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