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(); 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.isJavaScript() && 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.isJavaScript() && 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.isJavaScript() && 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 (&&FunctionReference)(void); 318 // void (^ObjCBlock)(void); 319 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression; 320 bool ProbablyFunctionType = 321 CurrentToken->isOneOf(tok::star, tok::amp, tok::ampamp, tok::caret); 322 bool HasMultipleLines = false; 323 bool HasMultipleParametersOnALine = false; 324 bool MightBeObjCForRangeLoop = 325 Left->Previous && Left->Previous->is(tok::kw_for); 326 FormatToken *PossibleObjCForInToken = nullptr; 327 while (CurrentToken) { 328 // LookForDecls is set when "if (" has been seen. Check for 329 // 'identifier' '*' 'identifier' followed by not '=' -- this 330 // '*' has to be a binary operator but determineStarAmpUsage() will 331 // categorize it as an unary operator, so set the right type here. 332 if (LookForDecls && CurrentToken->Next) { 333 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 334 if (Prev) { 335 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 336 FormatToken *Next = CurrentToken->Next; 337 if (PrevPrev && PrevPrev->is(tok::identifier) && 338 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 339 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 340 Prev->setType(TT_BinaryOperator); 341 LookForDecls = false; 342 } 343 } 344 } 345 346 if (CurrentToken->Previous->is(TT_PointerOrReference) && 347 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 348 tok::coloncolon)) 349 ProbablyFunctionType = true; 350 if (CurrentToken->is(tok::comma)) 351 MightBeFunctionType = false; 352 if (CurrentToken->Previous->is(TT_BinaryOperator)) 353 Contexts.back().IsExpression = true; 354 if (CurrentToken->is(tok::r_paren)) { 355 if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next && 356 (CurrentToken->Next->is(tok::l_paren) || 357 (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration))) 358 Left->setType(Left->Next->is(tok::caret) ? TT_ObjCBlockLParen 359 : TT_FunctionTypeLParen); 360 Left->MatchingParen = CurrentToken; 361 CurrentToken->MatchingParen = Left; 362 363 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) && 364 Left->Previous && Left->Previous->is(tok::l_paren)) { 365 // Detect the case where macros are used to generate lambdas or 366 // function bodies, e.g.: 367 // auto my_lambda = MACRO((Type *type, int i) { .. body .. }); 368 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) { 369 if (Tok->is(TT_BinaryOperator) && 370 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) 371 Tok->setType(TT_PointerOrReference); 372 } 373 } 374 375 if (StartsObjCMethodExpr) { 376 CurrentToken->setType(TT_ObjCMethodExpr); 377 if (Contexts.back().FirstObjCSelectorName) { 378 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 379 Contexts.back().LongestObjCSelectorName; 380 } 381 } 382 383 if (Left->is(TT_AttributeParen)) 384 CurrentToken->setType(TT_AttributeParen); 385 if (Left->is(TT_TypeDeclarationParen)) 386 CurrentToken->setType(TT_TypeDeclarationParen); 387 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation)) 388 CurrentToken->setType(TT_JavaAnnotation); 389 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation)) 390 CurrentToken->setType(TT_LeadingJavaAnnotation); 391 if (Left->Previous && Left->Previous->is(TT_AttributeSquare)) 392 CurrentToken->setType(TT_AttributeSquare); 393 394 if (!HasMultipleLines) 395 Left->setPackingKind(PPK_Inconclusive); 396 else if (HasMultipleParametersOnALine) 397 Left->setPackingKind(PPK_BinPacked); 398 else 399 Left->setPackingKind(PPK_OnePerLine); 400 401 next(); 402 return true; 403 } 404 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 405 return false; 406 407 if (CurrentToken->is(tok::l_brace)) 408 Left->setType(TT_Unknown); // Not TT_ObjCBlockLParen 409 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 410 !CurrentToken->Next->HasUnescapedNewline && 411 !CurrentToken->Next->isTrailingComment()) 412 HasMultipleParametersOnALine = true; 413 bool ProbablyFunctionTypeLParen = 414 (CurrentToken->is(tok::l_paren) && CurrentToken->Next && 415 CurrentToken->Next->isOneOf(tok::star, tok::amp, tok::caret)); 416 if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) || 417 CurrentToken->Previous->isSimpleTypeSpecifier()) && 418 !(CurrentToken->is(tok::l_brace) || 419 (CurrentToken->is(tok::l_paren) && !ProbablyFunctionTypeLParen))) 420 Contexts.back().IsExpression = false; 421 if (CurrentToken->isOneOf(tok::semi, tok::colon)) { 422 MightBeObjCForRangeLoop = false; 423 if (PossibleObjCForInToken) { 424 PossibleObjCForInToken->setType(TT_Unknown); 425 PossibleObjCForInToken = nullptr; 426 } 427 } 428 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) { 429 PossibleObjCForInToken = CurrentToken; 430 PossibleObjCForInToken->setType(TT_ObjCForIn); 431 } 432 // When we discover a 'new', we set CanBeExpression to 'false' in order to 433 // parse the type correctly. Reset that after a comma. 434 if (CurrentToken->is(tok::comma)) 435 Contexts.back().CanBeExpression = true; 436 437 FormatToken *Tok = CurrentToken; 438 if (!consumeToken()) 439 return false; 440 updateParameterCount(Left, Tok); 441 if (CurrentToken && CurrentToken->HasUnescapedNewline) 442 HasMultipleLines = true; 443 } 444 return false; 445 } 446 447 bool isCSharpAttributeSpecifier(const FormatToken &Tok) { 448 if (!Style.isCSharp()) 449 return false; 450 451 // `identifier[i]` is not an attribute. 452 if (Tok.Previous && Tok.Previous->is(tok::identifier)) 453 return false; 454 455 // Chains of [] in `identifier[i][j][k]` are not attributes. 456 if (Tok.Previous && Tok.Previous->is(tok::r_square)) { 457 auto *MatchingParen = Tok.Previous->MatchingParen; 458 if (!MatchingParen || MatchingParen->is(TT_ArraySubscriptLSquare)) 459 return false; 460 } 461 462 const FormatToken *AttrTok = Tok.Next; 463 if (!AttrTok) 464 return false; 465 466 // Just an empty declaration e.g. string []. 467 if (AttrTok->is(tok::r_square)) 468 return false; 469 470 // Move along the tokens inbetween the '[' and ']' e.g. [STAThread]. 471 while (AttrTok && AttrTok->isNot(tok::r_square)) { 472 AttrTok = AttrTok->Next; 473 } 474 475 if (!AttrTok) 476 return false; 477 478 // Allow an attribute to be the only content of a file. 479 AttrTok = AttrTok->Next; 480 if (!AttrTok) 481 return true; 482 483 // Limit this to being an access modifier that follows. 484 if (AttrTok->isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected, 485 tok::comment, tok::kw_class, tok::kw_static, 486 tok::l_square, Keywords.kw_internal)) { 487 return true; 488 } 489 490 // incase its a [XXX] retval func(.... 491 if (AttrTok->Next && 492 AttrTok->Next->startsSequence(tok::identifier, tok::l_paren)) 493 return true; 494 495 return false; 496 } 497 498 bool isCpp11AttributeSpecifier(const FormatToken &Tok) { 499 if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square)) 500 return false; 501 // The first square bracket is part of an ObjC array literal 502 if (Tok.Previous && Tok.Previous->is(tok::at)) { 503 return false; 504 } 505 const FormatToken *AttrTok = Tok.Next->Next; 506 if (!AttrTok) 507 return false; 508 // C++17 '[[using ns: foo, bar(baz, blech)]]' 509 // We assume nobody will name an ObjC variable 'using'. 510 if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon)) 511 return true; 512 if (AttrTok->isNot(tok::identifier)) 513 return false; 514 while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) { 515 // ObjC message send. We assume nobody will use : in a C++11 attribute 516 // specifier parameter, although this is technically valid: 517 // [[foo(:)]]. 518 if (AttrTok->is(tok::colon) || 519 AttrTok->startsSequence(tok::identifier, tok::identifier) || 520 AttrTok->startsSequence(tok::r_paren, tok::identifier)) 521 return false; 522 if (AttrTok->is(tok::ellipsis)) 523 return true; 524 AttrTok = AttrTok->Next; 525 } 526 return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square); 527 } 528 529 bool parseSquare() { 530 if (!CurrentToken) 531 return false; 532 533 // A '[' could be an index subscript (after an identifier or after 534 // ')' or ']'), it could be the start of an Objective-C method 535 // expression, it could the start of an Objective-C array literal, 536 // or it could be a C++ attribute specifier [[foo::bar]]. 537 FormatToken *Left = CurrentToken->Previous; 538 Left->ParentBracket = Contexts.back().ContextKind; 539 FormatToken *Parent = Left->getPreviousNonComment(); 540 541 // Cases where '>' is followed by '['. 542 // In C++, this can happen either in array of templates (foo<int>[10]) 543 // or when array is a nested template type (unique_ptr<type1<type2>[]>). 544 bool CppArrayTemplates = 545 Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) && 546 (Contexts.back().CanBeExpression || Contexts.back().IsExpression || 547 Contexts.back().InTemplateArgument); 548 549 bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) || 550 Contexts.back().InCpp11AttributeSpecifier; 551 552 // Treat C# Attributes [STAThread] much like C++ attributes [[...]]. 553 bool IsCSharpAttributeSpecifier = 554 isCSharpAttributeSpecifier(*Left) || 555 Contexts.back().InCSharpAttributeSpecifier; 556 557 bool InsideInlineASM = Line.startsWith(tok::kw_asm); 558 bool IsCppStructuredBinding = Left->isCppStructuredBinding(Style); 559 bool StartsObjCMethodExpr = 560 !IsCppStructuredBinding && !InsideInlineASM && !CppArrayTemplates && 561 Style.isCpp() && !IsCpp11AttributeSpecifier && 562 !IsCSharpAttributeSpecifier && Contexts.back().CanBeExpression && 563 Left->isNot(TT_LambdaLSquare) && 564 !CurrentToken->isOneOf(tok::l_brace, tok::r_square) && 565 (!Parent || 566 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 567 tok::kw_return, tok::kw_throw) || 568 Parent->isUnaryOperator() || 569 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 570 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) || 571 (getBinOpPrecedence(Parent->Tok.getKind(), true, true) > 572 prec::Unknown)); 573 bool ColonFound = false; 574 575 unsigned BindingIncrease = 1; 576 if (IsCppStructuredBinding) { 577 Left->setType(TT_StructuredBindingLSquare); 578 } else if (Left->is(TT_Unknown)) { 579 if (StartsObjCMethodExpr) { 580 Left->setType(TT_ObjCMethodExpr); 581 } else if (InsideInlineASM) { 582 Left->setType(TT_InlineASMSymbolicNameLSquare); 583 } else if (IsCpp11AttributeSpecifier) { 584 Left->setType(TT_AttributeSquare); 585 } else if (Style.isJavaScript() && Parent && 586 Contexts.back().ContextKind == tok::l_brace && 587 Parent->isOneOf(tok::l_brace, tok::comma)) { 588 Left->setType(TT_JsComputedPropertyName); 589 } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace && 590 Parent && Parent->isOneOf(tok::l_brace, tok::comma)) { 591 Left->setType(TT_DesignatedInitializerLSquare); 592 } else if (IsCSharpAttributeSpecifier) { 593 Left->setType(TT_AttributeSquare); 594 } else if (CurrentToken->is(tok::r_square) && Parent && 595 Parent->is(TT_TemplateCloser)) { 596 Left->setType(TT_ArraySubscriptLSquare); 597 } else if (Style.Language == FormatStyle::LK_Proto || 598 Style.Language == FormatStyle::LK_TextProto) { 599 // Square braces in LK_Proto can either be message field attributes: 600 // 601 // optional Aaa aaa = 1 [ 602 // (aaa) = aaa 603 // ]; 604 // 605 // extensions 123 [ 606 // (aaa) = aaa 607 // ]; 608 // 609 // or text proto extensions (in options): 610 // 611 // option (Aaa.options) = { 612 // [type.type/type] { 613 // key: value 614 // } 615 // } 616 // 617 // or repeated fields (in options): 618 // 619 // option (Aaa.options) = { 620 // keys: [ 1, 2, 3 ] 621 // } 622 // 623 // In the first and the third case we want to spread the contents inside 624 // the square braces; in the second we want to keep them inline. 625 Left->setType(TT_ArrayInitializerLSquare); 626 if (!Left->endsSequence(tok::l_square, tok::numeric_constant, 627 tok::equal) && 628 !Left->endsSequence(tok::l_square, tok::numeric_constant, 629 tok::identifier) && 630 !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) { 631 Left->setType(TT_ProtoExtensionLSquare); 632 BindingIncrease = 10; 633 } 634 } else if (!CppArrayTemplates && Parent && 635 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at, 636 tok::comma, tok::l_paren, tok::l_square, 637 tok::question, tok::colon, tok::kw_return, 638 // Should only be relevant to JavaScript: 639 tok::kw_default)) { 640 Left->setType(TT_ArrayInitializerLSquare); 641 } else { 642 BindingIncrease = 10; 643 Left->setType(TT_ArraySubscriptLSquare); 644 } 645 } 646 647 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease); 648 Contexts.back().IsExpression = true; 649 if (Style.isJavaScript() && Parent && 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.isJavaScript() && 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) || Style.isJavaScript()) 811 Left->setType(TT_DictLiteral); 812 } 813 if (CurrentToken->is(tok::comma)) { 814 if (Style.isJavaScript()) 815 Left->setType(TT_DictLiteral); 816 ++CommaCount; 817 } 818 if (!consumeToken()) 819 return false; 820 } 821 } 822 return true; 823 } 824 825 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 826 // For ObjC methods, the number of parameters is calculated differently as 827 // method declarations have a different structure (the parameters are not 828 // inside a bracket scope). 829 if (Current->is(tok::l_brace) && Current->is(BK_Block)) 830 ++Left->BlockParameterCount; 831 if (Current->is(tok::comma)) { 832 ++Left->ParameterCount; 833 if (!Left->Role) 834 Left->Role.reset(new CommaSeparatedList(Style)); 835 Left->Role->CommaFound(Current); 836 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 837 Left->ParameterCount = 1; 838 } 839 } 840 841 bool parseConditional() { 842 while (CurrentToken) { 843 if (CurrentToken->is(tok::colon)) { 844 CurrentToken->setType(TT_ConditionalExpr); 845 next(); 846 return true; 847 } 848 if (!consumeToken()) 849 return false; 850 } 851 return false; 852 } 853 854 bool parseTemplateDeclaration() { 855 if (CurrentToken && CurrentToken->is(tok::less)) { 856 CurrentToken->setType(TT_TemplateOpener); 857 next(); 858 if (!parseAngle()) 859 return false; 860 if (CurrentToken) 861 CurrentToken->Previous->ClosesTemplateDeclaration = true; 862 return true; 863 } 864 return false; 865 } 866 867 bool consumeToken() { 868 FormatToken *Tok = CurrentToken; 869 next(); 870 switch (Tok->Tok.getKind()) { 871 case tok::plus: 872 case tok::minus: 873 if (!Tok->Previous && Line.MustBeDeclaration) 874 Tok->setType(TT_ObjCMethodSpecifier); 875 break; 876 case tok::colon: 877 if (!Tok->Previous) 878 return false; 879 // Colons from ?: are handled in parseConditional(). 880 if (Style.isJavaScript()) { 881 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop 882 (Contexts.size() == 1 && // switch/case labels 883 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) || 884 Contexts.back().ContextKind == tok::l_paren || // function params 885 Contexts.back().ContextKind == tok::l_square || // array type 886 (!Contexts.back().IsExpression && 887 Contexts.back().ContextKind == tok::l_brace) || // object type 888 (Contexts.size() == 1 && 889 Line.MustBeDeclaration)) { // method/property declaration 890 Contexts.back().IsExpression = false; 891 Tok->setType(TT_JsTypeColon); 892 break; 893 } 894 } else if (Style.isCSharp()) { 895 if (Contexts.back().InCSharpAttributeSpecifier) { 896 Tok->setType(TT_AttributeColon); 897 break; 898 } 899 if (Contexts.back().ContextKind == tok::l_paren) { 900 Tok->setType(TT_CSharpNamedArgumentColon); 901 break; 902 } 903 } 904 if (Line.First->isOneOf(Keywords.kw_module, Keywords.kw_import) || 905 Line.First->startsSequence(tok::kw_export, Keywords.kw_module) || 906 Line.First->startsSequence(tok::kw_export, Keywords.kw_import)) { 907 Tok->setType(TT_ModulePartitionColon); 908 } else if (Contexts.back().ColonIsDictLiteral || 909 Style.Language == FormatStyle::LK_Proto || 910 Style.Language == FormatStyle::LK_TextProto) { 911 Tok->setType(TT_DictLiteral); 912 if (Style.Language == FormatStyle::LK_TextProto) { 913 if (FormatToken *Previous = Tok->getPreviousNonComment()) 914 Previous->setType(TT_SelectorName); 915 } 916 } else if (Contexts.back().ColonIsObjCMethodExpr || 917 Line.startsWith(TT_ObjCMethodSpecifier)) { 918 Tok->setType(TT_ObjCMethodExpr); 919 const FormatToken *BeforePrevious = Tok->Previous->Previous; 920 // Ensure we tag all identifiers in method declarations as 921 // TT_SelectorName. 922 bool UnknownIdentifierInMethodDeclaration = 923 Line.startsWith(TT_ObjCMethodSpecifier) && 924 Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown); 925 if (!BeforePrevious || 926 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 927 !(BeforePrevious->is(TT_CastRParen) || 928 (BeforePrevious->is(TT_ObjCMethodExpr) && 929 BeforePrevious->is(tok::colon))) || 930 BeforePrevious->is(tok::r_square) || 931 Contexts.back().LongestObjCSelectorName == 0 || 932 UnknownIdentifierInMethodDeclaration) { 933 Tok->Previous->setType(TT_SelectorName); 934 if (!Contexts.back().FirstObjCSelectorName) 935 Contexts.back().FirstObjCSelectorName = Tok->Previous; 936 else if (Tok->Previous->ColumnWidth > 937 Contexts.back().LongestObjCSelectorName) 938 Contexts.back().LongestObjCSelectorName = 939 Tok->Previous->ColumnWidth; 940 Tok->Previous->ParameterIndex = 941 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 942 ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 943 } 944 } else if (Contexts.back().ColonIsForRangeExpr) { 945 Tok->setType(TT_RangeBasedForLoopColon); 946 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 947 Tok->setType(TT_BitFieldColon); 948 } else if (Contexts.size() == 1 && 949 !Line.First->isOneOf(tok::kw_enum, tok::kw_case, 950 tok::kw_default)) { 951 FormatToken *Prev = Tok->getPreviousNonComment(); 952 if (!Prev) 953 break; 954 if (Prev->isOneOf(tok::r_paren, tok::kw_noexcept)) 955 Tok->setType(TT_CtorInitializerColon); 956 else if (Prev->is(tok::kw_try)) { 957 // Member initializer list within function try block. 958 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 959 if (!PrevPrev) 960 break; 961 if (PrevPrev && PrevPrev->isOneOf(tok::r_paren, tok::kw_noexcept)) 962 Tok->setType(TT_CtorInitializerColon); 963 } else 964 Tok->setType(TT_InheritanceColon); 965 } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next && 966 (Tok->Next->isOneOf(tok::r_paren, tok::comma) || 967 (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next && 968 Tok->Next->Next->is(tok::colon)))) { 969 // This handles a special macro in ObjC code where selectors including 970 // the colon are passed as macro arguments. 971 Tok->setType(TT_ObjCMethodExpr); 972 } else if (Contexts.back().ContextKind == tok::l_paren) { 973 Tok->setType(TT_InlineASMColon); 974 } 975 break; 976 case tok::pipe: 977 case tok::amp: 978 // | and & in declarations/type expressions represent union and 979 // intersection types, respectively. 980 if (Style.isJavaScript() && !Contexts.back().IsExpression) 981 Tok->setType(TT_JsTypeOperator); 982 break; 983 case tok::kw_if: 984 case tok::kw_while: 985 if (Tok->is(tok::kw_if) && CurrentToken && 986 CurrentToken->isOneOf(tok::kw_constexpr, tok::identifier)) 987 next(); 988 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 989 next(); 990 if (!parseParens(/*LookForDecls=*/true)) 991 return false; 992 } 993 break; 994 case tok::kw_for: 995 if (Style.isJavaScript()) { 996 // x.for and {for: ...} 997 if ((Tok->Previous && Tok->Previous->is(tok::period)) || 998 (Tok->Next && Tok->Next->is(tok::colon))) 999 break; 1000 // JS' for await ( ... 1001 if (CurrentToken && CurrentToken->is(Keywords.kw_await)) 1002 next(); 1003 } 1004 if (Style.isCpp() && CurrentToken && CurrentToken->is(tok::kw_co_await)) 1005 next(); 1006 Contexts.back().ColonIsForRangeExpr = true; 1007 next(); 1008 if (!parseParens()) 1009 return false; 1010 break; 1011 case tok::l_paren: 1012 // When faced with 'operator()()', the kw_operator handler incorrectly 1013 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make 1014 // the first two parens OverloadedOperators and the second l_paren an 1015 // OverloadedOperatorLParen. 1016 if (Tok->Previous && Tok->Previous->is(tok::r_paren) && 1017 Tok->Previous->MatchingParen && 1018 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) { 1019 Tok->Previous->setType(TT_OverloadedOperator); 1020 Tok->Previous->MatchingParen->setType(TT_OverloadedOperator); 1021 Tok->setType(TT_OverloadedOperatorLParen); 1022 } 1023 1024 if (!parseParens()) 1025 return false; 1026 if (Line.MustBeDeclaration && Contexts.size() == 1 && 1027 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) && 1028 !Tok->is(TT_TypeDeclarationParen) && 1029 (!Tok->Previous || !Tok->Previous->isOneOf(tok::kw___attribute, 1030 TT_LeadingJavaAnnotation))) 1031 Line.MightBeFunctionDecl = true; 1032 break; 1033 case tok::l_square: 1034 if (!parseSquare()) 1035 return false; 1036 break; 1037 case tok::l_brace: 1038 if (Style.Language == FormatStyle::LK_TextProto) { 1039 FormatToken *Previous = Tok->getPreviousNonComment(); 1040 if (Previous && Previous->getType() != TT_DictLiteral) 1041 Previous->setType(TT_SelectorName); 1042 } 1043 if (!parseBrace()) 1044 return false; 1045 break; 1046 case tok::less: 1047 if (parseAngle()) { 1048 Tok->setType(TT_TemplateOpener); 1049 // In TT_Proto, we must distignuish between: 1050 // map<key, value> 1051 // msg < item: data > 1052 // msg: < item: data > 1053 // In TT_TextProto, map<key, value> does not occur. 1054 if (Style.Language == FormatStyle::LK_TextProto || 1055 (Style.Language == FormatStyle::LK_Proto && Tok->Previous && 1056 Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) { 1057 Tok->setType(TT_DictLiteral); 1058 FormatToken *Previous = Tok->getPreviousNonComment(); 1059 if (Previous && Previous->getType() != TT_DictLiteral) 1060 Previous->setType(TT_SelectorName); 1061 } 1062 } else { 1063 Tok->setType(TT_BinaryOperator); 1064 NonTemplateLess.insert(Tok); 1065 CurrentToken = Tok; 1066 next(); 1067 } 1068 break; 1069 case tok::r_paren: 1070 case tok::r_square: 1071 return false; 1072 case tok::r_brace: 1073 // Lines can start with '}'. 1074 if (Tok->Previous) 1075 return false; 1076 break; 1077 case tok::greater: 1078 if (Style.Language != FormatStyle::LK_TextProto) 1079 Tok->setType(TT_BinaryOperator); 1080 if (Tok->Previous && Tok->Previous->is(TT_TemplateCloser)) 1081 Tok->SpacesRequiredBefore = 1; 1082 break; 1083 case tok::kw_operator: 1084 if (Style.Language == FormatStyle::LK_TextProto || 1085 Style.Language == FormatStyle::LK_Proto) 1086 break; 1087 while (CurrentToken && 1088 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 1089 if (CurrentToken->isOneOf(tok::star, tok::amp)) 1090 CurrentToken->setType(TT_PointerOrReference); 1091 consumeToken(); 1092 if (CurrentToken && CurrentToken->is(tok::comma) && 1093 CurrentToken->Previous->isNot(tok::kw_operator)) 1094 break; 1095 if (CurrentToken && CurrentToken->Previous->isOneOf( 1096 TT_BinaryOperator, TT_UnaryOperator, tok::comma, 1097 tok::star, tok::arrow, tok::amp, tok::ampamp)) 1098 CurrentToken->Previous->setType(TT_OverloadedOperator); 1099 } 1100 if (CurrentToken && CurrentToken->is(tok::l_paren)) 1101 CurrentToken->setType(TT_OverloadedOperatorLParen); 1102 if (CurrentToken && CurrentToken->Previous->is(TT_BinaryOperator)) 1103 CurrentToken->Previous->setType(TT_OverloadedOperator); 1104 break; 1105 case tok::question: 1106 if (Style.isJavaScript() && Tok->Next && 1107 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren, 1108 tok::r_brace)) { 1109 // Question marks before semicolons, colons, etc. indicate optional 1110 // types (fields, parameters), e.g. 1111 // function(x?: string, y?) {...} 1112 // class X { y?; } 1113 Tok->setType(TT_JsTypeOptionalQuestion); 1114 break; 1115 } 1116 // Declarations cannot be conditional expressions, this can only be part 1117 // of a type declaration. 1118 if (Line.MustBeDeclaration && !Contexts.back().IsExpression && 1119 Style.isJavaScript()) 1120 break; 1121 if (Style.isCSharp()) { 1122 // `Type?)`, `Type?>`, `Type? name;` and `Type? name =` can only be 1123 // nullable types. 1124 // Line.MustBeDeclaration will be true for `Type? name;`. 1125 if ((!Contexts.back().IsExpression && Line.MustBeDeclaration) || 1126 (Tok->Next && Tok->Next->isOneOf(tok::r_paren, tok::greater)) || 1127 (Tok->Next && Tok->Next->is(tok::identifier) && Tok->Next->Next && 1128 Tok->Next->Next->is(tok::equal))) { 1129 Tok->setType(TT_CSharpNullable); 1130 break; 1131 } 1132 } 1133 parseConditional(); 1134 break; 1135 case tok::kw_template: 1136 parseTemplateDeclaration(); 1137 break; 1138 case tok::comma: 1139 if (Contexts.back().InCtorInitializer) 1140 Tok->setType(TT_CtorInitializerComma); 1141 else if (Contexts.back().InInheritanceList) 1142 Tok->setType(TT_InheritanceComma); 1143 else if (Contexts.back().FirstStartOfName && 1144 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) { 1145 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 1146 Line.IsMultiVariableDeclStmt = true; 1147 } 1148 if (Contexts.back().IsForEachMacro) 1149 Contexts.back().IsExpression = true; 1150 break; 1151 case tok::identifier: 1152 if (Tok->isOneOf(Keywords.kw___has_include, 1153 Keywords.kw___has_include_next)) { 1154 parseHasInclude(); 1155 } 1156 if (Style.isCSharp() && Tok->is(Keywords.kw_where) && Tok->Next && 1157 Tok->Next->isNot(tok::l_paren)) { 1158 Tok->setType(TT_CSharpGenericTypeConstraint); 1159 parseCSharpGenericTypeConstraint(); 1160 } 1161 break; 1162 default: 1163 break; 1164 } 1165 return true; 1166 } 1167 1168 void parseCSharpGenericTypeConstraint() { 1169 int OpenAngleBracketsCount = 0; 1170 while (CurrentToken) { 1171 if (CurrentToken->is(tok::less)) { 1172 // parseAngle is too greedy and will consume the whole line. 1173 CurrentToken->setType(TT_TemplateOpener); 1174 ++OpenAngleBracketsCount; 1175 next(); 1176 } else if (CurrentToken->is(tok::greater)) { 1177 CurrentToken->setType(TT_TemplateCloser); 1178 --OpenAngleBracketsCount; 1179 next(); 1180 } else if (CurrentToken->is(tok::comma) && OpenAngleBracketsCount == 0) { 1181 // We allow line breaks after GenericTypeConstraintComma's 1182 // so do not flag commas in Generics as GenericTypeConstraintComma's. 1183 CurrentToken->setType(TT_CSharpGenericTypeConstraintComma); 1184 next(); 1185 } else if (CurrentToken->is(Keywords.kw_where)) { 1186 CurrentToken->setType(TT_CSharpGenericTypeConstraint); 1187 next(); 1188 } else if (CurrentToken->is(tok::colon)) { 1189 CurrentToken->setType(TT_CSharpGenericTypeConstraintColon); 1190 next(); 1191 } else { 1192 next(); 1193 } 1194 } 1195 } 1196 1197 void parseIncludeDirective() { 1198 if (CurrentToken && CurrentToken->is(tok::less)) { 1199 next(); 1200 while (CurrentToken) { 1201 // Mark tokens up to the trailing line comments as implicit string 1202 // literals. 1203 if (CurrentToken->isNot(tok::comment) && 1204 !CurrentToken->TokenText.startswith("//")) 1205 CurrentToken->setType(TT_ImplicitStringLiteral); 1206 next(); 1207 } 1208 } 1209 } 1210 1211 void parseWarningOrError() { 1212 next(); 1213 // We still want to format the whitespace left of the first token of the 1214 // warning or error. 1215 next(); 1216 while (CurrentToken) { 1217 CurrentToken->setType(TT_ImplicitStringLiteral); 1218 next(); 1219 } 1220 } 1221 1222 void parsePragma() { 1223 next(); // Consume "pragma". 1224 if (CurrentToken && 1225 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) { 1226 bool IsMark = CurrentToken->is(Keywords.kw_mark); 1227 next(); // Consume "mark". 1228 next(); // Consume first token (so we fix leading whitespace). 1229 while (CurrentToken) { 1230 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator)) 1231 CurrentToken->setType(TT_ImplicitStringLiteral); 1232 next(); 1233 } 1234 } 1235 } 1236 1237 void parseHasInclude() { 1238 if (!CurrentToken || !CurrentToken->is(tok::l_paren)) 1239 return; 1240 next(); // '(' 1241 parseIncludeDirective(); 1242 next(); // ')' 1243 } 1244 1245 LineType parsePreprocessorDirective() { 1246 bool IsFirstToken = CurrentToken->IsFirst; 1247 LineType Type = LT_PreprocessorDirective; 1248 next(); 1249 if (!CurrentToken) 1250 return Type; 1251 1252 if (Style.isJavaScript() && IsFirstToken) { 1253 // JavaScript files can contain shebang lines of the form: 1254 // #!/usr/bin/env node 1255 // Treat these like C++ #include directives. 1256 while (CurrentToken) { 1257 // Tokens cannot be comments here. 1258 CurrentToken->setType(TT_ImplicitStringLiteral); 1259 next(); 1260 } 1261 return LT_ImportStatement; 1262 } 1263 1264 if (CurrentToken->Tok.is(tok::numeric_constant)) { 1265 CurrentToken->SpacesRequiredBefore = 1; 1266 return Type; 1267 } 1268 // Hashes in the middle of a line can lead to any strange token 1269 // sequence. 1270 if (!CurrentToken->Tok.getIdentifierInfo()) 1271 return Type; 1272 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 1273 case tok::pp_include: 1274 case tok::pp_include_next: 1275 case tok::pp_import: 1276 next(); 1277 parseIncludeDirective(); 1278 Type = LT_ImportStatement; 1279 break; 1280 case tok::pp_error: 1281 case tok::pp_warning: 1282 parseWarningOrError(); 1283 break; 1284 case tok::pp_pragma: 1285 parsePragma(); 1286 break; 1287 case tok::pp_if: 1288 case tok::pp_elif: 1289 Contexts.back().IsExpression = true; 1290 next(); 1291 parseLine(); 1292 break; 1293 default: 1294 break; 1295 } 1296 while (CurrentToken) { 1297 FormatToken *Tok = CurrentToken; 1298 next(); 1299 if (Tok->is(tok::l_paren)) 1300 parseParens(); 1301 else if (Tok->isOneOf(Keywords.kw___has_include, 1302 Keywords.kw___has_include_next)) 1303 parseHasInclude(); 1304 } 1305 return Type; 1306 } 1307 1308 public: 1309 LineType parseLine() { 1310 if (!CurrentToken) 1311 return LT_Invalid; 1312 NonTemplateLess.clear(); 1313 if (CurrentToken->is(tok::hash)) 1314 return parsePreprocessorDirective(); 1315 1316 // Directly allow to 'import <string-literal>' to support protocol buffer 1317 // definitions (github.com/google/protobuf) or missing "#" (either way we 1318 // should not break the line). 1319 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 1320 if ((Style.Language == FormatStyle::LK_Java && 1321 CurrentToken->is(Keywords.kw_package)) || 1322 (Info && Info->getPPKeywordID() == tok::pp_import && 1323 CurrentToken->Next && 1324 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier, 1325 tok::kw_static))) { 1326 next(); 1327 parseIncludeDirective(); 1328 return LT_ImportStatement; 1329 } 1330 1331 // If this line starts and ends in '<' and '>', respectively, it is likely 1332 // part of "#define <a/b.h>". 1333 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 1334 parseIncludeDirective(); 1335 return LT_ImportStatement; 1336 } 1337 1338 // In .proto files, top-level options and package statements are very 1339 // similar to import statements and should not be line-wrapped. 1340 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 && 1341 CurrentToken->isOneOf(Keywords.kw_option, Keywords.kw_package)) { 1342 next(); 1343 if (CurrentToken && CurrentToken->is(tok::identifier)) { 1344 while (CurrentToken) 1345 next(); 1346 return LT_ImportStatement; 1347 } 1348 } 1349 1350 bool KeywordVirtualFound = false; 1351 bool ImportStatement = false; 1352 1353 // import {...} from '...'; 1354 if (Style.isJavaScript() && CurrentToken->is(Keywords.kw_import)) 1355 ImportStatement = true; 1356 1357 while (CurrentToken) { 1358 if (CurrentToken->is(tok::kw_virtual)) 1359 KeywordVirtualFound = true; 1360 if (Style.isJavaScript()) { 1361 // export {...} from '...'; 1362 // An export followed by "from 'some string';" is a re-export from 1363 // another module identified by a URI and is treated as a 1364 // LT_ImportStatement (i.e. prevent wraps on it for long URIs). 1365 // Just "export {...};" or "export class ..." should not be treated as 1366 // an import in this sense. 1367 if (Line.First->is(tok::kw_export) && 1368 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next && 1369 CurrentToken->Next->isStringLiteral()) 1370 ImportStatement = true; 1371 if (isClosureImportStatement(*CurrentToken)) 1372 ImportStatement = true; 1373 } 1374 if (!consumeToken()) 1375 return LT_Invalid; 1376 } 1377 if (KeywordVirtualFound) 1378 return LT_VirtualFunctionDecl; 1379 if (ImportStatement) 1380 return LT_ImportStatement; 1381 1382 if (Line.startsWith(TT_ObjCMethodSpecifier)) { 1383 if (Contexts.back().FirstObjCSelectorName) 1384 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 1385 Contexts.back().LongestObjCSelectorName; 1386 return LT_ObjCMethodDecl; 1387 } 1388 1389 for (const auto &ctx : Contexts) { 1390 if (ctx.InStructArrayInitializer) { 1391 return LT_ArrayOfStructInitializer; 1392 } 1393 } 1394 1395 return LT_Other; 1396 } 1397 1398 private: 1399 bool isClosureImportStatement(const FormatToken &Tok) { 1400 // FIXME: Closure-library specific stuff should not be hard-coded but be 1401 // configurable. 1402 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) && 1403 Tok.Next->Next && 1404 (Tok.Next->Next->TokenText == "module" || 1405 Tok.Next->Next->TokenText == "provide" || 1406 Tok.Next->Next->TokenText == "require" || 1407 Tok.Next->Next->TokenText == "requireType" || 1408 Tok.Next->Next->TokenText == "forwardDeclare") && 1409 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren); 1410 } 1411 1412 void resetTokenMetadata() { 1413 if (!CurrentToken) 1414 return; 1415 1416 // Reset token type in case we have already looked at it and then 1417 // recovered from an error (e.g. failure to find the matching >). 1418 if (!CurrentToken->isOneOf( 1419 TT_LambdaLSquare, TT_LambdaLBrace, TT_AttributeMacro, TT_IfMacro, 1420 TT_ForEachMacro, TT_TypenameMacro, TT_FunctionLBrace, 1421 TT_ImplicitStringLiteral, TT_InlineASMBrace, TT_FatArrow, 1422 TT_LambdaArrow, TT_NamespaceMacro, TT_OverloadedOperator, 1423 TT_RegexLiteral, TT_TemplateString, TT_ObjCStringLiteral, 1424 TT_UntouchableMacroFunc, TT_ConstraintJunctions, 1425 TT_StatementAttributeLikeMacro)) 1426 CurrentToken->setType(TT_Unknown); 1427 CurrentToken->Role.reset(); 1428 CurrentToken->MatchingParen = nullptr; 1429 CurrentToken->FakeLParens.clear(); 1430 CurrentToken->FakeRParens = 0; 1431 } 1432 1433 void next() { 1434 if (!CurrentToken) 1435 return; 1436 1437 CurrentToken->NestingLevel = Contexts.size() - 1; 1438 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 1439 modifyContext(*CurrentToken); 1440 determineTokenType(*CurrentToken); 1441 CurrentToken = CurrentToken->Next; 1442 1443 resetTokenMetadata(); 1444 } 1445 1446 /// A struct to hold information valid in a specific context, e.g. 1447 /// a pair of parenthesis. 1448 struct Context { 1449 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 1450 bool IsExpression) 1451 : ContextKind(ContextKind), BindingStrength(BindingStrength), 1452 IsExpression(IsExpression) {} 1453 1454 tok::TokenKind ContextKind; 1455 unsigned BindingStrength; 1456 bool IsExpression; 1457 unsigned LongestObjCSelectorName = 0; 1458 bool ColonIsForRangeExpr = false; 1459 bool ColonIsDictLiteral = false; 1460 bool ColonIsObjCMethodExpr = false; 1461 FormatToken *FirstObjCSelectorName = nullptr; 1462 FormatToken *FirstStartOfName = nullptr; 1463 bool CanBeExpression = true; 1464 bool InTemplateArgument = false; 1465 bool InCtorInitializer = false; 1466 bool InInheritanceList = false; 1467 bool CaretFound = false; 1468 bool IsForEachMacro = false; 1469 bool InCpp11AttributeSpecifier = false; 1470 bool InCSharpAttributeSpecifier = false; 1471 bool InStructArrayInitializer = false; 1472 }; 1473 1474 /// Puts a new \c Context onto the stack \c Contexts for the lifetime 1475 /// of each instance. 1476 struct ScopedContextCreator { 1477 AnnotatingParser &P; 1478 1479 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 1480 unsigned Increase) 1481 : P(P) { 1482 P.Contexts.push_back(Context(ContextKind, 1483 P.Contexts.back().BindingStrength + Increase, 1484 P.Contexts.back().IsExpression)); 1485 } 1486 1487 ~ScopedContextCreator() { 1488 if (P.Style.AlignArrayOfStructures != FormatStyle::AIAS_None) { 1489 if (P.Contexts.back().InStructArrayInitializer) { 1490 P.Contexts.pop_back(); 1491 P.Contexts.back().InStructArrayInitializer = true; 1492 return; 1493 } 1494 } 1495 P.Contexts.pop_back(); 1496 } 1497 }; 1498 1499 void modifyContext(const FormatToken &Current) { 1500 if (Current.getPrecedence() == prec::Assignment && 1501 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) && 1502 // Type aliases use `type X = ...;` in TypeScript and can be exported 1503 // using `export type ...`. 1504 !(Style.isJavaScript() && 1505 (Line.startsWith(Keywords.kw_type, tok::identifier) || 1506 Line.startsWith(tok::kw_export, Keywords.kw_type, 1507 tok::identifier))) && 1508 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 1509 Contexts.back().IsExpression = true; 1510 if (!Line.startsWith(TT_UnaryOperator)) { 1511 for (FormatToken *Previous = Current.Previous; 1512 Previous && Previous->Previous && 1513 !Previous->Previous->isOneOf(tok::comma, tok::semi); 1514 Previous = Previous->Previous) { 1515 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 1516 Previous = Previous->MatchingParen; 1517 if (!Previous) 1518 break; 1519 } 1520 if (Previous->opensScope()) 1521 break; 1522 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) && 1523 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) && 1524 Previous->Previous && Previous->Previous->isNot(tok::equal)) 1525 Previous->setType(TT_PointerOrReference); 1526 } 1527 } 1528 } else if (Current.is(tok::lessless) && 1529 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) { 1530 Contexts.back().IsExpression = true; 1531 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 1532 Contexts.back().IsExpression = true; 1533 } else if (Current.is(TT_TrailingReturnArrow)) { 1534 Contexts.back().IsExpression = false; 1535 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) { 1536 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java; 1537 } else if (Current.Previous && 1538 Current.Previous->is(TT_CtorInitializerColon)) { 1539 Contexts.back().IsExpression = true; 1540 Contexts.back().InCtorInitializer = true; 1541 } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) { 1542 Contexts.back().InInheritanceList = true; 1543 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 1544 for (FormatToken *Previous = Current.Previous; 1545 Previous && Previous->isOneOf(tok::star, tok::amp); 1546 Previous = Previous->Previous) 1547 Previous->setType(TT_PointerOrReference); 1548 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer) 1549 Contexts.back().IsExpression = false; 1550 } else if (Current.is(tok::kw_new)) { 1551 Contexts.back().CanBeExpression = false; 1552 } else if (Current.is(tok::semi) || 1553 (Current.is(tok::exclaim) && Current.Previous && 1554 !Current.Previous->is(tok::kw_operator))) { 1555 // This should be the condition or increment in a for-loop. 1556 // But not operator !() (can't use TT_OverloadedOperator here as its not 1557 // been annotated yet). 1558 Contexts.back().IsExpression = true; 1559 } 1560 } 1561 1562 static FormatToken *untilMatchingParen(FormatToken *Current) { 1563 // Used when `MatchingParen` is not yet established. 1564 int ParenLevel = 0; 1565 while (Current) { 1566 if (Current->is(tok::l_paren)) 1567 ParenLevel++; 1568 if (Current->is(tok::r_paren)) 1569 ParenLevel--; 1570 if (ParenLevel < 1) 1571 break; 1572 Current = Current->Next; 1573 } 1574 return Current; 1575 } 1576 1577 static bool isDeductionGuide(FormatToken &Current) { 1578 // Look for a deduction guide template<T> A(...) -> A<...>; 1579 if (Current.Previous && Current.Previous->is(tok::r_paren) && 1580 Current.startsSequence(tok::arrow, tok::identifier, tok::less)) { 1581 // Find the TemplateCloser. 1582 FormatToken *TemplateCloser = Current.Next->Next; 1583 int NestingLevel = 0; 1584 while (TemplateCloser) { 1585 // Skip over an expressions in parens A<(3 < 2)>; 1586 if (TemplateCloser->is(tok::l_paren)) { 1587 // No Matching Paren yet so skip to matching paren 1588 TemplateCloser = untilMatchingParen(TemplateCloser); 1589 if (!TemplateCloser) 1590 break; 1591 } 1592 if (TemplateCloser->is(tok::less)) 1593 NestingLevel++; 1594 if (TemplateCloser->is(tok::greater)) 1595 NestingLevel--; 1596 if (NestingLevel < 1) 1597 break; 1598 TemplateCloser = TemplateCloser->Next; 1599 } 1600 // Assuming we have found the end of the template ensure its followed 1601 // with a semi-colon. 1602 if (TemplateCloser && TemplateCloser->Next && 1603 TemplateCloser->Next->is(tok::semi) && 1604 Current.Previous->MatchingParen) { 1605 // Determine if the identifier `A` prior to the A<..>; is the same as 1606 // prior to the A(..) 1607 FormatToken *LeadingIdentifier = 1608 Current.Previous->MatchingParen->Previous; 1609 1610 // Differentiate a deduction guide by seeing the 1611 // > of the template prior to the leading identifier. 1612 if (LeadingIdentifier) { 1613 FormatToken *PriorLeadingIdentifier = LeadingIdentifier->Previous; 1614 // Skip back past explicit decoration 1615 if (PriorLeadingIdentifier && 1616 PriorLeadingIdentifier->is(tok::kw_explicit)) 1617 PriorLeadingIdentifier = PriorLeadingIdentifier->Previous; 1618 1619 return (PriorLeadingIdentifier && 1620 PriorLeadingIdentifier->is(TT_TemplateCloser) && 1621 LeadingIdentifier->TokenText == Current.Next->TokenText); 1622 } 1623 } 1624 } 1625 return false; 1626 } 1627 1628 void determineTokenType(FormatToken &Current) { 1629 if (!Current.is(TT_Unknown)) 1630 // The token type is already known. 1631 return; 1632 1633 if ((Style.isJavaScript() || Style.isCSharp()) && 1634 Current.is(tok::exclaim)) { 1635 if (Current.Previous) { 1636 bool IsIdentifier = 1637 Style.isJavaScript() 1638 ? Keywords.IsJavaScriptIdentifier( 1639 *Current.Previous, /* AcceptIdentifierName= */ true) 1640 : Current.Previous->is(tok::identifier); 1641 if (IsIdentifier || 1642 Current.Previous->isOneOf( 1643 tok::kw_namespace, tok::r_paren, tok::r_square, tok::r_brace, 1644 tok::kw_false, tok::kw_true, Keywords.kw_type, Keywords.kw_get, 1645 Keywords.kw_set) || 1646 Current.Previous->Tok.isLiteral()) { 1647 Current.setType(TT_NonNullAssertion); 1648 return; 1649 } 1650 } 1651 if (Current.Next && 1652 Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) { 1653 Current.setType(TT_NonNullAssertion); 1654 return; 1655 } 1656 } 1657 1658 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1659 // function declaration have been found. In this case, 'Current' is a 1660 // trailing token of this declaration and thus cannot be a name. 1661 if (Current.is(Keywords.kw_instanceof)) { 1662 Current.setType(TT_BinaryOperator); 1663 } else if (isStartOfName(Current) && 1664 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 1665 Contexts.back().FirstStartOfName = &Current; 1666 Current.setType(TT_StartOfName); 1667 } else if (Current.is(tok::semi)) { 1668 // Reset FirstStartOfName after finding a semicolon so that a for loop 1669 // with multiple increment statements is not confused with a for loop 1670 // having multiple variable declarations. 1671 Contexts.back().FirstStartOfName = nullptr; 1672 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) { 1673 AutoFound = true; 1674 } else if (Current.is(tok::arrow) && 1675 Style.Language == FormatStyle::LK_Java) { 1676 Current.setType(TT_LambdaArrow); 1677 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration && 1678 Current.NestingLevel == 0 && 1679 !Current.Previous->isOneOf(tok::kw_operator, tok::identifier)) { 1680 // not auto operator->() -> xxx; 1681 Current.setType(TT_TrailingReturnArrow); 1682 } else if (Current.is(tok::arrow) && Current.Previous && 1683 Current.Previous->is(tok::r_brace)) { 1684 // Concept implicit conversion constraint needs to be treated like 1685 // a trailing return type ... } -> <type>. 1686 Current.setType(TT_TrailingReturnArrow); 1687 } else if (isDeductionGuide(Current)) { 1688 // Deduction guides trailing arrow " A(...) -> A<T>;". 1689 Current.setType(TT_TrailingReturnArrow); 1690 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 1691 Current.setType(determineStarAmpUsage( 1692 Current, 1693 Contexts.back().CanBeExpression && Contexts.back().IsExpression, 1694 Contexts.back().InTemplateArgument)); 1695 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 1696 Current.setType(determinePlusMinusCaretUsage(Current)); 1697 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret)) 1698 Contexts.back().CaretFound = true; 1699 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 1700 Current.setType(determineIncrementUsage(Current)); 1701 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 1702 Current.setType(TT_UnaryOperator); 1703 } else if (Current.is(tok::question)) { 1704 if (Style.isJavaScript() && Line.MustBeDeclaration && 1705 !Contexts.back().IsExpression) { 1706 // In JavaScript, `interface X { foo?(): bar; }` is an optional method 1707 // on the interface, not a ternary expression. 1708 Current.setType(TT_JsTypeOptionalQuestion); 1709 } else { 1710 Current.setType(TT_ConditionalExpr); 1711 } 1712 } else if (Current.isBinaryOperator() && 1713 (!Current.Previous || Current.Previous->isNot(tok::l_square)) && 1714 (!Current.is(tok::greater) && 1715 Style.Language != FormatStyle::LK_TextProto)) { 1716 Current.setType(TT_BinaryOperator); 1717 } else if (Current.is(tok::comment)) { 1718 if (Current.TokenText.startswith("/*")) { 1719 if (Current.TokenText.endswith("*/")) 1720 Current.setType(TT_BlockComment); 1721 else 1722 // The lexer has for some reason determined a comment here. But we 1723 // cannot really handle it, if it isn't properly terminated. 1724 Current.Tok.setKind(tok::unknown); 1725 } else { 1726 Current.setType(TT_LineComment); 1727 } 1728 } else if (Current.is(tok::r_paren)) { 1729 if (rParenEndsCast(Current)) 1730 Current.setType(TT_CastRParen); 1731 if (Current.MatchingParen && Current.Next && 1732 !Current.Next->isBinaryOperator() && 1733 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace, 1734 tok::comma, tok::period, tok::arrow, 1735 tok::coloncolon)) 1736 if (FormatToken *AfterParen = Current.MatchingParen->Next) { 1737 // Make sure this isn't the return type of an Obj-C block declaration 1738 if (AfterParen->Tok.isNot(tok::caret)) { 1739 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) 1740 if (BeforeParen->is(tok::identifier) && 1741 !BeforeParen->is(TT_TypenameMacro) && 1742 BeforeParen->TokenText == BeforeParen->TokenText.upper() && 1743 (!BeforeParen->Previous || 1744 BeforeParen->Previous->ClosesTemplateDeclaration)) 1745 Current.setType(TT_FunctionAnnotationRParen); 1746 } 1747 } 1748 } else if (Current.is(tok::at) && Current.Next && !Style.isJavaScript() && 1749 Style.Language != FormatStyle::LK_Java) { 1750 // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it 1751 // marks declarations and properties that need special formatting. 1752 switch (Current.Next->Tok.getObjCKeywordID()) { 1753 case tok::objc_interface: 1754 case tok::objc_implementation: 1755 case tok::objc_protocol: 1756 Current.setType(TT_ObjCDecl); 1757 break; 1758 case tok::objc_property: 1759 Current.setType(TT_ObjCProperty); 1760 break; 1761 default: 1762 break; 1763 } 1764 } else if (Current.is(tok::period)) { 1765 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 1766 if (PreviousNoComment && 1767 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 1768 Current.setType(TT_DesignatedInitializerPeriod); 1769 else if (Style.Language == FormatStyle::LK_Java && Current.Previous && 1770 Current.Previous->isOneOf(TT_JavaAnnotation, 1771 TT_LeadingJavaAnnotation)) { 1772 Current.setType(Current.Previous->getType()); 1773 } 1774 } else if (canBeObjCSelectorComponent(Current) && 1775 // FIXME(bug 36976): ObjC return types shouldn't use 1776 // TT_CastRParen. 1777 Current.Previous && Current.Previous->is(TT_CastRParen) && 1778 Current.Previous->MatchingParen && 1779 Current.Previous->MatchingParen->Previous && 1780 Current.Previous->MatchingParen->Previous->is( 1781 TT_ObjCMethodSpecifier)) { 1782 // This is the first part of an Objective-C selector name. (If there's no 1783 // colon after this, this is the only place which annotates the identifier 1784 // as a selector.) 1785 Current.setType(TT_SelectorName); 1786 } else if (Current.isOneOf(tok::identifier, tok::kw_const, tok::kw_noexcept, 1787 tok::kw_requires) && 1788 Current.Previous && 1789 !Current.Previous->isOneOf(tok::equal, tok::at) && 1790 Line.MightBeFunctionDecl && Contexts.size() == 1) { 1791 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1792 // function declaration have been found. 1793 Current.setType(TT_TrailingAnnotation); 1794 } else if ((Style.Language == FormatStyle::LK_Java || 1795 Style.isJavaScript()) && 1796 Current.Previous) { 1797 if (Current.Previous->is(tok::at) && 1798 Current.isNot(Keywords.kw_interface)) { 1799 const FormatToken &AtToken = *Current.Previous; 1800 const FormatToken *Previous = AtToken.getPreviousNonComment(); 1801 if (!Previous || Previous->is(TT_LeadingJavaAnnotation)) 1802 Current.setType(TT_LeadingJavaAnnotation); 1803 else 1804 Current.setType(TT_JavaAnnotation); 1805 } else if (Current.Previous->is(tok::period) && 1806 Current.Previous->isOneOf(TT_JavaAnnotation, 1807 TT_LeadingJavaAnnotation)) { 1808 Current.setType(Current.Previous->getType()); 1809 } 1810 } 1811 } 1812 1813 /// Take a guess at whether \p Tok starts a name of a function or 1814 /// variable declaration. 1815 /// 1816 /// This is a heuristic based on whether \p Tok is an identifier following 1817 /// something that is likely a type. 1818 bool isStartOfName(const FormatToken &Tok) { 1819 if (Tok.isNot(tok::identifier) || !Tok.Previous) 1820 return false; 1821 1822 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof, 1823 Keywords.kw_as)) 1824 return false; 1825 if (Style.isJavaScript() && Tok.Previous->is(Keywords.kw_in)) 1826 return false; 1827 1828 // Skip "const" as it does not have an influence on whether this is a name. 1829 FormatToken *PreviousNotConst = Tok.getPreviousNonComment(); 1830 1831 // For javascript const can be like "let" or "var" 1832 if (!Style.isJavaScript()) 1833 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 1834 PreviousNotConst = PreviousNotConst->getPreviousNonComment(); 1835 1836 if (!PreviousNotConst) 1837 return false; 1838 1839 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 1840 PreviousNotConst->Previous && 1841 PreviousNotConst->Previous->is(tok::hash); 1842 1843 if (PreviousNotConst->is(TT_TemplateCloser)) 1844 return PreviousNotConst && PreviousNotConst->MatchingParen && 1845 PreviousNotConst->MatchingParen->Previous && 1846 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) && 1847 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 1848 1849 if (PreviousNotConst->is(tok::r_paren) && 1850 PreviousNotConst->is(TT_TypeDeclarationParen)) 1851 return true; 1852 1853 // If is a preprocess keyword like #define. 1854 if (IsPPKeyword) 1855 return false; 1856 1857 // int a or auto a. 1858 if (PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) 1859 return true; 1860 1861 // *a or &a or &&a. 1862 if (PreviousNotConst->is(TT_PointerOrReference)) 1863 return true; 1864 1865 // MyClass a; 1866 if (PreviousNotConst->isSimpleTypeSpecifier()) 1867 return true; 1868 1869 // const a = in JavaScript. 1870 return (Style.isJavaScript() && PreviousNotConst->is(tok::kw_const)); 1871 } 1872 1873 /// Determine whether ')' is ending a cast. 1874 bool rParenEndsCast(const FormatToken &Tok) { 1875 // C-style casts are only used in C++, C# and Java. 1876 if (!Style.isCSharp() && !Style.isCpp() && 1877 Style.Language != FormatStyle::LK_Java) 1878 return false; 1879 1880 // Empty parens aren't casts and there are no casts at the end of the line. 1881 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen) 1882 return false; 1883 1884 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 1885 if (LeftOfParens) { 1886 // If there is a closing parenthesis left of the current parentheses, 1887 // look past it as these might be chained casts. 1888 if (LeftOfParens->is(tok::r_paren)) { 1889 if (!LeftOfParens->MatchingParen || 1890 !LeftOfParens->MatchingParen->Previous) 1891 return false; 1892 LeftOfParens = LeftOfParens->MatchingParen->Previous; 1893 } 1894 1895 // If there is an identifier (or with a few exceptions a keyword) right 1896 // before the parentheses, this is unlikely to be a cast. 1897 if (LeftOfParens->Tok.getIdentifierInfo() && 1898 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case, 1899 tok::kw_delete)) 1900 return false; 1901 1902 // Certain other tokens right before the parentheses are also signals that 1903 // this cannot be a cast. 1904 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator, 1905 TT_TemplateCloser, tok::ellipsis)) 1906 return false; 1907 } 1908 1909 if (Tok.Next->is(tok::question)) 1910 return false; 1911 1912 // `foreach((A a, B b) in someList)` should not be seen as a cast. 1913 if (Tok.Next->is(Keywords.kw_in) && Style.isCSharp()) 1914 return false; 1915 1916 // Functions which end with decorations like volatile, noexcept are unlikely 1917 // to be casts. 1918 if (Tok.Next->isOneOf(tok::kw_noexcept, tok::kw_volatile, tok::kw_const, 1919 tok::kw_requires, tok::kw_throw, tok::arrow, 1920 Keywords.kw_override, Keywords.kw_final) || 1921 isCpp11AttributeSpecifier(*Tok.Next)) 1922 return false; 1923 1924 // As Java has no function types, a "(" after the ")" likely means that this 1925 // is a cast. 1926 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren)) 1927 return true; 1928 1929 // If a (non-string) literal follows, this is likely a cast. 1930 if (Tok.Next->isNot(tok::string_literal) && 1931 (Tok.Next->Tok.isLiteral() || 1932 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 1933 return true; 1934 1935 // Heuristically try to determine whether the parentheses contain a type. 1936 auto IsQualifiedPointerOrReference = [](FormatToken *T) { 1937 // This is used to handle cases such as x = (foo *const)&y; 1938 assert(!T->isSimpleTypeSpecifier() && "Should have already been checked"); 1939 // Strip trailing qualifiers such as const or volatile when checking 1940 // whether the parens could be a cast to a pointer/reference type. 1941 while (T) { 1942 if (T->is(TT_AttributeParen)) { 1943 // Handle `x = (foo *__attribute__((foo)))&v;`: 1944 if (T->MatchingParen && T->MatchingParen->Previous && 1945 T->MatchingParen->Previous->is(tok::kw___attribute)) { 1946 T = T->MatchingParen->Previous->Previous; 1947 continue; 1948 } 1949 } else if (T->is(TT_AttributeSquare)) { 1950 // Handle `x = (foo *[[clang::foo]])&v;`: 1951 if (T->MatchingParen && T->MatchingParen->Previous) { 1952 T = T->MatchingParen->Previous; 1953 continue; 1954 } 1955 } else if (T->canBePointerOrReferenceQualifier()) { 1956 T = T->Previous; 1957 continue; 1958 } 1959 break; 1960 } 1961 return T && T->is(TT_PointerOrReference); 1962 }; 1963 bool ParensAreType = 1964 !Tok.Previous || 1965 Tok.Previous->isOneOf(TT_TemplateCloser, TT_TypeDeclarationParen) || 1966 Tok.Previous->isSimpleTypeSpecifier() || 1967 IsQualifiedPointerOrReference(Tok.Previous); 1968 bool ParensCouldEndDecl = 1969 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater); 1970 if (ParensAreType && !ParensCouldEndDecl) 1971 return true; 1972 1973 // At this point, we heuristically assume that there are no casts at the 1974 // start of the line. We assume that we have found most cases where there 1975 // are by the logic above, e.g. "(void)x;". 1976 if (!LeftOfParens) 1977 return false; 1978 1979 // Certain token types inside the parentheses mean that this can't be a 1980 // cast. 1981 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok; 1982 Token = Token->Next) 1983 if (Token->is(TT_BinaryOperator)) 1984 return false; 1985 1986 // If the following token is an identifier or 'this', this is a cast. All 1987 // cases where this can be something else are handled above. 1988 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this)) 1989 return true; 1990 1991 // Look for a cast `( x ) (`. 1992 if (Tok.Next->is(tok::l_paren) && Tok.Previous && Tok.Previous->Previous) { 1993 if (Tok.Previous->is(tok::identifier) && 1994 Tok.Previous->Previous->is(tok::l_paren)) 1995 return true; 1996 } 1997 1998 if (!Tok.Next->Next) 1999 return false; 2000 2001 // If the next token after the parenthesis is a unary operator, assume 2002 // that this is cast, unless there are unexpected tokens inside the 2003 // parenthesis. 2004 bool NextIsUnary = 2005 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star); 2006 if (!NextIsUnary || Tok.Next->is(tok::plus) || 2007 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) 2008 return false; 2009 // Search for unexpected tokens. 2010 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen; 2011 Prev = Prev->Previous) { 2012 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon)) 2013 return false; 2014 } 2015 return true; 2016 } 2017 2018 /// Return the type of the given token assuming it is * or &. 2019 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 2020 bool InTemplateArgument) { 2021 if (Style.isJavaScript()) 2022 return TT_BinaryOperator; 2023 2024 // && in C# must be a binary operator. 2025 if (Style.isCSharp() && Tok.is(tok::ampamp)) 2026 return TT_BinaryOperator; 2027 2028 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 2029 if (!PrevToken) 2030 return TT_UnaryOperator; 2031 2032 const FormatToken *NextToken = Tok.getNextNonComment(); 2033 if (!NextToken || 2034 NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_noexcept) || 2035 NextToken->canBePointerOrReferenceQualifier() || 2036 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) 2037 return TT_PointerOrReference; 2038 2039 if (PrevToken->is(tok::coloncolon)) 2040 return TT_PointerOrReference; 2041 2042 if (PrevToken->is(tok::r_paren) && PrevToken->is(TT_TypeDeclarationParen)) 2043 return TT_PointerOrReference; 2044 2045 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 2046 tok::comma, tok::semi, tok::kw_return, tok::colon, 2047 tok::kw_co_return, tok::kw_co_await, 2048 tok::kw_co_yield, tok::equal, tok::kw_delete, 2049 tok::kw_sizeof, tok::kw_throw, TT_BinaryOperator, 2050 TT_ConditionalExpr, TT_UnaryOperator, TT_CastRParen)) 2051 return TT_UnaryOperator; 2052 2053 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare)) 2054 return TT_PointerOrReference; 2055 if (NextToken->is(tok::kw_operator) && !IsExpression) 2056 return TT_PointerOrReference; 2057 if (NextToken->isOneOf(tok::comma, tok::semi)) 2058 return TT_PointerOrReference; 2059 2060 if (PrevToken->Tok.isLiteral() || 2061 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 2062 tok::kw_false, tok::r_brace) || 2063 NextToken->Tok.isLiteral() || 2064 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 2065 NextToken->isUnaryOperator() || 2066 // If we know we're in a template argument, there are no named 2067 // declarations. Thus, having an identifier on the right-hand side 2068 // indicates a binary operator. 2069 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 2070 return TT_BinaryOperator; 2071 2072 // "&&(" is quite unlikely to be two successive unary "&". 2073 if (Tok.is(tok::ampamp) && NextToken->is(tok::l_paren)) 2074 return TT_BinaryOperator; 2075 2076 // This catches some cases where evaluation order is used as control flow: 2077 // aaa && aaa->f(); 2078 if (NextToken->Tok.isAnyIdentifier()) { 2079 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 2080 if (NextNextToken && NextNextToken->is(tok::arrow)) 2081 return TT_BinaryOperator; 2082 } 2083 2084 // It is very unlikely that we are going to find a pointer or reference type 2085 // definition on the RHS of an assignment. 2086 if (IsExpression && !Contexts.back().CaretFound) 2087 return TT_BinaryOperator; 2088 2089 return TT_PointerOrReference; 2090 } 2091 2092 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 2093 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 2094 if (!PrevToken) 2095 return TT_UnaryOperator; 2096 2097 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator)) 2098 // This must be a sequence of leading unary operators. 2099 return TT_UnaryOperator; 2100 2101 // Use heuristics to recognize unary operators. 2102 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 2103 tok::question, tok::colon, tok::kw_return, 2104 tok::kw_case, tok::at, tok::l_brace, tok::kw_throw, 2105 tok::kw_co_return, tok::kw_co_yield)) 2106 return TT_UnaryOperator; 2107 2108 // There can't be two consecutive binary operators. 2109 if (PrevToken->is(TT_BinaryOperator)) 2110 return TT_UnaryOperator; 2111 2112 // Fall back to marking the token as binary operator. 2113 return TT_BinaryOperator; 2114 } 2115 2116 /// Determine whether ++/-- are pre- or post-increments/-decrements. 2117 TokenType determineIncrementUsage(const FormatToken &Tok) { 2118 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 2119 if (!PrevToken || PrevToken->is(TT_CastRParen)) 2120 return TT_UnaryOperator; 2121 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 2122 return TT_TrailingUnaryOperator; 2123 2124 return TT_UnaryOperator; 2125 } 2126 2127 SmallVector<Context, 8> Contexts; 2128 2129 const FormatStyle &Style; 2130 AnnotatedLine &Line; 2131 FormatToken *CurrentToken; 2132 bool AutoFound; 2133 const AdditionalKeywords &Keywords; 2134 2135 // Set of "<" tokens that do not open a template parameter list. If parseAngle 2136 // determines that a specific token can't be a template opener, it will make 2137 // same decision irrespective of the decisions for tokens leading up to it. 2138 // Store this information to prevent this from causing exponential runtime. 2139 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess; 2140 }; 2141 2142 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1; 2143 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 2144 2145 /// Parses binary expressions by inserting fake parenthesis based on 2146 /// operator precedence. 2147 class ExpressionParser { 2148 public: 2149 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords, 2150 AnnotatedLine &Line) 2151 : Style(Style), Keywords(Keywords), Current(Line.First) {} 2152 2153 /// Parse expressions with the given operator precedence. 2154 void parse(int Precedence = 0) { 2155 // Skip 'return' and ObjC selector colons as they are not part of a binary 2156 // expression. 2157 while (Current && (Current->is(tok::kw_return) || 2158 (Current->is(tok::colon) && 2159 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) 2160 next(); 2161 2162 if (!Current || Precedence > PrecedenceArrowAndPeriod) 2163 return; 2164 2165 // Conditional expressions need to be parsed separately for proper nesting. 2166 if (Precedence == prec::Conditional) { 2167 parseConditionalExpr(); 2168 return; 2169 } 2170 2171 // Parse unary operators, which all have a higher precedence than binary 2172 // operators. 2173 if (Precedence == PrecedenceUnaryOperator) { 2174 parseUnaryOperator(); 2175 return; 2176 } 2177 2178 FormatToken *Start = Current; 2179 FormatToken *LatestOperator = nullptr; 2180 unsigned OperatorIndex = 0; 2181 2182 while (Current) { 2183 // Consume operators with higher precedence. 2184 parse(Precedence + 1); 2185 2186 int CurrentPrecedence = getCurrentPrecedence(); 2187 2188 if (Precedence == CurrentPrecedence && Current && 2189 Current->is(TT_SelectorName)) { 2190 if (LatestOperator) 2191 addFakeParenthesis(Start, prec::Level(Precedence)); 2192 Start = Current; 2193 } 2194 2195 // At the end of the line or when an operator with higher precedence is 2196 // found, insert fake parenthesis and return. 2197 if (!Current || 2198 (Current->closesScope() && 2199 (Current->MatchingParen || Current->is(TT_TemplateString))) || 2200 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) || 2201 (CurrentPrecedence == prec::Conditional && 2202 Precedence == prec::Assignment && Current->is(tok::colon))) { 2203 break; 2204 } 2205 2206 // Consume scopes: (), [], <> and {} 2207 if (Current->opensScope()) { 2208 // In fragment of a JavaScript template string can look like '}..${' and 2209 // thus close a scope and open a new one at the same time. 2210 while (Current && (!Current->closesScope() || Current->opensScope())) { 2211 next(); 2212 parse(); 2213 } 2214 next(); 2215 } else { 2216 // Operator found. 2217 if (CurrentPrecedence == Precedence) { 2218 if (LatestOperator) 2219 LatestOperator->NextOperator = Current; 2220 LatestOperator = Current; 2221 Current->OperatorIndex = OperatorIndex; 2222 ++OperatorIndex; 2223 } 2224 next(/*SkipPastLeadingComments=*/Precedence > 0); 2225 } 2226 } 2227 2228 if (LatestOperator && (Current || Precedence > 0)) { 2229 // LatestOperator->LastOperator = true; 2230 if (Precedence == PrecedenceArrowAndPeriod) { 2231 // Call expressions don't have a binary operator precedence. 2232 addFakeParenthesis(Start, prec::Unknown); 2233 } else { 2234 addFakeParenthesis(Start, prec::Level(Precedence)); 2235 } 2236 } 2237 } 2238 2239 private: 2240 /// Gets the precedence (+1) of the given token for binary operators 2241 /// and other tokens that we treat like binary operators. 2242 int getCurrentPrecedence() { 2243 if (Current) { 2244 const FormatToken *NextNonComment = Current->getNextNonComment(); 2245 if (Current->is(TT_ConditionalExpr)) 2246 return prec::Conditional; 2247 if (NextNonComment && Current->is(TT_SelectorName) && 2248 (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) || 2249 ((Style.Language == FormatStyle::LK_Proto || 2250 Style.Language == FormatStyle::LK_TextProto) && 2251 NextNonComment->is(tok::less)))) 2252 return prec::Assignment; 2253 if (Current->is(TT_JsComputedPropertyName)) 2254 return prec::Assignment; 2255 if (Current->is(TT_LambdaArrow)) 2256 return prec::Comma; 2257 if (Current->is(TT_FatArrow)) 2258 return prec::Assignment; 2259 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) || 2260 (Current->is(tok::comment) && NextNonComment && 2261 NextNonComment->is(TT_SelectorName))) 2262 return 0; 2263 if (Current->is(TT_RangeBasedForLoopColon)) 2264 return prec::Comma; 2265 if ((Style.Language == FormatStyle::LK_Java || Style.isJavaScript()) && 2266 Current->is(Keywords.kw_instanceof)) 2267 return prec::Relational; 2268 if (Style.isJavaScript() && 2269 Current->isOneOf(Keywords.kw_in, Keywords.kw_as)) 2270 return prec::Relational; 2271 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma)) 2272 return Current->getPrecedence(); 2273 if (Current->isOneOf(tok::period, tok::arrow)) 2274 return PrecedenceArrowAndPeriod; 2275 if ((Style.Language == FormatStyle::LK_Java || Style.isJavaScript()) && 2276 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements, 2277 Keywords.kw_throws)) 2278 return 0; 2279 } 2280 return -1; 2281 } 2282 2283 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 2284 Start->FakeLParens.push_back(Precedence); 2285 if (Precedence > prec::Unknown) 2286 Start->StartsBinaryExpression = true; 2287 if (Current) { 2288 FormatToken *Previous = Current->Previous; 2289 while (Previous->is(tok::comment) && Previous->Previous) 2290 Previous = Previous->Previous; 2291 ++Previous->FakeRParens; 2292 if (Precedence > prec::Unknown) 2293 Previous->EndsBinaryExpression = true; 2294 } 2295 } 2296 2297 /// Parse unary operator expressions and surround them with fake 2298 /// parentheses if appropriate. 2299 void parseUnaryOperator() { 2300 llvm::SmallVector<FormatToken *, 2> Tokens; 2301 while (Current && Current->is(TT_UnaryOperator)) { 2302 Tokens.push_back(Current); 2303 next(); 2304 } 2305 parse(PrecedenceArrowAndPeriod); 2306 for (FormatToken *Token : llvm::reverse(Tokens)) 2307 // The actual precedence doesn't matter. 2308 addFakeParenthesis(Token, prec::Unknown); 2309 } 2310 2311 void parseConditionalExpr() { 2312 while (Current && Current->isTrailingComment()) { 2313 next(); 2314 } 2315 FormatToken *Start = Current; 2316 parse(prec::LogicalOr); 2317 if (!Current || !Current->is(tok::question)) 2318 return; 2319 next(); 2320 parse(prec::Assignment); 2321 if (!Current || Current->isNot(TT_ConditionalExpr)) 2322 return; 2323 next(); 2324 parse(prec::Assignment); 2325 addFakeParenthesis(Start, prec::Conditional); 2326 } 2327 2328 void next(bool SkipPastLeadingComments = true) { 2329 if (Current) 2330 Current = Current->Next; 2331 while (Current && 2332 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 2333 Current->isTrailingComment()) 2334 Current = Current->Next; 2335 } 2336 2337 const FormatStyle &Style; 2338 const AdditionalKeywords &Keywords; 2339 FormatToken *Current; 2340 }; 2341 2342 } // end anonymous namespace 2343 2344 void TokenAnnotator::setCommentLineLevels( 2345 SmallVectorImpl<AnnotatedLine *> &Lines) { 2346 const AnnotatedLine *NextNonCommentLine = nullptr; 2347 for (AnnotatedLine *AL : llvm::reverse(Lines)) { 2348 bool CommentLine = true; 2349 for (const FormatToken *Tok = AL->First; Tok; Tok = Tok->Next) { 2350 if (!Tok->is(tok::comment)) { 2351 CommentLine = false; 2352 break; 2353 } 2354 } 2355 2356 // If the comment is currently aligned with the line immediately following 2357 // it, that's probably intentional and we should keep it. 2358 if (NextNonCommentLine && CommentLine && 2359 NextNonCommentLine->First->NewlinesBefore <= 1 && 2360 NextNonCommentLine->First->OriginalColumn == 2361 AL->First->OriginalColumn) { 2362 // Align comments for preprocessor lines with the # in column 0 if 2363 // preprocessor lines are not indented. Otherwise, align with the next 2364 // line. 2365 AL->Level = (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash && 2366 (NextNonCommentLine->Type == LT_PreprocessorDirective || 2367 NextNonCommentLine->Type == LT_ImportStatement)) 2368 ? 0 2369 : NextNonCommentLine->Level; 2370 } else { 2371 NextNonCommentLine = AL->First->isNot(tok::r_brace) ? AL : nullptr; 2372 } 2373 2374 setCommentLineLevels(AL->Children); 2375 } 2376 } 2377 2378 static unsigned maxNestingDepth(const AnnotatedLine &Line) { 2379 unsigned Result = 0; 2380 for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next) 2381 Result = std::max(Result, Tok->NestingLevel); 2382 return Result; 2383 } 2384 2385 void TokenAnnotator::annotate(AnnotatedLine &Line) { 2386 for (auto &Child : Line.Children) 2387 annotate(*Child); 2388 2389 AnnotatingParser Parser(Style, Line, Keywords); 2390 Line.Type = Parser.parseLine(); 2391 2392 // With very deep nesting, ExpressionParser uses lots of stack and the 2393 // formatting algorithm is very slow. We're not going to do a good job here 2394 // anyway - it's probably generated code being formatted by mistake. 2395 // Just skip the whole line. 2396 if (maxNestingDepth(Line) > 50) 2397 Line.Type = LT_Invalid; 2398 2399 if (Line.Type == LT_Invalid) 2400 return; 2401 2402 ExpressionParser ExprParser(Style, Keywords, Line); 2403 ExprParser.parse(); 2404 2405 if (Line.startsWith(TT_ObjCMethodSpecifier)) 2406 Line.Type = LT_ObjCMethodDecl; 2407 else if (Line.startsWith(TT_ObjCDecl)) 2408 Line.Type = LT_ObjCDecl; 2409 else if (Line.startsWith(TT_ObjCProperty)) 2410 Line.Type = LT_ObjCProperty; 2411 2412 Line.First->SpacesRequiredBefore = 1; 2413 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 2414 } 2415 2416 // This function heuristically determines whether 'Current' starts the name of a 2417 // function declaration. 2418 static bool isFunctionDeclarationName(bool IsCpp, const FormatToken &Current, 2419 const AnnotatedLine &Line) { 2420 auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * { 2421 for (; Next; Next = Next->Next) { 2422 if (Next->is(TT_OverloadedOperatorLParen)) 2423 return Next; 2424 if (Next->is(TT_OverloadedOperator)) 2425 continue; 2426 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) { 2427 // For 'new[]' and 'delete[]'. 2428 if (Next->Next && 2429 Next->Next->startsSequence(tok::l_square, tok::r_square)) 2430 Next = Next->Next->Next; 2431 continue; 2432 } 2433 if (Next->startsSequence(tok::l_square, tok::r_square)) { 2434 // For operator[](). 2435 Next = Next->Next; 2436 continue; 2437 } 2438 if ((Next->isSimpleTypeSpecifier() || Next->is(tok::identifier)) && 2439 Next->Next && Next->Next->isOneOf(tok::star, tok::amp, tok::ampamp)) { 2440 // For operator void*(), operator char*(), operator Foo*(). 2441 Next = Next->Next; 2442 continue; 2443 } 2444 if (Next->is(TT_TemplateOpener) && Next->MatchingParen) { 2445 Next = Next->MatchingParen; 2446 continue; 2447 } 2448 2449 break; 2450 } 2451 return nullptr; 2452 }; 2453 2454 // Find parentheses of parameter list. 2455 const FormatToken *Next = Current.Next; 2456 if (Current.is(tok::kw_operator)) { 2457 if (Current.Previous && Current.Previous->is(tok::coloncolon)) 2458 return false; 2459 Next = skipOperatorName(Next); 2460 } else { 2461 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0) 2462 return false; 2463 for (; Next; Next = Next->Next) { 2464 if (Next->is(TT_TemplateOpener)) { 2465 Next = Next->MatchingParen; 2466 } else if (Next->is(tok::coloncolon)) { 2467 Next = Next->Next; 2468 if (!Next) 2469 return false; 2470 if (Next->is(tok::kw_operator)) { 2471 Next = skipOperatorName(Next->Next); 2472 break; 2473 } 2474 if (!Next->is(tok::identifier)) 2475 return false; 2476 } else if (Next->is(tok::l_paren)) { 2477 break; 2478 } else { 2479 return false; 2480 } 2481 } 2482 } 2483 2484 // Check whether parameter list can belong to a function declaration. 2485 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen) 2486 return false; 2487 // If the lines ends with "{", this is likely a function definition. 2488 if (Line.Last->is(tok::l_brace)) 2489 return true; 2490 if (Next->Next == Next->MatchingParen) 2491 return true; // Empty parentheses. 2492 // If there is an &/&& after the r_paren, this is likely a function. 2493 if (Next->MatchingParen->Next && 2494 Next->MatchingParen->Next->is(TT_PointerOrReference)) 2495 return true; 2496 2497 // Check for K&R C function definitions (and C++ function definitions with 2498 // unnamed parameters), e.g.: 2499 // int f(i) 2500 // { 2501 // return i + 1; 2502 // } 2503 // bool g(size_t = 0, bool b = false) 2504 // { 2505 // return !b; 2506 // } 2507 if (IsCpp && Next->Next && Next->Next->is(tok::identifier) && 2508 !Line.endsWith(tok::semi)) 2509 return true; 2510 2511 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen; 2512 Tok = Tok->Next) { 2513 if (Tok->is(TT_TypeDeclarationParen)) 2514 return true; 2515 if (Tok->isOneOf(tok::l_paren, TT_TemplateOpener) && Tok->MatchingParen) { 2516 Tok = Tok->MatchingParen; 2517 continue; 2518 } 2519 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 2520 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis)) 2521 return true; 2522 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) || 2523 Tok->Tok.isLiteral()) 2524 return false; 2525 } 2526 return false; 2527 } 2528 2529 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const { 2530 assert(Line.MightBeFunctionDecl); 2531 2532 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel || 2533 Style.AlwaysBreakAfterReturnType == 2534 FormatStyle::RTBS_TopLevelDefinitions) && 2535 Line.Level > 0) 2536 return false; 2537 2538 switch (Style.AlwaysBreakAfterReturnType) { 2539 case FormatStyle::RTBS_None: 2540 return false; 2541 case FormatStyle::RTBS_All: 2542 case FormatStyle::RTBS_TopLevel: 2543 return true; 2544 case FormatStyle::RTBS_AllDefinitions: 2545 case FormatStyle::RTBS_TopLevelDefinitions: 2546 return Line.mightBeFunctionDefinition(); 2547 } 2548 2549 return false; 2550 } 2551 2552 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 2553 for (AnnotatedLine *ChildLine : Line.Children) 2554 calculateFormattingInformation(*ChildLine); 2555 2556 Line.First->TotalLength = 2557 Line.First->IsMultiline ? Style.ColumnLimit 2558 : Line.FirstStartColumn + Line.First->ColumnWidth; 2559 FormatToken *Current = Line.First->Next; 2560 bool InFunctionDecl = Line.MightBeFunctionDecl; 2561 bool AlignArrayOfStructures = 2562 (Style.AlignArrayOfStructures != FormatStyle::AIAS_None && 2563 Line.Type == LT_ArrayOfStructInitializer); 2564 if (AlignArrayOfStructures) 2565 calculateArrayInitializerColumnList(Line); 2566 2567 while (Current) { 2568 if (isFunctionDeclarationName(Style.isCpp(), *Current, Line)) 2569 Current->setType(TT_FunctionDeclarationName); 2570 const FormatToken *Prev = Current->Previous; 2571 if (Current->is(TT_LineComment)) { 2572 if (Prev->is(BK_BracedInit) && Prev->opensScope()) 2573 Current->SpacesRequiredBefore = 2574 (Style.Cpp11BracedListStyle && !Style.SpacesInParentheses) ? 0 : 1; 2575 else 2576 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 2577 2578 // If we find a trailing comment, iterate backwards to determine whether 2579 // it seems to relate to a specific parameter. If so, break before that 2580 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 2581 // to the previous line in: 2582 // SomeFunction(a, 2583 // b, // comment 2584 // c); 2585 if (!Current->HasUnescapedNewline) { 2586 for (FormatToken *Parameter = Current->Previous; Parameter; 2587 Parameter = Parameter->Previous) { 2588 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 2589 break; 2590 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 2591 if (!Parameter->Previous->is(TT_CtorInitializerComma) && 2592 Parameter->HasUnescapedNewline) 2593 Parameter->MustBreakBefore = true; 2594 break; 2595 } 2596 } 2597 } 2598 } else if (Current->SpacesRequiredBefore == 0 && 2599 spaceRequiredBefore(Line, *Current)) { 2600 Current->SpacesRequiredBefore = 1; 2601 } 2602 2603 Current->MustBreakBefore = 2604 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 2605 2606 if (!Current->MustBreakBefore && InFunctionDecl && 2607 Current->is(TT_FunctionDeclarationName)) 2608 Current->MustBreakBefore = mustBreakForReturnType(Line); 2609 2610 Current->CanBreakBefore = 2611 Current->MustBreakBefore || canBreakBefore(Line, *Current); 2612 unsigned ChildSize = 0; 2613 if (Prev->Children.size() == 1) { 2614 FormatToken &LastOfChild = *Prev->Children[0]->Last; 2615 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 2616 : LastOfChild.TotalLength + 1; 2617 } 2618 if (Current->MustBreakBefore || Prev->Children.size() > 1 || 2619 (Prev->Children.size() == 1 && 2620 Prev->Children[0]->First->MustBreakBefore) || 2621 Current->IsMultiline) 2622 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit; 2623 else 2624 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth + 2625 ChildSize + Current->SpacesRequiredBefore; 2626 2627 if (Current->is(TT_CtorInitializerColon)) 2628 InFunctionDecl = false; 2629 2630 // FIXME: Only calculate this if CanBreakBefore is true once static 2631 // initializers etc. are sorted out. 2632 // FIXME: Move magic numbers to a better place. 2633 2634 // Reduce penalty for aligning ObjC method arguments using the colon 2635 // alignment as this is the canonical way (still prefer fitting everything 2636 // into one line if possible). Trying to fit a whole expression into one 2637 // line should not force other line breaks (e.g. when ObjC method 2638 // expression is a part of other expression). 2639 Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl); 2640 if (Style.Language == FormatStyle::LK_ObjC && 2641 Current->is(TT_SelectorName) && Current->ParameterIndex > 0) { 2642 if (Current->ParameterIndex == 1) 2643 Current->SplitPenalty += 5 * Current->BindingStrength; 2644 } else { 2645 Current->SplitPenalty += 20 * Current->BindingStrength; 2646 } 2647 2648 Current = Current->Next; 2649 } 2650 2651 calculateUnbreakableTailLengths(Line); 2652 unsigned IndentLevel = Line.Level; 2653 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 2654 if (Current->Role) 2655 Current->Role->precomputeFormattingInfos(Current); 2656 if (Current->MatchingParen && 2657 Current->MatchingParen->opensBlockOrBlockTypeList(Style) && 2658 IndentLevel > 0) { 2659 --IndentLevel; 2660 } 2661 Current->IndentLevel = IndentLevel; 2662 if (Current->opensBlockOrBlockTypeList(Style)) 2663 ++IndentLevel; 2664 } 2665 2666 LLVM_DEBUG({ printDebugInfo(Line); }); 2667 } 2668 2669 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 2670 unsigned UnbreakableTailLength = 0; 2671 FormatToken *Current = Line.Last; 2672 while (Current) { 2673 Current->UnbreakableTailLength = UnbreakableTailLength; 2674 if (Current->CanBreakBefore || 2675 Current->isOneOf(tok::comment, tok::string_literal)) { 2676 UnbreakableTailLength = 0; 2677 } else { 2678 UnbreakableTailLength += 2679 Current->ColumnWidth + Current->SpacesRequiredBefore; 2680 } 2681 Current = Current->Previous; 2682 } 2683 } 2684 2685 void TokenAnnotator::calculateArrayInitializerColumnList(AnnotatedLine &Line) { 2686 if (Line.First == Line.Last) { 2687 return; 2688 } 2689 auto *CurrentToken = Line.First; 2690 CurrentToken->ArrayInitializerLineStart = true; 2691 unsigned Depth = 0; 2692 while (CurrentToken != nullptr && CurrentToken != Line.Last) { 2693 if (CurrentToken->is(tok::l_brace)) { 2694 CurrentToken->IsArrayInitializer = true; 2695 if (CurrentToken->Next != nullptr) 2696 CurrentToken->Next->MustBreakBefore = true; 2697 CurrentToken = 2698 calculateInitializerColumnList(Line, CurrentToken->Next, Depth + 1); 2699 } else { 2700 CurrentToken = CurrentToken->Next; 2701 } 2702 } 2703 } 2704 2705 FormatToken *TokenAnnotator::calculateInitializerColumnList( 2706 AnnotatedLine &Line, FormatToken *CurrentToken, unsigned Depth) { 2707 while (CurrentToken != nullptr && CurrentToken != Line.Last) { 2708 if (CurrentToken->is(tok::l_brace)) 2709 ++Depth; 2710 else if (CurrentToken->is(tok::r_brace)) 2711 --Depth; 2712 if (Depth == 2 && CurrentToken->isOneOf(tok::l_brace, tok::comma)) { 2713 CurrentToken = CurrentToken->Next; 2714 if (CurrentToken == nullptr) 2715 break; 2716 CurrentToken->StartsColumn = true; 2717 CurrentToken = CurrentToken->Previous; 2718 } 2719 CurrentToken = CurrentToken->Next; 2720 } 2721 return CurrentToken; 2722 } 2723 2724 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 2725 const FormatToken &Tok, 2726 bool InFunctionDecl) { 2727 const FormatToken &Left = *Tok.Previous; 2728 const FormatToken &Right = Tok; 2729 2730 if (Left.is(tok::semi)) 2731 return 0; 2732 2733 if (Style.Language == FormatStyle::LK_Java) { 2734 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws)) 2735 return 1; 2736 if (Right.is(Keywords.kw_implements)) 2737 return 2; 2738 if (Left.is(tok::comma) && Left.NestingLevel == 0) 2739 return 3; 2740 } else if (Style.isJavaScript()) { 2741 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma)) 2742 return 100; 2743 if (Left.is(TT_JsTypeColon)) 2744 return 35; 2745 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2746 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2747 return 100; 2748 // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()". 2749 if (Left.opensScope() && Right.closesScope()) 2750 return 200; 2751 } 2752 2753 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2754 return 1; 2755 if (Right.is(tok::l_square)) { 2756 if (Style.Language == FormatStyle::LK_Proto) 2757 return 1; 2758 if (Left.is(tok::r_square)) 2759 return 200; 2760 // Slightly prefer formatting local lambda definitions like functions. 2761 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal)) 2762 return 35; 2763 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2764 TT_ArrayInitializerLSquare, 2765 TT_DesignatedInitializerLSquare, TT_AttributeSquare)) 2766 return 500; 2767 } 2768 2769 if (Left.is(tok::coloncolon) || 2770 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 2771 return 500; 2772 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2773 Right.is(tok::kw_operator)) { 2774 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 2775 return 3; 2776 if (Left.is(TT_StartOfName)) 2777 return 110; 2778 if (InFunctionDecl && Right.NestingLevel == 0) 2779 return Style.PenaltyReturnTypeOnItsOwnLine; 2780 return 200; 2781 } 2782 if (Right.is(TT_PointerOrReference)) 2783 return 190; 2784 if (Right.is(TT_LambdaArrow)) 2785 return 110; 2786 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 2787 return 160; 2788 if (Left.is(TT_CastRParen)) 2789 return 100; 2790 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 2791 return 5000; 2792 if (Left.is(tok::comment)) 2793 return 1000; 2794 2795 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon, 2796 TT_CtorInitializerColon)) 2797 return 2; 2798 2799 if (Right.isMemberAccess()) { 2800 // Breaking before the "./->" of a chained call/member access is reasonably 2801 // cheap, as formatting those with one call per line is generally 2802 // desirable. In particular, it should be cheaper to break before the call 2803 // than it is to break inside a call's parameters, which could lead to weird 2804 // "hanging" indents. The exception is the very last "./->" to support this 2805 // frequent pattern: 2806 // 2807 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc( 2808 // dddddddd); 2809 // 2810 // which might otherwise be blown up onto many lines. Here, clang-format 2811 // won't produce "hanging" indents anyway as there is no other trailing 2812 // call. 2813 // 2814 // Also apply higher penalty is not a call as that might lead to a wrapping 2815 // like: 2816 // 2817 // aaaaaaa 2818 // .aaaaaaaaa.bbbbbbbb(cccccccc); 2819 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope() 2820 ? 150 2821 : 35; 2822 } 2823 2824 if (Right.is(TT_TrailingAnnotation) && 2825 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 2826 // Moving trailing annotations to the next line is fine for ObjC method 2827 // declarations. 2828 if (Line.startsWith(TT_ObjCMethodSpecifier)) 2829 return 10; 2830 // Generally, breaking before a trailing annotation is bad unless it is 2831 // function-like. It seems to be especially preferable to keep standard 2832 // annotations (i.e. "const", "final" and "override") on the same line. 2833 // Use a slightly higher penalty after ")" so that annotations like 2834 // "const override" are kept together. 2835 bool is_short_annotation = Right.TokenText.size() < 10; 2836 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 2837 } 2838 2839 // In for-loops, prefer breaking at ',' and ';'. 2840 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal)) 2841 return 4; 2842 2843 // In Objective-C method expressions, prefer breaking before "param:" over 2844 // breaking after it. 2845 if (Right.is(TT_SelectorName)) 2846 return 0; 2847 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr)) 2848 return Line.MightBeFunctionDecl ? 50 : 500; 2849 2850 // In Objective-C type declarations, avoid breaking after the category's 2851 // open paren (we'll prefer breaking after the protocol list's opening 2852 // angle bracket, if present). 2853 if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous && 2854 Left.Previous->isOneOf(tok::identifier, tok::greater)) 2855 return 500; 2856 2857 if (Left.is(tok::l_paren) && Style.PenaltyBreakOpenParenthesis != 0) 2858 return Style.PenaltyBreakOpenParenthesis; 2859 if (Left.is(tok::l_paren) && InFunctionDecl && 2860 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) 2861 return 100; 2862 if (Left.is(tok::l_paren) && Left.Previous && 2863 (Left.Previous->is(tok::kw_for) || Left.Previous->isIf())) 2864 return 1000; 2865 if (Left.is(tok::equal) && InFunctionDecl) 2866 return 110; 2867 if (Right.is(tok::r_brace)) 2868 return 1; 2869 if (Left.is(TT_TemplateOpener)) 2870 return 100; 2871 if (Left.opensScope()) { 2872 // If we aren't aligning after opening parens/braces we can always break 2873 // here unless the style does not want us to place all arguments on the 2874 // next line. 2875 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign && 2876 (Left.ParameterCount <= 1 || Style.AllowAllArgumentsOnNextLine)) 2877 return 0; 2878 if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle) 2879 return 19; 2880 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 2881 : 19; 2882 } 2883 if (Left.is(TT_JavaAnnotation)) 2884 return 50; 2885 2886 if (Left.is(TT_UnaryOperator)) 2887 return 60; 2888 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous && 2889 Left.Previous->isLabelString() && 2890 (Left.NextOperator || Left.OperatorIndex != 0)) 2891 return 50; 2892 if (Right.is(tok::plus) && Left.isLabelString() && 2893 (Right.NextOperator || Right.OperatorIndex != 0)) 2894 return 25; 2895 if (Left.is(tok::comma)) 2896 return 1; 2897 if (Right.is(tok::lessless) && Left.isLabelString() && 2898 (Right.NextOperator || Right.OperatorIndex != 1)) 2899 return 25; 2900 if (Right.is(tok::lessless)) { 2901 // Breaking at a << is really cheap. 2902 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0) 2903 // Slightly prefer to break before the first one in log-like statements. 2904 return 2; 2905 return 1; 2906 } 2907 if (Left.ClosesTemplateDeclaration) 2908 return Style.PenaltyBreakTemplateDeclaration; 2909 if (Left.is(TT_ConditionalExpr)) 2910 return prec::Conditional; 2911 prec::Level Level = Left.getPrecedence(); 2912 if (Level == prec::Unknown) 2913 Level = Right.getPrecedence(); 2914 if (Level == prec::Assignment) 2915 return Style.PenaltyBreakAssignment; 2916 if (Level != prec::Unknown) 2917 return Level; 2918 2919 return 3; 2920 } 2921 2922 bool TokenAnnotator::spaceRequiredBeforeParens(const FormatToken &Right) const { 2923 if (Style.SpaceBeforeParens == FormatStyle::SBPO_Always) 2924 return true; 2925 if (Right.is(TT_OverloadedOperatorLParen) && 2926 Style.SpaceBeforeParensOptions.AfterOverloadedOperator) 2927 return true; 2928 if (Style.SpaceBeforeParensOptions.BeforeNonEmptyParentheses && 2929 Right.ParameterCount > 0) 2930 return true; 2931 return false; 2932 } 2933 2934 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 2935 const FormatToken &Left, 2936 const FormatToken &Right) { 2937 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 2938 return true; 2939 if (Style.isJson() && Left.is(tok::string_literal) && Right.is(tok::colon)) 2940 return false; 2941 if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java) 2942 return true; 2943 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 2944 Left.Tok.getObjCKeywordID() == tok::objc_property) 2945 return true; 2946 if (Right.is(tok::hashhash)) 2947 return Left.is(tok::hash); 2948 if (Left.isOneOf(tok::hashhash, tok::hash)) 2949 return Right.is(tok::hash); 2950 if ((Left.is(tok::l_paren) && Right.is(tok::r_paren)) || 2951 (Left.is(tok::l_brace) && Left.isNot(BK_Block) && 2952 Right.is(tok::r_brace) && Right.isNot(BK_Block))) 2953 return Style.SpaceInEmptyParentheses; 2954 if (Style.SpacesInConditionalStatement) { 2955 if (Left.is(tok::l_paren) && Left.Previous && 2956 isKeywordWithCondition(*Left.Previous)) 2957 return true; 2958 if (Right.is(tok::r_paren) && Right.MatchingParen && 2959 Right.MatchingParen->Previous && 2960 isKeywordWithCondition(*Right.MatchingParen->Previous)) 2961 return true; 2962 } 2963 2964 // auto{x} auto(x) 2965 if (Left.is(tok::kw_auto) && Right.isOneOf(tok::l_paren, tok::l_brace)) 2966 return false; 2967 2968 // operator co_await(x) 2969 if (Right.is(tok::l_paren) && Left.is(tok::kw_co_await) && Left.Previous && 2970 Left.Previous->is(tok::kw_operator)) 2971 return false; 2972 // co_await (x), co_yield (x), co_return (x) 2973 if (Left.isOneOf(tok::kw_co_await, tok::kw_co_yield, tok::kw_co_return) && 2974 Right.isNot(tok::semi)) 2975 return true; 2976 // requires clause Concept1<T> && Concept2<T> 2977 if (Left.is(TT_ConstraintJunctions) && Right.is(tok::identifier)) 2978 return true; 2979 2980 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 2981 return (Right.is(TT_CastRParen) || 2982 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen))) 2983 ? Style.SpacesInCStyleCastParentheses 2984 : Style.SpacesInParentheses; 2985 if (Right.isOneOf(tok::semi, tok::comma)) 2986 return false; 2987 if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) { 2988 bool IsLightweightGeneric = Right.MatchingParen && 2989 Right.MatchingParen->Next && 2990 Right.MatchingParen->Next->is(tok::colon); 2991 return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList; 2992 } 2993 if (Right.is(tok::less) && Left.is(tok::kw_template)) 2994 return Style.SpaceAfterTemplateKeyword; 2995 if (Left.isOneOf(tok::exclaim, tok::tilde)) 2996 return false; 2997 if (Left.is(tok::at) && 2998 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 2999 tok::numeric_constant, tok::l_paren, tok::l_brace, 3000 tok::kw_true, tok::kw_false)) 3001 return false; 3002 if (Left.is(tok::colon)) 3003 return !Left.is(TT_ObjCMethodExpr); 3004 if (Left.is(tok::coloncolon)) 3005 return false; 3006 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) { 3007 if (Style.Language == FormatStyle::LK_TextProto || 3008 (Style.Language == FormatStyle::LK_Proto && 3009 (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) { 3010 // Format empty list as `<>`. 3011 if (Left.is(tok::less) && Right.is(tok::greater)) 3012 return false; 3013 return !Style.Cpp11BracedListStyle; 3014 } 3015 return false; 3016 } 3017 if (Right.is(tok::ellipsis)) 3018 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous && 3019 Left.Previous->is(tok::kw_case)); 3020 if (Left.is(tok::l_square) && Right.is(tok::amp)) 3021 return Style.SpacesInSquareBrackets; 3022 if (Right.is(TT_PointerOrReference)) { 3023 if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) { 3024 if (!Left.MatchingParen) 3025 return true; 3026 FormatToken *TokenBeforeMatchingParen = 3027 Left.MatchingParen->getPreviousNonComment(); 3028 if (!TokenBeforeMatchingParen || !Left.is(TT_TypeDeclarationParen)) 3029 return true; 3030 } 3031 // Add a space if the previous token is a pointer qualifier or the closing 3032 // parenthesis of __attribute__(()) expression and the style requires spaces 3033 // after pointer qualifiers. 3034 if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_After || 3035 Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) && 3036 (Left.is(TT_AttributeParen) || Left.canBePointerOrReferenceQualifier())) 3037 return true; 3038 if (Left.Tok.isLiteral()) 3039 return true; 3040 // for (auto a = 0, b = 0; const auto & c : {1, 2, 3}) 3041 if (Left.isTypeOrIdentifier() && Right.Next && Right.Next->Next && 3042 Right.Next->Next->is(TT_RangeBasedForLoopColon)) 3043 return getTokenPointerOrReferenceAlignment(Right) != 3044 FormatStyle::PAS_Left; 3045 return ( 3046 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) && 3047 (getTokenPointerOrReferenceAlignment(Right) != FormatStyle::PAS_Left || 3048 (Line.IsMultiVariableDeclStmt && 3049 (Left.NestingLevel == 0 || 3050 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for))))))); 3051 } 3052 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) && 3053 (!Left.is(TT_PointerOrReference) || 3054 (getTokenPointerOrReferenceAlignment(Left) != FormatStyle::PAS_Right && 3055 !Line.IsMultiVariableDeclStmt))) 3056 return true; 3057 if (Left.is(TT_PointerOrReference)) { 3058 // Add a space if the next token is a pointer qualifier and the style 3059 // requires spaces before pointer qualifiers. 3060 if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Before || 3061 Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) && 3062 Right.canBePointerOrReferenceQualifier()) 3063 return true; 3064 // & 1 3065 if (Right.Tok.isLiteral()) 3066 return true; 3067 // & /* comment 3068 if (Right.is(TT_BlockComment)) 3069 return true; 3070 // foo() -> const Bar * override/final 3071 if (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) && 3072 !Right.is(TT_StartOfName)) 3073 return true; 3074 // & { 3075 if (Right.is(tok::l_brace) && Right.is(BK_Block)) 3076 return true; 3077 // for (auto a = 0, b = 0; const auto& c : {1, 2, 3}) 3078 if (Left.Previous && Left.Previous->isTypeOrIdentifier() && Right.Next && 3079 Right.Next->is(TT_RangeBasedForLoopColon)) 3080 return getTokenPointerOrReferenceAlignment(Left) != 3081 FormatStyle::PAS_Right; 3082 return !Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare, 3083 tok::l_paren) && 3084 (getTokenPointerOrReferenceAlignment(Left) != 3085 FormatStyle::PAS_Right && 3086 !Line.IsMultiVariableDeclStmt) && 3087 Left.Previous && 3088 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon, 3089 tok::l_square); 3090 } 3091 // Ensure right pointer alignment with ellipsis e.g. int *...P 3092 if (Left.is(tok::ellipsis) && Left.Previous && 3093 Left.Previous->isOneOf(tok::star, tok::amp, tok::ampamp)) 3094 return Style.PointerAlignment != FormatStyle::PAS_Right; 3095 3096 if (Right.is(tok::star) && Left.is(tok::l_paren)) 3097 return false; 3098 if (Left.is(tok::star) && Right.isOneOf(tok::star, tok::amp, tok::ampamp)) 3099 return false; 3100 if (Right.isOneOf(tok::star, tok::amp, tok::ampamp)) { 3101 const FormatToken *Previous = &Left; 3102 while (Previous && !Previous->is(tok::kw_operator)) { 3103 if (Previous->is(tok::identifier) || Previous->isSimpleTypeSpecifier()) { 3104 Previous = Previous->getPreviousNonComment(); 3105 continue; 3106 } 3107 if (Previous->is(TT_TemplateCloser) && Previous->MatchingParen) { 3108 Previous = Previous->MatchingParen->getPreviousNonComment(); 3109 continue; 3110 } 3111 if (Previous->is(tok::coloncolon)) { 3112 Previous = Previous->getPreviousNonComment(); 3113 continue; 3114 } 3115 break; 3116 } 3117 // Space between the type and the * in: 3118 // operator void*() 3119 // operator char*() 3120 // operator void const*() 3121 // operator void volatile*() 3122 // operator /*comment*/ const char*() 3123 // operator volatile /*comment*/ char*() 3124 // operator Foo*() 3125 // operator C<T>*() 3126 // operator std::Foo*() 3127 // operator C<T>::D<U>*() 3128 // dependent on PointerAlignment style. 3129 if (Previous) { 3130 if (Previous->endsSequence(tok::kw_operator)) 3131 return (Style.PointerAlignment != FormatStyle::PAS_Left); 3132 if (Previous->is(tok::kw_const) || Previous->is(tok::kw_volatile)) 3133 return (Style.PointerAlignment != FormatStyle::PAS_Left) || 3134 (Style.SpaceAroundPointerQualifiers == 3135 FormatStyle::SAPQ_After) || 3136 (Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both); 3137 } 3138 } 3139 const auto SpaceRequiredForArrayInitializerLSquare = 3140 [](const FormatToken &LSquareTok, const FormatStyle &Style) { 3141 return Style.SpacesInContainerLiterals || 3142 ((Style.Language == FormatStyle::LK_Proto || 3143 Style.Language == FormatStyle::LK_TextProto) && 3144 !Style.Cpp11BracedListStyle && 3145 LSquareTok.endsSequence(tok::l_square, tok::colon, 3146 TT_SelectorName)); 3147 }; 3148 if (Left.is(tok::l_square)) 3149 return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) && 3150 SpaceRequiredForArrayInitializerLSquare(Left, Style)) || 3151 (Left.isOneOf(TT_ArraySubscriptLSquare, TT_StructuredBindingLSquare, 3152 TT_LambdaLSquare) && 3153 Style.SpacesInSquareBrackets && Right.isNot(tok::r_square)); 3154 if (Right.is(tok::r_square)) 3155 return Right.MatchingParen && 3156 ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) && 3157 SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen, 3158 Style)) || 3159 (Style.SpacesInSquareBrackets && 3160 Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare, 3161 TT_StructuredBindingLSquare, 3162 TT_LambdaLSquare)) || 3163 Right.MatchingParen->is(TT_AttributeParen)); 3164 if (Right.is(tok::l_square) && 3165 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 3166 TT_DesignatedInitializerLSquare, 3167 TT_StructuredBindingLSquare, TT_AttributeSquare) && 3168 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral) && 3169 !(!Left.is(tok::r_square) && Style.SpaceBeforeSquareBrackets && 3170 Right.is(TT_ArraySubscriptLSquare))) 3171 return false; 3172 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 3173 return !Left.Children.empty(); // No spaces in "{}". 3174 if ((Left.is(tok::l_brace) && Left.isNot(BK_Block)) || 3175 (Right.is(tok::r_brace) && Right.MatchingParen && 3176 Right.MatchingParen->isNot(BK_Block))) 3177 return Style.Cpp11BracedListStyle ? Style.SpacesInParentheses : true; 3178 if (Left.is(TT_BlockComment)) 3179 // No whitespace in x(/*foo=*/1), except for JavaScript. 3180 return Style.isJavaScript() || !Left.TokenText.endswith("=*/"); 3181 3182 // Space between template and attribute. 3183 // e.g. template <typename T> [[nodiscard]] ... 3184 if (Left.is(TT_TemplateCloser) && Right.is(TT_AttributeSquare)) 3185 return true; 3186 // Space before parentheses common for all languages 3187 if (Right.is(tok::l_paren)) { 3188 if (Left.is(TT_TemplateCloser) && Right.isNot(TT_FunctionTypeLParen)) 3189 return spaceRequiredBeforeParens(Right); 3190 if (Left.is(tok::kw_requires)) 3191 return spaceRequiredBeforeParens(Right); 3192 if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) || 3193 (Left.is(tok::r_square) && Left.is(TT_AttributeSquare))) 3194 return true; 3195 if (Left.is(TT_ForEachMacro)) 3196 return (Style.SpaceBeforeParensOptions.AfterForeachMacros || 3197 spaceRequiredBeforeParens(Right)); 3198 if (Left.is(TT_IfMacro)) 3199 return (Style.SpaceBeforeParensOptions.AfterIfMacros || 3200 spaceRequiredBeforeParens(Right)); 3201 if (Line.Type == LT_ObjCDecl) 3202 return true; 3203 if (Left.is(tok::semi)) 3204 return true; 3205 if (Left.isOneOf(tok::pp_elif, tok::kw_for, tok::kw_while, tok::kw_switch, 3206 tok::kw_case, TT_ForEachMacro, TT_ObjCForIn)) 3207 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3208 spaceRequiredBeforeParens(Right); 3209 if (Left.isIf(Line.Type != LT_PreprocessorDirective)) 3210 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3211 spaceRequiredBeforeParens(Right); 3212 3213 // TODO add Operator overloading specific Options to 3214 // SpaceBeforeParensOptions 3215 if (Right.is(TT_OverloadedOperatorLParen)) 3216 return spaceRequiredBeforeParens(Right); 3217 // Function declaration or definition 3218 if (Line.MightBeFunctionDecl && (Left.is(TT_FunctionDeclarationName))) { 3219 if (Line.mightBeFunctionDefinition()) 3220 return Style.SpaceBeforeParensOptions.AfterFunctionDefinitionName || 3221 spaceRequiredBeforeParens(Right); 3222 else 3223 return Style.SpaceBeforeParensOptions.AfterFunctionDeclarationName || 3224 spaceRequiredBeforeParens(Right); 3225 } 3226 // Lambda 3227 if (Line.Type != LT_PreprocessorDirective && Left.is(tok::r_square) && 3228 Left.MatchingParen && Left.MatchingParen->is(TT_LambdaLSquare)) 3229 return Style.SpaceBeforeParensOptions.AfterFunctionDefinitionName || 3230 spaceRequiredBeforeParens(Right); 3231 if (!Left.Previous || Left.Previous->isNot(tok::period)) { 3232 if (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch)) 3233 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3234 spaceRequiredBeforeParens(Right); 3235 if (Left.isOneOf(tok::kw_new, tok::kw_delete)) 3236 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never || 3237 spaceRequiredBeforeParens(Right); 3238 } 3239 if (Line.Type != LT_PreprocessorDirective && 3240 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() || 3241 Left.is(tok::r_paren) || Left.isSimpleTypeSpecifier())) 3242 return spaceRequiredBeforeParens(Right); 3243 return false; 3244 } 3245 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 3246 return false; 3247 if (Right.is(TT_UnaryOperator)) 3248 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 3249 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr)); 3250 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 3251 tok::r_paren) || 3252 Left.isSimpleTypeSpecifier()) && 3253 Right.is(tok::l_brace) && Right.getNextNonComment() && 3254 Right.isNot(BK_Block)) 3255 return false; 3256 if (Left.is(tok::period) || Right.is(tok::period)) 3257 return false; 3258 // u#str, U#str, L#str, u8#str 3259 // uR#str, UR#str, LR#str, u8R#str 3260 if (Right.is(tok::hash) && Left.is(tok::identifier) && 3261 (Left.TokenText == "L" || Left.TokenText == "u" || 3262 Left.TokenText == "U" || Left.TokenText == "u8" || 3263 Left.TokenText == "LR" || Left.TokenText == "uR" || 3264 Left.TokenText == "UR" || Left.TokenText == "u8R")) 3265 return false; 3266 if (Left.is(TT_TemplateCloser) && Left.MatchingParen && 3267 Left.MatchingParen->Previous && 3268 (Left.MatchingParen->Previous->is(tok::period) || 3269 Left.MatchingParen->Previous->is(tok::coloncolon))) 3270 // Java call to generic function with explicit type: 3271 // A.<B<C<...>>>DoSomething(); 3272 // A::<B<C<...>>>DoSomething(); // With a Java 8 method reference. 3273 return false; 3274 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square)) 3275 return false; 3276 if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at)) 3277 // Objective-C dictionary literal -> no space after opening brace. 3278 return false; 3279 if (Right.is(tok::r_brace) && Right.MatchingParen && 3280 Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at)) 3281 // Objective-C dictionary literal -> no space before closing brace. 3282 return false; 3283 if (Right.getType() == TT_TrailingAnnotation && 3284 Right.isOneOf(tok::amp, tok::ampamp) && 3285 Left.isOneOf(tok::kw_const, tok::kw_volatile) && 3286 (!Right.Next || Right.Next->is(tok::semi))) 3287 // Match const and volatile ref-qualifiers without any additional 3288 // qualifiers such as 3289 // void Fn() const &; 3290 return getTokenReferenceAlignment(Right) != FormatStyle::PAS_Left; 3291 3292 return true; 3293 } 3294 3295 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 3296 const FormatToken &Right) { 3297 const FormatToken &Left = *Right.Previous; 3298 auto HasExistingWhitespace = [&Right]() { 3299 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 3300 }; 3301 3302 // If the token is finalized don't touch it (as it could be in a 3303 // clang-format-off section). 3304 if (Left.Finalized) 3305 return HasExistingWhitespace(); 3306 3307 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo()) 3308 return true; // Never ever merge two identifiers. 3309 3310 // Leave a space between * and /* to avoid C4138 `comment end` found outside 3311 // of comment. 3312 if (Left.is(tok::star) && Right.is(tok::comment)) 3313 return true; 3314 3315 if (Style.isCpp()) { 3316 // Space between import <iostream>. 3317 // or import .....; 3318 if (Left.is(Keywords.kw_import) && Right.isOneOf(tok::less, tok::ellipsis)) 3319 return true; 3320 // No space between module :. 3321 if (Left.isOneOf(Keywords.kw_module, Keywords.kw_import) && 3322 Right.is(TT_ModulePartitionColon)) 3323 return true; 3324 // No space between import foo:bar but keep a space between import :bar; 3325 if (Left.is(tok::identifier) && Right.is(TT_ModulePartitionColon)) 3326 return false; 3327 // No space between :bar; 3328 if (Left.is(TT_ModulePartitionColon) && 3329 Right.isOneOf(tok::identifier, tok::kw_private)) 3330 return false; 3331 if (Left.is(tok::ellipsis) && Right.is(tok::identifier) && 3332 Line.First->is(Keywords.kw_import)) 3333 return false; 3334 3335 if (Left.is(tok::kw_operator)) 3336 return Right.is(tok::coloncolon); 3337 if (Right.is(tok::l_brace) && Right.is(BK_BracedInit) && 3338 !Left.opensScope() && Style.SpaceBeforeCpp11BracedList) 3339 return true; 3340 } else if (Style.Language == FormatStyle::LK_Proto || 3341 Style.Language == FormatStyle::LK_TextProto) { 3342 if (Right.is(tok::period) && 3343 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required, 3344 Keywords.kw_repeated, Keywords.kw_extend)) 3345 return true; 3346 if (Right.is(tok::l_paren) && 3347 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) 3348 return true; 3349 if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName)) 3350 return true; 3351 // Slashes occur in text protocol extension syntax: [type/type] { ... }. 3352 if (Left.is(tok::slash) || Right.is(tok::slash)) 3353 return false; 3354 if (Left.MatchingParen && 3355 Left.MatchingParen->is(TT_ProtoExtensionLSquare) && 3356 Right.isOneOf(tok::l_brace, tok::less)) 3357 return !Style.Cpp11BracedListStyle; 3358 // A percent is probably part of a formatting specification, such as %lld. 3359 if (Left.is(tok::percent)) 3360 return false; 3361 // Preserve the existence of a space before a percent for cases like 0x%04x 3362 // and "%d %d" 3363 if (Left.is(tok::numeric_constant) && Right.is(tok::percent)) 3364 return HasExistingWhitespace(); 3365 } else if (Style.isJson()) { 3366 if (Right.is(tok::colon)) 3367 return false; 3368 } else if (Style.isCSharp()) { 3369 // Require spaces around '{' and before '}' unless they appear in 3370 // interpolated strings. Interpolated strings are merged into a single token 3371 // so cannot have spaces inserted by this function. 3372 3373 // No space between 'this' and '[' 3374 if (Left.is(tok::kw_this) && Right.is(tok::l_square)) 3375 return false; 3376 3377 // No space between 'new' and '(' 3378 if (Left.is(tok::kw_new) && Right.is(tok::l_paren)) 3379 return false; 3380 3381 // Space before { (including space within '{ {'). 3382 if (Right.is(tok::l_brace)) 3383 return true; 3384 3385 // Spaces inside braces. 3386 if (Left.is(tok::l_brace) && Right.isNot(tok::r_brace)) 3387 return true; 3388 3389 if (Left.isNot(tok::l_brace) && Right.is(tok::r_brace)) 3390 return true; 3391 3392 // Spaces around '=>'. 3393 if (Left.is(TT_FatArrow) || Right.is(TT_FatArrow)) 3394 return true; 3395 3396 // No spaces around attribute target colons 3397 if (Left.is(TT_AttributeColon) || Right.is(TT_AttributeColon)) 3398 return false; 3399 3400 // space between type and variable e.g. Dictionary<string,string> foo; 3401 if (Left.is(TT_TemplateCloser) && Right.is(TT_StartOfName)) 3402 return true; 3403 3404 // spaces inside square brackets. 3405 if (Left.is(tok::l_square) || Right.is(tok::r_square)) 3406 return Style.SpacesInSquareBrackets; 3407 3408 // No space before ? in nullable types. 3409 if (Right.is(TT_CSharpNullable)) 3410 return false; 3411 3412 // No space before null forgiving '!'. 3413 if (Right.is(TT_NonNullAssertion)) 3414 return false; 3415 3416 // No space between consecutive commas '[,,]'. 3417 if (Left.is(tok::comma) && Right.is(tok::comma)) 3418 return false; 3419 3420 // space after var in `var (key, value)` 3421 if (Left.is(Keywords.kw_var) && Right.is(tok::l_paren)) 3422 return true; 3423 3424 // space between keywords and paren e.g. "using (" 3425 if (Right.is(tok::l_paren)) 3426 if (Left.isOneOf(tok::kw_using, Keywords.kw_async, Keywords.kw_when, 3427 Keywords.kw_lock)) 3428 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3429 spaceRequiredBeforeParens(Right); 3430 3431 // space between method modifier and opening parenthesis of a tuple return 3432 // type 3433 if (Left.isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected, 3434 tok::kw_virtual, tok::kw_extern, tok::kw_static, 3435 Keywords.kw_internal, Keywords.kw_abstract, 3436 Keywords.kw_sealed, Keywords.kw_override, 3437 Keywords.kw_async, Keywords.kw_unsafe) && 3438 Right.is(tok::l_paren)) 3439 return true; 3440 } else if (Style.isJavaScript()) { 3441 if (Left.is(TT_FatArrow)) 3442 return true; 3443 // for await ( ... 3444 if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous && 3445 Left.Previous->is(tok::kw_for)) 3446 return true; 3447 if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) && 3448 Right.MatchingParen) { 3449 const FormatToken *Next = Right.MatchingParen->getNextNonComment(); 3450 // An async arrow function, for example: `x = async () => foo();`, 3451 // as opposed to calling a function called async: `x = async();` 3452 if (Next && Next->is(TT_FatArrow)) 3453 return true; 3454 } 3455 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 3456 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 3457 return false; 3458 // In tagged template literals ("html`bar baz`"), there is no space between 3459 // the tag identifier and the template string. 3460 if (Keywords.IsJavaScriptIdentifier(Left, 3461 /* AcceptIdentifierName= */ false) && 3462 Right.is(TT_TemplateString)) 3463 return false; 3464 if (Right.is(tok::star) && 3465 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) 3466 return false; 3467 if (Right.isOneOf(tok::l_brace, tok::l_square) && 3468 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield, 3469 Keywords.kw_extends, Keywords.kw_implements)) 3470 return true; 3471 if (Right.is(tok::l_paren)) { 3472 // JS methods can use some keywords as names (e.g. `delete()`). 3473 if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo()) 3474 return false; 3475 // Valid JS method names can include keywords, e.g. `foo.delete()` or 3476 // `bar.instanceof()`. Recognize call positions by preceding period. 3477 if (Left.Previous && Left.Previous->is(tok::period) && 3478 Left.Tok.getIdentifierInfo()) 3479 return false; 3480 // Additional unary JavaScript operators that need a space after. 3481 if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof, 3482 tok::kw_void)) 3483 return true; 3484 } 3485 // `foo as const;` casts into a const type. 3486 if (Left.endsSequence(tok::kw_const, Keywords.kw_as)) { 3487 return false; 3488 } 3489 if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in, 3490 tok::kw_const) || 3491 // "of" is only a keyword if it appears after another identifier 3492 // (e.g. as "const x of y" in a for loop), or after a destructuring 3493 // operation (const [x, y] of z, const {a, b} of c). 3494 (Left.is(Keywords.kw_of) && Left.Previous && 3495 (Left.Previous->Tok.is(tok::identifier) || 3496 Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) && 3497 (!Left.Previous || !Left.Previous->is(tok::period))) 3498 return true; 3499 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous && 3500 Left.Previous->is(tok::period) && Right.is(tok::l_paren)) 3501 return false; 3502 if (Left.is(Keywords.kw_as) && 3503 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren)) 3504 return true; 3505 if (Left.is(tok::kw_default) && Left.Previous && 3506 Left.Previous->is(tok::kw_export)) 3507 return true; 3508 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace)) 3509 return true; 3510 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion)) 3511 return false; 3512 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator)) 3513 return false; 3514 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) && 3515 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) 3516 return false; 3517 if (Left.is(tok::ellipsis)) 3518 return false; 3519 if (Left.is(TT_TemplateCloser) && 3520 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square, 3521 Keywords.kw_implements, Keywords.kw_extends)) 3522 // Type assertions ('<type>expr') are not followed by whitespace. Other 3523 // locations that should have whitespace following are identified by the 3524 // above set of follower tokens. 3525 return false; 3526 if (Right.is(TT_NonNullAssertion)) 3527 return false; 3528 if (Left.is(TT_NonNullAssertion) && 3529 Right.isOneOf(Keywords.kw_as, Keywords.kw_in)) 3530 return true; // "x! as string", "x! in y" 3531 } else if (Style.Language == FormatStyle::LK_Java) { 3532 if (Left.is(tok::r_square) && Right.is(tok::l_brace)) 3533 return true; 3534 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) 3535 return Style.SpaceBeforeParensOptions.AfterControlStatements || 3536 spaceRequiredBeforeParens(Right); 3537 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private, 3538 tok::kw_protected) || 3539 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract, 3540 Keywords.kw_native)) && 3541 Right.is(TT_TemplateOpener)) 3542 return true; 3543 } 3544 if (Left.is(TT_ImplicitStringLiteral)) 3545 return HasExistingWhitespace(); 3546 if (Line.Type == LT_ObjCMethodDecl) { 3547 if (Left.is(TT_ObjCMethodSpecifier)) 3548 return true; 3549 if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right)) 3550 // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a 3551 // keyword in Objective-C, and '+ (instancetype)new;' is a standard class 3552 // method declaration. 3553 return false; 3554 } 3555 if (Line.Type == LT_ObjCProperty && 3556 (Right.is(tok::equal) || Left.is(tok::equal))) 3557 return false; 3558 3559 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) || 3560 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) 3561 return true; 3562 if (Left.is(tok::comma) && !Right.is(TT_OverloadedOperatorLParen)) 3563 return true; 3564 if (Right.is(tok::comma)) 3565 return false; 3566 if (Right.is(TT_ObjCBlockLParen)) 3567 return true; 3568 if (Right.is(TT_CtorInitializerColon)) 3569 return Style.SpaceBeforeCtorInitializerColon; 3570 if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon) 3571 return false; 3572 if (Right.is(TT_RangeBasedForLoopColon) && 3573 !Style.SpaceBeforeRangeBasedForLoopColon) 3574 return false; 3575 if (Left.is(TT_BitFieldColon)) 3576 return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both || 3577 Style.BitFieldColonSpacing == FormatStyle::BFCS_After; 3578 if (Right.is(tok::colon)) { 3579 if (Line.First->isOneOf(tok::kw_default, tok::kw_case)) 3580 return Style.SpaceBeforeCaseColon; 3581 if (!Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi)) 3582 return false; 3583 if (Right.is(TT_ObjCMethodExpr)) 3584 return false; 3585 if (Left.is(tok::question)) 3586 return false; 3587 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon)) 3588 return false; 3589 if (Right.is(TT_DictLiteral)) 3590 return Style.SpacesInContainerLiterals; 3591 if (Right.is(TT_AttributeColon)) 3592 return false; 3593 if (Right.is(TT_CSharpNamedArgumentColon)) 3594 return false; 3595 if (Right.is(TT_BitFieldColon)) 3596 return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both || 3597 Style.BitFieldColonSpacing == FormatStyle::BFCS_Before; 3598 return true; 3599 } 3600 // Do not merge "- -" into "--". 3601 if ((Left.isOneOf(tok::minus, tok::minusminus) && 3602 Right.isOneOf(tok::minus, tok::minusminus)) || 3603 (Left.isOneOf(tok::plus, tok::plusplus) && 3604 Right.isOneOf(tok::plus, tok::plusplus))) 3605 return true; 3606 if (Left.is(TT_UnaryOperator)) { 3607 if (!Right.is(tok::l_paren)) { 3608 // The alternative operators for ~ and ! are "compl" and "not". 3609 // If they are used instead, we do not want to combine them with 3610 // the token to the right, unless that is a left paren. 3611 if (Left.is(tok::exclaim) && Left.TokenText == "not") 3612 return true; 3613 if (Left.is(tok::tilde) && Left.TokenText == "compl") 3614 return true; 3615 // Lambda captures allow for a lone &, so "&]" needs to be properly 3616 // handled. 3617 if (Left.is(tok::amp) && Right.is(tok::r_square)) 3618 return Style.SpacesInSquareBrackets; 3619 } 3620 return (Style.SpaceAfterLogicalNot && Left.is(tok::exclaim)) || 3621 Right.is(TT_BinaryOperator); 3622 } 3623 3624 // If the next token is a binary operator or a selector name, we have 3625 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly. 3626 if (Left.is(TT_CastRParen)) 3627 return Style.SpaceAfterCStyleCast || 3628 Right.isOneOf(TT_BinaryOperator, TT_SelectorName); 3629 3630 auto ShouldAddSpacesInAngles = [this, &HasExistingWhitespace]() { 3631 if (this->Style.SpacesInAngles == FormatStyle::SIAS_Always) 3632 return true; 3633 if (this->Style.SpacesInAngles == FormatStyle::SIAS_Leave) 3634 return HasExistingWhitespace(); 3635 return false; 3636 }; 3637 3638 if (Left.is(tok::greater) && Right.is(tok::greater)) { 3639 if (Style.Language == FormatStyle::LK_TextProto || 3640 (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral))) 3641 return !Style.Cpp11BracedListStyle; 3642 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) && 3643 ((Style.Standard < FormatStyle::LS_Cpp11) || 3644 ShouldAddSpacesInAngles()); 3645 } 3646 if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) || 3647 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) || 3648 (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod))) 3649 return false; 3650 if (!Style.SpaceBeforeAssignmentOperators && Left.isNot(TT_TemplateCloser) && 3651 Right.getPrecedence() == prec::Assignment) 3652 return false; 3653 if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) && 3654 (Left.is(tok::identifier) || Left.is(tok::kw_this))) 3655 return false; 3656 if (Right.is(tok::coloncolon) && Left.is(tok::identifier)) 3657 // Generally don't remove existing spaces between an identifier and "::". 3658 // The identifier might actually be a macro name such as ALWAYS_INLINE. If 3659 // this turns out to be too lenient, add analysis of the identifier itself. 3660 return HasExistingWhitespace(); 3661 if (Right.is(tok::coloncolon) && 3662 !Left.isOneOf(tok::l_brace, tok::comment, tok::l_paren)) 3663 // Put a space between < and :: in vector< ::std::string > 3664 return (Left.is(TT_TemplateOpener) && 3665 ((Style.Standard < FormatStyle::LS_Cpp11) || 3666 ShouldAddSpacesInAngles())) || 3667 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square, 3668 tok::kw___super, TT_TemplateOpener, 3669 TT_TemplateCloser)) || 3670 (Left.is(tok::l_paren) && Style.SpacesInParentheses); 3671 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser))) 3672 return ShouldAddSpacesInAngles(); 3673 // Space before TT_StructuredBindingLSquare. 3674 if (Right.is(TT_StructuredBindingLSquare)) 3675 return !Left.isOneOf(tok::amp, tok::ampamp) || 3676 getTokenReferenceAlignment(Left) != FormatStyle::PAS_Right; 3677 // Space before & or && following a TT_StructuredBindingLSquare. 3678 if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) && 3679 Right.isOneOf(tok::amp, tok::ampamp)) 3680 return getTokenReferenceAlignment(Right) != FormatStyle::PAS_Left; 3681 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) || 3682 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && 3683 !Right.is(tok::r_paren))) 3684 return true; 3685 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) && 3686 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen)) 3687 return false; 3688 if (Right.is(tok::less) && Left.isNot(tok::l_paren) && 3689 Line.startsWith(tok::hash)) 3690 return true; 3691 if (Right.is(TT_TrailingUnaryOperator)) 3692 return false; 3693 if (Left.is(TT_RegexLiteral)) 3694 return false; 3695 return spaceRequiredBetween(Line, Left, Right); 3696 } 3697 3698 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 3699 static bool isAllmanBrace(const FormatToken &Tok) { 3700 return Tok.is(tok::l_brace) && Tok.is(BK_Block) && 3701 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_LambdaLBrace, TT_DictLiteral); 3702 } 3703 3704 // Returns 'true' if 'Tok' is a function argument. 3705 static bool IsFunctionArgument(const FormatToken &Tok) { 3706 return Tok.MatchingParen && Tok.MatchingParen->Next && 3707 Tok.MatchingParen->Next->isOneOf(tok::comma, tok::r_paren); 3708 } 3709 3710 static bool 3711 isItAnEmptyLambdaAllowed(const FormatToken &Tok, 3712 FormatStyle::ShortLambdaStyle ShortLambdaOption) { 3713 return Tok.Children.empty() && ShortLambdaOption != FormatStyle::SLS_None; 3714 } 3715 3716 static bool isAllmanLambdaBrace(const FormatToken &Tok) { 3717 return (Tok.is(tok::l_brace) && Tok.is(BK_Block) && 3718 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral)); 3719 } 3720 3721 // Returns the first token on the line that is not a comment. 3722 static const FormatToken *getFirstNonComment(const AnnotatedLine &Line) { 3723 const FormatToken *Next = Line.First; 3724 if (!Next) 3725 return Next; 3726 if (Next->is(tok::comment)) 3727 Next = Next->getNextNonComment(); 3728 return Next; 3729 } 3730 3731 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 3732 const FormatToken &Right) { 3733 const FormatToken &Left = *Right.Previous; 3734 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0) 3735 return true; 3736 3737 if (Style.isCSharp()) { 3738 if (Left.is(TT_FatArrow) && Right.is(tok::l_brace) && 3739 Style.BraceWrapping.AfterFunction) 3740 return true; 3741 if (Right.is(TT_CSharpNamedArgumentColon) || 3742 Left.is(TT_CSharpNamedArgumentColon)) 3743 return false; 3744 if (Right.is(TT_CSharpGenericTypeConstraint)) 3745 return true; 3746 if (Right.Next && Right.Next->is(TT_FatArrow) && 3747 (Right.is(tok::numeric_constant) || 3748 (Right.is(tok::identifier) && Right.TokenText == "_"))) 3749 return true; 3750 3751 // Break after C# [...] and before public/protected/private/internal. 3752 if (Left.is(TT_AttributeSquare) && Left.is(tok::r_square) && 3753 (Right.isAccessSpecifier(/*ColonRequired=*/false) || 3754 Right.is(Keywords.kw_internal))) 3755 return true; 3756 // Break between ] and [ but only when there are really 2 attributes. 3757 if (Left.is(TT_AttributeSquare) && Right.is(TT_AttributeSquare) && 3758 Left.is(tok::r_square) && Right.is(tok::l_square)) 3759 return true; 3760 3761 } else if (Style.isJavaScript()) { 3762 // FIXME: This might apply to other languages and token kinds. 3763 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous && 3764 Left.Previous->is(tok::string_literal)) 3765 return true; 3766 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 && 3767 Left.Previous && Left.Previous->is(tok::equal) && 3768 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export, 3769 tok::kw_const) && 3770 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match 3771 // above. 3772 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) 3773 // Object literals on the top level of a file are treated as "enum-style". 3774 // Each key/value pair is put on a separate line, instead of bin-packing. 3775 return true; 3776 if (Left.is(tok::l_brace) && Line.Level == 0 && 3777 (Line.startsWith(tok::kw_enum) || 3778 Line.startsWith(tok::kw_const, tok::kw_enum) || 3779 Line.startsWith(tok::kw_export, tok::kw_enum) || 3780 Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum))) 3781 // JavaScript top-level enum key/value pairs are put on separate lines 3782 // instead of bin-packing. 3783 return true; 3784 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && Left.Previous && 3785 Left.Previous->is(TT_FatArrow)) { 3786 // JS arrow function (=> {...}). 3787 switch (Style.AllowShortLambdasOnASingleLine) { 3788 case FormatStyle::SLS_All: 3789 return false; 3790 case FormatStyle::SLS_None: 3791 return true; 3792 case FormatStyle::SLS_Empty: 3793 return !Left.Children.empty(); 3794 case FormatStyle::SLS_Inline: 3795 // allow one-lining inline (e.g. in function call args) and empty arrow 3796 // functions. 3797 return (Left.NestingLevel == 0 && Line.Level == 0) && 3798 !Left.Children.empty(); 3799 } 3800 llvm_unreachable("Unknown FormatStyle::ShortLambdaStyle enum"); 3801 } 3802 3803 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && 3804 !Left.Children.empty()) 3805 // Support AllowShortFunctionsOnASingleLine for JavaScript. 3806 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 3807 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty || 3808 (Left.NestingLevel == 0 && Line.Level == 0 && 3809 Style.AllowShortFunctionsOnASingleLine & 3810 FormatStyle::SFS_InlineOnly); 3811 } else if (Style.Language == FormatStyle::LK_Java) { 3812 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next && 3813 Right.Next->is(tok::string_literal)) 3814 return true; 3815 } else if (Style.Language == FormatStyle::LK_Cpp || 3816 Style.Language == FormatStyle::LK_ObjC || 3817 Style.Language == FormatStyle::LK_Proto || 3818 Style.Language == FormatStyle::LK_TableGen || 3819 Style.Language == FormatStyle::LK_TextProto) { 3820 if (Left.isStringLiteral() && Right.isStringLiteral()) 3821 return true; 3822 } 3823 3824 // Basic JSON newline processing. 3825 if (Style.isJson()) { 3826 // Always break after a JSON record opener. 3827 // { 3828 // } 3829 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace)) 3830 return true; 3831 // Always break after a JSON array opener. 3832 // [ 3833 // ] 3834 if (Left.is(TT_ArrayInitializerLSquare) && Left.is(tok::l_square) && 3835 !Right.is(tok::r_square)) 3836 return true; 3837 // Always break after successive entries. 3838 // 1, 3839 // 2 3840 if (Left.is(tok::comma)) 3841 return true; 3842 } 3843 3844 // If the last token before a '}', ']', or ')' is a comma or a trailing 3845 // comment, the intention is to insert a line break after it in order to make 3846 // shuffling around entries easier. Import statements, especially in 3847 // JavaScript, can be an exception to this rule. 3848 if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) { 3849 const FormatToken *BeforeClosingBrace = nullptr; 3850 if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) || 3851 (Style.isJavaScript() && Left.is(tok::l_paren))) && 3852 Left.isNot(BK_Block) && Left.MatchingParen) 3853 BeforeClosingBrace = Left.MatchingParen->Previous; 3854 else if (Right.MatchingParen && 3855 (Right.MatchingParen->isOneOf(tok::l_brace, 3856 TT_ArrayInitializerLSquare) || 3857 (Style.isJavaScript() && Right.MatchingParen->is(tok::l_paren)))) 3858 BeforeClosingBrace = &Left; 3859 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 3860 BeforeClosingBrace->isTrailingComment())) 3861 return true; 3862 } 3863 3864 if (Right.is(tok::comment)) 3865 return Left.isNot(BK_BracedInit) && Left.isNot(TT_CtorInitializerColon) && 3866 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 3867 if (Left.isTrailingComment()) 3868 return true; 3869 if (Left.IsUnterminatedLiteral) 3870 return true; 3871 if (Right.is(tok::lessless) && Right.Next && Left.is(tok::string_literal) && 3872 Right.Next->is(tok::string_literal)) 3873 return true; 3874 // Can break after template<> declaration 3875 if (Left.ClosesTemplateDeclaration && Left.MatchingParen && 3876 Left.MatchingParen->NestingLevel == 0) { 3877 // Put concepts on the next line e.g. 3878 // template<typename T> 3879 // concept ... 3880 if (Right.is(tok::kw_concept)) 3881 return Style.BreakBeforeConceptDeclarations; 3882 return (Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes); 3883 } 3884 if (Style.PackConstructorInitializers == FormatStyle::PCIS_Never) { 3885 if (Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeColon && 3886 (Left.is(TT_CtorInitializerComma) || Right.is(TT_CtorInitializerColon))) 3887 return true; 3888 3889 if (Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon && 3890 Left.isOneOf(TT_CtorInitializerColon, TT_CtorInitializerComma)) 3891 return true; 3892 } 3893 if (Style.PackConstructorInitializers < FormatStyle::PCIS_CurrentLine && 3894 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 3895 Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon)) 3896 return true; 3897 // Break only if we have multiple inheritance. 3898 if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma && 3899 Right.is(TT_InheritanceComma)) 3900 return true; 3901 if (Style.BreakInheritanceList == FormatStyle::BILS_AfterComma && 3902 Left.is(TT_InheritanceComma)) 3903 return true; 3904 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) 3905 // Multiline raw string literals are special wrt. line breaks. The author 3906 // has made a deliberate choice and might have aligned the contents of the 3907 // string literal accordingly. Thus, we try keep existing line breaks. 3908 return Right.IsMultiline && Right.NewlinesBefore > 0; 3909 if ((Left.is(tok::l_brace) || (Left.is(tok::less) && Left.Previous && 3910 Left.Previous->is(tok::equal))) && 3911 Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) { 3912 // Don't put enums or option definitions onto single lines in protocol 3913 // buffers. 3914 return true; 3915 } 3916 if (Right.is(TT_InlineASMBrace)) 3917 return Right.HasUnescapedNewline; 3918 3919 if (isAllmanBrace(Left) || isAllmanBrace(Right)) { 3920 auto FirstNonComment = getFirstNonComment(Line); 3921 bool AccessSpecifier = 3922 FirstNonComment && 3923 FirstNonComment->isOneOf(Keywords.kw_internal, tok::kw_public, 3924 tok::kw_private, tok::kw_protected); 3925 3926 if (Style.BraceWrapping.AfterEnum) { 3927 if (Line.startsWith(tok::kw_enum) || 3928 Line.startsWith(tok::kw_typedef, tok::kw_enum)) 3929 return true; 3930 // Ensure BraceWrapping for `public enum A {`. 3931 if (AccessSpecifier && FirstNonComment->Next && 3932 FirstNonComment->Next->is(tok::kw_enum)) 3933 return true; 3934 } 3935 3936 // Ensure BraceWrapping for `public interface A {`. 3937 if (Style.BraceWrapping.AfterClass && 3938 ((AccessSpecifier && FirstNonComment->Next && 3939 FirstNonComment->Next->is(Keywords.kw_interface)) || 3940 Line.startsWith(Keywords.kw_interface))) 3941 return true; 3942 3943 return (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) || 3944 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct); 3945 } 3946 3947 if (Left.is(TT_ObjCBlockLBrace) && 3948 Style.AllowShortBlocksOnASingleLine == FormatStyle::SBS_Never) 3949 return true; 3950 3951 // Ensure wrapping after __attribute__((XX)) and @interface etc. 3952 if (Left.is(TT_AttributeParen) && Right.is(TT_ObjCDecl)) 3953 return true; 3954 3955 if (Left.is(TT_LambdaLBrace)) { 3956 if (IsFunctionArgument(Left) && 3957 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline) 3958 return false; 3959 3960 if (Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_None || 3961 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline || 3962 (!Left.Children.empty() && 3963 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Empty)) 3964 return true; 3965 } 3966 3967 if (Style.BraceWrapping.BeforeLambdaBody && Right.is(TT_LambdaLBrace) && 3968 Left.isOneOf(tok::star, tok::amp, tok::ampamp, TT_TemplateCloser)) { 3969 return true; 3970 } 3971 3972 // Put multiple Java annotation on a new line. 3973 if ((Style.Language == FormatStyle::LK_Java || Style.isJavaScript()) && 3974 Left.is(TT_LeadingJavaAnnotation) && 3975 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) && 3976 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) 3977 return true; 3978 3979 if (Right.is(TT_ProtoExtensionLSquare)) 3980 return true; 3981 3982 // In text proto instances if a submessage contains at least 2 entries and at 3983 // least one of them is a submessage, like A { ... B { ... } ... }, 3984 // put all of the entries of A on separate lines by forcing the selector of 3985 // the submessage B to be put on a newline. 3986 // 3987 // Example: these can stay on one line: 3988 // a { scalar_1: 1 scalar_2: 2 } 3989 // a { b { key: value } } 3990 // 3991 // and these entries need to be on a new line even if putting them all in one 3992 // line is under the column limit: 3993 // a { 3994 // scalar: 1 3995 // b { key: value } 3996 // } 3997 // 3998 // We enforce this by breaking before a submessage field that has previous 3999 // siblings, *and* breaking before a field that follows a submessage field. 4000 // 4001 // Be careful to exclude the case [proto.ext] { ... } since the `]` is 4002 // the TT_SelectorName there, but we don't want to break inside the brackets. 4003 // 4004 // Another edge case is @submessage { key: value }, which is a common 4005 // substitution placeholder. In this case we want to keep `@` and `submessage` 4006 // together. 4007 // 4008 // We ensure elsewhere that extensions are always on their own line. 4009 if ((Style.Language == FormatStyle::LK_Proto || 4010 Style.Language == FormatStyle::LK_TextProto) && 4011 Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) { 4012 // Keep `@submessage` together in: 4013 // @submessage { key: value } 4014 if (Left.is(tok::at)) 4015 return false; 4016 // Look for the scope opener after selector in cases like: 4017 // selector { ... 4018 // selector: { ... 4019 // selector: @base { ... 4020 FormatToken *LBrace = Right.Next; 4021 if (LBrace && LBrace->is(tok::colon)) { 4022 LBrace = LBrace->Next; 4023 if (LBrace && LBrace->is(tok::at)) { 4024 LBrace = LBrace->Next; 4025 if (LBrace) 4026 LBrace = LBrace->Next; 4027 } 4028 } 4029 if (LBrace && 4030 // The scope opener is one of {, [, <: 4031 // selector { ... } 4032 // selector [ ... ] 4033 // selector < ... > 4034 // 4035 // In case of selector { ... }, the l_brace is TT_DictLiteral. 4036 // In case of an empty selector {}, the l_brace is not TT_DictLiteral, 4037 // so we check for immediately following r_brace. 4038 ((LBrace->is(tok::l_brace) && 4039 (LBrace->is(TT_DictLiteral) || 4040 (LBrace->Next && LBrace->Next->is(tok::r_brace)))) || 4041 LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) { 4042 // If Left.ParameterCount is 0, then this submessage entry is not the 4043 // first in its parent submessage, and we want to break before this entry. 4044 // If Left.ParameterCount is greater than 0, then its parent submessage 4045 // might contain 1 or more entries and we want to break before this entry 4046 // if it contains at least 2 entries. We deal with this case later by 4047 // detecting and breaking before the next entry in the parent submessage. 4048 if (Left.ParameterCount == 0) 4049 return true; 4050 // However, if this submessage is the first entry in its parent 4051 // submessage, Left.ParameterCount might be 1 in some cases. 4052 // We deal with this case later by detecting an entry 4053 // following a closing paren of this submessage. 4054 } 4055 4056 // If this is an entry immediately following a submessage, it will be 4057 // preceded by a closing paren of that submessage, like in: 4058 // left---. .---right 4059 // v v 4060 // sub: { ... } key: value 4061 // If there was a comment between `}` an `key` above, then `key` would be 4062 // put on a new line anyways. 4063 if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square)) 4064 return true; 4065 } 4066 4067 // Deal with lambda arguments in C++ - we want consistent line breaks whether 4068 // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced 4069 // as aggressive line breaks are placed when the lambda is not the last arg. 4070 if ((Style.Language == FormatStyle::LK_Cpp || 4071 Style.Language == FormatStyle::LK_ObjC) && 4072 Left.is(tok::l_paren) && Left.BlockParameterCount > 0 && 4073 !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) { 4074 // Multiple lambdas in the same function call force line breaks. 4075 if (Left.BlockParameterCount > 1) 4076 return true; 4077 4078 // A lambda followed by another arg forces a line break. 4079 if (!Left.Role) 4080 return false; 4081 auto Comma = Left.Role->lastComma(); 4082 if (!Comma) 4083 return false; 4084 auto Next = Comma->getNextNonComment(); 4085 if (!Next) 4086 return false; 4087 if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret)) 4088 return true; 4089 } 4090 4091 return false; 4092 } 4093 4094 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 4095 const FormatToken &Right) { 4096 const FormatToken &Left = *Right.Previous; 4097 // Language-specific stuff. 4098 if (Style.isCSharp()) { 4099 if (Left.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon) || 4100 Right.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon)) 4101 return false; 4102 // Only break after commas for generic type constraints. 4103 if (Line.First->is(TT_CSharpGenericTypeConstraint)) 4104 return Left.is(TT_CSharpGenericTypeConstraintComma); 4105 // Keep nullable operators attached to their identifiers. 4106 if (Right.is(TT_CSharpNullable)) { 4107 return false; 4108 } 4109 } else if (Style.Language == FormatStyle::LK_Java) { 4110 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 4111 Keywords.kw_implements)) 4112 return false; 4113 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 4114 Keywords.kw_implements)) 4115 return true; 4116 } else if (Style.isJavaScript()) { 4117 const FormatToken *NonComment = Right.getPreviousNonComment(); 4118 if (NonComment && 4119 NonComment->isOneOf( 4120 tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break, 4121 tok::kw_throw, Keywords.kw_interface, Keywords.kw_type, 4122 tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected, 4123 Keywords.kw_readonly, Keywords.kw_override, Keywords.kw_abstract, 4124 Keywords.kw_get, Keywords.kw_set, Keywords.kw_async, 4125 Keywords.kw_await)) 4126 return false; // Otherwise automatic semicolon insertion would trigger. 4127 if (Right.NestingLevel == 0 && 4128 (Left.Tok.getIdentifierInfo() || 4129 Left.isOneOf(tok::r_square, tok::r_paren)) && 4130 Right.isOneOf(tok::l_square, tok::l_paren)) 4131 return false; // Otherwise automatic semicolon insertion would trigger. 4132 if (NonComment && NonComment->is(tok::identifier) && 4133 NonComment->TokenText == "asserts") 4134 return false; 4135 if (Left.is(TT_FatArrow) && Right.is(tok::l_brace)) 4136 return false; 4137 if (Left.is(TT_JsTypeColon)) 4138 return true; 4139 // Don't wrap between ":" and "!" of a strict prop init ("field!: type;"). 4140 if (Left.is(tok::exclaim) && Right.is(tok::colon)) 4141 return false; 4142 // Look for is type annotations like: 4143 // function f(): a is B { ... } 4144 // Do not break before is in these cases. 4145 if (Right.is(Keywords.kw_is)) { 4146 const FormatToken *Next = Right.getNextNonComment(); 4147 // If `is` is followed by a colon, it's likely that it's a dict key, so 4148 // ignore it for this check. 4149 // For example this is common in Polymer: 4150 // Polymer({ 4151 // is: 'name', 4152 // ... 4153 // }); 4154 if (!Next || !Next->is(tok::colon)) 4155 return false; 4156 } 4157 if (Left.is(Keywords.kw_in)) 4158 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None; 4159 if (Right.is(Keywords.kw_in)) 4160 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None; 4161 if (Right.is(Keywords.kw_as)) 4162 return false; // must not break before as in 'x as type' casts 4163 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) { 4164 // extends and infer can appear as keywords in conditional types: 4165 // https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types 4166 // do not break before them, as the expressions are subject to ASI. 4167 return false; 4168 } 4169 if (Left.is(Keywords.kw_as)) 4170 return true; 4171 if (Left.is(TT_NonNullAssertion)) 4172 return true; 4173 if (Left.is(Keywords.kw_declare) && 4174 Right.isOneOf(Keywords.kw_module, tok::kw_namespace, 4175 Keywords.kw_function, tok::kw_class, tok::kw_enum, 4176 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var, 4177 Keywords.kw_let, tok::kw_const)) 4178 // See grammar for 'declare' statements at: 4179 // https://github.com/Microsoft/TypeScript/blob/main/doc/spec-ARCHIVED.md#A.10 4180 return false; 4181 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) && 4182 Right.isOneOf(tok::identifier, tok::string_literal)) 4183 return false; // must not break in "module foo { ...}" 4184 if (Right.is(TT_TemplateString) && Right.closesScope()) 4185 return false; 4186 // Don't split tagged template literal so there is a break between the tag 4187 // identifier and template string. 4188 if (Left.is(tok::identifier) && Right.is(TT_TemplateString)) { 4189 return false; 4190 } 4191 if (Left.is(TT_TemplateString) && Left.opensScope()) 4192 return true; 4193 } 4194 4195 if (Left.is(tok::at)) 4196 return false; 4197 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 4198 return false; 4199 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation)) 4200 return !Right.is(tok::l_paren); 4201 if (Right.is(TT_PointerOrReference)) 4202 return Line.IsMultiVariableDeclStmt || 4203 (getTokenPointerOrReferenceAlignment(Right) == 4204 FormatStyle::PAS_Right && 4205 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName))); 4206 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 4207 Right.is(tok::kw_operator)) 4208 return true; 4209 if (Left.is(TT_PointerOrReference)) 4210 return false; 4211 if (Right.isTrailingComment()) 4212 // We rely on MustBreakBefore being set correctly here as we should not 4213 // change the "binding" behavior of a comment. 4214 // The first comment in a braced lists is always interpreted as belonging to 4215 // the first list element. Otherwise, it should be placed outside of the 4216 // list. 4217 return Left.is(BK_BracedInit) || 4218 (Left.is(TT_CtorInitializerColon) && 4219 Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon); 4220 if (Left.is(tok::question) && Right.is(tok::colon)) 4221 return false; 4222 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question)) 4223 return Style.BreakBeforeTernaryOperators; 4224 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question)) 4225 return !Style.BreakBeforeTernaryOperators; 4226 if (Left.is(TT_InheritanceColon)) 4227 return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon; 4228 if (Right.is(TT_InheritanceColon)) 4229 return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon; 4230 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) && 4231 Left.isNot(TT_SelectorName)) 4232 return true; 4233 4234 if (Right.is(tok::colon) && 4235 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) 4236 return false; 4237 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) { 4238 if (Style.Language == FormatStyle::LK_Proto || 4239 Style.Language == FormatStyle::LK_TextProto) { 4240 if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral()) 4241 return false; 4242 // Prevent cases like: 4243 // 4244 // submessage: 4245 // { key: valueeeeeeeeeeee } 4246 // 4247 // when the snippet does not fit into one line. 4248 // Prefer: 4249 // 4250 // submessage: { 4251 // key: valueeeeeeeeeeee 4252 // } 4253 // 4254 // instead, even if it is longer by one line. 4255 // 4256 // Note that this allows allows the "{" to go over the column limit 4257 // when the column limit is just between ":" and "{", but that does 4258 // not happen too often and alternative formattings in this case are 4259 // not much better. 4260 // 4261 // The code covers the cases: 4262 // 4263 // submessage: { ... } 4264 // submessage: < ... > 4265 // repeated: [ ... ] 4266 if (((Right.is(tok::l_brace) || Right.is(tok::less)) && 4267 Right.is(TT_DictLiteral)) || 4268 Right.is(TT_ArrayInitializerLSquare)) 4269 return false; 4270 } 4271 return true; 4272 } 4273 if (Right.is(tok::r_square) && Right.MatchingParen && 4274 Right.MatchingParen->is(TT_ProtoExtensionLSquare)) 4275 return false; 4276 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next && 4277 Right.Next->is(TT_ObjCMethodExpr))) 4278 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls. 4279 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 4280 return true; 4281 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen)) 4282 return true; 4283 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen, 4284 TT_OverloadedOperator)) 4285 return false; 4286 if (Left.is(TT_RangeBasedForLoopColon)) 4287 return true; 4288 if (Right.is(TT_RangeBasedForLoopColon)) 4289 return false; 4290 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener)) 4291 return true; 4292 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) || 4293 Left.is(tok::kw_operator)) 4294 return false; 4295 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) && 4296 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) 4297 return false; 4298 if (Left.is(tok::equal) && Right.is(tok::l_brace) && 4299 !Style.Cpp11BracedListStyle) 4300 return false; 4301 if (Left.is(tok::l_paren) && 4302 Left.isOneOf(TT_AttributeParen, TT_TypeDeclarationParen)) 4303 return false; 4304 if (Left.is(tok::l_paren) && Left.Previous && 4305 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) 4306 return false; 4307 if (Right.is(TT_ImplicitStringLiteral)) 4308 return false; 4309 4310 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser)) 4311 return false; 4312 if (Right.is(tok::r_square) && Right.MatchingParen && 4313 Right.MatchingParen->is(TT_LambdaLSquare)) 4314 return false; 4315 4316 // We only break before r_brace if there was a corresponding break before 4317 // the l_brace, which is tracked by BreakBeforeClosingBrace. 4318 if (Right.is(tok::r_brace)) 4319 return Right.MatchingParen && Right.MatchingParen->is(BK_Block); 4320 4321 // Allow breaking after a trailing annotation, e.g. after a method 4322 // declaration. 4323 if (Left.is(TT_TrailingAnnotation)) 4324 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 4325 tok::less, tok::coloncolon); 4326 4327 if (Right.is(tok::kw___attribute) || 4328 (Right.is(tok::l_square) && Right.is(TT_AttributeSquare))) 4329 return !Left.is(TT_AttributeSquare); 4330 4331 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 4332 return true; 4333 4334 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 4335 return true; 4336 4337 if (Left.is(TT_CtorInitializerColon)) 4338 return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon; 4339 if (Right.is(TT_CtorInitializerColon)) 4340 return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon; 4341 if (Left.is(TT_CtorInitializerComma) && 4342 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 4343 return false; 4344 if (Right.is(TT_CtorInitializerComma) && 4345 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 4346 return true; 4347 if (Left.is(TT_InheritanceComma) && 4348 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) 4349 return false; 4350 if (Right.is(TT_InheritanceComma) && 4351 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) 4352 return true; 4353 if ((Left.is(tok::greater) && Right.is(tok::greater)) || 4354 (Left.is(tok::less) && Right.is(tok::less))) 4355 return false; 4356 if (Right.is(TT_BinaryOperator) && 4357 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 4358 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 4359 Right.getPrecedence() != prec::Assignment)) 4360 return true; 4361 if (Left.is(TT_ArrayInitializerLSquare)) 4362 return true; 4363 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 4364 return true; 4365 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) && 4366 !Left.isOneOf(tok::arrowstar, tok::lessless) && 4367 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 4368 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 4369 Left.getPrecedence() == prec::Assignment)) 4370 return true; 4371 if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) || 4372 (Left.is(tok::r_square) && Right.is(TT_AttributeSquare))) 4373 return false; 4374 4375 auto ShortLambdaOption = Style.AllowShortLambdasOnASingleLine; 4376 if (Style.BraceWrapping.BeforeLambdaBody && Right.is(TT_LambdaLBrace)) { 4377 if (isAllmanLambdaBrace(Left)) 4378 return !isItAnEmptyLambdaAllowed(Left, ShortLambdaOption); 4379 if (isAllmanLambdaBrace(Right)) 4380 return !isItAnEmptyLambdaAllowed(Right, ShortLambdaOption); 4381 } 4382 4383 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 4384 tok::kw_class, tok::kw_struct, tok::comment) || 4385 Right.isMemberAccess() || 4386 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless, 4387 tok::colon, tok::l_square, tok::at) || 4388 (Left.is(tok::r_paren) && 4389 Right.isOneOf(tok::identifier, tok::kw_const)) || 4390 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) || 4391 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser)); 4392 } 4393 4394 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 4395 llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n"; 4396 const FormatToken *Tok = Line.First; 4397 while (Tok) { 4398 llvm::errs() << " M=" << Tok->MustBreakBefore 4399 << " C=" << Tok->CanBreakBefore 4400 << " T=" << getTokenTypeName(Tok->getType()) 4401 << " S=" << Tok->SpacesRequiredBefore 4402 << " F=" << Tok->Finalized << " B=" << Tok->BlockParameterCount 4403 << " BK=" << Tok->getBlockKind() << " P=" << Tok->SplitPenalty 4404 << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength 4405 << " PPK=" << Tok->getPackingKind() << " FakeLParens="; 4406 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 4407 llvm::errs() << Tok->FakeLParens[i] << "/"; 4408 llvm::errs() << " FakeRParens=" << Tok->FakeRParens; 4409 llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo(); 4410 llvm::errs() << " Text='" << Tok->TokenText << "'\n"; 4411 if (!Tok->Next) 4412 assert(Tok == Line.Last); 4413 Tok = Tok->Next; 4414 } 4415 llvm::errs() << "----\n"; 4416 } 4417 4418 FormatStyle::PointerAlignmentStyle 4419 TokenAnnotator::getTokenReferenceAlignment(const FormatToken &Reference) { 4420 assert(Reference.isOneOf(tok::amp, tok::ampamp)); 4421 switch (Style.ReferenceAlignment) { 4422 case FormatStyle::RAS_Pointer: 4423 return Style.PointerAlignment; 4424 case FormatStyle::RAS_Left: 4425 return FormatStyle::PAS_Left; 4426 case FormatStyle::RAS_Right: 4427 return FormatStyle::PAS_Right; 4428 case FormatStyle::RAS_Middle: 4429 return FormatStyle::PAS_Middle; 4430 } 4431 assert(0); //"Unhandled value of ReferenceAlignment" 4432 return Style.PointerAlignment; 4433 } 4434 4435 FormatStyle::PointerAlignmentStyle 4436 TokenAnnotator::getTokenPointerOrReferenceAlignment( 4437 const FormatToken &PointerOrReference) { 4438 if (PointerOrReference.isOneOf(tok::amp, tok::ampamp)) { 4439 switch (Style.ReferenceAlignment) { 4440 case FormatStyle::RAS_Pointer: 4441 return Style.PointerAlignment; 4442 case FormatStyle::RAS_Left: 4443 return FormatStyle::PAS_Left; 4444 case FormatStyle::RAS_Right: 4445 return FormatStyle::PAS_Right; 4446 case FormatStyle::RAS_Middle: 4447 return FormatStyle::PAS_Middle; 4448 } 4449 } 4450 assert(PointerOrReference.is(tok::star)); 4451 return Style.PointerAlignment; 4452 } 4453 4454 } // namespace format 4455 } // namespace clang 4456