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 "llvm/ADT/SmallPtrSet.h" 19 #include "llvm/Support/Debug.h" 20 21 #define DEBUG_TYPE "format-token-annotator" 22 23 namespace clang { 24 namespace format { 25 26 namespace { 27 28 /// Returns \c true if the token can be used as an identifier in 29 /// an Objective-C \c @selector, \c false otherwise. 30 /// 31 /// Because getFormattingLangOpts() always lexes source code as 32 /// Objective-C++, C++ keywords like \c new and \c delete are 33 /// lexed as tok::kw_*, not tok::identifier, even for Objective-C. 34 /// 35 /// For Objective-C and Objective-C++, both identifiers and keywords 36 /// are valid inside @selector(...) (or a macro which 37 /// invokes @selector(...)). So, we allow treat any identifier or 38 /// keyword as a potential Objective-C selector component. 39 static bool canBeObjCSelectorComponent(const FormatToken &Tok) { 40 return Tok.Tok.getIdentifierInfo() != nullptr; 41 } 42 43 /// With `Left` being '(', check if we're at either `[...](` or 44 /// `[...]<...>(`, where the [ opens a lambda capture list. 45 static bool isLambdaParameterList(const FormatToken *Left) { 46 // Skip <...> if present. 47 if (Left->Previous && Left->Previous->is(tok::greater) && 48 Left->Previous->MatchingParen && 49 Left->Previous->MatchingParen->is(TT_TemplateOpener)) 50 Left = Left->Previous->MatchingParen; 51 52 // Check for `[...]`. 53 return Left->Previous && Left->Previous->is(tok::r_square) && 54 Left->Previous->MatchingParen && 55 Left->Previous->MatchingParen->is(TT_LambdaLSquare); 56 } 57 58 /// A parser that gathers additional information about tokens. 59 /// 60 /// The \c TokenAnnotator tries to match parenthesis and square brakets and 61 /// store a parenthesis levels. It also tries to resolve matching "<" and ">" 62 /// into template parameter lists. 63 class AnnotatingParser { 64 public: 65 AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line, 66 const AdditionalKeywords &Keywords) 67 : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false), 68 TrailingReturnFound(false), Keywords(Keywords) { 69 Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false)); 70 resetTokenMetadata(CurrentToken); 71 } 72 73 private: 74 bool parseAngle() { 75 if (!CurrentToken || !CurrentToken->Previous) 76 return false; 77 if (NonTemplateLess.count(CurrentToken->Previous)) 78 return false; 79 80 const FormatToken &Previous = *CurrentToken->Previous; // The '<'. 81 if (Previous.Previous) { 82 if (Previous.Previous->Tok.isLiteral()) 83 return false; 84 if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 && 85 (!Previous.Previous->MatchingParen || 86 !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen))) 87 return false; 88 } 89 90 FormatToken *Left = CurrentToken->Previous; 91 Left->ParentBracket = Contexts.back().ContextKind; 92 ScopedContextCreator ContextCreator(*this, tok::less, 12); 93 94 // If this angle is in the context of an expression, we need to be more 95 // hesitant to detect it as opening template parameters. 96 bool InExprContext = Contexts.back().IsExpression; 97 98 Contexts.back().IsExpression = false; 99 // If there's a template keyword before the opening angle bracket, this is a 100 // template parameter, not an argument. 101 Contexts.back().InTemplateArgument = 102 Left->Previous && Left->Previous->Tok.isNot(tok::kw_template); 103 104 if (Style.Language == FormatStyle::LK_Java && 105 CurrentToken->is(tok::question)) 106 next(); 107 108 while (CurrentToken) { 109 if (CurrentToken->is(tok::greater)) { 110 Left->MatchingParen = CurrentToken; 111 CurrentToken->MatchingParen = Left; 112 // In TT_Proto, we must distignuish between: 113 // map<key, value> 114 // msg < item: data > 115 // msg: < item: data > 116 // In TT_TextProto, map<key, value> does not occur. 117 if (Style.Language == FormatStyle::LK_TextProto || 118 (Style.Language == FormatStyle::LK_Proto && Left->Previous && 119 Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) 120 CurrentToken->Type = TT_DictLiteral; 121 else 122 CurrentToken->Type = TT_TemplateCloser; 123 next(); 124 return true; 125 } 126 if (CurrentToken->is(tok::question) && 127 Style.Language == FormatStyle::LK_Java) { 128 next(); 129 continue; 130 } 131 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) || 132 (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext && 133 Style.Language != FormatStyle::LK_Proto && 134 Style.Language != FormatStyle::LK_TextProto)) 135 return false; 136 // If a && or || is found and interpreted as a binary operator, this set 137 // of angles is likely part of something like "a < b && c > d". If the 138 // angles are inside an expression, the ||/&& might also be a binary 139 // operator that was misinterpreted because we are parsing template 140 // parameters. 141 // FIXME: This is getting out of hand, write a decent parser. 142 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) && 143 CurrentToken->Previous->is(TT_BinaryOperator) && 144 Contexts[Contexts.size() - 2].IsExpression && 145 !Line.startsWith(tok::kw_template)) 146 return false; 147 updateParameterCount(Left, CurrentToken); 148 if (Style.Language == FormatStyle::LK_Proto) { 149 if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) { 150 if (CurrentToken->is(tok::colon) || 151 (CurrentToken->isOneOf(tok::l_brace, tok::less) && 152 Previous->isNot(tok::colon))) 153 Previous->Type = TT_SelectorName; 154 } 155 } 156 if (!consumeToken()) 157 return false; 158 } 159 return false; 160 } 161 162 bool parseParens(bool LookForDecls = false) { 163 if (!CurrentToken) 164 return false; 165 FormatToken *Left = CurrentToken->Previous; 166 Left->ParentBracket = Contexts.back().ContextKind; 167 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1); 168 169 // FIXME: This is a bit of a hack. Do better. 170 Contexts.back().ColonIsForRangeExpr = 171 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr; 172 173 bool StartsObjCMethodExpr = false; 174 if (FormatToken *MaybeSel = Left->Previous) { 175 // @selector( starts a selector. 176 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous && 177 MaybeSel->Previous->is(tok::at)) { 178 StartsObjCMethodExpr = true; 179 } 180 } 181 182 if (Left->is(TT_OverloadedOperatorLParen)) { 183 Contexts.back().IsExpression = false; 184 } else if (Style.Language == FormatStyle::LK_JavaScript && 185 (Line.startsWith(Keywords.kw_type, tok::identifier) || 186 Line.startsWith(tok::kw_export, Keywords.kw_type, 187 tok::identifier))) { 188 // type X = (...); 189 // export type X = (...); 190 Contexts.back().IsExpression = false; 191 } else if (Left->Previous && 192 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype, 193 tok::kw_while, tok::l_paren, 194 tok::comma) || 195 Left->Previous->isIf() || 196 Left->Previous->is(TT_BinaryOperator))) { 197 // static_assert, if and while usually contain expressions. 198 Contexts.back().IsExpression = true; 199 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 200 (Left->Previous->is(Keywords.kw_function) || 201 (Left->Previous->endsSequence(tok::identifier, 202 Keywords.kw_function)))) { 203 // function(...) or function f(...) 204 Contexts.back().IsExpression = false; 205 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 206 Left->Previous->is(TT_JsTypeColon)) { 207 // let x: (SomeType); 208 Contexts.back().IsExpression = false; 209 } else if (isLambdaParameterList(Left)) { 210 // This is a parameter list of a lambda expression. 211 Contexts.back().IsExpression = false; 212 } else if (Line.InPPDirective && 213 (!Left->Previous || !Left->Previous->is(tok::identifier))) { 214 Contexts.back().IsExpression = true; 215 } else if (Contexts[Contexts.size() - 2].CaretFound) { 216 // This is the parameter list of an ObjC block. 217 Contexts.back().IsExpression = false; 218 } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) { 219 Left->Type = TT_AttributeParen; 220 } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) { 221 // The first argument to a foreach macro is a declaration. 222 Contexts.back().IsForEachMacro = true; 223 Contexts.back().IsExpression = false; 224 } else if (Left->Previous && Left->Previous->MatchingParen && 225 Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) { 226 Contexts.back().IsExpression = false; 227 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) { 228 bool IsForOrCatch = 229 Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch); 230 Contexts.back().IsExpression = !IsForOrCatch; 231 } 232 233 if (StartsObjCMethodExpr) { 234 Contexts.back().ColonIsObjCMethodExpr = true; 235 Left->Type = TT_ObjCMethodExpr; 236 } 237 238 // MightBeFunctionType and ProbablyFunctionType are used for 239 // function pointer and reference types as well as Objective-C 240 // block types: 241 // 242 // void (*FunctionPointer)(void); 243 // void (&FunctionReference)(void); 244 // void (^ObjCBlock)(void); 245 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression; 246 bool ProbablyFunctionType = 247 CurrentToken->isOneOf(tok::star, tok::amp, tok::caret); 248 bool HasMultipleLines = false; 249 bool HasMultipleParametersOnALine = false; 250 bool MightBeObjCForRangeLoop = 251 Left->Previous && Left->Previous->is(tok::kw_for); 252 FormatToken *PossibleObjCForInToken = nullptr; 253 while (CurrentToken) { 254 // LookForDecls is set when "if (" has been seen. Check for 255 // 'identifier' '*' 'identifier' followed by not '=' -- this 256 // '*' has to be a binary operator but determineStarAmpUsage() will 257 // categorize it as an unary operator, so set the right type here. 258 if (LookForDecls && CurrentToken->Next) { 259 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 260 if (Prev) { 261 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 262 FormatToken *Next = CurrentToken->Next; 263 if (PrevPrev && PrevPrev->is(tok::identifier) && 264 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 265 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 266 Prev->Type = TT_BinaryOperator; 267 LookForDecls = false; 268 } 269 } 270 } 271 272 if (CurrentToken->Previous->is(TT_PointerOrReference) && 273 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 274 tok::coloncolon)) 275 ProbablyFunctionType = true; 276 if (CurrentToken->is(tok::comma)) 277 MightBeFunctionType = false; 278 if (CurrentToken->Previous->is(TT_BinaryOperator)) 279 Contexts.back().IsExpression = true; 280 if (CurrentToken->is(tok::r_paren)) { 281 if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next && 282 (CurrentToken->Next->is(tok::l_paren) || 283 (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration))) 284 Left->Type = Left->Next->is(tok::caret) ? TT_ObjCBlockLParen 285 : TT_FunctionTypeLParen; 286 Left->MatchingParen = CurrentToken; 287 CurrentToken->MatchingParen = Left; 288 289 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) && 290 Left->Previous && Left->Previous->is(tok::l_paren)) { 291 // Detect the case where macros are used to generate lambdas or 292 // function bodies, e.g.: 293 // auto my_lambda = MARCO((Type *type, int i) { .. body .. }); 294 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) { 295 if (Tok->is(TT_BinaryOperator) && 296 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) 297 Tok->Type = TT_PointerOrReference; 298 } 299 } 300 301 if (StartsObjCMethodExpr) { 302 CurrentToken->Type = TT_ObjCMethodExpr; 303 if (Contexts.back().FirstObjCSelectorName) { 304 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 305 Contexts.back().LongestObjCSelectorName; 306 } 307 } 308 309 if (Left->is(TT_AttributeParen)) 310 CurrentToken->Type = TT_AttributeParen; 311 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation)) 312 CurrentToken->Type = TT_JavaAnnotation; 313 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation)) 314 CurrentToken->Type = TT_LeadingJavaAnnotation; 315 if (Left->Previous && Left->Previous->is(TT_AttributeSquare)) 316 CurrentToken->Type = TT_AttributeSquare; 317 318 if (!HasMultipleLines) 319 Left->PackingKind = PPK_Inconclusive; 320 else if (HasMultipleParametersOnALine) 321 Left->PackingKind = PPK_BinPacked; 322 else 323 Left->PackingKind = PPK_OnePerLine; 324 325 next(); 326 return true; 327 } 328 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 329 return false; 330 331 if (CurrentToken->is(tok::l_brace)) 332 Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen 333 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 334 !CurrentToken->Next->HasUnescapedNewline && 335 !CurrentToken->Next->isTrailingComment()) 336 HasMultipleParametersOnALine = true; 337 if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) || 338 CurrentToken->Previous->isSimpleTypeSpecifier()) && 339 !CurrentToken->is(tok::l_brace)) 340 Contexts.back().IsExpression = false; 341 if (CurrentToken->isOneOf(tok::semi, tok::colon)) { 342 MightBeObjCForRangeLoop = false; 343 if (PossibleObjCForInToken) { 344 PossibleObjCForInToken->Type = TT_Unknown; 345 PossibleObjCForInToken = nullptr; 346 } 347 } 348 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) { 349 PossibleObjCForInToken = CurrentToken; 350 PossibleObjCForInToken->Type = TT_ObjCForIn; 351 } 352 // When we discover a 'new', we set CanBeExpression to 'false' in order to 353 // parse the type correctly. Reset that after a comma. 354 if (CurrentToken->is(tok::comma)) 355 Contexts.back().CanBeExpression = true; 356 357 FormatToken *Tok = CurrentToken; 358 if (!consumeToken()) 359 return false; 360 updateParameterCount(Left, Tok); 361 if (CurrentToken && CurrentToken->HasUnescapedNewline) 362 HasMultipleLines = true; 363 } 364 return false; 365 } 366 367 bool isCSharpAttributeSpecifier(const FormatToken &Tok) { 368 if (!Style.isCSharp()) 369 return false; 370 371 const FormatToken *AttrTok = Tok.Next; 372 if (!AttrTok) 373 return false; 374 375 // Just an empty declaration e.g. string []. 376 if (AttrTok->is(tok::r_square)) 377 return false; 378 379 // Move along the tokens inbetween the '[' and ']' e.g. [STAThread]. 380 while (AttrTok && AttrTok->isNot(tok::r_square)) { 381 AttrTok = AttrTok->Next; 382 } 383 384 if (!AttrTok) 385 return false; 386 387 // Move past the end of ']'. 388 AttrTok = AttrTok->Next; 389 if (!AttrTok) 390 return false; 391 392 // Limit this to being an access modifier that follows. 393 if (AttrTok->isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected, 394 tok::kw_class, tok::kw_static, tok::l_square, 395 Keywords.kw_internal)) { 396 return true; 397 } 398 399 // incase its a [XXX] retval func(.... 400 if (AttrTok->Next && 401 AttrTok->Next->startsSequence(tok::identifier, tok::l_paren)) 402 return true; 403 404 return false; 405 } 406 407 bool isCpp11AttributeSpecifier(const FormatToken &Tok) { 408 if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square)) 409 return false; 410 // The first square bracket is part of an ObjC array literal 411 if (Tok.Previous && Tok.Previous->is(tok::at)) { 412 return false; 413 } 414 const FormatToken *AttrTok = Tok.Next->Next; 415 if (!AttrTok) 416 return false; 417 // C++17 '[[using ns: foo, bar(baz, blech)]]' 418 // We assume nobody will name an ObjC variable 'using'. 419 if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon)) 420 return true; 421 if (AttrTok->isNot(tok::identifier)) 422 return false; 423 while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) { 424 // ObjC message send. We assume nobody will use : in a C++11 attribute 425 // specifier parameter, although this is technically valid: 426 // [[foo(:)]]. 427 if (AttrTok->is(tok::colon) || 428 AttrTok->startsSequence(tok::identifier, tok::identifier) || 429 AttrTok->startsSequence(tok::r_paren, tok::identifier)) 430 return false; 431 if (AttrTok->is(tok::ellipsis)) 432 return true; 433 AttrTok = AttrTok->Next; 434 } 435 return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square); 436 } 437 438 bool parseSquare() { 439 if (!CurrentToken) 440 return false; 441 442 // A '[' could be an index subscript (after an identifier or after 443 // ')' or ']'), it could be the start of an Objective-C method 444 // expression, it could the start of an Objective-C array literal, 445 // or it could be a C++ attribute specifier [[foo::bar]]. 446 FormatToken *Left = CurrentToken->Previous; 447 Left->ParentBracket = Contexts.back().ContextKind; 448 FormatToken *Parent = Left->getPreviousNonComment(); 449 450 // Cases where '>' is followed by '['. 451 // In C++, this can happen either in array of templates (foo<int>[10]) 452 // or when array is a nested template type (unique_ptr<type1<type2>[]>). 453 bool CppArrayTemplates = 454 Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) && 455 (Contexts.back().CanBeExpression || Contexts.back().IsExpression || 456 Contexts.back().InTemplateArgument); 457 458 bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) || 459 Contexts.back().InCpp11AttributeSpecifier; 460 461 // Treat C# Attributes [STAThread] much like C++ attributes [[...]]. 462 bool IsCSharp11AttributeSpecifier = 463 isCSharpAttributeSpecifier(*Left) || 464 Contexts.back().InCSharpAttributeSpecifier; 465 466 bool InsideInlineASM = Line.startsWith(tok::kw_asm); 467 bool IsCppStructuredBinding = Left->isCppStructuredBinding(Style); 468 bool StartsObjCMethodExpr = 469 !IsCppStructuredBinding && !InsideInlineASM && !CppArrayTemplates && 470 Style.isCpp() && !IsCpp11AttributeSpecifier && 471 Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) && 472 !CurrentToken->isOneOf(tok::l_brace, tok::r_square) && 473 (!Parent || 474 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 475 tok::kw_return, tok::kw_throw) || 476 Parent->isUnaryOperator() || 477 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 478 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) || 479 (getBinOpPrecedence(Parent->Tok.getKind(), true, true) > 480 prec::Unknown)); 481 bool ColonFound = false; 482 483 unsigned BindingIncrease = 1; 484 if (IsCppStructuredBinding) { 485 Left->Type = TT_StructuredBindingLSquare; 486 } else if (Left->is(TT_Unknown)) { 487 if (StartsObjCMethodExpr) { 488 Left->Type = TT_ObjCMethodExpr; 489 } else if (IsCpp11AttributeSpecifier) { 490 Left->Type = TT_AttributeSquare; 491 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent && 492 Contexts.back().ContextKind == tok::l_brace && 493 Parent->isOneOf(tok::l_brace, tok::comma)) { 494 Left->Type = TT_JsComputedPropertyName; 495 } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace && 496 Parent && Parent->isOneOf(tok::l_brace, tok::comma)) { 497 Left->Type = TT_DesignatedInitializerLSquare; 498 } else if (IsCSharp11AttributeSpecifier) { 499 Left->Type = TT_AttributeSquare; 500 } else if (CurrentToken->is(tok::r_square) && Parent && 501 Parent->is(TT_TemplateCloser)) { 502 Left->Type = TT_ArraySubscriptLSquare; 503 } else if (Style.Language == FormatStyle::LK_Proto || 504 Style.Language == FormatStyle::LK_TextProto) { 505 // Square braces in LK_Proto can either be message field attributes: 506 // 507 // optional Aaa aaa = 1 [ 508 // (aaa) = aaa 509 // ]; 510 // 511 // extensions 123 [ 512 // (aaa) = aaa 513 // ]; 514 // 515 // or text proto extensions (in options): 516 // 517 // option (Aaa.options) = { 518 // [type.type/type] { 519 // key: value 520 // } 521 // } 522 // 523 // or repeated fields (in options): 524 // 525 // option (Aaa.options) = { 526 // keys: [ 1, 2, 3 ] 527 // } 528 // 529 // In the first and the third case we want to spread the contents inside 530 // the square braces; in the second we want to keep them inline. 531 Left->Type = TT_ArrayInitializerLSquare; 532 if (!Left->endsSequence(tok::l_square, tok::numeric_constant, 533 tok::equal) && 534 !Left->endsSequence(tok::l_square, tok::numeric_constant, 535 tok::identifier) && 536 !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) { 537 Left->Type = TT_ProtoExtensionLSquare; 538 BindingIncrease = 10; 539 } 540 } else if (!CppArrayTemplates && Parent && 541 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at, 542 tok::comma, tok::l_paren, tok::l_square, 543 tok::question, tok::colon, tok::kw_return, 544 // Should only be relevant to JavaScript: 545 tok::kw_default)) { 546 Left->Type = TT_ArrayInitializerLSquare; 547 } else { 548 BindingIncrease = 10; 549 Left->Type = TT_ArraySubscriptLSquare; 550 } 551 } 552 553 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease); 554 Contexts.back().IsExpression = true; 555 if (Style.Language == FormatStyle::LK_JavaScript && Parent && 556 Parent->is(TT_JsTypeColon)) 557 Contexts.back().IsExpression = false; 558 559 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr; 560 Contexts.back().InCpp11AttributeSpecifier = IsCpp11AttributeSpecifier; 561 Contexts.back().InCSharpAttributeSpecifier = IsCSharp11AttributeSpecifier; 562 563 while (CurrentToken) { 564 if (CurrentToken->is(tok::r_square)) { 565 if (IsCpp11AttributeSpecifier) 566 CurrentToken->Type = TT_AttributeSquare; 567 if (IsCSharp11AttributeSpecifier) 568 CurrentToken->Type = TT_AttributeSquare; 569 else if (((CurrentToken->Next && 570 CurrentToken->Next->is(tok::l_paren)) || 571 (CurrentToken->Previous && 572 CurrentToken->Previous->Previous == Left)) && 573 Left->is(TT_ObjCMethodExpr)) { 574 // An ObjC method call is rarely followed by an open parenthesis. It 575 // also can't be composed of just one token, unless it's a macro that 576 // will be expanded to more tokens. 577 // FIXME: Do we incorrectly label ":" with this? 578 StartsObjCMethodExpr = false; 579 Left->Type = TT_Unknown; 580 } 581 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) { 582 CurrentToken->Type = TT_ObjCMethodExpr; 583 // If we haven't seen a colon yet, make sure the last identifier 584 // before the r_square is tagged as a selector name component. 585 if (!ColonFound && CurrentToken->Previous && 586 CurrentToken->Previous->is(TT_Unknown) && 587 canBeObjCSelectorComponent(*CurrentToken->Previous)) 588 CurrentToken->Previous->Type = TT_SelectorName; 589 // determineStarAmpUsage() thinks that '*' '[' is allocating an 590 // array of pointers, but if '[' starts a selector then '*' is a 591 // binary operator. 592 if (Parent && Parent->is(TT_PointerOrReference)) 593 Parent->Type = TT_BinaryOperator; 594 } 595 // An arrow after an ObjC method expression is not a lambda arrow. 596 if (CurrentToken->Type == TT_ObjCMethodExpr && CurrentToken->Next && 597 CurrentToken->Next->is(TT_LambdaArrow)) 598 CurrentToken->Next->Type = TT_Unknown; 599 Left->MatchingParen = CurrentToken; 600 CurrentToken->MatchingParen = Left; 601 // FirstObjCSelectorName is set when a colon is found. This does 602 // not work, however, when the method has no parameters. 603 // Here, we set FirstObjCSelectorName when the end of the method call is 604 // reached, in case it was not set already. 605 if (!Contexts.back().FirstObjCSelectorName) { 606 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 607 if (Previous && Previous->is(TT_SelectorName)) { 608 Previous->ObjCSelectorNameParts = 1; 609 Contexts.back().FirstObjCSelectorName = Previous; 610 } 611 } else { 612 Left->ParameterCount = 613 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 614 } 615 if (Contexts.back().FirstObjCSelectorName) { 616 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 617 Contexts.back().LongestObjCSelectorName; 618 if (Left->BlockParameterCount > 1) 619 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0; 620 } 621 next(); 622 return true; 623 } 624 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace)) 625 return false; 626 if (CurrentToken->is(tok::colon)) { 627 if (IsCpp11AttributeSpecifier && 628 CurrentToken->endsSequence(tok::colon, tok::identifier, 629 tok::kw_using)) { 630 // Remember that this is a [[using ns: foo]] C++ attribute, so we 631 // don't add a space before the colon (unlike other colons). 632 CurrentToken->Type = TT_AttributeColon; 633 } else if (Left->isOneOf(TT_ArraySubscriptLSquare, 634 TT_DesignatedInitializerLSquare)) { 635 Left->Type = TT_ObjCMethodExpr; 636 StartsObjCMethodExpr = true; 637 Contexts.back().ColonIsObjCMethodExpr = true; 638 if (Parent && Parent->is(tok::r_paren)) 639 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 640 Parent->Type = TT_CastRParen; 641 } 642 ColonFound = true; 643 } 644 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) && 645 !ColonFound) 646 Left->Type = TT_ArrayInitializerLSquare; 647 FormatToken *Tok = CurrentToken; 648 if (!consumeToken()) 649 return false; 650 updateParameterCount(Left, Tok); 651 } 652 return false; 653 } 654 655 bool parseBrace() { 656 if (CurrentToken) { 657 FormatToken *Left = CurrentToken->Previous; 658 Left->ParentBracket = Contexts.back().ContextKind; 659 660 if (Contexts.back().CaretFound) 661 Left->Type = TT_ObjCBlockLBrace; 662 Contexts.back().CaretFound = false; 663 664 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1); 665 Contexts.back().ColonIsDictLiteral = true; 666 if (Left->BlockKind == BK_BracedInit) 667 Contexts.back().IsExpression = true; 668 if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 669 Left->Previous->is(TT_JsTypeColon)) 670 Contexts.back().IsExpression = false; 671 672 while (CurrentToken) { 673 if (CurrentToken->is(tok::r_brace)) { 674 Left->MatchingParen = CurrentToken; 675 CurrentToken->MatchingParen = Left; 676 next(); 677 return true; 678 } 679 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square)) 680 return false; 681 updateParameterCount(Left, CurrentToken); 682 if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) { 683 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 684 if (Previous->is(TT_JsTypeOptionalQuestion)) 685 Previous = Previous->getPreviousNonComment(); 686 if ((CurrentToken->is(tok::colon) && 687 (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) || 688 Style.Language == FormatStyle::LK_Proto || 689 Style.Language == FormatStyle::LK_TextProto) { 690 Left->Type = TT_DictLiteral; 691 if (Previous->Tok.getIdentifierInfo() || 692 Previous->is(tok::string_literal)) 693 Previous->Type = TT_SelectorName; 694 } 695 if (CurrentToken->is(tok::colon) || 696 Style.Language == FormatStyle::LK_JavaScript) 697 Left->Type = TT_DictLiteral; 698 } 699 if (CurrentToken->is(tok::comma) && 700 Style.Language == FormatStyle::LK_JavaScript) 701 Left->Type = TT_DictLiteral; 702 if (!consumeToken()) 703 return false; 704 } 705 } 706 return true; 707 } 708 709 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 710 // For ObjC methods, the number of parameters is calculated differently as 711 // method declarations have a different structure (the parameters are not 712 // inside a bracket scope). 713 if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block) 714 ++Left->BlockParameterCount; 715 if (Current->is(tok::comma)) { 716 ++Left->ParameterCount; 717 if (!Left->Role) 718 Left->Role.reset(new CommaSeparatedList(Style)); 719 Left->Role->CommaFound(Current); 720 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 721 Left->ParameterCount = 1; 722 } 723 } 724 725 bool parseConditional() { 726 while (CurrentToken) { 727 if (CurrentToken->is(tok::colon)) { 728 CurrentToken->Type = TT_ConditionalExpr; 729 next(); 730 return true; 731 } 732 if (!consumeToken()) 733 return false; 734 } 735 return false; 736 } 737 738 bool parseTemplateDeclaration() { 739 if (CurrentToken && CurrentToken->is(tok::less)) { 740 CurrentToken->Type = TT_TemplateOpener; 741 next(); 742 if (!parseAngle()) 743 return false; 744 if (CurrentToken) 745 CurrentToken->Previous->ClosesTemplateDeclaration = true; 746 return true; 747 } 748 return false; 749 } 750 751 bool consumeToken() { 752 FormatToken *Tok = CurrentToken; 753 next(); 754 switch (Tok->Tok.getKind()) { 755 case tok::plus: 756 case tok::minus: 757 if (!Tok->Previous && Line.MustBeDeclaration) 758 Tok->Type = TT_ObjCMethodSpecifier; 759 break; 760 case tok::colon: 761 if (!Tok->Previous) 762 return false; 763 // Colons from ?: are handled in parseConditional(). 764 if (Style.Language == FormatStyle::LK_JavaScript) { 765 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop 766 (Contexts.size() == 1 && // switch/case labels 767 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) || 768 Contexts.back().ContextKind == tok::l_paren || // function params 769 Contexts.back().ContextKind == tok::l_square || // array type 770 (!Contexts.back().IsExpression && 771 Contexts.back().ContextKind == tok::l_brace) || // object type 772 (Contexts.size() == 1 && 773 Line.MustBeDeclaration)) { // method/property declaration 774 Contexts.back().IsExpression = false; 775 Tok->Type = TT_JsTypeColon; 776 break; 777 } 778 } 779 if (Contexts.back().ColonIsDictLiteral || 780 Style.Language == FormatStyle::LK_Proto || 781 Style.Language == FormatStyle::LK_TextProto) { 782 Tok->Type = TT_DictLiteral; 783 if (Style.Language == FormatStyle::LK_TextProto) { 784 if (FormatToken *Previous = Tok->getPreviousNonComment()) 785 Previous->Type = TT_SelectorName; 786 } 787 } else if (Contexts.back().ColonIsObjCMethodExpr || 788 Line.startsWith(TT_ObjCMethodSpecifier)) { 789 Tok->Type = TT_ObjCMethodExpr; 790 const FormatToken *BeforePrevious = Tok->Previous->Previous; 791 // Ensure we tag all identifiers in method declarations as 792 // TT_SelectorName. 793 bool UnknownIdentifierInMethodDeclaration = 794 Line.startsWith(TT_ObjCMethodSpecifier) && 795 Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown); 796 if (!BeforePrevious || 797 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 798 !(BeforePrevious->is(TT_CastRParen) || 799 (BeforePrevious->is(TT_ObjCMethodExpr) && 800 BeforePrevious->is(tok::colon))) || 801 BeforePrevious->is(tok::r_square) || 802 Contexts.back().LongestObjCSelectorName == 0 || 803 UnknownIdentifierInMethodDeclaration) { 804 Tok->Previous->Type = TT_SelectorName; 805 if (!Contexts.back().FirstObjCSelectorName) 806 Contexts.back().FirstObjCSelectorName = Tok->Previous; 807 else if (Tok->Previous->ColumnWidth > 808 Contexts.back().LongestObjCSelectorName) 809 Contexts.back().LongestObjCSelectorName = 810 Tok->Previous->ColumnWidth; 811 Tok->Previous->ParameterIndex = 812 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 813 ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 814 } 815 } else if (Contexts.back().ColonIsForRangeExpr) { 816 Tok->Type = TT_RangeBasedForLoopColon; 817 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 818 Tok->Type = TT_BitFieldColon; 819 } else if (Contexts.size() == 1 && 820 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) { 821 if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren, 822 tok::kw_noexcept)) 823 Tok->Type = TT_CtorInitializerColon; 824 else 825 Tok->Type = TT_InheritanceColon; 826 } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next && 827 (Tok->Next->isOneOf(tok::r_paren, tok::comma) || 828 (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next && 829 Tok->Next->Next->is(tok::colon)))) { 830 // This handles a special macro in ObjC code where selectors including 831 // the colon are passed as macro arguments. 832 Tok->Type = TT_ObjCMethodExpr; 833 } else if (Contexts.back().ContextKind == tok::l_paren) { 834 Tok->Type = TT_InlineASMColon; 835 } 836 break; 837 case tok::pipe: 838 case tok::amp: 839 // | and & in declarations/type expressions represent union and 840 // intersection types, respectively. 841 if (Style.Language == FormatStyle::LK_JavaScript && 842 !Contexts.back().IsExpression) 843 Tok->Type = TT_JsTypeOperator; 844 break; 845 case tok::kw_if: 846 case tok::kw_while: 847 assert(!Line.startsWith(tok::hash)); 848 if (Tok->is(tok::kw_if) && CurrentToken && 849 CurrentToken->isOneOf(tok::kw_constexpr, tok::identifier)) 850 next(); 851 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 852 next(); 853 if (!parseParens(/*LookForDecls=*/true)) 854 return false; 855 } 856 break; 857 case tok::kw_for: 858 if (Style.Language == FormatStyle::LK_JavaScript) { 859 // x.for and {for: ...} 860 if ((Tok->Previous && Tok->Previous->is(tok::period)) || 861 (Tok->Next && Tok->Next->is(tok::colon))) 862 break; 863 // JS' for await ( ... 864 if (CurrentToken && CurrentToken->is(Keywords.kw_await)) 865 next(); 866 } 867 Contexts.back().ColonIsForRangeExpr = true; 868 next(); 869 if (!parseParens()) 870 return false; 871 break; 872 case tok::l_paren: 873 // When faced with 'operator()()', the kw_operator handler incorrectly 874 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make 875 // the first two parens OverloadedOperators and the second l_paren an 876 // OverloadedOperatorLParen. 877 if (Tok->Previous && Tok->Previous->is(tok::r_paren) && 878 Tok->Previous->MatchingParen && 879 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) { 880 Tok->Previous->Type = TT_OverloadedOperator; 881 Tok->Previous->MatchingParen->Type = TT_OverloadedOperator; 882 Tok->Type = TT_OverloadedOperatorLParen; 883 } 884 885 if (!parseParens()) 886 return false; 887 if (Line.MustBeDeclaration && Contexts.size() == 1 && 888 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) && 889 (!Tok->Previous || 890 !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute, 891 TT_LeadingJavaAnnotation))) 892 Line.MightBeFunctionDecl = true; 893 break; 894 case tok::l_square: 895 if (!parseSquare()) 896 return false; 897 break; 898 case tok::l_brace: 899 if (Style.Language == FormatStyle::LK_TextProto) { 900 FormatToken *Previous = Tok->getPreviousNonComment(); 901 if (Previous && Previous->Type != TT_DictLiteral) 902 Previous->Type = TT_SelectorName; 903 } 904 if (!parseBrace()) 905 return false; 906 break; 907 case tok::less: 908 if (parseAngle()) { 909 Tok->Type = TT_TemplateOpener; 910 // In TT_Proto, we must distignuish between: 911 // map<key, value> 912 // msg < item: data > 913 // msg: < item: data > 914 // In TT_TextProto, map<key, value> does not occur. 915 if (Style.Language == FormatStyle::LK_TextProto || 916 (Style.Language == FormatStyle::LK_Proto && Tok->Previous && 917 Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) { 918 Tok->Type = TT_DictLiteral; 919 FormatToken *Previous = Tok->getPreviousNonComment(); 920 if (Previous && Previous->Type != TT_DictLiteral) 921 Previous->Type = TT_SelectorName; 922 } 923 } else { 924 Tok->Type = TT_BinaryOperator; 925 NonTemplateLess.insert(Tok); 926 CurrentToken = Tok; 927 next(); 928 } 929 break; 930 case tok::r_paren: 931 case tok::r_square: 932 return false; 933 case tok::r_brace: 934 // Lines can start with '}'. 935 if (Tok->Previous) 936 return false; 937 break; 938 case tok::greater: 939 if (Style.Language != FormatStyle::LK_TextProto) 940 Tok->Type = TT_BinaryOperator; 941 if (Tok->Previous && Tok->Previous->is(TT_TemplateCloser)) 942 Tok->SpacesRequiredBefore = 1; 943 break; 944 case tok::kw_operator: 945 if (Style.Language == FormatStyle::LK_TextProto || 946 Style.Language == FormatStyle::LK_Proto) 947 break; 948 while (CurrentToken && 949 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 950 if (CurrentToken->isOneOf(tok::star, tok::amp)) 951 CurrentToken->Type = TT_PointerOrReference; 952 consumeToken(); 953 if (CurrentToken && 954 CurrentToken->Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator, 955 tok::comma)) 956 CurrentToken->Previous->Type = TT_OverloadedOperator; 957 } 958 if (CurrentToken) { 959 CurrentToken->Type = TT_OverloadedOperatorLParen; 960 if (CurrentToken->Previous->is(TT_BinaryOperator)) 961 CurrentToken->Previous->Type = TT_OverloadedOperator; 962 } 963 break; 964 case tok::question: 965 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next && 966 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren, 967 tok::r_brace)) { 968 // Question marks before semicolons, colons, etc. indicate optional 969 // types (fields, parameters), e.g. 970 // function(x?: string, y?) {...} 971 // class X { y?; } 972 Tok->Type = TT_JsTypeOptionalQuestion; 973 break; 974 } 975 // Declarations cannot be conditional expressions, this can only be part 976 // of a type declaration. 977 if (Line.MustBeDeclaration && !Contexts.back().IsExpression && 978 Style.Language == FormatStyle::LK_JavaScript) 979 break; 980 parseConditional(); 981 break; 982 case tok::kw_template: 983 parseTemplateDeclaration(); 984 break; 985 case tok::comma: 986 if (Contexts.back().InCtorInitializer) 987 Tok->Type = TT_CtorInitializerComma; 988 else if (Contexts.back().InInheritanceList) 989 Tok->Type = TT_InheritanceComma; 990 else if (Contexts.back().FirstStartOfName && 991 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) { 992 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 993 Line.IsMultiVariableDeclStmt = true; 994 } 995 if (Contexts.back().IsForEachMacro) 996 Contexts.back().IsExpression = true; 997 break; 998 case tok::identifier: 999 if (Tok->isOneOf(Keywords.kw___has_include, 1000 Keywords.kw___has_include_next)) { 1001 parseHasInclude(); 1002 } 1003 break; 1004 default: 1005 break; 1006 } 1007 return true; 1008 } 1009 1010 void parseIncludeDirective() { 1011 if (CurrentToken && CurrentToken->is(tok::less)) { 1012 next(); 1013 while (CurrentToken) { 1014 // Mark tokens up to the trailing line comments as implicit string 1015 // literals. 1016 if (CurrentToken->isNot(tok::comment) && 1017 !CurrentToken->TokenText.startswith("//")) 1018 CurrentToken->Type = TT_ImplicitStringLiteral; 1019 next(); 1020 } 1021 } 1022 } 1023 1024 void parseWarningOrError() { 1025 next(); 1026 // We still want to format the whitespace left of the first token of the 1027 // warning or error. 1028 next(); 1029 while (CurrentToken) { 1030 CurrentToken->Type = TT_ImplicitStringLiteral; 1031 next(); 1032 } 1033 } 1034 1035 void parsePragma() { 1036 next(); // Consume "pragma". 1037 if (CurrentToken && 1038 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) { 1039 bool IsMark = CurrentToken->is(Keywords.kw_mark); 1040 next(); // Consume "mark". 1041 next(); // Consume first token (so we fix leading whitespace). 1042 while (CurrentToken) { 1043 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator)) 1044 CurrentToken->Type = TT_ImplicitStringLiteral; 1045 next(); 1046 } 1047 } 1048 } 1049 1050 void parseHasInclude() { 1051 if (!CurrentToken || !CurrentToken->is(tok::l_paren)) 1052 return; 1053 next(); // '(' 1054 parseIncludeDirective(); 1055 next(); // ')' 1056 } 1057 1058 LineType parsePreprocessorDirective() { 1059 bool IsFirstToken = CurrentToken->IsFirst; 1060 LineType Type = LT_PreprocessorDirective; 1061 next(); 1062 if (!CurrentToken) 1063 return Type; 1064 1065 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) { 1066 // JavaScript files can contain shebang lines of the form: 1067 // #!/usr/bin/env node 1068 // Treat these like C++ #include directives. 1069 while (CurrentToken) { 1070 // Tokens cannot be comments here. 1071 CurrentToken->Type = TT_ImplicitStringLiteral; 1072 next(); 1073 } 1074 return LT_ImportStatement; 1075 } 1076 1077 if (CurrentToken->Tok.is(tok::numeric_constant)) { 1078 CurrentToken->SpacesRequiredBefore = 1; 1079 return Type; 1080 } 1081 // Hashes in the middle of a line can lead to any strange token 1082 // sequence. 1083 if (!CurrentToken->Tok.getIdentifierInfo()) 1084 return Type; 1085 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 1086 case tok::pp_include: 1087 case tok::pp_include_next: 1088 case tok::pp_import: 1089 next(); 1090 parseIncludeDirective(); 1091 Type = LT_ImportStatement; 1092 break; 1093 case tok::pp_error: 1094 case tok::pp_warning: 1095 parseWarningOrError(); 1096 break; 1097 case tok::pp_pragma: 1098 parsePragma(); 1099 break; 1100 case tok::pp_if: 1101 case tok::pp_elif: 1102 Contexts.back().IsExpression = true; 1103 next(); 1104 parseLine(); 1105 break; 1106 default: 1107 break; 1108 } 1109 while (CurrentToken) { 1110 FormatToken *Tok = CurrentToken; 1111 next(); 1112 if (Tok->is(tok::l_paren)) 1113 parseParens(); 1114 else if (Tok->isOneOf(Keywords.kw___has_include, 1115 Keywords.kw___has_include_next)) 1116 parseHasInclude(); 1117 } 1118 return Type; 1119 } 1120 1121 public: 1122 LineType parseLine() { 1123 if (!CurrentToken) 1124 return LT_Invalid; 1125 NonTemplateLess.clear(); 1126 if (CurrentToken->is(tok::hash)) 1127 return parsePreprocessorDirective(); 1128 1129 // Directly allow to 'import <string-literal>' to support protocol buffer 1130 // definitions (github.com/google/protobuf) or missing "#" (either way we 1131 // should not break the line). 1132 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 1133 if ((Style.Language == FormatStyle::LK_Java && 1134 CurrentToken->is(Keywords.kw_package)) || 1135 (Info && Info->getPPKeywordID() == tok::pp_import && 1136 CurrentToken->Next && 1137 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier, 1138 tok::kw_static))) { 1139 next(); 1140 parseIncludeDirective(); 1141 return LT_ImportStatement; 1142 } 1143 1144 // If this line starts and ends in '<' and '>', respectively, it is likely 1145 // part of "#define <a/b.h>". 1146 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 1147 parseIncludeDirective(); 1148 return LT_ImportStatement; 1149 } 1150 1151 // In .proto files, top-level options and package statements are very 1152 // similar to import statements and should not be line-wrapped. 1153 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 && 1154 CurrentToken->isOneOf(Keywords.kw_option, Keywords.kw_package)) { 1155 next(); 1156 if (CurrentToken && CurrentToken->is(tok::identifier)) { 1157 while (CurrentToken) 1158 next(); 1159 return LT_ImportStatement; 1160 } 1161 } 1162 1163 bool KeywordVirtualFound = false; 1164 bool ImportStatement = false; 1165 1166 // import {...} from '...'; 1167 if (Style.Language == FormatStyle::LK_JavaScript && 1168 CurrentToken->is(Keywords.kw_import)) 1169 ImportStatement = true; 1170 1171 while (CurrentToken) { 1172 if (CurrentToken->is(tok::kw_virtual)) 1173 KeywordVirtualFound = true; 1174 if (Style.Language == FormatStyle::LK_JavaScript) { 1175 // export {...} from '...'; 1176 // An export followed by "from 'some string';" is a re-export from 1177 // another module identified by a URI and is treated as a 1178 // LT_ImportStatement (i.e. prevent wraps on it for long URIs). 1179 // Just "export {...};" or "export class ..." should not be treated as 1180 // an import in this sense. 1181 if (Line.First->is(tok::kw_export) && 1182 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next && 1183 CurrentToken->Next->isStringLiteral()) 1184 ImportStatement = true; 1185 if (isClosureImportStatement(*CurrentToken)) 1186 ImportStatement = true; 1187 } 1188 if (!consumeToken()) 1189 return LT_Invalid; 1190 } 1191 if (KeywordVirtualFound) 1192 return LT_VirtualFunctionDecl; 1193 if (ImportStatement) 1194 return LT_ImportStatement; 1195 1196 if (Line.startsWith(TT_ObjCMethodSpecifier)) { 1197 if (Contexts.back().FirstObjCSelectorName) 1198 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 1199 Contexts.back().LongestObjCSelectorName; 1200 return LT_ObjCMethodDecl; 1201 } 1202 1203 return LT_Other; 1204 } 1205 1206 private: 1207 bool isClosureImportStatement(const FormatToken &Tok) { 1208 // FIXME: Closure-library specific stuff should not be hard-coded but be 1209 // configurable. 1210 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) && 1211 Tok.Next->Next && 1212 (Tok.Next->Next->TokenText == "module" || 1213 Tok.Next->Next->TokenText == "provide" || 1214 Tok.Next->Next->TokenText == "require" || 1215 Tok.Next->Next->TokenText == "requireType" || 1216 Tok.Next->Next->TokenText == "forwardDeclare") && 1217 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren); 1218 } 1219 1220 void resetTokenMetadata(FormatToken *Token) { 1221 if (!Token) 1222 return; 1223 1224 // Reset token type in case we have already looked at it and then 1225 // recovered from an error (e.g. failure to find the matching >). 1226 if (!CurrentToken->isOneOf( 1227 TT_LambdaLSquare, TT_LambdaLBrace, TT_ForEachMacro, 1228 TT_TypenameMacro, TT_FunctionLBrace, TT_ImplicitStringLiteral, 1229 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow, TT_NamespaceMacro, 1230 TT_OverloadedOperator, TT_RegexLiteral, TT_TemplateString, 1231 TT_ObjCStringLiteral)) 1232 CurrentToken->Type = TT_Unknown; 1233 CurrentToken->Role.reset(); 1234 CurrentToken->MatchingParen = nullptr; 1235 CurrentToken->FakeLParens.clear(); 1236 CurrentToken->FakeRParens = 0; 1237 } 1238 1239 void next() { 1240 if (CurrentToken) { 1241 CurrentToken->NestingLevel = Contexts.size() - 1; 1242 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 1243 modifyContext(*CurrentToken); 1244 determineTokenType(*CurrentToken); 1245 CurrentToken = CurrentToken->Next; 1246 } 1247 1248 resetTokenMetadata(CurrentToken); 1249 } 1250 1251 /// A struct to hold information valid in a specific context, e.g. 1252 /// a pair of parenthesis. 1253 struct Context { 1254 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 1255 bool IsExpression) 1256 : ContextKind(ContextKind), BindingStrength(BindingStrength), 1257 IsExpression(IsExpression) {} 1258 1259 tok::TokenKind ContextKind; 1260 unsigned BindingStrength; 1261 bool IsExpression; 1262 unsigned LongestObjCSelectorName = 0; 1263 bool ColonIsForRangeExpr = false; 1264 bool ColonIsDictLiteral = false; 1265 bool ColonIsObjCMethodExpr = false; 1266 FormatToken *FirstObjCSelectorName = nullptr; 1267 FormatToken *FirstStartOfName = nullptr; 1268 bool CanBeExpression = true; 1269 bool InTemplateArgument = false; 1270 bool InCtorInitializer = false; 1271 bool InInheritanceList = false; 1272 bool CaretFound = false; 1273 bool IsForEachMacro = false; 1274 bool InCpp11AttributeSpecifier = false; 1275 bool InCSharpAttributeSpecifier = false; 1276 }; 1277 1278 /// Puts a new \c Context onto the stack \c Contexts for the lifetime 1279 /// of each instance. 1280 struct ScopedContextCreator { 1281 AnnotatingParser &P; 1282 1283 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 1284 unsigned Increase) 1285 : P(P) { 1286 P.Contexts.push_back(Context(ContextKind, 1287 P.Contexts.back().BindingStrength + Increase, 1288 P.Contexts.back().IsExpression)); 1289 } 1290 1291 ~ScopedContextCreator() { P.Contexts.pop_back(); } 1292 }; 1293 1294 void modifyContext(const FormatToken &Current) { 1295 if (Current.getPrecedence() == prec::Assignment && 1296 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) && 1297 // Type aliases use `type X = ...;` in TypeScript and can be exported 1298 // using `export type ...`. 1299 !(Style.Language == FormatStyle::LK_JavaScript && 1300 (Line.startsWith(Keywords.kw_type, tok::identifier) || 1301 Line.startsWith(tok::kw_export, Keywords.kw_type, 1302 tok::identifier))) && 1303 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 1304 Contexts.back().IsExpression = true; 1305 if (!Line.startsWith(TT_UnaryOperator)) { 1306 for (FormatToken *Previous = Current.Previous; 1307 Previous && Previous->Previous && 1308 !Previous->Previous->isOneOf(tok::comma, tok::semi); 1309 Previous = Previous->Previous) { 1310 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 1311 Previous = Previous->MatchingParen; 1312 if (!Previous) 1313 break; 1314 } 1315 if (Previous->opensScope()) 1316 break; 1317 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) && 1318 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) && 1319 Previous->Previous && Previous->Previous->isNot(tok::equal)) 1320 Previous->Type = TT_PointerOrReference; 1321 } 1322 } 1323 } else if (Current.is(tok::lessless) && 1324 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) { 1325 Contexts.back().IsExpression = true; 1326 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 1327 Contexts.back().IsExpression = true; 1328 } else if (Current.is(TT_TrailingReturnArrow)) { 1329 Contexts.back().IsExpression = false; 1330 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) { 1331 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java; 1332 } else if (Current.Previous && 1333 Current.Previous->is(TT_CtorInitializerColon)) { 1334 Contexts.back().IsExpression = true; 1335 Contexts.back().InCtorInitializer = true; 1336 } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) { 1337 Contexts.back().InInheritanceList = true; 1338 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 1339 for (FormatToken *Previous = Current.Previous; 1340 Previous && Previous->isOneOf(tok::star, tok::amp); 1341 Previous = Previous->Previous) 1342 Previous->Type = TT_PointerOrReference; 1343 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer) 1344 Contexts.back().IsExpression = false; 1345 } else if (Current.is(tok::kw_new)) { 1346 Contexts.back().CanBeExpression = false; 1347 } else if (Current.isOneOf(tok::semi, tok::exclaim)) { 1348 // This should be the condition or increment in a for-loop. 1349 Contexts.back().IsExpression = true; 1350 } 1351 } 1352 1353 void determineTokenType(FormatToken &Current) { 1354 if (!Current.is(TT_Unknown)) 1355 // The token type is already known. 1356 return; 1357 1358 if (Style.Language == FormatStyle::LK_JavaScript) { 1359 if (Current.is(tok::exclaim)) { 1360 if (Current.Previous && 1361 (Current.Previous->isOneOf(tok::identifier, tok::kw_namespace, 1362 tok::r_paren, tok::r_square, 1363 tok::r_brace) || 1364 Current.Previous->Tok.isLiteral())) { 1365 Current.Type = TT_JsNonNullAssertion; 1366 return; 1367 } 1368 if (Current.Next && 1369 Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) { 1370 Current.Type = TT_JsNonNullAssertion; 1371 return; 1372 } 1373 } 1374 } 1375 1376 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1377 // function declaration have been found. In this case, 'Current' is a 1378 // trailing token of this declaration and thus cannot be a name. 1379 if (Current.is(Keywords.kw_instanceof)) { 1380 Current.Type = TT_BinaryOperator; 1381 } else if (isStartOfName(Current) && 1382 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 1383 Contexts.back().FirstStartOfName = &Current; 1384 Current.Type = TT_StartOfName; 1385 } else if (Current.is(tok::semi)) { 1386 // Reset FirstStartOfName after finding a semicolon so that a for loop 1387 // with multiple increment statements is not confused with a for loop 1388 // having multiple variable declarations. 1389 Contexts.back().FirstStartOfName = nullptr; 1390 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) { 1391 AutoFound = true; 1392 } else if (Current.is(tok::arrow) && 1393 Style.Language == FormatStyle::LK_Java) { 1394 Current.Type = TT_LambdaArrow; 1395 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration && 1396 Current.NestingLevel == 0 && 1397 !Current.Previous->is(tok::kw_operator)) { 1398 // not auto operator->() -> xxx; 1399 Current.Type = TT_TrailingReturnArrow; 1400 TrailingReturnFound = true; 1401 } else if (Current.is(tok::star) || 1402 (Current.isOneOf(tok::amp, tok::ampamp) && 1403 (Current.NestingLevel != 0 || !Line.MightBeFunctionDecl || 1404 TrailingReturnFound))) { 1405 Current.Type = determineStarAmpUsage(Current, 1406 Contexts.back().CanBeExpression && 1407 Contexts.back().IsExpression, 1408 Contexts.back().InTemplateArgument); 1409 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 1410 Current.Type = determinePlusMinusCaretUsage(Current); 1411 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret)) 1412 Contexts.back().CaretFound = true; 1413 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 1414 Current.Type = determineIncrementUsage(Current); 1415 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 1416 Current.Type = TT_UnaryOperator; 1417 } else if (Current.is(tok::question)) { 1418 if (Style.Language == FormatStyle::LK_JavaScript && 1419 Line.MustBeDeclaration && !Contexts.back().IsExpression) { 1420 // In JavaScript, `interface X { foo?(): bar; }` is an optional method 1421 // on the interface, not a ternary expression. 1422 Current.Type = TT_JsTypeOptionalQuestion; 1423 } else { 1424 Current.Type = TT_ConditionalExpr; 1425 } 1426 } else if (Current.isBinaryOperator() && 1427 !(Line.MightBeFunctionDecl && Current.NestingLevel == 0 && 1428 Current.isOneOf(tok::amp, tok::ampamp)) && 1429 (!Current.Previous || Current.Previous->isNot(tok::l_square)) && 1430 (!Current.is(tok::greater) && 1431 Style.Language != FormatStyle::LK_TextProto)) { 1432 Current.Type = TT_BinaryOperator; 1433 } else if (Current.is(tok::comment)) { 1434 if (Current.TokenText.startswith("/*")) { 1435 if (Current.TokenText.endswith("*/")) 1436 Current.Type = TT_BlockComment; 1437 else 1438 // The lexer has for some reason determined a comment here. But we 1439 // cannot really handle it, if it isn't properly terminated. 1440 Current.Tok.setKind(tok::unknown); 1441 } else { 1442 Current.Type = TT_LineComment; 1443 } 1444 } else if (Current.is(tok::r_paren)) { 1445 if (rParenEndsCast(Current)) 1446 Current.Type = TT_CastRParen; 1447 if (Current.MatchingParen && Current.Next && 1448 !Current.Next->isBinaryOperator() && 1449 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace, 1450 tok::comma, tok::period, tok::arrow, 1451 tok::coloncolon)) 1452 if (FormatToken *AfterParen = Current.MatchingParen->Next) { 1453 // Make sure this isn't the return type of an Obj-C block declaration 1454 if (AfterParen->Tok.isNot(tok::caret)) { 1455 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) 1456 if (BeforeParen->is(tok::identifier) && 1457 !BeforeParen->is(TT_TypenameMacro) && 1458 BeforeParen->TokenText == BeforeParen->TokenText.upper() && 1459 (!BeforeParen->Previous || 1460 BeforeParen->Previous->ClosesTemplateDeclaration)) 1461 Current.Type = TT_FunctionAnnotationRParen; 1462 } 1463 } 1464 } else if (Current.is(tok::at) && Current.Next && 1465 Style.Language != FormatStyle::LK_JavaScript && 1466 Style.Language != FormatStyle::LK_Java) { 1467 // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it 1468 // marks declarations and properties that need special formatting. 1469 switch (Current.Next->Tok.getObjCKeywordID()) { 1470 case tok::objc_interface: 1471 case tok::objc_implementation: 1472 case tok::objc_protocol: 1473 Current.Type = TT_ObjCDecl; 1474 break; 1475 case tok::objc_property: 1476 Current.Type = TT_ObjCProperty; 1477 break; 1478 default: 1479 break; 1480 } 1481 } else if (Current.is(tok::period)) { 1482 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 1483 if (PreviousNoComment && 1484 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 1485 Current.Type = TT_DesignatedInitializerPeriod; 1486 else if (Style.Language == FormatStyle::LK_Java && Current.Previous && 1487 Current.Previous->isOneOf(TT_JavaAnnotation, 1488 TT_LeadingJavaAnnotation)) { 1489 Current.Type = Current.Previous->Type; 1490 } 1491 } else if (canBeObjCSelectorComponent(Current) && 1492 // FIXME(bug 36976): ObjC return types shouldn't use 1493 // TT_CastRParen. 1494 Current.Previous && Current.Previous->is(TT_CastRParen) && 1495 Current.Previous->MatchingParen && 1496 Current.Previous->MatchingParen->Previous && 1497 Current.Previous->MatchingParen->Previous->is( 1498 TT_ObjCMethodSpecifier)) { 1499 // This is the first part of an Objective-C selector name. (If there's no 1500 // colon after this, this is the only place which annotates the identifier 1501 // as a selector.) 1502 Current.Type = TT_SelectorName; 1503 } else if (Current.isOneOf(tok::identifier, tok::kw_const, tok::amp, 1504 tok::ampamp) && 1505 Current.Previous && 1506 !Current.Previous->isOneOf(tok::equal, tok::at) && 1507 Line.MightBeFunctionDecl && !TrailingReturnFound && 1508 Contexts.size() == 1) { 1509 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1510 // function declaration have been found. 1511 Current.Type = TT_TrailingAnnotation; 1512 } else if ((Style.Language == FormatStyle::LK_Java || 1513 Style.Language == FormatStyle::LK_JavaScript) && 1514 Current.Previous) { 1515 if (Current.Previous->is(tok::at) && 1516 Current.isNot(Keywords.kw_interface)) { 1517 const FormatToken &AtToken = *Current.Previous; 1518 const FormatToken *Previous = AtToken.getPreviousNonComment(); 1519 if (!Previous || Previous->is(TT_LeadingJavaAnnotation)) 1520 Current.Type = TT_LeadingJavaAnnotation; 1521 else 1522 Current.Type = TT_JavaAnnotation; 1523 } else if (Current.Previous->is(tok::period) && 1524 Current.Previous->isOneOf(TT_JavaAnnotation, 1525 TT_LeadingJavaAnnotation)) { 1526 Current.Type = Current.Previous->Type; 1527 } 1528 } 1529 } 1530 1531 /// Take a guess at whether \p Tok starts a name of a function or 1532 /// variable declaration. 1533 /// 1534 /// This is a heuristic based on whether \p Tok is an identifier following 1535 /// something that is likely a type. 1536 bool isStartOfName(const FormatToken &Tok) { 1537 if (Tok.isNot(tok::identifier) || !Tok.Previous) 1538 return false; 1539 1540 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof, 1541 Keywords.kw_as)) 1542 return false; 1543 if (Style.Language == FormatStyle::LK_JavaScript && 1544 Tok.Previous->is(Keywords.kw_in)) 1545 return false; 1546 1547 // Skip "const" as it does not have an influence on whether this is a name. 1548 FormatToken *PreviousNotConst = Tok.getPreviousNonComment(); 1549 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 1550 PreviousNotConst = PreviousNotConst->getPreviousNonComment(); 1551 1552 if (!PreviousNotConst) 1553 return false; 1554 1555 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 1556 PreviousNotConst->Previous && 1557 PreviousNotConst->Previous->is(tok::hash); 1558 1559 if (PreviousNotConst->is(TT_TemplateCloser)) 1560 return PreviousNotConst && PreviousNotConst->MatchingParen && 1561 PreviousNotConst->MatchingParen->Previous && 1562 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) && 1563 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 1564 1565 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen && 1566 PreviousNotConst->MatchingParen->Previous && 1567 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype)) 1568 return true; 1569 1570 return (!IsPPKeyword && 1571 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) || 1572 PreviousNotConst->is(TT_PointerOrReference) || 1573 PreviousNotConst->isSimpleTypeSpecifier(); 1574 } 1575 1576 /// Determine whether ')' is ending a cast. 1577 bool rParenEndsCast(const FormatToken &Tok) { 1578 // C-style casts are only used in C++ and Java. 1579 if (!Style.isCpp() && Style.Language != FormatStyle::LK_Java) 1580 return false; 1581 1582 // Empty parens aren't casts and there are no casts at the end of the line. 1583 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen) 1584 return false; 1585 1586 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 1587 if (LeftOfParens) { 1588 // If there is a closing parenthesis left of the current parentheses, 1589 // look past it as these might be chained casts. 1590 if (LeftOfParens->is(tok::r_paren)) { 1591 if (!LeftOfParens->MatchingParen || 1592 !LeftOfParens->MatchingParen->Previous) 1593 return false; 1594 LeftOfParens = LeftOfParens->MatchingParen->Previous; 1595 } 1596 1597 // If there is an identifier (or with a few exceptions a keyword) right 1598 // before the parentheses, this is unlikely to be a cast. 1599 if (LeftOfParens->Tok.getIdentifierInfo() && 1600 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case, 1601 tok::kw_delete)) 1602 return false; 1603 1604 // Certain other tokens right before the parentheses are also signals that 1605 // this cannot be a cast. 1606 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator, 1607 TT_TemplateCloser, tok::ellipsis)) 1608 return false; 1609 } 1610 1611 if (Tok.Next->is(tok::question)) 1612 return false; 1613 1614 // As Java has no function types, a "(" after the ")" likely means that this 1615 // is a cast. 1616 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren)) 1617 return true; 1618 1619 // If a (non-string) literal follows, this is likely a cast. 1620 if (Tok.Next->isNot(tok::string_literal) && 1621 (Tok.Next->Tok.isLiteral() || 1622 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 1623 return true; 1624 1625 // Heuristically try to determine whether the parentheses contain a type. 1626 bool ParensAreType = 1627 !Tok.Previous || 1628 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) || 1629 Tok.Previous->isSimpleTypeSpecifier(); 1630 bool ParensCouldEndDecl = 1631 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater); 1632 if (ParensAreType && !ParensCouldEndDecl) 1633 return true; 1634 1635 // At this point, we heuristically assume that there are no casts at the 1636 // start of the line. We assume that we have found most cases where there 1637 // are by the logic above, e.g. "(void)x;". 1638 if (!LeftOfParens) 1639 return false; 1640 1641 // Certain token types inside the parentheses mean that this can't be a 1642 // cast. 1643 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok; 1644 Token = Token->Next) 1645 if (Token->is(TT_BinaryOperator)) 1646 return false; 1647 1648 // If the following token is an identifier or 'this', this is a cast. All 1649 // cases where this can be something else are handled above. 1650 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this)) 1651 return true; 1652 1653 if (!Tok.Next->Next) 1654 return false; 1655 1656 // If the next token after the parenthesis is a unary operator, assume 1657 // that this is cast, unless there are unexpected tokens inside the 1658 // parenthesis. 1659 bool NextIsUnary = 1660 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star); 1661 if (!NextIsUnary || Tok.Next->is(tok::plus) || 1662 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) 1663 return false; 1664 // Search for unexpected tokens. 1665 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen; 1666 Prev = Prev->Previous) { 1667 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon)) 1668 return false; 1669 } 1670 return true; 1671 } 1672 1673 /// Return the type of the given token assuming it is * or &. 1674 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 1675 bool InTemplateArgument) { 1676 if (Style.Language == FormatStyle::LK_JavaScript) 1677 return TT_BinaryOperator; 1678 1679 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1680 if (!PrevToken) 1681 return TT_UnaryOperator; 1682 1683 const FormatToken *NextToken = Tok.getNextNonComment(); 1684 if (!NextToken || 1685 NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_const) || 1686 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) 1687 return TT_PointerOrReference; 1688 1689 if (PrevToken->is(tok::coloncolon)) 1690 return TT_PointerOrReference; 1691 1692 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 1693 tok::comma, tok::semi, tok::kw_return, tok::colon, 1694 tok::equal, tok::kw_delete, tok::kw_sizeof, 1695 tok::kw_throw) || 1696 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr, 1697 TT_UnaryOperator, TT_CastRParen)) 1698 return TT_UnaryOperator; 1699 1700 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare)) 1701 return TT_PointerOrReference; 1702 if (NextToken->is(tok::kw_operator) && !IsExpression) 1703 return TT_PointerOrReference; 1704 if (NextToken->isOneOf(tok::comma, tok::semi)) 1705 return TT_PointerOrReference; 1706 1707 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen) { 1708 FormatToken *TokenBeforeMatchingParen = 1709 PrevToken->MatchingParen->getPreviousNonComment(); 1710 if (TokenBeforeMatchingParen && 1711 TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype, 1712 TT_TypenameMacro)) 1713 return TT_PointerOrReference; 1714 } 1715 1716 if (PrevToken->Tok.isLiteral() || 1717 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 1718 tok::kw_false, tok::r_brace) || 1719 NextToken->Tok.isLiteral() || 1720 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 1721 NextToken->isUnaryOperator() || 1722 // If we know we're in a template argument, there are no named 1723 // declarations. Thus, having an identifier on the right-hand side 1724 // indicates a binary operator. 1725 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 1726 return TT_BinaryOperator; 1727 1728 // "&&(" is quite unlikely to be two successive unary "&". 1729 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren)) 1730 return TT_BinaryOperator; 1731 1732 // This catches some cases where evaluation order is used as control flow: 1733 // aaa && aaa->f(); 1734 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 1735 if (NextNextToken && NextNextToken->is(tok::arrow)) 1736 return TT_BinaryOperator; 1737 1738 // It is very unlikely that we are going to find a pointer or reference type 1739 // definition on the RHS of an assignment. 1740 if (IsExpression && !Contexts.back().CaretFound) 1741 return TT_BinaryOperator; 1742 1743 return TT_PointerOrReference; 1744 } 1745 1746 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 1747 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1748 if (!PrevToken) 1749 return TT_UnaryOperator; 1750 1751 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator)) 1752 // This must be a sequence of leading unary operators. 1753 return TT_UnaryOperator; 1754 1755 // Use heuristics to recognize unary operators. 1756 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 1757 tok::question, tok::colon, tok::kw_return, 1758 tok::kw_case, tok::at, tok::l_brace)) 1759 return TT_UnaryOperator; 1760 1761 // There can't be two consecutive binary operators. 1762 if (PrevToken->is(TT_BinaryOperator)) 1763 return TT_UnaryOperator; 1764 1765 // Fall back to marking the token as binary operator. 1766 return TT_BinaryOperator; 1767 } 1768 1769 /// Determine whether ++/-- are pre- or post-increments/-decrements. 1770 TokenType determineIncrementUsage(const FormatToken &Tok) { 1771 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1772 if (!PrevToken || PrevToken->is(TT_CastRParen)) 1773 return TT_UnaryOperator; 1774 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 1775 return TT_TrailingUnaryOperator; 1776 1777 return TT_UnaryOperator; 1778 } 1779 1780 SmallVector<Context, 8> Contexts; 1781 1782 const FormatStyle &Style; 1783 AnnotatedLine &Line; 1784 FormatToken *CurrentToken; 1785 bool AutoFound; 1786 bool TrailingReturnFound; 1787 const AdditionalKeywords &Keywords; 1788 1789 // Set of "<" tokens that do not open a template parameter list. If parseAngle 1790 // determines that a specific token can't be a template opener, it will make 1791 // same decision irrespective of the decisions for tokens leading up to it. 1792 // Store this information to prevent this from causing exponential runtime. 1793 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess; 1794 }; 1795 1796 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1; 1797 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 1798 1799 /// Parses binary expressions by inserting fake parenthesis based on 1800 /// operator precedence. 1801 class ExpressionParser { 1802 public: 1803 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords, 1804 AnnotatedLine &Line) 1805 : Style(Style), Keywords(Keywords), Current(Line.First) {} 1806 1807 /// Parse expressions with the given operator precedence. 1808 void parse(int Precedence = 0) { 1809 // Skip 'return' and ObjC selector colons as they are not part of a binary 1810 // expression. 1811 while (Current && (Current->is(tok::kw_return) || 1812 (Current->is(tok::colon) && 1813 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) 1814 next(); 1815 1816 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1817 return; 1818 1819 // Conditional expressions need to be parsed separately for proper nesting. 1820 if (Precedence == prec::Conditional) { 1821 parseConditionalExpr(); 1822 return; 1823 } 1824 1825 // Parse unary operators, which all have a higher precedence than binary 1826 // operators. 1827 if (Precedence == PrecedenceUnaryOperator) { 1828 parseUnaryOperator(); 1829 return; 1830 } 1831 1832 FormatToken *Start = Current; 1833 FormatToken *LatestOperator = nullptr; 1834 unsigned OperatorIndex = 0; 1835 1836 while (Current) { 1837 // Consume operators with higher precedence. 1838 parse(Precedence + 1); 1839 1840 int CurrentPrecedence = getCurrentPrecedence(); 1841 1842 if (Current && Current->is(TT_SelectorName) && 1843 Precedence == CurrentPrecedence) { 1844 if (LatestOperator) 1845 addFakeParenthesis(Start, prec::Level(Precedence)); 1846 Start = Current; 1847 } 1848 1849 // At the end of the line or when an operator with higher precedence is 1850 // found, insert fake parenthesis and return. 1851 if (!Current || 1852 (Current->closesScope() && 1853 (Current->MatchingParen || Current->is(TT_TemplateString))) || 1854 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) || 1855 (CurrentPrecedence == prec::Conditional && 1856 Precedence == prec::Assignment && Current->is(tok::colon))) { 1857 break; 1858 } 1859 1860 // Consume scopes: (), [], <> and {} 1861 if (Current->opensScope()) { 1862 // In fragment of a JavaScript template string can look like '}..${' and 1863 // thus close a scope and open a new one at the same time. 1864 while (Current && (!Current->closesScope() || Current->opensScope())) { 1865 next(); 1866 parse(); 1867 } 1868 next(); 1869 } else { 1870 // Operator found. 1871 if (CurrentPrecedence == Precedence) { 1872 if (LatestOperator) 1873 LatestOperator->NextOperator = Current; 1874 LatestOperator = Current; 1875 Current->OperatorIndex = OperatorIndex; 1876 ++OperatorIndex; 1877 } 1878 next(/*SkipPastLeadingComments=*/Precedence > 0); 1879 } 1880 } 1881 1882 if (LatestOperator && (Current || Precedence > 0)) { 1883 // LatestOperator->LastOperator = true; 1884 if (Precedence == PrecedenceArrowAndPeriod) { 1885 // Call expressions don't have a binary operator precedence. 1886 addFakeParenthesis(Start, prec::Unknown); 1887 } else { 1888 addFakeParenthesis(Start, prec::Level(Precedence)); 1889 } 1890 } 1891 } 1892 1893 private: 1894 /// Gets the precedence (+1) of the given token for binary operators 1895 /// and other tokens that we treat like binary operators. 1896 int getCurrentPrecedence() { 1897 if (Current) { 1898 const FormatToken *NextNonComment = Current->getNextNonComment(); 1899 if (Current->is(TT_ConditionalExpr)) 1900 return prec::Conditional; 1901 if (NextNonComment && Current->is(TT_SelectorName) && 1902 (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) || 1903 ((Style.Language == FormatStyle::LK_Proto || 1904 Style.Language == FormatStyle::LK_TextProto) && 1905 NextNonComment->is(tok::less)))) 1906 return prec::Assignment; 1907 if (Current->is(TT_JsComputedPropertyName)) 1908 return prec::Assignment; 1909 if (Current->is(TT_LambdaArrow)) 1910 return prec::Comma; 1911 if (Current->is(TT_JsFatArrow)) 1912 return prec::Assignment; 1913 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) || 1914 (Current->is(tok::comment) && NextNonComment && 1915 NextNonComment->is(TT_SelectorName))) 1916 return 0; 1917 if (Current->is(TT_RangeBasedForLoopColon)) 1918 return prec::Comma; 1919 if ((Style.Language == FormatStyle::LK_Java || 1920 Style.Language == FormatStyle::LK_JavaScript) && 1921 Current->is(Keywords.kw_instanceof)) 1922 return prec::Relational; 1923 if (Style.Language == FormatStyle::LK_JavaScript && 1924 Current->isOneOf(Keywords.kw_in, Keywords.kw_as)) 1925 return prec::Relational; 1926 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma)) 1927 return Current->getPrecedence(); 1928 if (Current->isOneOf(tok::period, tok::arrow)) 1929 return PrecedenceArrowAndPeriod; 1930 if ((Style.Language == FormatStyle::LK_Java || 1931 Style.Language == FormatStyle::LK_JavaScript) && 1932 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements, 1933 Keywords.kw_throws)) 1934 return 0; 1935 } 1936 return -1; 1937 } 1938 1939 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1940 Start->FakeLParens.push_back(Precedence); 1941 if (Precedence > prec::Unknown) 1942 Start->StartsBinaryExpression = true; 1943 if (Current) { 1944 FormatToken *Previous = Current->Previous; 1945 while (Previous->is(tok::comment) && Previous->Previous) 1946 Previous = Previous->Previous; 1947 ++Previous->FakeRParens; 1948 if (Precedence > prec::Unknown) 1949 Previous->EndsBinaryExpression = true; 1950 } 1951 } 1952 1953 /// Parse unary operator expressions and surround them with fake 1954 /// parentheses if appropriate. 1955 void parseUnaryOperator() { 1956 llvm::SmallVector<FormatToken *, 2> Tokens; 1957 while (Current && Current->is(TT_UnaryOperator)) { 1958 Tokens.push_back(Current); 1959 next(); 1960 } 1961 parse(PrecedenceArrowAndPeriod); 1962 for (FormatToken *Token : llvm::reverse(Tokens)) 1963 // The actual precedence doesn't matter. 1964 addFakeParenthesis(Token, prec::Unknown); 1965 } 1966 1967 void parseConditionalExpr() { 1968 while (Current && Current->isTrailingComment()) { 1969 next(); 1970 } 1971 FormatToken *Start = Current; 1972 parse(prec::LogicalOr); 1973 if (!Current || !Current->is(tok::question)) 1974 return; 1975 next(); 1976 parse(prec::Assignment); 1977 if (!Current || Current->isNot(TT_ConditionalExpr)) 1978 return; 1979 next(); 1980 parse(prec::Assignment); 1981 addFakeParenthesis(Start, prec::Conditional); 1982 } 1983 1984 void next(bool SkipPastLeadingComments = true) { 1985 if (Current) 1986 Current = Current->Next; 1987 while (Current && 1988 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 1989 Current->isTrailingComment()) 1990 Current = Current->Next; 1991 } 1992 1993 const FormatStyle &Style; 1994 const AdditionalKeywords &Keywords; 1995 FormatToken *Current; 1996 }; 1997 1998 } // end anonymous namespace 1999 2000 void TokenAnnotator::setCommentLineLevels( 2001 SmallVectorImpl<AnnotatedLine *> &Lines) { 2002 const AnnotatedLine *NextNonCommentLine = nullptr; 2003 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 2004 E = Lines.rend(); 2005 I != E; ++I) { 2006 bool CommentLine = true; 2007 for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) { 2008 if (!Tok->is(tok::comment)) { 2009 CommentLine = false; 2010 break; 2011 } 2012 } 2013 2014 // If the comment is currently aligned with the line immediately following 2015 // it, that's probably intentional and we should keep it. 2016 if (NextNonCommentLine && CommentLine && 2017 NextNonCommentLine->First->NewlinesBefore <= 1 && 2018 NextNonCommentLine->First->OriginalColumn == 2019 (*I)->First->OriginalColumn) { 2020 // Align comments for preprocessor lines with the # in column 0 if 2021 // preprocessor lines are not indented. Otherwise, align with the next 2022 // line. 2023 (*I)->Level = 2024 (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash && 2025 (NextNonCommentLine->Type == LT_PreprocessorDirective || 2026 NextNonCommentLine->Type == LT_ImportStatement)) 2027 ? 0 2028 : NextNonCommentLine->Level; 2029 } else { 2030 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 2031 } 2032 2033 setCommentLineLevels((*I)->Children); 2034 } 2035 } 2036 2037 static unsigned maxNestingDepth(const AnnotatedLine &Line) { 2038 unsigned Result = 0; 2039 for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next) 2040 Result = std::max(Result, Tok->NestingLevel); 2041 return Result; 2042 } 2043 2044 void TokenAnnotator::annotate(AnnotatedLine &Line) { 2045 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 2046 E = Line.Children.end(); 2047 I != E; ++I) { 2048 annotate(**I); 2049 } 2050 AnnotatingParser Parser(Style, Line, Keywords); 2051 Line.Type = Parser.parseLine(); 2052 2053 // With very deep nesting, ExpressionParser uses lots of stack and the 2054 // formatting algorithm is very slow. We're not going to do a good job here 2055 // anyway - it's probably generated code being formatted by mistake. 2056 // Just skip the whole line. 2057 if (maxNestingDepth(Line) > 50) 2058 Line.Type = LT_Invalid; 2059 2060 if (Line.Type == LT_Invalid) 2061 return; 2062 2063 ExpressionParser ExprParser(Style, Keywords, Line); 2064 ExprParser.parse(); 2065 2066 if (Line.startsWith(TT_ObjCMethodSpecifier)) 2067 Line.Type = LT_ObjCMethodDecl; 2068 else if (Line.startsWith(TT_ObjCDecl)) 2069 Line.Type = LT_ObjCDecl; 2070 else if (Line.startsWith(TT_ObjCProperty)) 2071 Line.Type = LT_ObjCProperty; 2072 2073 Line.First->SpacesRequiredBefore = 1; 2074 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 2075 } 2076 2077 // This function heuristically determines whether 'Current' starts the name of a 2078 // function declaration. 2079 static bool isFunctionDeclarationName(const FormatToken &Current, 2080 const AnnotatedLine &Line) { 2081 auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * { 2082 for (; Next; Next = Next->Next) { 2083 if (Next->is(TT_OverloadedOperatorLParen)) 2084 return Next; 2085 if (Next->is(TT_OverloadedOperator)) 2086 continue; 2087 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) { 2088 // For 'new[]' and 'delete[]'. 2089 if (Next->Next && Next->Next->is(tok::l_square) && Next->Next->Next && 2090 Next->Next->Next->is(tok::r_square)) 2091 Next = Next->Next->Next; 2092 continue; 2093 } 2094 2095 break; 2096 } 2097 return nullptr; 2098 }; 2099 2100 // Find parentheses of parameter list. 2101 const FormatToken *Next = Current.Next; 2102 if (Current.is(tok::kw_operator)) { 2103 if (Current.Previous && Current.Previous->is(tok::coloncolon)) 2104 return false; 2105 Next = skipOperatorName(Next); 2106 } else { 2107 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0) 2108 return false; 2109 for (; Next; Next = Next->Next) { 2110 if (Next->is(TT_TemplateOpener)) { 2111 Next = Next->MatchingParen; 2112 } else if (Next->is(tok::coloncolon)) { 2113 Next = Next->Next; 2114 if (!Next) 2115 return false; 2116 if (Next->is(tok::kw_operator)) { 2117 Next = skipOperatorName(Next->Next); 2118 break; 2119 } 2120 if (!Next->is(tok::identifier)) 2121 return false; 2122 } else if (Next->is(tok::l_paren)) { 2123 break; 2124 } else { 2125 return false; 2126 } 2127 } 2128 } 2129 2130 // Check whether parameter list can belong to a function declaration. 2131 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen) 2132 return false; 2133 // If the lines ends with "{", this is likely an function definition. 2134 if (Line.Last->is(tok::l_brace)) 2135 return true; 2136 if (Next->Next == Next->MatchingParen) 2137 return true; // Empty parentheses. 2138 // If there is an &/&& after the r_paren, this is likely a function. 2139 if (Next->MatchingParen->Next && 2140 Next->MatchingParen->Next->is(TT_PointerOrReference)) 2141 return true; 2142 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen; 2143 Tok = Tok->Next) { 2144 if (Tok->isOneOf(tok::l_paren, TT_TemplateOpener) && Tok->MatchingParen) { 2145 Tok = Tok->MatchingParen; 2146 continue; 2147 } 2148 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 2149 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis)) 2150 return true; 2151 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) || 2152 Tok->Tok.isLiteral()) 2153 return false; 2154 } 2155 return false; 2156 } 2157 2158 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const { 2159 assert(Line.MightBeFunctionDecl); 2160 2161 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel || 2162 Style.AlwaysBreakAfterReturnType == 2163 FormatStyle::RTBS_TopLevelDefinitions) && 2164 Line.Level > 0) 2165 return false; 2166 2167 switch (Style.AlwaysBreakAfterReturnType) { 2168 case FormatStyle::RTBS_None: 2169 return false; 2170 case FormatStyle::RTBS_All: 2171 case FormatStyle::RTBS_TopLevel: 2172 return true; 2173 case FormatStyle::RTBS_AllDefinitions: 2174 case FormatStyle::RTBS_TopLevelDefinitions: 2175 return Line.mightBeFunctionDefinition(); 2176 } 2177 2178 return false; 2179 } 2180 2181 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 2182 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 2183 E = Line.Children.end(); 2184 I != E; ++I) { 2185 calculateFormattingInformation(**I); 2186 } 2187 2188 Line.First->TotalLength = 2189 Line.First->IsMultiline ? Style.ColumnLimit 2190 : Line.FirstStartColumn + Line.First->ColumnWidth; 2191 FormatToken *Current = Line.First->Next; 2192 bool InFunctionDecl = Line.MightBeFunctionDecl; 2193 while (Current) { 2194 if (isFunctionDeclarationName(*Current, Line)) 2195 Current->Type = TT_FunctionDeclarationName; 2196 if (Current->is(TT_LineComment)) { 2197 if (Current->Previous->BlockKind == BK_BracedInit && 2198 Current->Previous->opensScope()) 2199 Current->SpacesRequiredBefore = 2200 (Style.Cpp11BracedListStyle && !Style.SpacesInParentheses) ? 0 : 1; 2201 else 2202 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 2203 2204 // If we find a trailing comment, iterate backwards to determine whether 2205 // it seems to relate to a specific parameter. If so, break before that 2206 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 2207 // to the previous line in: 2208 // SomeFunction(a, 2209 // b, // comment 2210 // c); 2211 if (!Current->HasUnescapedNewline) { 2212 for (FormatToken *Parameter = Current->Previous; Parameter; 2213 Parameter = Parameter->Previous) { 2214 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 2215 break; 2216 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 2217 if (!Parameter->Previous->is(TT_CtorInitializerComma) && 2218 Parameter->HasUnescapedNewline) 2219 Parameter->MustBreakBefore = true; 2220 break; 2221 } 2222 } 2223 } 2224 } else if (Current->SpacesRequiredBefore == 0 && 2225 spaceRequiredBefore(Line, *Current)) { 2226 Current->SpacesRequiredBefore = 1; 2227 } 2228 2229 Current->MustBreakBefore = 2230 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 2231 2232 if (!Current->MustBreakBefore && InFunctionDecl && 2233 Current->is(TT_FunctionDeclarationName)) 2234 Current->MustBreakBefore = mustBreakForReturnType(Line); 2235 2236 Current->CanBreakBefore = 2237 Current->MustBreakBefore || canBreakBefore(Line, *Current); 2238 unsigned ChildSize = 0; 2239 if (Current->Previous->Children.size() == 1) { 2240 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 2241 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 2242 : LastOfChild.TotalLength + 1; 2243 } 2244 const FormatToken *Prev = Current->Previous; 2245 if (Current->MustBreakBefore || Prev->Children.size() > 1 || 2246 (Prev->Children.size() == 1 && 2247 Prev->Children[0]->First->MustBreakBefore) || 2248 Current->IsMultiline) 2249 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit; 2250 else 2251 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth + 2252 ChildSize + Current->SpacesRequiredBefore; 2253 2254 if (Current->is(TT_CtorInitializerColon)) 2255 InFunctionDecl = false; 2256 2257 // FIXME: Only calculate this if CanBreakBefore is true once static 2258 // initializers etc. are sorted out. 2259 // FIXME: Move magic numbers to a better place. 2260 2261 // Reduce penalty for aligning ObjC method arguments using the colon 2262 // alignment as this is the canonical way (still prefer fitting everything 2263 // into one line if possible). Trying to fit a whole expression into one 2264 // line should not force other line breaks (e.g. when ObjC method 2265 // expression is a part of other expression). 2266 Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl); 2267 if (Style.Language == FormatStyle::LK_ObjC && 2268 Current->is(TT_SelectorName) && Current->ParameterIndex > 0) { 2269 if (Current->ParameterIndex == 1) 2270 Current->SplitPenalty += 5 * Current->BindingStrength; 2271 } else { 2272 Current->SplitPenalty += 20 * Current->BindingStrength; 2273 } 2274 2275 Current = Current->Next; 2276 } 2277 2278 calculateUnbreakableTailLengths(Line); 2279 unsigned IndentLevel = Line.Level; 2280 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 2281 if (Current->Role) 2282 Current->Role->precomputeFormattingInfos(Current); 2283 if (Current->MatchingParen && 2284 Current->MatchingParen->opensBlockOrBlockTypeList(Style)) { 2285 assert(IndentLevel > 0); 2286 --IndentLevel; 2287 } 2288 Current->IndentLevel = IndentLevel; 2289 if (Current->opensBlockOrBlockTypeList(Style)) 2290 ++IndentLevel; 2291 } 2292 2293 LLVM_DEBUG({ printDebugInfo(Line); }); 2294 } 2295 2296 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 2297 unsigned UnbreakableTailLength = 0; 2298 FormatToken *Current = Line.Last; 2299 while (Current) { 2300 Current->UnbreakableTailLength = UnbreakableTailLength; 2301 if (Current->CanBreakBefore || 2302 Current->isOneOf(tok::comment, tok::string_literal)) { 2303 UnbreakableTailLength = 0; 2304 } else { 2305 UnbreakableTailLength += 2306 Current->ColumnWidth + Current->SpacesRequiredBefore; 2307 } 2308 Current = Current->Previous; 2309 } 2310 } 2311 2312 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 2313 const FormatToken &Tok, 2314 bool InFunctionDecl) { 2315 const FormatToken &Left = *Tok.Previous; 2316 const FormatToken &Right = Tok; 2317 2318 if (Left.is(tok::semi)) 2319 return 0; 2320 2321 if (Style.Language == FormatStyle::LK_Java) { 2322 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws)) 2323 return 1; 2324 if (Right.is(Keywords.kw_implements)) 2325 return 2; 2326 if (Left.is(tok::comma) && Left.NestingLevel == 0) 2327 return 3; 2328 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2329 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma)) 2330 return 100; 2331 if (Left.is(TT_JsTypeColon)) 2332 return 35; 2333 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2334 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2335 return 100; 2336 // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()". 2337 if (Left.opensScope() && Right.closesScope()) 2338 return 200; 2339 } 2340 2341 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2342 return 1; 2343 if (Right.is(tok::l_square)) { 2344 if (Style.Language == FormatStyle::LK_Proto) 2345 return 1; 2346 if (Left.is(tok::r_square)) 2347 return 200; 2348 // Slightly prefer formatting local lambda definitions like functions. 2349 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal)) 2350 return 35; 2351 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2352 TT_ArrayInitializerLSquare, 2353 TT_DesignatedInitializerLSquare, TT_AttributeSquare)) 2354 return 500; 2355 } 2356 2357 if (Left.is(tok::coloncolon) || 2358 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 2359 return 500; 2360 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2361 Right.is(tok::kw_operator)) { 2362 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 2363 return 3; 2364 if (Left.is(TT_StartOfName)) 2365 return 110; 2366 if (InFunctionDecl && Right.NestingLevel == 0) 2367 return Style.PenaltyReturnTypeOnItsOwnLine; 2368 return 200; 2369 } 2370 if (Right.is(TT_PointerOrReference)) 2371 return 190; 2372 if (Right.is(TT_LambdaArrow)) 2373 return 110; 2374 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 2375 return 160; 2376 if (Left.is(TT_CastRParen)) 2377 return 100; 2378 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 2379 return 5000; 2380 if (Left.is(tok::comment)) 2381 return 1000; 2382 2383 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon, 2384 TT_CtorInitializerColon)) 2385 return 2; 2386 2387 if (Right.isMemberAccess()) { 2388 // Breaking before the "./->" of a chained call/member access is reasonably 2389 // cheap, as formatting those with one call per line is generally 2390 // desirable. In particular, it should be cheaper to break before the call 2391 // than it is to break inside a call's parameters, which could lead to weird 2392 // "hanging" indents. The exception is the very last "./->" to support this 2393 // frequent pattern: 2394 // 2395 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc( 2396 // dddddddd); 2397 // 2398 // which might otherwise be blown up onto many lines. Here, clang-format 2399 // won't produce "hanging" indents anyway as there is no other trailing 2400 // call. 2401 // 2402 // Also apply higher penalty is not a call as that might lead to a wrapping 2403 // like: 2404 // 2405 // aaaaaaa 2406 // .aaaaaaaaa.bbbbbbbb(cccccccc); 2407 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope() 2408 ? 150 2409 : 35; 2410 } 2411 2412 if (Right.is(TT_TrailingAnnotation) && 2413 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 2414 // Moving trailing annotations to the next line is fine for ObjC method 2415 // declarations. 2416 if (Line.startsWith(TT_ObjCMethodSpecifier)) 2417 return 10; 2418 // Generally, breaking before a trailing annotation is bad unless it is 2419 // function-like. It seems to be especially preferable to keep standard 2420 // annotations (i.e. "const", "final" and "override") on the same line. 2421 // Use a slightly higher penalty after ")" so that annotations like 2422 // "const override" are kept together. 2423 bool is_short_annotation = Right.TokenText.size() < 10; 2424 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 2425 } 2426 2427 // In for-loops, prefer breaking at ',' and ';'. 2428 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal)) 2429 return 4; 2430 2431 // In Objective-C method expressions, prefer breaking before "param:" over 2432 // breaking after it. 2433 if (Right.is(TT_SelectorName)) 2434 return 0; 2435 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr)) 2436 return Line.MightBeFunctionDecl ? 50 : 500; 2437 2438 // In Objective-C type declarations, avoid breaking after the category's 2439 // open paren (we'll prefer breaking after the protocol list's opening 2440 // angle bracket, if present). 2441 if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous && 2442 Left.Previous->isOneOf(tok::identifier, tok::greater)) 2443 return 500; 2444 2445 if (Left.is(tok::l_paren) && InFunctionDecl && 2446 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) 2447 return 100; 2448 if (Left.is(tok::l_paren) && Left.Previous && 2449 (Left.Previous->is(tok::kw_for) || Left.Previous->isIf())) 2450 return 1000; 2451 if (Left.is(tok::equal) && InFunctionDecl) 2452 return 110; 2453 if (Right.is(tok::r_brace)) 2454 return 1; 2455 if (Left.is(TT_TemplateOpener)) 2456 return 100; 2457 if (Left.opensScope()) { 2458 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign) 2459 return 0; 2460 if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle) 2461 return 19; 2462 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 2463 : 19; 2464 } 2465 if (Left.is(TT_JavaAnnotation)) 2466 return 50; 2467 2468 if (Left.is(TT_UnaryOperator)) 2469 return 60; 2470 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous && 2471 Left.Previous->isLabelString() && 2472 (Left.NextOperator || Left.OperatorIndex != 0)) 2473 return 50; 2474 if (Right.is(tok::plus) && Left.isLabelString() && 2475 (Right.NextOperator || Right.OperatorIndex != 0)) 2476 return 25; 2477 if (Left.is(tok::comma)) 2478 return 1; 2479 if (Right.is(tok::lessless) && Left.isLabelString() && 2480 (Right.NextOperator || Right.OperatorIndex != 1)) 2481 return 25; 2482 if (Right.is(tok::lessless)) { 2483 // Breaking at a << is really cheap. 2484 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0) 2485 // Slightly prefer to break before the first one in log-like statements. 2486 return 2; 2487 return 1; 2488 } 2489 if (Left.ClosesTemplateDeclaration) 2490 return Style.PenaltyBreakTemplateDeclaration; 2491 if (Left.is(TT_ConditionalExpr)) 2492 return prec::Conditional; 2493 prec::Level Level = Left.getPrecedence(); 2494 if (Level == prec::Unknown) 2495 Level = Right.getPrecedence(); 2496 if (Level == prec::Assignment) 2497 return Style.PenaltyBreakAssignment; 2498 if (Level != prec::Unknown) 2499 return Level; 2500 2501 return 3; 2502 } 2503 2504 bool TokenAnnotator::spaceRequiredBeforeParens(const FormatToken &Right) const { 2505 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always || 2506 (Style.SpaceBeforeParens == FormatStyle::SBPO_NonEmptyParentheses && 2507 Right.ParameterCount > 0); 2508 } 2509 2510 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 2511 const FormatToken &Left, 2512 const FormatToken &Right) { 2513 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 2514 return true; 2515 if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java) 2516 return true; 2517 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 2518 Left.Tok.getObjCKeywordID() == tok::objc_property) 2519 return true; 2520 if (Right.is(tok::hashhash)) 2521 return Left.is(tok::hash); 2522 if (Left.isOneOf(tok::hashhash, tok::hash)) 2523 return Right.is(tok::hash); 2524 if ((Left.is(tok::l_paren) && Right.is(tok::r_paren)) || 2525 (Left.is(tok::l_brace) && Left.BlockKind != BK_Block && 2526 Right.is(tok::r_brace) && Right.BlockKind != BK_Block)) 2527 return Style.SpaceInEmptyParentheses; 2528 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 2529 return (Right.is(TT_CastRParen) || 2530 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen))) 2531 ? Style.SpacesInCStyleCastParentheses 2532 : Style.SpacesInParentheses; 2533 if (Right.isOneOf(tok::semi, tok::comma)) 2534 return false; 2535 if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) { 2536 bool IsLightweightGeneric = Right.MatchingParen && 2537 Right.MatchingParen->Next && 2538 Right.MatchingParen->Next->is(tok::colon); 2539 return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList; 2540 } 2541 if (Right.is(tok::less) && Left.is(tok::kw_template)) 2542 return Style.SpaceAfterTemplateKeyword; 2543 if (Left.isOneOf(tok::exclaim, tok::tilde)) 2544 return false; 2545 if (Left.is(tok::at) && 2546 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 2547 tok::numeric_constant, tok::l_paren, tok::l_brace, 2548 tok::kw_true, tok::kw_false)) 2549 return false; 2550 if (Left.is(tok::colon)) 2551 return !Left.is(TT_ObjCMethodExpr); 2552 if (Left.is(tok::coloncolon)) 2553 return false; 2554 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) { 2555 if (Style.Language == FormatStyle::LK_TextProto || 2556 (Style.Language == FormatStyle::LK_Proto && 2557 (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) { 2558 // Format empty list as `<>`. 2559 if (Left.is(tok::less) && Right.is(tok::greater)) 2560 return false; 2561 return !Style.Cpp11BracedListStyle; 2562 } 2563 return false; 2564 } 2565 if (Right.is(tok::ellipsis)) 2566 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous && 2567 Left.Previous->is(tok::kw_case)); 2568 if (Left.is(tok::l_square) && Right.is(tok::amp)) 2569 return false; 2570 if (Right.is(TT_PointerOrReference)) { 2571 if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) { 2572 if (!Left.MatchingParen) 2573 return true; 2574 FormatToken *TokenBeforeMatchingParen = 2575 Left.MatchingParen->getPreviousNonComment(); 2576 if (!TokenBeforeMatchingParen || 2577 !TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype, 2578 TT_TypenameMacro)) 2579 return true; 2580 } 2581 return (Left.Tok.isLiteral() || 2582 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) && 2583 (Style.PointerAlignment != FormatStyle::PAS_Left || 2584 (Line.IsMultiVariableDeclStmt && 2585 (Left.NestingLevel == 0 || 2586 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for))))))); 2587 } 2588 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) && 2589 (!Left.is(TT_PointerOrReference) || 2590 (Style.PointerAlignment != FormatStyle::PAS_Right && 2591 !Line.IsMultiVariableDeclStmt))) 2592 return true; 2593 if (Left.is(TT_PointerOrReference)) 2594 return Right.Tok.isLiteral() || Right.is(TT_BlockComment) || 2595 (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) && 2596 !Right.is(TT_StartOfName)) || 2597 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) || 2598 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare, 2599 tok::l_paren) && 2600 (Style.PointerAlignment != FormatStyle::PAS_Right && 2601 !Line.IsMultiVariableDeclStmt) && 2602 Left.Previous && 2603 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon, 2604 tok::l_square)); 2605 if (Right.is(tok::star) && Left.is(tok::l_paren)) 2606 return false; 2607 const auto SpaceRequiredForArrayInitializerLSquare = 2608 [](const FormatToken &LSquareTok, const FormatStyle &Style) { 2609 return Style.SpacesInContainerLiterals || 2610 ((Style.Language == FormatStyle::LK_Proto || 2611 Style.Language == FormatStyle::LK_TextProto) && 2612 !Style.Cpp11BracedListStyle && 2613 LSquareTok.endsSequence(tok::l_square, tok::colon, 2614 TT_SelectorName)); 2615 }; 2616 if (Left.is(tok::l_square)) 2617 return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) && 2618 SpaceRequiredForArrayInitializerLSquare(Left, Style)) || 2619 (Left.isOneOf(TT_ArraySubscriptLSquare, 2620 TT_StructuredBindingLSquare) && 2621 Style.SpacesInSquareBrackets && Right.isNot(tok::r_square)); 2622 if (Right.is(tok::r_square)) 2623 return Right.MatchingParen && 2624 ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) && 2625 SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen, 2626 Style)) || 2627 (Style.SpacesInSquareBrackets && 2628 Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare, 2629 TT_StructuredBindingLSquare)) || 2630 Right.MatchingParen->is(TT_AttributeParen)); 2631 if (Right.is(tok::l_square) && 2632 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2633 TT_DesignatedInitializerLSquare, 2634 TT_StructuredBindingLSquare, TT_AttributeSquare) && 2635 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral)) 2636 return false; 2637 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 2638 return !Left.Children.empty(); // No spaces in "{}". 2639 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 2640 (Right.is(tok::r_brace) && Right.MatchingParen && 2641 Right.MatchingParen->BlockKind != BK_Block)) 2642 return Style.Cpp11BracedListStyle ? Style.SpacesInParentheses : true; 2643 if (Left.is(TT_BlockComment)) 2644 // No whitespace in x(/*foo=*/1), except for JavaScript. 2645 return Style.Language == FormatStyle::LK_JavaScript || 2646 !Left.TokenText.endswith("=*/"); 2647 if (Right.is(tok::l_paren)) { 2648 if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) || 2649 (Left.is(tok::r_square) && Left.is(TT_AttributeSquare))) 2650 return true; 2651 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) || 2652 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 2653 (Left.isOneOf(tok::pp_elif, tok::kw_for, tok::kw_while, 2654 tok::kw_switch, tok::kw_case, TT_ForEachMacro, 2655 TT_ObjCForIn) || 2656 Left.isIf(Line.Type != LT_PreprocessorDirective) || 2657 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch, 2658 tok::kw_new, tok::kw_delete) && 2659 (!Left.Previous || Left.Previous->isNot(tok::period))))) || 2660 (spaceRequiredBeforeParens(Right) && 2661 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() || 2662 Left.is(tok::r_paren) || Left.isSimpleTypeSpecifier() || 2663 (Left.is(tok::r_square) && Left.MatchingParen && 2664 Left.MatchingParen->is(TT_LambdaLSquare))) && 2665 Line.Type != LT_PreprocessorDirective); 2666 } 2667 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 2668 return false; 2669 if (Right.is(TT_UnaryOperator)) 2670 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 2671 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr)); 2672 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 2673 tok::r_paren) || 2674 Left.isSimpleTypeSpecifier()) && 2675 Right.is(tok::l_brace) && Right.getNextNonComment() && 2676 Right.BlockKind != BK_Block) 2677 return false; 2678 if (Left.is(tok::period) || Right.is(tok::period)) 2679 return false; 2680 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 2681 return false; 2682 if (Left.is(TT_TemplateCloser) && Left.MatchingParen && 2683 Left.MatchingParen->Previous && 2684 (Left.MatchingParen->Previous->is(tok::period) || 2685 Left.MatchingParen->Previous->is(tok::coloncolon))) 2686 // Java call to generic function with explicit type: 2687 // A.<B<C<...>>>DoSomething(); 2688 // A::<B<C<...>>>DoSomething(); // With a Java 8 method reference. 2689 return false; 2690 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square)) 2691 return false; 2692 if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at)) 2693 // Objective-C dictionary literal -> no space after opening brace. 2694 return false; 2695 if (Right.is(tok::r_brace) && Right.MatchingParen && 2696 Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at)) 2697 // Objective-C dictionary literal -> no space before closing brace. 2698 return false; 2699 if (Right.Type == TT_TrailingAnnotation && 2700 Right.isOneOf(tok::amp, tok::ampamp) && 2701 Left.isOneOf(tok::kw_const, tok::kw_volatile) && 2702 (!Right.Next || Right.Next->is(tok::semi))) 2703 // Match const and volatile ref-qualifiers without any additional 2704 // qualifiers such as 2705 // void Fn() const &; 2706 return Style.PointerAlignment != FormatStyle::PAS_Left; 2707 return true; 2708 } 2709 2710 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 2711 const FormatToken &Right) { 2712 const FormatToken &Left = *Right.Previous; 2713 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo()) 2714 return true; // Never ever merge two identifiers. 2715 if (Style.isCpp()) { 2716 if (Left.is(tok::kw_operator)) 2717 return Right.is(tok::coloncolon); 2718 if (Right.is(tok::l_brace) && Right.BlockKind == BK_BracedInit && 2719 !Left.opensScope() && Style.SpaceBeforeCpp11BracedList) 2720 return true; 2721 } else if (Style.Language == FormatStyle::LK_Proto || 2722 Style.Language == FormatStyle::LK_TextProto) { 2723 if (Right.is(tok::period) && 2724 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required, 2725 Keywords.kw_repeated, Keywords.kw_extend)) 2726 return true; 2727 if (Right.is(tok::l_paren) && 2728 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) 2729 return true; 2730 if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName)) 2731 return true; 2732 // Slashes occur in text protocol extension syntax: [type/type] { ... }. 2733 if (Left.is(tok::slash) || Right.is(tok::slash)) 2734 return false; 2735 if (Left.MatchingParen && 2736 Left.MatchingParen->is(TT_ProtoExtensionLSquare) && 2737 Right.isOneOf(tok::l_brace, tok::less)) 2738 return !Style.Cpp11BracedListStyle; 2739 // A percent is probably part of a formatting specification, such as %lld. 2740 if (Left.is(tok::percent)) 2741 return false; 2742 // Preserve the existence of a space before a percent for cases like 0x%04x 2743 // and "%d %d" 2744 if (Left.is(tok::numeric_constant) && Right.is(tok::percent)) 2745 return Right.WhitespaceRange.getEnd() != Right.WhitespaceRange.getBegin(); 2746 } else if (Style.isCSharp()) { 2747 // space between type and variable e.g. Dictionary<string,string> foo; 2748 if (Left.is(TT_TemplateCloser) && Right.is(TT_StartOfName)) 2749 return true; 2750 // space between keywords and paren e.g. "using (" 2751 if (Right.is(tok::l_paren)) 2752 if (Left.is(tok::kw_using)) 2753 return spaceRequiredBeforeParens(Left); 2754 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2755 if (Left.is(TT_JsFatArrow)) 2756 return true; 2757 // for await ( ... 2758 if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous && 2759 Left.Previous->is(tok::kw_for)) 2760 return true; 2761 if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) && 2762 Right.MatchingParen) { 2763 const FormatToken *Next = Right.MatchingParen->getNextNonComment(); 2764 // An async arrow function, for example: `x = async () => foo();`, 2765 // as opposed to calling a function called async: `x = async();` 2766 if (Next && Next->is(TT_JsFatArrow)) 2767 return true; 2768 } 2769 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2770 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2771 return false; 2772 // In tagged template literals ("html`bar baz`"), there is no space between 2773 // the tag identifier and the template string. getIdentifierInfo makes sure 2774 // that the identifier is not a pseudo keyword like `yield`, either. 2775 if (Left.is(tok::identifier) && Keywords.IsJavaScriptIdentifier(Left) && 2776 Right.is(TT_TemplateString)) 2777 return false; 2778 if (Right.is(tok::star) && 2779 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) 2780 return false; 2781 if (Right.isOneOf(tok::l_brace, tok::l_square) && 2782 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield, 2783 Keywords.kw_extends, Keywords.kw_implements)) 2784 return true; 2785 if (Right.is(tok::l_paren)) { 2786 // JS methods can use some keywords as names (e.g. `delete()`). 2787 if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo()) 2788 return false; 2789 // Valid JS method names can include keywords, e.g. `foo.delete()` or 2790 // `bar.instanceof()`. Recognize call positions by preceding period. 2791 if (Left.Previous && Left.Previous->is(tok::period) && 2792 Left.Tok.getIdentifierInfo()) 2793 return false; 2794 // Additional unary JavaScript operators that need a space after. 2795 if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof, 2796 tok::kw_void)) 2797 return true; 2798 } 2799 // `foo as const;` casts into a const type. 2800 if (Left.endsSequence(tok::kw_const, Keywords.kw_as)) { 2801 return false; 2802 } 2803 if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in, 2804 tok::kw_const) || 2805 // "of" is only a keyword if it appears after another identifier 2806 // (e.g. as "const x of y" in a for loop), or after a destructuring 2807 // operation (const [x, y] of z, const {a, b} of c). 2808 (Left.is(Keywords.kw_of) && Left.Previous && 2809 (Left.Previous->Tok.is(tok::identifier) || 2810 Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) && 2811 (!Left.Previous || !Left.Previous->is(tok::period))) 2812 return true; 2813 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous && 2814 Left.Previous->is(tok::period) && Right.is(tok::l_paren)) 2815 return false; 2816 if (Left.is(Keywords.kw_as) && 2817 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren)) 2818 return true; 2819 if (Left.is(tok::kw_default) && Left.Previous && 2820 Left.Previous->is(tok::kw_export)) 2821 return true; 2822 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace)) 2823 return true; 2824 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion)) 2825 return false; 2826 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator)) 2827 return false; 2828 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) && 2829 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) 2830 return false; 2831 if (Left.is(tok::ellipsis)) 2832 return false; 2833 if (Left.is(TT_TemplateCloser) && 2834 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square, 2835 Keywords.kw_implements, Keywords.kw_extends)) 2836 // Type assertions ('<type>expr') are not followed by whitespace. Other 2837 // locations that should have whitespace following are identified by the 2838 // above set of follower tokens. 2839 return false; 2840 if (Right.is(TT_JsNonNullAssertion)) 2841 return false; 2842 if (Left.is(TT_JsNonNullAssertion) && 2843 Right.isOneOf(Keywords.kw_as, Keywords.kw_in)) 2844 return true; // "x! as string", "x! in y" 2845 } else if (Style.Language == FormatStyle::LK_Java) { 2846 if (Left.is(tok::r_square) && Right.is(tok::l_brace)) 2847 return true; 2848 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) 2849 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never; 2850 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private, 2851 tok::kw_protected) || 2852 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract, 2853 Keywords.kw_native)) && 2854 Right.is(TT_TemplateOpener)) 2855 return true; 2856 } 2857 if (Left.is(TT_ImplicitStringLiteral)) 2858 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2859 if (Line.Type == LT_ObjCMethodDecl) { 2860 if (Left.is(TT_ObjCMethodSpecifier)) 2861 return true; 2862 if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right)) 2863 // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a 2864 // keyword in Objective-C, and '+ (instancetype)new;' is a standard class 2865 // method declaration. 2866 return false; 2867 } 2868 if (Line.Type == LT_ObjCProperty && 2869 (Right.is(tok::equal) || Left.is(tok::equal))) 2870 return false; 2871 2872 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) || 2873 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) 2874 return true; 2875 if (Right.is(TT_OverloadedOperatorLParen)) 2876 return spaceRequiredBeforeParens(Right); 2877 if (Left.is(tok::comma)) 2878 return true; 2879 if (Right.is(tok::comma)) 2880 return false; 2881 if (Right.is(TT_ObjCBlockLParen)) 2882 return true; 2883 if (Right.is(TT_CtorInitializerColon)) 2884 return Style.SpaceBeforeCtorInitializerColon; 2885 if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon) 2886 return false; 2887 if (Right.is(TT_RangeBasedForLoopColon) && 2888 !Style.SpaceBeforeRangeBasedForLoopColon) 2889 return false; 2890 if (Right.is(tok::colon)) { 2891 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) || 2892 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi)) 2893 return false; 2894 if (Right.is(TT_ObjCMethodExpr)) 2895 return false; 2896 if (Left.is(tok::question)) 2897 return false; 2898 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon)) 2899 return false; 2900 if (Right.is(TT_DictLiteral)) 2901 return Style.SpacesInContainerLiterals; 2902 if (Right.is(TT_AttributeColon)) 2903 return false; 2904 return true; 2905 } 2906 if (Left.is(TT_UnaryOperator)) { 2907 // The alternative operators for ~ and ! are "compl" and "not". 2908 // If they are used instead, we do not want to combine them with 2909 // the token to the right, unless that is a left paren. 2910 if (!Right.is(tok::l_paren)) { 2911 if (Left.is(tok::exclaim) && Left.TokenText == "not") 2912 return true; 2913 if (Left.is(tok::tilde) && Left.TokenText == "compl") 2914 return true; 2915 } 2916 return (Style.SpaceAfterLogicalNot && Left.is(tok::exclaim)) || 2917 Right.is(TT_BinaryOperator); 2918 } 2919 2920 // If the next token is a binary operator or a selector name, we have 2921 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly. 2922 if (Left.is(TT_CastRParen)) 2923 return Style.SpaceAfterCStyleCast || 2924 Right.isOneOf(TT_BinaryOperator, TT_SelectorName); 2925 2926 if (Left.is(tok::greater) && Right.is(tok::greater)) { 2927 if (Style.Language == FormatStyle::LK_TextProto || 2928 (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral))) 2929 return !Style.Cpp11BracedListStyle; 2930 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) && 2931 (Style.Standard < FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 2932 } 2933 if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) || 2934 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) || 2935 (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod))) 2936 return false; 2937 if (!Style.SpaceBeforeAssignmentOperators && Left.isNot(TT_TemplateCloser) && 2938 Right.getPrecedence() == prec::Assignment) 2939 return false; 2940 if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) && 2941 (Left.is(tok::identifier) || Left.is(tok::kw_this))) 2942 return false; 2943 if (Right.is(tok::coloncolon) && Left.is(tok::identifier)) 2944 // Generally don't remove existing spaces between an identifier and "::". 2945 // The identifier might actually be a macro name such as ALWAYS_INLINE. If 2946 // this turns out to be too lenient, add analysis of the identifier itself. 2947 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2948 if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment)) 2949 return (Left.is(TT_TemplateOpener) && 2950 Style.Standard < FormatStyle::LS_Cpp11) || 2951 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square, 2952 tok::kw___super, TT_TemplateCloser, 2953 TT_TemplateOpener)) || 2954 (Left.is(tok ::l_paren) && Style.SpacesInParentheses); 2955 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser))) 2956 return Style.SpacesInAngles; 2957 // Space before TT_StructuredBindingLSquare. 2958 if (Right.is(TT_StructuredBindingLSquare)) 2959 return !Left.isOneOf(tok::amp, tok::ampamp) || 2960 Style.PointerAlignment != FormatStyle::PAS_Right; 2961 // Space before & or && following a TT_StructuredBindingLSquare. 2962 if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) && 2963 Right.isOneOf(tok::amp, tok::ampamp)) 2964 return Style.PointerAlignment != FormatStyle::PAS_Left; 2965 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) || 2966 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && 2967 !Right.is(tok::r_paren))) 2968 return true; 2969 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) && 2970 Right.isNot(TT_FunctionTypeLParen)) 2971 return spaceRequiredBeforeParens(Right); 2972 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) && 2973 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen)) 2974 return false; 2975 if (Right.is(tok::less) && Left.isNot(tok::l_paren) && 2976 Line.startsWith(tok::hash)) 2977 return true; 2978 if (Right.is(TT_TrailingUnaryOperator)) 2979 return false; 2980 if (Left.is(TT_RegexLiteral)) 2981 return false; 2982 return spaceRequiredBetween(Line, Left, Right); 2983 } 2984 2985 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 2986 static bool isAllmanBrace(const FormatToken &Tok) { 2987 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 2988 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_LambdaLBrace, TT_DictLiteral); 2989 } 2990 2991 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 2992 const FormatToken &Right) { 2993 const FormatToken &Left = *Right.Previous; 2994 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0) 2995 return true; 2996 2997 if (Style.Language == FormatStyle::LK_JavaScript) { 2998 // FIXME: This might apply to other languages and token kinds. 2999 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous && 3000 Left.Previous->is(tok::string_literal)) 3001 return true; 3002 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 && 3003 Left.Previous && Left.Previous->is(tok::equal) && 3004 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export, 3005 tok::kw_const) && 3006 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match 3007 // above. 3008 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) 3009 // Object literals on the top level of a file are treated as "enum-style". 3010 // Each key/value pair is put on a separate line, instead of bin-packing. 3011 return true; 3012 if (Left.is(tok::l_brace) && Line.Level == 0 && 3013 (Line.startsWith(tok::kw_enum) || 3014 Line.startsWith(tok::kw_const, tok::kw_enum) || 3015 Line.startsWith(tok::kw_export, tok::kw_enum) || 3016 Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum))) 3017 // JavaScript top-level enum key/value pairs are put on separate lines 3018 // instead of bin-packing. 3019 return true; 3020 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && 3021 !Left.Children.empty()) 3022 // Support AllowShortFunctionsOnASingleLine for JavaScript. 3023 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 3024 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty || 3025 (Left.NestingLevel == 0 && Line.Level == 0 && 3026 Style.AllowShortFunctionsOnASingleLine & 3027 FormatStyle::SFS_InlineOnly); 3028 } else if (Style.Language == FormatStyle::LK_Java) { 3029 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next && 3030 Right.Next->is(tok::string_literal)) 3031 return true; 3032 } else if (Style.Language == FormatStyle::LK_Cpp || 3033 Style.Language == FormatStyle::LK_ObjC || 3034 Style.Language == FormatStyle::LK_Proto || 3035 Style.Language == FormatStyle::LK_TableGen || 3036 Style.Language == FormatStyle::LK_TextProto) { 3037 if (Left.isStringLiteral() && Right.isStringLiteral()) 3038 return true; 3039 } 3040 3041 // If the last token before a '}', ']', or ')' is a comma or a trailing 3042 // comment, the intention is to insert a line break after it in order to make 3043 // shuffling around entries easier. Import statements, especially in 3044 // JavaScript, can be an exception to this rule. 3045 if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) { 3046 const FormatToken *BeforeClosingBrace = nullptr; 3047 if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) || 3048 (Style.Language == FormatStyle::LK_JavaScript && 3049 Left.is(tok::l_paren))) && 3050 Left.BlockKind != BK_Block && Left.MatchingParen) 3051 BeforeClosingBrace = Left.MatchingParen->Previous; 3052 else if (Right.MatchingParen && 3053 (Right.MatchingParen->isOneOf(tok::l_brace, 3054 TT_ArrayInitializerLSquare) || 3055 (Style.Language == FormatStyle::LK_JavaScript && 3056 Right.MatchingParen->is(tok::l_paren)))) 3057 BeforeClosingBrace = &Left; 3058 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 3059 BeforeClosingBrace->isTrailingComment())) 3060 return true; 3061 } 3062 3063 if (Right.is(tok::comment)) 3064 return Left.BlockKind != BK_BracedInit && 3065 Left.isNot(TT_CtorInitializerColon) && 3066 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 3067 if (Left.isTrailingComment()) 3068 return true; 3069 if (Right.Previous->IsUnterminatedLiteral) 3070 return true; 3071 if (Right.is(tok::lessless) && Right.Next && 3072 Right.Previous->is(tok::string_literal) && 3073 Right.Next->is(tok::string_literal)) 3074 return true; 3075 if (Right.Previous->ClosesTemplateDeclaration && 3076 Right.Previous->MatchingParen && 3077 Right.Previous->MatchingParen->NestingLevel == 0 && 3078 Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes) 3079 return true; 3080 if (Right.is(TT_CtorInitializerComma) && 3081 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 3082 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 3083 return true; 3084 if (Right.is(TT_CtorInitializerColon) && 3085 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 3086 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 3087 return true; 3088 // Break only if we have multiple inheritance. 3089 if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma && 3090 Right.is(TT_InheritanceComma)) 3091 return true; 3092 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) 3093 // Multiline raw string literals are special wrt. line breaks. The author 3094 // has made a deliberate choice and might have aligned the contents of the 3095 // string literal accordingly. Thus, we try keep existing line breaks. 3096 return Right.IsMultiline && Right.NewlinesBefore > 0; 3097 if ((Right.Previous->is(tok::l_brace) || 3098 (Right.Previous->is(tok::less) && Right.Previous->Previous && 3099 Right.Previous->Previous->is(tok::equal))) && 3100 Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) { 3101 // Don't put enums or option definitions onto single lines in protocol 3102 // buffers. 3103 return true; 3104 } 3105 if (Right.is(TT_InlineASMBrace)) 3106 return Right.HasUnescapedNewline; 3107 if (isAllmanBrace(Left) || isAllmanBrace(Right)) 3108 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) || 3109 (Line.startsWith(tok::kw_typedef, tok::kw_enum) && 3110 Style.BraceWrapping.AfterEnum) || 3111 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) || 3112 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct); 3113 if (Left.is(TT_ObjCBlockLBrace) && 3114 Style.AllowShortBlocksOnASingleLine == FormatStyle::SBS_Never) 3115 return true; 3116 3117 if (Left.is(TT_LambdaLBrace)) { 3118 if (Left.MatchingParen && Left.MatchingParen->Next && 3119 Left.MatchingParen->Next->isOneOf(tok::comma, tok::r_paren) && 3120 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline) 3121 return false; 3122 3123 if (Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_None || 3124 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline || 3125 (!Left.Children.empty() && 3126 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Empty)) 3127 return true; 3128 } 3129 3130 // Put multiple C# attributes on a new line. 3131 if (Style.isCSharp() && 3132 ((Left.is(TT_AttributeSquare) && Left.is(tok::r_square)) || 3133 (Left.is(tok::r_square) && Right.is(TT_AttributeSquare) && 3134 Right.is(tok::l_square)))) 3135 return true; 3136 3137 // Put multiple Java annotation on a new line. 3138 if ((Style.Language == FormatStyle::LK_Java || 3139 Style.Language == FormatStyle::LK_JavaScript) && 3140 Left.is(TT_LeadingJavaAnnotation) && 3141 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) && 3142 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) 3143 return true; 3144 3145 if (Right.is(TT_ProtoExtensionLSquare)) 3146 return true; 3147 3148 // In text proto instances if a submessage contains at least 2 entries and at 3149 // least one of them is a submessage, like A { ... B { ... } ... }, 3150 // put all of the entries of A on separate lines by forcing the selector of 3151 // the submessage B to be put on a newline. 3152 // 3153 // Example: these can stay on one line: 3154 // a { scalar_1: 1 scalar_2: 2 } 3155 // a { b { key: value } } 3156 // 3157 // and these entries need to be on a new line even if putting them all in one 3158 // line is under the column limit: 3159 // a { 3160 // scalar: 1 3161 // b { key: value } 3162 // } 3163 // 3164 // We enforce this by breaking before a submessage field that has previous 3165 // siblings, *and* breaking before a field that follows a submessage field. 3166 // 3167 // Be careful to exclude the case [proto.ext] { ... } since the `]` is 3168 // the TT_SelectorName there, but we don't want to break inside the brackets. 3169 // 3170 // Another edge case is @submessage { key: value }, which is a common 3171 // substitution placeholder. In this case we want to keep `@` and `submessage` 3172 // together. 3173 // 3174 // We ensure elsewhere that extensions are always on their own line. 3175 if ((Style.Language == FormatStyle::LK_Proto || 3176 Style.Language == FormatStyle::LK_TextProto) && 3177 Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) { 3178 // Keep `@submessage` together in: 3179 // @submessage { key: value } 3180 if (Right.Previous && Right.Previous->is(tok::at)) 3181 return false; 3182 // Look for the scope opener after selector in cases like: 3183 // selector { ... 3184 // selector: { ... 3185 // selector: @base { ... 3186 FormatToken *LBrace = Right.Next; 3187 if (LBrace && LBrace->is(tok::colon)) { 3188 LBrace = LBrace->Next; 3189 if (LBrace && LBrace->is(tok::at)) { 3190 LBrace = LBrace->Next; 3191 if (LBrace) 3192 LBrace = LBrace->Next; 3193 } 3194 } 3195 if (LBrace && 3196 // The scope opener is one of {, [, <: 3197 // selector { ... } 3198 // selector [ ... ] 3199 // selector < ... > 3200 // 3201 // In case of selector { ... }, the l_brace is TT_DictLiteral. 3202 // In case of an empty selector {}, the l_brace is not TT_DictLiteral, 3203 // so we check for immediately following r_brace. 3204 ((LBrace->is(tok::l_brace) && 3205 (LBrace->is(TT_DictLiteral) || 3206 (LBrace->Next && LBrace->Next->is(tok::r_brace)))) || 3207 LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) { 3208 // If Left.ParameterCount is 0, then this submessage entry is not the 3209 // first in its parent submessage, and we want to break before this entry. 3210 // If Left.ParameterCount is greater than 0, then its parent submessage 3211 // might contain 1 or more entries and we want to break before this entry 3212 // if it contains at least 2 entries. We deal with this case later by 3213 // detecting and breaking before the next entry in the parent submessage. 3214 if (Left.ParameterCount == 0) 3215 return true; 3216 // However, if this submessage is the first entry in its parent 3217 // submessage, Left.ParameterCount might be 1 in some cases. 3218 // We deal with this case later by detecting an entry 3219 // following a closing paren of this submessage. 3220 } 3221 3222 // If this is an entry immediately following a submessage, it will be 3223 // preceded by a closing paren of that submessage, like in: 3224 // left---. .---right 3225 // v v 3226 // sub: { ... } key: value 3227 // If there was a comment between `}` an `key` above, then `key` would be 3228 // put on a new line anyways. 3229 if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square)) 3230 return true; 3231 } 3232 3233 // Deal with lambda arguments in C++ - we want consistent line breaks whether 3234 // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced 3235 // as aggressive line breaks are placed when the lambda is not the last arg. 3236 if ((Style.Language == FormatStyle::LK_Cpp || 3237 Style.Language == FormatStyle::LK_ObjC) && 3238 Left.is(tok::l_paren) && Left.BlockParameterCount > 0 && 3239 !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) { 3240 // Multiple lambdas in the same function call force line breaks. 3241 if (Left.BlockParameterCount > 1) 3242 return true; 3243 3244 // A lambda followed by another arg forces a line break. 3245 if (!Left.Role) 3246 return false; 3247 auto Comma = Left.Role->lastComma(); 3248 if (!Comma) 3249 return false; 3250 auto Next = Comma->getNextNonComment(); 3251 if (!Next) 3252 return false; 3253 if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret)) 3254 return true; 3255 } 3256 3257 return false; 3258 } 3259 3260 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 3261 const FormatToken &Right) { 3262 const FormatToken &Left = *Right.Previous; 3263 3264 // Language-specific stuff. 3265 if (Style.Language == FormatStyle::LK_Java) { 3266 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 3267 Keywords.kw_implements)) 3268 return false; 3269 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 3270 Keywords.kw_implements)) 3271 return true; 3272 } else if (Style.Language == FormatStyle::LK_JavaScript) { 3273 const FormatToken *NonComment = Right.getPreviousNonComment(); 3274 if (NonComment && 3275 NonComment->isOneOf( 3276 tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break, 3277 tok::kw_throw, Keywords.kw_interface, Keywords.kw_type, 3278 tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected, 3279 Keywords.kw_readonly, Keywords.kw_abstract, Keywords.kw_get, 3280 Keywords.kw_set, Keywords.kw_async, Keywords.kw_await)) 3281 return false; // Otherwise automatic semicolon insertion would trigger. 3282 if (Right.NestingLevel == 0 && 3283 (Left.Tok.getIdentifierInfo() || 3284 Left.isOneOf(tok::r_square, tok::r_paren)) && 3285 Right.isOneOf(tok::l_square, tok::l_paren)) 3286 return false; // Otherwise automatic semicolon insertion would trigger. 3287 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace)) 3288 return false; 3289 if (Left.is(TT_JsTypeColon)) 3290 return true; 3291 // Don't wrap between ":" and "!" of a strict prop init ("field!: type;"). 3292 if (Left.is(tok::exclaim) && Right.is(tok::colon)) 3293 return false; 3294 // Look for is type annotations like: 3295 // function f(): a is B { ... } 3296 // Do not break before is in these cases. 3297 if (Right.is(Keywords.kw_is)) { 3298 const FormatToken *Next = Right.getNextNonComment(); 3299 // If `is` is followed by a colon, it's likely that it's a dict key, so 3300 // ignore it for this check. 3301 // For example this is common in Polymer: 3302 // Polymer({ 3303 // is: 'name', 3304 // ... 3305 // }); 3306 if (!Next || !Next->is(tok::colon)) 3307 return false; 3308 } 3309 if (Left.is(Keywords.kw_in)) 3310 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None; 3311 if (Right.is(Keywords.kw_in)) 3312 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None; 3313 if (Right.is(Keywords.kw_as)) 3314 return false; // must not break before as in 'x as type' casts 3315 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) { 3316 // extends and infer can appear as keywords in conditional types: 3317 // https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types 3318 // do not break before them, as the expressions are subject to ASI. 3319 return false; 3320 } 3321 if (Left.is(Keywords.kw_as)) 3322 return true; 3323 if (Left.is(TT_JsNonNullAssertion)) 3324 return true; 3325 if (Left.is(Keywords.kw_declare) && 3326 Right.isOneOf(Keywords.kw_module, tok::kw_namespace, 3327 Keywords.kw_function, tok::kw_class, tok::kw_enum, 3328 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var, 3329 Keywords.kw_let, tok::kw_const)) 3330 // See grammar for 'declare' statements at: 3331 // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10 3332 return false; 3333 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) && 3334 Right.isOneOf(tok::identifier, tok::string_literal)) 3335 return false; // must not break in "module foo { ...}" 3336 if (Right.is(TT_TemplateString) && Right.closesScope()) 3337 return false; 3338 // Don't split tagged template literal so there is a break between the tag 3339 // identifier and template string. 3340 if (Left.is(tok::identifier) && Right.is(TT_TemplateString)) { 3341 return false; 3342 } 3343 if (Left.is(TT_TemplateString) && Left.opensScope()) 3344 return true; 3345 } 3346 3347 if (Left.is(tok::at)) 3348 return false; 3349 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 3350 return false; 3351 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation)) 3352 return !Right.is(tok::l_paren); 3353 if (Right.is(TT_PointerOrReference)) 3354 return Line.IsMultiVariableDeclStmt || 3355 (Style.PointerAlignment == FormatStyle::PAS_Right && 3356 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName))); 3357 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 3358 Right.is(tok::kw_operator)) 3359 return true; 3360 if (Left.is(TT_PointerOrReference)) 3361 return false; 3362 if (Right.isTrailingComment()) 3363 // We rely on MustBreakBefore being set correctly here as we should not 3364 // change the "binding" behavior of a comment. 3365 // The first comment in a braced lists is always interpreted as belonging to 3366 // the first list element. Otherwise, it should be placed outside of the 3367 // list. 3368 return Left.BlockKind == BK_BracedInit || 3369 (Left.is(TT_CtorInitializerColon) && 3370 Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon); 3371 if (Left.is(tok::question) && Right.is(tok::colon)) 3372 return false; 3373 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question)) 3374 return Style.BreakBeforeTernaryOperators; 3375 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question)) 3376 return !Style.BreakBeforeTernaryOperators; 3377 if (Left.is(TT_InheritanceColon)) 3378 return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon; 3379 if (Right.is(TT_InheritanceColon)) 3380 return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon; 3381 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) && 3382 Left.isNot(TT_SelectorName)) 3383 return true; 3384 3385 if (Right.is(tok::colon) && 3386 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) 3387 return false; 3388 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) { 3389 if (Style.Language == FormatStyle::LK_Proto || 3390 Style.Language == FormatStyle::LK_TextProto) { 3391 if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral()) 3392 return false; 3393 // Prevent cases like: 3394 // 3395 // submessage: 3396 // { key: valueeeeeeeeeeee } 3397 // 3398 // when the snippet does not fit into one line. 3399 // Prefer: 3400 // 3401 // submessage: { 3402 // key: valueeeeeeeeeeee 3403 // } 3404 // 3405 // instead, even if it is longer by one line. 3406 // 3407 // Note that this allows allows the "{" to go over the column limit 3408 // when the column limit is just between ":" and "{", but that does 3409 // not happen too often and alternative formattings in this case are 3410 // not much better. 3411 // 3412 // The code covers the cases: 3413 // 3414 // submessage: { ... } 3415 // submessage: < ... > 3416 // repeated: [ ... ] 3417 if (((Right.is(tok::l_brace) || Right.is(tok::less)) && 3418 Right.is(TT_DictLiteral)) || 3419 Right.is(TT_ArrayInitializerLSquare)) 3420 return false; 3421 } 3422 return true; 3423 } 3424 if (Right.is(tok::r_square) && Right.MatchingParen && 3425 Right.MatchingParen->is(TT_ProtoExtensionLSquare)) 3426 return false; 3427 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next && 3428 Right.Next->is(TT_ObjCMethodExpr))) 3429 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls. 3430 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 3431 return true; 3432 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen)) 3433 return true; 3434 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen, 3435 TT_OverloadedOperator)) 3436 return false; 3437 if (Left.is(TT_RangeBasedForLoopColon)) 3438 return true; 3439 if (Right.is(TT_RangeBasedForLoopColon)) 3440 return false; 3441 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener)) 3442 return true; 3443 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) || 3444 Left.is(tok::kw_operator)) 3445 return false; 3446 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) && 3447 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) 3448 return false; 3449 if (Left.is(tok::equal) && Right.is(tok::l_brace) && 3450 !Style.Cpp11BracedListStyle) 3451 return false; 3452 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen)) 3453 return false; 3454 if (Left.is(tok::l_paren) && Left.Previous && 3455 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) 3456 return false; 3457 if (Right.is(TT_ImplicitStringLiteral)) 3458 return false; 3459 3460 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser)) 3461 return false; 3462 if (Right.is(tok::r_square) && Right.MatchingParen && 3463 Right.MatchingParen->is(TT_LambdaLSquare)) 3464 return false; 3465 3466 // We only break before r_brace if there was a corresponding break before 3467 // the l_brace, which is tracked by BreakBeforeClosingBrace. 3468 if (Right.is(tok::r_brace)) 3469 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 3470 3471 // Allow breaking after a trailing annotation, e.g. after a method 3472 // declaration. 3473 if (Left.is(TT_TrailingAnnotation)) 3474 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 3475 tok::less, tok::coloncolon); 3476 3477 if (Right.is(tok::kw___attribute) || 3478 (Right.is(tok::l_square) && Right.is(TT_AttributeSquare))) 3479 return true; 3480 3481 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 3482 return true; 3483 3484 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 3485 return true; 3486 3487 if (Left.is(TT_CtorInitializerColon)) 3488 return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon; 3489 if (Right.is(TT_CtorInitializerColon)) 3490 return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon; 3491 if (Left.is(TT_CtorInitializerComma) && 3492 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 3493 return false; 3494 if (Right.is(TT_CtorInitializerComma) && 3495 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 3496 return true; 3497 if (Left.is(TT_InheritanceComma) && 3498 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) 3499 return false; 3500 if (Right.is(TT_InheritanceComma) && 3501 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) 3502 return true; 3503 if ((Left.is(tok::greater) && Right.is(tok::greater)) || 3504 (Left.is(tok::less) && Right.is(tok::less))) 3505 return false; 3506 if (Right.is(TT_BinaryOperator) && 3507 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 3508 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 3509 Right.getPrecedence() != prec::Assignment)) 3510 return true; 3511 if (Left.is(TT_ArrayInitializerLSquare)) 3512 return true; 3513 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 3514 return true; 3515 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) && 3516 !Left.isOneOf(tok::arrowstar, tok::lessless) && 3517 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 3518 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 3519 Left.getPrecedence() == prec::Assignment)) 3520 return true; 3521 if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) || 3522 (Left.is(tok::r_square) && Right.is(TT_AttributeSquare))) 3523 return false; 3524 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 3525 tok::kw_class, tok::kw_struct, tok::comment) || 3526 Right.isMemberAccess() || 3527 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless, 3528 tok::colon, tok::l_square, tok::at) || 3529 (Left.is(tok::r_paren) && 3530 Right.isOneOf(tok::identifier, tok::kw_const)) || 3531 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) || 3532 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser)); 3533 } 3534 3535 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 3536 llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n"; 3537 const FormatToken *Tok = Line.First; 3538 while (Tok) { 3539 llvm::errs() << " M=" << Tok->MustBreakBefore 3540 << " C=" << Tok->CanBreakBefore 3541 << " T=" << getTokenTypeName(Tok->Type) 3542 << " S=" << Tok->SpacesRequiredBefore 3543 << " B=" << Tok->BlockParameterCount 3544 << " BK=" << Tok->BlockKind << " P=" << Tok->SplitPenalty 3545 << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength 3546 << " PPK=" << Tok->PackingKind << " FakeLParens="; 3547 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 3548 llvm::errs() << Tok->FakeLParens[i] << "/"; 3549 llvm::errs() << " FakeRParens=" << Tok->FakeRParens; 3550 llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo(); 3551 llvm::errs() << " Text='" << Tok->TokenText << "'\n"; 3552 if (!Tok->Next) 3553 assert(Tok == Line.Last); 3554 Tok = Tok->Next; 3555 } 3556 llvm::errs() << "----\n"; 3557 } 3558 3559 } // namespace format 3560 } // namespace clang 3561