1 //===--- TokenAnnotator.cpp - Format C++ code -----------------------------===// 2 // 3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 4 // See https://llvm.org/LICENSE.txt for license information. 5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 6 // 7 //===----------------------------------------------------------------------===// 8 /// 9 /// \file 10 /// This file implements a token annotator, i.e. creates 11 /// \c AnnotatedTokens out of \c FormatTokens with required extra information. 12 /// 13 //===----------------------------------------------------------------------===// 14 15 #include "TokenAnnotator.h" 16 #include "FormatToken.h" 17 #include "clang/Basic/SourceManager.h" 18 #include "clang/Basic/TokenKinds.h" 19 #include "llvm/ADT/SmallPtrSet.h" 20 #include "llvm/Support/Debug.h" 21 22 #define DEBUG_TYPE "format-token-annotator" 23 24 namespace clang { 25 namespace format { 26 27 namespace { 28 29 /// Returns \c true if the token can be used as an identifier in 30 /// an Objective-C \c \@selector, \c false otherwise. 31 /// 32 /// Because getFormattingLangOpts() always lexes source code as 33 /// Objective-C++, C++ keywords like \c new and \c delete are 34 /// lexed as tok::kw_*, not tok::identifier, even for Objective-C. 35 /// 36 /// For Objective-C and Objective-C++, both identifiers and keywords 37 /// are valid inside @selector(...) (or a macro which 38 /// invokes @selector(...)). So, we allow treat any identifier or 39 /// keyword as a potential Objective-C selector component. 40 static bool canBeObjCSelectorComponent(const FormatToken &Tok) { 41 return Tok.Tok.getIdentifierInfo() != nullptr; 42 } 43 44 /// With `Left` being '(', check if we're at either `[...](` or 45 /// `[...]<...>(`, where the [ opens a lambda capture list. 46 static bool isLambdaParameterList(const FormatToken *Left) { 47 // Skip <...> if present. 48 if (Left->Previous && Left->Previous->is(tok::greater) && 49 Left->Previous->MatchingParen && 50 Left->Previous->MatchingParen->is(TT_TemplateOpener)) 51 Left = Left->Previous->MatchingParen; 52 53 // Check for `[...]`. 54 return Left->Previous && Left->Previous->is(tok::r_square) && 55 Left->Previous->MatchingParen && 56 Left->Previous->MatchingParen->is(TT_LambdaLSquare); 57 } 58 59 /// Returns \c true if the token is followed by a boolean condition, \c false 60 /// otherwise. 61 static bool isKeywordWithCondition(const FormatToken &Tok) { 62 return Tok.isOneOf(tok::kw_if, tok::kw_for, tok::kw_while, tok::kw_switch, 63 tok::kw_constexpr, tok::kw_catch); 64 } 65 66 /// A parser that gathers additional information about tokens. 67 /// 68 /// The \c TokenAnnotator tries to match parenthesis and square brakets and 69 /// store a parenthesis levels. It also tries to resolve matching "<" and ">" 70 /// into template parameter lists. 71 class AnnotatingParser { 72 public: 73 AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line, 74 const AdditionalKeywords &Keywords) 75 : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false), 76 Keywords(Keywords) { 77 Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false)); 78 resetTokenMetadata(CurrentToken); 79 } 80 81 private: 82 bool parseAngle() { 83 if (!CurrentToken || !CurrentToken->Previous) 84 return false; 85 if (NonTemplateLess.count(CurrentToken->Previous)) 86 return false; 87 88 const FormatToken &Previous = *CurrentToken->Previous; // The '<'. 89 if (Previous.Previous) { 90 if (Previous.Previous->Tok.isLiteral()) 91 return false; 92 if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 && 93 (!Previous.Previous->MatchingParen || 94 !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen))) 95 return false; 96 } 97 98 FormatToken *Left = CurrentToken->Previous; 99 Left->ParentBracket = Contexts.back().ContextKind; 100 ScopedContextCreator ContextCreator(*this, tok::less, 12); 101 102 // If this angle is in the context of an expression, we need to be more 103 // hesitant to detect it as opening template parameters. 104 bool InExprContext = Contexts.back().IsExpression; 105 106 Contexts.back().IsExpression = false; 107 // If there's a template keyword before the opening angle bracket, this is a 108 // template parameter, not an argument. 109 Contexts.back().InTemplateArgument = 110 Left->Previous && Left->Previous->Tok.isNot(tok::kw_template); 111 112 if (Style.Language == FormatStyle::LK_Java && 113 CurrentToken->is(tok::question)) 114 next(); 115 116 while (CurrentToken) { 117 if (CurrentToken->is(tok::greater)) { 118 // Try to do a better job at looking for ">>" within the condition of 119 // a statement. Conservatively insert spaces between consecutive ">" 120 // tokens to prevent splitting right bitshift operators and potentially 121 // altering program semantics. This check is overly conservative and 122 // will prevent spaces from being inserted in select nested template 123 // parameter cases, but should not alter program semantics. 124 if (CurrentToken->Next && CurrentToken->Next->is(tok::greater) && 125 Left->ParentBracket != tok::less && 126 (isKeywordWithCondition(*Line.First) || 127 CurrentToken->getStartOfNonWhitespace() == 128 CurrentToken->Next->getStartOfNonWhitespace().getLocWithOffset( 129 -1))) 130 return false; 131 Left->MatchingParen = CurrentToken; 132 CurrentToken->MatchingParen = Left; 133 // In TT_Proto, we must distignuish between: 134 // map<key, value> 135 // msg < item: data > 136 // msg: < item: data > 137 // In TT_TextProto, map<key, value> does not occur. 138 if (Style.Language == FormatStyle::LK_TextProto || 139 (Style.Language == FormatStyle::LK_Proto && Left->Previous && 140 Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) 141 CurrentToken->setType(TT_DictLiteral); 142 else 143 CurrentToken->setType(TT_TemplateCloser); 144 next(); 145 return true; 146 } 147 if (CurrentToken->is(tok::question) && 148 Style.Language == FormatStyle::LK_Java) { 149 next(); 150 continue; 151 } 152 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) || 153 (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext && 154 !Style.isCSharp() && Style.Language != FormatStyle::LK_Proto && 155 Style.Language != FormatStyle::LK_TextProto)) 156 return false; 157 // If a && or || is found and interpreted as a binary operator, this set 158 // of angles is likely part of something like "a < b && c > d". If the 159 // angles are inside an expression, the ||/&& might also be a binary 160 // operator that was misinterpreted because we are parsing template 161 // parameters. 162 // FIXME: This is getting out of hand, write a decent parser. 163 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) && 164 CurrentToken->Previous->is(TT_BinaryOperator) && 165 Contexts[Contexts.size() - 2].IsExpression && 166 !Line.startsWith(tok::kw_template)) 167 return false; 168 updateParameterCount(Left, CurrentToken); 169 if (Style.Language == FormatStyle::LK_Proto) { 170 if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) { 171 if (CurrentToken->is(tok::colon) || 172 (CurrentToken->isOneOf(tok::l_brace, tok::less) && 173 Previous->isNot(tok::colon))) 174 Previous->setType(TT_SelectorName); 175 } 176 } 177 if (!consumeToken()) 178 return false; 179 } 180 return false; 181 } 182 183 bool parseUntouchableParens() { 184 while (CurrentToken) { 185 CurrentToken->Finalized = true; 186 switch (CurrentToken->Tok.getKind()) { 187 case tok::l_paren: 188 next(); 189 if (!parseUntouchableParens()) 190 return false; 191 continue; 192 case tok::r_paren: 193 next(); 194 return true; 195 default: 196 // no-op 197 break; 198 } 199 next(); 200 } 201 return false; 202 } 203 204 bool parseParens(bool LookForDecls = false) { 205 if (!CurrentToken) 206 return false; 207 FormatToken *Left = CurrentToken->Previous; 208 assert(Left && "Unknown previous token"); 209 FormatToken *PrevNonComment = Left->getPreviousNonComment(); 210 Left->ParentBracket = Contexts.back().ContextKind; 211 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1); 212 213 // FIXME: This is a bit of a hack. Do better. 214 Contexts.back().ColonIsForRangeExpr = 215 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr; 216 217 if (Left->Previous && Left->Previous->is(TT_UntouchableMacroFunc)) { 218 Left->Finalized = true; 219 return parseUntouchableParens(); 220 } 221 222 bool StartsObjCMethodExpr = false; 223 if (FormatToken *MaybeSel = Left->Previous) { 224 // @selector( starts a selector. 225 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous && 226 MaybeSel->Previous->is(tok::at)) { 227 StartsObjCMethodExpr = true; 228 } 229 } 230 231 if (Left->is(TT_OverloadedOperatorLParen)) { 232 // Find the previous kw_operator token. 233 FormatToken *Prev = Left; 234 while (!Prev->is(tok::kw_operator)) { 235 Prev = Prev->Previous; 236 assert(Prev && "Expect a kw_operator prior to the OperatorLParen!"); 237 } 238 239 // If faced with "a.operator*(argument)" or "a->operator*(argument)", 240 // i.e. the operator is called as a member function, 241 // then the argument must be an expression. 242 bool OperatorCalledAsMemberFunction = 243 Prev->Previous && Prev->Previous->isOneOf(tok::period, tok::arrow); 244 Contexts.back().IsExpression = OperatorCalledAsMemberFunction; 245 } else if (Style.Language == FormatStyle::LK_JavaScript && 246 (Line.startsWith(Keywords.kw_type, tok::identifier) || 247 Line.startsWith(tok::kw_export, Keywords.kw_type, 248 tok::identifier))) { 249 // type X = (...); 250 // export type X = (...); 251 Contexts.back().IsExpression = false; 252 } else if (Left->Previous && 253 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_while, 254 tok::l_paren, tok::comma) || 255 Left->Previous->isIf() || 256 Left->Previous->is(TT_BinaryOperator))) { 257 // static_assert, if and while usually contain expressions. 258 Contexts.back().IsExpression = true; 259 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 260 (Left->Previous->is(Keywords.kw_function) || 261 (Left->Previous->endsSequence(tok::identifier, 262 Keywords.kw_function)))) { 263 // function(...) or function f(...) 264 Contexts.back().IsExpression = false; 265 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 266 Left->Previous->is(TT_JsTypeColon)) { 267 // let x: (SomeType); 268 Contexts.back().IsExpression = false; 269 } else if (isLambdaParameterList(Left)) { 270 // This is a parameter list of a lambda expression. 271 Contexts.back().IsExpression = false; 272 } else if (Line.InPPDirective && 273 (!Left->Previous || !Left->Previous->is(tok::identifier))) { 274 Contexts.back().IsExpression = true; 275 } else if (Contexts[Contexts.size() - 2].CaretFound) { 276 // This is the parameter list of an ObjC block. 277 Contexts.back().IsExpression = false; 278 } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) { 279 // The first argument to a foreach macro is a declaration. 280 Contexts.back().IsForEachMacro = true; 281 Contexts.back().IsExpression = false; 282 } else if (Left->Previous && Left->Previous->MatchingParen && 283 Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) { 284 Contexts.back().IsExpression = false; 285 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) { 286 bool IsForOrCatch = 287 Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch); 288 Contexts.back().IsExpression = !IsForOrCatch; 289 } 290 291 // Infer the role of the l_paren based on the previous token if we haven't 292 // detected one one yet. 293 if (PrevNonComment && Left->is(TT_Unknown)) { 294 if (PrevNonComment->is(tok::kw___attribute)) { 295 Left->setType(TT_AttributeParen); 296 } else if (PrevNonComment->isOneOf(TT_TypenameMacro, tok::kw_decltype, 297 tok::kw_typeof, tok::kw__Atomic, 298 tok::kw___underlying_type)) { 299 Left->setType(TT_TypeDeclarationParen); 300 // decltype() and typeof() usually contain expressions. 301 if (PrevNonComment->isOneOf(tok::kw_decltype, tok::kw_typeof)) 302 Contexts.back().IsExpression = true; 303 } 304 } 305 306 if (StartsObjCMethodExpr) { 307 Contexts.back().ColonIsObjCMethodExpr = true; 308 Left->setType(TT_ObjCMethodExpr); 309 } 310 311 // MightBeFunctionType and ProbablyFunctionType are used for 312 // function pointer and reference types as well as Objective-C 313 // block types: 314 // 315 // void (*FunctionPointer)(void); 316 // void (&FunctionReference)(void); 317 // void (^ObjCBlock)(void); 318 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression; 319 bool ProbablyFunctionType = 320 CurrentToken->isOneOf(tok::star, tok::amp, tok::caret); 321 bool HasMultipleLines = false; 322 bool HasMultipleParametersOnALine = false; 323 bool MightBeObjCForRangeLoop = 324 Left->Previous && Left->Previous->is(tok::kw_for); 325 FormatToken *PossibleObjCForInToken = nullptr; 326 while (CurrentToken) { 327 // LookForDecls is set when "if (" has been seen. Check for 328 // 'identifier' '*' 'identifier' followed by not '=' -- this 329 // '*' has to be a binary operator but determineStarAmpUsage() will 330 // categorize it as an unary operator, so set the right type here. 331 if (LookForDecls && CurrentToken->Next) { 332 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 333 if (Prev) { 334 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 335 FormatToken *Next = CurrentToken->Next; 336 if (PrevPrev && PrevPrev->is(tok::identifier) && 337 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 338 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 339 Prev->setType(TT_BinaryOperator); 340 LookForDecls = false; 341 } 342 } 343 } 344 345 if (CurrentToken->Previous->is(TT_PointerOrReference) && 346 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 347 tok::coloncolon)) 348 ProbablyFunctionType = true; 349 if (CurrentToken->is(tok::comma)) 350 MightBeFunctionType = false; 351 if (CurrentToken->Previous->is(TT_BinaryOperator)) 352 Contexts.back().IsExpression = true; 353 if (CurrentToken->is(tok::r_paren)) { 354 if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next && 355 (CurrentToken->Next->is(tok::l_paren) || 356 (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration))) 357 Left->setType(Left->Next->is(tok::caret) ? TT_ObjCBlockLParen 358 : TT_FunctionTypeLParen); 359 Left->MatchingParen = CurrentToken; 360 CurrentToken->MatchingParen = Left; 361 362 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) && 363 Left->Previous && Left->Previous->is(tok::l_paren)) { 364 // Detect the case where macros are used to generate lambdas or 365 // function bodies, e.g.: 366 // auto my_lambda = MACRO((Type *type, int i) { .. body .. }); 367 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) { 368 if (Tok->is(TT_BinaryOperator) && 369 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) 370 Tok->setType(TT_PointerOrReference); 371 } 372 } 373 374 if (StartsObjCMethodExpr) { 375 CurrentToken->setType(TT_ObjCMethodExpr); 376 if (Contexts.back().FirstObjCSelectorName) { 377 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 378 Contexts.back().LongestObjCSelectorName; 379 } 380 } 381 382 if (Left->is(TT_AttributeParen)) 383 CurrentToken->setType(TT_AttributeParen); 384 if (Left->is(TT_TypeDeclarationParen)) 385 CurrentToken->setType(TT_TypeDeclarationParen); 386 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation)) 387 CurrentToken->setType(TT_JavaAnnotation); 388 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation)) 389 CurrentToken->setType(TT_LeadingJavaAnnotation); 390 if (Left->Previous && Left->Previous->is(TT_AttributeSquare)) 391 CurrentToken->setType(TT_AttributeSquare); 392 393 if (!HasMultipleLines) 394 Left->setPackingKind(PPK_Inconclusive); 395 else if (HasMultipleParametersOnALine) 396 Left->setPackingKind(PPK_BinPacked); 397 else 398 Left->setPackingKind(PPK_OnePerLine); 399 400 next(); 401 return true; 402 } 403 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 404 return false; 405 406 if (CurrentToken->is(tok::l_brace)) 407 Left->setType(TT_Unknown); // Not TT_ObjCBlockLParen 408 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 409 !CurrentToken->Next->HasUnescapedNewline && 410 !CurrentToken->Next->isTrailingComment()) 411 HasMultipleParametersOnALine = true; 412 bool ProbablyFunctionTypeLParen = 413 (CurrentToken->is(tok::l_paren) && CurrentToken->Next && 414 CurrentToken->Next->isOneOf(tok::star, tok::amp, tok::caret)); 415 if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) || 416 CurrentToken->Previous->isSimpleTypeSpecifier()) && 417 !(CurrentToken->is(tok::l_brace) || 418 (CurrentToken->is(tok::l_paren) && !ProbablyFunctionTypeLParen))) 419 Contexts.back().IsExpression = false; 420 if (CurrentToken->isOneOf(tok::semi, tok::colon)) { 421 MightBeObjCForRangeLoop = false; 422 if (PossibleObjCForInToken) { 423 PossibleObjCForInToken->setType(TT_Unknown); 424 PossibleObjCForInToken = nullptr; 425 } 426 } 427 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) { 428 PossibleObjCForInToken = CurrentToken; 429 PossibleObjCForInToken->setType(TT_ObjCForIn); 430 } 431 // When we discover a 'new', we set CanBeExpression to 'false' in order to 432 // parse the type correctly. Reset that after a comma. 433 if (CurrentToken->is(tok::comma)) 434 Contexts.back().CanBeExpression = true; 435 436 FormatToken *Tok = CurrentToken; 437 if (!consumeToken()) 438 return false; 439 updateParameterCount(Left, Tok); 440 if (CurrentToken && CurrentToken->HasUnescapedNewline) 441 HasMultipleLines = true; 442 } 443 return false; 444 } 445 446 bool isCSharpAttributeSpecifier(const FormatToken &Tok) { 447 if (!Style.isCSharp()) 448 return false; 449 450 // `identifier[i]` is not an attribute. 451 if (Tok.Previous && Tok.Previous->is(tok::identifier)) 452 return false; 453 454 // Chains of [] in `identifier[i][j][k]` are not attributes. 455 if (Tok.Previous && Tok.Previous->is(tok::r_square)) { 456 auto *MatchingParen = Tok.Previous->MatchingParen; 457 if (!MatchingParen || MatchingParen->is(TT_ArraySubscriptLSquare)) 458 return false; 459 } 460 461 const FormatToken *AttrTok = Tok.Next; 462 if (!AttrTok) 463 return false; 464 465 // Just an empty declaration e.g. string []. 466 if (AttrTok->is(tok::r_square)) 467 return false; 468 469 // Move along the tokens inbetween the '[' and ']' e.g. [STAThread]. 470 while (AttrTok && AttrTok->isNot(tok::r_square)) { 471 AttrTok = AttrTok->Next; 472 } 473 474 if (!AttrTok) 475 return false; 476 477 // Allow an attribute to be the only content of a file. 478 AttrTok = AttrTok->Next; 479 if (!AttrTok) 480 return true; 481 482 // Limit this to being an access modifier that follows. 483 if (AttrTok->isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected, 484 tok::comment, tok::kw_class, tok::kw_static, 485 tok::l_square, Keywords.kw_internal)) { 486 return true; 487 } 488 489 // incase its a [XXX] retval func(.... 490 if (AttrTok->Next && 491 AttrTok->Next->startsSequence(tok::identifier, tok::l_paren)) 492 return true; 493 494 return false; 495 } 496 497 bool isCpp11AttributeSpecifier(const FormatToken &Tok) { 498 if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square)) 499 return false; 500 // The first square bracket is part of an ObjC array literal 501 if (Tok.Previous && Tok.Previous->is(tok::at)) { 502 return false; 503 } 504 const FormatToken *AttrTok = Tok.Next->Next; 505 if (!AttrTok) 506 return false; 507 // C++17 '[[using ns: foo, bar(baz, blech)]]' 508 // We assume nobody will name an ObjC variable 'using'. 509 if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon)) 510 return true; 511 if (AttrTok->isNot(tok::identifier)) 512 return false; 513 while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) { 514 // ObjC message send. We assume nobody will use : in a C++11 attribute 515 // specifier parameter, although this is technically valid: 516 // [[foo(:)]]. 517 if (AttrTok->is(tok::colon) || 518 AttrTok->startsSequence(tok::identifier, tok::identifier) || 519 AttrTok->startsSequence(tok::r_paren, tok::identifier)) 520 return false; 521 if (AttrTok->is(tok::ellipsis)) 522 return true; 523 AttrTok = AttrTok->Next; 524 } 525 return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square); 526 } 527 528 bool parseSquare() { 529 if (!CurrentToken) 530 return false; 531 532 // A '[' could be an index subscript (after an identifier or after 533 // ')' or ']'), it could be the start of an Objective-C method 534 // expression, it could the start of an Objective-C array literal, 535 // or it could be a C++ attribute specifier [[foo::bar]]. 536 FormatToken *Left = CurrentToken->Previous; 537 Left->ParentBracket = Contexts.back().ContextKind; 538 FormatToken *Parent = Left->getPreviousNonComment(); 539 540 // Cases where '>' is followed by '['. 541 // In C++, this can happen either in array of templates (foo<int>[10]) 542 // or when array is a nested template type (unique_ptr<type1<type2>[]>). 543 bool CppArrayTemplates = 544 Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) && 545 (Contexts.back().CanBeExpression || Contexts.back().IsExpression || 546 Contexts.back().InTemplateArgument); 547 548 bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) || 549 Contexts.back().InCpp11AttributeSpecifier; 550 551 // Treat C# Attributes [STAThread] much like C++ attributes [[...]]. 552 bool IsCSharpAttributeSpecifier = 553 isCSharpAttributeSpecifier(*Left) || 554 Contexts.back().InCSharpAttributeSpecifier; 555 556 bool InsideInlineASM = Line.startsWith(tok::kw_asm); 557 bool IsCppStructuredBinding = Left->isCppStructuredBinding(Style); 558 bool StartsObjCMethodExpr = 559 !IsCppStructuredBinding && !InsideInlineASM && !CppArrayTemplates && 560 Style.isCpp() && !IsCpp11AttributeSpecifier && 561 !IsCSharpAttributeSpecifier && Contexts.back().CanBeExpression && 562 Left->isNot(TT_LambdaLSquare) && 563 !CurrentToken->isOneOf(tok::l_brace, tok::r_square) && 564 (!Parent || 565 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 566 tok::kw_return, tok::kw_throw) || 567 Parent->isUnaryOperator() || 568 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 569 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) || 570 (getBinOpPrecedence(Parent->Tok.getKind(), true, true) > 571 prec::Unknown)); 572 bool ColonFound = false; 573 574 unsigned BindingIncrease = 1; 575 if (IsCppStructuredBinding) { 576 Left->setType(TT_StructuredBindingLSquare); 577 } else if (Left->is(TT_Unknown)) { 578 if (StartsObjCMethodExpr) { 579 Left->setType(TT_ObjCMethodExpr); 580 } else if (InsideInlineASM) { 581 Left->setType(TT_InlineASMSymbolicNameLSquare); 582 } else if (IsCpp11AttributeSpecifier) { 583 Left->setType(TT_AttributeSquare); 584 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent && 585 Contexts.back().ContextKind == tok::l_brace && 586 Parent->isOneOf(tok::l_brace, tok::comma)) { 587 Left->setType(TT_JsComputedPropertyName); 588 } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace && 589 Parent && Parent->isOneOf(tok::l_brace, tok::comma)) { 590 Left->setType(TT_DesignatedInitializerLSquare); 591 } else if (IsCSharpAttributeSpecifier) { 592 Left->setType(TT_AttributeSquare); 593 } else if (CurrentToken->is(tok::r_square) && Parent && 594 Parent->is(TT_TemplateCloser)) { 595 Left->setType(TT_ArraySubscriptLSquare); 596 } else if (Style.Language == FormatStyle::LK_Proto || 597 Style.Language == FormatStyle::LK_TextProto) { 598 // Square braces in LK_Proto can either be message field attributes: 599 // 600 // optional Aaa aaa = 1 [ 601 // (aaa) = aaa 602 // ]; 603 // 604 // extensions 123 [ 605 // (aaa) = aaa 606 // ]; 607 // 608 // or text proto extensions (in options): 609 // 610 // option (Aaa.options) = { 611 // [type.type/type] { 612 // key: value 613 // } 614 // } 615 // 616 // or repeated fields (in options): 617 // 618 // option (Aaa.options) = { 619 // keys: [ 1, 2, 3 ] 620 // } 621 // 622 // In the first and the third case we want to spread the contents inside 623 // the square braces; in the second we want to keep them inline. 624 Left->setType(TT_ArrayInitializerLSquare); 625 if (!Left->endsSequence(tok::l_square, tok::numeric_constant, 626 tok::equal) && 627 !Left->endsSequence(tok::l_square, tok::numeric_constant, 628 tok::identifier) && 629 !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) { 630 Left->setType(TT_ProtoExtensionLSquare); 631 BindingIncrease = 10; 632 } 633 } else if (!CppArrayTemplates && Parent && 634 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at, 635 tok::comma, tok::l_paren, tok::l_square, 636 tok::question, tok::colon, tok::kw_return, 637 // Should only be relevant to JavaScript: 638 tok::kw_default)) { 639 Left->setType(TT_ArrayInitializerLSquare); 640 } else { 641 BindingIncrease = 10; 642 Left->setType(TT_ArraySubscriptLSquare); 643 } 644 } 645 646 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease); 647 Contexts.back().IsExpression = true; 648 if (Style.Language == FormatStyle::LK_JavaScript && Parent && 649 Parent->is(TT_JsTypeColon)) 650 Contexts.back().IsExpression = false; 651 652 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr; 653 Contexts.back().InCpp11AttributeSpecifier = IsCpp11AttributeSpecifier; 654 Contexts.back().InCSharpAttributeSpecifier = IsCSharpAttributeSpecifier; 655 656 while (CurrentToken) { 657 if (CurrentToken->is(tok::r_square)) { 658 if (IsCpp11AttributeSpecifier) 659 CurrentToken->setType(TT_AttributeSquare); 660 if (IsCSharpAttributeSpecifier) 661 CurrentToken->setType(TT_AttributeSquare); 662 else if (((CurrentToken->Next && 663 CurrentToken->Next->is(tok::l_paren)) || 664 (CurrentToken->Previous && 665 CurrentToken->Previous->Previous == Left)) && 666 Left->is(TT_ObjCMethodExpr)) { 667 // An ObjC method call is rarely followed by an open parenthesis. It 668 // also can't be composed of just one token, unless it's a macro that 669 // will be expanded to more tokens. 670 // FIXME: Do we incorrectly label ":" with this? 671 StartsObjCMethodExpr = false; 672 Left->setType(TT_Unknown); 673 } 674 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) { 675 CurrentToken->setType(TT_ObjCMethodExpr); 676 // If we haven't seen a colon yet, make sure the last identifier 677 // before the r_square is tagged as a selector name component. 678 if (!ColonFound && CurrentToken->Previous && 679 CurrentToken->Previous->is(TT_Unknown) && 680 canBeObjCSelectorComponent(*CurrentToken->Previous)) 681 CurrentToken->Previous->setType(TT_SelectorName); 682 // determineStarAmpUsage() thinks that '*' '[' is allocating an 683 // array of pointers, but if '[' starts a selector then '*' is a 684 // binary operator. 685 if (Parent && Parent->is(TT_PointerOrReference)) 686 Parent->setType(TT_BinaryOperator); 687 } 688 // An arrow after an ObjC method expression is not a lambda arrow. 689 if (CurrentToken->getType() == TT_ObjCMethodExpr && 690 CurrentToken->Next && CurrentToken->Next->is(TT_LambdaArrow)) 691 CurrentToken->Next->setType(TT_Unknown); 692 Left->MatchingParen = CurrentToken; 693 CurrentToken->MatchingParen = Left; 694 // FirstObjCSelectorName is set when a colon is found. This does 695 // not work, however, when the method has no parameters. 696 // Here, we set FirstObjCSelectorName when the end of the method call is 697 // reached, in case it was not set already. 698 if (!Contexts.back().FirstObjCSelectorName) { 699 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 700 if (Previous && Previous->is(TT_SelectorName)) { 701 Previous->ObjCSelectorNameParts = 1; 702 Contexts.back().FirstObjCSelectorName = Previous; 703 } 704 } else { 705 Left->ParameterCount = 706 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 707 } 708 if (Contexts.back().FirstObjCSelectorName) { 709 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 710 Contexts.back().LongestObjCSelectorName; 711 if (Left->BlockParameterCount > 1) 712 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0; 713 } 714 next(); 715 return true; 716 } 717 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace)) 718 return false; 719 if (CurrentToken->is(tok::colon)) { 720 if (IsCpp11AttributeSpecifier && 721 CurrentToken->endsSequence(tok::colon, tok::identifier, 722 tok::kw_using)) { 723 // Remember that this is a [[using ns: foo]] C++ attribute, so we 724 // don't add a space before the colon (unlike other colons). 725 CurrentToken->setType(TT_AttributeColon); 726 } else if (Left->isOneOf(TT_ArraySubscriptLSquare, 727 TT_DesignatedInitializerLSquare)) { 728 Left->setType(TT_ObjCMethodExpr); 729 StartsObjCMethodExpr = true; 730 Contexts.back().ColonIsObjCMethodExpr = true; 731 if (Parent && Parent->is(tok::r_paren)) 732 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 733 Parent->setType(TT_CastRParen); 734 } 735 ColonFound = true; 736 } 737 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) && 738 !ColonFound) 739 Left->setType(TT_ArrayInitializerLSquare); 740 FormatToken *Tok = CurrentToken; 741 if (!consumeToken()) 742 return false; 743 updateParameterCount(Left, Tok); 744 } 745 return false; 746 } 747 748 bool couldBeInStructArrayInitializer() const { 749 if (Contexts.size() < 2) 750 return false; 751 // We want to back up no more then 2 context levels i.e. 752 // . { { <- 753 const auto End = std::next(Contexts.rbegin(), 2); 754 auto Last = Contexts.rbegin(); 755 unsigned Depth = 0; 756 for (; Last != End; ++Last) { 757 if (Last->ContextKind == tok::l_brace) 758 ++Depth; 759 } 760 return Depth == 2 && Last->ContextKind != tok::l_brace; 761 } 762 763 bool parseBrace() { 764 if (CurrentToken) { 765 FormatToken *Left = CurrentToken->Previous; 766 Left->ParentBracket = Contexts.back().ContextKind; 767 768 if (Contexts.back().CaretFound) 769 Left->setType(TT_ObjCBlockLBrace); 770 Contexts.back().CaretFound = false; 771 772 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1); 773 Contexts.back().ColonIsDictLiteral = true; 774 if (Left->is(BK_BracedInit)) 775 Contexts.back().IsExpression = true; 776 if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 777 Left->Previous->is(TT_JsTypeColon)) 778 Contexts.back().IsExpression = false; 779 780 unsigned CommaCount = 0; 781 while (CurrentToken) { 782 if (CurrentToken->is(tok::r_brace)) { 783 Left->MatchingParen = CurrentToken; 784 CurrentToken->MatchingParen = Left; 785 if (Style.AlignArrayOfStructures != FormatStyle::AIAS_None) { 786 if (Left->ParentBracket == tok::l_brace && 787 couldBeInStructArrayInitializer() && CommaCount > 0) { 788 Contexts.back().InStructArrayInitializer = true; 789 } 790 } 791 next(); 792 return true; 793 } 794 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square)) 795 return false; 796 updateParameterCount(Left, CurrentToken); 797 if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) { 798 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 799 if (Previous->is(TT_JsTypeOptionalQuestion)) 800 Previous = Previous->getPreviousNonComment(); 801 if ((CurrentToken->is(tok::colon) && 802 (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) || 803 Style.Language == FormatStyle::LK_Proto || 804 Style.Language == FormatStyle::LK_TextProto) { 805 Left->setType(TT_DictLiteral); 806 if (Previous->Tok.getIdentifierInfo() || 807 Previous->is(tok::string_literal)) 808 Previous->setType(TT_SelectorName); 809 } 810 if (CurrentToken->is(tok::colon) || 811 Style.Language == FormatStyle::LK_JavaScript) 812 Left->setType(TT_DictLiteral); 813 } 814 if (CurrentToken->is(tok::comma)) { 815 if (Style.Language == FormatStyle::LK_JavaScript) 816 Left->setType(TT_DictLiteral); 817 ++CommaCount; 818 } 819 if (!consumeToken()) 820 return false; 821 } 822 } 823 return true; 824 } 825 826 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 827 // For ObjC methods, the number of parameters is calculated differently as 828 // method declarations have a different structure (the parameters are not 829 // inside a bracket scope). 830 if (Current->is(tok::l_brace) && Current->is(BK_Block)) 831 ++Left->BlockParameterCount; 832 if (Current->is(tok::comma)) { 833 ++Left->ParameterCount; 834 if (!Left->Role) 835 Left->Role.reset(new CommaSeparatedList(Style)); 836 Left->Role->CommaFound(Current); 837 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 838 Left->ParameterCount = 1; 839 } 840 } 841 842 bool parseConditional() { 843 while (CurrentToken) { 844 if (CurrentToken->is(tok::colon)) { 845 CurrentToken->setType(TT_ConditionalExpr); 846 next(); 847 return true; 848 } 849 if (!consumeToken()) 850 return false; 851 } 852 return false; 853 } 854 855 bool parseTemplateDeclaration() { 856 if (CurrentToken && CurrentToken->is(tok::less)) { 857 CurrentToken->setType(TT_TemplateOpener); 858 next(); 859 if (!parseAngle()) 860 return false; 861 if (CurrentToken) 862 CurrentToken->Previous->ClosesTemplateDeclaration = true; 863 return true; 864 } 865 return false; 866 } 867 868 bool consumeToken() { 869 FormatToken *Tok = CurrentToken; 870 next(); 871 switch (Tok->Tok.getKind()) { 872 case tok::plus: 873 case tok::minus: 874 if (!Tok->Previous && Line.MustBeDeclaration) 875 Tok->setType(TT_ObjCMethodSpecifier); 876 break; 877 case tok::colon: 878 if (!Tok->Previous) 879 return false; 880 // Colons from ?: are handled in parseConditional(). 881 if (Style.Language == FormatStyle::LK_JavaScript) { 882 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop 883 (Contexts.size() == 1 && // switch/case labels 884 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) || 885 Contexts.back().ContextKind == tok::l_paren || // function params 886 Contexts.back().ContextKind == tok::l_square || // array type 887 (!Contexts.back().IsExpression && 888 Contexts.back().ContextKind == tok::l_brace) || // object type 889 (Contexts.size() == 1 && 890 Line.MustBeDeclaration)) { // method/property declaration 891 Contexts.back().IsExpression = false; 892 Tok->setType(TT_JsTypeColon); 893 break; 894 } 895 } else if (Style.isCSharp()) { 896 if (Contexts.back().InCSharpAttributeSpecifier) { 897 Tok->setType(TT_AttributeColon); 898 break; 899 } 900 if (Contexts.back().ContextKind == tok::l_paren) { 901 Tok->setType(TT_CSharpNamedArgumentColon); 902 break; 903 } 904 } 905 if (Contexts.back().ColonIsDictLiteral || 906 Style.Language == FormatStyle::LK_Proto || 907 Style.Language == FormatStyle::LK_TextProto) { 908 Tok->setType(TT_DictLiteral); 909 if (Style.Language == FormatStyle::LK_TextProto) { 910 if (FormatToken *Previous = Tok->getPreviousNonComment()) 911 Previous->setType(TT_SelectorName); 912 } 913 } else if (Contexts.back().ColonIsObjCMethodExpr || 914 Line.startsWith(TT_ObjCMethodSpecifier)) { 915 Tok->setType(TT_ObjCMethodExpr); 916 const FormatToken *BeforePrevious = Tok->Previous->Previous; 917 // Ensure we tag all identifiers in method declarations as 918 // TT_SelectorName. 919 bool UnknownIdentifierInMethodDeclaration = 920 Line.startsWith(TT_ObjCMethodSpecifier) && 921 Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown); 922 if (!BeforePrevious || 923 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 924 !(BeforePrevious->is(TT_CastRParen) || 925 (BeforePrevious->is(TT_ObjCMethodExpr) && 926 BeforePrevious->is(tok::colon))) || 927 BeforePrevious->is(tok::r_square) || 928 Contexts.back().LongestObjCSelectorName == 0 || 929 UnknownIdentifierInMethodDeclaration) { 930 Tok->Previous->setType(TT_SelectorName); 931 if (!Contexts.back().FirstObjCSelectorName) 932 Contexts.back().FirstObjCSelectorName = Tok->Previous; 933 else if (Tok->Previous->ColumnWidth > 934 Contexts.back().LongestObjCSelectorName) 935 Contexts.back().LongestObjCSelectorName = 936 Tok->Previous->ColumnWidth; 937 Tok->Previous->ParameterIndex = 938 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 939 ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 940 } 941 } else if (Contexts.back().ColonIsForRangeExpr) { 942 Tok->setType(TT_RangeBasedForLoopColon); 943 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 944 Tok->setType(TT_BitFieldColon); 945 } else if (Contexts.size() == 1 && 946 !Line.First->isOneOf(tok::kw_enum, tok::kw_case, 947 tok::kw_default)) { 948 FormatToken *Prev = Tok->getPreviousNonComment(); 949 if (!Prev) 950 break; 951 if (Prev->isOneOf(tok::r_paren, tok::kw_noexcept)) 952 Tok->setType(TT_CtorInitializerColon); 953 else if (Prev->is(tok::kw_try)) { 954 // Member initializer list within function try block. 955 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 956 if (!PrevPrev) 957 break; 958 if (PrevPrev && PrevPrev->isOneOf(tok::r_paren, tok::kw_noexcept)) 959 Tok->setType(TT_CtorInitializerColon); 960 } else 961 Tok->setType(TT_InheritanceColon); 962 } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next && 963 (Tok->Next->isOneOf(tok::r_paren, tok::comma) || 964 (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next && 965 Tok->Next->Next->is(tok::colon)))) { 966 // This handles a special macro in ObjC code where selectors including 967 // the colon are passed as macro arguments. 968 Tok->setType(TT_ObjCMethodExpr); 969 } else if (Contexts.back().ContextKind == tok::l_paren) { 970 Tok->setType(TT_InlineASMColon); 971 } 972 break; 973 case tok::pipe: 974 case tok::amp: 975 // | and & in declarations/type expressions represent union and 976 // intersection types, respectively. 977 if (Style.Language == FormatStyle::LK_JavaScript && 978 !Contexts.back().IsExpression) 979 Tok->setType(TT_JsTypeOperator); 980 break; 981 case tok::kw_if: 982 case tok::kw_while: 983 if (Tok->is(tok::kw_if) && CurrentToken && 984 CurrentToken->isOneOf(tok::kw_constexpr, tok::identifier)) 985 next(); 986 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 987 next(); 988 if (!parseParens(/*LookForDecls=*/true)) 989 return false; 990 } 991 break; 992 case tok::kw_for: 993 if (Style.Language == FormatStyle::LK_JavaScript) { 994 // x.for and {for: ...} 995 if ((Tok->Previous && Tok->Previous->is(tok::period)) || 996 (Tok->Next && Tok->Next->is(tok::colon))) 997 break; 998 // JS' for await ( ... 999 if (CurrentToken && CurrentToken->is(Keywords.kw_await)) 1000 next(); 1001 } 1002 Contexts.back().ColonIsForRangeExpr = true; 1003 next(); 1004 if (!parseParens()) 1005 return false; 1006 break; 1007 case tok::l_paren: 1008 // When faced with 'operator()()', the kw_operator handler incorrectly 1009 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make 1010 // the first two parens OverloadedOperators and the second l_paren an 1011 // OverloadedOperatorLParen. 1012 if (Tok->Previous && Tok->Previous->is(tok::r_paren) && 1013 Tok->Previous->MatchingParen && 1014 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) { 1015 Tok->Previous->setType(TT_OverloadedOperator); 1016 Tok->Previous->MatchingParen->setType(TT_OverloadedOperator); 1017 Tok->setType(TT_OverloadedOperatorLParen); 1018 } 1019 1020 if (!parseParens()) 1021 return false; 1022 if (Line.MustBeDeclaration && Contexts.size() == 1 && 1023 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) && 1024 !Tok->is(TT_TypeDeclarationParen) && 1025 (!Tok->Previous || !Tok->Previous->isOneOf(tok::kw___attribute, 1026 TT_LeadingJavaAnnotation))) 1027 Line.MightBeFunctionDecl = true; 1028 break; 1029 case tok::l_square: 1030 if (!parseSquare()) 1031 return false; 1032 break; 1033 case tok::l_brace: 1034 if (Style.Language == FormatStyle::LK_TextProto) { 1035 FormatToken *Previous = Tok->getPreviousNonComment(); 1036 if (Previous && Previous->getType() != TT_DictLiteral) 1037 Previous->setType(TT_SelectorName); 1038 } 1039 if (!parseBrace()) 1040 return false; 1041 break; 1042 case tok::less: 1043 if (parseAngle()) { 1044 Tok->setType(TT_TemplateOpener); 1045 // In TT_Proto, we must distignuish between: 1046 // map<key, value> 1047 // msg < item: data > 1048 // msg: < item: data > 1049 // In TT_TextProto, map<key, value> does not occur. 1050 if (Style.Language == FormatStyle::LK_TextProto || 1051 (Style.Language == FormatStyle::LK_Proto && Tok->Previous && 1052 Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) { 1053 Tok->setType(TT_DictLiteral); 1054 FormatToken *Previous = Tok->getPreviousNonComment(); 1055 if (Previous && Previous->getType() != TT_DictLiteral) 1056 Previous->setType(TT_SelectorName); 1057 } 1058 } else { 1059 Tok->setType(TT_BinaryOperator); 1060 NonTemplateLess.insert(Tok); 1061 CurrentToken = Tok; 1062 next(); 1063 } 1064 break; 1065 case tok::r_paren: 1066 case tok::r_square: 1067 return false; 1068 case tok::r_brace: 1069 // Lines can start with '}'. 1070 if (Tok->Previous) 1071 return false; 1072 break; 1073 case tok::greater: 1074 if (Style.Language != FormatStyle::LK_TextProto) 1075 Tok->setType(TT_BinaryOperator); 1076 if (Tok->Previous && Tok->Previous->is(TT_TemplateCloser)) 1077 Tok->SpacesRequiredBefore = 1; 1078 break; 1079 case tok::kw_operator: 1080 if (Style.Language == FormatStyle::LK_TextProto || 1081 Style.Language == FormatStyle::LK_Proto) 1082 break; 1083 while (CurrentToken && 1084 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 1085 if (CurrentToken->isOneOf(tok::star, tok::amp)) 1086 CurrentToken->setType(TT_PointerOrReference); 1087 consumeToken(); 1088 if (CurrentToken && CurrentToken->is(tok::comma) && 1089 CurrentToken->Previous->isNot(tok::kw_operator)) 1090 break; 1091 if (CurrentToken && CurrentToken->Previous->isOneOf( 1092 TT_BinaryOperator, TT_UnaryOperator, tok::comma, 1093 tok::star, tok::arrow, tok::amp, tok::ampamp)) 1094 CurrentToken->Previous->setType(TT_OverloadedOperator); 1095 } 1096 if (CurrentToken && CurrentToken->is(tok::l_paren)) 1097 CurrentToken->setType(TT_OverloadedOperatorLParen); 1098 if (CurrentToken && CurrentToken->Previous->is(TT_BinaryOperator)) 1099 CurrentToken->Previous->setType(TT_OverloadedOperator); 1100 break; 1101 case tok::question: 1102 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next && 1103 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren, 1104 tok::r_brace)) { 1105 // Question marks before semicolons, colons, etc. indicate optional 1106 // types (fields, parameters), e.g. 1107 // function(x?: string, y?) {...} 1108 // class X { y?; } 1109 Tok->setType(TT_JsTypeOptionalQuestion); 1110 break; 1111 } 1112 // Declarations cannot be conditional expressions, this can only be part 1113 // of a type declaration. 1114 if (Line.MustBeDeclaration && !Contexts.back().IsExpression && 1115 Style.Language == FormatStyle::LK_JavaScript) 1116 break; 1117 if (Style.isCSharp()) { 1118 // `Type?)`, `Type?>`, `Type? name;` and `Type? name =` can only be 1119 // nullable types. 1120 // Line.MustBeDeclaration will be true for `Type? name;`. 1121 if ((!Contexts.back().IsExpression && Line.MustBeDeclaration) || 1122 (Tok->Next && Tok->Next->isOneOf(tok::r_paren, tok::greater)) || 1123 (Tok->Next && Tok->Next->is(tok::identifier) && Tok->Next->Next && 1124 Tok->Next->Next->is(tok::equal))) { 1125 Tok->setType(TT_CSharpNullable); 1126 break; 1127 } 1128 } 1129 parseConditional(); 1130 break; 1131 case tok::kw_template: 1132 parseTemplateDeclaration(); 1133 break; 1134 case tok::comma: 1135 if (Contexts.back().InCtorInitializer) 1136 Tok->setType(TT_CtorInitializerComma); 1137 else if (Contexts.back().InInheritanceList) 1138 Tok->setType(TT_InheritanceComma); 1139 else if (Contexts.back().FirstStartOfName && 1140 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) { 1141 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 1142 Line.IsMultiVariableDeclStmt = true; 1143 } 1144 if (Contexts.back().IsForEachMacro) 1145 Contexts.back().IsExpression = true; 1146 break; 1147 case tok::identifier: 1148 if (Tok->isOneOf(Keywords.kw___has_include, 1149 Keywords.kw___has_include_next)) { 1150 parseHasInclude(); 1151 } 1152 if (Style.isCSharp() && Tok->is(Keywords.kw_where) && Tok->Next && 1153 Tok->Next->isNot(tok::l_paren)) { 1154 Tok->setType(TT_CSharpGenericTypeConstraint); 1155 parseCSharpGenericTypeConstraint(); 1156 } 1157 break; 1158 default: 1159 break; 1160 } 1161 return true; 1162 } 1163 1164 void parseCSharpGenericTypeConstraint() { 1165 int OpenAngleBracketsCount = 0; 1166 while (CurrentToken) { 1167 if (CurrentToken->is(tok::less)) { 1168 // parseAngle is too greedy and will consume the whole line. 1169 CurrentToken->setType(TT_TemplateOpener); 1170 ++OpenAngleBracketsCount; 1171 next(); 1172 } else if (CurrentToken->is(tok::greater)) { 1173 CurrentToken->setType(TT_TemplateCloser); 1174 --OpenAngleBracketsCount; 1175 next(); 1176 } else if (CurrentToken->is(tok::comma) && OpenAngleBracketsCount == 0) { 1177 // We allow line breaks after GenericTypeConstraintComma's 1178 // so do not flag commas in Generics as GenericTypeConstraintComma's. 1179 CurrentToken->setType(TT_CSharpGenericTypeConstraintComma); 1180 next(); 1181 } else if (CurrentToken->is(Keywords.kw_where)) { 1182 CurrentToken->setType(TT_CSharpGenericTypeConstraint); 1183 next(); 1184 } else if (CurrentToken->is(tok::colon)) { 1185 CurrentToken->setType(TT_CSharpGenericTypeConstraintColon); 1186 next(); 1187 } else { 1188 next(); 1189 } 1190 } 1191 } 1192 1193 void parseIncludeDirective() { 1194 if (CurrentToken && CurrentToken->is(tok::less)) { 1195 next(); 1196 while (CurrentToken) { 1197 // Mark tokens up to the trailing line comments as implicit string 1198 // literals. 1199 if (CurrentToken->isNot(tok::comment) && 1200 !CurrentToken->TokenText.startswith("//")) 1201 CurrentToken->setType(TT_ImplicitStringLiteral); 1202 next(); 1203 } 1204 } 1205 } 1206 1207 void parseWarningOrError() { 1208 next(); 1209 // We still want to format the whitespace left of the first token of the 1210 // warning or error. 1211 next(); 1212 while (CurrentToken) { 1213 CurrentToken->setType(TT_ImplicitStringLiteral); 1214 next(); 1215 } 1216 } 1217 1218 void parsePragma() { 1219 next(); // Consume "pragma". 1220 if (CurrentToken && 1221 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) { 1222 bool IsMark = CurrentToken->is(Keywords.kw_mark); 1223 next(); // Consume "mark". 1224 next(); // Consume first token (so we fix leading whitespace). 1225 while (CurrentToken) { 1226 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator)) 1227 CurrentToken->setType(TT_ImplicitStringLiteral); 1228 next(); 1229 } 1230 } 1231 } 1232 1233 void parseHasInclude() { 1234 if (!CurrentToken || !CurrentToken->is(tok::l_paren)) 1235 return; 1236 next(); // '(' 1237 parseIncludeDirective(); 1238 next(); // ')' 1239 } 1240 1241 LineType parsePreprocessorDirective() { 1242 bool IsFirstToken = CurrentToken->IsFirst; 1243 LineType Type = LT_PreprocessorDirective; 1244 next(); 1245 if (!CurrentToken) 1246 return Type; 1247 1248 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) { 1249 // JavaScript files can contain shebang lines of the form: 1250 // #!/usr/bin/env node 1251 // Treat these like C++ #include directives. 1252 while (CurrentToken) { 1253 // Tokens cannot be comments here. 1254 CurrentToken->setType(TT_ImplicitStringLiteral); 1255 next(); 1256 } 1257 return LT_ImportStatement; 1258 } 1259 1260 if (CurrentToken->Tok.is(tok::numeric_constant)) { 1261 CurrentToken->SpacesRequiredBefore = 1; 1262 return Type; 1263 } 1264 // Hashes in the middle of a line can lead to any strange token 1265 // sequence. 1266 if (!CurrentToken->Tok.getIdentifierInfo()) 1267 return Type; 1268 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 1269 case tok::pp_include: 1270 case tok::pp_include_next: 1271 case tok::pp_import: 1272 next(); 1273 parseIncludeDirective(); 1274 Type = LT_ImportStatement; 1275 break; 1276 case tok::pp_error: 1277 case tok::pp_warning: 1278 parseWarningOrError(); 1279 break; 1280 case tok::pp_pragma: 1281 parsePragma(); 1282 break; 1283 case tok::pp_if: 1284 case tok::pp_elif: 1285 Contexts.back().IsExpression = true; 1286 next(); 1287 parseLine(); 1288 break; 1289 default: 1290 break; 1291 } 1292 while (CurrentToken) { 1293 FormatToken *Tok = CurrentToken; 1294 next(); 1295 if (Tok->is(tok::l_paren)) 1296 parseParens(); 1297 else if (Tok->isOneOf(Keywords.kw___has_include, 1298 Keywords.kw___has_include_next)) 1299 parseHasInclude(); 1300 } 1301 return Type; 1302 } 1303 1304 public: 1305 LineType parseLine() { 1306 if (!CurrentToken) 1307 return LT_Invalid; 1308 NonTemplateLess.clear(); 1309 if (CurrentToken->is(tok::hash)) 1310 return parsePreprocessorDirective(); 1311 1312 // Directly allow to 'import <string-literal>' to support protocol buffer 1313 // definitions (github.com/google/protobuf) or missing "#" (either way we 1314 // should not break the line). 1315 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 1316 if ((Style.Language == FormatStyle::LK_Java && 1317 CurrentToken->is(Keywords.kw_package)) || 1318 (Info && Info->getPPKeywordID() == tok::pp_import && 1319 CurrentToken->Next && 1320 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier, 1321 tok::kw_static))) { 1322 next(); 1323 parseIncludeDirective(); 1324 return LT_ImportStatement; 1325 } 1326 1327 // If this line starts and ends in '<' and '>', respectively, it is likely 1328 // part of "#define <a/b.h>". 1329 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 1330 parseIncludeDirective(); 1331 return LT_ImportStatement; 1332 } 1333 1334 // In .proto files, top-level options and package statements are very 1335 // similar to import statements and should not be line-wrapped. 1336 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 && 1337 CurrentToken->isOneOf(Keywords.kw_option, Keywords.kw_package)) { 1338 next(); 1339 if (CurrentToken && CurrentToken->is(tok::identifier)) { 1340 while (CurrentToken) 1341 next(); 1342 return LT_ImportStatement; 1343 } 1344 } 1345 1346 bool KeywordVirtualFound = false; 1347 bool ImportStatement = false; 1348 1349 // import {...} from '...'; 1350 if (Style.Language == FormatStyle::LK_JavaScript && 1351 CurrentToken->is(Keywords.kw_import)) 1352 ImportStatement = true; 1353 1354 while (CurrentToken) { 1355 if (CurrentToken->is(tok::kw_virtual)) 1356 KeywordVirtualFound = true; 1357 if (Style.Language == FormatStyle::LK_JavaScript) { 1358 // export {...} from '...'; 1359 // An export followed by "from 'some string';" is a re-export from 1360 // another module identified by a URI and is treated as a 1361 // LT_ImportStatement (i.e. prevent wraps on it for long URIs). 1362 // Just "export {...};" or "export class ..." should not be treated as 1363 // an import in this sense. 1364 if (Line.First->is(tok::kw_export) && 1365 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next && 1366 CurrentToken->Next->isStringLiteral()) 1367 ImportStatement = true; 1368 if (isClosureImportStatement(*CurrentToken)) 1369 ImportStatement = true; 1370 } 1371 if (!consumeToken()) 1372 return LT_Invalid; 1373 } 1374 if (KeywordVirtualFound) 1375 return LT_VirtualFunctionDecl; 1376 if (ImportStatement) 1377 return LT_ImportStatement; 1378 1379 if (Line.startsWith(TT_ObjCMethodSpecifier)) { 1380 if (Contexts.back().FirstObjCSelectorName) 1381 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 1382 Contexts.back().LongestObjCSelectorName; 1383 return LT_ObjCMethodDecl; 1384 } 1385 1386 for (const auto &ctx : Contexts) { 1387 if (ctx.InStructArrayInitializer) { 1388 return LT_ArrayOfStructInitializer; 1389 } 1390 } 1391 1392 return LT_Other; 1393 } 1394 1395 private: 1396 bool isClosureImportStatement(const FormatToken &Tok) { 1397 // FIXME: Closure-library specific stuff should not be hard-coded but be 1398 // configurable. 1399 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) && 1400 Tok.Next->Next && 1401 (Tok.Next->Next->TokenText == "module" || 1402 Tok.Next->Next->TokenText == "provide" || 1403 Tok.Next->Next->TokenText == "require" || 1404 Tok.Next->Next->TokenText == "requireType" || 1405 Tok.Next->Next->TokenText == "forwardDeclare") && 1406 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren); 1407 } 1408 1409 void resetTokenMetadata(FormatToken *Token) { 1410 if (!Token) 1411 return; 1412 1413 // Reset token type in case we have already looked at it and then 1414 // recovered from an error (e.g. failure to find the matching >). 1415 if (!CurrentToken->isOneOf( 1416 TT_LambdaLSquare, TT_LambdaLBrace, TT_AttributeMacro, TT_IfMacro, 1417 TT_ForEachMacro, TT_TypenameMacro, TT_FunctionLBrace, 1418 TT_ImplicitStringLiteral, TT_InlineASMBrace, TT_FatArrow, 1419 TT_LambdaArrow, TT_NamespaceMacro, TT_OverloadedOperator, 1420 TT_RegexLiteral, TT_TemplateString, TT_ObjCStringLiteral, 1421 TT_UntouchableMacroFunc, TT_ConstraintJunctions, 1422 TT_StatementAttributeLikeMacro)) 1423 CurrentToken->setType(TT_Unknown); 1424 CurrentToken->Role.reset(); 1425 CurrentToken->MatchingParen = nullptr; 1426 CurrentToken->FakeLParens.clear(); 1427 CurrentToken->FakeRParens = 0; 1428 } 1429 1430 void next() { 1431 if (CurrentToken) { 1432 CurrentToken->NestingLevel = Contexts.size() - 1; 1433 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 1434 modifyContext(*CurrentToken); 1435 determineTokenType(*CurrentToken); 1436 CurrentToken = CurrentToken->Next; 1437 } 1438 1439 resetTokenMetadata(CurrentToken); 1440 } 1441 1442 /// A struct to hold information valid in a specific context, e.g. 1443 /// a pair of parenthesis. 1444 struct Context { 1445 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 1446 bool IsExpression) 1447 : ContextKind(ContextKind), BindingStrength(BindingStrength), 1448 IsExpression(IsExpression) {} 1449 1450 tok::TokenKind ContextKind; 1451 unsigned BindingStrength; 1452 bool IsExpression; 1453 unsigned LongestObjCSelectorName = 0; 1454 bool ColonIsForRangeExpr = false; 1455 bool ColonIsDictLiteral = false; 1456 bool ColonIsObjCMethodExpr = false; 1457 FormatToken *FirstObjCSelectorName = nullptr; 1458 FormatToken *FirstStartOfName = nullptr; 1459 bool CanBeExpression = true; 1460 bool InTemplateArgument = false; 1461 bool InCtorInitializer = false; 1462 bool InInheritanceList = false; 1463 bool CaretFound = false; 1464 bool IsForEachMacro = false; 1465 bool InCpp11AttributeSpecifier = false; 1466 bool InCSharpAttributeSpecifier = false; 1467 bool InStructArrayInitializer = false; 1468 }; 1469 1470 /// Puts a new \c Context onto the stack \c Contexts for the lifetime 1471 /// of each instance. 1472 struct ScopedContextCreator { 1473 AnnotatingParser &P; 1474 1475 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 1476 unsigned Increase) 1477 : P(P) { 1478 P.Contexts.push_back(Context(ContextKind, 1479 P.Contexts.back().BindingStrength + Increase, 1480 P.Contexts.back().IsExpression)); 1481 } 1482 1483 ~ScopedContextCreator() { 1484 if (P.Style.AlignArrayOfStructures != FormatStyle::AIAS_None) { 1485 if (P.Contexts.back().InStructArrayInitializer) { 1486 P.Contexts.pop_back(); 1487 P.Contexts.back().InStructArrayInitializer = true; 1488 return; 1489 } 1490 } 1491 P.Contexts.pop_back(); 1492 } 1493 }; 1494 1495 void modifyContext(const FormatToken &Current) { 1496 if (Current.getPrecedence() == prec::Assignment && 1497 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) && 1498 // Type aliases use `type X = ...;` in TypeScript and can be exported 1499 // using `export type ...`. 1500 !(Style.Language == FormatStyle::LK_JavaScript && 1501 (Line.startsWith(Keywords.kw_type, tok::identifier) || 1502 Line.startsWith(tok::kw_export, Keywords.kw_type, 1503 tok::identifier))) && 1504 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 1505 Contexts.back().IsExpression = true; 1506 if (!Line.startsWith(TT_UnaryOperator)) { 1507 for (FormatToken *Previous = Current.Previous; 1508 Previous && Previous->Previous && 1509 !Previous->Previous->isOneOf(tok::comma, tok::semi); 1510 Previous = Previous->Previous) { 1511 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 1512 Previous = Previous->MatchingParen; 1513 if (!Previous) 1514 break; 1515 } 1516 if (Previous->opensScope()) 1517 break; 1518 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) && 1519 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) && 1520 Previous->Previous && Previous->Previous->isNot(tok::equal)) 1521 Previous->setType(TT_PointerOrReference); 1522 } 1523 } 1524 } else if (Current.is(tok::lessless) && 1525 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) { 1526 Contexts.back().IsExpression = true; 1527 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 1528 Contexts.back().IsExpression = true; 1529 } else if (Current.is(TT_TrailingReturnArrow)) { 1530 Contexts.back().IsExpression = false; 1531 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) { 1532 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java; 1533 } else if (Current.Previous && 1534 Current.Previous->is(TT_CtorInitializerColon)) { 1535 Contexts.back().IsExpression = true; 1536 Contexts.back().InCtorInitializer = true; 1537 } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) { 1538 Contexts.back().InInheritanceList = true; 1539 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 1540 for (FormatToken *Previous = Current.Previous; 1541 Previous && Previous->isOneOf(tok::star, tok::amp); 1542 Previous = Previous->Previous) 1543 Previous->setType(TT_PointerOrReference); 1544 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer) 1545 Contexts.back().IsExpression = false; 1546 } else if (Current.is(tok::kw_new)) { 1547 Contexts.back().CanBeExpression = false; 1548 } else if (Current.is(tok::semi) || 1549 (Current.is(tok::exclaim) && Current.Previous && 1550 !Current.Previous->is(tok::kw_operator))) { 1551 // This should be the condition or increment in a for-loop. 1552 // But not operator !() (can't use TT_OverloadedOperator here as its not 1553 // been annotated yet). 1554 Contexts.back().IsExpression = true; 1555 } 1556 } 1557 1558 static FormatToken *untilMatchingParen(FormatToken *Current) { 1559 // Used when `MatchingParen` is not yet established. 1560 int ParenLevel = 0; 1561 while (Current) { 1562 if (Current->is(tok::l_paren)) 1563 ParenLevel++; 1564 if (Current->is(tok::r_paren)) 1565 ParenLevel--; 1566 if (ParenLevel < 1) 1567 break; 1568 Current = Current->Next; 1569 } 1570 return Current; 1571 } 1572 1573 static bool isDeductionGuide(FormatToken &Current) { 1574 // Look for a deduction guide template<T> A(...) -> A<...>; 1575 if (Current.Previous && Current.Previous->is(tok::r_paren) && 1576 Current.startsSequence(tok::arrow, tok::identifier, tok::less)) { 1577 // Find the TemplateCloser. 1578 FormatToken *TemplateCloser = Current.Next->Next; 1579 int NestingLevel = 0; 1580 while (TemplateCloser) { 1581 // Skip over an expressions in parens A<(3 < 2)>; 1582 if (TemplateCloser->is(tok::l_paren)) { 1583 // No Matching Paren yet so skip to matching paren 1584 TemplateCloser = untilMatchingParen(TemplateCloser); 1585 if (!TemplateCloser) 1586 break; 1587 } 1588 if (TemplateCloser->is(tok::less)) 1589 NestingLevel++; 1590 if (TemplateCloser->is(tok::greater)) 1591 NestingLevel--; 1592 if (NestingLevel < 1) 1593 break; 1594 TemplateCloser = TemplateCloser->Next; 1595 } 1596 // Assuming we have found the end of the template ensure its followed 1597 // with a semi-colon. 1598 if (TemplateCloser && TemplateCloser->Next && 1599 TemplateCloser->Next->is(tok::semi) && 1600 Current.Previous->MatchingParen) { 1601 // Determine if the identifier `A` prior to the A<..>; is the same as 1602 // prior to the A(..) 1603 FormatToken *LeadingIdentifier = 1604 Current.Previous->MatchingParen->Previous; 1605 1606 // Differentiate a deduction guide by seeing the 1607 // > of the template prior to the leading identifier. 1608 if (LeadingIdentifier) { 1609 FormatToken *PriorLeadingIdentifier = LeadingIdentifier->Previous; 1610 // Skip back past explicit decoration 1611 if (PriorLeadingIdentifier && 1612 PriorLeadingIdentifier->is(tok::kw_explicit)) 1613 PriorLeadingIdentifier = PriorLeadingIdentifier->Previous; 1614 1615 return (PriorLeadingIdentifier && 1616 PriorLeadingIdentifier->is(TT_TemplateCloser) && 1617 LeadingIdentifier->TokenText == Current.Next->TokenText); 1618 } 1619 } 1620 } 1621 return false; 1622 } 1623 1624 void determineTokenType(FormatToken &Current) { 1625 if (!Current.is(TT_Unknown)) 1626 // The token type is already known. 1627 return; 1628 1629 if ((Style.Language == FormatStyle::LK_JavaScript || Style.isCSharp()) && 1630 Current.is(tok::exclaim)) { 1631 if (Current.Previous) { 1632 bool IsIdentifier = 1633 Style.Language == FormatStyle::LK_JavaScript 1634 ? Keywords.IsJavaScriptIdentifier( 1635 *Current.Previous, /* AcceptIdentifierName= */ true) 1636 : Current.Previous->is(tok::identifier); 1637 if (IsIdentifier || 1638 Current.Previous->isOneOf( 1639 tok::kw_namespace, tok::r_paren, tok::r_square, tok::r_brace, 1640 tok::kw_false, tok::kw_true, Keywords.kw_type, Keywords.kw_get, 1641 Keywords.kw_set) || 1642 Current.Previous->Tok.isLiteral()) { 1643 Current.setType(TT_NonNullAssertion); 1644 return; 1645 } 1646 } 1647 if (Current.Next && 1648 Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) { 1649 Current.setType(TT_NonNullAssertion); 1650 return; 1651 } 1652 } 1653 1654 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1655 // function declaration have been found. In this case, 'Current' is a 1656 // trailing token of this declaration and thus cannot be a name. 1657 if (Current.is(Keywords.kw_instanceof)) { 1658 Current.setType(TT_BinaryOperator); 1659 } else if (isStartOfName(Current) && 1660 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 1661 Contexts.back().FirstStartOfName = &Current; 1662 Current.setType(TT_StartOfName); 1663 } else if (Current.is(tok::semi)) { 1664 // Reset FirstStartOfName after finding a semicolon so that a for loop 1665 // with multiple increment statements is not confused with a for loop 1666 // having multiple variable declarations. 1667 Contexts.back().FirstStartOfName = nullptr; 1668 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) { 1669 AutoFound = true; 1670 } else if (Current.is(tok::arrow) && 1671 Style.Language == FormatStyle::LK_Java) { 1672 Current.setType(TT_LambdaArrow); 1673 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration && 1674 Current.NestingLevel == 0 && 1675 !Current.Previous->is(tok::kw_operator)) { 1676 // not auto operator->() -> xxx; 1677 Current.setType(TT_TrailingReturnArrow); 1678 } else if (Current.is(tok::arrow) && Current.Previous && 1679 Current.Previous->is(tok::r_brace)) { 1680 // Concept implicit conversion constraint needs to be treated like 1681 // a trailing return type ... } -> <type>. 1682 Current.setType(TT_TrailingReturnArrow); 1683 } else if (isDeductionGuide(Current)) { 1684 // Deduction guides trailing arrow " A(...) -> A<T>;". 1685 Current.setType(TT_TrailingReturnArrow); 1686 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 1687 Current.setType(determineStarAmpUsage( 1688 Current, 1689 Contexts.back().CanBeExpression && Contexts.back().IsExpression, 1690 Contexts.back().InTemplateArgument)); 1691 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 1692 Current.setType(determinePlusMinusCaretUsage(Current)); 1693 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret)) 1694 Contexts.back().CaretFound = true; 1695 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 1696 Current.setType(determineIncrementUsage(Current)); 1697 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 1698 Current.setType(TT_UnaryOperator); 1699 } else if (Current.is(tok::question)) { 1700 if (Style.Language == FormatStyle::LK_JavaScript && 1701 Line.MustBeDeclaration && !Contexts.back().IsExpression) { 1702 // In JavaScript, `interface X { foo?(): bar; }` is an optional method 1703 // on the interface, not a ternary expression. 1704 Current.setType(TT_JsTypeOptionalQuestion); 1705 } else { 1706 Current.setType(TT_ConditionalExpr); 1707 } 1708 } else if (Current.isBinaryOperator() && 1709 (!Current.Previous || Current.Previous->isNot(tok::l_square)) && 1710 (!Current.is(tok::greater) && 1711 Style.Language != FormatStyle::LK_TextProto)) { 1712 Current.setType(TT_BinaryOperator); 1713 } else if (Current.is(tok::comment)) { 1714 if (Current.TokenText.startswith("/*")) { 1715 if (Current.TokenText.endswith("*/")) 1716 Current.setType(TT_BlockComment); 1717 else 1718 // The lexer has for some reason determined a comment here. But we 1719 // cannot really handle it, if it isn't properly terminated. 1720 Current.Tok.setKind(tok::unknown); 1721 } else { 1722 Current.setType(TT_LineComment); 1723 } 1724 } else if (Current.is(tok::r_paren)) { 1725 if (rParenEndsCast(Current)) 1726 Current.setType(TT_CastRParen); 1727 if (Current.MatchingParen && Current.Next && 1728 !Current.Next->isBinaryOperator() && 1729 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace, 1730 tok::comma, tok::period, tok::arrow, 1731 tok::coloncolon)) 1732 if (FormatToken *AfterParen = Current.MatchingParen->Next) { 1733 // Make sure this isn't the return type of an Obj-C block declaration 1734 if (AfterParen->Tok.isNot(tok::caret)) { 1735 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) 1736 if (BeforeParen->is(tok::identifier) && 1737 !BeforeParen->is(TT_TypenameMacro) && 1738 BeforeParen->TokenText == BeforeParen->TokenText.upper() && 1739 (!BeforeParen->Previous || 1740 BeforeParen->Previous->ClosesTemplateDeclaration)) 1741 Current.setType(TT_FunctionAnnotationRParen); 1742 } 1743 } 1744 } else if (Current.is(tok::at) && Current.Next && 1745 Style.Language != FormatStyle::LK_JavaScript && 1746 Style.Language != FormatStyle::LK_Java) { 1747 // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it 1748 // marks declarations and properties that need special formatting. 1749 switch (Current.Next->Tok.getObjCKeywordID()) { 1750 case tok::objc_interface: 1751 case tok::objc_implementation: 1752 case tok::objc_protocol: 1753 Current.setType(TT_ObjCDecl); 1754 break; 1755 case tok::objc_property: 1756 Current.setType(TT_ObjCProperty); 1757 break; 1758 default: 1759 break; 1760 } 1761 } else if (Current.is(tok::period)) { 1762 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 1763 if (PreviousNoComment && 1764 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 1765 Current.setType(TT_DesignatedInitializerPeriod); 1766 else if (Style.Language == FormatStyle::LK_Java && Current.Previous && 1767 Current.Previous->isOneOf(TT_JavaAnnotation, 1768 TT_LeadingJavaAnnotation)) { 1769 Current.setType(Current.Previous->getType()); 1770 } 1771 } else if (canBeObjCSelectorComponent(Current) && 1772 // FIXME(bug 36976): ObjC return types shouldn't use 1773 // TT_CastRParen. 1774 Current.Previous && Current.Previous->is(TT_CastRParen) && 1775 Current.Previous->MatchingParen && 1776 Current.Previous->MatchingParen->Previous && 1777 Current.Previous->MatchingParen->Previous->is( 1778 TT_ObjCMethodSpecifier)) { 1779 // This is the first part of an Objective-C selector name. (If there's no 1780 // colon after this, this is the only place which annotates the identifier 1781 // as a selector.) 1782 Current.setType(TT_SelectorName); 1783 } else if (Current.isOneOf(tok::identifier, tok::kw_const, tok::kw_noexcept, 1784 tok::kw_requires) && 1785 Current.Previous && 1786 !Current.Previous->isOneOf(tok::equal, tok::at) && 1787 Line.MightBeFunctionDecl && Contexts.size() == 1) { 1788 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1789 // function declaration have been found. 1790 Current.setType(TT_TrailingAnnotation); 1791 } else if ((Style.Language == FormatStyle::LK_Java || 1792 Style.Language == FormatStyle::LK_JavaScript) && 1793 Current.Previous) { 1794 if (Current.Previous->is(tok::at) && 1795 Current.isNot(Keywords.kw_interface)) { 1796 const FormatToken &AtToken = *Current.Previous; 1797 const FormatToken *Previous = AtToken.getPreviousNonComment(); 1798 if (!Previous || Previous->is(TT_LeadingJavaAnnotation)) 1799 Current.setType(TT_LeadingJavaAnnotation); 1800 else 1801 Current.setType(TT_JavaAnnotation); 1802 } else if (Current.Previous->is(tok::period) && 1803 Current.Previous->isOneOf(TT_JavaAnnotation, 1804 TT_LeadingJavaAnnotation)) { 1805 Current.setType(Current.Previous->getType()); 1806 } 1807 } 1808 } 1809 1810 /// Take a guess at whether \p Tok starts a name of a function or 1811 /// variable declaration. 1812 /// 1813 /// This is a heuristic based on whether \p Tok is an identifier following 1814 /// something that is likely a type. 1815 bool isStartOfName(const FormatToken &Tok) { 1816 if (Tok.isNot(tok::identifier) || !Tok.Previous) 1817 return false; 1818 1819 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof, 1820 Keywords.kw_as)) 1821 return false; 1822 if (Style.Language == FormatStyle::LK_JavaScript && 1823 Tok.Previous->is(Keywords.kw_in)) 1824 return false; 1825 1826 // Skip "const" as it does not have an influence on whether this is a name. 1827 FormatToken *PreviousNotConst = Tok.getPreviousNonComment(); 1828 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 1829 PreviousNotConst = PreviousNotConst->getPreviousNonComment(); 1830 1831 if (!PreviousNotConst) 1832 return false; 1833 1834 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 1835 PreviousNotConst->Previous && 1836 PreviousNotConst->Previous->is(tok::hash); 1837 1838 if (PreviousNotConst->is(TT_TemplateCloser)) 1839 return PreviousNotConst && PreviousNotConst->MatchingParen && 1840 PreviousNotConst->MatchingParen->Previous && 1841 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) && 1842 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 1843 1844 if (PreviousNotConst->is(tok::r_paren) && 1845 PreviousNotConst->is(TT_TypeDeclarationParen)) 1846 return true; 1847 1848 return (!IsPPKeyword && 1849 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) || 1850 PreviousNotConst->is(TT_PointerOrReference) || 1851 PreviousNotConst->isSimpleTypeSpecifier(); 1852 } 1853 1854 /// Determine whether ')' is ending a cast. 1855 bool rParenEndsCast(const FormatToken &Tok) { 1856 // C-style casts are only used in C++, C# and Java. 1857 if (!Style.isCSharp() && !Style.isCpp() && 1858 Style.Language != FormatStyle::LK_Java) 1859 return false; 1860 1861 // Empty parens aren't casts and there are no casts at the end of the line. 1862 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen) 1863 return false; 1864 1865 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 1866 if (LeftOfParens) { 1867 // If there is a closing parenthesis left of the current parentheses, 1868 // look past it as these might be chained casts. 1869 if (LeftOfParens->is(tok::r_paren)) { 1870 if (!LeftOfParens->MatchingParen || 1871 !LeftOfParens->MatchingParen->Previous) 1872 return false; 1873 LeftOfParens = LeftOfParens->MatchingParen->Previous; 1874 } 1875 1876 // If there is an identifier (or with a few exceptions a keyword) right 1877 // before the parentheses, this is unlikely to be a cast. 1878 if (LeftOfParens->Tok.getIdentifierInfo() && 1879 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case, 1880 tok::kw_delete)) 1881 return false; 1882 1883 // Certain other tokens right before the parentheses are also signals that 1884 // this cannot be a cast. 1885 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator, 1886 TT_TemplateCloser, tok::ellipsis)) 1887 return false; 1888 } 1889 1890 if (Tok.Next->is(tok::question)) 1891 return false; 1892 1893 // `foreach((A a, B b) in someList)` should not be seen as a cast. 1894 if (Tok.Next->is(Keywords.kw_in) && Style.isCSharp()) 1895 return false; 1896 1897 // Functions which end with decorations like volatile, noexcept are unlikely 1898 // to be casts. 1899 if (Tok.Next->isOneOf(tok::kw_noexcept, tok::kw_volatile, tok::kw_const, 1900 tok::kw_requires, tok::kw_throw, tok::arrow, 1901 Keywords.kw_override, Keywords.kw_final) || 1902 isCpp11AttributeSpecifier(*Tok.Next)) 1903 return false; 1904 1905 // As Java has no function types, a "(" after the ")" likely means that this 1906 // is a cast. 1907 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren)) 1908 return true; 1909 1910 // If a (non-string) literal follows, this is likely a cast. 1911 if (Tok.Next->isNot(tok::string_literal) && 1912 (Tok.Next->Tok.isLiteral() || 1913 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 1914 return true; 1915 1916 // Heuristically try to determine whether the parentheses contain a type. 1917 auto IsQualifiedPointerOrReference = [](FormatToken *T) { 1918 // This is used to handle cases such as x = (foo *const)&y; 1919 assert(!T->isSimpleTypeSpecifier() && "Should have already been checked"); 1920 // Strip trailing qualifiers such as const or volatile when checking 1921 // whether the parens could be a cast to a pointer/reference type. 1922 while (T) { 1923 if (T->is(TT_AttributeParen)) { 1924 // Handle `x = (foo *__attribute__((foo)))&v;`: 1925 if (T->MatchingParen && T->MatchingParen->Previous && 1926 T->MatchingParen->Previous->is(tok::kw___attribute)) { 1927 T = T->MatchingParen->Previous->Previous; 1928 continue; 1929 } 1930 } else if (T->is(TT_AttributeSquare)) { 1931 // Handle `x = (foo *[[clang::foo]])&v;`: 1932 if (T->MatchingParen && T->MatchingParen->Previous) { 1933 T = T->MatchingParen->Previous; 1934 continue; 1935 } 1936 } else if (T->canBePointerOrReferenceQualifier()) { 1937 T = T->Previous; 1938 continue; 1939 } 1940 break; 1941 } 1942 return T && T->is(TT_PointerOrReference); 1943 }; 1944 bool ParensAreType = 1945 !Tok.Previous || 1946 Tok.Previous->isOneOf(TT_TemplateCloser, TT_TypeDeclarationParen) || 1947 Tok.Previous->isSimpleTypeSpecifier() || 1948 IsQualifiedPointerOrReference(Tok.Previous); 1949 bool ParensCouldEndDecl = 1950 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater); 1951 if (ParensAreType && !ParensCouldEndDecl) 1952 return true; 1953 1954 // At this point, we heuristically assume that there are no casts at the 1955 // start of the line. We assume that we have found most cases where there 1956 // are by the logic above, e.g. "(void)x;". 1957 if (!LeftOfParens) 1958 return false; 1959 1960 // Certain token types inside the parentheses mean that this can't be a 1961 // cast. 1962 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok; 1963 Token = Token->Next) 1964 if (Token->is(TT_BinaryOperator)) 1965 return false; 1966 1967 // If the following token is an identifier or 'this', this is a cast. All 1968 // cases where this can be something else are handled above. 1969 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this)) 1970 return true; 1971 1972 // Look for a cast `( x ) (`. 1973 if (Tok.Next->is(tok::l_paren) && Tok.Previous && Tok.Previous->Previous) { 1974 if (Tok.Previous->is(tok::identifier) && 1975 Tok.Previous->Previous->is(tok::l_paren)) 1976 return true; 1977 } 1978 1979 if (!Tok.Next->Next) 1980 return false; 1981 1982 // If the next token after the parenthesis is a unary operator, assume 1983 // that this is cast, unless there are unexpected tokens inside the 1984 // parenthesis. 1985 bool NextIsUnary = 1986 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star); 1987 if (!NextIsUnary || Tok.Next->is(tok::plus) || 1988 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) 1989 return false; 1990 // Search for unexpected tokens. 1991 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen; 1992 Prev = Prev->Previous) { 1993 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon)) 1994 return false; 1995 } 1996 return true; 1997 } 1998 1999 /// Return the type of the given token assuming it is * or &. 2000 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 2001 bool InTemplateArgument) { 2002 if (Style.Language == FormatStyle::LK_JavaScript) 2003 return TT_BinaryOperator; 2004 2005 // && in C# must be a binary operator. 2006 if (Style.isCSharp() && Tok.is(tok::ampamp)) 2007 return TT_BinaryOperator; 2008 2009 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 2010 if (!PrevToken) 2011 return TT_UnaryOperator; 2012 2013 const FormatToken *NextToken = Tok.getNextNonComment(); 2014 if (!NextToken || 2015 NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_noexcept) || 2016 NextToken->canBePointerOrReferenceQualifier() || 2017 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) 2018 return TT_PointerOrReference; 2019 2020 if (PrevToken->is(tok::coloncolon)) 2021 return TT_PointerOrReference; 2022 2023 if (PrevToken->is(tok::r_paren) && PrevToken->is(TT_TypeDeclarationParen)) 2024 return TT_PointerOrReference; 2025 2026 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 2027 tok::comma, tok::semi, tok::kw_return, tok::colon, 2028 tok::kw_co_return, tok::kw_co_await, 2029 tok::kw_co_yield, tok::equal, tok::kw_delete, 2030 tok::kw_sizeof, tok::kw_throw) || 2031 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr, 2032 TT_UnaryOperator, TT_CastRParen)) 2033 return TT_UnaryOperator; 2034 2035 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare)) 2036 return TT_PointerOrReference; 2037 if (NextToken->is(tok::kw_operator) && !IsExpression) 2038 return TT_PointerOrReference; 2039 if (NextToken->isOneOf(tok::comma, tok::semi)) 2040 return TT_PointerOrReference; 2041 2042 if (PrevToken->Tok.isLiteral() || 2043 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 2044 tok::kw_false, tok::r_brace) || 2045 NextToken->Tok.isLiteral() || 2046 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 2047 NextToken->isUnaryOperator() || 2048 // If we know we're in a template argument, there are no named 2049 // declarations. Thus, having an identifier on the right-hand side 2050 // indicates a binary operator. 2051 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 2052 return TT_BinaryOperator; 2053 2054 // "&&(" is quite unlikely to be two successive unary "&". 2055 if (Tok.is(tok::ampamp) && NextToken->is(tok::l_paren)) 2056 return TT_BinaryOperator; 2057 2058 // This catches some cases where evaluation order is used as control flow: 2059 // aaa && aaa->f(); 2060 if (NextToken->Tok.isAnyIdentifier()) { 2061 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 2062 if (NextNextToken && NextNextToken->is(tok::arrow)) 2063 return TT_BinaryOperator; 2064 } 2065 2066 // It is very unlikely that we are going to find a pointer or reference type 2067 // definition on the RHS of an assignment. 2068 if (IsExpression && !Contexts.back().CaretFound) 2069 return TT_BinaryOperator; 2070 2071 return TT_PointerOrReference; 2072 } 2073 2074 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 2075 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 2076 if (!PrevToken) 2077 return TT_UnaryOperator; 2078 2079 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator)) 2080 // This must be a sequence of leading unary operators. 2081 return TT_UnaryOperator; 2082 2083 // Use heuristics to recognize unary operators. 2084 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 2085 tok::question, tok::colon, tok::kw_return, 2086 tok::kw_case, tok::at, tok::l_brace, tok::kw_throw, 2087 tok::kw_co_return, tok::kw_co_yield)) 2088 return TT_UnaryOperator; 2089 2090 // There can't be two consecutive binary operators. 2091 if (PrevToken->is(TT_BinaryOperator)) 2092 return TT_UnaryOperator; 2093 2094 // Fall back to marking the token as binary operator. 2095 return TT_BinaryOperator; 2096 } 2097 2098 /// Determine whether ++/-- are pre- or post-increments/-decrements. 2099 TokenType determineIncrementUsage(const FormatToken &Tok) { 2100 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 2101 if (!PrevToken || PrevToken->is(TT_CastRParen)) 2102 return TT_UnaryOperator; 2103 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 2104 return TT_TrailingUnaryOperator; 2105 2106 return TT_UnaryOperator; 2107 } 2108 2109 SmallVector<Context, 8> Contexts; 2110 2111 const FormatStyle &Style; 2112 AnnotatedLine &Line; 2113 FormatToken *CurrentToken; 2114 bool AutoFound; 2115 const AdditionalKeywords &Keywords; 2116 2117 // Set of "<" tokens that do not open a template parameter list. If parseAngle 2118 // determines that a specific token can't be a template opener, it will make 2119 // same decision irrespective of the decisions for tokens leading up to it. 2120 // Store this information to prevent this from causing exponential runtime. 2121 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess; 2122 }; 2123 2124 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1; 2125 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 2126 2127 /// Parses binary expressions by inserting fake parenthesis based on 2128 /// operator precedence. 2129 class ExpressionParser { 2130 public: 2131 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords, 2132 AnnotatedLine &Line) 2133 : Style(Style), Keywords(Keywords), Current(Line.First) {} 2134 2135 /// Parse expressions with the given operator precedence. 2136 void parse(int Precedence = 0) { 2137 // Skip 'return' and ObjC selector colons as they are not part of a binary 2138 // expression. 2139 while (Current && (Current->is(tok::kw_return) || 2140 (Current->is(tok::colon) && 2141 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) 2142 next(); 2143 2144 if (!Current || Precedence > PrecedenceArrowAndPeriod) 2145 return; 2146 2147 // Conditional expressions need to be parsed separately for proper nesting. 2148 if (Precedence == prec::Conditional) { 2149 parseConditionalExpr(); 2150 return; 2151 } 2152 2153 // Parse unary operators, which all have a higher precedence than binary 2154 // operators. 2155 if (Precedence == PrecedenceUnaryOperator) { 2156 parseUnaryOperator(); 2157 return; 2158 } 2159 2160 FormatToken *Start = Current; 2161 FormatToken *LatestOperator = nullptr; 2162 unsigned OperatorIndex = 0; 2163 2164 while (Current) { 2165 // Consume operators with higher precedence. 2166 parse(Precedence + 1); 2167 2168 int CurrentPrecedence = getCurrentPrecedence(); 2169 2170 if (Current && Current->is(TT_SelectorName) && 2171 Precedence == CurrentPrecedence) { 2172 if (LatestOperator) 2173 addFakeParenthesis(Start, prec::Level(Precedence)); 2174 Start = Current; 2175 } 2176 2177 // At the end of the line or when an operator with higher precedence is 2178 // found, insert fake parenthesis and return. 2179 if (!Current || 2180 (Current->closesScope() && 2181 (Current->MatchingParen || Current->is(TT_TemplateString))) || 2182 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) || 2183 (CurrentPrecedence == prec::Conditional && 2184 Precedence == prec::Assignment && Current->is(tok::colon))) { 2185 break; 2186 } 2187 2188 // Consume scopes: (), [], <> and {} 2189 if (Current->opensScope()) { 2190 // In fragment of a JavaScript template string can look like '}..${' and 2191 // thus close a scope and open a new one at the same time. 2192 while (Current && (!Current->closesScope() || Current->opensScope())) { 2193 next(); 2194 parse(); 2195 } 2196 next(); 2197 } else { 2198 // Operator found. 2199 if (CurrentPrecedence == Precedence) { 2200 if (LatestOperator) 2201 LatestOperator->NextOperator = Current; 2202 LatestOperator = Current; 2203 Current->OperatorIndex = OperatorIndex; 2204 ++OperatorIndex; 2205 } 2206 next(/*SkipPastLeadingComments=*/Precedence > 0); 2207 } 2208 } 2209 2210 if (LatestOperator && (Current || Precedence > 0)) { 2211 // LatestOperator->LastOperator = true; 2212 if (Precedence == PrecedenceArrowAndPeriod) { 2213 // Call expressions don't have a binary operator precedence. 2214 addFakeParenthesis(Start, prec::Unknown); 2215 } else { 2216 addFakeParenthesis(Start, prec::Level(Precedence)); 2217 } 2218 } 2219 } 2220 2221 private: 2222 /// Gets the precedence (+1) of the given token for binary operators 2223 /// and other tokens that we treat like binary operators. 2224 int getCurrentPrecedence() { 2225 if (Current) { 2226 const FormatToken *NextNonComment = Current->getNextNonComment(); 2227 if (Current->is(TT_ConditionalExpr)) 2228 return prec::Conditional; 2229 if (NextNonComment && Current->is(TT_SelectorName) && 2230 (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) || 2231 ((Style.Language == FormatStyle::LK_Proto || 2232 Style.Language == FormatStyle::LK_TextProto) && 2233 NextNonComment->is(tok::less)))) 2234 return prec::Assignment; 2235 if (Current->is(TT_JsComputedPropertyName)) 2236 return prec::Assignment; 2237 if (Current->is(TT_LambdaArrow)) 2238 return prec::Comma; 2239 if (Current->is(TT_FatArrow)) 2240 return prec::Assignment; 2241 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) || 2242 (Current->is(tok::comment) && NextNonComment && 2243 NextNonComment->is(TT_SelectorName))) 2244 return 0; 2245 if (Current->is(TT_RangeBasedForLoopColon)) 2246 return prec::Comma; 2247 if ((Style.Language == FormatStyle::LK_Java || 2248 Style.Language == FormatStyle::LK_JavaScript) && 2249 Current->is(Keywords.kw_instanceof)) 2250 return prec::Relational; 2251 if (Style.Language == FormatStyle::LK_JavaScript && 2252 Current->isOneOf(Keywords.kw_in, Keywords.kw_as)) 2253 return prec::Relational; 2254 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma)) 2255 return Current->getPrecedence(); 2256 if (Current->isOneOf(tok::period, tok::arrow)) 2257 return PrecedenceArrowAndPeriod; 2258 if ((Style.Language == FormatStyle::LK_Java || 2259 Style.Language == FormatStyle::LK_JavaScript) && 2260 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements, 2261 Keywords.kw_throws)) 2262 return 0; 2263 } 2264 return -1; 2265 } 2266 2267 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 2268 Start->FakeLParens.push_back(Precedence); 2269 if (Precedence > prec::Unknown) 2270 Start->StartsBinaryExpression = true; 2271 if (Current) { 2272 FormatToken *Previous = Current->Previous; 2273 while (Previous->is(tok::comment) && Previous->Previous) 2274 Previous = Previous->Previous; 2275 ++Previous->FakeRParens; 2276 if (Precedence > prec::Unknown) 2277 Previous->EndsBinaryExpression = true; 2278 } 2279 } 2280 2281 /// Parse unary operator expressions and surround them with fake 2282 /// parentheses if appropriate. 2283 void parseUnaryOperator() { 2284 llvm::SmallVector<FormatToken *, 2> Tokens; 2285 while (Current && Current->is(TT_UnaryOperator)) { 2286 Tokens.push_back(Current); 2287 next(); 2288 } 2289 parse(PrecedenceArrowAndPeriod); 2290 for (FormatToken *Token : llvm::reverse(Tokens)) 2291 // The actual precedence doesn't matter. 2292 addFakeParenthesis(Token, prec::Unknown); 2293 } 2294 2295 void parseConditionalExpr() { 2296 while (Current && Current->isTrailingComment()) { 2297 next(); 2298 } 2299 FormatToken *Start = Current; 2300 parse(prec::LogicalOr); 2301 if (!Current || !Current->is(tok::question)) 2302 return; 2303 next(); 2304 parse(prec::Assignment); 2305 if (!Current || Current->isNot(TT_ConditionalExpr)) 2306 return; 2307 next(); 2308 parse(prec::Assignment); 2309 addFakeParenthesis(Start, prec::Conditional); 2310 } 2311 2312 void next(bool SkipPastLeadingComments = true) { 2313 if (Current) 2314 Current = Current->Next; 2315 while (Current && 2316 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 2317 Current->isTrailingComment()) 2318 Current = Current->Next; 2319 } 2320 2321 const FormatStyle &Style; 2322 const AdditionalKeywords &Keywords; 2323 FormatToken *Current; 2324 }; 2325 2326 } // end anonymous namespace 2327 2328 void TokenAnnotator::setCommentLineLevels( 2329 SmallVectorImpl<AnnotatedLine *> &Lines) { 2330 const AnnotatedLine *NextNonCommentLine = nullptr; 2331 for (AnnotatedLine *AL : llvm::reverse(Lines)) { 2332 bool CommentLine = true; 2333 for (const FormatToken *Tok = AL->First; Tok; Tok = Tok->Next) { 2334 if (!Tok->is(tok::comment)) { 2335 CommentLine = false; 2336 break; 2337 } 2338 } 2339 2340 // If the comment is currently aligned with the line immediately following 2341 // it, that's probably intentional and we should keep it. 2342 if (NextNonCommentLine && CommentLine && 2343 NextNonCommentLine->First->NewlinesBefore <= 1 && 2344 NextNonCommentLine->First->OriginalColumn == 2345 AL->First->OriginalColumn) { 2346 // Align comments for preprocessor lines with the # in column 0 if 2347 // preprocessor lines are not indented. Otherwise, align with the next 2348 // line. 2349 AL->Level = 2350 (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash && 2351 (NextNonCommentLine->Type == LT_PreprocessorDirective || 2352 NextNonCommentLine->Type == LT_ImportStatement)) 2353 ? 0 2354 : NextNonCommentLine->Level; 2355 } else { 2356 NextNonCommentLine = AL->First->isNot(tok::r_brace) ? AL : nullptr; 2357 } 2358 2359 setCommentLineLevels(AL->Children); 2360 } 2361 } 2362 2363 static unsigned maxNestingDepth(const AnnotatedLine &Line) { 2364 unsigned Result = 0; 2365 for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next) 2366 Result = std::max(Result, Tok->NestingLevel); 2367 return Result; 2368 } 2369 2370 void TokenAnnotator::annotate(AnnotatedLine &Line) { 2371 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 2372 E = Line.Children.end(); 2373 I != E; ++I) { 2374 annotate(**I); 2375 } 2376 AnnotatingParser Parser(Style, Line, Keywords); 2377 Line.Type = Parser.parseLine(); 2378 2379 // With very deep nesting, ExpressionParser uses lots of stack and the 2380 // formatting algorithm is very slow. We're not going to do a good job here 2381 // anyway - it's probably generated code being formatted by mistake. 2382 // Just skip the whole line. 2383 if (maxNestingDepth(Line) > 50) 2384 Line.Type = LT_Invalid; 2385 2386 if (Line.Type == LT_Invalid) 2387 return; 2388 2389 ExpressionParser ExprParser(Style, Keywords, Line); 2390 ExprParser.parse(); 2391 2392 if (Line.startsWith(TT_ObjCMethodSpecifier)) 2393 Line.Type = LT_ObjCMethodDecl; 2394 else if (Line.startsWith(TT_ObjCDecl)) 2395 Line.Type = LT_ObjCDecl; 2396 else if (Line.startsWith(TT_ObjCProperty)) 2397 Line.Type = LT_ObjCProperty; 2398 2399 Line.First->SpacesRequiredBefore = 1; 2400 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 2401 } 2402 2403 // This function heuristically determines whether 'Current' starts the name of a 2404 // function declaration. 2405 static bool isFunctionDeclarationName(bool IsCpp, const FormatToken &Current, 2406 const AnnotatedLine &Line) { 2407 auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * { 2408 for (; Next; Next = Next->Next) { 2409 if (Next->is(TT_OverloadedOperatorLParen)) 2410 return Next; 2411 if (Next->is(TT_OverloadedOperator)) 2412 continue; 2413 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) { 2414 // For 'new[]' and 'delete[]'. 2415 if (Next->Next && 2416 Next->Next->startsSequence(tok::l_square, tok::r_square)) 2417 Next = Next->Next->Next; 2418 continue; 2419 } 2420 if (Next->startsSequence(tok::l_square, tok::r_square)) { 2421 // For operator[](). 2422 Next = Next->Next; 2423 continue; 2424 } 2425 if ((Next->isSimpleTypeSpecifier() || Next->is(tok::identifier)) && 2426 Next->Next && Next->Next->isOneOf(tok::star, tok::amp, tok::ampamp)) { 2427 // For operator void*(), operator char*(), operator Foo*(). 2428 Next = Next->Next; 2429 continue; 2430 } 2431 if (Next->is(TT_TemplateOpener) && Next->MatchingParen) { 2432 Next = Next->MatchingParen; 2433 continue; 2434 } 2435 2436 break; 2437 } 2438 return nullptr; 2439 }; 2440 2441 // Find parentheses of parameter list. 2442 const FormatToken *Next = Current.Next; 2443 if (Current.is(tok::kw_operator)) { 2444 if (Current.Previous && Current.Previous->is(tok::coloncolon)) 2445 return false; 2446 Next = skipOperatorName(Next); 2447 } else { 2448 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0) 2449 return false; 2450 for (; Next; Next = Next->Next) { 2451 if (Next->is(TT_TemplateOpener)) { 2452 Next = Next->MatchingParen; 2453 } else if (Next->is(tok::coloncolon)) { 2454 Next = Next->Next; 2455 if (!Next) 2456 return false; 2457 if (Next->is(tok::kw_operator)) { 2458 Next = skipOperatorName(Next->Next); 2459 break; 2460 } 2461 if (!Next->is(tok::identifier)) 2462 return false; 2463 } else if (Next->is(tok::l_paren)) { 2464 break; 2465 } else { 2466 return false; 2467 } 2468 } 2469 } 2470 2471 // Check whether parameter list can belong to a function declaration. 2472 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen) 2473 return false; 2474 // If the lines ends with "{", this is likely a function definition. 2475 if (Line.Last->is(tok::l_brace)) 2476 return true; 2477 if (Next->Next == Next->MatchingParen) 2478 return true; // Empty parentheses. 2479 // If there is an &/&& after the r_paren, this is likely a function. 2480 if (Next->MatchingParen->Next && 2481 Next->MatchingParen->Next->is(TT_PointerOrReference)) 2482 return true; 2483 2484 // Check for K&R C function definitions (and C++ function definitions with 2485 // unnamed parameters), e.g.: 2486 // int f(i) 2487 // { 2488 // return i + 1; 2489 // } 2490 // bool g(size_t = 0, bool b = false) 2491 // { 2492 // return !b; 2493 // } 2494 if (IsCpp && Next->Next && Next->Next->is(tok::identifier) && 2495 !Line.endsWith(tok::semi)) 2496 return true; 2497 2498 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen; 2499 Tok = Tok->Next) { 2500 if (Tok->is(TT_TypeDeclarationParen)) 2501 return true; 2502 if (Tok->isOneOf(tok::l_paren, TT_TemplateOpener) && Tok->MatchingParen) { 2503 Tok = Tok->MatchingParen; 2504 continue; 2505 } 2506 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 2507 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis)) 2508 return true; 2509 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) || 2510 Tok->Tok.isLiteral()) 2511 return false; 2512 } 2513 return false; 2514 } 2515 2516 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const { 2517 assert(Line.MightBeFunctionDecl); 2518 2519 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel || 2520 Style.AlwaysBreakAfterReturnType == 2521 FormatStyle::RTBS_TopLevelDefinitions) && 2522 Line.Level > 0) 2523 return false; 2524 2525 switch (Style.AlwaysBreakAfterReturnType) { 2526 case FormatStyle::RTBS_None: 2527 return false; 2528 case FormatStyle::RTBS_All: 2529 case FormatStyle::RTBS_TopLevel: 2530 return true; 2531 case FormatStyle::RTBS_AllDefinitions: 2532 case FormatStyle::RTBS_TopLevelDefinitions: 2533 return Line.mightBeFunctionDefinition(); 2534 } 2535 2536 return false; 2537 } 2538 2539 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 2540 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 2541 E = Line.Children.end(); 2542 I != E; ++I) { 2543 calculateFormattingInformation(**I); 2544 } 2545 2546 Line.First->TotalLength = 2547 Line.First->IsMultiline ? Style.ColumnLimit 2548 : Line.FirstStartColumn + Line.First->ColumnWidth; 2549 FormatToken *Current = Line.First->Next; 2550 bool InFunctionDecl = Line.MightBeFunctionDecl; 2551 bool AlignArrayOfStructures = 2552 (Style.AlignArrayOfStructures != FormatStyle::AIAS_None && 2553 Line.Type == LT_ArrayOfStructInitializer); 2554 if (AlignArrayOfStructures) 2555 calculateArrayInitializerColumnList(Line); 2556 2557 while (Current) { 2558 if (isFunctionDeclarationName(Style.isCpp(), *Current, Line)) 2559 Current->setType(TT_FunctionDeclarationName); 2560 if (Current->is(TT_LineComment)) { 2561 if (Current->Previous->is(BK_BracedInit) && 2562 Current->Previous->opensScope()) 2563 Current->SpacesRequiredBefore = 2564 (Style.Cpp11BracedListStyle && !Style.SpacesInParentheses) ? 0 : 1; 2565 else 2566 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 2567 2568 // If we find a trailing comment, iterate backwards to determine whether 2569 // it seems to relate to a specific parameter. If so, break before that 2570 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 2571 // to the previous line in: 2572 // SomeFunction(a, 2573 // b, // comment 2574 // c); 2575 if (!Current->HasUnescapedNewline) { 2576 for (FormatToken *Parameter = Current->Previous; Parameter; 2577 Parameter = Parameter->Previous) { 2578 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 2579 break; 2580 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 2581 if (!Parameter->Previous->is(TT_CtorInitializerComma) && 2582 Parameter->HasUnescapedNewline) 2583 Parameter->MustBreakBefore = true; 2584 break; 2585 } 2586 } 2587 } 2588 } else if (Current->SpacesRequiredBefore == 0 && 2589 spaceRequiredBefore(Line, *Current)) { 2590 Current->SpacesRequiredBefore = 1; 2591 } 2592 2593 Current->MustBreakBefore = 2594 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 2595 2596 if (!Current->MustBreakBefore && InFunctionDecl && 2597 Current->is(TT_FunctionDeclarationName)) 2598 Current->MustBreakBefore = mustBreakForReturnType(Line); 2599 2600 Current->CanBreakBefore = 2601 Current->MustBreakBefore || canBreakBefore(Line, *Current); 2602 unsigned ChildSize = 0; 2603 if (Current->Previous->Children.size() == 1) { 2604 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 2605 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 2606 : LastOfChild.TotalLength + 1; 2607 } 2608 const FormatToken *Prev = Current->Previous; 2609 if (Current->MustBreakBefore || Prev->Children.size() > 1 || 2610 (Prev->Children.size() == 1 && 2611 Prev->Children[0]->First->MustBreakBefore) || 2612 Current->IsMultiline) 2613 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit; 2614 else 2615 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth + 2616 ChildSize + Current->SpacesRequiredBefore; 2617 2618 if (Current->is(TT_CtorInitializerColon)) 2619 InFunctionDecl = false; 2620 2621 // FIXME: Only calculate this if CanBreakBefore is true once static 2622 // initializers etc. are sorted out. 2623 // FIXME: Move magic numbers to a better place. 2624 2625 // Reduce penalty for aligning ObjC method arguments using the colon 2626 // alignment as this is the canonical way (still prefer fitting everything 2627 // into one line if possible). Trying to fit a whole expression into one 2628 // line should not force other line breaks (e.g. when ObjC method 2629 // expression is a part of other expression). 2630 Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl); 2631 if (Style.Language == FormatStyle::LK_ObjC && 2632 Current->is(TT_SelectorName) && Current->ParameterIndex > 0) { 2633 if (Current->ParameterIndex == 1) 2634 Current->SplitPenalty += 5 * Current->BindingStrength; 2635 } else { 2636 Current->SplitPenalty += 20 * Current->BindingStrength; 2637 } 2638 2639 Current = Current->Next; 2640 } 2641 2642 calculateUnbreakableTailLengths(Line); 2643 unsigned IndentLevel = Line.Level; 2644 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 2645 if (Current->Role) 2646 Current->Role->precomputeFormattingInfos(Current); 2647 if (Current->MatchingParen && 2648 Current->MatchingParen->opensBlockOrBlockTypeList(Style) && 2649 IndentLevel > 0) { 2650 --IndentLevel; 2651 } 2652 Current->IndentLevel = IndentLevel; 2653 if (Current->opensBlockOrBlockTypeList(Style)) 2654 ++IndentLevel; 2655 } 2656 2657 LLVM_DEBUG({ printDebugInfo(Line); }); 2658 } 2659 2660 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 2661 unsigned UnbreakableTailLength = 0; 2662 FormatToken *Current = Line.Last; 2663 while (Current) { 2664 Current->UnbreakableTailLength = UnbreakableTailLength; 2665 if (Current->CanBreakBefore || 2666 Current->isOneOf(tok::comment, tok::string_literal)) { 2667 UnbreakableTailLength = 0; 2668 } else { 2669 UnbreakableTailLength += 2670 Current->ColumnWidth + Current->SpacesRequiredBefore; 2671 } 2672 Current = Current->Previous; 2673 } 2674 } 2675 2676 void TokenAnnotator::calculateArrayInitializerColumnList(AnnotatedLine &Line) { 2677 if (Line.First == Line.Last) { 2678 return; 2679 } 2680 auto *CurrentToken = Line.First; 2681 CurrentToken->ArrayInitializerLineStart = true; 2682 unsigned Depth = 0; 2683 while (CurrentToken != nullptr && CurrentToken != Line.Last) { 2684 if (CurrentToken->is(tok::l_brace)) { 2685 CurrentToken->IsArrayInitializer = true; 2686 if (CurrentToken->Next != nullptr) 2687 CurrentToken->Next->MustBreakBefore = true; 2688 CurrentToken = 2689 calculateInitializerColumnList(Line, CurrentToken->Next, Depth + 1); 2690 } else { 2691 CurrentToken = CurrentToken->Next; 2692 } 2693 } 2694 } 2695 2696 FormatToken *TokenAnnotator::calculateInitializerColumnList( 2697 AnnotatedLine &Line, FormatToken *CurrentToken, unsigned Depth) { 2698 while (CurrentToken != nullptr && CurrentToken != Line.Last) { 2699 if (CurrentToken->is(tok::l_brace)) 2700 ++Depth; 2701 else if (CurrentToken->is(tok::r_brace)) 2702 --Depth; 2703 if (Depth == 2 && CurrentToken->isOneOf(tok::l_brace, tok::comma)) { 2704 CurrentToken = CurrentToken->Next; 2705 if (CurrentToken == nullptr) 2706 break; 2707 CurrentToken->StartsColumn = true; 2708 CurrentToken = CurrentToken->Previous; 2709 } 2710 CurrentToken = CurrentToken->Next; 2711 } 2712 return CurrentToken; 2713 } 2714 2715 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 2716 const FormatToken &Tok, 2717 bool InFunctionDecl) { 2718 const FormatToken &Left = *Tok.Previous; 2719 const FormatToken &Right = Tok; 2720 2721 if (Left.is(tok::semi)) 2722 return 0; 2723 2724 if (Style.Language == FormatStyle::LK_Java) { 2725 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws)) 2726 return 1; 2727 if (Right.is(Keywords.kw_implements)) 2728 return 2; 2729 if (Left.is(tok::comma) && Left.NestingLevel == 0) 2730 return 3; 2731 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2732 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma)) 2733 return 100; 2734 if (Left.is(TT_JsTypeColon)) 2735 return 35; 2736 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2737 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2738 return 100; 2739 // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()". 2740 if (Left.opensScope() && Right.closesScope()) 2741 return 200; 2742 } 2743 2744 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2745 return 1; 2746 if (Right.is(tok::l_square)) { 2747 if (Style.Language == FormatStyle::LK_Proto) 2748 return 1; 2749 if (Left.is(tok::r_square)) 2750 return 200; 2751 // Slightly prefer formatting local lambda definitions like functions. 2752 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal)) 2753 return 35; 2754 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2755 TT_ArrayInitializerLSquare, 2756 TT_DesignatedInitializerLSquare, TT_AttributeSquare)) 2757 return 500; 2758 } 2759 2760 if (Left.is(tok::coloncolon) || 2761 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 2762 return 500; 2763 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2764 Right.is(tok::kw_operator)) { 2765 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 2766 return 3; 2767 if (Left.is(TT_StartOfName)) 2768 return 110; 2769 if (InFunctionDecl && Right.NestingLevel == 0) 2770 return Style.PenaltyReturnTypeOnItsOwnLine; 2771 return 200; 2772 } 2773 if (Right.is(TT_PointerOrReference)) 2774 return 190; 2775 if (Right.is(TT_LambdaArrow)) 2776 return 110; 2777 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 2778 return 160; 2779 if (Left.is(TT_CastRParen)) 2780 return 100; 2781 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 2782 return 5000; 2783 if (Left.is(tok::comment)) 2784 return 1000; 2785 2786 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon, 2787 TT_CtorInitializerColon)) 2788 return 2; 2789 2790 if (Right.isMemberAccess()) { 2791 // Breaking before the "./->" of a chained call/member access is reasonably 2792 // cheap, as formatting those with one call per line is generally 2793 // desirable. In particular, it should be cheaper to break before the call 2794 // than it is to break inside a call's parameters, which could lead to weird 2795 // "hanging" indents. The exception is the very last "./->" to support this 2796 // frequent pattern: 2797 // 2798 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc( 2799 // dddddddd); 2800 // 2801 // which might otherwise be blown up onto many lines. Here, clang-format 2802 // won't produce "hanging" indents anyway as there is no other trailing 2803 // call. 2804 // 2805 // Also apply higher penalty is not a call as that might lead to a wrapping 2806 // like: 2807 // 2808 // aaaaaaa 2809 // .aaaaaaaaa.bbbbbbbb(cccccccc); 2810 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope() 2811 ? 150 2812 : 35; 2813 } 2814 2815 if (Right.is(TT_TrailingAnnotation) && 2816 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 2817 // Moving trailing annotations to the next line is fine for ObjC method 2818 // declarations. 2819 if (Line.startsWith(TT_ObjCMethodSpecifier)) 2820 return 10; 2821 // Generally, breaking before a trailing annotation is bad unless it is 2822 // function-like. It seems to be especially preferable to keep standard 2823 // annotations (i.e. "const", "final" and "override") on the same line. 2824 // Use a slightly higher penalty after ")" so that annotations like 2825 // "const override" are kept together. 2826 bool is_short_annotation = Right.TokenText.size() < 10; 2827 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 2828 } 2829 2830 // In for-loops, prefer breaking at ',' and ';'. 2831 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal)) 2832 return 4; 2833 2834 // In Objective-C method expressions, prefer breaking before "param:" over 2835 // breaking after it. 2836 if (Right.is(TT_SelectorName)) 2837 return 0; 2838 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr)) 2839 return Line.MightBeFunctionDecl ? 50 : 500; 2840 2841 // In Objective-C type declarations, avoid breaking after the category's 2842 // open paren (we'll prefer breaking after the protocol list's opening 2843 // angle bracket, if present). 2844 if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous && 2845 Left.Previous->isOneOf(tok::identifier, tok::greater)) 2846 return 500; 2847 2848 if (Left.is(tok::l_paren) && InFunctionDecl && 2849 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) 2850 return 100; 2851 if (Left.is(tok::l_paren) && Left.Previous && 2852 (Left.Previous->is(tok::kw_for) || Left.Previous->isIf())) 2853 return 1000; 2854 if (Left.is(tok::equal) && InFunctionDecl) 2855 return 110; 2856 if (Right.is(tok::r_brace)) 2857 return 1; 2858 if (Left.is(TT_TemplateOpener)) 2859 return 100; 2860 if (Left.opensScope()) { 2861 // If we aren't aligning after opening parens/braces we can always break 2862 // here unless the style does not want us to place all arguments on the 2863 // next line. 2864 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign && 2865 (Left.ParameterCount <= 1 || Style.AllowAllArgumentsOnNextLine)) 2866 return 0; 2867 if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle) 2868 return 19; 2869 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 2870 : 19; 2871 } 2872 if (Left.is(TT_JavaAnnotation)) 2873 return 50; 2874 2875 if (Left.is(TT_UnaryOperator)) 2876 return 60; 2877 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous && 2878 Left.Previous->isLabelString() && 2879 (Left.NextOperator || Left.OperatorIndex != 0)) 2880 return 50; 2881 if (Right.is(tok::plus) && Left.isLabelString() && 2882 (Right.NextOperator || Right.OperatorIndex != 0)) 2883 return 25; 2884 if (Left.is(tok::comma)) 2885 return 1; 2886 if (Right.is(tok::lessless) && Left.isLabelString() && 2887 (Right.NextOperator || Right.OperatorIndex != 1)) 2888 return 25; 2889 if (Right.is(tok::lessless)) { 2890 // Breaking at a << is really cheap. 2891 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0) 2892 // Slightly prefer to break before the first one in log-like statements. 2893 return 2; 2894 return 1; 2895 } 2896 if (Left.ClosesTemplateDeclaration) 2897 return Style.PenaltyBreakTemplateDeclaration; 2898 if (Left.is(TT_ConditionalExpr)) 2899 return prec::Conditional; 2900 prec::Level Level = Left.getPrecedence(); 2901 if (Level == prec::Unknown) 2902 Level = Right.getPrecedence(); 2903 if (Level == prec::Assignment) 2904 return Style.PenaltyBreakAssignment; 2905 if (Level != prec::Unknown) 2906 return Level; 2907 2908 return 3; 2909 } 2910 2911 bool TokenAnnotator::spaceRequiredBeforeParens(const FormatToken &Right) const { 2912 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always || 2913 (Style.SpaceBeforeParensOptions.BeforeNonEmptyParentheses && 2914 Right.ParameterCount > 0); 2915 } 2916 2917 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 2918 const FormatToken &Left, 2919 const FormatToken &Right) { 2920 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 2921 return true; 2922 if (Style.isJson() && Left.is(tok::string_literal) && Right.is(tok::colon)) 2923 return false; 2924 if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java) 2925 return true; 2926 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 2927 Left.Tok.getObjCKeywordID() == tok::objc_property) 2928 return true; 2929 if (Right.is(tok::hashhash)) 2930 return Left.is(tok::hash); 2931 if (Left.isOneOf(tok::hashhash, tok::hash)) 2932 return Right.is(tok::hash); 2933 if ((Left.is(tok::l_paren) && Right.is(tok::r_paren)) || 2934 (Left.is(tok::l_brace) && Left.isNot(BK_Block) && 2935 Right.is(tok::r_brace) && Right.isNot(BK_Block))) 2936 return Style.SpaceInEmptyParentheses; 2937 if (Style.SpacesInConditionalStatement) { 2938 if (Left.is(tok::l_paren) && Left.Previous && 2939 isKeywordWithCondition(*Left.Previous)) 2940 return true; 2941 if (Right.is(tok::r_paren) && Right.MatchingParen && 2942 Right.MatchingParen->Previous && 2943 isKeywordWithCondition(*Right.MatchingParen->Previous)) 2944 return true; 2945 } 2946 2947 // auto{x} auto(x) 2948 if (Left.is(tok::kw_auto) && Right.isOneOf(tok::l_paren, tok::l_brace)) 2949 return false; 2950 2951 // requires clause Concept1<T> && Concept2<T> 2952 if (Left.is(TT_ConstraintJunctions) && Right.is(tok::identifier)) 2953 return true; 2954 2955 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 2956 return (Right.is(TT_CastRParen) || 2957 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen))) 2958 ? Style.SpacesInCStyleCastParentheses 2959 : Style.SpacesInParentheses; 2960 if (Right.isOneOf(tok::semi, tok::comma)) 2961 return false; 2962 if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) { 2963 bool IsLightweightGeneric = Right.MatchingParen && 2964 Right.MatchingParen->Next && 2965 Right.MatchingParen->Next->is(tok::colon); 2966 return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList; 2967 } 2968 if (Right.is(tok::less) && Left.is(tok::kw_template)) 2969 return Style.SpaceAfterTemplateKeyword; 2970 if (Left.isOneOf(tok::exclaim, tok::tilde)) 2971 return false; 2972 if (Left.is(tok::at) && 2973 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 2974 tok::numeric_constant, tok::l_paren, tok::l_brace, 2975 tok::kw_true, tok::kw_false)) 2976 return false; 2977 if (Left.is(tok::colon)) 2978 return !Left.is(TT_ObjCMethodExpr); 2979 if (Left.is(tok::coloncolon)) 2980 return false; 2981 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) { 2982 if (Style.Language == FormatStyle::LK_TextProto || 2983 (Style.Language == FormatStyle::LK_Proto && 2984 (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) { 2985 // Format empty list as `<>`. 2986 if (Left.is(tok::less) && Right.is(tok::greater)) 2987 return false; 2988 return !Style.Cpp11BracedListStyle; 2989 } 2990 return false; 2991 } 2992 if (Right.is(tok::ellipsis)) 2993 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous && 2994 Left.Previous->is(tok::kw_case)); 2995 if (Left.is(tok::l_square) && Right.is(tok::amp)) 2996 return Style.SpacesInSquareBrackets; 2997 if (Right.is(TT_PointerOrReference)) { 2998 if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) { 2999 if (!Left.MatchingParen) 3000 return true; 3001 FormatToken *TokenBeforeMatchingParen = 3002 Left.MatchingParen->getPreviousNonComment(); 3003 if (!TokenBeforeMatchingParen || !Left.is(TT_TypeDeclarationParen)) 3004 return true; 3005 } 3006 // Add a space if the previous token is a pointer qualifier or the closing 3007 // parenthesis of __attribute__(()) expression and the style requires spaces 3008 // after pointer qualifiers. 3009 if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_After || 3010 Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) && 3011 (Left.is(TT_AttributeParen) || Left.canBePointerOrReferenceQualifier())) 3012 return true; 3013 return ( 3014 Left.Tok.isLiteral() || 3015 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) && 3016 (getTokenPointerOrReferenceAlignment(Right) != FormatStyle::PAS_Left || 3017 (Line.IsMultiVariableDeclStmt && 3018 (Left.NestingLevel == 0 || 3019 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for))))))); 3020 } 3021 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) && 3022 (!Left.is(TT_PointerOrReference) || 3023 (getTokenPointerOrReferenceAlignment(Left) != FormatStyle::PAS_Right && 3024 !Line.IsMultiVariableDeclStmt))) 3025 return true; 3026 if (Left.is(TT_PointerOrReference)) { 3027 // Add a space if the next token is a pointer qualifier and the style 3028 // requires spaces before pointer qualifiers. 3029 if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Before || 3030 Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) && 3031 Right.canBePointerOrReferenceQualifier()) 3032 return true; 3033 return Right.Tok.isLiteral() || Right.is(TT_BlockComment) || 3034 (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) && 3035 !Right.is(TT_StartOfName)) || 3036 (Right.is(tok::l_brace) && Right.is(BK_Block)) || 3037 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare, 3038 tok::l_paren) && 3039 (getTokenPointerOrReferenceAlignment(Left) != 3040 FormatStyle::PAS_Right && 3041 !Line.IsMultiVariableDeclStmt) && 3042 Left.Previous && 3043 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon, 3044 tok::l_square)); 3045 } 3046 // Ensure right pointer alignment with ellipsis e.g. int *...P 3047 if (Left.is(tok::ellipsis) && Left.Previous && 3048 Left.Previous->isOneOf(tok::star, tok::amp, tok::ampamp)) 3049 return Style.PointerAlignment != FormatStyle::PAS_Right; 3050 3051 if (Right.is(tok::star) && Left.is(tok::l_paren)) 3052 return false; 3053 if (Left.is(tok::star) && Right.isOneOf(tok::star, tok::amp, tok::ampamp)) 3054 return false; 3055 if (Right.isOneOf(tok::star, tok::amp, tok::ampamp)) { 3056 const FormatToken *Previous = &Left; 3057 while (Previous && !Previous->is(tok::kw_operator)) { 3058 if (Previous->is(tok::identifier) || Previous->isSimpleTypeSpecifier()) { 3059 Previous = Previous->getPreviousNonComment(); 3060 continue; 3061 } 3062 if (Previous->is(TT_TemplateCloser) && Previous->MatchingParen) { 3063 Previous = Previous->MatchingParen->getPreviousNonComment(); 3064 continue; 3065 } 3066 if (Previous->is(tok::coloncolon)) { 3067 Previous = Previous->getPreviousNonComment(); 3068 continue; 3069 } 3070 break; 3071 } 3072 // Space between the type and the * in: 3073 // operator void*() 3074 // operator char*() 3075 // operator void const*() 3076 // operator void volatile*() 3077 // operator /*comment*/ const char*() 3078 // operator volatile /*comment*/ char*() 3079 // operator Foo*() 3080 // operator C<T>*() 3081 // operator std::Foo*() 3082 // operator C<T>::D<U>*() 3083 // dependent on PointerAlignment style. 3084 if (Previous) { 3085 if (Previous->endsSequence(tok::kw_operator)) 3086 return (Style.PointerAlignment != FormatStyle::PAS_Left); 3087 if (Previous->is(tok::kw_const) || Previous->is(tok::kw_volatile)) 3088 return (Style.PointerAlignment != FormatStyle::PAS_Left) || 3089 (Style.SpaceAroundPointerQualifiers == 3090 FormatStyle::SAPQ_After) || 3091 (Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both); 3092 } 3093 } 3094 const auto SpaceRequiredForArrayInitializerLSquare = 3095 [](const FormatToken &LSquareTok, const FormatStyle &Style) { 3096 return Style.SpacesInContainerLiterals || 3097 ((Style.Language == FormatStyle::LK_Proto || 3098 Style.Language == FormatStyle::LK_TextProto) && 3099 !Style.Cpp11BracedListStyle && 3100 LSquareTok.endsSequence(tok::l_square, tok::colon, 3101 TT_SelectorName)); 3102 }; 3103 if (Left.is(tok::l_square)) 3104 return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) && 3105 SpaceRequiredForArrayInitializerLSquare(Left, Style)) || 3106 (Left.isOneOf(TT_ArraySubscriptLSquare, TT_StructuredBindingLSquare, 3107 TT_LambdaLSquare) && 3108 Style.SpacesInSquareBrackets && Right.isNot(tok::r_square)); 3109 if (Right.is(tok::r_square)) 3110 return Right.MatchingParen && 3111 ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) && 3112 SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen, 3113 Style)) || 3114 (Style.SpacesInSquareBrackets && 3115 Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare, 3116 TT_StructuredBindingLSquare, 3117 TT_LambdaLSquare)) || 3118 Right.MatchingParen->is(TT_AttributeParen)); 3119 if (Right.is(tok::l_square) && 3120 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 3121 TT_DesignatedInitializerLSquare, 3122 TT_StructuredBindingLSquare, TT_AttributeSquare) && 3123 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral) && 3124 !(!Left.is(tok::r_square) && Style.SpaceBeforeSquareBrackets && 3125 Right.is(TT_ArraySubscriptLSquare))) 3126 return false; 3127 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 3128 return !Left.Children.empty(); // No spaces in "{}". 3129 if ((Left.is(tok::l_brace) && Left.isNot(BK_Block)) || 3130 (Right.is(tok::r_brace) && Right.MatchingParen && 3131 Right.MatchingParen->isNot(BK_Block))) 3132 return Style.Cpp11BracedListStyle ? Style.SpacesInParentheses : true; 3133 if (Left.is(TT_BlockComment)) 3134 // No whitespace in x(/*foo=*/1), except for JavaScript. 3135 return Style.Language == FormatStyle::LK_JavaScript || 3136 !Left.TokenText.endswith("=*/"); 3137 3138 // Space between template and attribute. 3139 // e.g. template <typename T> [[nodiscard]] ... 3140 if (Left.is(TT_TemplateCloser) && Right.is(TT_AttributeSquare)) 3141 return true; 3142 // Space before parentheses common for all languages 3143 if (Right.is(tok::l_paren)) { 3144 if (Left.is(TT_TemplateCloser) && Right.isNot(TT_FunctionTypeLParen)) 3145 return spaceRequiredBeforeParens(Right); 3146 if (Left.is(tok::kw_requires)) 3147 return spaceRequiredBeforeParens(Right); 3148 if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) || 3149 (Left.is(tok::r_square) && Left.is(TT_AttributeSquare))) 3150 return true; 3151 if (Left.is(TT_ForEachMacro)) 3152 return (Style.SpaceBeforeParensOptions.AfterForeachMacros || 3153 spaceRequiredBeforeParens(Right)); 3154 if (Left.is(TT_IfMacro)) 3155 return (Style.SpaceBeforeParensOptions.AfterIfMacros || 3156 spaceRequiredBeforeParens(Right)); 3157 if (Line.Type == LT_ObjCDecl) 3158 return true; 3159 if (Left.is(tok::semi)) 3160 return true; 3161 if (Left.isOneOf(tok::pp_elif, tok::kw_for, tok::kw_while, tok::kw_switch, 3162 tok::kw_case, TT_ForEachMacro, TT_ObjCForIn)) 3163 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3164 spaceRequiredBeforeParens(Right); 3165 if (Left.isIf(Line.Type != LT_PreprocessorDirective)) 3166 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3167 spaceRequiredBeforeParens(Right); 3168 // Function declaration or definition 3169 if (Line.MightBeFunctionDecl && (Left.is(TT_FunctionDeclarationName) || 3170 Right.is(TT_OverloadedOperatorLParen))) { 3171 if (Line.mightBeFunctionDefinition()) 3172 return Style.SpaceBeforeParensOptions.AfterFunctionDefinitionName || 3173 spaceRequiredBeforeParens(Right); 3174 else 3175 return Style.SpaceBeforeParensOptions.AfterFunctionDeclarationName || 3176 spaceRequiredBeforeParens(Right); 3177 } 3178 // Lambda 3179 if (Line.Type != LT_PreprocessorDirective && Left.is(tok::r_square) && 3180 Left.MatchingParen && Left.MatchingParen->is(TT_LambdaLSquare)) 3181 return Style.SpaceBeforeParensOptions.AfterFunctionDefinitionName || 3182 spaceRequiredBeforeParens(Right); 3183 if (!Left.Previous || Left.Previous->isNot(tok::period)) { 3184 if (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch)) 3185 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3186 spaceRequiredBeforeParens(Right); 3187 if (Left.isOneOf(tok::kw_new, tok::kw_delete)) 3188 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never || 3189 spaceRequiredBeforeParens(Right); 3190 } 3191 if (Line.Type != LT_PreprocessorDirective && 3192 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() || 3193 Left.is(tok::r_paren) || Left.isSimpleTypeSpecifier())) 3194 return spaceRequiredBeforeParens(Right); 3195 return false; 3196 } 3197 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 3198 return false; 3199 if (Right.is(TT_UnaryOperator)) 3200 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 3201 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr)); 3202 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 3203 tok::r_paren) || 3204 Left.isSimpleTypeSpecifier()) && 3205 Right.is(tok::l_brace) && Right.getNextNonComment() && 3206 Right.isNot(BK_Block)) 3207 return false; 3208 if (Left.is(tok::period) || Right.is(tok::period)) 3209 return false; 3210 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 3211 return false; 3212 if (Left.is(TT_TemplateCloser) && Left.MatchingParen && 3213 Left.MatchingParen->Previous && 3214 (Left.MatchingParen->Previous->is(tok::period) || 3215 Left.MatchingParen->Previous->is(tok::coloncolon))) 3216 // Java call to generic function with explicit type: 3217 // A.<B<C<...>>>DoSomething(); 3218 // A::<B<C<...>>>DoSomething(); // With a Java 8 method reference. 3219 return false; 3220 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square)) 3221 return false; 3222 if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at)) 3223 // Objective-C dictionary literal -> no space after opening brace. 3224 return false; 3225 if (Right.is(tok::r_brace) && Right.MatchingParen && 3226 Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at)) 3227 // Objective-C dictionary literal -> no space before closing brace. 3228 return false; 3229 if (Right.getType() == TT_TrailingAnnotation && 3230 Right.isOneOf(tok::amp, tok::ampamp) && 3231 Left.isOneOf(tok::kw_const, tok::kw_volatile) && 3232 (!Right.Next || Right.Next->is(tok::semi))) 3233 // Match const and volatile ref-qualifiers without any additional 3234 // qualifiers such as 3235 // void Fn() const &; 3236 return getTokenReferenceAlignment(Right) != FormatStyle::PAS_Left; 3237 3238 return true; 3239 } 3240 3241 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 3242 const FormatToken &Right) { 3243 const FormatToken &Left = *Right.Previous; 3244 auto HasExistingWhitespace = [&Right]() { 3245 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 3246 }; 3247 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo()) 3248 return true; // Never ever merge two identifiers. 3249 if (Style.isCpp()) { 3250 if (Left.is(tok::kw_operator)) 3251 return Right.is(tok::coloncolon); 3252 if (Right.is(tok::l_brace) && Right.is(BK_BracedInit) && 3253 !Left.opensScope() && Style.SpaceBeforeCpp11BracedList) 3254 return true; 3255 } else if (Style.Language == FormatStyle::LK_Proto || 3256 Style.Language == FormatStyle::LK_TextProto) { 3257 if (Right.is(tok::period) && 3258 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required, 3259 Keywords.kw_repeated, Keywords.kw_extend)) 3260 return true; 3261 if (Right.is(tok::l_paren) && 3262 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) 3263 return true; 3264 if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName)) 3265 return true; 3266 // Slashes occur in text protocol extension syntax: [type/type] { ... }. 3267 if (Left.is(tok::slash) || Right.is(tok::slash)) 3268 return false; 3269 if (Left.MatchingParen && 3270 Left.MatchingParen->is(TT_ProtoExtensionLSquare) && 3271 Right.isOneOf(tok::l_brace, tok::less)) 3272 return !Style.Cpp11BracedListStyle; 3273 // A percent is probably part of a formatting specification, such as %lld. 3274 if (Left.is(tok::percent)) 3275 return false; 3276 // Preserve the existence of a space before a percent for cases like 0x%04x 3277 // and "%d %d" 3278 if (Left.is(tok::numeric_constant) && Right.is(tok::percent)) 3279 return HasExistingWhitespace(); 3280 } else if (Style.isJson()) { 3281 if (Right.is(tok::colon)) 3282 return false; 3283 } else if (Style.isCSharp()) { 3284 // Require spaces around '{' and before '}' unless they appear in 3285 // interpolated strings. Interpolated strings are merged into a single token 3286 // so cannot have spaces inserted by this function. 3287 3288 // No space between 'this' and '[' 3289 if (Left.is(tok::kw_this) && Right.is(tok::l_square)) 3290 return false; 3291 3292 // No space between 'new' and '(' 3293 if (Left.is(tok::kw_new) && Right.is(tok::l_paren)) 3294 return false; 3295 3296 // Space before { (including space within '{ {'). 3297 if (Right.is(tok::l_brace)) 3298 return true; 3299 3300 // Spaces inside braces. 3301 if (Left.is(tok::l_brace) && Right.isNot(tok::r_brace)) 3302 return true; 3303 3304 if (Left.isNot(tok::l_brace) && Right.is(tok::r_brace)) 3305 return true; 3306 3307 // Spaces around '=>'. 3308 if (Left.is(TT_FatArrow) || Right.is(TT_FatArrow)) 3309 return true; 3310 3311 // No spaces around attribute target colons 3312 if (Left.is(TT_AttributeColon) || Right.is(TT_AttributeColon)) 3313 return false; 3314 3315 // space between type and variable e.g. Dictionary<string,string> foo; 3316 if (Left.is(TT_TemplateCloser) && Right.is(TT_StartOfName)) 3317 return true; 3318 3319 // spaces inside square brackets. 3320 if (Left.is(tok::l_square) || Right.is(tok::r_square)) 3321 return Style.SpacesInSquareBrackets; 3322 3323 // No space before ? in nullable types. 3324 if (Right.is(TT_CSharpNullable)) 3325 return false; 3326 3327 // No space before null forgiving '!'. 3328 if (Right.is(TT_NonNullAssertion)) 3329 return false; 3330 3331 // No space between consecutive commas '[,,]'. 3332 if (Left.is(tok::comma) && Right.is(tok::comma)) 3333 return false; 3334 3335 // space after var in `var (key, value)` 3336 if (Left.is(Keywords.kw_var) && Right.is(tok::l_paren)) 3337 return true; 3338 3339 // space between keywords and paren e.g. "using (" 3340 if (Right.is(tok::l_paren)) 3341 if (Left.isOneOf(tok::kw_using, Keywords.kw_async, Keywords.kw_when, 3342 Keywords.kw_lock)) 3343 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3344 spaceRequiredBeforeParens(Right); 3345 3346 // space between method modifier and opening parenthesis of a tuple return 3347 // type 3348 if (Left.isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected, 3349 tok::kw_virtual, tok::kw_extern, tok::kw_static, 3350 Keywords.kw_internal, Keywords.kw_abstract, 3351 Keywords.kw_sealed, Keywords.kw_override, 3352 Keywords.kw_async, Keywords.kw_unsafe) && 3353 Right.is(tok::l_paren)) 3354 return true; 3355 } else if (Style.Language == FormatStyle::LK_JavaScript) { 3356 if (Left.is(TT_FatArrow)) 3357 return true; 3358 // for await ( ... 3359 if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous && 3360 Left.Previous->is(tok::kw_for)) 3361 return true; 3362 if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) && 3363 Right.MatchingParen) { 3364 const FormatToken *Next = Right.MatchingParen->getNextNonComment(); 3365 // An async arrow function, for example: `x = async () => foo();`, 3366 // as opposed to calling a function called async: `x = async();` 3367 if (Next && Next->is(TT_FatArrow)) 3368 return true; 3369 } 3370 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 3371 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 3372 return false; 3373 // In tagged template literals ("html`bar baz`"), there is no space between 3374 // the tag identifier and the template string. 3375 if (Keywords.IsJavaScriptIdentifier(Left, 3376 /* AcceptIdentifierName= */ false) && 3377 Right.is(TT_TemplateString)) 3378 return false; 3379 if (Right.is(tok::star) && 3380 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) 3381 return false; 3382 if (Right.isOneOf(tok::l_brace, tok::l_square) && 3383 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield, 3384 Keywords.kw_extends, Keywords.kw_implements)) 3385 return true; 3386 if (Right.is(tok::l_paren)) { 3387 // JS methods can use some keywords as names (e.g. `delete()`). 3388 if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo()) 3389 return false; 3390 // Valid JS method names can include keywords, e.g. `foo.delete()` or 3391 // `bar.instanceof()`. Recognize call positions by preceding period. 3392 if (Left.Previous && Left.Previous->is(tok::period) && 3393 Left.Tok.getIdentifierInfo()) 3394 return false; 3395 // Additional unary JavaScript operators that need a space after. 3396 if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof, 3397 tok::kw_void)) 3398 return true; 3399 } 3400 // `foo as const;` casts into a const type. 3401 if (Left.endsSequence(tok::kw_const, Keywords.kw_as)) { 3402 return false; 3403 } 3404 if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in, 3405 tok::kw_const) || 3406 // "of" is only a keyword if it appears after another identifier 3407 // (e.g. as "const x of y" in a for loop), or after a destructuring 3408 // operation (const [x, y] of z, const {a, b} of c). 3409 (Left.is(Keywords.kw_of) && Left.Previous && 3410 (Left.Previous->Tok.is(tok::identifier) || 3411 Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) && 3412 (!Left.Previous || !Left.Previous->is(tok::period))) 3413 return true; 3414 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous && 3415 Left.Previous->is(tok::period) && Right.is(tok::l_paren)) 3416 return false; 3417 if (Left.is(Keywords.kw_as) && 3418 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren)) 3419 return true; 3420 if (Left.is(tok::kw_default) && Left.Previous && 3421 Left.Previous->is(tok::kw_export)) 3422 return true; 3423 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace)) 3424 return true; 3425 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion)) 3426 return false; 3427 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator)) 3428 return false; 3429 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) && 3430 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) 3431 return false; 3432 if (Left.is(tok::ellipsis)) 3433 return false; 3434 if (Left.is(TT_TemplateCloser) && 3435 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square, 3436 Keywords.kw_implements, Keywords.kw_extends)) 3437 // Type assertions ('<type>expr') are not followed by whitespace. Other 3438 // locations that should have whitespace following are identified by the 3439 // above set of follower tokens. 3440 return false; 3441 if (Right.is(TT_NonNullAssertion)) 3442 return false; 3443 if (Left.is(TT_NonNullAssertion) && 3444 Right.isOneOf(Keywords.kw_as, Keywords.kw_in)) 3445 return true; // "x! as string", "x! in y" 3446 } else if (Style.Language == FormatStyle::LK_Java) { 3447 if (Left.is(tok::r_square) && Right.is(tok::l_brace)) 3448 return true; 3449 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) 3450 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3451 spaceRequiredBeforeParens(Right); 3452 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private, 3453 tok::kw_protected) || 3454 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract, 3455 Keywords.kw_native)) && 3456 Right.is(TT_TemplateOpener)) 3457 return true; 3458 } 3459 if (Left.is(TT_ImplicitStringLiteral)) 3460 return HasExistingWhitespace(); 3461 if (Line.Type == LT_ObjCMethodDecl) { 3462 if (Left.is(TT_ObjCMethodSpecifier)) 3463 return true; 3464 if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right)) 3465 // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a 3466 // keyword in Objective-C, and '+ (instancetype)new;' is a standard class 3467 // method declaration. 3468 return false; 3469 } 3470 if (Line.Type == LT_ObjCProperty && 3471 (Right.is(tok::equal) || Left.is(tok::equal))) 3472 return false; 3473 3474 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) || 3475 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) 3476 return true; 3477 if (Left.is(tok::comma) && !Right.is(TT_OverloadedOperatorLParen)) 3478 return true; 3479 if (Right.is(tok::comma)) 3480 return false; 3481 if (Right.is(TT_ObjCBlockLParen)) 3482 return true; 3483 if (Right.is(TT_CtorInitializerColon)) 3484 return Style.SpaceBeforeCtorInitializerColon; 3485 if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon) 3486 return false; 3487 if (Right.is(TT_RangeBasedForLoopColon) && 3488 !Style.SpaceBeforeRangeBasedForLoopColon) 3489 return false; 3490 if (Left.is(TT_BitFieldColon)) 3491 return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both || 3492 Style.BitFieldColonSpacing == FormatStyle::BFCS_After; 3493 if (Right.is(tok::colon)) { 3494 if (Line.First->isOneOf(tok::kw_default, tok::kw_case)) 3495 return Style.SpaceBeforeCaseColon; 3496 if (!Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi)) 3497 return false; 3498 if (Right.is(TT_ObjCMethodExpr)) 3499 return false; 3500 if (Left.is(tok::question)) 3501 return false; 3502 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon)) 3503 return false; 3504 if (Right.is(TT_DictLiteral)) 3505 return Style.SpacesInContainerLiterals; 3506 if (Right.is(TT_AttributeColon)) 3507 return false; 3508 if (Right.is(TT_CSharpNamedArgumentColon)) 3509 return false; 3510 if (Right.is(TT_BitFieldColon)) 3511 return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both || 3512 Style.BitFieldColonSpacing == FormatStyle::BFCS_Before; 3513 return true; 3514 } 3515 // Do not merge "- -" into "--". 3516 if ((Left.isOneOf(tok::minus, tok::minusminus) && 3517 Right.isOneOf(tok::minus, tok::minusminus)) || 3518 (Left.isOneOf(tok::plus, tok::plusplus) && 3519 Right.isOneOf(tok::plus, tok::plusplus))) 3520 return true; 3521 if (Left.is(TT_UnaryOperator)) { 3522 if (!Right.is(tok::l_paren)) { 3523 // The alternative operators for ~ and ! are "compl" and "not". 3524 // If they are used instead, we do not want to combine them with 3525 // the token to the right, unless that is a left paren. 3526 if (Left.is(tok::exclaim) && Left.TokenText == "not") 3527 return true; 3528 if (Left.is(tok::tilde) && Left.TokenText == "compl") 3529 return true; 3530 // Lambda captures allow for a lone &, so "&]" needs to be properly 3531 // handled. 3532 if (Left.is(tok::amp) && Right.is(tok::r_square)) 3533 return Style.SpacesInSquareBrackets; 3534 } 3535 return (Style.SpaceAfterLogicalNot && Left.is(tok::exclaim)) || 3536 Right.is(TT_BinaryOperator); 3537 } 3538 3539 // If the next token is a binary operator or a selector name, we have 3540 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly. 3541 if (Left.is(TT_CastRParen)) 3542 return Style.SpaceAfterCStyleCast || 3543 Right.isOneOf(TT_BinaryOperator, TT_SelectorName); 3544 3545 auto ShouldAddSpacesInAngles = [this, &HasExistingWhitespace]() { 3546 if (this->Style.SpacesInAngles == FormatStyle::SIAS_Always) 3547 return true; 3548 if (this->Style.SpacesInAngles == FormatStyle::SIAS_Leave) 3549 return HasExistingWhitespace(); 3550 return false; 3551 }; 3552 3553 if (Left.is(tok::greater) && Right.is(tok::greater)) { 3554 if (Style.Language == FormatStyle::LK_TextProto || 3555 (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral))) 3556 return !Style.Cpp11BracedListStyle; 3557 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) && 3558 ((Style.Standard < FormatStyle::LS_Cpp11) || 3559 ShouldAddSpacesInAngles()); 3560 } 3561 if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) || 3562 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) || 3563 (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod))) 3564 return false; 3565 if (!Style.SpaceBeforeAssignmentOperators && Left.isNot(TT_TemplateCloser) && 3566 Right.getPrecedence() == prec::Assignment) 3567 return false; 3568 if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) && 3569 (Left.is(tok::identifier) || Left.is(tok::kw_this))) 3570 return false; 3571 if (Right.is(tok::coloncolon) && Left.is(tok::identifier)) 3572 // Generally don't remove existing spaces between an identifier and "::". 3573 // The identifier might actually be a macro name such as ALWAYS_INLINE. If 3574 // this turns out to be too lenient, add analysis of the identifier itself. 3575 return HasExistingWhitespace(); 3576 if (Right.is(tok::coloncolon) && 3577 !Left.isOneOf(tok::l_brace, tok::comment, tok::l_paren)) 3578 // Put a space between < and :: in vector< ::std::string > 3579 return (Left.is(TT_TemplateOpener) && 3580 ((Style.Standard < FormatStyle::LS_Cpp11) || 3581 ShouldAddSpacesInAngles())) || 3582 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square, 3583 tok::kw___super, TT_TemplateOpener, 3584 TT_TemplateCloser)) || 3585 (Left.is(tok::l_paren) && Style.SpacesInParentheses); 3586 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser))) 3587 return ShouldAddSpacesInAngles(); 3588 // Space before TT_StructuredBindingLSquare. 3589 if (Right.is(TT_StructuredBindingLSquare)) 3590 return !Left.isOneOf(tok::amp, tok::ampamp) || 3591 getTokenReferenceAlignment(Left) != FormatStyle::PAS_Right; 3592 // Space before & or && following a TT_StructuredBindingLSquare. 3593 if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) && 3594 Right.isOneOf(tok::amp, tok::ampamp)) 3595 return getTokenReferenceAlignment(Right) != FormatStyle::PAS_Left; 3596 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) || 3597 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && 3598 !Right.is(tok::r_paren))) 3599 return true; 3600 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) && 3601 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen)) 3602 return false; 3603 if (Right.is(tok::less) && Left.isNot(tok::l_paren) && 3604 Line.startsWith(tok::hash)) 3605 return true; 3606 if (Right.is(TT_TrailingUnaryOperator)) 3607 return false; 3608 if (Left.is(TT_RegexLiteral)) 3609 return false; 3610 return spaceRequiredBetween(Line, Left, Right); 3611 } 3612 3613 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 3614 static bool isAllmanBrace(const FormatToken &Tok) { 3615 return Tok.is(tok::l_brace) && Tok.is(BK_Block) && 3616 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_LambdaLBrace, TT_DictLiteral); 3617 } 3618 3619 // Returns 'true' if 'Tok' is a function argument. 3620 static bool IsFunctionArgument(const FormatToken &Tok) { 3621 return Tok.MatchingParen && Tok.MatchingParen->Next && 3622 Tok.MatchingParen->Next->isOneOf(tok::comma, tok::r_paren); 3623 } 3624 3625 static bool 3626 isItAnEmptyLambdaAllowed(const FormatToken &Tok, 3627 FormatStyle::ShortLambdaStyle ShortLambdaOption) { 3628 return Tok.Children.empty() && ShortLambdaOption != FormatStyle::SLS_None; 3629 } 3630 3631 static bool isAllmanLambdaBrace(const FormatToken &Tok) { 3632 return (Tok.is(tok::l_brace) && Tok.is(BK_Block) && 3633 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral)); 3634 } 3635 3636 // Returns the first token on the line that is not a comment. 3637 static const FormatToken *getFirstNonComment(const AnnotatedLine &Line) { 3638 const FormatToken *Next = Line.First; 3639 if (!Next) 3640 return Next; 3641 if (Next->is(tok::comment)) 3642 Next = Next->getNextNonComment(); 3643 return Next; 3644 } 3645 3646 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 3647 const FormatToken &Right) { 3648 const FormatToken &Left = *Right.Previous; 3649 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0) 3650 return true; 3651 3652 if (Style.isCSharp()) { 3653 if (Right.is(TT_CSharpNamedArgumentColon) || 3654 Left.is(TT_CSharpNamedArgumentColon)) 3655 return false; 3656 if (Right.is(TT_CSharpGenericTypeConstraint)) 3657 return true; 3658 3659 // Break after C# [...] and before public/protected/private/internal. 3660 if (Left.is(TT_AttributeSquare) && Left.is(tok::r_square) && 3661 (Right.isAccessSpecifier(/*ColonRequired=*/false) || 3662 Right.is(Keywords.kw_internal))) 3663 return true; 3664 // Break between ] and [ but only when there are really 2 attributes. 3665 if (Left.is(TT_AttributeSquare) && Right.is(TT_AttributeSquare) && 3666 Left.is(tok::r_square) && Right.is(tok::l_square)) 3667 return true; 3668 3669 } else if (Style.Language == FormatStyle::LK_JavaScript) { 3670 // FIXME: This might apply to other languages and token kinds. 3671 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous && 3672 Left.Previous->is(tok::string_literal)) 3673 return true; 3674 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 && 3675 Left.Previous && Left.Previous->is(tok::equal) && 3676 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export, 3677 tok::kw_const) && 3678 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match 3679 // above. 3680 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) 3681 // Object literals on the top level of a file are treated as "enum-style". 3682 // Each key/value pair is put on a separate line, instead of bin-packing. 3683 return true; 3684 if (Left.is(tok::l_brace) && Line.Level == 0 && 3685 (Line.startsWith(tok::kw_enum) || 3686 Line.startsWith(tok::kw_const, tok::kw_enum) || 3687 Line.startsWith(tok::kw_export, tok::kw_enum) || 3688 Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum))) 3689 // JavaScript top-level enum key/value pairs are put on separate lines 3690 // instead of bin-packing. 3691 return true; 3692 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && Left.Previous && 3693 Left.Previous->is(TT_FatArrow)) { 3694 // JS arrow function (=> {...}). 3695 switch (Style.AllowShortLambdasOnASingleLine) { 3696 case FormatStyle::SLS_All: 3697 return false; 3698 case FormatStyle::SLS_None: 3699 return true; 3700 case FormatStyle::SLS_Empty: 3701 return !Left.Children.empty(); 3702 case FormatStyle::SLS_Inline: 3703 // allow one-lining inline (e.g. in function call args) and empty arrow 3704 // functions. 3705 return (Left.NestingLevel == 0 && Line.Level == 0) && 3706 !Left.Children.empty(); 3707 } 3708 llvm_unreachable("Unknown FormatStyle::ShortLambdaStyle enum"); 3709 } 3710 3711 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && 3712 !Left.Children.empty()) 3713 // Support AllowShortFunctionsOnASingleLine for JavaScript. 3714 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 3715 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty || 3716 (Left.NestingLevel == 0 && Line.Level == 0 && 3717 Style.AllowShortFunctionsOnASingleLine & 3718 FormatStyle::SFS_InlineOnly); 3719 } else if (Style.Language == FormatStyle::LK_Java) { 3720 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next && 3721 Right.Next->is(tok::string_literal)) 3722 return true; 3723 } else if (Style.Language == FormatStyle::LK_Cpp || 3724 Style.Language == FormatStyle::LK_ObjC || 3725 Style.Language == FormatStyle::LK_Proto || 3726 Style.Language == FormatStyle::LK_TableGen || 3727 Style.Language == FormatStyle::LK_TextProto) { 3728 if (Left.isStringLiteral() && Right.isStringLiteral()) 3729 return true; 3730 } 3731 3732 // Basic JSON newline processing. 3733 if (Style.isJson()) { 3734 // Always break after a JSON record opener. 3735 // { 3736 // } 3737 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace)) 3738 return true; 3739 // Always break after a JSON array opener. 3740 // [ 3741 // ] 3742 if (Left.is(TT_ArrayInitializerLSquare) && Left.is(tok::l_square) && 3743 !Right.is(tok::r_square)) 3744 return true; 3745 // Always break after successive entries. 3746 // 1, 3747 // 2 3748 if (Left.is(tok::comma)) 3749 return true; 3750 } 3751 3752 // If the last token before a '}', ']', or ')' is a comma or a trailing 3753 // comment, the intention is to insert a line break after it in order to make 3754 // shuffling around entries easier. Import statements, especially in 3755 // JavaScript, can be an exception to this rule. 3756 if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) { 3757 const FormatToken *BeforeClosingBrace = nullptr; 3758 if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) || 3759 (Style.Language == FormatStyle::LK_JavaScript && 3760 Left.is(tok::l_paren))) && 3761 Left.isNot(BK_Block) && Left.MatchingParen) 3762 BeforeClosingBrace = Left.MatchingParen->Previous; 3763 else if (Right.MatchingParen && 3764 (Right.MatchingParen->isOneOf(tok::l_brace, 3765 TT_ArrayInitializerLSquare) || 3766 (Style.Language == FormatStyle::LK_JavaScript && 3767 Right.MatchingParen->is(tok::l_paren)))) 3768 BeforeClosingBrace = &Left; 3769 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 3770 BeforeClosingBrace->isTrailingComment())) 3771 return true; 3772 } 3773 3774 if (Right.is(tok::comment)) 3775 return Left.isNot(BK_BracedInit) && Left.isNot(TT_CtorInitializerColon) && 3776 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 3777 if (Left.isTrailingComment()) 3778 return true; 3779 if (Right.Previous->IsUnterminatedLiteral) 3780 return true; 3781 if (Right.is(tok::lessless) && Right.Next && 3782 Right.Previous->is(tok::string_literal) && 3783 Right.Next->is(tok::string_literal)) 3784 return true; 3785 // Can break after template<> declaration 3786 if (Right.Previous->ClosesTemplateDeclaration && 3787 Right.Previous->MatchingParen && 3788 Right.Previous->MatchingParen->NestingLevel == 0) { 3789 // Put concepts on the next line e.g. 3790 // template<typename T> 3791 // concept ... 3792 if (Right.is(tok::kw_concept)) 3793 return Style.BreakBeforeConceptDeclarations; 3794 return (Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes); 3795 } 3796 if (Style.PackConstructorInitializers == FormatStyle::PCIS_Never) { 3797 if (Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeColon && 3798 (Left.is(TT_CtorInitializerComma) || Right.is(TT_CtorInitializerColon))) 3799 return true; 3800 3801 if (Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon && 3802 Left.isOneOf(TT_CtorInitializerColon, TT_CtorInitializerComma)) 3803 return true; 3804 } 3805 if (Style.PackConstructorInitializers < FormatStyle::PCIS_CurrentLine && 3806 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 3807 Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon)) 3808 return true; 3809 // Break only if we have multiple inheritance. 3810 if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma && 3811 Right.is(TT_InheritanceComma)) 3812 return true; 3813 if (Style.BreakInheritanceList == FormatStyle::BILS_AfterComma && 3814 Left.is(TT_InheritanceComma)) 3815 return true; 3816 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) 3817 // Multiline raw string literals are special wrt. line breaks. The author 3818 // has made a deliberate choice and might have aligned the contents of the 3819 // string literal accordingly. Thus, we try keep existing line breaks. 3820 return Right.IsMultiline && Right.NewlinesBefore > 0; 3821 if ((Right.Previous->is(tok::l_brace) || 3822 (Right.Previous->is(tok::less) && Right.Previous->Previous && 3823 Right.Previous->Previous->is(tok::equal))) && 3824 Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) { 3825 // Don't put enums or option definitions onto single lines in protocol 3826 // buffers. 3827 return true; 3828 } 3829 if (Right.is(TT_InlineASMBrace)) 3830 return Right.HasUnescapedNewline; 3831 3832 if (isAllmanBrace(Left) || isAllmanBrace(Right)) { 3833 auto FirstNonComment = getFirstNonComment(Line); 3834 bool AccessSpecifier = 3835 FirstNonComment && 3836 FirstNonComment->isOneOf(Keywords.kw_internal, tok::kw_public, 3837 tok::kw_private, tok::kw_protected); 3838 3839 if (Style.BraceWrapping.AfterEnum) { 3840 if (Line.startsWith(tok::kw_enum) || 3841 Line.startsWith(tok::kw_typedef, tok::kw_enum)) 3842 return true; 3843 // Ensure BraceWrapping for `public enum A {`. 3844 if (AccessSpecifier && FirstNonComment->Next && 3845 FirstNonComment->Next->is(tok::kw_enum)) 3846 return true; 3847 } 3848 3849 // Ensure BraceWrapping for `public interface A {`. 3850 if (Style.BraceWrapping.AfterClass && 3851 ((AccessSpecifier && FirstNonComment->Next && 3852 FirstNonComment->Next->is(Keywords.kw_interface)) || 3853 Line.startsWith(Keywords.kw_interface))) 3854 return true; 3855 3856 return (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) || 3857 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct); 3858 } 3859 3860 if (Left.is(TT_ObjCBlockLBrace) && 3861 Style.AllowShortBlocksOnASingleLine == FormatStyle::SBS_Never) 3862 return true; 3863 3864 // Ensure wrapping after __attribute__((XX)) and @interface etc. 3865 if (Left.is(TT_AttributeParen) && Right.is(TT_ObjCDecl)) 3866 return true; 3867 3868 if (Left.is(TT_LambdaLBrace)) { 3869 if (IsFunctionArgument(Left) && 3870 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline) 3871 return false; 3872 3873 if (Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_None || 3874 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline || 3875 (!Left.Children.empty() && 3876 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Empty)) 3877 return true; 3878 } 3879 3880 if (Style.BraceWrapping.BeforeLambdaBody && Right.is(TT_LambdaLBrace) && 3881 Left.isOneOf(tok::star, tok::amp, tok::ampamp, TT_TemplateCloser)) { 3882 return true; 3883 } 3884 3885 // Put multiple Java annotation on a new line. 3886 if ((Style.Language == FormatStyle::LK_Java || 3887 Style.Language == FormatStyle::LK_JavaScript) && 3888 Left.is(TT_LeadingJavaAnnotation) && 3889 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) && 3890 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) 3891 return true; 3892 3893 if (Right.is(TT_ProtoExtensionLSquare)) 3894 return true; 3895 3896 // In text proto instances if a submessage contains at least 2 entries and at 3897 // least one of them is a submessage, like A { ... B { ... } ... }, 3898 // put all of the entries of A on separate lines by forcing the selector of 3899 // the submessage B to be put on a newline. 3900 // 3901 // Example: these can stay on one line: 3902 // a { scalar_1: 1 scalar_2: 2 } 3903 // a { b { key: value } } 3904 // 3905 // and these entries need to be on a new line even if putting them all in one 3906 // line is under the column limit: 3907 // a { 3908 // scalar: 1 3909 // b { key: value } 3910 // } 3911 // 3912 // We enforce this by breaking before a submessage field that has previous 3913 // siblings, *and* breaking before a field that follows a submessage field. 3914 // 3915 // Be careful to exclude the case [proto.ext] { ... } since the `]` is 3916 // the TT_SelectorName there, but we don't want to break inside the brackets. 3917 // 3918 // Another edge case is @submessage { key: value }, which is a common 3919 // substitution placeholder. In this case we want to keep `@` and `submessage` 3920 // together. 3921 // 3922 // We ensure elsewhere that extensions are always on their own line. 3923 if ((Style.Language == FormatStyle::LK_Proto || 3924 Style.Language == FormatStyle::LK_TextProto) && 3925 Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) { 3926 // Keep `@submessage` together in: 3927 // @submessage { key: value } 3928 if (Right.Previous && Right.Previous->is(tok::at)) 3929 return false; 3930 // Look for the scope opener after selector in cases like: 3931 // selector { ... 3932 // selector: { ... 3933 // selector: @base { ... 3934 FormatToken *LBrace = Right.Next; 3935 if (LBrace && LBrace->is(tok::colon)) { 3936 LBrace = LBrace->Next; 3937 if (LBrace && LBrace->is(tok::at)) { 3938 LBrace = LBrace->Next; 3939 if (LBrace) 3940 LBrace = LBrace->Next; 3941 } 3942 } 3943 if (LBrace && 3944 // The scope opener is one of {, [, <: 3945 // selector { ... } 3946 // selector [ ... ] 3947 // selector < ... > 3948 // 3949 // In case of selector { ... }, the l_brace is TT_DictLiteral. 3950 // In case of an empty selector {}, the l_brace is not TT_DictLiteral, 3951 // so we check for immediately following r_brace. 3952 ((LBrace->is(tok::l_brace) && 3953 (LBrace->is(TT_DictLiteral) || 3954 (LBrace->Next && LBrace->Next->is(tok::r_brace)))) || 3955 LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) { 3956 // If Left.ParameterCount is 0, then this submessage entry is not the 3957 // first in its parent submessage, and we want to break before this entry. 3958 // If Left.ParameterCount is greater than 0, then its parent submessage 3959 // might contain 1 or more entries and we want to break before this entry 3960 // if it contains at least 2 entries. We deal with this case later by 3961 // detecting and breaking before the next entry in the parent submessage. 3962 if (Left.ParameterCount == 0) 3963 return true; 3964 // However, if this submessage is the first entry in its parent 3965 // submessage, Left.ParameterCount might be 1 in some cases. 3966 // We deal with this case later by detecting an entry 3967 // following a closing paren of this submessage. 3968 } 3969 3970 // If this is an entry immediately following a submessage, it will be 3971 // preceded by a closing paren of that submessage, like in: 3972 // left---. .---right 3973 // v v 3974 // sub: { ... } key: value 3975 // If there was a comment between `}` an `key` above, then `key` would be 3976 // put on a new line anyways. 3977 if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square)) 3978 return true; 3979 } 3980 3981 // Deal with lambda arguments in C++ - we want consistent line breaks whether 3982 // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced 3983 // as aggressive line breaks are placed when the lambda is not the last arg. 3984 if ((Style.Language == FormatStyle::LK_Cpp || 3985 Style.Language == FormatStyle::LK_ObjC) && 3986 Left.is(tok::l_paren) && Left.BlockParameterCount > 0 && 3987 !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) { 3988 // Multiple lambdas in the same function call force line breaks. 3989 if (Left.BlockParameterCount > 1) 3990 return true; 3991 3992 // A lambda followed by another arg forces a line break. 3993 if (!Left.Role) 3994 return false; 3995 auto Comma = Left.Role->lastComma(); 3996 if (!Comma) 3997 return false; 3998 auto Next = Comma->getNextNonComment(); 3999 if (!Next) 4000 return false; 4001 if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret)) 4002 return true; 4003 } 4004 4005 return false; 4006 } 4007 4008 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 4009 const FormatToken &Right) { 4010 const FormatToken &Left = *Right.Previous; 4011 // Language-specific stuff. 4012 if (Style.isCSharp()) { 4013 if (Left.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon) || 4014 Right.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon)) 4015 return false; 4016 // Only break after commas for generic type constraints. 4017 if (Line.First->is(TT_CSharpGenericTypeConstraint)) 4018 return Left.is(TT_CSharpGenericTypeConstraintComma); 4019 // Keep nullable operators attached to their identifiers. 4020 if (Right.is(TT_CSharpNullable)) { 4021 return false; 4022 } 4023 } else if (Style.Language == FormatStyle::LK_Java) { 4024 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 4025 Keywords.kw_implements)) 4026 return false; 4027 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 4028 Keywords.kw_implements)) 4029 return true; 4030 } else if (Style.Language == FormatStyle::LK_JavaScript) { 4031 const FormatToken *NonComment = Right.getPreviousNonComment(); 4032 if (NonComment && 4033 NonComment->isOneOf( 4034 tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break, 4035 tok::kw_throw, Keywords.kw_interface, Keywords.kw_type, 4036 tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected, 4037 Keywords.kw_readonly, Keywords.kw_override, Keywords.kw_abstract, 4038 Keywords.kw_get, Keywords.kw_set, Keywords.kw_async, 4039 Keywords.kw_await)) 4040 return false; // Otherwise automatic semicolon insertion would trigger. 4041 if (Right.NestingLevel == 0 && 4042 (Left.Tok.getIdentifierInfo() || 4043 Left.isOneOf(tok::r_square, tok::r_paren)) && 4044 Right.isOneOf(tok::l_square, tok::l_paren)) 4045 return false; // Otherwise automatic semicolon insertion would trigger. 4046 if (NonComment && NonComment->is(tok::identifier) && 4047 NonComment->TokenText == "asserts") 4048 return false; 4049 if (Left.is(TT_FatArrow) && Right.is(tok::l_brace)) 4050 return false; 4051 if (Left.is(TT_JsTypeColon)) 4052 return true; 4053 // Don't wrap between ":" and "!" of a strict prop init ("field!: type;"). 4054 if (Left.is(tok::exclaim) && Right.is(tok::colon)) 4055 return false; 4056 // Look for is type annotations like: 4057 // function f(): a is B { ... } 4058 // Do not break before is in these cases. 4059 if (Right.is(Keywords.kw_is)) { 4060 const FormatToken *Next = Right.getNextNonComment(); 4061 // If `is` is followed by a colon, it's likely that it's a dict key, so 4062 // ignore it for this check. 4063 // For example this is common in Polymer: 4064 // Polymer({ 4065 // is: 'name', 4066 // ... 4067 // }); 4068 if (!Next || !Next->is(tok::colon)) 4069 return false; 4070 } 4071 if (Left.is(Keywords.kw_in)) 4072 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None; 4073 if (Right.is(Keywords.kw_in)) 4074 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None; 4075 if (Right.is(Keywords.kw_as)) 4076 return false; // must not break before as in 'x as type' casts 4077 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) { 4078 // extends and infer can appear as keywords in conditional types: 4079 // https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types 4080 // do not break before them, as the expressions are subject to ASI. 4081 return false; 4082 } 4083 if (Left.is(Keywords.kw_as)) 4084 return true; 4085 if (Left.is(TT_NonNullAssertion)) 4086 return true; 4087 if (Left.is(Keywords.kw_declare) && 4088 Right.isOneOf(Keywords.kw_module, tok::kw_namespace, 4089 Keywords.kw_function, tok::kw_class, tok::kw_enum, 4090 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var, 4091 Keywords.kw_let, tok::kw_const)) 4092 // See grammar for 'declare' statements at: 4093 // https://github.com/Microsoft/TypeScript/blob/main/doc/spec-ARCHIVED.md#A.10 4094 return false; 4095 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) && 4096 Right.isOneOf(tok::identifier, tok::string_literal)) 4097 return false; // must not break in "module foo { ...}" 4098 if (Right.is(TT_TemplateString) && Right.closesScope()) 4099 return false; 4100 // Don't split tagged template literal so there is a break between the tag 4101 // identifier and template string. 4102 if (Left.is(tok::identifier) && Right.is(TT_TemplateString)) { 4103 return false; 4104 } 4105 if (Left.is(TT_TemplateString) && Left.opensScope()) 4106 return true; 4107 } 4108 4109 if (Left.is(tok::at)) 4110 return false; 4111 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 4112 return false; 4113 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation)) 4114 return !Right.is(tok::l_paren); 4115 if (Right.is(TT_PointerOrReference)) 4116 return Line.IsMultiVariableDeclStmt || 4117 (getTokenPointerOrReferenceAlignment(Right) == 4118 FormatStyle::PAS_Right && 4119 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName))); 4120 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 4121 Right.is(tok::kw_operator)) 4122 return true; 4123 if (Left.is(TT_PointerOrReference)) 4124 return false; 4125 if (Right.isTrailingComment()) 4126 // We rely on MustBreakBefore being set correctly here as we should not 4127 // change the "binding" behavior of a comment. 4128 // The first comment in a braced lists is always interpreted as belonging to 4129 // the first list element. Otherwise, it should be placed outside of the 4130 // list. 4131 return Left.is(BK_BracedInit) || 4132 (Left.is(TT_CtorInitializerColon) && 4133 Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon); 4134 if (Left.is(tok::question) && Right.is(tok::colon)) 4135 return false; 4136 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question)) 4137 return Style.BreakBeforeTernaryOperators; 4138 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question)) 4139 return !Style.BreakBeforeTernaryOperators; 4140 if (Left.is(TT_InheritanceColon)) 4141 return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon; 4142 if (Right.is(TT_InheritanceColon)) 4143 return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon; 4144 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) && 4145 Left.isNot(TT_SelectorName)) 4146 return true; 4147 4148 if (Right.is(tok::colon) && 4149 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) 4150 return false; 4151 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) { 4152 if (Style.Language == FormatStyle::LK_Proto || 4153 Style.Language == FormatStyle::LK_TextProto) { 4154 if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral()) 4155 return false; 4156 // Prevent cases like: 4157 // 4158 // submessage: 4159 // { key: valueeeeeeeeeeee } 4160 // 4161 // when the snippet does not fit into one line. 4162 // Prefer: 4163 // 4164 // submessage: { 4165 // key: valueeeeeeeeeeee 4166 // } 4167 // 4168 // instead, even if it is longer by one line. 4169 // 4170 // Note that this allows allows the "{" to go over the column limit 4171 // when the column limit is just between ":" and "{", but that does 4172 // not happen too often and alternative formattings in this case are 4173 // not much better. 4174 // 4175 // The code covers the cases: 4176 // 4177 // submessage: { ... } 4178 // submessage: < ... > 4179 // repeated: [ ... ] 4180 if (((Right.is(tok::l_brace) || Right.is(tok::less)) && 4181 Right.is(TT_DictLiteral)) || 4182 Right.is(TT_ArrayInitializerLSquare)) 4183 return false; 4184 } 4185 return true; 4186 } 4187 if (Right.is(tok::r_square) && Right.MatchingParen && 4188 Right.MatchingParen->is(TT_ProtoExtensionLSquare)) 4189 return false; 4190 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next && 4191 Right.Next->is(TT_ObjCMethodExpr))) 4192 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls. 4193 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 4194 return true; 4195 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen)) 4196 return true; 4197 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen, 4198 TT_OverloadedOperator)) 4199 return false; 4200 if (Left.is(TT_RangeBasedForLoopColon)) 4201 return true; 4202 if (Right.is(TT_RangeBasedForLoopColon)) 4203 return false; 4204 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener)) 4205 return true; 4206 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) || 4207 Left.is(tok::kw_operator)) 4208 return false; 4209 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) && 4210 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) 4211 return false; 4212 if (Left.is(tok::equal) && Right.is(tok::l_brace) && 4213 !Style.Cpp11BracedListStyle) 4214 return false; 4215 if (Left.is(tok::l_paren) && 4216 Left.isOneOf(TT_AttributeParen, TT_TypeDeclarationParen)) 4217 return false; 4218 if (Left.is(tok::l_paren) && Left.Previous && 4219 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) 4220 return false; 4221 if (Right.is(TT_ImplicitStringLiteral)) 4222 return false; 4223 4224 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser)) 4225 return false; 4226 if (Right.is(tok::r_square) && Right.MatchingParen && 4227 Right.MatchingParen->is(TT_LambdaLSquare)) 4228 return false; 4229 4230 // We only break before r_brace if there was a corresponding break before 4231 // the l_brace, which is tracked by BreakBeforeClosingBrace. 4232 if (Right.is(tok::r_brace)) 4233 return Right.MatchingParen && Right.MatchingParen->is(BK_Block); 4234 4235 // Allow breaking after a trailing annotation, e.g. after a method 4236 // declaration. 4237 if (Left.is(TT_TrailingAnnotation)) 4238 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 4239 tok::less, tok::coloncolon); 4240 4241 if (Right.is(tok::kw___attribute) || 4242 (Right.is(tok::l_square) && Right.is(TT_AttributeSquare))) 4243 return !Left.is(TT_AttributeSquare); 4244 4245 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 4246 return true; 4247 4248 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 4249 return true; 4250 4251 if (Left.is(TT_CtorInitializerColon)) 4252 return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon; 4253 if (Right.is(TT_CtorInitializerColon)) 4254 return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon; 4255 if (Left.is(TT_CtorInitializerComma) && 4256 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 4257 return false; 4258 if (Right.is(TT_CtorInitializerComma) && 4259 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 4260 return true; 4261 if (Left.is(TT_InheritanceComma) && 4262 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) 4263 return false; 4264 if (Right.is(TT_InheritanceComma) && 4265 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) 4266 return true; 4267 if ((Left.is(tok::greater) && Right.is(tok::greater)) || 4268 (Left.is(tok::less) && Right.is(tok::less))) 4269 return false; 4270 if (Right.is(TT_BinaryOperator) && 4271 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 4272 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 4273 Right.getPrecedence() != prec::Assignment)) 4274 return true; 4275 if (Left.is(TT_ArrayInitializerLSquare)) 4276 return true; 4277 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 4278 return true; 4279 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) && 4280 !Left.isOneOf(tok::arrowstar, tok::lessless) && 4281 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 4282 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 4283 Left.getPrecedence() == prec::Assignment)) 4284 return true; 4285 if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) || 4286 (Left.is(tok::r_square) && Right.is(TT_AttributeSquare))) 4287 return false; 4288 4289 auto ShortLambdaOption = Style.AllowShortLambdasOnASingleLine; 4290 if (Style.BraceWrapping.BeforeLambdaBody && Right.is(TT_LambdaLBrace)) { 4291 if (isAllmanLambdaBrace(Left)) 4292 return !isItAnEmptyLambdaAllowed(Left, ShortLambdaOption); 4293 if (isAllmanLambdaBrace(Right)) 4294 return !isItAnEmptyLambdaAllowed(Right, ShortLambdaOption); 4295 } 4296 4297 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 4298 tok::kw_class, tok::kw_struct, tok::comment) || 4299 Right.isMemberAccess() || 4300 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless, 4301 tok::colon, tok::l_square, tok::at) || 4302 (Left.is(tok::r_paren) && 4303 Right.isOneOf(tok::identifier, tok::kw_const)) || 4304 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) || 4305 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser)); 4306 } 4307 4308 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 4309 llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n"; 4310 const FormatToken *Tok = Line.First; 4311 while (Tok) { 4312 llvm::errs() << " M=" << Tok->MustBreakBefore 4313 << " C=" << Tok->CanBreakBefore 4314 << " T=" << getTokenTypeName(Tok->getType()) 4315 << " S=" << Tok->SpacesRequiredBefore 4316 << " F=" << Tok->Finalized << " B=" << Tok->BlockParameterCount 4317 << " BK=" << Tok->getBlockKind() << " P=" << Tok->SplitPenalty 4318 << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength 4319 << " PPK=" << Tok->getPackingKind() << " FakeLParens="; 4320 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 4321 llvm::errs() << Tok->FakeLParens[i] << "/"; 4322 llvm::errs() << " FakeRParens=" << Tok->FakeRParens; 4323 llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo(); 4324 llvm::errs() << " Text='" << Tok->TokenText << "'\n"; 4325 if (!Tok->Next) 4326 assert(Tok == Line.Last); 4327 Tok = Tok->Next; 4328 } 4329 llvm::errs() << "----\n"; 4330 } 4331 4332 FormatStyle::PointerAlignmentStyle 4333 TokenAnnotator::getTokenReferenceAlignment(const FormatToken &Reference) { 4334 assert(Reference.isOneOf(tok::amp, tok::ampamp)); 4335 switch (Style.ReferenceAlignment) { 4336 case FormatStyle::RAS_Pointer: 4337 return Style.PointerAlignment; 4338 case FormatStyle::RAS_Left: 4339 return FormatStyle::PAS_Left; 4340 case FormatStyle::RAS_Right: 4341 return FormatStyle::PAS_Right; 4342 case FormatStyle::RAS_Middle: 4343 return FormatStyle::PAS_Middle; 4344 } 4345 assert(0); //"Unhandled value of ReferenceAlignment" 4346 return Style.PointerAlignment; 4347 } 4348 4349 FormatStyle::PointerAlignmentStyle 4350 TokenAnnotator::getTokenPointerOrReferenceAlignment( 4351 const FormatToken &PointerOrReference) { 4352 if (PointerOrReference.isOneOf(tok::amp, tok::ampamp)) { 4353 switch (Style.ReferenceAlignment) { 4354 case FormatStyle::RAS_Pointer: 4355 return Style.PointerAlignment; 4356 case FormatStyle::RAS_Left: 4357 return FormatStyle::PAS_Left; 4358 case FormatStyle::RAS_Right: 4359 return FormatStyle::PAS_Right; 4360 case FormatStyle::RAS_Middle: 4361 return FormatStyle::PAS_Middle; 4362 } 4363 } 4364 assert(PointerOrReference.is(tok::star)); 4365 return Style.PointerAlignment; 4366 } 4367 4368 } // namespace format 4369 } // namespace clang 4370