) { 1096 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 1097 if (PreviousNoComment && 1098 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 1099 Current.Type = TT_DesignatedInitializerPeriod; 1100 else if (Style.Language == FormatStyle::LK_Java && Current.Previous && 1101 Current.Previous->isOneOf(TT_JavaAnnotation, 1102 TT_LeadingJavaAnnotation)) { 1103 Current.Type = Current.Previous->Type; 1104 } 1105 } else if (Current.isOneOf(tok::identifier, tok::kw_const) && 1106 Current.Previous && 1107 !Current.Previous->isOneOf(tok::equal, tok::at) && 1108 Line.MightBeFunctionDecl && Contexts.size() == 1) { 1109 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1110 // function declaration have been found. 1111 Current.Type = TT_TrailingAnnotation; 1112 } else if ((Style.Language == FormatStyle::LK_Java || 1113 Style.Language == FormatStyle::LK_JavaScript) && 1114 Current.Previous) { 1115 if (Current.Previous->is(tok::at) && 1116 Current.isNot(Keywords.kw_interface)) { 1117 const FormatToken &AtToken = *Current.Previous; 1118 const FormatToken *Previous = AtToken.getPreviousNonComment(); 1119 if (!Previous || Previous->is(TT_LeadingJavaAnnotation)) 1120 Current.Type = TT_LeadingJavaAnnotation; 1121 else 1122 Current.Type = TT_JavaAnnotation; 1123 } else if (Current.Previous->is(tok::period) && 1124 Current.Previous->isOneOf(TT_JavaAnnotation, 1125 TT_LeadingJavaAnnotation)) { 1126 Current.Type = Current.Previous->Type; 1127 } 1128 } 1129 } 1130 1131 /// \brief Take a guess at whether \p Tok starts a name of a function or 1132 /// variable declaration. 1133 /// 1134 /// This is a heuristic based on whether \p Tok is an identifier following 1135 /// something that is likely a type. 1136 bool isStartOfName(const FormatToken &Tok) { 1137 if (Tok.isNot(tok::identifier) || !Tok.Previous) 1138 return false; 1139 1140 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof)) 1141 return false; 1142 if (Style.Language == FormatStyle::LK_JavaScript && 1143 Tok.Previous->is(Keywords.kw_in)) 1144 return false; 1145 1146 // Skip "const" as it does not have an influence on whether this is a name. 1147 FormatToken *PreviousNotConst = Tok.Previous; 1148 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 1149 PreviousNotConst = PreviousNotConst->Previous; 1150 1151 if (!PreviousNotConst) 1152 return false; 1153 1154 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 1155 PreviousNotConst->Previous && 1156 PreviousNotConst->Previous->is(tok::hash); 1157 1158 if (PreviousNotConst->is(TT_TemplateCloser)) 1159 return PreviousNotConst && PreviousNotConst->MatchingParen && 1160 PreviousNotConst->MatchingParen->Previous && 1161 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) && 1162 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 1163 1164 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen && 1165 PreviousNotConst->MatchingParen->Previous && 1166 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype)) 1167 return true; 1168 1169 return (!IsPPKeyword && 1170 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) || 1171 PreviousNotConst->is(TT_PointerOrReference) || 1172 PreviousNotConst->isSimpleTypeSpecifier(); 1173 } 1174 1175 /// \brief Determine whether ')' is ending a cast. 1176 bool rParenEndsCast(const FormatToken &Tok) { 1177 // C-style casts are only used in C++ and Java. 1178 if (Style.Language != FormatStyle::LK_Cpp && 1179 Style.Language != FormatStyle::LK_ObjC && 1180 Style.Language != FormatStyle::LK_Java) 1181 return false; 1182 1183 // Empty parens aren't casts and there are no casts at the end of the line. 1184 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen) 1185 return false; 1186 1187 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 1188 if (LeftOfParens) { 1189 // If there is a closing parenthesis left of the current parentheses, 1190 // look past it as these might be chained casts. 1191 if (LeftOfParens->is(tok::r_paren)) { 1192 if (!LeftOfParens->MatchingParen || 1193 !LeftOfParens->MatchingParen->Previous) 1194 return false; 1195 LeftOfParens = LeftOfParens->MatchingParen->Previous; 1196 } 1197 1198 // If there is an identifier (or with a few exceptions a keyword) right 1199 // before the parentheses, this is unlikely to be a cast. 1200 if (LeftOfParens->Tok.getIdentifierInfo() && 1201 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case, 1202 tok::kw_delete)) 1203 return false; 1204 1205 // Certain other tokens right before the parentheses are also signals that 1206 // this cannot be a cast. 1207 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator, 1208 TT_TemplateCloser, tok::ellipsis)) 1209 return false; 1210 } 1211 1212 if (Tok.Next->is(tok::question)) 1213 return false; 1214 1215 // As Java has no function types, a "(" after the ")" likely means that this 1216 // is a cast. 1217 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren)) 1218 return true; 1219 1220 // If a (non-string) literal follows, this is likely a cast. 1221 if (Tok.Next->isNot(tok::string_literal) && 1222 (Tok.Next->Tok.isLiteral() || 1223 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 1224 return true; 1225 1226 // Heuristically try to determine whether the parentheses contain a type. 1227 bool ParensAreType = 1228 !Tok.Previous || 1229 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) || 1230 Tok.Previous->isSimpleTypeSpecifier(); 1231 bool ParensCouldEndDecl = 1232 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater); 1233 if (ParensAreType && !ParensCouldEndDecl) 1234 return true; 1235 1236 // At this point, we heuristically assume that there are no casts at the 1237 // start of the line. We assume that we have found most cases where there 1238 // are by the logic above, e.g. "(void)x;". 1239 if (!LeftOfParens) 1240 return false; 1241 1242 // Certain token types inside the parentheses mean that this can't be a 1243 // cast. 1244 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok; 1245 Token = Token->Next) 1246 if (Token->is(TT_BinaryOperator)) 1247 return false; 1248 1249 // If the following token is an identifier or 'this', this is a cast. All 1250 // cases where this can be something else are handled above. 1251 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this)) 1252 return true; 1253 1254 if (!Tok.Next->Next) 1255 return false; 1256 1257 // If the next token after the parenthesis is a unary operator, assume 1258 // that this is cast, unless there are unexpected tokens inside the 1259 // parenthesis. 1260 bool NextIsUnary = 1261 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star); 1262 if (!NextIsUnary || Tok.Next->is(tok::plus) || 1263 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) 1264 return false; 1265 // Search for unexpected tokens. 1266 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen; 1267 Prev = Prev->Previous) { 1268 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon)) 1269 return false; 1270 } 1271 return true; 1272 } 1273 1274 /// \brief Return the type of the given token assuming it is * or &. 1275 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 1276 bool InTemplateArgument) { 1277 if (Style.Language == FormatStyle::LK_JavaScript) 1278 return TT_BinaryOperator; 1279 1280 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1281 if (!PrevToken) 1282 return TT_UnaryOperator; 1283 1284 const FormatToken *NextToken = Tok.getNextNonComment(); 1285 if (!NextToken || NextToken->isOneOf(tok::arrow, tok::equal) || 1286 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) 1287 return TT_PointerOrReference; 1288 1289 if (PrevToken->is(tok::coloncolon)) 1290 return TT_PointerOrReference; 1291 1292 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 1293 tok::comma, tok::semi, tok::kw_return, tok::colon, 1294 tok::equal, tok::kw_delete, tok::kw_sizeof) || 1295 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr, 1296 TT_UnaryOperator, TT_CastRParen)) 1297 return TT_UnaryOperator; 1298 1299 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare)) 1300 return TT_PointerOrReference; 1301 if (NextToken->is(tok::kw_operator) && !IsExpression) 1302 return TT_PointerOrReference; 1303 if (NextToken->isOneOf(tok::comma, tok::semi)) 1304 return TT_PointerOrReference; 1305 1306 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen && 1307 PrevToken->MatchingParen->Previous && 1308 PrevToken->MatchingParen->Previous->isOneOf(tok::kw_typeof, 1309 tok::kw_decltype)) 1310 return TT_PointerOrReference; 1311 1312 if (PrevToken->Tok.isLiteral() || 1313 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 1314 tok::kw_false, tok::r_brace) || 1315 NextToken->Tok.isLiteral() || 1316 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 1317 NextToken->isUnaryOperator() || 1318 // If we know we're in a template argument, there are no named 1319 // declarations. Thus, having an identifier on the right-hand side 1320 // indicates a binary operator. 1321 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 1322 return TT_BinaryOperator; 1323 1324 // "&&(" is quite unlikely to be two successive unary "&". 1325 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren)) 1326 return TT_BinaryOperator; 1327 1328 // This catches some cases where evaluation order is used as control flow: 1329 // aaa && aaa->f(); 1330 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 1331 if (NextNextToken && NextNextToken->is(tok::arrow)) 1332 return TT_BinaryOperator; 1333 1334 // It is very unlikely that we are going to find a pointer or reference type 1335 // definition on the RHS of an assignment. 1336 if (IsExpression && !Contexts.back().CaretFound) 1337 return TT_BinaryOperator; 1338 1339 return TT_PointerOrReference; 1340 } 1341 1342 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 1343 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1344 if (!PrevToken) 1345 return TT_UnaryOperator; 1346 1347 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator) && 1348 !PrevToken->is(tok::exclaim)) 1349 // There aren't any trailing unary operators except for TypeScript's 1350 // non-null operator (!). Thus, this must be squence of leading operators. 1351 return TT_UnaryOperator; 1352 1353 // Use heuristics to recognize unary operators. 1354 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 1355 tok::question, tok::colon, tok::kw_return, 1356 tok::kw_case, tok::at, tok::l_brace)) 1357 return TT_UnaryOperator; 1358 1359 // There can't be two consecutive binary operators. 1360 if (PrevToken->is(TT_BinaryOperator)) 1361 return TT_UnaryOperator; 1362 1363 // Fall back to marking the token as binary operator. 1364 return TT_BinaryOperator; 1365 } 1366 1367 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements. 1368 TokenType determineIncrementUsage(const FormatToken &Tok) { 1369 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1370 if (!PrevToken || PrevToken->is(TT_CastRParen)) 1371 return TT_UnaryOperator; 1372 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 1373 return TT_TrailingUnaryOperator; 1374 1375 return TT_UnaryOperator; 1376 } 1377 1378 SmallVector<Context, 8> Contexts; 1379 1380 const FormatStyle &Style; 1381 AnnotatedLine &Line; 1382 FormatToken *CurrentToken; 1383 bool AutoFound; 1384 const AdditionalKeywords &Keywords; 1385 1386 // Set of "<" tokens that do not open a template parameter list. If parseAngle 1387 // determines that a specific token can't be a template opener, it will make 1388 // same decision irrespective of the decisions for tokens leading up to it. 1389 // Store this information to prevent this from causing exponential runtime. 1390 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess; 1391 }; 1392 1393 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1; 1394 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 1395 1396 /// \brief Parses binary expressions by inserting fake parenthesis based on 1397 /// operator precedence. 1398 class ExpressionParser { 1399 public: 1400 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords, 1401 AnnotatedLine &Line) 1402 : Style(Style), Keywords(Keywords), Current(Line.First) {} 1403 1404 /// \brief Parse expressions with the given operatore precedence. 1405 void parse(int Precedence = 0) { 1406 // Skip 'return' and ObjC selector colons as they are not part of a binary 1407 // expression. 1408 while (Current && (Current->is(tok::kw_return) || 1409 (Current->is(tok::colon) && 1410 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) 1411 next(); 1412 1413 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1414 return; 1415 1416 // Conditional expressions need to be parsed separately for proper nesting. 1417 if (Precedence == prec::Conditional) { 1418 parseConditionalExpr(); 1419 return; 1420 } 1421 1422 // Parse unary operators, which all have a higher precedence than binary 1423 // operators. 1424 if (Precedence == PrecedenceUnaryOperator) { 1425 parseUnaryOperator(); 1426 return; 1427 } 1428 1429 FormatToken *Start = Current; 1430 FormatToken *LatestOperator = nullptr; 1431 unsigned OperatorIndex = 0; 1432 1433 while (Current) { 1434 // Consume operators with higher precedence. 1435 parse(Precedence + 1); 1436 1437 int CurrentPrecedence = getCurrentPrecedence(); 1438 1439 if (Current && Current->is(TT_SelectorName) && 1440 Precedence == CurrentPrecedence) { 1441 if (LatestOperator) 1442 addFakeParenthesis(Start, prec::Level(Precedence)); 1443 Start = Current; 1444 } 1445 1446 // At the end of the line or when an operator with higher precedence is 1447 // found, insert fake parenthesis and return. 1448 if (!Current || (Current->closesScope() && Current->MatchingParen) || 1449 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) || 1450 (CurrentPrecedence == prec::Conditional && 1451 Precedence == prec::Assignment && Current->is(tok::colon))) { 1452 break; 1453 } 1454 1455 // Consume scopes: (), [], <> and {} 1456 if (Current->opensScope()) { 1457 while (Current && !Current->closesScope()) { 1458 next(); 1459 parse(); 1460 } 1461 next(); 1462 } else { 1463 // Operator found. 1464 if (CurrentPrecedence == Precedence) { 1465 if (LatestOperator) 1466 LatestOperator->NextOperator = Current; 1467 LatestOperator = Current; 1468 Current->OperatorIndex = OperatorIndex; 1469 ++OperatorIndex; 1470 } 1471 next(/*SkipPastLeadingComments=*/Precedence > 0); 1472 } 1473 } 1474 1475 if (LatestOperator && (Current || Precedence > 0)) { 1476 // LatestOperator->LastOperator = true; 1477 if (Precedence == PrecedenceArrowAndPeriod) { 1478 // Call expressions don't have a binary operator precedence. 1479 addFakeParenthesis(Start, prec::Unknown); 1480 } else { 1481 addFakeParenthesis(Start, prec::Level(Precedence)); 1482 } 1483 } 1484 } 1485 1486 private: 1487 /// \brief Gets the precedence (+1) of the given token for binary operators 1488 /// and other tokens that we treat like binary operators. 1489 int getCurrentPrecedence() { 1490 if (Current) { 1491 const FormatToken *NextNonComment = Current->getNextNonComment(); 1492 if (Current->is(TT_ConditionalExpr)) 1493 return prec::Conditional; 1494 if (NextNonComment && NextNonComment->is(tok::colon) && 1495 NextNonComment->is(TT_DictLiteral)) 1496 return prec::Comma; 1497 if (Current->is(TT_LambdaArrow)) 1498 return prec::Comma; 1499 if (Current->is(TT_JsFatArrow)) 1500 return prec::Assignment; 1501 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName, 1502 TT_JsComputedPropertyName) || 1503 (Current->is(tok::comment) && NextNonComment && 1504 NextNonComment->is(TT_SelectorName))) 1505 return 0; 1506 if (Current->is(TT_RangeBasedForLoopColon)) 1507 return prec::Comma; 1508 if ((Style.Language == FormatStyle::LK_Java || 1509 Style.Language == FormatStyle::LK_JavaScript) && 1510 Current->is(Keywords.kw_instanceof)) 1511 return prec::Relational; 1512 if (Style.Language == FormatStyle::LK_JavaScript && 1513 Current->is(Keywords.kw_in)) 1514 return prec::Relational; 1515 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma)) 1516 return Current->getPrecedence(); 1517 if (Current->isOneOf(tok::period, tok::arrow)) 1518 return PrecedenceArrowAndPeriod; 1519 if ((Style.Language == FormatStyle::LK_Java || 1520 Style.Language == FormatStyle::LK_JavaScript) && 1521 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements, 1522 Keywords.kw_throws)) 1523 return 0; 1524 } 1525 return -1; 1526 } 1527 1528 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1529 Start->FakeLParens.push_back(Precedence); 1530 if (Precedence > prec::Unknown) 1531 Start->StartsBinaryExpression = true; 1532 if (Current) { 1533 FormatToken *Previous = Current->Previous; 1534 while (Previous->is(tok::comment) && Previous->Previous) 1535 Previous = Previous->Previous; 1536 ++Previous->FakeRParens; 1537 if (Precedence > prec::Unknown) 1538 Previous->EndsBinaryExpression = true; 1539 } 1540 } 1541 1542 /// \brief Parse unary operator expressions and surround them with fake 1543 /// parentheses if appropriate. 1544 void parseUnaryOperator() { 1545 if (!Current || Current->isNot(TT_UnaryOperator)) { 1546 parse(PrecedenceArrowAndPeriod); 1547 return; 1548 } 1549 1550 FormatToken *Start = Current; 1551 next(); 1552 parseUnaryOperator(); 1553 1554 // The actual precedence doesn't matter. 1555 addFakeParenthesis(Start, prec::Unknown); 1556 } 1557 1558 void parseConditionalExpr() { 1559 while (Current && Current->isTrailingComment()) { 1560 next(); 1561 } 1562 FormatToken *Start = Current; 1563 parse(prec::LogicalOr); 1564 if (!Current || !Current->is(tok::question)) 1565 return; 1566 next(); 1567 parse(prec::Assignment); 1568 if (!Current || Current->isNot(TT_ConditionalExpr)) 1569 return; 1570 next(); 1571 parse(prec::Assignment); 1572 addFakeParenthesis(Start, prec::Conditional); 1573 } 1574 1575 void next(bool SkipPastLeadingComments = true) { 1576 if (Current) 1577 Current = Current->Next; 1578 while (Current && 1579 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 1580 Current->isTrailingComment()) 1581 Current = Current->Next; 1582 } 1583 1584 const FormatStyle &Style; 1585 const AdditionalKeywords &Keywords; 1586 FormatToken *Current; 1587 }; 1588 1589 } // end anonymous namespace 1590 1591 void TokenAnnotator::setCommentLineLevels( 1592 SmallVectorImpl<AnnotatedLine *> &Lines) { 1593 const AnnotatedLine *NextNonCommentLine = nullptr; 1594 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 1595 E = Lines.rend(); 1596 I != E; ++I) { 1597 bool CommentLine = (*I)->First; 1598 for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) { 1599 if (!Tok->is(tok::comment)) { 1600 CommentLine = false; 1601 break; 1602 } 1603 } 1604 if (NextNonCommentLine && CommentLine) 1605 (*I)->Level = NextNonCommentLine->Level; 1606 else 1607 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 1608 1609 setCommentLineLevels((*I)->Children); 1610 } 1611 } 1612 1613 static unsigned maxNestingDepth(const AnnotatedLine &Line) { 1614 unsigned Result = 0; 1615 for (const auto* Tok = Line.First; Tok != nullptr; Tok = Tok->Next) 1616 Result = std::max(Result, Tok->NestingLevel); 1617 return Result; 1618 } 1619 1620 void TokenAnnotator::annotate(AnnotatedLine &Line) { 1621 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1622 E = Line.Children.end(); 1623 I != E; ++I) { 1624 annotate(**I); 1625 } 1626 AnnotatingParser Parser(Style, Line, Keywords); 1627 Line.Type = Parser.parseLine(); 1628 1629 // With very deep nesting, ExpressionParser uses lots of stack and the 1630 // formatting algorithm is very slow. We're not going to do a good job here 1631 // anyway - it's probably generated code being formatted by mistake. 1632 // Just skip the whole line. 1633 if (maxNestingDepth(Line) > 50) 1634 Line.Type = LT_Invalid; 1635 1636 if (Line.Type == LT_Invalid) 1637 return; 1638 1639 ExpressionParser ExprParser(Style, Keywords, Line); 1640 ExprParser.parse(); 1641 1642 if (Line.startsWith(TT_ObjCMethodSpecifier)) 1643 Line.Type = LT_ObjCMethodDecl; 1644 else if (Line.startsWith(TT_ObjCDecl)) 1645 Line.Type = LT_ObjCDecl; 1646 else if (Line.startsWith(TT_ObjCProperty)) 1647 Line.Type = LT_ObjCProperty; 1648 1649 Line.First->SpacesRequiredBefore = 1; 1650 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 1651 } 1652 1653 // This function heuristically determines whether 'Current' starts the name of a 1654 // function declaration. 1655 static bool isFunctionDeclarationName(const FormatToken &Current, 1656 const AnnotatedLine &Line) { 1657 auto skipOperatorName = [](const FormatToken* Next) -> const FormatToken* { 1658 for (; Next; Next = Next->Next) { 1659 if (Next->is(TT_OverloadedOperatorLParen)) 1660 return Next; 1661 if (Next->is(TT_OverloadedOperator)) 1662 continue; 1663 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) { 1664 // For 'new[]' and 'delete[]'. 1665 if (Next->Next && Next->Next->is(tok::l_square) && 1666 Next->Next->Next && Next->Next->Next->is(tok::r_square)) 1667 Next = Next->Next->Next; 1668 continue; 1669 } 1670 1671 break; 1672 } 1673 return nullptr; 1674 }; 1675 1676 // Find parentheses of parameter list. 1677 const FormatToken *Next = Current.Next; 1678 if (Current.is(tok::kw_operator)) { 1679 if (Current.Previous && Current.Previous->is(tok::coloncolon)) 1680 return false; 1681 Next = skipOperatorName(Next); 1682 } else { 1683 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0) 1684 return false; 1685 for (; Next; Next = Next->Next) { 1686 if (Next->is(TT_TemplateOpener)) { 1687 Next = Next->MatchingParen; 1688 } else if (Next->is(tok::coloncolon)) { 1689 Next = Next->Next; 1690 if (!Next) 1691 return false; 1692 if (Next->is(tok::kw_operator)) { 1693 Next = skipOperatorName(Next->Next); 1694 break; 1695 } 1696 if (!Next->is(tok::identifier)) 1697 return false; 1698 } else if (Next->is(tok::l_paren)) { 1699 break; 1700 } else { 1701 return false; 1702 } 1703 } 1704 } 1705 1706 // Check whether parameter list can be long to a function declaration. 1707 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen) 1708 return false; 1709 // If the lines ends with "{", this is likely an function definition. 1710 if (Line.Last->is(tok::l_brace)) 1711 return true; 1712 if (Next->Next == Next->MatchingParen) 1713 return true; // Empty parentheses. 1714 // If there is an &/&& after the r_paren, this is likely a function. 1715 if (Next->MatchingParen->Next && 1716 Next->MatchingParen->Next->is(TT_PointerOrReference)) 1717 return true; 1718 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen; 1719 Tok = Tok->Next) { 1720 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 1721 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis)) 1722 return true; 1723 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) || 1724 Tok->Tok.isLiteral()) 1725 return false; 1726 } 1727 return false; 1728 } 1729 1730 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const { 1731 assert(Line.MightBeFunctionDecl); 1732 1733 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel || 1734 Style.AlwaysBreakAfterReturnType == 1735 FormatStyle::RTBS_TopLevelDefinitions) && 1736 Line.Level > 0) 1737 return false; 1738 1739 switch (Style.AlwaysBreakAfterReturnType) { 1740 case FormatStyle::RTBS_None: 1741 return false; 1742 case FormatStyle::RTBS_All: 1743 case FormatStyle::RTBS_TopLevel: 1744 return true; 1745 case FormatStyle::RTBS_AllDefinitions: 1746 case FormatStyle::RTBS_TopLevelDefinitions: 1747 return Line.mightBeFunctionDefinition(); 1748 } 1749 1750 return false; 1751 } 1752 1753 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 1754 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1755 E = Line.Children.end(); 1756 I != E; ++I) { 1757 calculateFormattingInformation(**I); 1758 } 1759 1760 Line.First->TotalLength = 1761 Line.First->IsMultiline ? Style.ColumnLimit : Line.First->ColumnWidth; 1762 if (!Line.First->Next) 1763 return; 1764 FormatToken *Current = Line.First->Next; 1765 bool InFunctionDecl = Line.MightBeFunctionDecl; 1766 while (Current) { 1767 if (isFunctionDeclarationName(*Current, Line)) 1768 Current->Type = TT_FunctionDeclarationName; 1769 if (Current->is(TT_LineComment)) { 1770 if (Current->Previous->BlockKind == BK_BracedInit && 1771 Current->Previous->opensScope()) 1772 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1; 1773 else 1774 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 1775 1776 // If we find a trailing comment, iterate backwards to determine whether 1777 // it seems to relate to a specific parameter. If so, break before that 1778 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 1779 // to the previous line in: 1780 // SomeFunction(a, 1781 // b, // comment 1782 // c); 1783 if (!Current->HasUnescapedNewline) { 1784 for (FormatToken *Parameter = Current->Previous; Parameter; 1785 Parameter = Parameter->Previous) { 1786 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 1787 break; 1788 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 1789 if (!Parameter->Previous->is(TT_CtorInitializerComma) && 1790 Parameter->HasUnescapedNewline) 1791 Parameter->MustBreakBefore = true; 1792 break; 1793 } 1794 } 1795 } 1796 } else if (Current->SpacesRequiredBefore == 0 && 1797 spaceRequiredBefore(Line, *Current)) { 1798 Current->SpacesRequiredBefore = 1; 1799 } 1800 1801 Current->MustBreakBefore = 1802 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 1803 1804 if (!Current->MustBreakBefore && InFunctionDecl && 1805 Current->is(TT_FunctionDeclarationName)) 1806 Current->MustBreakBefore = mustBreakForReturnType(Line); 1807 1808 Current->CanBreakBefore = 1809 Current->MustBreakBefore || canBreakBefore(Line, *Current); 1810 unsigned ChildSize = 0; 1811 if (Current->Previous->Children.size() == 1) { 1812 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 1813 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 1814 : LastOfChild.TotalLength + 1; 1815 } 1816 const FormatToken *Prev = Current->Previous; 1817 if (Current->MustBreakBefore || Prev->Children.size() > 1 || 1818 (Prev->Children.size() == 1 && 1819 Prev->Children[0]->First->MustBreakBefore) || 1820 Current->IsMultiline) 1821 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit; 1822 else 1823 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth + 1824 ChildSize + Current->SpacesRequiredBefore; 1825 1826 if (Current->is(TT_CtorInitializerColon)) 1827 InFunctionDecl = false; 1828 1829 // FIXME: Only calculate this if CanBreakBefore is true once static 1830 // initializers etc. are sorted out. 1831 // FIXME: Move magic numbers to a better place. 1832 Current->SplitPenalty = 20 * Current->BindingStrength + 1833 splitPenalty(Line, *Current, InFunctionDecl); 1834 1835 Current = Current->Next; 1836 } 1837 1838 calculateUnbreakableTailLengths(Line); 1839 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 1840 if (Current->Role) 1841 Current->Role->precomputeFormattingInfos(Current); 1842 } 1843 1844 DEBUG({ printDebugInfo(Line); }); 1845 } 1846 1847 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 1848 unsigned UnbreakableTailLength = 0; 1849 FormatToken *Current = Line.Last; 1850 while (Current) { 1851 Current->UnbreakableTailLength = UnbreakableTailLength; 1852 if (Current->CanBreakBefore || 1853 Current->isOneOf(tok::comment, tok::string_literal)) { 1854 UnbreakableTailLength = 0; 1855 } else { 1856 UnbreakableTailLength += 1857 Current->ColumnWidth + Current->SpacesRequiredBefore; 1858 } 1859 Current = Current->Previous; 1860 } 1861 } 1862 1863 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 1864 const FormatToken &Tok, 1865 bool InFunctionDecl) { 1866 const FormatToken &Left = *Tok.Previous; 1867 const FormatToken &Right = Tok; 1868 1869 if (Left.is(tok::semi)) 1870 return 0; 1871 1872 if (Style.Language == FormatStyle::LK_Java) { 1873 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws)) 1874 return 1; 1875 if (Right.is(Keywords.kw_implements)) 1876 return 2; 1877 if (Left.is(tok::comma) && Left.NestingLevel == 0) 1878 return 3; 1879 } else if (Style.Language == FormatStyle::LK_JavaScript) { 1880 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma)) 1881 return 100; 1882 if (Left.is(TT_JsTypeColon)) 1883 return 35; 1884 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 1885 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 1886 return 100; 1887 } 1888 1889 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 1890 return 1; 1891 if (Right.is(tok::l_square)) { 1892 if (Style.Language == FormatStyle::LK_Proto) 1893 return 1; 1894 if (Left.is(tok::r_square)) 1895 return 200; 1896 // Slightly prefer formatting local lambda definitions like functions. 1897 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal)) 1898 return 35; 1899 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 1900 TT_ArrayInitializerLSquare)) 1901 return 500; 1902 } 1903 1904 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 1905 Right.is(tok::kw_operator)) { 1906 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 1907 return 3; 1908 if (Left.is(TT_StartOfName)) 1909 return 110; 1910 if (InFunctionDecl && Right.NestingLevel == 0) 1911 return Style.PenaltyReturnTypeOnItsOwnLine; 1912 return 200; 1913 } 1914 if (Right.is(TT_PointerOrReference)) 1915 return 190; 1916 if (Right.is(TT_LambdaArrow)) 1917 return 110; 1918 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 1919 return 150; 1920 if (Left.is(TT_CastRParen)) 1921 return 100; 1922 if (Left.is(tok::coloncolon) || 1923 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 1924 return 500; 1925 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 1926 return 5000; 1927 if (Left.is(tok::comment)) 1928 return 1000; 1929 1930 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon)) 1931 return 2; 1932 1933 if (Right.isMemberAccess()) { 1934 // Breaking before the "./->" of a chained call/member access is reasonably 1935 // cheap, as formatting those with one call per line is generally 1936 // desirable. In particular, it should be cheaper to break before the call 1937 // than it is to break inside a call's parameters, which could lead to weird 1938 // "hanging" indents. The exception is the very last "./->" to support this 1939 // frequent pattern: 1940 // 1941 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc( 1942 // dddddddd); 1943 // 1944 // which might otherwise be blown up onto many lines. Here, clang-format 1945 // won't produce "hanging" indents anyway as there is no other trailing 1946 // call. 1947 // 1948 // Also apply higher penalty is not a call as that might lead to a wrapping 1949 // like: 1950 // 1951 // aaaaaaa 1952 // .aaaaaaaaa.bbbbbbbb(cccccccc); 1953 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope() 1954 ? 150 1955 : 35; 1956 } 1957 1958 if (Right.is(TT_TrailingAnnotation) && 1959 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 1960 // Moving trailing annotations to the next line is fine for ObjC method 1961 // declarations. 1962 if (Line.startsWith(TT_ObjCMethodSpecifier)) 1963 return 10; 1964 // Generally, breaking before a trailing annotation is bad unless it is 1965 // function-like. It seems to be especially preferable to keep standard 1966 // annotations (i.e. "const", "final" and "override") on the same line. 1967 // Use a slightly higher penalty after ")" so that annotations like 1968 // "const override" are kept together. 1969 bool is_short_annotation = Right.TokenText.size() < 10; 1970 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 1971 } 1972 1973 // In for-loops, prefer breaking at ',' and ';'. 1974 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal)) 1975 return 4; 1976 1977 // In Objective-C method expressions, prefer breaking before "param:" over 1978 // breaking after it. 1979 if (Right.is(TT_SelectorName)) 1980 return 0; 1981 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr)) 1982 return Line.MightBeFunctionDecl ? 50 : 500; 1983 1984 if (Left.is(tok::l_paren) && InFunctionDecl && 1985 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) 1986 return 100; 1987 if (Left.is(tok::l_paren) && Left.Previous && 1988 Left.Previous->isOneOf(tok::kw_if, tok::kw_for)) 1989 return 1000; 1990 if (Left.is(tok::equal) && InFunctionDecl) 1991 return 110; 1992 if (Right.is(tok::r_brace)) 1993 return 1; 1994 if (Left.is(TT_TemplateOpener)) 1995 return 100; 1996 if (Left.opensScope()) { 1997 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign) 1998 return 0; 1999 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 2000 : 19; 2001 } 2002 if (Left.is(TT_JavaAnnotation)) 2003 return 50; 2004 2005 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous && 2006 Left.Previous->isLabelString() && 2007 (Left.NextOperator || Left.OperatorIndex != 0)) 2008 return 45; 2009 if (Right.is(tok::plus) && Left.isLabelString() && 2010 (Right.NextOperator || Right.OperatorIndex != 0)) 2011 return 25; 2012 if (Left.is(tok::comma)) 2013 return 1; 2014 if (Right.is(tok::lessless) && Left.isLabelString() && 2015 (Right.NextOperator || Right.OperatorIndex != 1)) 2016 return 25; 2017 if (Right.is(tok::lessless)) { 2018 // Breaking at a << is really cheap. 2019 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0) 2020 // Slightly prefer to break before the first one in log-like statements. 2021 return 2; 2022 return 1; 2023 } 2024 if (Left.is(TT_ConditionalExpr)) 2025 return prec::Conditional; 2026 prec::Level Level = Left.getPrecedence(); 2027 if (Level != prec::Unknown) 2028 return Level; 2029 Level = Right.getPrecedence(); 2030 if (Level != prec::Unknown) 2031 return Level; 2032 2033 return 3; 2034 } 2035 2036 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 2037 const FormatToken &Left, 2038 const FormatToken &Right) { 2039 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 2040 return true; 2041 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 2042 Left.Tok.getObjCKeywordID() == tok::objc_property) 2043 return true; 2044 if (Right.is(tok::hashhash)) 2045 return Left.is(tok::hash); 2046 if (Left.isOneOf(tok::hashhash, tok::hash)) 2047 return Right.is(tok::hash); 2048 if (Left.is(tok::l_paren) && Right.is(tok::r_paren)) 2049 return Style.SpaceInEmptyParentheses; 2050 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 2051 return (Right.is(TT_CastRParen) || 2052 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen))) 2053 ? Style.SpacesInCStyleCastParentheses 2054 : Style.SpacesInParentheses; 2055 if (Right.isOneOf(tok::semi, tok::comma)) 2056 return false; 2057 if (Right.is(tok::less) && 2058 Line.Type == LT_ObjCDecl && Style.ObjCSpaceBeforeProtocolList) 2059 return true; 2060 if (Right.is(tok::less) && Left.is(tok::kw_template)) 2061 return Style.SpaceAfterTemplateKeyword; 2062 if (Left.isOneOf(tok::exclaim, tok::tilde)) 2063 return false; 2064 if (Left.is(tok::at) && 2065 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 2066 tok::numeric_constant, tok::l_paren, tok::l_brace, 2067 tok::kw_true, tok::kw_false)) 2068 return false; 2069 if (Left.is(tok::colon)) 2070 return !Left.is(TT_ObjCMethodExpr); 2071 if (Left.is(tok::coloncolon)) 2072 return false; 2073 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) 2074 return false; 2075 if (Right.is(tok::ellipsis)) 2076 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous && 2077 Left.Previous->is(tok::kw_case)); 2078 if (Left.is(tok::l_square) && Right.is(tok::amp)) 2079 return false; 2080 if (Right.is(TT_PointerOrReference)) 2081 return (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) || 2082 (Left.Tok.isLiteral() || (Left.is(tok::kw_const) && Left.Previous && 2083 Left.Previous->is(tok::r_paren)) || 2084 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) && 2085 (Style.PointerAlignment != FormatStyle::PAS_Left || 2086 (Line.IsMultiVariableDeclStmt && 2087 (Left.NestingLevel == 0 || 2088 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for))))))); 2089 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) && 2090 (!Left.is(TT_PointerOrReference) || 2091 (Style.PointerAlignment != FormatStyle::PAS_Right && 2092 !Line.IsMultiVariableDeclStmt))) 2093 return true; 2094 if (Left.is(TT_PointerOrReference)) 2095 return Right.Tok.isLiteral() || Right.is(TT_BlockComment) || 2096 (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) && 2097 !Right.is(TT_StartOfName)) || 2098 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) || 2099 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare, 2100 tok::l_paren) && 2101 (Style.PointerAlignment != FormatStyle::PAS_Right && 2102 !Line.IsMultiVariableDeclStmt) && 2103 Left.Previous && 2104 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon)); 2105 if (Right.is(tok::star) && Left.is(tok::l_paren)) 2106 return false; 2107 if (Left.is(tok::l_square)) 2108 return (Left.is(TT_ArrayInitializerLSquare) && 2109 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square)) || 2110 (Left.is(TT_ArraySubscriptLSquare) && Style.SpacesInSquareBrackets && 2111 Right.isNot(tok::r_square)); 2112 if (Right.is(tok::r_square)) 2113 return Right.MatchingParen && 2114 ((Style.SpacesInContainerLiterals && 2115 Right.MatchingParen->is(TT_ArrayInitializerLSquare)) || 2116 (Style.SpacesInSquareBrackets && 2117 Right.MatchingParen->is(TT_ArraySubscriptLSquare))); 2118 if (Right.is(tok::l_square) && 2119 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare) && 2120 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral)) 2121 return false; 2122 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 2123 return !Left.Children.empty(); // No spaces in "{}". 2124 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 2125 (Right.is(tok::r_brace) && Right.MatchingParen && 2126 Right.MatchingParen->BlockKind != BK_Block)) 2127 return !Style.Cpp11BracedListStyle; 2128 if (Left.is(TT_BlockComment)) 2129 return !Left.TokenText.endswith("=*/"); 2130 if (Right.is(tok::l_paren)) { 2131 if (Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) 2132 return true; 2133 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) || 2134 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 2135 (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while, 2136 tok::kw_switch, tok::kw_case, TT_ForEachMacro, 2137 TT_ObjCForIn) || 2138 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch, 2139 tok::kw_new, tok::kw_delete) && 2140 (!Left.Previous || Left.Previous->isNot(tok::period))))) || 2141 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always && 2142 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() || 2143 Left.is(tok::r_paren)) && 2144 Line.Type != LT_PreprocessorDirective); 2145 } 2146 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 2147 return false; 2148 if (Right.is(TT_UnaryOperator)) 2149 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 2150 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr)); 2151 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 2152 tok::r_paren) || 2153 Left.isSimpleTypeSpecifier()) && 2154 Right.is(tok::l_brace) && Right.getNextNonComment() && 2155 Right.BlockKind != BK_Block) 2156 return false; 2157 if (Left.is(tok::period) || Right.is(tok::period)) 2158 return false; 2159 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 2160 return false; 2161 if (Left.is(TT_TemplateCloser) && Left.MatchingParen && 2162 Left.MatchingParen->Previous && 2163 Left.MatchingParen->Previous->is(tok::period)) 2164 // A.<B>DoSomething(); 2165 return false; 2166 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square)) 2167 return false; 2168 return true; 2169 } 2170 2171 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 2172 const FormatToken &Right) { 2173 const FormatToken &Left = *Right.Previous; 2174 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo()) 2175 return true; // Never ever merge two identifiers. 2176 if (Style.Language == FormatStyle::LK_Cpp) { 2177 if (Left.is(tok::kw_operator)) 2178 return Right.is(tok::coloncolon); 2179 } else if (Style.Language == FormatStyle::LK_Proto) { 2180 if (Right.is(tok::period) && 2181 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required, 2182 Keywords.kw_repeated, Keywords.kw_extend)) 2183 return true; 2184 if (Right.is(tok::l_paren) && 2185 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) 2186 return true; 2187 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2188 if (Left.is(TT_JsFatArrow)) 2189 return true; 2190 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2191 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2192 return false; 2193 if (Left.is(tok::identifier) && Right.is(TT_TemplateString)) 2194 return false; 2195 if (Right.is(tok::star) && 2196 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) 2197 return false; 2198 if (Right.isOneOf(tok::l_brace, tok::l_square) && 2199 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) 2200 return true; 2201 // JS methods can use some keywords as names (e.g. `delete()`). 2202 if (Right.is(tok::l_paren) && Line.MustBeDeclaration && 2203 Left.Tok.getIdentifierInfo()) 2204 return false; 2205 if (Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in, 2206 Keywords.kw_of, tok::kw_const) && 2207 (!Left.Previous || !Left.Previous->is(tok::period))) 2208 return true; 2209 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous && 2210 Left.Previous->is(tok::period) && Right.is(tok::l_paren)) 2211 return false; 2212 if (Left.is(Keywords.kw_as) && 2213 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren)) 2214 return true; 2215 if (Left.is(tok::kw_default) && Left.Previous && 2216 Left.Previous->is(tok::kw_export)) 2217 return true; 2218 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace)) 2219 return true; 2220 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion)) 2221 return false; 2222 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator)) 2223 return false; 2224 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) && 2225 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) 2226 return false; 2227 if (Left.is(tok::ellipsis)) 2228 return false; 2229 if (Left.is(TT_TemplateCloser) && 2230 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square, 2231 Keywords.kw_implements, Keywords.kw_extends)) 2232 // Type assertions ('<type>expr') are not followed by whitespace. Other 2233 // locations that should have whitespace following are identified by the 2234 // above set of follower tokens. 2235 return false; 2236 // Postfix non-null assertion operator, as in `foo!.bar()`. 2237 if (Right.is(tok::exclaim) && (Left.isOneOf(tok::identifier, tok::r_paren, 2238 tok::r_square, tok::r_brace) || 2239 Left.Tok.isLiteral())) 2240 return false; 2241 if (Left.is(tok::exclaim) && Right.is(Keywords.kw_as)) 2242 return true; // "x! as string" 2243 } else if (Style.Language == FormatStyle::LK_Java) { 2244 if (Left.is(tok::r_square) && Right.is(tok::l_brace)) 2245 return true; 2246 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) 2247 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never; 2248 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private, 2249 tok::kw_protected) || 2250 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract, 2251 Keywords.kw_native)) && 2252 Right.is(TT_TemplateOpener)) 2253 return true; 2254 } 2255 if (Left.is(TT_ImplicitStringLiteral)) 2256 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2257 if (Line.Type == LT_ObjCMethodDecl) { 2258 if (Left.is(TT_ObjCMethodSpecifier)) 2259 return true; 2260 if (Left.is(tok::r_paren) && Right.is(tok::identifier)) 2261 // Don't space between ')' and <id> 2262 return false; 2263 } 2264 if (Line.Type == LT_ObjCProperty && 2265 (Right.is(tok::equal) || Left.is(tok::equal))) 2266 return false; 2267 2268 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) || 2269 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) 2270 return true; 2271 if (Right.is(TT_OverloadedOperatorLParen)) 2272 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2273 if (Left.is(tok::comma)) 2274 return true; 2275 if (Right.is(tok::comma)) 2276 return false; 2277 if (Right.isOneOf(TT_CtorInitializerColon, TT_ObjCBlockLParen)) 2278 return true; 2279 if (Right.is(tok::colon)) { 2280 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) || 2281 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi)) 2282 return false; 2283 if (Right.is(TT_ObjCMethodExpr)) 2284 return false; 2285 if (Left.is(tok::question)) 2286 return false; 2287 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon)) 2288 return false; 2289 if (Right.is(TT_DictLiteral)) 2290 return Style.SpacesInContainerLiterals; 2291 return true; 2292 } 2293 if (Left.is(TT_UnaryOperator)) 2294 return Right.is(TT_BinaryOperator); 2295 2296 // If the next token is a binary operator or a selector name, we have 2297 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly. 2298 if (Left.is(TT_CastRParen)) 2299 return Style.SpaceAfterCStyleCast || 2300 Right.isOneOf(TT_BinaryOperator, TT_SelectorName); 2301 2302 if (Left.is(tok::greater) && Right.is(tok::greater)) 2303 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) && 2304 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 2305 if (Right.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) || 2306 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar)) 2307 return false; 2308 if (!Style.SpaceBeforeAssignmentOperators && 2309 Right.getPrecedence() == prec::Assignment) 2310 return false; 2311 if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment)) 2312 return (Left.is(TT_TemplateOpener) && 2313 Style.Standard == FormatStyle::LS_Cpp03) || 2314 !(Left.isOneOf(tok::identifier, tok::l_paren, tok::r_paren, 2315 tok::l_square) || 2316 Left.isOneOf(TT_TemplateCloser, TT_TemplateOpener)); 2317 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser))) 2318 return Style.SpacesInAngles; 2319 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) || 2320 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && 2321 !Right.is(tok::r_paren))) 2322 return true; 2323 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) && 2324 Right.isNot(TT_FunctionTypeLParen)) 2325 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2326 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) && 2327 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen)) 2328 return false; 2329 if (Right.is(tok::less) && Left.isNot(tok::l_paren) && 2330 Line.startsWith(tok::hash)) 2331 return true; 2332 if (Right.is(TT_TrailingUnaryOperator)) 2333 return false; 2334 if (Left.is(TT_RegexLiteral)) 2335 return false; 2336 return spaceRequiredBetween(Line, Left, Right); 2337 } 2338 2339 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 2340 static bool isAllmanBrace(const FormatToken &Tok) { 2341 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 2342 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral); 2343 } 2344 2345 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 2346 const FormatToken &Right) { 2347 const FormatToken &Left = *Right.Previous; 2348 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0) 2349 return true; 2350 2351 if (Style.Language == FormatStyle::LK_JavaScript) { 2352 // FIXME: This might apply to other languages and token kinds. 2353 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous && 2354 Left.Previous->is(tok::string_literal)) 2355 return true; 2356 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 && 2357 Left.Previous && Left.Previous->is(tok::equal) && 2358 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export, 2359 tok::kw_const) && 2360 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match 2361 // above. 2362 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) 2363 // Object literals on the top level of a file are treated as "enum-style". 2364 // Each key/value pair is put on a separate line, instead of bin-packing. 2365 return true; 2366 if (Left.is(tok::l_brace) && Line.Level == 0 && 2367 (Line.startsWith(tok::kw_enum) || 2368 Line.startsWith(tok::kw_export, tok::kw_enum))) 2369 // JavaScript top-level enum key/value pairs are put on separate lines 2370 // instead of bin-packing. 2371 return true; 2372 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && 2373 !Left.Children.empty()) 2374 // Support AllowShortFunctionsOnASingleLine for JavaScript. 2375 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 2376 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty || 2377 (Left.NestingLevel == 0 && Line.Level == 0 && 2378 Style.AllowShortFunctionsOnASingleLine == 2379 FormatStyle::SFS_Inline); 2380 } else if (Style.Language == FormatStyle::LK_Java) { 2381 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next && 2382 Right.Next->is(tok::string_literal)) 2383 return true; 2384 } 2385 2386 // If the last token before a '}' is a comma or a trailing comment, the 2387 // intention is to insert a line break after it in order to make shuffling 2388 // around entries easier. 2389 const FormatToken *BeforeClosingBrace = nullptr; 2390 if (Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) && 2391 Left.BlockKind != BK_Block && Left.MatchingParen) 2392 BeforeClosingBrace = Left.MatchingParen->Previous; 2393 else if (Right.MatchingParen && 2394 Right.MatchingParen->isOneOf(tok::l_brace, 2395 TT_ArrayInitializerLSquare)) 2396 BeforeClosingBrace = &Left; 2397 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 2398 BeforeClosingBrace->isTrailingComment())) 2399 return true; 2400 2401 if (Right.is(tok::comment)) 2402 return Left.BlockKind != BK_BracedInit && 2403 Left.isNot(TT_CtorInitializerColon) && 2404 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 2405 if (Left.isTrailingComment()) 2406 return true; 2407 if (Left.isStringLiteral() && 2408 (Right.isStringLiteral() || Right.is(TT_ObjCStringLiteral))) 2409 return true; 2410 if (Right.Previous->IsUnterminatedLiteral) 2411 return true; 2412 if (Right.is(tok::lessless) && Right.Next && 2413 Right.Previous->is(tok::string_literal) && 2414 Right.Next->is(tok::string_literal)) 2415 return true; 2416 if (Right.Previous->ClosesTemplateDeclaration && 2417 Right.Previous->MatchingParen && 2418 Right.Previous->MatchingParen->NestingLevel == 0 && 2419 Style.AlwaysBreakTemplateDeclarations) 2420 return true; 2421 if ((Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon)) && 2422 Style.BreakConstructorInitializersBeforeComma && 2423 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 2424 return true; 2425 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) 2426 // Raw string literals are special wrt. line breaks. The author has made a 2427 // deliberate choice and might have aligned the contents of the string 2428 // literal accordingly. Thus, we try keep existing line breaks. 2429 return Right.NewlinesBefore > 0; 2430 if (Right.Previous->is(tok::l_brace) && Right.NestingLevel == 1 && 2431 Style.Language == FormatStyle::LK_Proto) 2432 // Don't put enums onto single lines in protocol buffers. 2433 return true; 2434 if (Right.is(TT_InlineASMBrace)) 2435 return Right.HasUnescapedNewline; 2436 if (isAllmanBrace(Left) || isAllmanBrace(Right)) 2437 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) || 2438 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) || 2439 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct); 2440 if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine) 2441 return true; 2442 2443 if ((Style.Language == FormatStyle::LK_Java || 2444 Style.Language == FormatStyle::LK_JavaScript) && 2445 Left.is(TT_LeadingJavaAnnotation) && 2446 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) && 2447 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) 2448 return true; 2449 2450 return false; 2451 } 2452 2453 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 2454 const FormatToken &Right) { 2455 const FormatToken &Left = *Right.Previous; 2456 2457 // Language-specific stuff. 2458 if (Style.Language == FormatStyle::LK_Java) { 2459 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2460 Keywords.kw_implements)) 2461 return false; 2462 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2463 Keywords.kw_implements)) 2464 return true; 2465 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2466 const FormatToken *NonComment = Right.getPreviousNonComment(); 2467 if (Left.isOneOf(tok::kw_return, tok::kw_continue, tok::kw_break, 2468 tok::kw_throw) || 2469 (NonComment && 2470 NonComment->isOneOf(tok::kw_return, tok::kw_continue, tok::kw_break, 2471 tok::kw_throw))) 2472 return false; // Otherwise a semicolon is inserted. 2473 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace)) 2474 return false; 2475 if (Left.is(TT_JsTypeColon)) 2476 return true; 2477 if (Right.NestingLevel == 0 && Right.is(Keywords.kw_is)) 2478 return false; 2479 if (Left.is(Keywords.kw_in)) 2480 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None; 2481 if (Right.is(Keywords.kw_in)) 2482 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None; 2483 if (Right.is(Keywords.kw_as)) 2484 return false; // must not break before as in 'x as type' casts 2485 if (Left.is(Keywords.kw_declare) && 2486 Right.isOneOf(Keywords.kw_module, tok::kw_namespace, 2487 Keywords.kw_function, tok::kw_class, tok::kw_enum, 2488 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var, 2489 Keywords.kw_let, tok::kw_const)) 2490 // See grammar for 'declare' statements at: 2491 // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10 2492 return false; 2493 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) && 2494 Right.isOneOf(tok::identifier, tok::string_literal)) { 2495 return false; // must not break in "module foo { ...}" 2496 } 2497 } 2498 2499 if (Left.is(tok::at)) 2500 return false; 2501 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 2502 return false; 2503 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation)) 2504 return !Right.is(tok::l_paren); 2505 if (Right.is(TT_PointerOrReference)) 2506 return Line.IsMultiVariableDeclStmt || 2507 (Style.PointerAlignment == FormatStyle::PAS_Right && 2508 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName))); 2509 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2510 Right.is(tok::kw_operator)) 2511 return true; 2512 if (Left.is(TT_PointerOrReference)) 2513 return false; 2514 if (Right.isTrailingComment()) 2515 // We rely on MustBreakBefore being set correctly here as we should not 2516 // change the "binding" behavior of a comment. 2517 // The first comment in a braced lists is always interpreted as belonging to 2518 // the first list element. Otherwise, it should be placed outside of the 2519 // list. 2520 return Left.BlockKind == BK_BracedInit; 2521 if (Left.is(tok::question) && Right.is(tok::colon)) 2522 return false; 2523 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question)) 2524 return Style.BreakBeforeTernaryOperators; 2525 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question)) 2526 return !Style.BreakBeforeTernaryOperators; 2527 if (Right.is(TT_InheritanceColon)) 2528 return true; 2529 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) && 2530 Left.isNot(TT_SelectorName)) 2531 return true; 2532 if (Right.is(tok::colon) && 2533 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) 2534 return false; 2535 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) 2536 return true; 2537 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next && 2538 Right.Next->is(TT_ObjCMethodExpr))) 2539 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls. 2540 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 2541 return true; 2542 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen)) 2543 return true; 2544 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen, 2545 TT_OverloadedOperator)) 2546 return false; 2547 if (Left.is(TT_RangeBasedForLoopColon)) 2548 return true; 2549 if (Right.is(TT_RangeBasedForLoopColon)) 2550 return false; 2551 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener)) 2552 return true; 2553 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) || 2554 Left.is(tok::kw_operator)) 2555 return false; 2556 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) && 2557 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) 2558 return false; 2559 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen)) 2560 return false; 2561 if (Left.is(tok::l_paren) && Left.Previous && 2562 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) 2563 return false; 2564 if (Right.is(TT_ImplicitStringLiteral)) 2565 return false; 2566 2567 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser)) 2568 return false; 2569 if (Right.is(tok::r_square) && Right.MatchingParen && 2570 Right.MatchingParen->is(TT_LambdaLSquare)) 2571 return false; 2572 2573 // We only break before r_brace if there was a corresponding break before 2574 // the l_brace, which is tracked by BreakBeforeClosingBrace. 2575 if (Right.is(tok::r_brace)) 2576 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 2577 2578 // Allow breaking after a trailing annotation, e.g. after a method 2579 // declaration. 2580 if (Left.is(TT_TrailingAnnotation)) 2581 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 2582 tok::less, tok::coloncolon); 2583 2584 if (Right.is(tok::kw___attribute)) 2585 return true; 2586 2587 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 2588 return true; 2589 2590 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2591 return true; 2592 2593 if (Left.is(TT_CtorInitializerComma) && 2594 Style.BreakConstructorInitializersBeforeComma) 2595 return false; 2596 if (Right.is(TT_CtorInitializerComma) && 2597 Style.BreakConstructorInitializersBeforeComma) 2598 return true; 2599 if ((Left.is(tok::greater) && Right.is(tok::greater)) || 2600 (Left.is(tok::less) && Right.is(tok::less))) 2601 return false; 2602 if (Right.is(TT_BinaryOperator) && 2603 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 2604 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 2605 Right.getPrecedence() != prec::Assignment)) 2606 return true; 2607 if (Left.is(TT_ArrayInitializerLSquare)) 2608 return true; 2609 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 2610 return true; 2611 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) && 2612 !Left.isOneOf(tok::arrowstar, tok::lessless) && 2613 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 2614 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 2615 Left.getPrecedence() == prec::Assignment)) 2616 return true; 2617 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 2618 tok::kw_class, tok::kw_struct, tok::comment) || 2619 Right.isMemberAccess() || 2620 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless, 2621 tok::colon, tok::l_square, tok::at) || 2622 (Left.is(tok::r_paren) && 2623 Right.isOneOf(tok::identifier, tok::kw_const)) || 2624 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)); 2625 } 2626 2627 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 2628 llvm::errs() << "AnnotatedTokens:\n"; 2629 const FormatToken *Tok = Line.First; 2630 while (Tok) { 2631 llvm::errs() << " M=" << Tok->MustBreakBefore 2632 << " C=" << Tok->CanBreakBefore 2633 << " T=" << getTokenTypeName(Tok->Type) 2634 << " S=" << Tok->SpacesRequiredBefore 2635 << " B=" << Tok->BlockParameterCount 2636 << " P=" << Tok->SplitPenalty << " Name=" << Tok->Tok.getName() 2637 << " L=" << Tok->TotalLength << " PPK=" << Tok->PackingKind 2638 << " FakeLParens="; 2639 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 2640 llvm::errs() << Tok->FakeLParens[i] << "/"; 2641 llvm::errs() << " FakeRParens=" << Tok->FakeRParens; 2642 llvm::errs() << " Text='" << Tok->TokenText << "'\n"; 2643 if (!Tok->Next) 2644 assert(Tok == Line.Last); 2645 Tok = Tok->Next; 2646 } 2647 llvm::errs() << "----\n"; 2648 } 2649 2650 } // namespace format 2651 } // namespace clang 2652