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; // The '<'. 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 // In TT_Proto, we must distignuish between: 83 // map<key, value> 84 // msg < item: data > 85 // msg: < item: data > 86 // In TT_TextProto, map<key, value> does not occur. 87 if (Style.Language == FormatStyle::LK_TextProto || 88 (Style.Language == FormatStyle::LK_Proto && Left->Previous && 89 Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) 90 CurrentToken->Type = TT_DictLiteral; 91 else 92 CurrentToken->Type = TT_TemplateCloser; 93 next(); 94 return true; 95 } 96 if (CurrentToken->is(tok::question) && 97 Style.Language == FormatStyle::LK_Java) { 98 next(); 99 continue; 100 } 101 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) || 102 (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext && 103 Style.Language != FormatStyle::LK_Proto && 104 Style.Language != FormatStyle::LK_TextProto)) 105 return false; 106 // If a && or || is found and interpreted as a binary operator, this set 107 // of angles is likely part of something like "a < b && c > d". If the 108 // angles are inside an expression, the ||/&& might also be a binary 109 // operator that was misinterpreted because we are parsing template 110 // parameters. 111 // FIXME: This is getting out of hand, write a decent parser. 112 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) && 113 CurrentToken->Previous->is(TT_BinaryOperator) && 114 Contexts[Contexts.size() - 2].IsExpression && 115 !Line.startsWith(tok::kw_template)) 116 return false; 117 updateParameterCount(Left, CurrentToken); 118 if (Style.Language == FormatStyle::LK_Proto) { 119 if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) { 120 if (CurrentToken->is(tok::colon) || 121 (CurrentToken->isOneOf(tok::l_brace, tok::less) && 122 Previous->isNot(tok::colon))) 123 Previous->Type = TT_SelectorName; 124 } 125 } 126 if (!consumeToken()) 127 return false; 128 } 129 return false; 130 } 131 132 bool parseParens(bool LookForDecls = false) { 133 if (!CurrentToken) 134 return false; 135 FormatToken *Left = CurrentToken->Previous; 136 Left->ParentBracket = Contexts.back().ContextKind; 137 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1); 138 139 // FIXME: This is a bit of a hack. Do better. 140 Contexts.back().ColonIsForRangeExpr = 141 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr; 142 143 bool StartsObjCMethodExpr = false; 144 if (CurrentToken->is(tok::caret)) { 145 // (^ can start a block type. 146 Left->Type = TT_ObjCBlockLParen; 147 } else if (FormatToken *MaybeSel = Left->Previous) { 148 // @selector( starts a selector. 149 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous && 150 MaybeSel->Previous->is(tok::at)) { 151 StartsObjCMethodExpr = true; 152 } 153 } 154 155 if (Left->is(TT_OverloadedOperatorLParen)) { 156 Contexts.back().IsExpression = false; 157 } else if (Style.Language == FormatStyle::LK_JavaScript && 158 (Line.startsWith(Keywords.kw_type, tok::identifier) || 159 Line.startsWith(tok::kw_export, Keywords.kw_type, 160 tok::identifier))) { 161 // type X = (...); 162 // export type X = (...); 163 Contexts.back().IsExpression = false; 164 } else if (Left->Previous && 165 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype, 166 tok::kw_if, tok::kw_while, tok::l_paren, 167 tok::comma) || 168 Left->Previous->endsSequence(tok::kw_constexpr, tok::kw_if) || 169 Left->Previous->is(TT_BinaryOperator))) { 170 // static_assert, if and while usually contain expressions. 171 Contexts.back().IsExpression = true; 172 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 173 (Left->Previous->is(Keywords.kw_function) || 174 (Left->Previous->endsSequence(tok::identifier, 175 Keywords.kw_function)))) { 176 // function(...) or function f(...) 177 Contexts.back().IsExpression = false; 178 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 179 Left->Previous->is(TT_JsTypeColon)) { 180 // let x: (SomeType); 181 Contexts.back().IsExpression = false; 182 } else if (Left->Previous && Left->Previous->is(tok::r_square) && 183 Left->Previous->MatchingParen && 184 Left->Previous->MatchingParen->is(TT_LambdaLSquare)) { 185 // This is a parameter list of a lambda expression. 186 Contexts.back().IsExpression = false; 187 } else if (Line.InPPDirective && 188 (!Left->Previous || !Left->Previous->is(tok::identifier))) { 189 Contexts.back().IsExpression = true; 190 } else if (Contexts[Contexts.size() - 2].CaretFound) { 191 // This is the parameter list of an ObjC block. 192 Contexts.back().IsExpression = false; 193 } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) { 194 Left->Type = TT_AttributeParen; 195 } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) { 196 // The first argument to a foreach macro is a declaration. 197 Contexts.back().IsForEachMacro = true; 198 Contexts.back().IsExpression = false; 199 } else if (Left->Previous && Left->Previous->MatchingParen && 200 Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) { 201 Contexts.back().IsExpression = false; 202 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) { 203 bool IsForOrCatch = 204 Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch); 205 Contexts.back().IsExpression = !IsForOrCatch; 206 } 207 208 if (StartsObjCMethodExpr) { 209 Contexts.back().ColonIsObjCMethodExpr = true; 210 Left->Type = TT_ObjCMethodExpr; 211 } 212 213 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression; 214 bool ProbablyFunctionType = CurrentToken->isOneOf(tok::star, tok::amp); 215 bool HasMultipleLines = false; 216 bool HasMultipleParametersOnALine = false; 217 bool MightBeObjCForRangeLoop = 218 Left->Previous && Left->Previous->is(tok::kw_for); 219 FormatToken *PossibleObjCForInToken = nullptr; 220 while (CurrentToken) { 221 // LookForDecls is set when "if (" has been seen. Check for 222 // 'identifier' '*' 'identifier' followed by not '=' -- this 223 // '*' has to be a binary operator but determineStarAmpUsage() will 224 // categorize it as an unary operator, so set the right type here. 225 if (LookForDecls && CurrentToken->Next) { 226 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 227 if (Prev) { 228 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 229 FormatToken *Next = CurrentToken->Next; 230 if (PrevPrev && PrevPrev->is(tok::identifier) && 231 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 232 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 233 Prev->Type = TT_BinaryOperator; 234 LookForDecls = false; 235 } 236 } 237 } 238 239 if (CurrentToken->Previous->is(TT_PointerOrReference) && 240 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 241 tok::coloncolon)) 242 ProbablyFunctionType = true; 243 if (CurrentToken->is(tok::comma)) 244 MightBeFunctionType = false; 245 if (CurrentToken->Previous->is(TT_BinaryOperator)) 246 Contexts.back().IsExpression = true; 247 if (CurrentToken->is(tok::r_paren)) { 248 if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next && 249 (CurrentToken->Next->is(tok::l_paren) || 250 (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration))) 251 Left->Type = TT_FunctionTypeLParen; 252 Left->MatchingParen = CurrentToken; 253 CurrentToken->MatchingParen = Left; 254 255 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) && 256 Left->Previous && Left->Previous->is(tok::l_paren)) { 257 // Detect the case where macros are used to generate lambdas or 258 // function bodies, e.g.: 259 // auto my_lambda = MARCO((Type *type, int i) { .. body .. }); 260 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) { 261 if (Tok->is(TT_BinaryOperator) && 262 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) 263 Tok->Type = TT_PointerOrReference; 264 } 265 } 266 267 if (StartsObjCMethodExpr) { 268 CurrentToken->Type = TT_ObjCMethodExpr; 269 if (Contexts.back().FirstObjCSelectorName) { 270 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 271 Contexts.back().LongestObjCSelectorName; 272 } 273 } 274 275 if (Left->is(TT_AttributeParen)) 276 CurrentToken->Type = TT_AttributeParen; 277 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation)) 278 CurrentToken->Type = TT_JavaAnnotation; 279 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation)) 280 CurrentToken->Type = TT_LeadingJavaAnnotation; 281 282 if (!HasMultipleLines) 283 Left->PackingKind = PPK_Inconclusive; 284 else if (HasMultipleParametersOnALine) 285 Left->PackingKind = PPK_BinPacked; 286 else 287 Left->PackingKind = PPK_OnePerLine; 288 289 next(); 290 return true; 291 } 292 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 293 return false; 294 295 if (CurrentToken->is(tok::l_brace)) 296 Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen 297 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 298 !CurrentToken->Next->HasUnescapedNewline && 299 !CurrentToken->Next->isTrailingComment()) 300 HasMultipleParametersOnALine = true; 301 if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) || 302 CurrentToken->Previous->isSimpleTypeSpecifier()) && 303 !CurrentToken->is(tok::l_brace)) 304 Contexts.back().IsExpression = false; 305 if (CurrentToken->isOneOf(tok::semi, tok::colon)) { 306 MightBeObjCForRangeLoop = false; 307 if (PossibleObjCForInToken) { 308 PossibleObjCForInToken->Type = TT_Unknown; 309 PossibleObjCForInToken = nullptr; 310 } 311 } 312 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) { 313 PossibleObjCForInToken = CurrentToken; 314 PossibleObjCForInToken->Type = TT_ObjCForIn; 315 } 316 // When we discover a 'new', we set CanBeExpression to 'false' in order to 317 // parse the type correctly. Reset that after a comma. 318 if (CurrentToken->is(tok::comma)) 319 Contexts.back().CanBeExpression = true; 320 321 FormatToken *Tok = CurrentToken; 322 if (!consumeToken()) 323 return false; 324 updateParameterCount(Left, Tok); 325 if (CurrentToken && CurrentToken->HasUnescapedNewline) 326 HasMultipleLines = true; 327 } 328 return false; 329 } 330 331 bool parseSquare() { 332 if (!CurrentToken) 333 return false; 334 335 // A '[' could be an index subscript (after an identifier or after 336 // ')' or ']'), it could be the start of an Objective-C method 337 // expression, or it could the start of an Objective-C array literal. 338 FormatToken *Left = CurrentToken->Previous; 339 Left->ParentBracket = Contexts.back().ContextKind; 340 FormatToken *Parent = Left->getPreviousNonComment(); 341 342 // Cases where '>' is followed by '['. 343 // In C++, this can happen either in array of templates (foo<int>[10]) 344 // or when array is a nested template type (unique_ptr<type1<type2>[]>). 345 bool CppArrayTemplates = 346 Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) && 347 (Contexts.back().CanBeExpression || Contexts.back().IsExpression || 348 Contexts.back().InTemplateArgument); 349 350 bool StartsObjCMethodExpr = 351 !CppArrayTemplates && Style.isCpp() && 352 Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) && 353 CurrentToken->isNot(tok::l_brace) && 354 (!Parent || 355 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 356 tok::kw_return, tok::kw_throw) || 357 Parent->isUnaryOperator() || 358 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) || 359 getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown); 360 bool ColonFound = false; 361 362 unsigned BindingIncrease = 1; 363 if (Left->isCppStructuredBinding(Style)) { 364 Left->Type = TT_StructuredBindingLSquare; 365 } else if (Left->is(TT_Unknown)) { 366 if (StartsObjCMethodExpr) { 367 Left->Type = TT_ObjCMethodExpr; 368 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent && 369 Contexts.back().ContextKind == tok::l_brace && 370 Parent->isOneOf(tok::l_brace, tok::comma)) { 371 Left->Type = TT_JsComputedPropertyName; 372 } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace && 373 Parent && Parent->isOneOf(tok::l_brace, tok::comma)) { 374 Left->Type = TT_DesignatedInitializerLSquare; 375 } else if (CurrentToken->is(tok::r_square) && Parent && 376 Parent->is(TT_TemplateCloser)) { 377 Left->Type = TT_ArraySubscriptLSquare; 378 } else if (Style.Language == FormatStyle::LK_Proto || 379 Style.Language == FormatStyle::LK_TextProto) { 380 // Square braces in LK_Proto can either be message field attributes: 381 // 382 // optional Aaa aaa = 1 [ 383 // (aaa) = aaa 384 // ]; 385 // 386 // extensions 123 [ 387 // (aaa) = aaa 388 // ]; 389 // 390 // or text proto extensions (in options): 391 // 392 // option (Aaa.options) = { 393 // [type.type/type] { 394 // key: value 395 // } 396 // } 397 // 398 // or repeated fields (in options): 399 // 400 // option (Aaa.options) = { 401 // keys: [ 1, 2, 3 ] 402 // } 403 // 404 // In the first and the third case we want to spread the contents inside 405 // the square braces; in the second we want to keep them inline. 406 Left->Type = TT_ArrayInitializerLSquare; 407 if (!Left->endsSequence(tok::l_square, tok::numeric_constant, 408 tok::equal) && 409 !Left->endsSequence(tok::l_square, tok::numeric_constant, 410 tok::identifier) && 411 !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) { 412 Left->Type = TT_ProtoExtensionLSquare; 413 BindingIncrease = 10; 414 } 415 } else if (!CppArrayTemplates && Parent && 416 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at, 417 tok::comma, tok::l_paren, tok::l_square, 418 tok::question, tok::colon, tok::kw_return, 419 // Should only be relevant to JavaScript: 420 tok::kw_default)) { 421 Left->Type = TT_ArrayInitializerLSquare; 422 } else { 423 BindingIncrease = 10; 424 Left->Type = TT_ArraySubscriptLSquare; 425 } 426 } 427 428 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease); 429 Contexts.back().IsExpression = true; 430 if (Style.Language == FormatStyle::LK_JavaScript && Parent && 431 Parent->is(TT_JsTypeColon)) 432 Contexts.back().IsExpression = false; 433 434 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr; 435 436 while (CurrentToken) { 437 if (CurrentToken->is(tok::r_square)) { 438 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_paren) && 439 Left->is(TT_ObjCMethodExpr)) { 440 // An ObjC method call is rarely followed by an open parenthesis. 441 // FIXME: Do we incorrectly label ":" with this? 442 StartsObjCMethodExpr = false; 443 Left->Type = TT_Unknown; 444 } 445 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) { 446 CurrentToken->Type = TT_ObjCMethodExpr; 447 // determineStarAmpUsage() thinks that '*' '[' is allocating an 448 // array of pointers, but if '[' starts a selector then '*' is a 449 // binary operator. 450 if (Parent && Parent->is(TT_PointerOrReference)) 451 Parent->Type = TT_BinaryOperator; 452 } 453 Left->MatchingParen = CurrentToken; 454 CurrentToken->MatchingParen = Left; 455 if (Contexts.back().FirstObjCSelectorName) { 456 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 457 Contexts.back().LongestObjCSelectorName; 458 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts = 459 Left->ParameterCount; 460 if (Left->BlockParameterCount > 1) 461 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0; 462 } 463 next(); 464 return true; 465 } 466 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace)) 467 return false; 468 if (CurrentToken->is(tok::colon)) { 469 if (Left->isOneOf(TT_ArraySubscriptLSquare, 470 TT_DesignatedInitializerLSquare)) { 471 Left->Type = TT_ObjCMethodExpr; 472 StartsObjCMethodExpr = true; 473 // ParameterCount might have been set to 1 before expression was 474 // recognized as ObjCMethodExpr (as '1 + number of commas' formula is 475 // used for other expression types). Parameter counter has to be, 476 // therefore, reset to 0. 477 Left->ParameterCount = 0; 478 Contexts.back().ColonIsObjCMethodExpr = true; 479 if (Parent && Parent->is(tok::r_paren)) 480 Parent->Type = TT_CastRParen; 481 } 482 ColonFound = true; 483 } 484 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) && 485 !ColonFound) 486 Left->Type = TT_ArrayInitializerLSquare; 487 FormatToken *Tok = CurrentToken; 488 if (!consumeToken()) 489 return false; 490 updateParameterCount(Left, Tok); 491 } 492 return false; 493 } 494 495 bool parseBrace() { 496 if (CurrentToken) { 497 FormatToken *Left = CurrentToken->Previous; 498 Left->ParentBracket = Contexts.back().ContextKind; 499 500 if (Contexts.back().CaretFound) 501 Left->Type = TT_ObjCBlockLBrace; 502 Contexts.back().CaretFound = false; 503 504 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1); 505 Contexts.back().ColonIsDictLiteral = true; 506 if (Left->BlockKind == BK_BracedInit) 507 Contexts.back().IsExpression = true; 508 if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 509 Left->Previous->is(TT_JsTypeColon)) 510 Contexts.back().IsExpression = false; 511 512 while (CurrentToken) { 513 if (CurrentToken->is(tok::r_brace)) { 514 Left->MatchingParen = CurrentToken; 515 CurrentToken->MatchingParen = Left; 516 next(); 517 return true; 518 } 519 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square)) 520 return false; 521 updateParameterCount(Left, CurrentToken); 522 if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) { 523 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 524 if (Previous->is(TT_JsTypeOptionalQuestion)) 525 Previous = Previous->getPreviousNonComment(); 526 if ((CurrentToken->is(tok::colon) && 527 (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) || 528 Style.Language == FormatStyle::LK_Proto || 529 Style.Language == FormatStyle::LK_TextProto) { 530 Left->Type = TT_DictLiteral; 531 if (Previous->Tok.getIdentifierInfo() || 532 Previous->is(tok::string_literal)) 533 Previous->Type = TT_SelectorName; 534 } 535 if (CurrentToken->is(tok::colon) || 536 Style.Language == FormatStyle::LK_JavaScript) 537 Left->Type = TT_DictLiteral; 538 } 539 if (CurrentToken->is(tok::comma) && 540 Style.Language == FormatStyle::LK_JavaScript) 541 Left->Type = TT_DictLiteral; 542 if (!consumeToken()) 543 return false; 544 } 545 } 546 return true; 547 } 548 549 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 550 if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block) 551 ++Left->BlockParameterCount; 552 if (Left->Type == TT_ObjCMethodExpr) { 553 if (Current->is(tok::colon)) 554 ++Left->ParameterCount; 555 } else if (Current->is(tok::comma)) { 556 ++Left->ParameterCount; 557 if (!Left->Role) 558 Left->Role.reset(new CommaSeparatedList(Style)); 559 Left->Role->CommaFound(Current); 560 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 561 Left->ParameterCount = 1; 562 } 563 } 564 565 bool parseConditional() { 566 while (CurrentToken) { 567 if (CurrentToken->is(tok::colon)) { 568 CurrentToken->Type = TT_ConditionalExpr; 569 next(); 570 return true; 571 } 572 if (!consumeToken()) 573 return false; 574 } 575 return false; 576 } 577 578 bool parseTemplateDeclaration() { 579 if (CurrentToken && CurrentToken->is(tok::less)) { 580 CurrentToken->Type = TT_TemplateOpener; 581 next(); 582 if (!parseAngle()) 583 return false; 584 if (CurrentToken) 585 CurrentToken->Previous->ClosesTemplateDeclaration = true; 586 return true; 587 } 588 return false; 589 } 590 591 bool consumeToken() { 592 FormatToken *Tok = CurrentToken; 593 next(); 594 switch (Tok->Tok.getKind()) { 595 case tok::plus: 596 case tok::minus: 597 if (!Tok->Previous && Line.MustBeDeclaration) 598 Tok->Type = TT_ObjCMethodSpecifier; 599 break; 600 case tok::colon: 601 if (!Tok->Previous) 602 return false; 603 // Colons from ?: are handled in parseConditional(). 604 if (Style.Language == FormatStyle::LK_JavaScript) { 605 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop 606 (Contexts.size() == 1 && // switch/case labels 607 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) || 608 Contexts.back().ContextKind == tok::l_paren || // function params 609 Contexts.back().ContextKind == tok::l_square || // array type 610 (!Contexts.back().IsExpression && 611 Contexts.back().ContextKind == tok::l_brace) || // object type 612 (Contexts.size() == 1 && 613 Line.MustBeDeclaration)) { // method/property declaration 614 Contexts.back().IsExpression = false; 615 Tok->Type = TT_JsTypeColon; 616 break; 617 } 618 } 619 if (Contexts.back().ColonIsDictLiteral || 620 Style.Language == FormatStyle::LK_Proto || 621 Style.Language == FormatStyle::LK_TextProto) { 622 Tok->Type = TT_DictLiteral; 623 if (Style.Language == FormatStyle::LK_TextProto) { 624 if (FormatToken *Previous = Tok->getPreviousNonComment()) 625 Previous->Type = TT_SelectorName; 626 } 627 } else if (Contexts.back().ColonIsObjCMethodExpr || 628 Line.startsWith(TT_ObjCMethodSpecifier)) { 629 Tok->Type = TT_ObjCMethodExpr; 630 const FormatToken *BeforePrevious = Tok->Previous->Previous; 631 if (!BeforePrevious || 632 !(BeforePrevious->is(TT_CastRParen) || 633 (BeforePrevious->is(TT_ObjCMethodExpr) && 634 BeforePrevious->is(tok::colon))) || 635 BeforePrevious->is(tok::r_square) || 636 Contexts.back().LongestObjCSelectorName == 0) { 637 Tok->Previous->Type = TT_SelectorName; 638 if (!Contexts.back().FirstObjCSelectorName) 639 Contexts.back().FirstObjCSelectorName = Tok->Previous; 640 else if (Tok->Previous->ColumnWidth > 641 Contexts.back().LongestObjCSelectorName) 642 Contexts.back().LongestObjCSelectorName = 643 Tok->Previous->ColumnWidth; 644 } 645 } else if (Contexts.back().ColonIsForRangeExpr) { 646 Tok->Type = TT_RangeBasedForLoopColon; 647 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 648 Tok->Type = TT_BitFieldColon; 649 } else if (Contexts.size() == 1 && 650 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) { 651 if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren, 652 tok::kw_noexcept)) 653 Tok->Type = TT_CtorInitializerColon; 654 else 655 Tok->Type = TT_InheritanceColon; 656 } else if (Tok->Previous->is(tok::identifier) && Tok->Next && 657 Tok->Next->isOneOf(tok::r_paren, tok::comma)) { 658 // This handles a special macro in ObjC code where selectors including 659 // the colon are passed as macro arguments. 660 Tok->Type = TT_ObjCMethodExpr; 661 } else if (Contexts.back().ContextKind == tok::l_paren) { 662 Tok->Type = TT_InlineASMColon; 663 } 664 break; 665 case tok::pipe: 666 case tok::amp: 667 // | and & in declarations/type expressions represent union and 668 // intersection types, respectively. 669 if (Style.Language == FormatStyle::LK_JavaScript && 670 !Contexts.back().IsExpression) 671 Tok->Type = TT_JsTypeOperator; 672 break; 673 case tok::kw_if: 674 case tok::kw_while: 675 if (Tok->is(tok::kw_if) && CurrentToken && 676 CurrentToken->is(tok::kw_constexpr)) 677 next(); 678 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 679 next(); 680 if (!parseParens(/*LookForDecls=*/true)) 681 return false; 682 } 683 break; 684 case tok::kw_for: 685 if (Style.Language == FormatStyle::LK_JavaScript) { 686 // x.for and {for: ...} 687 if ((Tok->Previous && Tok->Previous->is(tok::period)) || 688 (Tok->Next && Tok->Next->is(tok::colon))) 689 break; 690 // JS' for await ( ... 691 if (CurrentToken && CurrentToken->is(Keywords.kw_await)) 692 next(); 693 } 694 Contexts.back().ColonIsForRangeExpr = true; 695 next(); 696 if (!parseParens()) 697 return false; 698 break; 699 case tok::l_paren: 700 // When faced with 'operator()()', the kw_operator handler incorrectly 701 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make 702 // the first two parens OverloadedOperators and the second l_paren an 703 // OverloadedOperatorLParen. 704 if (Tok->Previous && Tok->Previous->is(tok::r_paren) && 705 Tok->Previous->MatchingParen && 706 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) { 707 Tok->Previous->Type = TT_OverloadedOperator; 708 Tok->Previous->MatchingParen->Type = TT_OverloadedOperator; 709 Tok->Type = TT_OverloadedOperatorLParen; 710 } 711 712 if (!parseParens()) 713 return false; 714 if (Line.MustBeDeclaration && Contexts.size() == 1 && 715 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) && 716 (!Tok->Previous || 717 !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute, 718 TT_LeadingJavaAnnotation))) 719 Line.MightBeFunctionDecl = true; 720 break; 721 case tok::l_square: 722 if (!parseSquare()) 723 return false; 724 break; 725 case tok::l_brace: 726 if (Style.Language == FormatStyle::LK_TextProto) { 727 FormatToken *Previous = Tok->getPreviousNonComment(); 728 if (Previous && Previous->Type != TT_DictLiteral) 729 Previous->Type = TT_SelectorName; 730 } 731 if (!parseBrace()) 732 return false; 733 break; 734 case tok::less: 735 if (parseAngle()) { 736 Tok->Type = TT_TemplateOpener; 737 // In TT_Proto, we must distignuish between: 738 // map<key, value> 739 // msg < item: data > 740 // msg: < item: data > 741 // In TT_TextProto, map<key, value> does not occur. 742 if (Style.Language == FormatStyle::LK_TextProto || 743 (Style.Language == FormatStyle::LK_Proto && Tok->Previous && 744 Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) { 745 Tok->Type = TT_DictLiteral; 746 FormatToken *Previous = Tok->getPreviousNonComment(); 747 if (Previous && Previous->Type != TT_DictLiteral) 748 Previous->Type = TT_SelectorName; 749 } 750 } else { 751 Tok->Type = TT_BinaryOperator; 752 NonTemplateLess.insert(Tok); 753 CurrentToken = Tok; 754 next(); 755 } 756 break; 757 case tok::r_paren: 758 case tok::r_square: 759 return false; 760 case tok::r_brace: 761 // Lines can start with '}'. 762 if (Tok->Previous) 763 return false; 764 break; 765 case tok::greater: 766 if (Style.Language != FormatStyle::LK_TextProto) 767 Tok->Type = TT_BinaryOperator; 768 break; 769 case tok::kw_operator: 770 if (Style.Language == FormatStyle::LK_TextProto || 771 Style.Language == FormatStyle::LK_Proto) 772 break; 773 while (CurrentToken && 774 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 775 if (CurrentToken->isOneOf(tok::star, tok::amp)) 776 CurrentToken->Type = TT_PointerOrReference; 777 consumeToken(); 778 if (CurrentToken && 779 CurrentToken->Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator, 780 tok::comma)) 781 CurrentToken->Previous->Type = TT_OverloadedOperator; 782 } 783 if (CurrentToken) { 784 CurrentToken->Type = TT_OverloadedOperatorLParen; 785 if (CurrentToken->Previous->is(TT_BinaryOperator)) 786 CurrentToken->Previous->Type = TT_OverloadedOperator; 787 } 788 break; 789 case tok::question: 790 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next && 791 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren, 792 tok::r_brace)) { 793 // Question marks before semicolons, colons, etc. indicate optional 794 // types (fields, parameters), e.g. 795 // function(x?: string, y?) {...} 796 // class X { y?; } 797 Tok->Type = TT_JsTypeOptionalQuestion; 798 break; 799 } 800 // Declarations cannot be conditional expressions, this can only be part 801 // of a type declaration. 802 if (Line.MustBeDeclaration && !Contexts.back().IsExpression && 803 Style.Language == FormatStyle::LK_JavaScript) 804 break; 805 parseConditional(); 806 break; 807 case tok::kw_template: 808 parseTemplateDeclaration(); 809 break; 810 case tok::comma: 811 if (Contexts.back().InCtorInitializer) 812 Tok->Type = TT_CtorInitializerComma; 813 else if (Contexts.back().InInheritanceList) 814 Tok->Type = TT_InheritanceComma; 815 else if (Contexts.back().FirstStartOfName && 816 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) { 817 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 818 Line.IsMultiVariableDeclStmt = true; 819 } 820 if (Contexts.back().IsForEachMacro) 821 Contexts.back().IsExpression = true; 822 break; 823 case tok::identifier: 824 if (Tok->isOneOf(Keywords.kw___has_include, 825 Keywords.kw___has_include_next)) { 826 parseHasInclude(); 827 } 828 break; 829 default: 830 break; 831 } 832 return true; 833 } 834 835 void parseIncludeDirective() { 836 if (CurrentToken && CurrentToken->is(tok::less)) { 837 next(); 838 while (CurrentToken) { 839 // Mark tokens up to the trailing line comments as implicit string 840 // literals. 841 if (CurrentToken->isNot(tok::comment) && 842 !CurrentToken->TokenText.startswith("//")) 843 CurrentToken->Type = TT_ImplicitStringLiteral; 844 next(); 845 } 846 } 847 } 848 849 void parseWarningOrError() { 850 next(); 851 // We still want to format the whitespace left of the first token of the 852 // warning or error. 853 next(); 854 while (CurrentToken) { 855 CurrentToken->Type = TT_ImplicitStringLiteral; 856 next(); 857 } 858 } 859 860 void parsePragma() { 861 next(); // Consume "pragma". 862 if (CurrentToken && 863 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) { 864 bool IsMark = CurrentToken->is(Keywords.kw_mark); 865 next(); // Consume "mark". 866 next(); // Consume first token (so we fix leading whitespace). 867 while (CurrentToken) { 868 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator)) 869 CurrentToken->Type = TT_ImplicitStringLiteral; 870 next(); 871 } 872 } 873 } 874 875 void parseHasInclude() { 876 if (!CurrentToken || !CurrentToken->is(tok::l_paren)) 877 return; 878 next(); // '(' 879 parseIncludeDirective(); 880 next(); // ')' 881 } 882 883 LineType parsePreprocessorDirective() { 884 bool IsFirstToken = CurrentToken->IsFirst; 885 LineType Type = LT_PreprocessorDirective; 886 next(); 887 if (!CurrentToken) 888 return Type; 889 890 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) { 891 // JavaScript files can contain shebang lines of the form: 892 // #!/usr/bin/env node 893 // Treat these like C++ #include directives. 894 while (CurrentToken) { 895 // Tokens cannot be comments here. 896 CurrentToken->Type = TT_ImplicitStringLiteral; 897 next(); 898 } 899 return LT_ImportStatement; 900 } 901 902 if (CurrentToken->Tok.is(tok::numeric_constant)) { 903 CurrentToken->SpacesRequiredBefore = 1; 904 return Type; 905 } 906 // Hashes in the middle of a line can lead to any strange token 907 // sequence. 908 if (!CurrentToken->Tok.getIdentifierInfo()) 909 return Type; 910 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 911 case tok::pp_include: 912 case tok::pp_include_next: 913 case tok::pp_import: 914 next(); 915 parseIncludeDirective(); 916 Type = LT_ImportStatement; 917 break; 918 case tok::pp_error: 919 case tok::pp_warning: 920 parseWarningOrError(); 921 break; 922 case tok::pp_pragma: 923 parsePragma(); 924 break; 925 case tok::pp_if: 926 case tok::pp_elif: 927 Contexts.back().IsExpression = true; 928 parseLine(); 929 break; 930 default: 931 break; 932 } 933 while (CurrentToken) { 934 FormatToken *Tok = CurrentToken; 935 next(); 936 if (Tok->is(tok::l_paren)) 937 parseParens(); 938 else if (Tok->isOneOf(Keywords.kw___has_include, 939 Keywords.kw___has_include_next)) 940 parseHasInclude(); 941 } 942 return Type; 943 } 944 945 public: 946 LineType parseLine() { 947 NonTemplateLess.clear(); 948 if (CurrentToken->is(tok::hash)) 949 return parsePreprocessorDirective(); 950 951 // Directly allow to 'import <string-literal>' to support protocol buffer 952 // definitions (github.com/google/protobuf) or missing "#" (either way we 953 // should not break the line). 954 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 955 if ((Style.Language == FormatStyle::LK_Java && 956 CurrentToken->is(Keywords.kw_package)) || 957 (Info && Info->getPPKeywordID() == tok::pp_import && 958 CurrentToken->Next && 959 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier, 960 tok::kw_static))) { 961 next(); 962 parseIncludeDirective(); 963 return LT_ImportStatement; 964 } 965 966 // If this line starts and ends in '<' and '>', respectively, it is likely 967 // part of "#define <a/b.h>". 968 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 969 parseIncludeDirective(); 970 return LT_ImportStatement; 971 } 972 973 // In .proto files, top-level options are very similar to import statements 974 // and should not be line-wrapped. 975 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 && 976 CurrentToken->is(Keywords.kw_option)) { 977 next(); 978 if (CurrentToken && CurrentToken->is(tok::identifier)) 979 return LT_ImportStatement; 980 } 981 982 bool KeywordVirtualFound = false; 983 bool ImportStatement = false; 984 985 // import {...} from '...'; 986 if (Style.Language == FormatStyle::LK_JavaScript && 987 CurrentToken->is(Keywords.kw_import)) 988 ImportStatement = true; 989 990 while (CurrentToken) { 991 if (CurrentToken->is(tok::kw_virtual)) 992 KeywordVirtualFound = true; 993 if (Style.Language == FormatStyle::LK_JavaScript) { 994 // export {...} from '...'; 995 // An export followed by "from 'some string';" is a re-export from 996 // another module identified by a URI and is treated as a 997 // LT_ImportStatement (i.e. prevent wraps on it for long URIs). 998 // Just "export {...};" or "export class ..." should not be treated as 999 // an import in this sense. 1000 if (Line.First->is(tok::kw_export) && 1001 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next && 1002 CurrentToken->Next->isStringLiteral()) 1003 ImportStatement = true; 1004 if (isClosureImportStatement(*CurrentToken)) 1005 ImportStatement = true; 1006 } 1007 if (!consumeToken()) 1008 return LT_Invalid; 1009 } 1010 if (KeywordVirtualFound) 1011 return LT_VirtualFunctionDecl; 1012 if (ImportStatement) 1013 return LT_ImportStatement; 1014 1015 if (Line.startsWith(TT_ObjCMethodSpecifier)) { 1016 if (Contexts.back().FirstObjCSelectorName) 1017 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 1018 Contexts.back().LongestObjCSelectorName; 1019 return LT_ObjCMethodDecl; 1020 } 1021 1022 return LT_Other; 1023 } 1024 1025 private: 1026 bool isClosureImportStatement(const FormatToken &Tok) { 1027 // FIXME: Closure-library specific stuff should not be hard-coded but be 1028 // configurable. 1029 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) && 1030 Tok.Next->Next && 1031 (Tok.Next->Next->TokenText == "module" || 1032 Tok.Next->Next->TokenText == "provide" || 1033 Tok.Next->Next->TokenText == "require" || 1034 Tok.Next->Next->TokenText == "forwardDeclare") && 1035 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren); 1036 } 1037 1038 void resetTokenMetadata(FormatToken *Token) { 1039 if (!Token) 1040 return; 1041 1042 // Reset token type in case we have already looked at it and then 1043 // recovered from an error (e.g. failure to find the matching >). 1044 if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro, 1045 TT_FunctionLBrace, TT_ImplicitStringLiteral, 1046 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow, 1047 TT_OverloadedOperator, TT_RegexLiteral, 1048 TT_TemplateString, TT_ObjCStringLiteral)) 1049 CurrentToken->Type = TT_Unknown; 1050 CurrentToken->Role.reset(); 1051 CurrentToken->MatchingParen = nullptr; 1052 CurrentToken->FakeLParens.clear(); 1053 CurrentToken->FakeRParens = 0; 1054 } 1055 1056 void next() { 1057 if (CurrentToken) { 1058 CurrentToken->NestingLevel = Contexts.size() - 1; 1059 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 1060 modifyContext(*CurrentToken); 1061 determineTokenType(*CurrentToken); 1062 CurrentToken = CurrentToken->Next; 1063 } 1064 1065 resetTokenMetadata(CurrentToken); 1066 } 1067 1068 /// \brief A struct to hold information valid in a specific context, e.g. 1069 /// a pair of parenthesis. 1070 struct Context { 1071 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 1072 bool IsExpression) 1073 : ContextKind(ContextKind), BindingStrength(BindingStrength), 1074 IsExpression(IsExpression) {} 1075 1076 tok::TokenKind ContextKind; 1077 unsigned BindingStrength; 1078 bool IsExpression; 1079 unsigned LongestObjCSelectorName = 0; 1080 bool ColonIsForRangeExpr = false; 1081 bool ColonIsDictLiteral = false; 1082 bool ColonIsObjCMethodExpr = false; 1083 FormatToken *FirstObjCSelectorName = nullptr; 1084 FormatToken *FirstStartOfName = nullptr; 1085 bool CanBeExpression = true; 1086 bool InTemplateArgument = false; 1087 bool InCtorInitializer = false; 1088 bool InInheritanceList = false; 1089 bool CaretFound = false; 1090 bool IsForEachMacro = false; 1091 }; 1092 1093 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime 1094 /// of each instance. 1095 struct ScopedContextCreator { 1096 AnnotatingParser &P; 1097 1098 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 1099 unsigned Increase) 1100 : P(P) { 1101 P.Contexts.push_back(Context(ContextKind, 1102 P.Contexts.back().BindingStrength + Increase, 1103 P.Contexts.back().IsExpression)); 1104 } 1105 1106 ~ScopedContextCreator() { P.Contexts.pop_back(); } 1107 }; 1108 1109 void modifyContext(const FormatToken &Current) { 1110 if (Current.getPrecedence() == prec::Assignment && 1111 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) && 1112 // Type aliases use `type X = ...;` in TypeScript and can be exported 1113 // using `export type ...`. 1114 !(Style.Language == FormatStyle::LK_JavaScript && 1115 (Line.startsWith(Keywords.kw_type, tok::identifier) || 1116 Line.startsWith(tok::kw_export, Keywords.kw_type, 1117 tok::identifier))) && 1118 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 1119 Contexts.back().IsExpression = true; 1120 if (!Line.startsWith(TT_UnaryOperator)) { 1121 for (FormatToken *Previous = Current.Previous; 1122 Previous && Previous->Previous && 1123 !Previous->Previous->isOneOf(tok::comma, tok::semi); 1124 Previous = Previous->Previous) { 1125 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 1126 Previous = Previous->MatchingParen; 1127 if (!Previous) 1128 break; 1129 } 1130 if (Previous->opensScope()) 1131 break; 1132 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) && 1133 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) && 1134 Previous->Previous && Previous->Previous->isNot(tok::equal)) 1135 Previous->Type = TT_PointerOrReference; 1136 } 1137 } 1138 } else if (Current.is(tok::lessless) && 1139 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) { 1140 Contexts.back().IsExpression = true; 1141 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 1142 Contexts.back().IsExpression = true; 1143 } else if (Current.is(TT_TrailingReturnArrow)) { 1144 Contexts.back().IsExpression = false; 1145 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) { 1146 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java; 1147 } else if (Current.Previous && 1148 Current.Previous->is(TT_CtorInitializerColon)) { 1149 Contexts.back().IsExpression = true; 1150 Contexts.back().InCtorInitializer = true; 1151 } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) { 1152 Contexts.back().InInheritanceList = true; 1153 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 1154 for (FormatToken *Previous = Current.Previous; 1155 Previous && Previous->isOneOf(tok::star, tok::amp); 1156 Previous = Previous->Previous) 1157 Previous->Type = TT_PointerOrReference; 1158 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer) 1159 Contexts.back().IsExpression = false; 1160 } else if (Current.is(tok::kw_new)) { 1161 Contexts.back().CanBeExpression = false; 1162 } else if (Current.isOneOf(tok::semi, tok::exclaim)) { 1163 // This should be the condition or increment in a for-loop. 1164 Contexts.back().IsExpression = true; 1165 } 1166 } 1167 1168 void determineTokenType(FormatToken &Current) { 1169 if (!Current.is(TT_Unknown)) 1170 // The token type is already known. 1171 return; 1172 1173 if (Style.Language == FormatStyle::LK_JavaScript) { 1174 if (Current.is(tok::exclaim)) { 1175 if (Current.Previous && 1176 (Current.Previous->isOneOf(tok::identifier, tok::kw_namespace, 1177 tok::r_paren, tok::r_square, 1178 tok::r_brace) || 1179 Current.Previous->Tok.isLiteral())) { 1180 Current.Type = TT_JsNonNullAssertion; 1181 return; 1182 } 1183 if (Current.Next && 1184 Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) { 1185 Current.Type = TT_JsNonNullAssertion; 1186 return; 1187 } 1188 } 1189 } 1190 1191 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1192 // function declaration have been found. In this case, 'Current' is a 1193 // trailing token of this declaration and thus cannot be a name. 1194 if (Current.is(Keywords.kw_instanceof)) { 1195 Current.Type = TT_BinaryOperator; 1196 } else if (isStartOfName(Current) && 1197 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 1198 Contexts.back().FirstStartOfName = &Current; 1199 Current.Type = TT_StartOfName; 1200 } else if (Current.is(tok::semi)) { 1201 // Reset FirstStartOfName after finding a semicolon so that a for loop 1202 // with multiple increment statements is not confused with a for loop 1203 // having multiple variable declarations. 1204 Contexts.back().FirstStartOfName = nullptr; 1205 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) { 1206 AutoFound = true; 1207 } else if (Current.is(tok::arrow) && 1208 Style.Language == FormatStyle::LK_Java) { 1209 Current.Type = TT_LambdaArrow; 1210 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration && 1211 Current.NestingLevel == 0) { 1212 Current.Type = TT_TrailingReturnArrow; 1213 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 1214 Current.Type = determineStarAmpUsage(Current, 1215 Contexts.back().CanBeExpression && 1216 Contexts.back().IsExpression, 1217 Contexts.back().InTemplateArgument); 1218 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 1219 Current.Type = determinePlusMinusCaretUsage(Current); 1220 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret)) 1221 Contexts.back().CaretFound = true; 1222 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 1223 Current.Type = determineIncrementUsage(Current); 1224 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 1225 Current.Type = TT_UnaryOperator; 1226 } else if (Current.is(tok::question)) { 1227 if (Style.Language == FormatStyle::LK_JavaScript && 1228 Line.MustBeDeclaration && !Contexts.back().IsExpression) { 1229 // In JavaScript, `interface X { foo?(): bar; }` is an optional method 1230 // on the interface, not a ternary expression. 1231 Current.Type = TT_JsTypeOptionalQuestion; 1232 } else { 1233 Current.Type = TT_ConditionalExpr; 1234 } 1235 } else if (Current.isBinaryOperator() && 1236 (!Current.Previous || Current.Previous->isNot(tok::l_square)) && 1237 (!Current.is(tok::greater) && 1238 Style.Language != FormatStyle::LK_TextProto)) { 1239 Current.Type = TT_BinaryOperator; 1240 } else if (Current.is(tok::comment)) { 1241 if (Current.TokenText.startswith("/*")) { 1242 if (Current.TokenText.endswith("*/")) 1243 Current.Type = TT_BlockComment; 1244 else 1245 // The lexer has for some reason determined a comment here. But we 1246 // cannot really handle it, if it isn't properly terminated. 1247 Current.Tok.setKind(tok::unknown); 1248 } else { 1249 Current.Type = TT_LineComment; 1250 } 1251 } else if (Current.is(tok::r_paren)) { 1252 if (rParenEndsCast(Current)) 1253 Current.Type = TT_CastRParen; 1254 if (Current.MatchingParen && Current.Next && 1255 !Current.Next->isBinaryOperator() && 1256 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace, 1257 tok::comma, tok::period, tok::arrow, 1258 tok::coloncolon)) 1259 if (FormatToken *AfterParen = Current.MatchingParen->Next) { 1260 // Make sure this isn't the return type of an Obj-C block declaration 1261 if (AfterParen->Tok.isNot(tok::caret)) { 1262 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) 1263 if (BeforeParen->is(tok::identifier) && 1264 BeforeParen->TokenText == BeforeParen->TokenText.upper() && 1265 (!BeforeParen->Previous || 1266 BeforeParen->Previous->ClosesTemplateDeclaration)) 1267 Current.Type = TT_FunctionAnnotationRParen; 1268 } 1269 } 1270 } else if (Current.is(tok::at) && Current.Next && 1271 Style.Language != FormatStyle::LK_JavaScript && 1272 Style.Language != FormatStyle::LK_Java) { 1273 // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it 1274 // marks declarations and properties that need special formatting. 1275 switch (Current.Next->Tok.getObjCKeywordID()) { 1276 case tok::objc_interface: 1277 case tok::objc_implementation: 1278 case tok::objc_protocol: 1279 Current.Type = TT_ObjCDecl; 1280 break; 1281 case tok::objc_property: 1282 Current.Type = TT_ObjCProperty; 1283 break; 1284 default: 1285 break; 1286 } 1287 } else if (Current.is(tok::period)) { 1288 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 1289 if (PreviousNoComment && 1290 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 1291 Current.Type = TT_DesignatedInitializerPeriod; 1292 else if (Style.Language == FormatStyle::LK_Java && Current.Previous && 1293 Current.Previous->isOneOf(TT_JavaAnnotation, 1294 TT_LeadingJavaAnnotation)) { 1295 Current.Type = Current.Previous->Type; 1296 } 1297 } else if (Current.isOneOf(tok::identifier, tok::kw_const) && 1298 Current.Previous && 1299 !Current.Previous->isOneOf(tok::equal, tok::at) && 1300 Line.MightBeFunctionDecl && Contexts.size() == 1) { 1301 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1302 // function declaration have been found. 1303 Current.Type = TT_TrailingAnnotation; 1304 } else if ((Style.Language == FormatStyle::LK_Java || 1305 Style.Language == FormatStyle::LK_JavaScript) && 1306 Current.Previous) { 1307 if (Current.Previous->is(tok::at) && 1308 Current.isNot(Keywords.kw_interface)) { 1309 const FormatToken &AtToken = *Current.Previous; 1310 const FormatToken *Previous = AtToken.getPreviousNonComment(); 1311 if (!Previous || Previous->is(TT_LeadingJavaAnnotation)) 1312 Current.Type = TT_LeadingJavaAnnotation; 1313 else 1314 Current.Type = TT_JavaAnnotation; 1315 } else if (Current.Previous->is(tok::period) && 1316 Current.Previous->isOneOf(TT_JavaAnnotation, 1317 TT_LeadingJavaAnnotation)) { 1318 Current.Type = Current.Previous->Type; 1319 } 1320 } 1321 } 1322 1323 /// \brief Take a guess at whether \p Tok starts a name of a function or 1324 /// variable declaration. 1325 /// 1326 /// This is a heuristic based on whether \p Tok is an identifier following 1327 /// something that is likely a type. 1328 bool isStartOfName(const FormatToken &Tok) { 1329 if (Tok.isNot(tok::identifier) || !Tok.Previous) 1330 return false; 1331 1332 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof, 1333 Keywords.kw_as)) 1334 return false; 1335 if (Style.Language == FormatStyle::LK_JavaScript && 1336 Tok.Previous->is(Keywords.kw_in)) 1337 return false; 1338 1339 // Skip "const" as it does not have an influence on whether this is a name. 1340 FormatToken *PreviousNotConst = Tok.getPreviousNonComment(); 1341 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 1342 PreviousNotConst = PreviousNotConst->getPreviousNonComment(); 1343 1344 if (!PreviousNotConst) 1345 return false; 1346 1347 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 1348 PreviousNotConst->Previous && 1349 PreviousNotConst->Previous->is(tok::hash); 1350 1351 if (PreviousNotConst->is(TT_TemplateCloser)) 1352 return PreviousNotConst && PreviousNotConst->MatchingParen && 1353 PreviousNotConst->MatchingParen->Previous && 1354 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) && 1355 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 1356 1357 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen && 1358 PreviousNotConst->MatchingParen->Previous && 1359 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype)) 1360 return true; 1361 1362 return (!IsPPKeyword && 1363 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) || 1364 PreviousNotConst->is(TT_PointerOrReference) || 1365 PreviousNotConst->isSimpleTypeSpecifier(); 1366 } 1367 1368 /// \brief Determine whether ')' is ending a cast. 1369 bool rParenEndsCast(const FormatToken &Tok) { 1370 // C-style casts are only used in C++ and Java. 1371 if (!Style.isCpp() && Style.Language != FormatStyle::LK_Java) 1372 return false; 1373 1374 // Empty parens aren't casts and there are no casts at the end of the line. 1375 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen) 1376 return false; 1377 1378 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 1379 if (LeftOfParens) { 1380 // If there is a closing parenthesis left of the current parentheses, 1381 // look past it as these might be chained casts. 1382 if (LeftOfParens->is(tok::r_paren)) { 1383 if (!LeftOfParens->MatchingParen || 1384 !LeftOfParens->MatchingParen->Previous) 1385 return false; 1386 LeftOfParens = LeftOfParens->MatchingParen->Previous; 1387 } 1388 1389 // If there is an identifier (or with a few exceptions a keyword) right 1390 // before the parentheses, this is unlikely to be a cast. 1391 if (LeftOfParens->Tok.getIdentifierInfo() && 1392 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case, 1393 tok::kw_delete)) 1394 return false; 1395 1396 // Certain other tokens right before the parentheses are also signals that 1397 // this cannot be a cast. 1398 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator, 1399 TT_TemplateCloser, tok::ellipsis)) 1400 return false; 1401 } 1402 1403 if (Tok.Next->is(tok::question)) 1404 return false; 1405 1406 // As Java has no function types, a "(" after the ")" likely means that this 1407 // is a cast. 1408 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren)) 1409 return true; 1410 1411 // If a (non-string) literal follows, this is likely a cast. 1412 if (Tok.Next->isNot(tok::string_literal) && 1413 (Tok.Next->Tok.isLiteral() || 1414 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 1415 return true; 1416 1417 // Heuristically try to determine whether the parentheses contain a type. 1418 bool ParensAreType = 1419 !Tok.Previous || 1420 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) || 1421 Tok.Previous->isSimpleTypeSpecifier(); 1422 bool ParensCouldEndDecl = 1423 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater); 1424 if (ParensAreType && !ParensCouldEndDecl) 1425 return true; 1426 1427 // At this point, we heuristically assume that there are no casts at the 1428 // start of the line. We assume that we have found most cases where there 1429 // are by the logic above, e.g. "(void)x;". 1430 if (!LeftOfParens) 1431 return false; 1432 1433 // Certain token types inside the parentheses mean that this can't be a 1434 // cast. 1435 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok; 1436 Token = Token->Next) 1437 if (Token->is(TT_BinaryOperator)) 1438 return false; 1439 1440 // If the following token is an identifier or 'this', this is a cast. All 1441 // cases where this can be something else are handled above. 1442 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this)) 1443 return true; 1444 1445 if (!Tok.Next->Next) 1446 return false; 1447 1448 // If the next token after the parenthesis is a unary operator, assume 1449 // that this is cast, unless there are unexpected tokens inside the 1450 // parenthesis. 1451 bool NextIsUnary = 1452 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star); 1453 if (!NextIsUnary || Tok.Next->is(tok::plus) || 1454 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) 1455 return false; 1456 // Search for unexpected tokens. 1457 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen; 1458 Prev = Prev->Previous) { 1459 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon)) 1460 return false; 1461 } 1462 return true; 1463 } 1464 1465 /// \brief Return the type of the given token assuming it is * or &. 1466 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 1467 bool InTemplateArgument) { 1468 if (Style.Language == FormatStyle::LK_JavaScript) 1469 return TT_BinaryOperator; 1470 1471 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1472 if (!PrevToken) 1473 return TT_UnaryOperator; 1474 1475 const FormatToken *NextToken = Tok.getNextNonComment(); 1476 if (!NextToken || 1477 NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_const) || 1478 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) 1479 return TT_PointerOrReference; 1480 1481 if (PrevToken->is(tok::coloncolon)) 1482 return TT_PointerOrReference; 1483 1484 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 1485 tok::comma, tok::semi, tok::kw_return, tok::colon, 1486 tok::equal, tok::kw_delete, tok::kw_sizeof, 1487 tok::kw_throw) || 1488 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr, 1489 TT_UnaryOperator, TT_CastRParen)) 1490 return TT_UnaryOperator; 1491 1492 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare)) 1493 return TT_PointerOrReference; 1494 if (NextToken->is(tok::kw_operator) && !IsExpression) 1495 return TT_PointerOrReference; 1496 if (NextToken->isOneOf(tok::comma, tok::semi)) 1497 return TT_PointerOrReference; 1498 1499 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen) { 1500 FormatToken *TokenBeforeMatchingParen = 1501 PrevToken->MatchingParen->getPreviousNonComment(); 1502 if (TokenBeforeMatchingParen && 1503 TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype)) 1504 return TT_PointerOrReference; 1505 } 1506 1507 if (PrevToken->Tok.isLiteral() || 1508 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 1509 tok::kw_false, tok::r_brace) || 1510 NextToken->Tok.isLiteral() || 1511 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 1512 NextToken->isUnaryOperator() || 1513 // If we know we're in a template argument, there are no named 1514 // declarations. Thus, having an identifier on the right-hand side 1515 // indicates a binary operator. 1516 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 1517 return TT_BinaryOperator; 1518 1519 // "&&(" is quite unlikely to be two successive unary "&". 1520 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren)) 1521 return TT_BinaryOperator; 1522 1523 // This catches some cases where evaluation order is used as control flow: 1524 // aaa && aaa->f(); 1525 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 1526 if (NextNextToken && NextNextToken->is(tok::arrow)) 1527 return TT_BinaryOperator; 1528 1529 // It is very unlikely that we are going to find a pointer or reference type 1530 // definition on the RHS of an assignment. 1531 if (IsExpression && !Contexts.back().CaretFound) 1532 return TT_BinaryOperator; 1533 1534 return TT_PointerOrReference; 1535 } 1536 1537 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 1538 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1539 if (!PrevToken) 1540 return TT_UnaryOperator; 1541 1542 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator)) 1543 // This must be a sequence of leading unary operators. 1544 return TT_UnaryOperator; 1545 1546 // Use heuristics to recognize unary operators. 1547 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 1548 tok::question, tok::colon, tok::kw_return, 1549 tok::kw_case, tok::at, tok::l_brace)) 1550 return TT_UnaryOperator; 1551 1552 // There can't be two consecutive binary operators. 1553 if (PrevToken->is(TT_BinaryOperator)) 1554 return TT_UnaryOperator; 1555 1556 // Fall back to marking the token as binary operator. 1557 return TT_BinaryOperator; 1558 } 1559 1560 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements. 1561 TokenType determineIncrementUsage(const FormatToken &Tok) { 1562 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1563 if (!PrevToken || PrevToken->is(TT_CastRParen)) 1564 return TT_UnaryOperator; 1565 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 1566 return TT_TrailingUnaryOperator; 1567 1568 return TT_UnaryOperator; 1569 } 1570 1571 SmallVector<Context, 8> Contexts; 1572 1573 const FormatStyle &Style; 1574 AnnotatedLine &Line; 1575 FormatToken *CurrentToken; 1576 bool AutoFound; 1577 const AdditionalKeywords &Keywords; 1578 1579 // Set of "<" tokens that do not open a template parameter list. If parseAngle 1580 // determines that a specific token can't be a template opener, it will make 1581 // same decision irrespective of the decisions for tokens leading up to it. 1582 // Store this information to prevent this from causing exponential runtime. 1583 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess; 1584 }; 1585 1586 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1; 1587 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 1588 1589 /// \brief Parses binary expressions by inserting fake parenthesis based on 1590 /// operator precedence. 1591 class ExpressionParser { 1592 public: 1593 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords, 1594 AnnotatedLine &Line) 1595 : Style(Style), Keywords(Keywords), Current(Line.First) {} 1596 1597 /// \brief Parse expressions with the given operator precedence. 1598 void parse(int Precedence = 0) { 1599 // Skip 'return' and ObjC selector colons as they are not part of a binary 1600 // expression. 1601 while (Current && (Current->is(tok::kw_return) || 1602 (Current->is(tok::colon) && 1603 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) 1604 next(); 1605 1606 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1607 return; 1608 1609 // Conditional expressions need to be parsed separately for proper nesting. 1610 if (Precedence == prec::Conditional) { 1611 parseConditionalExpr(); 1612 return; 1613 } 1614 1615 // Parse unary operators, which all have a higher precedence than binary 1616 // operators. 1617 if (Precedence == PrecedenceUnaryOperator) { 1618 parseUnaryOperator(); 1619 return; 1620 } 1621 1622 FormatToken *Start = Current; 1623 FormatToken *LatestOperator = nullptr; 1624 unsigned OperatorIndex = 0; 1625 1626 while (Current) { 1627 // Consume operators with higher precedence. 1628 parse(Precedence + 1); 1629 1630 int CurrentPrecedence = getCurrentPrecedence(); 1631 1632 if (Current && Current->is(TT_SelectorName) && 1633 Precedence == CurrentPrecedence) { 1634 if (LatestOperator) 1635 addFakeParenthesis(Start, prec::Level(Precedence)); 1636 Start = Current; 1637 } 1638 1639 // At the end of the line or when an operator with higher precedence is 1640 // found, insert fake parenthesis and return. 1641 if (!Current || 1642 (Current->closesScope() && 1643 (Current->MatchingParen || Current->is(TT_TemplateString))) || 1644 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) || 1645 (CurrentPrecedence == prec::Conditional && 1646 Precedence == prec::Assignment && Current->is(tok::colon))) { 1647 break; 1648 } 1649 1650 // Consume scopes: (), [], <> and {} 1651 if (Current->opensScope()) { 1652 // In fragment of a JavaScript template string can look like '}..${' and 1653 // thus close a scope and open a new one at the same time. 1654 while (Current && (!Current->closesScope() || Current->opensScope())) { 1655 next(); 1656 parse(); 1657 } 1658 next(); 1659 } else { 1660 // Operator found. 1661 if (CurrentPrecedence == Precedence) { 1662 if (LatestOperator) 1663 LatestOperator->NextOperator = Current; 1664 LatestOperator = Current; 1665 Current->OperatorIndex = OperatorIndex; 1666 ++OperatorIndex; 1667 } 1668 next(/*SkipPastLeadingComments=*/Precedence > 0); 1669 } 1670 } 1671 1672 if (LatestOperator && (Current || Precedence > 0)) { 1673 // LatestOperator->LastOperator = true; 1674 if (Precedence == PrecedenceArrowAndPeriod) { 1675 // Call expressions don't have a binary operator precedence. 1676 addFakeParenthesis(Start, prec::Unknown); 1677 } else { 1678 addFakeParenthesis(Start, prec::Level(Precedence)); 1679 } 1680 } 1681 } 1682 1683 private: 1684 /// \brief Gets the precedence (+1) of the given token for binary operators 1685 /// and other tokens that we treat like binary operators. 1686 int getCurrentPrecedence() { 1687 if (Current) { 1688 const FormatToken *NextNonComment = Current->getNextNonComment(); 1689 if (Current->is(TT_ConditionalExpr)) 1690 return prec::Conditional; 1691 if (NextNonComment && Current->is(TT_SelectorName) && 1692 (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) || 1693 ((Style.Language == FormatStyle::LK_Proto || 1694 Style.Language == FormatStyle::LK_TextProto) && 1695 NextNonComment->is(tok::less)))) 1696 return prec::Assignment; 1697 if (Current->is(TT_JsComputedPropertyName)) 1698 return prec::Assignment; 1699 if (Current->is(TT_LambdaArrow)) 1700 return prec::Comma; 1701 if (Current->is(TT_JsFatArrow)) 1702 return prec::Assignment; 1703 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) || 1704 (Current->is(tok::comment) && NextNonComment && 1705 NextNonComment->is(TT_SelectorName))) 1706 return 0; 1707 if (Current->is(TT_RangeBasedForLoopColon)) 1708 return prec::Comma; 1709 if ((Style.Language == FormatStyle::LK_Java || 1710 Style.Language == FormatStyle::LK_JavaScript) && 1711 Current->is(Keywords.kw_instanceof)) 1712 return prec::Relational; 1713 if (Style.Language == FormatStyle::LK_JavaScript && 1714 Current->isOneOf(Keywords.kw_in, Keywords.kw_as)) 1715 return prec::Relational; 1716 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma)) 1717 return Current->getPrecedence(); 1718 if (Current->isOneOf(tok::period, tok::arrow)) 1719 return PrecedenceArrowAndPeriod; 1720 if ((Style.Language == FormatStyle::LK_Java || 1721 Style.Language == FormatStyle::LK_JavaScript) && 1722 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements, 1723 Keywords.kw_throws)) 1724 return 0; 1725 } 1726 return -1; 1727 } 1728 1729 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1730 Start->FakeLParens.push_back(Precedence); 1731 if (Precedence > prec::Unknown) 1732 Start->StartsBinaryExpression = true; 1733 if (Current) { 1734 FormatToken *Previous = Current->Previous; 1735 while (Previous->is(tok::comment) && Previous->Previous) 1736 Previous = Previous->Previous; 1737 ++Previous->FakeRParens; 1738 if (Precedence > prec::Unknown) 1739 Previous->EndsBinaryExpression = true; 1740 } 1741 } 1742 1743 /// \brief Parse unary operator expressions and surround them with fake 1744 /// parentheses if appropriate. 1745 void parseUnaryOperator() { 1746 llvm::SmallVector<FormatToken *, 2> Tokens; 1747 while (Current && Current->is(TT_UnaryOperator)) { 1748 Tokens.push_back(Current); 1749 next(); 1750 } 1751 parse(PrecedenceArrowAndPeriod); 1752 for (FormatToken *Token : llvm::reverse(Tokens)) 1753 // The actual precedence doesn't matter. 1754 addFakeParenthesis(Token, prec::Unknown); 1755 } 1756 1757 void parseConditionalExpr() { 1758 while (Current && Current->isTrailingComment()) { 1759 next(); 1760 } 1761 FormatToken *Start = Current; 1762 parse(prec::LogicalOr); 1763 if (!Current || !Current->is(tok::question)) 1764 return; 1765 next(); 1766 parse(prec::Assignment); 1767 if (!Current || Current->isNot(TT_ConditionalExpr)) 1768 return; 1769 next(); 1770 parse(prec::Assignment); 1771 addFakeParenthesis(Start, prec::Conditional); 1772 } 1773 1774 void next(bool SkipPastLeadingComments = true) { 1775 if (Current) 1776 Current = Current->Next; 1777 while (Current && 1778 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 1779 Current->isTrailingComment()) 1780 Current = Current->Next; 1781 } 1782 1783 const FormatStyle &Style; 1784 const AdditionalKeywords &Keywords; 1785 FormatToken *Current; 1786 }; 1787 1788 } // end anonymous namespace 1789 1790 void TokenAnnotator::setCommentLineLevels( 1791 SmallVectorImpl<AnnotatedLine *> &Lines) { 1792 const AnnotatedLine *NextNonCommentLine = nullptr; 1793 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 1794 E = Lines.rend(); 1795 I != E; ++I) { 1796 bool CommentLine = true; 1797 for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) { 1798 if (!Tok->is(tok::comment)) { 1799 CommentLine = false; 1800 break; 1801 } 1802 } 1803 1804 // If the comment is currently aligned with the line immediately following 1805 // it, that's probably intentional and we should keep it. 1806 if (NextNonCommentLine && CommentLine && 1807 NextNonCommentLine->First->NewlinesBefore <= 1 && 1808 NextNonCommentLine->First->OriginalColumn == 1809 (*I)->First->OriginalColumn) { 1810 // Align comments for preprocessor lines with the # in column 0. 1811 // Otherwise, align with the next line. 1812 (*I)->Level = (NextNonCommentLine->Type == LT_PreprocessorDirective || 1813 NextNonCommentLine->Type == LT_ImportStatement) 1814 ? 0 1815 : NextNonCommentLine->Level; 1816 } else { 1817 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 1818 } 1819 1820 setCommentLineLevels((*I)->Children); 1821 } 1822 } 1823 1824 static unsigned maxNestingDepth(const AnnotatedLine &Line) { 1825 unsigned Result = 0; 1826 for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next) 1827 Result = std::max(Result, Tok->NestingLevel); 1828 return Result; 1829 } 1830 1831 void TokenAnnotator::annotate(AnnotatedLine &Line) { 1832 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1833 E = Line.Children.end(); 1834 I != E; ++I) { 1835 annotate(**I); 1836 } 1837 AnnotatingParser Parser(Style, Line, Keywords); 1838 Line.Type = Parser.parseLine(); 1839 1840 // With very deep nesting, ExpressionParser uses lots of stack and the 1841 // formatting algorithm is very slow. We're not going to do a good job here 1842 // anyway - it's probably generated code being formatted by mistake. 1843 // Just skip the whole line. 1844 if (maxNestingDepth(Line) > 50) 1845 Line.Type = LT_Invalid; 1846 1847 if (Line.Type == LT_Invalid) 1848 return; 1849 1850 ExpressionParser ExprParser(Style, Keywords, Line); 1851 ExprParser.parse(); 1852 1853 if (Line.startsWith(TT_ObjCMethodSpecifier)) 1854 Line.Type = LT_ObjCMethodDecl; 1855 else if (Line.startsWith(TT_ObjCDecl)) 1856 Line.Type = LT_ObjCDecl; 1857 else if (Line.startsWith(TT_ObjCProperty)) 1858 Line.Type = LT_ObjCProperty; 1859 1860 Line.First->SpacesRequiredBefore = 1; 1861 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 1862 } 1863 1864 // This function heuristically determines whether 'Current' starts the name of a 1865 // function declaration. 1866 static bool isFunctionDeclarationName(const FormatToken &Current, 1867 const AnnotatedLine &Line) { 1868 auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * { 1869 for (; Next; Next = Next->Next) { 1870 if (Next->is(TT_OverloadedOperatorLParen)) 1871 return Next; 1872 if (Next->is(TT_OverloadedOperator)) 1873 continue; 1874 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) { 1875 // For 'new[]' and 'delete[]'. 1876 if (Next->Next && Next->Next->is(tok::l_square) && Next->Next->Next && 1877 Next->Next->Next->is(tok::r_square)) 1878 Next = Next->Next->Next; 1879 continue; 1880 } 1881 1882 break; 1883 } 1884 return nullptr; 1885 }; 1886 1887 // Find parentheses of parameter list. 1888 const FormatToken *Next = Current.Next; 1889 if (Current.is(tok::kw_operator)) { 1890 if (Current.Previous && Current.Previous->is(tok::coloncolon)) 1891 return false; 1892 Next = skipOperatorName(Next); 1893 } else { 1894 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0) 1895 return false; 1896 for (; Next; Next = Next->Next) { 1897 if (Next->is(TT_TemplateOpener)) { 1898 Next = Next->MatchingParen; 1899 } else if (Next->is(tok::coloncolon)) { 1900 Next = Next->Next; 1901 if (!Next) 1902 return false; 1903 if (Next->is(tok::kw_operator)) { 1904 Next = skipOperatorName(Next->Next); 1905 break; 1906 } 1907 if (!Next->is(tok::identifier)) 1908 return false; 1909 } else if (Next->is(tok::l_paren)) { 1910 break; 1911 } else { 1912 return false; 1913 } 1914 } 1915 } 1916 1917 // Check whether parameter list can belong to a function declaration. 1918 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen) 1919 return false; 1920 // If the lines ends with "{", this is likely an function definition. 1921 if (Line.Last->is(tok::l_brace)) 1922 return true; 1923 if (Next->Next == Next->MatchingParen) 1924 return true; // Empty parentheses. 1925 // If there is an &/&& after the r_paren, this is likely a function. 1926 if (Next->MatchingParen->Next && 1927 Next->MatchingParen->Next->is(TT_PointerOrReference)) 1928 return true; 1929 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen; 1930 Tok = Tok->Next) { 1931 if (Tok->is(tok::l_paren) && Tok->MatchingParen) { 1932 Tok = Tok->MatchingParen; 1933 continue; 1934 } 1935 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 1936 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis)) 1937 return true; 1938 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) || 1939 Tok->Tok.isLiteral()) 1940 return false; 1941 } 1942 return false; 1943 } 1944 1945 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const { 1946 assert(Line.MightBeFunctionDecl); 1947 1948 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel || 1949 Style.AlwaysBreakAfterReturnType == 1950 FormatStyle::RTBS_TopLevelDefinitions) && 1951 Line.Level > 0) 1952 return false; 1953 1954 switch (Style.AlwaysBreakAfterReturnType) { 1955 case FormatStyle::RTBS_None: 1956 return false; 1957 case FormatStyle::RTBS_All: 1958 case FormatStyle::RTBS_TopLevel: 1959 return true; 1960 case FormatStyle::RTBS_AllDefinitions: 1961 case FormatStyle::RTBS_TopLevelDefinitions: 1962 return Line.mightBeFunctionDefinition(); 1963 } 1964 1965 return false; 1966 } 1967 1968 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 1969 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1970 E = Line.Children.end(); 1971 I != E; ++I) { 1972 calculateFormattingInformation(**I); 1973 } 1974 1975 Line.First->TotalLength = 1976 Line.First->IsMultiline ? Style.ColumnLimit 1977 : Line.FirstStartColumn + Line.First->ColumnWidth; 1978 FormatToken *Current = Line.First->Next; 1979 bool InFunctionDecl = Line.MightBeFunctionDecl; 1980 while (Current) { 1981 if (isFunctionDeclarationName(*Current, Line)) 1982 Current->Type = TT_FunctionDeclarationName; 1983 if (Current->is(TT_LineComment)) { 1984 if (Current->Previous->BlockKind == BK_BracedInit && 1985 Current->Previous->opensScope()) 1986 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1; 1987 else 1988 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 1989 1990 // If we find a trailing comment, iterate backwards to determine whether 1991 // it seems to relate to a specific parameter. If so, break before that 1992 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 1993 // to the previous line in: 1994 // SomeFunction(a, 1995 // b, // comment 1996 // c); 1997 if (!Current->HasUnescapedNewline) { 1998 for (FormatToken *Parameter = Current->Previous; Parameter; 1999 Parameter = Parameter->Previous) { 2000 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 2001 break; 2002 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 2003 if (!Parameter->Previous->is(TT_CtorInitializerComma) && 2004 Parameter->HasUnescapedNewline) 2005 Parameter->MustBreakBefore = true; 2006 break; 2007 } 2008 } 2009 } 2010 } else if (Current->SpacesRequiredBefore == 0 && 2011 spaceRequiredBefore(Line, *Current)) { 2012 Current->SpacesRequiredBefore = 1; 2013 } 2014 2015 Current->MustBreakBefore = 2016 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 2017 2018 if (!Current->MustBreakBefore && InFunctionDecl && 2019 Current->is(TT_FunctionDeclarationName)) 2020 Current->MustBreakBefore = mustBreakForReturnType(Line); 2021 2022 Current->CanBreakBefore = 2023 Current->MustBreakBefore || canBreakBefore(Line, *Current); 2024 unsigned ChildSize = 0; 2025 if (Current->Previous->Children.size() == 1) { 2026 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 2027 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 2028 : LastOfChild.TotalLength + 1; 2029 } 2030 const FormatToken *Prev = Current->Previous; 2031 if (Current->MustBreakBefore || Prev->Children.size() > 1 || 2032 (Prev->Children.size() == 1 && 2033 Prev->Children[0]->First->MustBreakBefore) || 2034 Current->IsMultiline) 2035 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit; 2036 else 2037 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth + 2038 ChildSize + Current->SpacesRequiredBefore; 2039 2040 if (Current->is(TT_CtorInitializerColon)) 2041 InFunctionDecl = false; 2042 2043 // FIXME: Only calculate this if CanBreakBefore is true once static 2044 // initializers etc. are sorted out. 2045 // FIXME: Move magic numbers to a better place. 2046 Current->SplitPenalty = 20 * Current->BindingStrength + 2047 splitPenalty(Line, *Current, InFunctionDecl); 2048 2049 Current = Current->Next; 2050 } 2051 2052 calculateUnbreakableTailLengths(Line); 2053 unsigned IndentLevel = Line.Level; 2054 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 2055 if (Current->Role) 2056 Current->Role->precomputeFormattingInfos(Current); 2057 if (Current->MatchingParen && 2058 Current->MatchingParen->opensBlockOrBlockTypeList(Style)) { 2059 assert(IndentLevel > 0); 2060 --IndentLevel; 2061 } 2062 Current->IndentLevel = IndentLevel; 2063 if (Current->opensBlockOrBlockTypeList(Style)) 2064 ++IndentLevel; 2065 } 2066 2067 DEBUG({ printDebugInfo(Line); }); 2068 } 2069 2070 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 2071 unsigned UnbreakableTailLength = 0; 2072 FormatToken *Current = Line.Last; 2073 while (Current) { 2074 Current->UnbreakableTailLength = UnbreakableTailLength; 2075 if (Current->CanBreakBefore || 2076 Current->isOneOf(tok::comment, tok::string_literal)) { 2077 UnbreakableTailLength = 0; 2078 } else { 2079 UnbreakableTailLength += 2080 Current->ColumnWidth + Current->SpacesRequiredBefore; 2081 } 2082 Current = Current->Previous; 2083 } 2084 } 2085 2086 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 2087 const FormatToken &Tok, 2088 bool InFunctionDecl) { 2089 const FormatToken &Left = *Tok.Previous; 2090 const FormatToken &Right = Tok; 2091 2092 if (Left.is(tok::semi)) 2093 return 0; 2094 2095 if (Style.Language == FormatStyle::LK_Java) { 2096 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws)) 2097 return 1; 2098 if (Right.is(Keywords.kw_implements)) 2099 return 2; 2100 if (Left.is(tok::comma) && Left.NestingLevel == 0) 2101 return 3; 2102 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2103 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma)) 2104 return 100; 2105 if (Left.is(TT_JsTypeColon)) 2106 return 35; 2107 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2108 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2109 return 100; 2110 // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()". 2111 if (Left.opensScope() && Right.closesScope()) 2112 return 200; 2113 } 2114 2115 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2116 return 1; 2117 if (Right.is(tok::l_square)) { 2118 if (Style.Language == FormatStyle::LK_Proto) 2119 return 1; 2120 if (Left.is(tok::r_square)) 2121 return 200; 2122 // Slightly prefer formatting local lambda definitions like functions. 2123 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal)) 2124 return 35; 2125 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2126 TT_ArrayInitializerLSquare, 2127 TT_DesignatedInitializerLSquare)) 2128 return 500; 2129 } 2130 2131 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2132 Right.is(tok::kw_operator)) { 2133 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 2134 return 3; 2135 if (Left.is(TT_StartOfName)) 2136 return 110; 2137 if (InFunctionDecl && Right.NestingLevel == 0) 2138 return Style.PenaltyReturnTypeOnItsOwnLine; 2139 return 200; 2140 } 2141 if (Right.is(TT_PointerOrReference)) 2142 return 190; 2143 if (Right.is(TT_LambdaArrow)) 2144 return 110; 2145 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 2146 return 160; 2147 if (Left.is(TT_CastRParen)) 2148 return 100; 2149 if (Left.is(tok::coloncolon) || 2150 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 2151 return 500; 2152 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 2153 return 5000; 2154 if (Left.is(tok::comment)) 2155 return 1000; 2156 2157 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon, 2158 TT_CtorInitializerColon)) 2159 return 2; 2160 2161 if (Right.isMemberAccess()) { 2162 // Breaking before the "./->" of a chained call/member access is reasonably 2163 // cheap, as formatting those with one call per line is generally 2164 // desirable. In particular, it should be cheaper to break before the call 2165 // than it is to break inside a call's parameters, which could lead to weird 2166 // "hanging" indents. The exception is the very last "./->" to support this 2167 // frequent pattern: 2168 // 2169 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc( 2170 // dddddddd); 2171 // 2172 // which might otherwise be blown up onto many lines. Here, clang-format 2173 // won't produce "hanging" indents anyway as there is no other trailing 2174 // call. 2175 // 2176 // Also apply higher penalty is not a call as that might lead to a wrapping 2177 // like: 2178 // 2179 // aaaaaaa 2180 // .aaaaaaaaa.bbbbbbbb(cccccccc); 2181 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope() 2182 ? 150 2183 : 35; 2184 } 2185 2186 if (Right.is(TT_TrailingAnnotation) && 2187 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 2188 // Moving trailing annotations to the next line is fine for ObjC method 2189 // declarations. 2190 if (Line.startsWith(TT_ObjCMethodSpecifier)) 2191 return 10; 2192 // Generally, breaking before a trailing annotation is bad unless it is 2193 // function-like. It seems to be especially preferable to keep standard 2194 // annotations (i.e. "const", "final" and "override") on the same line. 2195 // Use a slightly higher penalty after ")" so that annotations like 2196 // "const override" are kept together. 2197 bool is_short_annotation = Right.TokenText.size() < 10; 2198 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 2199 } 2200 2201 // In for-loops, prefer breaking at ',' and ';'. 2202 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal)) 2203 return 4; 2204 2205 // In Objective-C method expressions, prefer breaking before "param:" over 2206 // breaking after it. 2207 if (Right.is(TT_SelectorName)) 2208 return 0; 2209 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr)) 2210 return Line.MightBeFunctionDecl ? 50 : 500; 2211 2212 if (Left.is(tok::l_paren) && InFunctionDecl && 2213 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) 2214 return 100; 2215 if (Left.is(tok::l_paren) && Left.Previous && 2216 (Left.Previous->isOneOf(tok::kw_if, tok::kw_for) || 2217 Left.Previous->endsSequence(tok::kw_constexpr, tok::kw_if))) 2218 return 1000; 2219 if (Left.is(tok::equal) && InFunctionDecl) 2220 return 110; 2221 if (Right.is(tok::r_brace)) 2222 return 1; 2223 if (Left.is(TT_TemplateOpener)) 2224 return 100; 2225 if (Left.opensScope()) { 2226 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign) 2227 return 0; 2228 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 2229 : 19; 2230 } 2231 if (Left.is(TT_JavaAnnotation)) 2232 return 50; 2233 2234 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous && 2235 Left.Previous->isLabelString() && 2236 (Left.NextOperator || Left.OperatorIndex != 0)) 2237 return 50; 2238 if (Right.is(tok::plus) && Left.isLabelString() && 2239 (Right.NextOperator || Right.OperatorIndex != 0)) 2240 return 25; 2241 if (Left.is(tok::comma)) 2242 return 1; 2243 if (Right.is(tok::lessless) && Left.isLabelString() && 2244 (Right.NextOperator || Right.OperatorIndex != 1)) 2245 return 25; 2246 if (Right.is(tok::lessless)) { 2247 // Breaking at a << is really cheap. 2248 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0) 2249 // Slightly prefer to break before the first one in log-like statements. 2250 return 2; 2251 return 1; 2252 } 2253 if (Left.is(TT_ConditionalExpr)) 2254 return prec::Conditional; 2255 prec::Level Level = Left.getPrecedence(); 2256 if (Level == prec::Unknown) 2257 Level = Right.getPrecedence(); 2258 if (Level == prec::Assignment) 2259 return Style.PenaltyBreakAssignment; 2260 if (Level != prec::Unknown) 2261 return Level; 2262 2263 return 3; 2264 } 2265 2266 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 2267 const FormatToken &Left, 2268 const FormatToken &Right) { 2269 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 2270 return true; 2271 if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java) 2272 return true; 2273 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 2274 Left.Tok.getObjCKeywordID() == tok::objc_property) 2275 return true; 2276 if (Right.is(tok::hashhash)) 2277 return Left.is(tok::hash); 2278 if (Left.isOneOf(tok::hashhash, tok::hash)) 2279 return Right.is(tok::hash); 2280 if (Left.is(tok::l_paren) && Right.is(tok::r_paren)) 2281 return Style.SpaceInEmptyParentheses; 2282 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 2283 return (Right.is(TT_CastRParen) || 2284 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen))) 2285 ? Style.SpacesInCStyleCastParentheses 2286 : Style.SpacesInParentheses; 2287 if (Right.isOneOf(tok::semi, tok::comma)) 2288 return false; 2289 if (Right.is(tok::less) && Line.Type == LT_ObjCDecl && 2290 Style.ObjCSpaceBeforeProtocolList) 2291 return true; 2292 if (Right.is(tok::less) && Left.is(tok::kw_template)) 2293 return Style.SpaceAfterTemplateKeyword; 2294 if (Left.isOneOf(tok::exclaim, tok::tilde)) 2295 return false; 2296 if (Left.is(tok::at) && 2297 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 2298 tok::numeric_constant, tok::l_paren, tok::l_brace, 2299 tok::kw_true, tok::kw_false)) 2300 return false; 2301 if (Left.is(tok::colon)) 2302 return !Left.is(TT_ObjCMethodExpr); 2303 if (Left.is(tok::coloncolon)) 2304 return false; 2305 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) { 2306 if (Style.Language == FormatStyle::LK_TextProto || 2307 (Style.Language == FormatStyle::LK_Proto && 2308 (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) { 2309 // Format empty list as `<>`. 2310 if (Left.is(tok::less) && Right.is(tok::greater)) 2311 return false; 2312 return !Style.Cpp11BracedListStyle; 2313 } 2314 return false; 2315 } 2316 if (Right.is(tok::ellipsis)) 2317 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous && 2318 Left.Previous->is(tok::kw_case)); 2319 if (Left.is(tok::l_square) && Right.is(tok::amp)) 2320 return false; 2321 if (Right.is(TT_PointerOrReference)) { 2322 if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) { 2323 if (!Left.MatchingParen) 2324 return true; 2325 FormatToken *TokenBeforeMatchingParen = 2326 Left.MatchingParen->getPreviousNonComment(); 2327 if (!TokenBeforeMatchingParen || 2328 !TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype)) 2329 return true; 2330 } 2331 return (Left.Tok.isLiteral() || 2332 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) && 2333 (Style.PointerAlignment != FormatStyle::PAS_Left || 2334 (Line.IsMultiVariableDeclStmt && 2335 (Left.NestingLevel == 0 || 2336 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for))))))); 2337 } 2338 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) && 2339 (!Left.is(TT_PointerOrReference) || 2340 (Style.PointerAlignment != FormatStyle::PAS_Right && 2341 !Line.IsMultiVariableDeclStmt))) 2342 return true; 2343 if (Left.is(TT_PointerOrReference)) 2344 return Right.Tok.isLiteral() || Right.is(TT_BlockComment) || 2345 (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) && 2346 !Right.is(TT_StartOfName)) || 2347 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) || 2348 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare, 2349 tok::l_paren) && 2350 (Style.PointerAlignment != FormatStyle::PAS_Right && 2351 !Line.IsMultiVariableDeclStmt) && 2352 Left.Previous && 2353 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon)); 2354 if (Right.is(tok::star) && Left.is(tok::l_paren)) 2355 return false; 2356 const auto SpaceRequiredForArrayInitializerLSquare = 2357 [](const FormatToken &LSquareTok, const FormatStyle &Style) { 2358 return Style.SpacesInContainerLiterals || 2359 ((Style.Language == FormatStyle::LK_Proto || 2360 Style.Language == FormatStyle::LK_TextProto) && 2361 !Style.Cpp11BracedListStyle && 2362 LSquareTok.endsSequence(tok::l_square, tok::colon, 2363 TT_SelectorName)); 2364 }; 2365 if (Left.is(tok::l_square)) 2366 return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) && 2367 SpaceRequiredForArrayInitializerLSquare(Left, Style)) || 2368 (Left.isOneOf(TT_ArraySubscriptLSquare, 2369 TT_StructuredBindingLSquare) && 2370 Style.SpacesInSquareBrackets && Right.isNot(tok::r_square)); 2371 if (Right.is(tok::r_square)) 2372 return Right.MatchingParen && 2373 ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) && 2374 SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen, 2375 Style)) || 2376 (Style.SpacesInSquareBrackets && 2377 Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare, 2378 TT_StructuredBindingLSquare))); 2379 if (Right.is(tok::l_square) && 2380 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2381 TT_DesignatedInitializerLSquare, 2382 TT_StructuredBindingLSquare) && 2383 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral)) 2384 return false; 2385 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 2386 return !Left.Children.empty(); // No spaces in "{}". 2387 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 2388 (Right.is(tok::r_brace) && Right.MatchingParen && 2389 Right.MatchingParen->BlockKind != BK_Block)) 2390 return !Style.Cpp11BracedListStyle; 2391 if (Left.is(TT_BlockComment)) 2392 return !Left.TokenText.endswith("=*/"); 2393 if (Right.is(tok::l_paren)) { 2394 if (Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) 2395 return true; 2396 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) || 2397 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 2398 (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while, 2399 tok::kw_switch, tok::kw_case, TT_ForEachMacro, 2400 TT_ObjCForIn) || 2401 Left.endsSequence(tok::kw_constexpr, tok::kw_if) || 2402 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch, 2403 tok::kw_new, tok::kw_delete) && 2404 (!Left.Previous || Left.Previous->isNot(tok::period))))) || 2405 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always && 2406 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() || 2407 Left.is(tok::r_paren)) && 2408 Line.Type != LT_PreprocessorDirective); 2409 } 2410 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 2411 return false; 2412 if (Right.is(TT_UnaryOperator)) 2413 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 2414 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr)); 2415 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 2416 tok::r_paren) || 2417 Left.isSimpleTypeSpecifier()) && 2418 Right.is(tok::l_brace) && Right.getNextNonComment() && 2419 Right.BlockKind != BK_Block) 2420 return false; 2421 if (Left.is(tok::period) || Right.is(tok::period)) 2422 return false; 2423 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 2424 return false; 2425 if (Left.is(TT_TemplateCloser) && Left.MatchingParen && 2426 Left.MatchingParen->Previous && 2427 Left.MatchingParen->Previous->is(tok::period)) 2428 // A.<B<C<...>>>DoSomething(); 2429 return false; 2430 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square)) 2431 return false; 2432 return true; 2433 } 2434 2435 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 2436 const FormatToken &Right) { 2437 const FormatToken &Left = *Right.Previous; 2438 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo()) 2439 return true; // Never ever merge two identifiers. 2440 if (Style.isCpp()) { 2441 if (Left.is(tok::kw_operator)) 2442 return Right.is(tok::coloncolon); 2443 } else if (Style.Language == FormatStyle::LK_Proto || 2444 Style.Language == FormatStyle::LK_TextProto) { 2445 if (Right.is(tok::period) && 2446 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required, 2447 Keywords.kw_repeated, Keywords.kw_extend)) 2448 return true; 2449 if (Right.is(tok::l_paren) && 2450 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) 2451 return true; 2452 if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName)) 2453 return true; 2454 // Slashes occur in text protocol extension syntax: [type/type] { ... }. 2455 if (Left.is(tok::slash) || Right.is(tok::slash)) 2456 return false; 2457 if (Left.MatchingParen && Left.MatchingParen->is(TT_ProtoExtensionLSquare) && 2458 Right.isOneOf(tok::l_brace, tok::less)) 2459 return !Style.Cpp11BracedListStyle; 2460 // A percent is probably part of a formatting specification, such as %lld. 2461 if (Left.is(tok::percent)) 2462 return false; 2463 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2464 if (Left.is(TT_JsFatArrow)) 2465 return true; 2466 // for await ( ... 2467 if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous && 2468 Left.Previous->is(tok::kw_for)) 2469 return true; 2470 if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) && 2471 Right.MatchingParen) { 2472 const FormatToken *Next = Right.MatchingParen->getNextNonComment(); 2473 // An async arrow function, for example: `x = async () => foo();`, 2474 // as opposed to calling a function called async: `x = async();` 2475 if (Next && Next->is(TT_JsFatArrow)) 2476 return true; 2477 } 2478 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2479 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2480 return false; 2481 // In tagged template literals ("html`bar baz`"), there is no space between 2482 // the tag identifier and the template string. getIdentifierInfo makes sure 2483 // that the identifier is not a pseudo keyword like `yield`, either. 2484 if (Left.is(tok::identifier) && Keywords.IsJavaScriptIdentifier(Left) && 2485 Right.is(TT_TemplateString)) 2486 return false; 2487 if (Right.is(tok::star) && 2488 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) 2489 return false; 2490 if (Right.isOneOf(tok::l_brace, tok::l_square) && 2491 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield, 2492 Keywords.kw_extends, Keywords.kw_implements)) 2493 return true; 2494 if (Right.is(tok::l_paren)) { 2495 // JS methods can use some keywords as names (e.g. `delete()`). 2496 if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo()) 2497 return false; 2498 // Valid JS method names can include keywords, e.g. `foo.delete()` or 2499 // `bar.instanceof()`. Recognize call positions by preceding period. 2500 if (Left.Previous && Left.Previous->is(tok::period) && 2501 Left.Tok.getIdentifierInfo()) 2502 return false; 2503 // Additional unary JavaScript operators that need a space after. 2504 if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof, 2505 tok::kw_void)) 2506 return true; 2507 } 2508 if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in, 2509 tok::kw_const) || 2510 // "of" is only a keyword if it appears after another identifier 2511 // (e.g. as "const x of y" in a for loop), or after a destructuring 2512 // operation (const [x, y] of z, const {a, b} of c). 2513 (Left.is(Keywords.kw_of) && Left.Previous && 2514 (Left.Previous->Tok.is(tok::identifier) || 2515 Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) && 2516 (!Left.Previous || !Left.Previous->is(tok::period))) 2517 return true; 2518 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous && 2519 Left.Previous->is(tok::period) && Right.is(tok::l_paren)) 2520 return false; 2521 if (Left.is(Keywords.kw_as) && 2522 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren)) 2523 return true; 2524 if (Left.is(tok::kw_default) && Left.Previous && 2525 Left.Previous->is(tok::kw_export)) 2526 return true; 2527 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace)) 2528 return true; 2529 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion)) 2530 return false; 2531 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator)) 2532 return false; 2533 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) && 2534 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) 2535 return false; 2536 if (Left.is(tok::ellipsis)) 2537 return false; 2538 if (Left.is(TT_TemplateCloser) && 2539 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square, 2540 Keywords.kw_implements, Keywords.kw_extends)) 2541 // Type assertions ('<type>expr') are not followed by whitespace. Other 2542 // locations that should have whitespace following are identified by the 2543 // above set of follower tokens. 2544 return false; 2545 if (Right.is(TT_JsNonNullAssertion)) 2546 return false; 2547 if (Left.is(TT_JsNonNullAssertion) && 2548 Right.isOneOf(Keywords.kw_as, Keywords.kw_in)) 2549 return true; // "x! as string", "x! in y" 2550 } else if (Style.Language == FormatStyle::LK_Java) { 2551 if (Left.is(tok::r_square) && Right.is(tok::l_brace)) 2552 return true; 2553 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) 2554 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never; 2555 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private, 2556 tok::kw_protected) || 2557 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract, 2558 Keywords.kw_native)) && 2559 Right.is(TT_TemplateOpener)) 2560 return true; 2561 } 2562 if (Left.is(TT_ImplicitStringLiteral)) 2563 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2564 if (Line.Type == LT_ObjCMethodDecl) { 2565 if (Left.is(TT_ObjCMethodSpecifier)) 2566 return true; 2567 if (Left.is(tok::r_paren) && Right.is(tok::identifier)) 2568 // Don't space between ')' and <id> 2569 return false; 2570 } 2571 if (Line.Type == LT_ObjCProperty && 2572 (Right.is(tok::equal) || Left.is(tok::equal))) 2573 return false; 2574 2575 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) || 2576 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) 2577 return true; 2578 if (Right.is(TT_OverloadedOperatorLParen)) 2579 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2580 if (Left.is(tok::comma)) 2581 return true; 2582 if (Right.is(tok::comma)) 2583 return false; 2584 if (Right.is(TT_ObjCBlockLParen)) 2585 return true; 2586 if (Right.is(TT_CtorInitializerColon)) 2587 return Style.SpaceBeforeCtorInitializerColon; 2588 if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon) 2589 return false; 2590 if (Right.is(TT_RangeBasedForLoopColon) && 2591 !Style.SpaceBeforeRangeBasedForLoopColon) 2592 return false; 2593 if (Right.is(tok::colon)) { 2594 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) || 2595 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi)) 2596 return false; 2597 if (Right.is(TT_ObjCMethodExpr)) 2598 return false; 2599 if (Left.is(tok::question)) 2600 return false; 2601 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon)) 2602 return false; 2603 if (Right.is(TT_DictLiteral)) 2604 return Style.SpacesInContainerLiterals; 2605 return true; 2606 } 2607 if (Left.is(TT_UnaryOperator)) 2608 return Right.is(TT_BinaryOperator); 2609 2610 // If the next token is a binary operator or a selector name, we have 2611 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly. 2612 if (Left.is(TT_CastRParen)) 2613 return Style.SpaceAfterCStyleCast || 2614 Right.isOneOf(TT_BinaryOperator, TT_SelectorName); 2615 2616 if (Left.is(tok::greater) && Right.is(tok::greater)) { 2617 if (Style.Language == FormatStyle::LK_TextProto || 2618 (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral))) 2619 return !Style.Cpp11BracedListStyle; 2620 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) && 2621 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 2622 } 2623 if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) || 2624 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) || 2625 (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod))) 2626 return false; 2627 if (!Style.SpaceBeforeAssignmentOperators && 2628 Right.getPrecedence() == prec::Assignment) 2629 return false; 2630 if (Right.is(tok::coloncolon) && Left.is(tok::identifier)) 2631 // Generally don't remove existing spaces between an identifier and "::". 2632 // The identifier might actually be a macro name such as ALWAYS_INLINE. If 2633 // this turns out to be too lenient, add analysis of the identifier itself. 2634 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2635 if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment)) 2636 return (Left.is(TT_TemplateOpener) && 2637 Style.Standard == FormatStyle::LS_Cpp03) || 2638 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square, 2639 tok::kw___super, TT_TemplateCloser, 2640 TT_TemplateOpener)); 2641 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser))) 2642 return Style.SpacesInAngles; 2643 // Space before TT_StructuredBindingLSquare. 2644 if (Right.is(TT_StructuredBindingLSquare)) 2645 return !Left.isOneOf(tok::amp, tok::ampamp) || 2646 Style.PointerAlignment != FormatStyle::PAS_Right; 2647 // Space before & or && following a TT_StructuredBindingLSquare. 2648 if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) && 2649 Right.isOneOf(tok::amp, tok::ampamp)) 2650 return Style.PointerAlignment != FormatStyle::PAS_Left; 2651 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) || 2652 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && 2653 !Right.is(tok::r_paren))) 2654 return true; 2655 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) && 2656 Right.isNot(TT_FunctionTypeLParen)) 2657 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2658 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) && 2659 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen)) 2660 return false; 2661 if (Right.is(tok::less) && Left.isNot(tok::l_paren) && 2662 Line.startsWith(tok::hash)) 2663 return true; 2664 if (Right.is(TT_TrailingUnaryOperator)) 2665 return false; 2666 if (Left.is(TT_RegexLiteral)) 2667 return false; 2668 return spaceRequiredBetween(Line, Left, Right); 2669 } 2670 2671 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 2672 static bool isAllmanBrace(const FormatToken &Tok) { 2673 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 2674 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral); 2675 } 2676 2677 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 2678 const FormatToken &Right) { 2679 const FormatToken &Left = *Right.Previous; 2680 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0) 2681 return true; 2682 2683 if (Style.Language == FormatStyle::LK_JavaScript) { 2684 // FIXME: This might apply to other languages and token kinds. 2685 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous && 2686 Left.Previous->is(tok::string_literal)) 2687 return true; 2688 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 && 2689 Left.Previous && Left.Previous->is(tok::equal) && 2690 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export, 2691 tok::kw_const) && 2692 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match 2693 // above. 2694 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) 2695 // Object literals on the top level of a file are treated as "enum-style". 2696 // Each key/value pair is put on a separate line, instead of bin-packing. 2697 return true; 2698 if (Left.is(tok::l_brace) && Line.Level == 0 && 2699 (Line.startsWith(tok::kw_enum) || 2700 Line.startsWith(tok::kw_const, tok::kw_enum) || 2701 Line.startsWith(tok::kw_export, tok::kw_enum) || 2702 Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum))) 2703 // JavaScript top-level enum key/value pairs are put on separate lines 2704 // instead of bin-packing. 2705 return true; 2706 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && 2707 !Left.Children.empty()) 2708 // Support AllowShortFunctionsOnASingleLine for JavaScript. 2709 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 2710 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty || 2711 (Left.NestingLevel == 0 && Line.Level == 0 && 2712 Style.AllowShortFunctionsOnASingleLine & 2713 FormatStyle::SFS_InlineOnly); 2714 } else if (Style.Language == FormatStyle::LK_Java) { 2715 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next && 2716 Right.Next->is(tok::string_literal)) 2717 return true; 2718 } else if (Style.Language == FormatStyle::LK_Cpp || 2719 Style.Language == FormatStyle::LK_ObjC || 2720 Style.Language == FormatStyle::LK_Proto) { 2721 if (Left.isStringLiteral() && Right.isStringLiteral()) 2722 return true; 2723 } 2724 2725 // If the last token before a '}', ']', or ')' is a comma or a trailing 2726 // comment, the intention is to insert a line break after it in order to make 2727 // shuffling around entries easier. Import statements, especially in 2728 // JavaScript, can be an exception to this rule. 2729 if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) { 2730 const FormatToken *BeforeClosingBrace = nullptr; 2731 if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) || 2732 (Style.Language == FormatStyle::LK_JavaScript && 2733 Left.is(tok::l_paren))) && 2734 Left.BlockKind != BK_Block && Left.MatchingParen) 2735 BeforeClosingBrace = Left.MatchingParen->Previous; 2736 else if (Right.MatchingParen && 2737 (Right.MatchingParen->isOneOf(tok::l_brace, 2738 TT_ArrayInitializerLSquare) || 2739 (Style.Language == FormatStyle::LK_JavaScript && 2740 Right.MatchingParen->is(tok::l_paren)))) 2741 BeforeClosingBrace = &Left; 2742 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 2743 BeforeClosingBrace->isTrailingComment())) 2744 return true; 2745 } 2746 2747 if (Right.is(tok::comment)) 2748 return Left.BlockKind != BK_BracedInit && 2749 Left.isNot(TT_CtorInitializerColon) && 2750 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 2751 if (Left.isTrailingComment()) 2752 return true; 2753 if (Right.Previous->IsUnterminatedLiteral) 2754 return true; 2755 if (Right.is(tok::lessless) && Right.Next && 2756 Right.Previous->is(tok::string_literal) && 2757 Right.Next->is(tok::string_literal)) 2758 return true; 2759 if (Right.Previous->ClosesTemplateDeclaration && 2760 Right.Previous->MatchingParen && 2761 Right.Previous->MatchingParen->NestingLevel == 0 && 2762 Style.AlwaysBreakTemplateDeclarations) 2763 return true; 2764 if (Right.is(TT_CtorInitializerComma) && 2765 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 2766 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 2767 return true; 2768 if (Right.is(TT_CtorInitializerColon) && 2769 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 2770 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 2771 return true; 2772 // Break only if we have multiple inheritance. 2773 if (Style.BreakBeforeInheritanceComma && Right.is(TT_InheritanceComma)) 2774 return true; 2775 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) 2776 // Raw string literals are special wrt. line breaks. The author has made a 2777 // deliberate choice and might have aligned the contents of the string 2778 // literal accordingly. Thus, we try keep existing line breaks. 2779 return Right.NewlinesBefore > 0; 2780 if ((Right.Previous->is(tok::l_brace) || 2781 (Right.Previous->is(tok::less) && Right.Previous->Previous && 2782 Right.Previous->Previous->is(tok::equal))) && 2783 Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) { 2784 // Don't put enums or option definitions onto single lines in protocol 2785 // buffers. 2786 return true; 2787 } 2788 if (Right.is(TT_InlineASMBrace)) 2789 return Right.HasUnescapedNewline; 2790 if (isAllmanBrace(Left) || isAllmanBrace(Right)) 2791 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) || 2792 (Line.startsWith(tok::kw_typedef, tok::kw_enum) && 2793 Style.BraceWrapping.AfterEnum) || 2794 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) || 2795 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct); 2796 if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine) 2797 return true; 2798 2799 if ((Style.Language == FormatStyle::LK_Java || 2800 Style.Language == FormatStyle::LK_JavaScript) && 2801 Left.is(TT_LeadingJavaAnnotation) && 2802 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) && 2803 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) 2804 return true; 2805 2806 if (Right.is(TT_ProtoExtensionLSquare)) 2807 return true; 2808 2809 return false; 2810 } 2811 2812 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 2813 const FormatToken &Right) { 2814 const FormatToken &Left = *Right.Previous; 2815 2816 // Language-specific stuff. 2817 if (Style.Language == FormatStyle::LK_Java) { 2818 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2819 Keywords.kw_implements)) 2820 return false; 2821 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2822 Keywords.kw_implements)) 2823 return true; 2824 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2825 const FormatToken *NonComment = Right.getPreviousNonComment(); 2826 if (NonComment && 2827 NonComment->isOneOf( 2828 tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break, 2829 tok::kw_throw, Keywords.kw_interface, Keywords.kw_type, 2830 tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected, 2831 Keywords.kw_readonly, Keywords.kw_abstract, Keywords.kw_get, 2832 Keywords.kw_set, Keywords.kw_async, Keywords.kw_await)) 2833 return false; // Otherwise automatic semicolon insertion would trigger. 2834 if (Right.NestingLevel == 0 && 2835 (Left.Tok.getIdentifierInfo() || 2836 Left.isOneOf(tok::r_square, tok::r_paren)) && 2837 Right.isOneOf(tok::l_square, tok::l_paren)) 2838 return false; // Otherwise automatic semicolon insertion would trigger. 2839 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace)) 2840 return false; 2841 if (Left.is(TT_JsTypeColon)) 2842 return true; 2843 if (Right.NestingLevel == 0 && Right.is(Keywords.kw_is)) 2844 return false; 2845 if (Left.is(Keywords.kw_in)) 2846 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None; 2847 if (Right.is(Keywords.kw_in)) 2848 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None; 2849 if (Right.is(Keywords.kw_as)) 2850 return false; // must not break before as in 'x as type' casts 2851 if (Left.is(Keywords.kw_as)) 2852 return true; 2853 if (Left.is(TT_JsNonNullAssertion)) 2854 return true; 2855 if (Left.is(Keywords.kw_declare) && 2856 Right.isOneOf(Keywords.kw_module, tok::kw_namespace, 2857 Keywords.kw_function, tok::kw_class, tok::kw_enum, 2858 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var, 2859 Keywords.kw_let, tok::kw_const)) 2860 // See grammar for 'declare' statements at: 2861 // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10 2862 return false; 2863 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) && 2864 Right.isOneOf(tok::identifier, tok::string_literal)) 2865 return false; // must not break in "module foo { ...}" 2866 if (Right.is(TT_TemplateString) && Right.closesScope()) 2867 return false; 2868 if (Left.is(TT_TemplateString) && Left.opensScope()) 2869 return true; 2870 } 2871 2872 if (Left.is(tok::at)) 2873 return false; 2874 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 2875 return false; 2876 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation)) 2877 return !Right.is(tok::l_paren); 2878 if (Right.is(TT_PointerOrReference)) 2879 return Line.IsMultiVariableDeclStmt || 2880 (Style.PointerAlignment == FormatStyle::PAS_Right && 2881 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName))); 2882 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2883 Right.is(tok::kw_operator)) 2884 return true; 2885 if (Left.is(TT_PointerOrReference)) 2886 return false; 2887 if (Right.isTrailingComment()) 2888 // We rely on MustBreakBefore being set correctly here as we should not 2889 // change the "binding" behavior of a comment. 2890 // The first comment in a braced lists is always interpreted as belonging to 2891 // the first list element. Otherwise, it should be placed outside of the 2892 // list. 2893 return Left.BlockKind == BK_BracedInit || 2894 (Left.is(TT_CtorInitializerColon) && 2895 Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon); 2896 if (Left.is(tok::question) && Right.is(tok::colon)) 2897 return false; 2898 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question)) 2899 return Style.BreakBeforeTernaryOperators; 2900 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question)) 2901 return !Style.BreakBeforeTernaryOperators; 2902 if (Right.is(TT_InheritanceColon)) 2903 return true; 2904 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) && 2905 Left.isNot(TT_SelectorName)) 2906 return true; 2907 2908 if (Right.is(tok::colon) && 2909 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) 2910 return false; 2911 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) { 2912 if ((Style.Language == FormatStyle::LK_Proto || 2913 Style.Language == FormatStyle::LK_TextProto) && 2914 Right.isStringLiteral()) 2915 return false; 2916 return true; 2917 } 2918 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next && 2919 Right.Next->is(TT_ObjCMethodExpr))) 2920 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls. 2921 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 2922 return true; 2923 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen)) 2924 return true; 2925 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen, 2926 TT_OverloadedOperator)) 2927 return false; 2928 if (Left.is(TT_RangeBasedForLoopColon)) 2929 return true; 2930 if (Right.is(TT_RangeBasedForLoopColon)) 2931 return false; 2932 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener)) 2933 return true; 2934 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) || 2935 Left.is(tok::kw_operator)) 2936 return false; 2937 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) && 2938 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) 2939 return false; 2940 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen)) 2941 return false; 2942 if (Left.is(tok::l_paren) && Left.Previous && 2943 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) 2944 return false; 2945 if (Right.is(TT_ImplicitStringLiteral)) 2946 return false; 2947 2948 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser)) 2949 return false; 2950 if (Right.is(tok::r_square) && Right.MatchingParen && 2951 Right.MatchingParen->is(TT_LambdaLSquare)) 2952 return false; 2953 2954 // We only break before r_brace if there was a corresponding break before 2955 // the l_brace, which is tracked by BreakBeforeClosingBrace. 2956 if (Right.is(tok::r_brace)) 2957 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 2958 2959 // Allow breaking after a trailing annotation, e.g. after a method 2960 // declaration. 2961 if (Left.is(TT_TrailingAnnotation)) 2962 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 2963 tok::less, tok::coloncolon); 2964 2965 if (Right.is(tok::kw___attribute)) 2966 return true; 2967 2968 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 2969 return true; 2970 2971 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2972 return true; 2973 2974 if (Left.is(TT_CtorInitializerColon)) 2975 return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon; 2976 if (Right.is(TT_CtorInitializerColon)) 2977 return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon; 2978 if (Left.is(TT_CtorInitializerComma) && 2979 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 2980 return false; 2981 if (Right.is(TT_CtorInitializerComma) && 2982 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 2983 return true; 2984 if (Left.is(TT_InheritanceComma) && Style.BreakBeforeInheritanceComma) 2985 return false; 2986 if (Right.is(TT_InheritanceComma) && Style.BreakBeforeInheritanceComma) 2987 return true; 2988 if ((Left.is(tok::greater) && Right.is(tok::greater)) || 2989 (Left.is(tok::less) && Right.is(tok::less))) 2990 return false; 2991 if (Right.is(TT_BinaryOperator) && 2992 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 2993 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 2994 Right.getPrecedence() != prec::Assignment)) 2995 return true; 2996 if (Left.is(TT_ArrayInitializerLSquare)) 2997 return true; 2998 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 2999 return true; 3000 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) && 3001 !Left.isOneOf(tok::arrowstar, tok::lessless) && 3002 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 3003 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 3004 Left.getPrecedence() == prec::Assignment)) 3005 return true; 3006 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 3007 tok::kw_class, tok::kw_struct, tok::comment) || 3008 Right.isMemberAccess() || 3009 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless, 3010 tok::colon, tok::l_square, tok::at) || 3011 (Left.is(tok::r_paren) && 3012 Right.isOneOf(tok::identifier, tok::kw_const)) || 3013 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) || 3014 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser)); 3015 } 3016 3017 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 3018 llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n"; 3019 const FormatToken *Tok = Line.First; 3020 while (Tok) { 3021 llvm::errs() << " M=" << Tok->MustBreakBefore 3022 << " C=" << Tok->CanBreakBefore 3023 << " T=" << getTokenTypeName(Tok->Type) 3024 << " S=" << Tok->SpacesRequiredBefore 3025 << " B=" << Tok->BlockParameterCount 3026 << " BK=" << Tok->BlockKind << " P=" << Tok->SplitPenalty 3027 << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength 3028 << " PPK=" << Tok->PackingKind << " FakeLParens="; 3029 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 3030 llvm::errs() << Tok->FakeLParens[i] << "/"; 3031 llvm::errs() << " FakeRParens=" << Tok->FakeRParens; 3032 llvm::errs() << " Text='" << Tok->TokenText << "'\n"; 3033 if (!Tok->Next) 3034 assert(Tok == Line.Last); 3035 Tok = Tok->Next; 3036 } 3037 llvm::errs() << "----\n"; 3038 } 3039 3040 } // namespace format 3041 } // namespace clang 3042