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