1 //===--- TokenAnnotator.cpp - Format C++ code -----------------------------===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is distributed under the University of Illinois Open Source 6 // License. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 /// 10 /// \file 11 /// \brief This file implements a token annotator, i.e. creates 12 /// \c AnnotatedTokens out of \c FormatTokens with required extra information. 13 /// 14 //===----------------------------------------------------------------------===// 15 16 #include "TokenAnnotator.h" 17 #include "clang/Basic/SourceManager.h" 18 #include "llvm/ADT/SmallPtrSet.h" 19 #include "llvm/Support/Debug.h" 20 21 #define DEBUG_TYPE "format-token-annotator" 22 23 namespace clang { 24 namespace format { 25 26 namespace { 27 28 /// \brief A parser that gathers additional information about tokens. 29 /// 30 /// The \c TokenAnnotator tries to match parenthesis and square brakets and 31 /// store a parenthesis levels. It also tries to resolve matching "<" and ">" 32 /// into template parameter lists. 33 class AnnotatingParser { 34 public: 35 AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line, 36 const AdditionalKeywords &Keywords) 37 : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false), 38 Keywords(Keywords) { 39 Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false)); 40 resetTokenMetadata(CurrentToken); 41 } 42 43 private: 44 bool parseAngle() { 45 if (!CurrentToken || !CurrentToken->Previous) 46 return false; 47 if (NonTemplateLess.count(CurrentToken->Previous)) 48 return false; 49 50 const FormatToken& Previous = *CurrentToken->Previous; 51 if (Previous.Previous) { 52 if (Previous.Previous->Tok.isLiteral()) 53 return false; 54 if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 && 55 (!Previous.Previous->MatchingParen || 56 !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen))) 57 return false; 58 } 59 60 FormatToken *Left = CurrentToken->Previous; 61 Left->ParentBracket = Contexts.back().ContextKind; 62 ScopedContextCreator ContextCreator(*this, tok::less, 12); 63 64 // If this angle is in the context of an expression, we need to be more 65 // hesitant to detect it as opening template parameters. 66 bool InExprContext = Contexts.back().IsExpression; 67 68 Contexts.back().IsExpression = false; 69 // If there's a template keyword before the opening angle bracket, this is a 70 // template parameter, not an argument. 71 Contexts.back().InTemplateArgument = 72 Left->Previous && Left->Previous->Tok.isNot(tok::kw_template); 73 74 if (Style.Language == FormatStyle::LK_Java && 75 CurrentToken->is(tok::question)) 76 next(); 77 78 while (CurrentToken) { 79 if (CurrentToken->is(tok::greater)) { 80 Left->MatchingParen = CurrentToken; 81 CurrentToken->MatchingParen = Left; 82 CurrentToken->Type = TT_TemplateCloser; 83 next(); 84 return true; 85 } 86 if (CurrentToken->is(tok::question) && 87 Style.Language == FormatStyle::LK_Java) { 88 next(); 89 continue; 90 } 91 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) || 92 (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext)) 93 return false; 94 // If a && or || is found and interpreted as a binary operator, this set 95 // of angles is likely part of something like "a < b && c > d". If the 96 // angles are inside an expression, the ||/&& might also be a binary 97 // operator that was misinterpreted because we are parsing template 98 // parameters. 99 // FIXME: This is getting out of hand, write a decent parser. 100 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) && 101 CurrentToken->Previous->is(TT_BinaryOperator) && 102 Contexts[Contexts.size() - 2].IsExpression && 103 !Line.startsWith(tok::kw_template)) 104 return false; 105 updateParameterCount(Left, CurrentToken); 106 if (!consumeToken()) 107 return false; 108 } 109 return false; 110 } 111 112 bool parseParens(bool LookForDecls = false) { 113 if (!CurrentToken) 114 return false; 115 FormatToken *Left = CurrentToken->Previous; 116 Left->ParentBracket = Contexts.back().ContextKind; 117 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1); 118 119 // FIXME: This is a bit of a hack. Do better. 120 Contexts.back().ColonIsForRangeExpr = 121 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr; 122 123 bool StartsObjCMethodExpr = false; 124 if (CurrentToken->is(tok::caret)) { 125 // (^ can start a block type. 126 Left->Type = TT_ObjCBlockLParen; 127 } else if (FormatToken *MaybeSel = Left->Previous) { 128 // @selector( starts a selector. 129 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous && 130 MaybeSel->Previous->is(tok::at)) { 131 StartsObjCMethodExpr = true; 132 } 133 } 134 135 if (Left->is(TT_OverloadedOperatorLParen)) { 136 Contexts.back().IsExpression = false; 137 } else if (Left->Previous && 138 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype, 139 tok::kw_if, tok::kw_while, tok::l_paren, 140 tok::comma) || 141 Left->Previous->is(TT_BinaryOperator))) { 142 // static_assert, if and while usually contain expressions. 143 Contexts.back().IsExpression = true; 144 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 145 Left->Previous->is(Keywords.kw_function)) { 146 Contexts.back().IsExpression = false; 147 } else if (Left->Previous && Left->Previous->is(tok::r_square) && 148 Left->Previous->MatchingParen && 149 Left->Previous->MatchingParen->is(TT_LambdaLSquare)) { 150 // This is a parameter list of a lambda expression. 151 Contexts.back().IsExpression = false; 152 } else if (Line.InPPDirective && 153 (!Left->Previous || !Left->Previous->is(tok::identifier))) { 154 Contexts.back().IsExpression = true; 155 } else if (Contexts[Contexts.size() - 2].CaretFound) { 156 // This is the parameter list of an ObjC block. 157 Contexts.back().IsExpression = false; 158 } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) { 159 Left->Type = TT_AttributeParen; 160 } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) { 161 // The first argument to a foreach macro is a declaration. 162 Contexts.back().IsForEachMacro = true; 163 Contexts.back().IsExpression = false; 164 } else if (Left->Previous && Left->Previous->MatchingParen && 165 Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) { 166 Contexts.back().IsExpression = false; 167 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) { 168 bool IsForOrCatch = 169 Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch); 170 Contexts.back().IsExpression = !IsForOrCatch; 171 } 172 173 if (StartsObjCMethodExpr) { 174 Contexts.back().ColonIsObjCMethodExpr = true; 175 Left->Type = TT_ObjCMethodExpr; 176 } 177 178 bool MightBeFunctionType = CurrentToken->isOneOf(tok::star, tok::amp) && 179 !Contexts[Contexts.size() - 2].IsExpression; 180 bool HasMultipleLines = false; 181 bool HasMultipleParametersOnALine = false; 182 bool MightBeObjCForRangeLoop = 183 Left->Previous && Left->Previous->is(tok::kw_for); 184 while (CurrentToken) { 185 // LookForDecls is set when "if (" has been seen. Check for 186 // 'identifier' '*' 'identifier' followed by not '=' -- this 187 // '*' has to be a binary operator but determineStarAmpUsage() will 188 // categorize it as an unary operator, so set the right type here. 189 if (LookForDecls && CurrentToken->Next) { 190 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 191 if (Prev) { 192 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 193 FormatToken *Next = CurrentToken->Next; 194 if (PrevPrev && PrevPrev->is(tok::identifier) && 195 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 196 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 197 Prev->Type = TT_BinaryOperator; 198 LookForDecls = false; 199 } 200 } 201 } 202 203 if (CurrentToken->Previous->is(TT_PointerOrReference) && 204 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 205 tok::coloncolon)) 206 MightBeFunctionType = true; 207 if (CurrentToken->Previous->is(TT_BinaryOperator)) 208 Contexts.back().IsExpression = true; 209 if (CurrentToken->is(tok::r_paren)) { 210 if (MightBeFunctionType && CurrentToken->Next && 211 (CurrentToken->Next->is(tok::l_paren) || 212 (CurrentToken->Next->is(tok::l_square) && 213 Line.MustBeDeclaration))) 214 Left->Type = TT_FunctionTypeLParen; 215 Left->MatchingParen = CurrentToken; 216 CurrentToken->MatchingParen = Left; 217 218 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) && 219 Left->Previous && Left->Previous->is(tok::l_paren)) { 220 // Detect the case where macros are used to generate lambdas or 221 // function bodies, e.g.: 222 // auto my_lambda = MARCO((Type *type, int i) { .. body .. }); 223 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) { 224 if (Tok->is(TT_BinaryOperator) && 225 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) 226 Tok->Type = TT_PointerOrReference; 227 } 228 } 229 230 if (StartsObjCMethodExpr) { 231 CurrentToken->Type = TT_ObjCMethodExpr; 232 if (Contexts.back().FirstObjCSelectorName) { 233 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 234 Contexts.back().LongestObjCSelectorName; 235 } 236 } 237 238 if (Left->is(TT_AttributeParen)) 239 CurrentToken->Type = TT_AttributeParen; 240 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation)) 241 CurrentToken->Type = TT_JavaAnnotation; 242 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation)) 243 CurrentToken->Type = TT_LeadingJavaAnnotation; 244 245 if (!HasMultipleLines) 246 Left->PackingKind = PPK_Inconclusive; 247 else if (HasMultipleParametersOnALine) 248 Left->PackingKind = PPK_BinPacked; 249 else 250 Left->PackingKind = PPK_OnePerLine; 251 252 next(); 253 return true; 254 } 255 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 256 return false; 257 258 if (CurrentToken->is(tok::l_brace)) 259 Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen 260 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 261 !CurrentToken->Next->HasUnescapedNewline && 262 !CurrentToken->Next->isTrailingComment()) 263 HasMultipleParametersOnALine = true; 264 if (CurrentToken->isOneOf(tok::kw_const, tok::kw_auto) || 265 CurrentToken->isSimpleTypeSpecifier()) 266 Contexts.back().IsExpression = false; 267 if (CurrentToken->isOneOf(tok::semi, tok::colon)) 268 MightBeObjCForRangeLoop = false; 269 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) 270 CurrentToken->Type = TT_ObjCForIn; 271 // When we discover a 'new', we set CanBeExpression to 'false' in order to 272 // parse the type correctly. Reset that after a comma. 273 if (CurrentToken->is(tok::comma)) 274 Contexts.back().CanBeExpression = true; 275 276 FormatToken *Tok = CurrentToken; 277 if (!consumeToken()) 278 return false; 279 updateParameterCount(Left, Tok); 280 if (CurrentToken && CurrentToken->HasUnescapedNewline) 281 HasMultipleLines = true; 282 } 283 return false; 284 } 285 286 bool parseSquare() { 287 if (!CurrentToken) 288 return false; 289 290 // A '[' could be an index subscript (after an identifier or after 291 // ')' or ']'), it could be the start of an Objective-C method 292 // expression, or it could the start of an Objective-C array literal. 293 FormatToken *Left = CurrentToken->Previous; 294 Left->ParentBracket = Contexts.back().ContextKind; 295 FormatToken *Parent = Left->getPreviousNonComment(); 296 bool StartsObjCMethodExpr = 297 Style.Language == FormatStyle::LK_Cpp && 298 Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) && 299 CurrentToken->isNot(tok::l_brace) && 300 (!Parent || 301 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 302 tok::kw_return, tok::kw_throw) || 303 Parent->isUnaryOperator() || 304 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) || 305 getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown); 306 bool ColonFound = false; 307 308 unsigned BindingIncrease = 1; 309 if (Left->is(TT_Unknown)) { 310 if (StartsObjCMethodExpr) { 311 Left->Type = TT_ObjCMethodExpr; 312 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent && 313 Contexts.back().ContextKind == tok::l_brace && 314 Parent->isOneOf(tok::l_brace, tok::comma)) { 315 Left->Type = TT_JsComputedPropertyName; 316 } else if (Style.Language == FormatStyle::LK_Proto || 317 (Parent && 318 Parent->isOneOf(TT_BinaryOperator, tok::at, tok::comma, 319 tok::l_paren, tok::l_square, tok::question, 320 tok::colon, tok::kw_return, 321 // Should only be relevant to JavaScript: 322 tok::kw_default))) { 323 Left->Type = TT_ArrayInitializerLSquare; 324 } else { 325 BindingIncrease = 10; 326 Left->Type = TT_ArraySubscriptLSquare; 327 } 328 } 329 330 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease); 331 Contexts.back().IsExpression = true; 332 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr; 333 334 while (CurrentToken) { 335 if (CurrentToken->is(tok::r_square)) { 336 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_paren) && 337 Left->is(TT_ObjCMethodExpr)) { 338 // An ObjC method call is rarely followed by an open parenthesis. 339 // FIXME: Do we incorrectly label ":" with this? 340 StartsObjCMethodExpr = false; 341 Left->Type = TT_Unknown; 342 } 343 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) { 344 CurrentToken->Type = TT_ObjCMethodExpr; 345 // determineStarAmpUsage() thinks that '*' '[' is allocating an 346 // array of pointers, but if '[' starts a selector then '*' is a 347 // binary operator. 348 if (Parent && Parent->is(TT_PointerOrReference)) 349 Parent->Type = TT_BinaryOperator; 350 } 351 Left->MatchingParen = CurrentToken; 352 CurrentToken->MatchingParen = Left; 353 if (Contexts.back().FirstObjCSelectorName) { 354 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 355 Contexts.back().LongestObjCSelectorName; 356 if (Left->BlockParameterCount > 1) 357 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0; 358 } 359 next(); 360 return true; 361 } 362 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace)) 363 return false; 364 if (CurrentToken->is(tok::colon)) { 365 if (Left->is(TT_ArraySubscriptLSquare)) { 366 Left->Type = TT_ObjCMethodExpr; 367 StartsObjCMethodExpr = true; 368 Contexts.back().ColonIsObjCMethodExpr = true; 369 if (Parent && Parent->is(tok::r_paren)) 370 Parent->Type = TT_CastRParen; 371 } 372 ColonFound = true; 373 } 374 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) && 375 !ColonFound) 376 Left->Type = TT_ArrayInitializerLSquare; 377 FormatToken *Tok = CurrentToken; 378 if (!consumeToken()) 379 return false; 380 updateParameterCount(Left, Tok); 381 } 382 return false; 383 } 384 385 bool parseBrace() { 386 if (CurrentToken) { 387 FormatToken *Left = CurrentToken->Previous; 388 Left->ParentBracket = Contexts.back().ContextKind; 389 390 if (Contexts.back().CaretFound) 391 Left->Type = TT_ObjCBlockLBrace; 392 Contexts.back().CaretFound = false; 393 394 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1); 395 Contexts.back().ColonIsDictLiteral = true; 396 if (Left->BlockKind == BK_BracedInit) 397 Contexts.back().IsExpression = true; 398 399 while (CurrentToken) { 400 if (CurrentToken->is(tok::r_brace)) { 401 Left->MatchingParen = CurrentToken; 402 CurrentToken->MatchingParen = Left; 403 next(); 404 return true; 405 } 406 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square)) 407 return false; 408 updateParameterCount(Left, CurrentToken); 409 if (CurrentToken->isOneOf(tok::colon, tok::l_brace)) { 410 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 411 if (((CurrentToken->is(tok::colon) && 412 (!Contexts.back().ColonIsDictLiteral || 413 Style.Language != FormatStyle::LK_Cpp)) || 414 Style.Language == FormatStyle::LK_Proto) && 415 (Previous->Tok.getIdentifierInfo() || 416 Previous->is(tok::string_literal))) 417 Previous->Type = TT_SelectorName; 418 if (CurrentToken->is(tok::colon) || 419 Style.Language == FormatStyle::LK_JavaScript) 420 Left->Type = TT_DictLiteral; 421 } 422 if (!consumeToken()) 423 return false; 424 } 425 } 426 return true; 427 } 428 429 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 430 if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block) 431 ++Left->BlockParameterCount; 432 if (Current->is(tok::comma)) { 433 ++Left->ParameterCount; 434 if (!Left->Role) 435 Left->Role.reset(new CommaSeparatedList(Style)); 436 Left->Role->CommaFound(Current); 437 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 438 Left->ParameterCount = 1; 439 } 440 } 441 442 bool parseConditional() { 443 while (CurrentToken) { 444 if (CurrentToken->is(tok::colon)) { 445 CurrentToken->Type = TT_ConditionalExpr; 446 next(); 447 return true; 448 } 449 if (!consumeToken()) 450 return false; 451 } 452 return false; 453 } 454 455 bool parseTemplateDeclaration() { 456 if (CurrentToken && CurrentToken->is(tok::less)) { 457 CurrentToken->Type = TT_TemplateOpener; 458 next(); 459 if (!parseAngle()) 460 return false; 461 if (CurrentToken) 462 CurrentToken->Previous->ClosesTemplateDeclaration = true; 463 return true; 464 } 465 return false; 466 } 467 468 bool consumeToken() { 469 FormatToken *Tok = CurrentToken; 470 next(); 471 switch (Tok->Tok.getKind()) { 472 case tok::plus: 473 case tok::minus: 474 if (!Tok->Previous && Line.MustBeDeclaration) 475 Tok->Type = TT_ObjCMethodSpecifier; 476 break; 477 case tok::colon: 478 if (!Tok->Previous) 479 return false; 480 // Colons from ?: are handled in parseConditional(). 481 if (Style.Language == FormatStyle::LK_JavaScript) { 482 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop 483 (Contexts.size() == 1 && // switch/case labels 484 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) || 485 Contexts.back().ContextKind == tok::l_paren || // function params 486 Contexts.back().ContextKind == tok::l_square || // array type 487 (Contexts.size() == 1 && 488 Line.MustBeDeclaration)) { // method/property declaration 489 Tok->Type = TT_JsTypeColon; 490 break; 491 } 492 } 493 if (Contexts.back().ColonIsDictLiteral || 494 Style.Language == FormatStyle::LK_Proto) { 495 Tok->Type = TT_DictLiteral; 496 } else if (Contexts.back().ColonIsObjCMethodExpr || 497 Line.startsWith(TT_ObjCMethodSpecifier)) { 498 Tok->Type = TT_ObjCMethodExpr; 499 Tok->Previous->Type = TT_SelectorName; 500 if (Tok->Previous->ColumnWidth > 501 Contexts.back().LongestObjCSelectorName) 502 Contexts.back().LongestObjCSelectorName = Tok->Previous->ColumnWidth; 503 if (!Contexts.back().FirstObjCSelectorName) 504 Contexts.back().FirstObjCSelectorName = Tok->Previous; 505 } else if (Contexts.back().ColonIsForRangeExpr) { 506 Tok->Type = TT_RangeBasedForLoopColon; 507 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 508 Tok->Type = TT_BitFieldColon; 509 } else if (Contexts.size() == 1 && 510 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) { 511 if (Tok->Previous->isOneOf(tok::r_paren, tok::kw_noexcept)) 512 Tok->Type = TT_CtorInitializerColon; 513 else 514 Tok->Type = TT_InheritanceColon; 515 } else if (Tok->Previous->is(tok::identifier) && Tok->Next && 516 Tok->Next->isOneOf(tok::r_paren, tok::comma)) { 517 // This handles a special macro in ObjC code where selectors including 518 // the colon are passed as macro arguments. 519 Tok->Type = TT_ObjCMethodExpr; 520 } else if (Contexts.back().ContextKind == tok::l_paren) { 521 Tok->Type = TT_InlineASMColon; 522 } 523 break; 524 case tok::pipe: 525 case tok::amp: 526 // | and & in declarations/type expressions represent union and 527 // intersection types, respectively. 528 if (Style.Language == FormatStyle::LK_JavaScript && 529 !Contexts.back().IsExpression) 530 Tok->Type = TT_JsTypeOperator; 531 break; 532 case tok::kw_if: 533 case tok::kw_while: 534 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 535 next(); 536 if (!parseParens(/*LookForDecls=*/true)) 537 return false; 538 } 539 break; 540 case tok::kw_for: 541 Contexts.back().ColonIsForRangeExpr = true; 542 next(); 543 if (!parseParens()) 544 return false; 545 break; 546 case tok::l_paren: 547 // When faced with 'operator()()', the kw_operator handler incorrectly 548 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make 549 // the first two parens OverloadedOperators and the second l_paren an 550 // OverloadedOperatorLParen. 551 if (Tok->Previous && 552 Tok->Previous->is(tok::r_paren) && 553 Tok->Previous->MatchingParen && 554 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) { 555 Tok->Previous->Type = TT_OverloadedOperator; 556 Tok->Previous->MatchingParen->Type = TT_OverloadedOperator; 557 Tok->Type = TT_OverloadedOperatorLParen; 558 } 559 560 if (!parseParens()) 561 return false; 562 if (Line.MustBeDeclaration && Contexts.size() == 1 && 563 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) && 564 (!Tok->Previous || 565 !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute, 566 TT_LeadingJavaAnnotation))) 567 Line.MightBeFunctionDecl = true; 568 break; 569 case tok::l_square: 570 if (!parseSquare()) 571 return false; 572 break; 573 case tok::l_brace: 574 if (!parseBrace()) 575 return false; 576 break; 577 case tok::less: 578 if (parseAngle()) { 579 Tok->Type = TT_TemplateOpener; 580 } else { 581 Tok->Type = TT_BinaryOperator; 582 NonTemplateLess.insert(Tok); 583 CurrentToken = Tok; 584 next(); 585 } 586 break; 587 case tok::r_paren: 588 case tok::r_square: 589 return false; 590 case tok::r_brace: 591 // Lines can start with '}'. 592 if (Tok->Previous) 593 return false; 594 break; 595 case tok::greater: 596 Tok->Type = TT_BinaryOperator; 597 break; 598 case tok::kw_operator: 599 while (CurrentToken && 600 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 601 if (CurrentToken->isOneOf(tok::star, tok::amp)) 602 CurrentToken->Type = TT_PointerOrReference; 603 consumeToken(); 604 if (CurrentToken && 605 CurrentToken->Previous->isOneOf(TT_BinaryOperator, tok::comma)) 606 CurrentToken->Previous->Type = TT_OverloadedOperator; 607 } 608 if (CurrentToken) { 609 CurrentToken->Type = TT_OverloadedOperatorLParen; 610 if (CurrentToken->Previous->is(TT_BinaryOperator)) 611 CurrentToken->Previous->Type = TT_OverloadedOperator; 612 } 613 break; 614 case tok::question: 615 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next && 616 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren, 617 tok::r_brace)) { 618 // Question marks before semicolons, colons, etc. indicate optional 619 // types (fields, parameters), e.g. 620 // function(x?: string, y?) {...} 621 // class X { y?; } 622 Tok->Type = TT_JsTypeOptionalQuestion; 623 break; 624 } 625 // Declarations cannot be conditional expressions, this can only be part 626 // of a type declaration. 627 if (Line.MustBeDeclaration && 628 Style.Language == FormatStyle::LK_JavaScript) 629 break; 630 parseConditional(); 631 break; 632 case tok::kw_template: 633 parseTemplateDeclaration(); 634 break; 635 case tok::comma: 636 if (Contexts.back().InCtorInitializer) 637 Tok->Type = TT_CtorInitializerComma; 638 else if (Contexts.back().FirstStartOfName && 639 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) { 640 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 641 Line.IsMultiVariableDeclStmt = true; 642 } 643 if (Contexts.back().IsForEachMacro) 644 Contexts.back().IsExpression = true; 645 break; 646 default: 647 break; 648 } 649 return true; 650 } 651 652 void parseIncludeDirective() { 653 if (CurrentToken && CurrentToken->is(tok::less)) { 654 next(); 655 while (CurrentToken) { 656 if (CurrentToken->isNot(tok::comment) || CurrentToken->Next) 657 CurrentToken->Type = TT_ImplicitStringLiteral; 658 next(); 659 } 660 } 661 } 662 663 void parseWarningOrError() { 664 next(); 665 // We still want to format the whitespace left of the first token of the 666 // warning or error. 667 next(); 668 while (CurrentToken) { 669 CurrentToken->Type = TT_ImplicitStringLiteral; 670 next(); 671 } 672 } 673 674 void parsePragma() { 675 next(); // Consume "pragma". 676 if (CurrentToken && 677 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) { 678 bool IsMark = CurrentToken->is(Keywords.kw_mark); 679 next(); // Consume "mark". 680 next(); // Consume first token (so we fix leading whitespace). 681 while (CurrentToken) { 682 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator)) 683 CurrentToken->Type = TT_ImplicitStringLiteral; 684 next(); 685 } 686 } 687 } 688 689 LineType parsePreprocessorDirective() { 690 LineType Type = LT_PreprocessorDirective; 691 next(); 692 if (!CurrentToken) 693 return Type; 694 if (CurrentToken->Tok.is(tok::numeric_constant)) { 695 CurrentToken->SpacesRequiredBefore = 1; 696 return Type; 697 } 698 // Hashes in the middle of a line can lead to any strange token 699 // sequence. 700 if (!CurrentToken->Tok.getIdentifierInfo()) 701 return Type; 702 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 703 case tok::pp_include: 704 case tok::pp_include_next: 705 case tok::pp_import: 706 next(); 707 parseIncludeDirective(); 708 Type = LT_ImportStatement; 709 break; 710 case tok::pp_error: 711 case tok::pp_warning: 712 parseWarningOrError(); 713 break; 714 case tok::pp_pragma: 715 parsePragma(); 716 break; 717 case tok::pp_if: 718 case tok::pp_elif: 719 Contexts.back().IsExpression = true; 720 parseLine(); 721 break; 722 default: 723 break; 724 } 725 while (CurrentToken) 726 next(); 727 return Type; 728 } 729 730 public: 731 LineType parseLine() { 732 NonTemplateLess.clear(); 733 if (CurrentToken->is(tok::hash)) 734 return parsePreprocessorDirective(); 735 736 // Directly allow to 'import <string-literal>' to support protocol buffer 737 // definitions (code.google.com/p/protobuf) or missing "#" (either way we 738 // should not break the line). 739 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 740 if ((Style.Language == FormatStyle::LK_Java && 741 CurrentToken->is(Keywords.kw_package)) || 742 (Info && Info->getPPKeywordID() == tok::pp_import && 743 CurrentToken->Next && 744 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier, 745 tok::kw_static))) { 746 next(); 747 parseIncludeDirective(); 748 return LT_ImportStatement; 749 } 750 751 // If this line starts and ends in '<' and '>', respectively, it is likely 752 // part of "#define <a/b.h>". 753 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 754 parseIncludeDirective(); 755 return LT_ImportStatement; 756 } 757 758 // In .proto files, top-level options are very similar to import statements 759 // and should not be line-wrapped. 760 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 && 761 CurrentToken->is(Keywords.kw_option)) { 762 next(); 763 if (CurrentToken && CurrentToken->is(tok::identifier)) 764 return LT_ImportStatement; 765 } 766 767 // import {...} from '...'; 768 if (Style.Language == FormatStyle::LK_JavaScript && 769 CurrentToken->is(Keywords.kw_import)) 770 return LT_ImportStatement; 771 772 bool KeywordVirtualFound = false; 773 bool ImportStatement = false; 774 while (CurrentToken) { 775 if (CurrentToken->is(tok::kw_virtual)) 776 KeywordVirtualFound = true; 777 if (Style.Language == FormatStyle::LK_JavaScript) { 778 // export {...} from '...'; 779 // An export followed by "from 'some string';" is a re-export from 780 // another module identified by a URI and is treated as a 781 // LT_ImportStatement (i.e. prevent wraps on it for long URIs). 782 // Just "export {...};" or "export class ..." should not be treated as 783 // an import in this sense. 784 if (Line.First->is(tok::kw_export) && 785 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next && 786 CurrentToken->Next->isStringLiteral()) 787 ImportStatement = true; 788 if (isClosureImportStatement(*CurrentToken)) 789 ImportStatement = true; 790 } 791 if (!consumeToken()) 792 return LT_Invalid; 793 } 794 if (KeywordVirtualFound) 795 return LT_VirtualFunctionDecl; 796 if (ImportStatement) 797 return LT_ImportStatement; 798 799 if (Line.startsWith(TT_ObjCMethodSpecifier)) { 800 if (Contexts.back().FirstObjCSelectorName) 801 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 802 Contexts.back().LongestObjCSelectorName; 803 return LT_ObjCMethodDecl; 804 } 805 806 return LT_Other; 807 } 808 809 private: 810 bool isClosureImportStatement(const FormatToken &Tok) { 811 // FIXME: Closure-library specific stuff should not be hard-coded but be 812 // configurable. 813 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) && 814 Tok.Next->Next && (Tok.Next->Next->TokenText == "module" || 815 Tok.Next->Next->TokenText == "provide" || 816 Tok.Next->Next->TokenText == "require" || 817 Tok.Next->Next->TokenText == "setTestOnly" || 818 Tok.Next->Next->TokenText == "forwardDeclare") && 819 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren); 820 } 821 822 void resetTokenMetadata(FormatToken *Token) { 823 if (!Token) 824 return; 825 826 // Reset token type in case we have already looked at it and then 827 // recovered from an error (e.g. failure to find the matching >). 828 if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro, 829 TT_FunctionLBrace, TT_ImplicitStringLiteral, 830 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow, 831 TT_RegexLiteral)) 832 CurrentToken->Type = TT_Unknown; 833 CurrentToken->Role.reset(); 834 CurrentToken->MatchingParen = nullptr; 835 CurrentToken->FakeLParens.clear(); 836 CurrentToken->FakeRParens = 0; 837 } 838 839 void next() { 840 if (CurrentToken) { 841 CurrentToken->NestingLevel = Contexts.size() - 1; 842 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 843 modifyContext(*CurrentToken); 844 determineTokenType(*CurrentToken); 845 CurrentToken = CurrentToken->Next; 846 } 847 848 resetTokenMetadata(CurrentToken); 849 } 850 851 /// \brief A struct to hold information valid in a specific context, e.g. 852 /// a pair of parenthesis. 853 struct Context { 854 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 855 bool IsExpression) 856 : ContextKind(ContextKind), BindingStrength(BindingStrength), 857 IsExpression(IsExpression) {} 858 859 tok::TokenKind ContextKind; 860 unsigned BindingStrength; 861 bool IsExpression; 862 unsigned LongestObjCSelectorName = 0; 863 bool ColonIsForRangeExpr = false; 864 bool ColonIsDictLiteral = false; 865 bool ColonIsObjCMethodExpr = false; 866 FormatToken *FirstObjCSelectorName = nullptr; 867 FormatToken *FirstStartOfName = nullptr; 868 bool CanBeExpression = true; 869 bool InTemplateArgument = false; 870 bool InCtorInitializer = false; 871 bool CaretFound = false; 872 bool IsForEachMacro = false; 873 }; 874 875 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime 876 /// of each instance. 877 struct ScopedContextCreator { 878 AnnotatingParser &P; 879 880 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 881 unsigned Increase) 882 : P(P) { 883 P.Contexts.push_back(Context(ContextKind, 884 P.Contexts.back().BindingStrength + Increase, 885 P.Contexts.back().IsExpression)); 886 } 887 888 ~ScopedContextCreator() { P.Contexts.pop_back(); } 889 }; 890 891 void modifyContext(const FormatToken &Current) { 892 if (Current.getPrecedence() == prec::Assignment && 893 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) && 894 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 895 Contexts.back().IsExpression = true; 896 if (!Line.startsWith(TT_UnaryOperator)) { 897 for (FormatToken *Previous = Current.Previous; 898 Previous && Previous->Previous && 899 !Previous->Previous->isOneOf(tok::comma, tok::semi); 900 Previous = Previous->Previous) { 901 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 902 Previous = Previous->MatchingParen; 903 if (!Previous) 904 break; 905 } 906 if (Previous->opensScope()) 907 break; 908 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) && 909 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) && 910 Previous->Previous && Previous->Previous->isNot(tok::equal)) 911 Previous->Type = TT_PointerOrReference; 912 } 913 } 914 } else if (Current.is(tok::lessless) && 915 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) { 916 Contexts.back().IsExpression = true; 917 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 918 Contexts.back().IsExpression = true; 919 } else if (Current.is(TT_TrailingReturnArrow)) { 920 Contexts.back().IsExpression = false; 921 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) { 922 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java; 923 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 924 for (FormatToken *Previous = Current.Previous; 925 Previous && Previous->isOneOf(tok::star, tok::amp); 926 Previous = Previous->Previous) 927 Previous->Type = TT_PointerOrReference; 928 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer) 929 Contexts.back().IsExpression = false; 930 } else if (Current.Previous && 931 Current.Previous->is(TT_CtorInitializerColon)) { 932 Contexts.back().IsExpression = true; 933 Contexts.back().InCtorInitializer = true; 934 } else if (Current.is(tok::kw_new)) { 935 Contexts.back().CanBeExpression = false; 936 } else if (Current.isOneOf(tok::semi, tok::exclaim)) { 937 // This should be the condition or increment in a for-loop. 938 Contexts.back().IsExpression = true; 939 } 940 } 941 942 void determineTokenType(FormatToken &Current) { 943 if (!Current.is(TT_Unknown)) 944 // The token type is already known. 945 return; 946 947 // Line.MightBeFunctionDecl can only be true after the parentheses of a 948 // function declaration have been found. In this case, 'Current' is a 949 // trailing token of this declaration and thus cannot be a name. 950 if (Current.is(Keywords.kw_instanceof)) { 951 Current.Type = TT_BinaryOperator; 952 } else if (isStartOfName(Current) && 953 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 954 Contexts.back().FirstStartOfName = &Current; 955 Current.Type = TT_StartOfName; 956 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) { 957 AutoFound = true; 958 } else if (Current.is(tok::arrow) && 959 Style.Language == FormatStyle::LK_Java) { 960 Current.Type = TT_LambdaArrow; 961 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration && 962 Current.NestingLevel == 0) { 963 Current.Type = TT_TrailingReturnArrow; 964 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 965 Current.Type = 966 determineStarAmpUsage(Current, Contexts.back().CanBeExpression && 967 Contexts.back().IsExpression, 968 Contexts.back().InTemplateArgument); 969 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 970 Current.Type = determinePlusMinusCaretUsage(Current); 971 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret)) 972 Contexts.back().CaretFound = true; 973 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 974 Current.Type = determineIncrementUsage(Current); 975 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 976 Current.Type = TT_UnaryOperator; 977 } else if (Current.is(tok::question)) { 978 if (Style.Language == FormatStyle::LK_JavaScript && 979 Line.MustBeDeclaration) { 980 // In JavaScript, `interface X { foo?(): bar; }` is an optional method 981 // on the interface, not a ternary expression. 982 Current.Type = TT_JsTypeOptionalQuestion; 983 } else { 984 Current.Type = TT_ConditionalExpr; 985 } 986 } else if (Current.isBinaryOperator() && 987 (!Current.Previous || Current.Previous->isNot(tok::l_square))) { 988 Current.Type = TT_BinaryOperator; 989 } else if (Current.is(tok::comment)) { 990 if (Current.TokenText.startswith("/*")) { 991 if (Current.TokenText.endswith("*/")) 992 Current.Type = TT_BlockComment; 993 else 994 // The lexer has for some reason determined a comment here. But we 995 // cannot really handle it, if it isn't properly terminated. 996 Current.Tok.setKind(tok::unknown); 997 } else { 998 Current.Type = TT_LineComment; 999 } 1000 } else if (Current.is(tok::r_paren)) { 1001 if (rParenEndsCast(Current)) 1002 Current.Type = TT_CastRParen; 1003 if (Current.MatchingParen && Current.Next && 1004 !Current.Next->isBinaryOperator() && 1005 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace)) 1006 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) 1007 if (BeforeParen->is(tok::identifier) && 1008 BeforeParen->TokenText == BeforeParen->TokenText.upper() && 1009 (!BeforeParen->Previous || 1010 BeforeParen->Previous->ClosesTemplateDeclaration)) 1011 Current.Type = TT_FunctionAnnotationRParen; 1012 } else if (Current.is(tok::at) && Current.Next) { 1013 if (Current.Next->isStringLiteral()) { 1014 Current.Type = TT_ObjCStringLiteral; 1015 } else { 1016 switch (Current.Next->Tok.getObjCKeywordID()) { 1017 case tok::objc_interface: 1018 case tok::objc_implementation: 1019 case tok::objc_protocol: 1020 Current.Type = TT_ObjCDecl; 1021 break; 1022 case tok::objc_property: 1023 Current.Type = TT_ObjCProperty; 1024 break; 1025 default: 1026 break; 1027 } 1028 } 1029 } else if (Current.is(tok::period)) { 1030 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 1031 if (PreviousNoComment && 1032 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 1033 Current.Type = TT_DesignatedInitializerPeriod; 1034 else if (Style.Language == FormatStyle::LK_Java && Current.Previous && 1035 Current.Previous->isOneOf(TT_JavaAnnotation, 1036 TT_LeadingJavaAnnotation)) { 1037 Current.Type = Current.Previous->Type; 1038 } 1039 } else if (Current.isOneOf(tok::identifier, tok::kw_const) && 1040 Current.Previous && 1041 !Current.Previous->isOneOf(tok::equal, tok::at) && 1042 Line.MightBeFunctionDecl && Contexts.size() == 1) { 1043 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1044 // function declaration have been found. 1045 Current.Type = TT_TrailingAnnotation; 1046 } else if ((Style.Language == FormatStyle::LK_Java || 1047 Style.Language == FormatStyle::LK_JavaScript) && 1048 Current.Previous) { 1049 if (Current.Previous->is(tok::at) && 1050 Current.isNot(Keywords.kw_interface)) { 1051 const FormatToken &AtToken = *Current.Previous; 1052 const FormatToken *Previous = AtToken.getPreviousNonComment(); 1053 if (!Previous || Previous->is(TT_LeadingJavaAnnotation)) 1054 Current.Type = TT_LeadingJavaAnnotation; 1055 else 1056 Current.Type = TT_JavaAnnotation; 1057 } else if (Current.Previous->is(tok::period) && 1058 Current.Previous->isOneOf(TT_JavaAnnotation, 1059 TT_LeadingJavaAnnotation)) { 1060 Current.Type = Current.Previous->Type; 1061 } 1062 } 1063 } 1064 1065 /// \brief Take a guess at whether \p Tok starts a name of a function or 1066 /// variable declaration. 1067 /// 1068 /// This is a heuristic based on whether \p Tok is an identifier following 1069 /// something that is likely a type. 1070 bool isStartOfName(const FormatToken &Tok) { 1071 if (Tok.isNot(tok::identifier) || !Tok.Previous) 1072 return false; 1073 1074 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof)) 1075 return false; 1076 if (Style.Language == FormatStyle::LK_JavaScript && 1077 Tok.Previous->is(Keywords.kw_in)) 1078 return false; 1079 1080 // Skip "const" as it does not have an influence on whether this is a name. 1081 FormatToken *PreviousNotConst = Tok.Previous; 1082 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 1083 PreviousNotConst = PreviousNotConst->Previous; 1084 1085 if (!PreviousNotConst) 1086 return false; 1087 1088 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 1089 PreviousNotConst->Previous && 1090 PreviousNotConst->Previous->is(tok::hash); 1091 1092 if (PreviousNotConst->is(TT_TemplateCloser)) 1093 return PreviousNotConst && PreviousNotConst->MatchingParen && 1094 PreviousNotConst->MatchingParen->Previous && 1095 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) && 1096 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 1097 1098 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen && 1099 PreviousNotConst->MatchingParen->Previous && 1100 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype)) 1101 return true; 1102 1103 return (!IsPPKeyword && 1104 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) || 1105 PreviousNotConst->is(TT_PointerOrReference) || 1106 PreviousNotConst->isSimpleTypeSpecifier(); 1107 } 1108 1109 /// \brief Determine whether ')' is ending a cast. 1110 bool rParenEndsCast(const FormatToken &Tok) { 1111 // C-style casts are only used in C++ and Java. 1112 if (Style.Language != FormatStyle::LK_Cpp && 1113 Style.Language != FormatStyle::LK_Java) 1114 return false; 1115 1116 // Empty parens aren't casts and there are no casts at the end of the line. 1117 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen) 1118 return false; 1119 1120 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 1121 if (LeftOfParens) { 1122 // If there is an opening parenthesis left of the current parentheses, 1123 // look past it as these might be chained casts. 1124 if (LeftOfParens->is(tok::r_paren)) { 1125 if (!LeftOfParens->MatchingParen || 1126 !LeftOfParens->MatchingParen->Previous) 1127 return false; 1128 LeftOfParens = LeftOfParens->MatchingParen->Previous; 1129 } 1130 1131 // If there is an identifier (or with a few exceptions a keyword) right 1132 // before the parentheses, this is unlikely to be a cast. 1133 if (LeftOfParens->Tok.getIdentifierInfo() && 1134 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case, 1135 tok::kw_delete)) 1136 return false; 1137 1138 // Certain other tokens right before the parentheses are also signals that 1139 // this cannot be a cast. 1140 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator, 1141 TT_TemplateCloser)) 1142 return false; 1143 } 1144 1145 if (Tok.Next->is(tok::question)) 1146 return false; 1147 1148 // As Java has no function types, a "(" after the ")" likely means that this 1149 // is a cast. 1150 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren)) 1151 return true; 1152 1153 // If a (non-string) literal follows, this is likely a cast. 1154 if (Tok.Next->isNot(tok::string_literal) && 1155 (Tok.Next->Tok.isLiteral() || 1156 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 1157 return true; 1158 1159 // Heuristically try to determine whether the parentheses contain a type. 1160 bool ParensAreType = 1161 !Tok.Previous || 1162 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) || 1163 Tok.Previous->isSimpleTypeSpecifier(); 1164 bool ParensCouldEndDecl = 1165 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater); 1166 if (ParensAreType && !ParensCouldEndDecl) 1167 return true; 1168 1169 // At this point, we heuristically assume that there are no casts at the 1170 // start of the line. We assume that we have found most cases where there 1171 // are by the logic above, e.g. "(void)x;". 1172 if (!LeftOfParens) 1173 return false; 1174 1175 // If the following token is an identifier, this is a cast. All cases where 1176 // this can be something else are handled above. 1177 if (Tok.Next->is(tok::identifier)) 1178 return true; 1179 1180 if (!Tok.Next->Next) 1181 return false; 1182 1183 // If the next token after the parenthesis is a unary operator, assume 1184 // that this is cast, unless there are unexpected tokens inside the 1185 // parenthesis. 1186 bool NextIsUnary = 1187 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star); 1188 if (!NextIsUnary || Tok.Next->is(tok::plus) || 1189 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) 1190 return false; 1191 // Search for unexpected tokens. 1192 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen; 1193 Prev = Prev->Previous) { 1194 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon)) 1195 return false; 1196 } 1197 return true; 1198 } 1199 1200 /// \brief Return the type of the given token assuming it is * or &. 1201 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 1202 bool InTemplateArgument) { 1203 if (Style.Language == FormatStyle::LK_JavaScript) 1204 return TT_BinaryOperator; 1205 1206 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1207 if (!PrevToken) 1208 return TT_UnaryOperator; 1209 1210 const FormatToken *NextToken = Tok.getNextNonComment(); 1211 if (!NextToken || 1212 NextToken->isOneOf(tok::arrow, Keywords.kw_final, 1213 Keywords.kw_override) || 1214 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) 1215 return TT_PointerOrReference; 1216 1217 if (PrevToken->is(tok::coloncolon)) 1218 return TT_PointerOrReference; 1219 1220 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 1221 tok::comma, tok::semi, tok::kw_return, tok::colon, 1222 tok::equal, tok::kw_delete, tok::kw_sizeof) || 1223 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr, 1224 TT_UnaryOperator, TT_CastRParen)) 1225 return TT_UnaryOperator; 1226 1227 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare)) 1228 return TT_PointerOrReference; 1229 if (NextToken->is(tok::kw_operator) && !IsExpression) 1230 return TT_PointerOrReference; 1231 if (NextToken->isOneOf(tok::comma, tok::semi)) 1232 return TT_PointerOrReference; 1233 1234 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen && 1235 PrevToken->MatchingParen->Previous && 1236 PrevToken->MatchingParen->Previous->isOneOf(tok::kw_typeof, 1237 tok::kw_decltype)) 1238 return TT_PointerOrReference; 1239 1240 if (PrevToken->Tok.isLiteral() || 1241 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 1242 tok::kw_false, tok::r_brace) || 1243 NextToken->Tok.isLiteral() || 1244 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 1245 NextToken->isUnaryOperator() || 1246 // If we know we're in a template argument, there are no named 1247 // declarations. Thus, having an identifier on the right-hand side 1248 // indicates a binary operator. 1249 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 1250 return TT_BinaryOperator; 1251 1252 // "&&(" is quite unlikely to be two successive unary "&". 1253 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren)) 1254 return TT_BinaryOperator; 1255 1256 // This catches some cases where evaluation order is used as control flow: 1257 // aaa && aaa->f(); 1258 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 1259 if (NextNextToken && NextNextToken->is(tok::arrow)) 1260 return TT_BinaryOperator; 1261 1262 // It is very unlikely that we are going to find a pointer or reference type 1263 // definition on the RHS of an assignment. 1264 if (IsExpression && !Contexts.back().CaretFound) 1265 return TT_BinaryOperator; 1266 1267 return TT_PointerOrReference; 1268 } 1269 1270 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 1271 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1272 if (!PrevToken || PrevToken->is(TT_CastRParen)) 1273 return TT_UnaryOperator; 1274 1275 // Use heuristics to recognize unary operators. 1276 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 1277 tok::question, tok::colon, tok::kw_return, 1278 tok::kw_case, tok::at, tok::l_brace)) 1279 return TT_UnaryOperator; 1280 1281 // There can't be two consecutive binary operators. 1282 if (PrevToken->is(TT_BinaryOperator)) 1283 return TT_UnaryOperator; 1284 1285 // Fall back to marking the token as binary operator. 1286 return TT_BinaryOperator; 1287 } 1288 1289 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements. 1290 TokenType determineIncrementUsage(const FormatToken &Tok) { 1291 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1292 if (!PrevToken || PrevToken->is(TT_CastRParen)) 1293 return TT_UnaryOperator; 1294 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 1295 return TT_TrailingUnaryOperator; 1296 1297 return TT_UnaryOperator; 1298 } 1299 1300 SmallVector<Context, 8> Contexts; 1301 1302 const FormatStyle &Style; 1303 AnnotatedLine &Line; 1304 FormatToken *CurrentToken; 1305 bool AutoFound; 1306 const AdditionalKeywords &Keywords; 1307 1308 // Set of "<" tokens that do not open a template parameter list. If parseAngle 1309 // determines that a specific token can't be a template opener, it will make 1310 // same decision irrespective of the decisions for tokens leading up to it. 1311 // Store this information to prevent this from causing exponential runtime. 1312 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess; 1313 }; 1314 1315 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1; 1316 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 1317 1318 /// \brief Parses binary expressions by inserting fake parenthesis based on 1319 /// operator precedence. 1320 class ExpressionParser { 1321 public: 1322 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords, 1323 AnnotatedLine &Line) 1324 : Style(Style), Keywords(Keywords), Current(Line.First) {} 1325 1326 /// \brief Parse expressions with the given operatore precedence. 1327 void parse(int Precedence = 0) { 1328 // Skip 'return' and ObjC selector colons as they are not part of a binary 1329 // expression. 1330 while (Current && (Current->is(tok::kw_return) || 1331 (Current->is(tok::colon) && 1332 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) 1333 next(); 1334 1335 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1336 return; 1337 1338 // Conditional expressions need to be parsed separately for proper nesting. 1339 if (Precedence == prec::Conditional) { 1340 parseConditionalExpr(); 1341 return; 1342 } 1343 1344 // Parse unary operators, which all have a higher precedence than binary 1345 // operators. 1346 if (Precedence == PrecedenceUnaryOperator) { 1347 parseUnaryOperator(); 1348 return; 1349 } 1350 1351 FormatToken *Start = Current; 1352 FormatToken *LatestOperator = nullptr; 1353 unsigned OperatorIndex = 0; 1354 1355 while (Current) { 1356 // Consume operators with higher precedence. 1357 parse(Precedence + 1); 1358 1359 int CurrentPrecedence = getCurrentPrecedence(); 1360 1361 if (Current && Current->is(TT_SelectorName) && 1362 Precedence == CurrentPrecedence) { 1363 if (LatestOperator) 1364 addFakeParenthesis(Start, prec::Level(Precedence)); 1365 Start = Current; 1366 } 1367 1368 // At the end of the line or when an operator with higher precedence is 1369 // found, insert fake parenthesis and return. 1370 if (!Current || (Current->closesScope() && Current->MatchingParen) || 1371 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) || 1372 (CurrentPrecedence == prec::Conditional && 1373 Precedence == prec::Assignment && Current->is(tok::colon))) { 1374 break; 1375 } 1376 1377 // Consume scopes: (), [], <> and {} 1378 if (Current->opensScope()) { 1379 while (Current && !Current->closesScope()) { 1380 next(); 1381 parse(); 1382 } 1383 next(); 1384 } else { 1385 // Operator found. 1386 if (CurrentPrecedence == Precedence) { 1387 if (LatestOperator) 1388 LatestOperator->NextOperator = Current; 1389 LatestOperator = Current; 1390 Current->OperatorIndex = OperatorIndex; 1391 ++OperatorIndex; 1392 } 1393 next(/*SkipPastLeadingComments=*/Precedence > 0); 1394 } 1395 } 1396 1397 if (LatestOperator && (Current || Precedence > 0)) { 1398 // LatestOperator->LastOperator = true; 1399 if (Precedence == PrecedenceArrowAndPeriod) { 1400 // Call expressions don't have a binary operator precedence. 1401 addFakeParenthesis(Start, prec::Unknown); 1402 } else { 1403 addFakeParenthesis(Start, prec::Level(Precedence)); 1404 } 1405 } 1406 } 1407 1408 private: 1409 /// \brief Gets the precedence (+1) of the given token for binary operators 1410 /// and other tokens that we treat like binary operators. 1411 int getCurrentPrecedence() { 1412 if (Current) { 1413 const FormatToken *NextNonComment = Current->getNextNonComment(); 1414 if (Current->is(TT_ConditionalExpr)) 1415 return prec::Conditional; 1416 if (NextNonComment && NextNonComment->is(tok::colon) && 1417 NextNonComment->is(TT_DictLiteral)) 1418 return prec::Comma; 1419 if (Current->is(TT_LambdaArrow)) 1420 return prec::Comma; 1421 if (Current->is(TT_JsFatArrow)) 1422 return prec::Assignment; 1423 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName, 1424 TT_JsComputedPropertyName) || 1425 (Current->is(tok::comment) && NextNonComment && 1426 NextNonComment->is(TT_SelectorName))) 1427 return 0; 1428 if (Current->is(TT_RangeBasedForLoopColon)) 1429 return prec::Comma; 1430 if ((Style.Language == FormatStyle::LK_Java || 1431 Style.Language == FormatStyle::LK_JavaScript) && 1432 Current->is(Keywords.kw_instanceof)) 1433 return prec::Relational; 1434 if (Style.Language == FormatStyle::LK_JavaScript && 1435 Current->is(Keywords.kw_in)) 1436 return prec::Relational; 1437 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma)) 1438 return Current->getPrecedence(); 1439 if (Current->isOneOf(tok::period, tok::arrow)) 1440 return PrecedenceArrowAndPeriod; 1441 if (Style.Language == FormatStyle::LK_Java && 1442 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements, 1443 Keywords.kw_throws)) 1444 return 0; 1445 } 1446 return -1; 1447 } 1448 1449 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1450 Start->FakeLParens.push_back(Precedence); 1451 if (Precedence > prec::Unknown) 1452 Start->StartsBinaryExpression = true; 1453 if (Current) { 1454 FormatToken *Previous = Current->Previous; 1455 while (Previous->is(tok::comment) && Previous->Previous) 1456 Previous = Previous->Previous; 1457 ++Previous->FakeRParens; 1458 if (Precedence > prec::Unknown) 1459 Previous->EndsBinaryExpression = true; 1460 } 1461 } 1462 1463 /// \brief Parse unary operator expressions and surround them with fake 1464 /// parentheses if appropriate. 1465 void parseUnaryOperator() { 1466 if (!Current || Current->isNot(TT_UnaryOperator)) { 1467 parse(PrecedenceArrowAndPeriod); 1468 return; 1469 } 1470 1471 FormatToken *Start = Current; 1472 next(); 1473 parseUnaryOperator(); 1474 1475 // The actual precedence doesn't matter. 1476 addFakeParenthesis(Start, prec::Unknown); 1477 } 1478 1479 void parseConditionalExpr() { 1480 while (Current && Current->isTrailingComment()) { 1481 next(); 1482 } 1483 FormatToken *Start = Current; 1484 parse(prec::LogicalOr); 1485 if (!Current || !Current->is(tok::question)) 1486 return; 1487 next(); 1488 parse(prec::Assignment); 1489 if (!Current || Current->isNot(TT_ConditionalExpr)) 1490 return; 1491 next(); 1492 parse(prec::Assignment); 1493 addFakeParenthesis(Start, prec::Conditional); 1494 } 1495 1496 void next(bool SkipPastLeadingComments = true) { 1497 if (Current) 1498 Current = Current->Next; 1499 while (Current && 1500 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 1501 Current->isTrailingComment()) 1502 Current = Current->Next; 1503 } 1504 1505 const FormatStyle &Style; 1506 const AdditionalKeywords &Keywords; 1507 FormatToken *Current; 1508 }; 1509 1510 } // end anonymous namespace 1511 1512 void TokenAnnotator::setCommentLineLevels( 1513 SmallVectorImpl<AnnotatedLine *> &Lines) { 1514 const AnnotatedLine *NextNonCommentLine = nullptr; 1515 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 1516 E = Lines.rend(); 1517 I != E; ++I) { 1518 if (NextNonCommentLine && (*I)->First->is(tok::comment) && 1519 (*I)->First->Next == nullptr) 1520 (*I)->Level = NextNonCommentLine->Level; 1521 else 1522 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 1523 1524 setCommentLineLevels((*I)->Children); 1525 } 1526 } 1527 1528 void TokenAnnotator::annotate(AnnotatedLine &Line) { 1529 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1530 E = Line.Children.end(); 1531 I != E; ++I) { 1532 annotate(**I); 1533 } 1534 AnnotatingParser Parser(Style, Line, Keywords); 1535 Line.Type = Parser.parseLine(); 1536 if (Line.Type == LT_Invalid) 1537 return; 1538 1539 ExpressionParser ExprParser(Style, Keywords, Line); 1540 ExprParser.parse(); 1541 1542 if (Line.startsWith(TT_ObjCMethodSpecifier)) 1543 Line.Type = LT_ObjCMethodDecl; 1544 else if (Line.startsWith(TT_ObjCDecl)) 1545 Line.Type = LT_ObjCDecl; 1546 else if (Line.startsWith(TT_ObjCProperty)) 1547 Line.Type = LT_ObjCProperty; 1548 1549 Line.First->SpacesRequiredBefore = 1; 1550 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 1551 } 1552 1553 // This function heuristically determines whether 'Current' starts the name of a 1554 // function declaration. 1555 static bool isFunctionDeclarationName(const FormatToken &Current) { 1556 auto skipOperatorName = [](const FormatToken* Next) -> const FormatToken* { 1557 for (; Next; Next = Next->Next) { 1558 if (Next->is(TT_OverloadedOperatorLParen)) 1559 return Next; 1560 if (Next->is(TT_OverloadedOperator)) 1561 continue; 1562 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) { 1563 // For 'new[]' and 'delete[]'. 1564 if (Next->Next && Next->Next->is(tok::l_square) && 1565 Next->Next->Next && Next->Next->Next->is(tok::r_square)) 1566 Next = Next->Next->Next; 1567 continue; 1568 } 1569 1570 break; 1571 } 1572 return nullptr; 1573 }; 1574 1575 const FormatToken *Next = Current.Next; 1576 if (Current.is(tok::kw_operator)) { 1577 if (Current.Previous && Current.Previous->is(tok::coloncolon)) 1578 return false; 1579 Next = skipOperatorName(Next); 1580 } else { 1581 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0) 1582 return false; 1583 for (; Next; Next = Next->Next) { 1584 if (Next->is(TT_TemplateOpener)) { 1585 Next = Next->MatchingParen; 1586 } else if (Next->is(tok::coloncolon)) { 1587 Next = Next->Next; 1588 if (!Next) 1589 return false; 1590 if (Next->is(tok::kw_operator)) { 1591 Next = skipOperatorName(Next->Next); 1592 break; 1593 } 1594 if (!Next->is(tok::identifier)) 1595 return false; 1596 } else if (Next->is(tok::l_paren)) { 1597 break; 1598 } else { 1599 return false; 1600 } 1601 } 1602 } 1603 1604 if (!Next || !Next->is(tok::l_paren)) 1605 return false; 1606 if (Next->Next == Next->MatchingParen) 1607 return true; 1608 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen; 1609 Tok = Tok->Next) { 1610 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 1611 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName)) 1612 return true; 1613 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) || 1614 Tok->Tok.isLiteral()) 1615 return false; 1616 } 1617 return false; 1618 } 1619 1620 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const { 1621 assert(Line.MightBeFunctionDecl); 1622 1623 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel || 1624 Style.AlwaysBreakAfterReturnType == 1625 FormatStyle::RTBS_TopLevelDefinitions) && 1626 Line.Level > 0) 1627 return false; 1628 1629 switch (Style.AlwaysBreakAfterReturnType) { 1630 case FormatStyle::RTBS_None: 1631 return false; 1632 case FormatStyle::RTBS_All: 1633 case FormatStyle::RTBS_TopLevel: 1634 return true; 1635 case FormatStyle::RTBS_AllDefinitions: 1636 case FormatStyle::RTBS_TopLevelDefinitions: 1637 return Line.mightBeFunctionDefinition(); 1638 } 1639 1640 return false; 1641 } 1642 1643 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 1644 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1645 E = Line.Children.end(); 1646 I != E; ++I) { 1647 calculateFormattingInformation(**I); 1648 } 1649 1650 Line.First->TotalLength = 1651 Line.First->IsMultiline ? Style.ColumnLimit : Line.First->ColumnWidth; 1652 if (!Line.First->Next) 1653 return; 1654 FormatToken *Current = Line.First->Next; 1655 bool InFunctionDecl = Line.MightBeFunctionDecl; 1656 while (Current) { 1657 if (isFunctionDeclarationName(*Current)) 1658 Current->Type = TT_FunctionDeclarationName; 1659 if (Current->is(TT_LineComment)) { 1660 if (Current->Previous->BlockKind == BK_BracedInit && 1661 Current->Previous->opensScope()) 1662 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1; 1663 else 1664 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 1665 1666 // If we find a trailing comment, iterate backwards to determine whether 1667 // it seems to relate to a specific parameter. If so, break before that 1668 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 1669 // to the previous line in: 1670 // SomeFunction(a, 1671 // b, // comment 1672 // c); 1673 if (!Current->HasUnescapedNewline) { 1674 for (FormatToken *Parameter = Current->Previous; Parameter; 1675 Parameter = Parameter->Previous) { 1676 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 1677 break; 1678 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 1679 if (!Parameter->Previous->is(TT_CtorInitializerComma) && 1680 Parameter->HasUnescapedNewline) 1681 Parameter->MustBreakBefore = true; 1682 break; 1683 } 1684 } 1685 } 1686 } else if (Current->SpacesRequiredBefore == 0 && 1687 spaceRequiredBefore(Line, *Current)) { 1688 Current->SpacesRequiredBefore = 1; 1689 } 1690 1691 Current->MustBreakBefore = 1692 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 1693 1694 if (!Current->MustBreakBefore && InFunctionDecl && 1695 Current->is(TT_FunctionDeclarationName)) 1696 Current->MustBreakBefore = mustBreakForReturnType(Line); 1697 1698 Current->CanBreakBefore = 1699 Current->MustBreakBefore || canBreakBefore(Line, *Current); 1700 unsigned ChildSize = 0; 1701 if (Current->Previous->Children.size() == 1) { 1702 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 1703 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 1704 : LastOfChild.TotalLength + 1; 1705 } 1706 const FormatToken *Prev = Current->Previous; 1707 if (Current->MustBreakBefore || Prev->Children.size() > 1 || 1708 (Prev->Children.size() == 1 && 1709 Prev->Children[0]->First->MustBreakBefore) || 1710 Current->IsMultiline) 1711 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit; 1712 else 1713 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth + 1714 ChildSize + Current->SpacesRequiredBefore; 1715 1716 if (Current->is(TT_CtorInitializerColon)) 1717 InFunctionDecl = false; 1718 1719 // FIXME: Only calculate this if CanBreakBefore is true once static 1720 // initializers etc. are sorted out. 1721 // FIXME: Move magic numbers to a better place. 1722 Current->SplitPenalty = 20 * Current->BindingStrength + 1723 splitPenalty(Line, *Current, InFunctionDecl); 1724 1725 Current = Current->Next; 1726 } 1727 1728 calculateUnbreakableTailLengths(Line); 1729 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 1730 if (Current->Role) 1731 Current->Role->precomputeFormattingInfos(Current); 1732 } 1733 1734 DEBUG({ printDebugInfo(Line); }); 1735 } 1736 1737 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 1738 unsigned UnbreakableTailLength = 0; 1739 FormatToken *Current = Line.Last; 1740 while (Current) { 1741 Current->UnbreakableTailLength = UnbreakableTailLength; 1742 if (Current->CanBreakBefore || 1743 Current->isOneOf(tok::comment, tok::string_literal)) { 1744 UnbreakableTailLength = 0; 1745 } else { 1746 UnbreakableTailLength += 1747 Current->ColumnWidth + Current->SpacesRequiredBefore; 1748 } 1749 Current = Current->Previous; 1750 } 1751 } 1752 1753 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 1754 const FormatToken &Tok, 1755 bool InFunctionDecl) { 1756 const FormatToken &Left = *Tok.Previous; 1757 const FormatToken &Right = Tok; 1758 1759 if (Left.is(tok::semi)) 1760 return 0; 1761 1762 if (Style.Language == FormatStyle::LK_Java) { 1763 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws)) 1764 return 1; 1765 if (Right.is(Keywords.kw_implements)) 1766 return 2; 1767 if (Left.is(tok::comma) && Left.NestingLevel == 0) 1768 return 3; 1769 } else if (Style.Language == FormatStyle::LK_JavaScript) { 1770 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma)) 1771 return 100; 1772 if (Left.is(TT_JsTypeColon)) 1773 return 35; 1774 } 1775 1776 if (Left.is(tok::comma) || (Right.is(tok::identifier) && Right.Next && 1777 Right.Next->is(TT_DictLiteral))) 1778 return 1; 1779 if (Right.is(tok::l_square)) { 1780 if (Style.Language == FormatStyle::LK_Proto) 1781 return 1; 1782 if (Left.is(tok::r_square)) 1783 return 200; 1784 // Slightly prefer formatting local lambda definitions like functions. 1785 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal)) 1786 return 35; 1787 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 1788 TT_ArrayInitializerLSquare)) 1789 return 500; 1790 } 1791 1792 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 1793 Right.is(tok::kw_operator)) { 1794 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 1795 return 3; 1796 if (Left.is(TT_StartOfName)) 1797 return 110; 1798 if (InFunctionDecl && Right.NestingLevel == 0) 1799 return Style.PenaltyReturnTypeOnItsOwnLine; 1800 return 200; 1801 } 1802 if (Right.is(TT_PointerOrReference)) 1803 return 190; 1804 if (Right.is(TT_LambdaArrow)) 1805 return 110; 1806 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 1807 return 150; 1808 if (Left.is(TT_CastRParen)) 1809 return 100; 1810 if (Left.is(tok::coloncolon) || 1811 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 1812 return 500; 1813 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 1814 return 5000; 1815 1816 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon)) 1817 return 2; 1818 1819 if (Right.isMemberAccess()) { 1820 // Breaking before the "./->" of a chained call/member access is reasonably 1821 // cheap, as formatting those with one call per line is generally 1822 // desirable. In particular, it should be cheaper to break before the call 1823 // than it is to break inside a call's parameters, which could lead to weird 1824 // "hanging" indents. The exception is the very last "./->" to support this 1825 // frequent pattern: 1826 // 1827 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc( 1828 // dddddddd); 1829 // 1830 // which might otherwise be blown up onto many lines. Here, clang-format 1831 // won't produce "hanging" indents anyway as there is no other trailing 1832 // call. 1833 // 1834 // Also apply higher penalty is not a call as that might lead to a wrapping 1835 // like: 1836 // 1837 // aaaaaaa 1838 // .aaaaaaaaa.bbbbbbbb(cccccccc); 1839 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope() 1840 ? 150 1841 : 35; 1842 } 1843 1844 if (Right.is(TT_TrailingAnnotation) && 1845 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 1846 // Moving trailing annotations to the next line is fine for ObjC method 1847 // declarations. 1848 if (Line.startsWith(TT_ObjCMethodSpecifier)) 1849 return 10; 1850 // Generally, breaking before a trailing annotation is bad unless it is 1851 // function-like. It seems to be especially preferable to keep standard 1852 // annotations (i.e. "const", "final" and "override") on the same line. 1853 // Use a slightly higher penalty after ")" so that annotations like 1854 // "const override" are kept together. 1855 bool is_short_annotation = Right.TokenText.size() < 10; 1856 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 1857 } 1858 1859 // In for-loops, prefer breaking at ',' and ';'. 1860 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal)) 1861 return 4; 1862 1863 // In Objective-C method expressions, prefer breaking before "param:" over 1864 // breaking after it. 1865 if (Right.is(TT_SelectorName)) 1866 return 0; 1867 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr)) 1868 return Line.MightBeFunctionDecl ? 50 : 500; 1869 1870 if (Left.is(tok::l_paren) && InFunctionDecl && 1871 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) 1872 return 100; 1873 if (Left.is(tok::l_paren) && Left.Previous && 1874 Left.Previous->isOneOf(tok::kw_if, tok::kw_for)) 1875 return 1000; 1876 if (Left.is(tok::equal) && InFunctionDecl) 1877 return 110; 1878 if (Right.is(tok::r_brace)) 1879 return 1; 1880 if (Left.is(TT_TemplateOpener)) 1881 return 100; 1882 if (Left.opensScope()) { 1883 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign) 1884 return 0; 1885 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 1886 : 19; 1887 } 1888 if (Left.is(TT_JavaAnnotation)) 1889 return 50; 1890 1891 if (Right.is(tok::lessless)) { 1892 if (Left.is(tok::string_literal) && 1893 (Right.NextOperator || Right.OperatorIndex != 1)) { 1894 StringRef Content = Left.TokenText; 1895 if (Content.startswith("\"")) 1896 Content = Content.drop_front(1); 1897 if (Content.endswith("\"")) 1898 Content = Content.drop_back(1); 1899 Content = Content.trim(); 1900 if (Content.size() > 1 && 1901 (Content.back() == ':' || Content.back() == '=')) 1902 return 25; 1903 } 1904 return 1; // Breaking at a << is really cheap. 1905 } 1906 if (Left.is(TT_ConditionalExpr)) 1907 return prec::Conditional; 1908 prec::Level Level = Left.getPrecedence(); 1909 if (Level != prec::Unknown) 1910 return Level; 1911 Level = Right.getPrecedence(); 1912 if (Level != prec::Unknown) 1913 return Level; 1914 1915 return 3; 1916 } 1917 1918 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 1919 const FormatToken &Left, 1920 const FormatToken &Right) { 1921 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 1922 return true; 1923 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 1924 Left.Tok.getObjCKeywordID() == tok::objc_property) 1925 return true; 1926 if (Right.is(tok::hashhash)) 1927 return Left.is(tok::hash); 1928 if (Left.isOneOf(tok::hashhash, tok::hash)) 1929 return Right.is(tok::hash); 1930 if (Left.is(tok::l_paren) && Right.is(tok::r_paren)) 1931 return Style.SpaceInEmptyParentheses; 1932 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 1933 return (Right.is(TT_CastRParen) || 1934 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen))) 1935 ? Style.SpacesInCStyleCastParentheses 1936 : Style.SpacesInParentheses; 1937 if (Right.isOneOf(tok::semi, tok::comma)) 1938 return false; 1939 if (Right.is(tok::less) && 1940 (Left.is(tok::kw_template) || 1941 (Line.Type == LT_ObjCDecl && Style.ObjCSpaceBeforeProtocolList))) 1942 return true; 1943 if (Left.isOneOf(tok::exclaim, tok::tilde)) 1944 return false; 1945 if (Left.is(tok::at) && 1946 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 1947 tok::numeric_constant, tok::l_paren, tok::l_brace, 1948 tok::kw_true, tok::kw_false)) 1949 return false; 1950 if (Left.is(tok::colon)) 1951 return !Left.is(TT_ObjCMethodExpr); 1952 if (Left.is(tok::coloncolon)) 1953 return false; 1954 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) 1955 return false; 1956 if (Right.is(tok::ellipsis)) 1957 return Left.Tok.isLiteral(); 1958 if (Left.is(tok::l_square) && Right.is(tok::amp)) 1959 return false; 1960 if (Right.is(TT_PointerOrReference)) 1961 return (Left.is(tok::r_paren) && Left.MatchingParen && 1962 (Left.MatchingParen->is(TT_OverloadedOperatorLParen) || 1963 (Left.MatchingParen->Previous && 1964 Left.MatchingParen->Previous->is(TT_FunctionDeclarationName)))) || 1965 (Left.Tok.isLiteral() || 1966 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) && 1967 (Style.PointerAlignment != FormatStyle::PAS_Left || 1968 Line.IsMultiVariableDeclStmt))); 1969 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) && 1970 (!Left.is(TT_PointerOrReference) || 1971 (Style.PointerAlignment != FormatStyle::PAS_Right && 1972 !Line.IsMultiVariableDeclStmt))) 1973 return true; 1974 if (Left.is(TT_PointerOrReference)) 1975 return Right.Tok.isLiteral() || 1976 Right.isOneOf(TT_BlockComment, Keywords.kw_final, 1977 Keywords.kw_override) || 1978 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) || 1979 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare, 1980 tok::l_paren) && 1981 (Style.PointerAlignment != FormatStyle::PAS_Right && 1982 !Line.IsMultiVariableDeclStmt) && 1983 Left.Previous && 1984 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon)); 1985 if (Right.is(tok::star) && Left.is(tok::l_paren)) 1986 return false; 1987 if (Left.is(tok::l_square)) 1988 return (Left.is(TT_ArrayInitializerLSquare) && 1989 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square)) || 1990 (Left.is(TT_ArraySubscriptLSquare) && Style.SpacesInSquareBrackets && 1991 Right.isNot(tok::r_square)); 1992 if (Right.is(tok::r_square)) 1993 return Right.MatchingParen && 1994 ((Style.SpacesInContainerLiterals && 1995 Right.MatchingParen->is(TT_ArrayInitializerLSquare)) || 1996 (Style.SpacesInSquareBrackets && 1997 Right.MatchingParen->is(TT_ArraySubscriptLSquare))); 1998 if (Right.is(tok::l_square) && 1999 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare) && 2000 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral)) 2001 return false; 2002 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 2003 return !Left.Children.empty(); // No spaces in "{}". 2004 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 2005 (Right.is(tok::r_brace) && Right.MatchingParen && 2006 Right.MatchingParen->BlockKind != BK_Block)) 2007 return !Style.Cpp11BracedListStyle; 2008 if (Left.is(TT_BlockComment)) 2009 return !Left.TokenText.endswith("=*/"); 2010 if (Right.is(tok::l_paren)) { 2011 if (Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) 2012 return true; 2013 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) || 2014 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 2015 (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while, 2016 tok::kw_switch, tok::kw_case, TT_ForEachMacro, 2017 TT_ObjCForIn) || 2018 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch, 2019 tok::kw_new, tok::kw_delete) && 2020 (!Left.Previous || Left.Previous->isNot(tok::period))))) || 2021 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always && 2022 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() || 2023 Left.is(tok::r_paren)) && 2024 Line.Type != LT_PreprocessorDirective); 2025 } 2026 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 2027 return false; 2028 if (Right.is(TT_UnaryOperator)) 2029 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 2030 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr)); 2031 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 2032 tok::r_paren) || 2033 Left.isSimpleTypeSpecifier()) && 2034 Right.is(tok::l_brace) && Right.getNextNonComment() && 2035 Right.BlockKind != BK_Block) 2036 return false; 2037 if (Left.is(tok::period) || Right.is(tok::period)) 2038 return false; 2039 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 2040 return false; 2041 if (Left.is(TT_TemplateCloser) && Left.MatchingParen && 2042 Left.MatchingParen->Previous && 2043 Left.MatchingParen->Previous->is(tok::period)) 2044 // A.<B>DoSomething(); 2045 return false; 2046 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square)) 2047 return false; 2048 return true; 2049 } 2050 2051 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 2052 const FormatToken &Right) { 2053 const FormatToken &Left = *Right.Previous; 2054 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo()) 2055 return true; // Never ever merge two identifiers. 2056 if (Style.Language == FormatStyle::LK_Cpp) { 2057 if (Left.is(tok::kw_operator)) 2058 return Right.is(tok::coloncolon); 2059 } else if (Style.Language == FormatStyle::LK_Proto) { 2060 if (Right.is(tok::period) && 2061 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required, 2062 Keywords.kw_repeated, Keywords.kw_extend)) 2063 return true; 2064 if (Right.is(tok::l_paren) && 2065 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) 2066 return true; 2067 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2068 if (Left.is(TT_JsFatArrow)) 2069 return true; 2070 if (Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in, 2071 Keywords.kw_of) && 2072 (!Left.Previous || !Left.Previous->is(tok::period))) 2073 return true; 2074 if (Left.is(tok::kw_default) && Left.Previous && 2075 Left.Previous->is(tok::kw_export)) 2076 return true; 2077 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace)) 2078 return true; 2079 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion)) 2080 return false; 2081 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator)) 2082 return false; 2083 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) && 2084 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) 2085 return false; 2086 if (Left.is(tok::ellipsis)) 2087 return false; 2088 if (Left.is(TT_TemplateCloser) && 2089 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square, 2090 Keywords.kw_implements, Keywords.kw_extends)) 2091 // Type assertions ('<type>expr') are not followed by whitespace. Other 2092 // locations that should have whitespace following are identified by the 2093 // above set of follower tokens. 2094 return false; 2095 } else if (Style.Language == FormatStyle::LK_Java) { 2096 if (Left.is(tok::r_square) && Right.is(tok::l_brace)) 2097 return true; 2098 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) 2099 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never; 2100 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private, 2101 tok::kw_protected) || 2102 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract, 2103 Keywords.kw_native)) && 2104 Right.is(TT_TemplateOpener)) 2105 return true; 2106 } 2107 if (Left.is(TT_ImplicitStringLiteral)) 2108 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2109 if (Line.Type == LT_ObjCMethodDecl) { 2110 if (Left.is(TT_ObjCMethodSpecifier)) 2111 return true; 2112 if (Left.is(tok::r_paren) && Right.is(tok::identifier)) 2113 // Don't space between ')' and <id> 2114 return false; 2115 } 2116 if (Line.Type == LT_ObjCProperty && 2117 (Right.is(tok::equal) || Left.is(tok::equal))) 2118 return false; 2119 2120 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) || 2121 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) 2122 return true; 2123 if (Right.is(TT_OverloadedOperatorLParen)) 2124 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2125 if (Left.is(tok::comma)) 2126 return true; 2127 if (Right.is(tok::comma)) 2128 return false; 2129 if (Right.isOneOf(TT_CtorInitializerColon, TT_ObjCBlockLParen)) 2130 return true; 2131 if (Right.is(tok::colon)) { 2132 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) || 2133 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi)) 2134 return false; 2135 if (Right.is(TT_ObjCMethodExpr)) 2136 return false; 2137 if (Left.is(tok::question)) 2138 return false; 2139 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon)) 2140 return false; 2141 if (Right.is(TT_DictLiteral)) 2142 return Style.SpacesInContainerLiterals; 2143 return true; 2144 } 2145 if (Left.is(TT_UnaryOperator)) 2146 return Right.is(TT_BinaryOperator); 2147 2148 // If the next token is a binary operator or a selector name, we have 2149 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly. 2150 if (Left.is(TT_CastRParen)) 2151 return Style.SpaceAfterCStyleCast || 2152 Right.isOneOf(TT_BinaryOperator, TT_SelectorName); 2153 2154 if (Left.is(tok::greater) && Right.is(tok::greater)) 2155 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) && 2156 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 2157 if (Right.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) || 2158 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar)) 2159 return false; 2160 if (!Style.SpaceBeforeAssignmentOperators && 2161 Right.getPrecedence() == prec::Assignment) 2162 return false; 2163 if (Right.is(tok::coloncolon) && Left.isNot(tok::l_brace)) 2164 return (Left.is(TT_TemplateOpener) && 2165 Style.Standard == FormatStyle::LS_Cpp03) || 2166 !(Left.isOneOf(tok::identifier, tok::l_paren, tok::r_paren, 2167 tok::l_square) || 2168 Left.isOneOf(TT_TemplateCloser, TT_TemplateOpener)); 2169 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser))) 2170 return Style.SpacesInAngles; 2171 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) || 2172 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && 2173 !Right.is(tok::r_paren))) 2174 return true; 2175 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) && 2176 Right.isNot(TT_FunctionTypeLParen)) 2177 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2178 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) && 2179 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen)) 2180 return false; 2181 if (Right.is(tok::less) && Left.isNot(tok::l_paren) && 2182 Line.startsWith(tok::hash)) 2183 return true; 2184 if (Right.is(TT_TrailingUnaryOperator)) 2185 return false; 2186 if (Left.is(TT_RegexLiteral)) 2187 return false; 2188 return spaceRequiredBetween(Line, Left, Right); 2189 } 2190 2191 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 2192 static bool isAllmanBrace(const FormatToken &Tok) { 2193 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 2194 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral); 2195 } 2196 2197 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 2198 const FormatToken &Right) { 2199 const FormatToken &Left = *Right.Previous; 2200 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0) 2201 return true; 2202 2203 if (Style.Language == FormatStyle::LK_JavaScript) { 2204 // FIXME: This might apply to other languages and token kinds. 2205 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous && 2206 Left.Previous->is(tok::string_literal)) 2207 return true; 2208 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 && 2209 Left.Previous && Left.Previous->is(tok::equal) && 2210 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export, 2211 tok::kw_const) && 2212 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match 2213 // above. 2214 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) 2215 // Object literals on the top level of a file are treated as "enum-style". 2216 // Each key/value pair is put on a separate line, instead of bin-packing. 2217 return true; 2218 if (Left.is(tok::l_brace) && Line.Level == 0 && 2219 (Line.startsWith(tok::kw_enum) || 2220 Line.startsWith(tok::kw_export, tok::kw_enum))) 2221 // JavaScript top-level enum key/value pairs are put on separate lines 2222 // instead of bin-packing. 2223 return true; 2224 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && 2225 !Left.Children.empty()) 2226 // Support AllowShortFunctionsOnASingleLine for JavaScript. 2227 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 2228 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty || 2229 (Left.NestingLevel == 0 && Line.Level == 0 && 2230 Style.AllowShortFunctionsOnASingleLine == 2231 FormatStyle::SFS_Inline); 2232 } else if (Style.Language == FormatStyle::LK_Java) { 2233 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next && 2234 Right.Next->is(tok::string_literal)) 2235 return true; 2236 } 2237 2238 // If the last token before a '}' is a comma or a trailing comment, the 2239 // intention is to insert a line break after it in order to make shuffling 2240 // around entries easier. 2241 const FormatToken *BeforeClosingBrace = nullptr; 2242 if (Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) && 2243 Left.BlockKind != BK_Block && Left.MatchingParen) 2244 BeforeClosingBrace = Left.MatchingParen->Previous; 2245 else if (Right.MatchingParen && 2246 Right.MatchingParen->isOneOf(tok::l_brace, 2247 TT_ArrayInitializerLSquare)) 2248 BeforeClosingBrace = &Left; 2249 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 2250 BeforeClosingBrace->isTrailingComment())) 2251 return true; 2252 2253 if (Right.is(tok::comment)) 2254 return Left.BlockKind != BK_BracedInit && 2255 Left.isNot(TT_CtorInitializerColon) && 2256 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 2257 if (Left.isTrailingComment()) 2258 return true; 2259 if (Left.isStringLiteral() && 2260 (Right.isStringLiteral() || Right.is(TT_ObjCStringLiteral))) 2261 return true; 2262 if (Right.Previous->IsUnterminatedLiteral) 2263 return true; 2264 if (Right.is(tok::lessless) && Right.Next && 2265 Right.Previous->is(tok::string_literal) && 2266 Right.Next->is(tok::string_literal)) 2267 return true; 2268 if (Right.Previous->ClosesTemplateDeclaration && 2269 Right.Previous->MatchingParen && 2270 Right.Previous->MatchingParen->NestingLevel == 0 && 2271 Style.AlwaysBreakTemplateDeclarations) 2272 return true; 2273 if ((Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon)) && 2274 Style.BreakConstructorInitializersBeforeComma && 2275 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 2276 return true; 2277 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) 2278 // Raw string literals are special wrt. line breaks. The author has made a 2279 // deliberate choice and might have aligned the contents of the string 2280 // literal accordingly. Thus, we try keep existing line breaks. 2281 return Right.NewlinesBefore > 0; 2282 if (Right.Previous->is(tok::l_brace) && Right.NestingLevel == 1 && 2283 Style.Language == FormatStyle::LK_Proto) 2284 // Don't put enums onto single lines in protocol buffers. 2285 return true; 2286 if (Right.is(TT_InlineASMBrace)) 2287 return Right.HasUnescapedNewline; 2288 if (isAllmanBrace(Left) || isAllmanBrace(Right)) 2289 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) || 2290 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) || 2291 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct); 2292 if (Style.Language == FormatStyle::LK_Proto && Left.isNot(tok::l_brace) && 2293 Right.is(TT_SelectorName)) 2294 return true; 2295 if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine) 2296 return true; 2297 2298 if ((Style.Language == FormatStyle::LK_Java || 2299 Style.Language == FormatStyle::LK_JavaScript) && 2300 Left.is(TT_LeadingJavaAnnotation) && 2301 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) && 2302 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) 2303 return true; 2304 2305 return false; 2306 } 2307 2308 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 2309 const FormatToken &Right) { 2310 const FormatToken &Left = *Right.Previous; 2311 2312 // Language-specific stuff. 2313 if (Style.Language == FormatStyle::LK_Java) { 2314 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2315 Keywords.kw_implements)) 2316 return false; 2317 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2318 Keywords.kw_implements)) 2319 return true; 2320 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2321 if (Left.is(tok::kw_return)) 2322 return false; // Otherwise a semicolon is inserted. 2323 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace)) 2324 return false; 2325 if (Left.is(TT_JsTypeColon)) 2326 return true; 2327 if (Right.NestingLevel == 0 && Right.is(Keywords.kw_is)) 2328 return false; 2329 if (Left.is(Keywords.kw_in)) 2330 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None; 2331 if (Right.is(Keywords.kw_in)) 2332 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None; 2333 } 2334 2335 if (Left.is(tok::at)) 2336 return false; 2337 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 2338 return false; 2339 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation)) 2340 return !Right.is(tok::l_paren); 2341 if (Right.is(TT_PointerOrReference)) 2342 return Line.IsMultiVariableDeclStmt || 2343 (Style.PointerAlignment == FormatStyle::PAS_Right && 2344 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName))); 2345 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2346 Right.is(tok::kw_operator)) 2347 return true; 2348 if (Left.is(TT_PointerOrReference)) 2349 return false; 2350 if (Right.isTrailingComment()) 2351 // We rely on MustBreakBefore being set correctly here as we should not 2352 // change the "binding" behavior of a comment. 2353 // The first comment in a braced lists is always interpreted as belonging to 2354 // the first list element. Otherwise, it should be placed outside of the 2355 // list. 2356 return Left.BlockKind == BK_BracedInit; 2357 if (Left.is(tok::question) && Right.is(tok::colon)) 2358 return false; 2359 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question)) 2360 return Style.BreakBeforeTernaryOperators; 2361 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question)) 2362 return !Style.BreakBeforeTernaryOperators; 2363 if (Right.is(TT_InheritanceColon)) 2364 return true; 2365 if (Right.is(tok::colon) && 2366 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) 2367 return false; 2368 if (Left.is(tok::colon) && (Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr))) 2369 return true; 2370 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next && 2371 Right.Next->is(TT_ObjCMethodExpr))) 2372 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls. 2373 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 2374 return true; 2375 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen)) 2376 return true; 2377 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen, 2378 TT_OverloadedOperator)) 2379 return false; 2380 if (Left.is(TT_RangeBasedForLoopColon)) 2381 return true; 2382 if (Right.is(TT_RangeBasedForLoopColon)) 2383 return false; 2384 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) || 2385 Left.is(tok::kw_operator)) 2386 return false; 2387 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) && 2388 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) 2389 return false; 2390 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen)) 2391 return false; 2392 if (Left.is(tok::l_paren) && Left.Previous && 2393 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) 2394 return false; 2395 if (Right.is(TT_ImplicitStringLiteral)) 2396 return false; 2397 2398 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser)) 2399 return false; 2400 if (Right.is(tok::r_square) && Right.MatchingParen && 2401 Right.MatchingParen->is(TT_LambdaLSquare)) 2402 return false; 2403 2404 // We only break before r_brace if there was a corresponding break before 2405 // the l_brace, which is tracked by BreakBeforeClosingBrace. 2406 if (Right.is(tok::r_brace)) 2407 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 2408 2409 // Allow breaking after a trailing annotation, e.g. after a method 2410 // declaration. 2411 if (Left.is(TT_TrailingAnnotation)) 2412 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 2413 tok::less, tok::coloncolon); 2414 2415 if (Right.is(tok::kw___attribute)) 2416 return true; 2417 2418 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 2419 return true; 2420 2421 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2422 return true; 2423 2424 if (Left.is(TT_CtorInitializerComma) && 2425 Style.BreakConstructorInitializersBeforeComma) 2426 return false; 2427 if (Right.is(TT_CtorInitializerComma) && 2428 Style.BreakConstructorInitializersBeforeComma) 2429 return true; 2430 if ((Left.is(tok::greater) && Right.is(tok::greater)) || 2431 (Left.is(tok::less) && Right.is(tok::less))) 2432 return false; 2433 if (Right.is(TT_BinaryOperator) && 2434 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 2435 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 2436 Right.getPrecedence() != prec::Assignment)) 2437 return true; 2438 if (Left.is(TT_ArrayInitializerLSquare)) 2439 return true; 2440 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 2441 return true; 2442 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) && 2443 !Left.isOneOf(tok::arrowstar, tok::lessless) && 2444 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 2445 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 2446 Left.getPrecedence() == prec::Assignment)) 2447 return true; 2448 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 2449 tok::kw_class, tok::kw_struct) || 2450 Right.isMemberAccess() || 2451 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless, 2452 tok::colon, tok::l_square, tok::at) || 2453 (Left.is(tok::r_paren) && 2454 Right.isOneOf(tok::identifier, tok::kw_const)) || 2455 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)); 2456 } 2457 2458 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 2459 llvm::errs() << "AnnotatedTokens:\n"; 2460 const FormatToken *Tok = Line.First; 2461 while (Tok) { 2462 llvm::errs() << " M=" << Tok->MustBreakBefore 2463 << " C=" << Tok->CanBreakBefore 2464 << " T=" << getTokenTypeName(Tok->Type) 2465 << " S=" << Tok->SpacesRequiredBefore 2466 << " B=" << Tok->BlockParameterCount 2467 << " P=" << Tok->SplitPenalty << " Name=" << Tok->Tok.getName() 2468 << " L=" << Tok->TotalLength << " PPK=" << Tok->PackingKind 2469 << " FakeLParens="; 2470 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 2471 llvm::errs() << Tok->FakeLParens[i] << "/"; 2472 llvm::errs() << " FakeRParens=" << Tok->FakeRParens << "\n"; 2473 if (!Tok->Next) 2474 assert(Tok == Line.Last); 2475 Tok = Tok->Next; 2476 } 2477 llvm::errs() << "----\n"; 2478 } 2479 2480 } // namespace format 2481 } // namespace clang 2482