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