1 //===--- TokenAnnotator.cpp - Format C++ code -----------------------------===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is distributed under the University of Illinois Open Source 6 // License. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 /// 10 /// \file 11 /// This file implements a token annotator, i.e. creates 12 /// \c AnnotatedTokens out of \c FormatTokens with required extra information. 13 /// 14 //===----------------------------------------------------------------------===// 15 16 #include "TokenAnnotator.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 /// A parser that gathers additional information about tokens. 44 /// 45 /// The \c TokenAnnotator tries to match parenthesis and square brakets and 46 /// store a parenthesis levels. It also tries to resolve matching "<" and ">" 47 /// into template parameter lists. 48 class AnnotatingParser { 49 public: 50 AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line, 51 const AdditionalKeywords &Keywords) 52 : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false), 53 Keywords(Keywords) { 54 Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false)); 55 resetTokenMetadata(CurrentToken); 56 } 57 58 private: 59 bool parseAngle() { 60 if (!CurrentToken || !CurrentToken->Previous) 61 return false; 62 if (NonTemplateLess.count(CurrentToken->Previous)) 63 return false; 64 65 const FormatToken &Previous = *CurrentToken->Previous; // The '<'. 66 if (Previous.Previous) { 67 if (Previous.Previous->Tok.isLiteral()) 68 return false; 69 if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 && 70 (!Previous.Previous->MatchingParen || 71 !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen))) 72 return false; 73 } 74 75 FormatToken *Left = CurrentToken->Previous; 76 Left->ParentBracket = Contexts.back().ContextKind; 77 ScopedContextCreator ContextCreator(*this, tok::less, 12); 78 79 // If this angle is in the context of an expression, we need to be more 80 // hesitant to detect it as opening template parameters. 81 bool InExprContext = Contexts.back().IsExpression; 82 83 Contexts.back().IsExpression = false; 84 // If there's a template keyword before the opening angle bracket, this is a 85 // template parameter, not an argument. 86 Contexts.back().InTemplateArgument = 87 Left->Previous && Left->Previous->Tok.isNot(tok::kw_template); 88 89 if (Style.Language == FormatStyle::LK_Java && 90 CurrentToken->is(tok::question)) 91 next(); 92 93 while (CurrentToken) { 94 if (CurrentToken->is(tok::greater)) { 95 Left->MatchingParen = CurrentToken; 96 CurrentToken->MatchingParen = Left; 97 // In TT_Proto, we must distignuish between: 98 // map<key, value> 99 // msg < item: data > 100 // msg: < item: data > 101 // In TT_TextProto, map<key, value> does not occur. 102 if (Style.Language == FormatStyle::LK_TextProto || 103 (Style.Language == FormatStyle::LK_Proto && Left->Previous && 104 Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) 105 CurrentToken->Type = TT_DictLiteral; 106 else 107 CurrentToken->Type = TT_TemplateCloser; 108 next(); 109 return true; 110 } 111 if (CurrentToken->is(tok::question) && 112 Style.Language == FormatStyle::LK_Java) { 113 next(); 114 continue; 115 } 116 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) || 117 (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext && 118 Style.Language != FormatStyle::LK_Proto && 119 Style.Language != FormatStyle::LK_TextProto)) 120 return false; 121 // If a && or || is found and interpreted as a binary operator, this set 122 // of angles is likely part of something like "a < b && c > d". If the 123 // angles are inside an expression, the ||/&& might also be a binary 124 // operator that was misinterpreted because we are parsing template 125 // parameters. 126 // FIXME: This is getting out of hand, write a decent parser. 127 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) && 128 CurrentToken->Previous->is(TT_BinaryOperator) && 129 Contexts[Contexts.size() - 2].IsExpression && 130 !Line.startsWith(tok::kw_template)) 131 return false; 132 updateParameterCount(Left, CurrentToken); 133 if (Style.Language == FormatStyle::LK_Proto) { 134 if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) { 135 if (CurrentToken->is(tok::colon) || 136 (CurrentToken->isOneOf(tok::l_brace, tok::less) && 137 Previous->isNot(tok::colon))) 138 Previous->Type = TT_SelectorName; 139 } 140 } 141 if (!consumeToken()) 142 return false; 143 } 144 return false; 145 } 146 147 bool parseParens(bool LookForDecls = false) { 148 if (!CurrentToken) 149 return false; 150 FormatToken *Left = CurrentToken->Previous; 151 Left->ParentBracket = Contexts.back().ContextKind; 152 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1); 153 154 // FIXME: This is a bit of a hack. Do better. 155 Contexts.back().ColonIsForRangeExpr = 156 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr; 157 158 bool StartsObjCMethodExpr = false; 159 if (FormatToken *MaybeSel = Left->Previous) { 160 // @selector( starts a selector. 161 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous && 162 MaybeSel->Previous->is(tok::at)) { 163 StartsObjCMethodExpr = true; 164 } 165 } 166 167 if (Left->is(TT_OverloadedOperatorLParen)) { 168 Contexts.back().IsExpression = false; 169 } else if (Style.Language == FormatStyle::LK_JavaScript && 170 (Line.startsWith(Keywords.kw_type, tok::identifier) || 171 Line.startsWith(tok::kw_export, Keywords.kw_type, 172 tok::identifier))) { 173 // type X = (...); 174 // export type X = (...); 175 Contexts.back().IsExpression = false; 176 } else if (Left->Previous && 177 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype, 178 tok::kw_if, tok::kw_while, tok::l_paren, 179 tok::comma) || 180 Left->Previous->endsSequence(tok::kw_constexpr, tok::kw_if) || 181 Left->Previous->is(TT_BinaryOperator))) { 182 // static_assert, if and while usually contain expressions. 183 Contexts.back().IsExpression = true; 184 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 185 (Left->Previous->is(Keywords.kw_function) || 186 (Left->Previous->endsSequence(tok::identifier, 187 Keywords.kw_function)))) { 188 // function(...) or function f(...) 189 Contexts.back().IsExpression = false; 190 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 191 Left->Previous->is(TT_JsTypeColon)) { 192 // let x: (SomeType); 193 Contexts.back().IsExpression = false; 194 } else if (Left->Previous && Left->Previous->is(tok::r_square) && 195 Left->Previous->MatchingParen && 196 Left->Previous->MatchingParen->is(TT_LambdaLSquare)) { 197 // This is a parameter list of a lambda expression. 198 Contexts.back().IsExpression = false; 199 } else if (Line.InPPDirective && 200 (!Left->Previous || !Left->Previous->is(tok::identifier))) { 201 Contexts.back().IsExpression = true; 202 } else if (Contexts[Contexts.size() - 2].CaretFound) { 203 // This is the parameter list of an ObjC block. 204 Contexts.back().IsExpression = false; 205 } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) { 206 Left->Type = TT_AttributeParen; 207 } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) { 208 // The first argument to a foreach macro is a declaration. 209 Contexts.back().IsForEachMacro = true; 210 Contexts.back().IsExpression = false; 211 } else if (Left->Previous && Left->Previous->MatchingParen && 212 Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) { 213 Contexts.back().IsExpression = false; 214 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) { 215 bool IsForOrCatch = 216 Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch); 217 Contexts.back().IsExpression = !IsForOrCatch; 218 } 219 220 if (StartsObjCMethodExpr) { 221 Contexts.back().ColonIsObjCMethodExpr = true; 222 Left->Type = TT_ObjCMethodExpr; 223 } 224 225 // MightBeFunctionType and ProbablyFunctionType are used for 226 // function pointer and reference types as well as Objective-C 227 // block types: 228 // 229 // void (*FunctionPointer)(void); 230 // void (&FunctionReference)(void); 231 // void (^ObjCBlock)(void); 232 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression; 233 bool ProbablyFunctionType = 234 CurrentToken->isOneOf(tok::star, tok::amp, tok::caret); 235 bool HasMultipleLines = false; 236 bool HasMultipleParametersOnALine = false; 237 bool MightBeObjCForRangeLoop = 238 Left->Previous && Left->Previous->is(tok::kw_for); 239 FormatToken *PossibleObjCForInToken = nullptr; 240 while (CurrentToken) { 241 // LookForDecls is set when "if (" has been seen. Check for 242 // 'identifier' '*' 'identifier' followed by not '=' -- this 243 // '*' has to be a binary operator but determineStarAmpUsage() will 244 // categorize it as an unary operator, so set the right type here. 245 if (LookForDecls && CurrentToken->Next) { 246 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 247 if (Prev) { 248 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 249 FormatToken *Next = CurrentToken->Next; 250 if (PrevPrev && PrevPrev->is(tok::identifier) && 251 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 252 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 253 Prev->Type = TT_BinaryOperator; 254 LookForDecls = false; 255 } 256 } 257 } 258 259 if (CurrentToken->Previous->is(TT_PointerOrReference) && 260 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 261 tok::coloncolon)) 262 ProbablyFunctionType = true; 263 if (CurrentToken->is(tok::comma)) 264 MightBeFunctionType = false; 265 if (CurrentToken->Previous->is(TT_BinaryOperator)) 266 Contexts.back().IsExpression = true; 267 if (CurrentToken->is(tok::r_paren)) { 268 if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next && 269 (CurrentToken->Next->is(tok::l_paren) || 270 (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration))) 271 Left->Type = Left->Next->is(tok::caret) ? TT_ObjCBlockLParen 272 : TT_FunctionTypeLParen; 273 Left->MatchingParen = CurrentToken; 274 CurrentToken->MatchingParen = Left; 275 276 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) && 277 Left->Previous && Left->Previous->is(tok::l_paren)) { 278 // Detect the case where macros are used to generate lambdas or 279 // function bodies, e.g.: 280 // auto my_lambda = MARCO((Type *type, int i) { .. body .. }); 281 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) { 282 if (Tok->is(TT_BinaryOperator) && 283 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) 284 Tok->Type = TT_PointerOrReference; 285 } 286 } 287 288 if (StartsObjCMethodExpr) { 289 CurrentToken->Type = TT_ObjCMethodExpr; 290 if (Contexts.back().FirstObjCSelectorName) { 291 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 292 Contexts.back().LongestObjCSelectorName; 293 } 294 } 295 296 if (Left->is(TT_AttributeParen)) 297 CurrentToken->Type = TT_AttributeParen; 298 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation)) 299 CurrentToken->Type = TT_JavaAnnotation; 300 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation)) 301 CurrentToken->Type = TT_LeadingJavaAnnotation; 302 303 if (!HasMultipleLines) 304 Left->PackingKind = PPK_Inconclusive; 305 else if (HasMultipleParametersOnALine) 306 Left->PackingKind = PPK_BinPacked; 307 else 308 Left->PackingKind = PPK_OnePerLine; 309 310 next(); 311 return true; 312 } 313 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 314 return false; 315 316 if (CurrentToken->is(tok::l_brace)) 317 Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen 318 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 319 !CurrentToken->Next->HasUnescapedNewline && 320 !CurrentToken->Next->isTrailingComment()) 321 HasMultipleParametersOnALine = true; 322 if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) || 323 CurrentToken->Previous->isSimpleTypeSpecifier()) && 324 !CurrentToken->is(tok::l_brace)) 325 Contexts.back().IsExpression = false; 326 if (CurrentToken->isOneOf(tok::semi, tok::colon)) { 327 MightBeObjCForRangeLoop = false; 328 if (PossibleObjCForInToken) { 329 PossibleObjCForInToken->Type = TT_Unknown; 330 PossibleObjCForInToken = nullptr; 331 } 332 } 333 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) { 334 PossibleObjCForInToken = CurrentToken; 335 PossibleObjCForInToken->Type = TT_ObjCForIn; 336 } 337 // When we discover a 'new', we set CanBeExpression to 'false' in order to 338 // parse the type correctly. Reset that after a comma. 339 if (CurrentToken->is(tok::comma)) 340 Contexts.back().CanBeExpression = true; 341 342 FormatToken *Tok = CurrentToken; 343 if (!consumeToken()) 344 return false; 345 updateParameterCount(Left, Tok); 346 if (CurrentToken && CurrentToken->HasUnescapedNewline) 347 HasMultipleLines = true; 348 } 349 return false; 350 } 351 352 bool isCpp11AttributeSpecifier(const FormatToken &Tok) { 353 if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square)) 354 return false; 355 const FormatToken *AttrTok = Tok.Next->Next; 356 if (!AttrTok) 357 return false; 358 // C++17 '[[using ns: foo, bar(baz, blech)]]' 359 // We assume nobody will name an ObjC variable 'using'. 360 if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon)) 361 return true; 362 if (AttrTok->isNot(tok::identifier)) 363 return false; 364 while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) { 365 // ObjC message send. We assume nobody will use : in a C++11 attribute 366 // specifier parameter, although this is technically valid: 367 // [[foo(:)]] 368 if (AttrTok->is(tok::colon) || 369 AttrTok->startsSequence(tok::identifier, tok::identifier) || 370 AttrTok->startsSequence(tok::r_paren, tok::identifier)) 371 return false; 372 if (AttrTok->is(tok::ellipsis)) 373 return true; 374 AttrTok = AttrTok->Next; 375 } 376 return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square); 377 } 378 379 bool parseSquare() { 380 if (!CurrentToken) 381 return false; 382 383 // A '[' could be an index subscript (after an identifier or after 384 // ')' or ']'), it could be the start of an Objective-C method 385 // expression, it could the start of an Objective-C array literal, 386 // or it could be a C++ attribute specifier [[foo::bar]]. 387 FormatToken *Left = CurrentToken->Previous; 388 Left->ParentBracket = Contexts.back().ContextKind; 389 FormatToken *Parent = Left->getPreviousNonComment(); 390 391 // Cases where '>' is followed by '['. 392 // In C++, this can happen either in array of templates (foo<int>[10]) 393 // or when array is a nested template type (unique_ptr<type1<type2>[]>). 394 bool CppArrayTemplates = 395 Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) && 396 (Contexts.back().CanBeExpression || Contexts.back().IsExpression || 397 Contexts.back().InTemplateArgument); 398 399 bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) || 400 Contexts.back().InCpp11AttributeSpecifier; 401 402 bool StartsObjCMethodExpr = 403 !CppArrayTemplates && Style.isCpp() && !IsCpp11AttributeSpecifier && 404 Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) && 405 !CurrentToken->isOneOf(tok::l_brace, tok::r_square) && 406 (!Parent || 407 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 408 tok::kw_return, tok::kw_throw) || 409 Parent->isUnaryOperator() || 410 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 411 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) || 412 getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown); 413 bool ColonFound = false; 414 415 unsigned BindingIncrease = 1; 416 if (Left->isCppStructuredBinding(Style)) { 417 Left->Type = TT_StructuredBindingLSquare; 418 } else if (Left->is(TT_Unknown)) { 419 if (StartsObjCMethodExpr) { 420 Left->Type = TT_ObjCMethodExpr; 421 } else if (IsCpp11AttributeSpecifier) { 422 Left->Type = TT_AttributeSquare; 423 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent && 424 Contexts.back().ContextKind == tok::l_brace && 425 Parent->isOneOf(tok::l_brace, tok::comma)) { 426 Left->Type = TT_JsComputedPropertyName; 427 } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace && 428 Parent && Parent->isOneOf(tok::l_brace, tok::comma)) { 429 Left->Type = TT_DesignatedInitializerLSquare; 430 } else if (CurrentToken->is(tok::r_square) && Parent && 431 Parent->is(TT_TemplateCloser)) { 432 Left->Type = TT_ArraySubscriptLSquare; 433 } else if (Style.Language == FormatStyle::LK_Proto || 434 Style.Language == FormatStyle::LK_TextProto) { 435 // Square braces in LK_Proto can either be message field attributes: 436 // 437 // optional Aaa aaa = 1 [ 438 // (aaa) = aaa 439 // ]; 440 // 441 // extensions 123 [ 442 // (aaa) = aaa 443 // ]; 444 // 445 // or text proto extensions (in options): 446 // 447 // option (Aaa.options) = { 448 // [type.type/type] { 449 // key: value 450 // } 451 // } 452 // 453 // or repeated fields (in options): 454 // 455 // option (Aaa.options) = { 456 // keys: [ 1, 2, 3 ] 457 // } 458 // 459 // In the first and the third case we want to spread the contents inside 460 // the square braces; in the second we want to keep them inline. 461 Left->Type = TT_ArrayInitializerLSquare; 462 if (!Left->endsSequence(tok::l_square, tok::numeric_constant, 463 tok::equal) && 464 !Left->endsSequence(tok::l_square, tok::numeric_constant, 465 tok::identifier) && 466 !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) { 467 Left->Type = TT_ProtoExtensionLSquare; 468 BindingIncrease = 10; 469 } 470 } else if (!CppArrayTemplates && Parent && 471 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at, 472 tok::comma, tok::l_paren, tok::l_square, 473 tok::question, tok::colon, tok::kw_return, 474 // Should only be relevant to JavaScript: 475 tok::kw_default)) { 476 Left->Type = TT_ArrayInitializerLSquare; 477 } else { 478 BindingIncrease = 10; 479 Left->Type = TT_ArraySubscriptLSquare; 480 } 481 } 482 483 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease); 484 Contexts.back().IsExpression = true; 485 if (Style.Language == FormatStyle::LK_JavaScript && Parent && 486 Parent->is(TT_JsTypeColon)) 487 Contexts.back().IsExpression = false; 488 489 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr; 490 Contexts.back().InCpp11AttributeSpecifier = IsCpp11AttributeSpecifier; 491 492 while (CurrentToken) { 493 if (CurrentToken->is(tok::r_square)) { 494 if (IsCpp11AttributeSpecifier) 495 CurrentToken->Type = TT_AttributeSquare; 496 else if (CurrentToken->Next && CurrentToken->Next->is(tok::l_paren) && 497 Left->is(TT_ObjCMethodExpr)) { 498 // An ObjC method call is rarely followed by an open parenthesis. 499 // FIXME: Do we incorrectly label ":" with this? 500 StartsObjCMethodExpr = false; 501 Left->Type = TT_Unknown; 502 } 503 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) { 504 CurrentToken->Type = TT_ObjCMethodExpr; 505 // If we haven't seen a colon yet, make sure the last identifier 506 // before the r_square is tagged as a selector name component. 507 if (!ColonFound && CurrentToken->Previous && 508 CurrentToken->Previous->is(TT_Unknown) && 509 canBeObjCSelectorComponent(*CurrentToken->Previous)) 510 CurrentToken->Previous->Type = TT_SelectorName; 511 // determineStarAmpUsage() thinks that '*' '[' is allocating an 512 // array of pointers, but if '[' starts a selector then '*' is a 513 // binary operator. 514 if (Parent && Parent->is(TT_PointerOrReference)) 515 Parent->Type = TT_BinaryOperator; 516 } 517 Left->MatchingParen = CurrentToken; 518 CurrentToken->MatchingParen = Left; 519 // FirstObjCSelectorName is set when a colon is found. This does 520 // not work, however, when the method has no parameters. 521 // Here, we set FirstObjCSelectorName when the end of the method call is 522 // reached, in case it was not set already. 523 if (!Contexts.back().FirstObjCSelectorName) { 524 FormatToken* Previous = CurrentToken->getPreviousNonComment(); 525 if (Previous && Previous->is(TT_SelectorName)) { 526 Previous->ObjCSelectorNameParts = 1; 527 Contexts.back().FirstObjCSelectorName = Previous; 528 } 529 } else { 530 Left->ParameterCount = 531 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 532 } 533 if (Contexts.back().FirstObjCSelectorName) { 534 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 535 Contexts.back().LongestObjCSelectorName; 536 if (Left->BlockParameterCount > 1) 537 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0; 538 } 539 next(); 540 return true; 541 } 542 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace)) 543 return false; 544 if (CurrentToken->is(tok::colon)) { 545 if (IsCpp11AttributeSpecifier && 546 CurrentToken->endsSequence(tok::colon, tok::identifier, 547 tok::kw_using)) { 548 // Remember that this is a [[using ns: foo]] C++ attribute, so we 549 // don't add a space before the colon (unlike other colons). 550 CurrentToken->Type = TT_AttributeColon; 551 } else if (Left->isOneOf(TT_ArraySubscriptLSquare, 552 TT_DesignatedInitializerLSquare)) { 553 Left->Type = TT_ObjCMethodExpr; 554 StartsObjCMethodExpr = true; 555 Contexts.back().ColonIsObjCMethodExpr = true; 556 if (Parent && Parent->is(tok::r_paren)) 557 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 558 Parent->Type = TT_CastRParen; 559 } 560 ColonFound = true; 561 } 562 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) && 563 !ColonFound) 564 Left->Type = TT_ArrayInitializerLSquare; 565 FormatToken *Tok = CurrentToken; 566 if (!consumeToken()) 567 return false; 568 updateParameterCount(Left, Tok); 569 } 570 return false; 571 } 572 573 bool parseBrace() { 574 if (CurrentToken) { 575 FormatToken *Left = CurrentToken->Previous; 576 Left->ParentBracket = Contexts.back().ContextKind; 577 578 if (Contexts.back().CaretFound) 579 Left->Type = TT_ObjCBlockLBrace; 580 Contexts.back().CaretFound = false; 581 582 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1); 583 Contexts.back().ColonIsDictLiteral = true; 584 if (Left->BlockKind == BK_BracedInit) 585 Contexts.back().IsExpression = true; 586 if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 587 Left->Previous->is(TT_JsTypeColon)) 588 Contexts.back().IsExpression = false; 589 590 while (CurrentToken) { 591 if (CurrentToken->is(tok::r_brace)) { 592 Left->MatchingParen = CurrentToken; 593 CurrentToken->MatchingParen = Left; 594 next(); 595 return true; 596 } 597 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square)) 598 return false; 599 updateParameterCount(Left, CurrentToken); 600 if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) { 601 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 602 if (Previous->is(TT_JsTypeOptionalQuestion)) 603 Previous = Previous->getPreviousNonComment(); 604 if ((CurrentToken->is(tok::colon) && 605 (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) || 606 Style.Language == FormatStyle::LK_Proto || 607 Style.Language == FormatStyle::LK_TextProto) { 608 Left->Type = TT_DictLiteral; 609 if (Previous->Tok.getIdentifierInfo() || 610 Previous->is(tok::string_literal)) 611 Previous->Type = TT_SelectorName; 612 } 613 if (CurrentToken->is(tok::colon) || 614 Style.Language == FormatStyle::LK_JavaScript) 615 Left->Type = TT_DictLiteral; 616 } 617 if (CurrentToken->is(tok::comma) && 618 Style.Language == FormatStyle::LK_JavaScript) 619 Left->Type = TT_DictLiteral; 620 if (!consumeToken()) 621 return false; 622 } 623 } 624 return true; 625 } 626 627 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 628 // For ObjC methods, the number of parameters is calculated differently as 629 // method declarations have a different structure (the parameters are not 630 // inside a bracket scope). 631 if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block) 632 ++Left->BlockParameterCount; 633 if (Current->is(tok::comma)) { 634 ++Left->ParameterCount; 635 if (!Left->Role) 636 Left->Role.reset(new CommaSeparatedList(Style)); 637 Left->Role->CommaFound(Current); 638 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 639 Left->ParameterCount = 1; 640 } 641 } 642 643 bool parseConditional() { 644 while (CurrentToken) { 645 if (CurrentToken->is(tok::colon)) { 646 CurrentToken->Type = TT_ConditionalExpr; 647 next(); 648 return true; 649 } 650 if (!consumeToken()) 651 return false; 652 } 653 return false; 654 } 655 656 bool parseTemplateDeclaration() { 657 if (CurrentToken && CurrentToken->is(tok::less)) { 658 CurrentToken->Type = TT_TemplateOpener; 659 next(); 660 if (!parseAngle()) 661 return false; 662 if (CurrentToken) 663 CurrentToken->Previous->ClosesTemplateDeclaration = true; 664 return true; 665 } 666 return false; 667 } 668 669 bool consumeToken() { 670 FormatToken *Tok = CurrentToken; 671 next(); 672 switch (Tok->Tok.getKind()) { 673 case tok::plus: 674 case tok::minus: 675 if (!Tok->Previous && Line.MustBeDeclaration) 676 Tok->Type = TT_ObjCMethodSpecifier; 677 break; 678 case tok::colon: 679 if (!Tok->Previous) 680 return false; 681 // Colons from ?: are handled in parseConditional(). 682 if (Style.Language == FormatStyle::LK_JavaScript) { 683 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop 684 (Contexts.size() == 1 && // switch/case labels 685 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) || 686 Contexts.back().ContextKind == tok::l_paren || // function params 687 Contexts.back().ContextKind == tok::l_square || // array type 688 (!Contexts.back().IsExpression && 689 Contexts.back().ContextKind == tok::l_brace) || // object type 690 (Contexts.size() == 1 && 691 Line.MustBeDeclaration)) { // method/property declaration 692 Contexts.back().IsExpression = false; 693 Tok->Type = TT_JsTypeColon; 694 break; 695 } 696 } 697 if (Contexts.back().ColonIsDictLiteral || 698 Style.Language == FormatStyle::LK_Proto || 699 Style.Language == FormatStyle::LK_TextProto) { 700 Tok->Type = TT_DictLiteral; 701 if (Style.Language == FormatStyle::LK_TextProto) { 702 if (FormatToken *Previous = Tok->getPreviousNonComment()) 703 Previous->Type = TT_SelectorName; 704 } 705 } else if (Contexts.back().ColonIsObjCMethodExpr || 706 Line.startsWith(TT_ObjCMethodSpecifier)) { 707 Tok->Type = TT_ObjCMethodExpr; 708 const FormatToken *BeforePrevious = Tok->Previous->Previous; 709 // Ensure we tag all identifiers in method declarations as 710 // TT_SelectorName. 711 bool UnknownIdentifierInMethodDeclaration = 712 Line.startsWith(TT_ObjCMethodSpecifier) && 713 Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown); 714 if (!BeforePrevious || 715 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 716 !(BeforePrevious->is(TT_CastRParen) || 717 (BeforePrevious->is(TT_ObjCMethodExpr) && 718 BeforePrevious->is(tok::colon))) || 719 BeforePrevious->is(tok::r_square) || 720 Contexts.back().LongestObjCSelectorName == 0 || 721 UnknownIdentifierInMethodDeclaration) { 722 Tok->Previous->Type = TT_SelectorName; 723 if (!Contexts.back().FirstObjCSelectorName) 724 Contexts.back().FirstObjCSelectorName = Tok->Previous; 725 else if (Tok->Previous->ColumnWidth > 726 Contexts.back().LongestObjCSelectorName) 727 Contexts.back().LongestObjCSelectorName = 728 Tok->Previous->ColumnWidth; 729 Tok->Previous->ParameterIndex = 730 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 731 ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts; 732 } 733 } else if (Contexts.back().ColonIsForRangeExpr) { 734 Tok->Type = TT_RangeBasedForLoopColon; 735 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 736 Tok->Type = TT_BitFieldColon; 737 } else if (Contexts.size() == 1 && 738 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) { 739 if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren, 740 tok::kw_noexcept)) 741 Tok->Type = TT_CtorInitializerColon; 742 else 743 Tok->Type = TT_InheritanceColon; 744 } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next && 745 (Tok->Next->isOneOf(tok::r_paren, tok::comma) || 746 (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next && 747 Tok->Next->Next->is(tok::colon)))) { 748 // This handles a special macro in ObjC code where selectors including 749 // the colon are passed as macro arguments. 750 Tok->Type = TT_ObjCMethodExpr; 751 } else if (Contexts.back().ContextKind == tok::l_paren) { 752 Tok->Type = TT_InlineASMColon; 753 } 754 break; 755 case tok::pipe: 756 case tok::amp: 757 // | and & in declarations/type expressions represent union and 758 // intersection types, respectively. 759 if (Style.Language == FormatStyle::LK_JavaScript && 760 !Contexts.back().IsExpression) 761 Tok->Type = TT_JsTypeOperator; 762 break; 763 case tok::kw_if: 764 case tok::kw_while: 765 if (Tok->is(tok::kw_if) && CurrentToken && 766 CurrentToken->is(tok::kw_constexpr)) 767 next(); 768 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 769 next(); 770 if (!parseParens(/*LookForDecls=*/true)) 771 return false; 772 } 773 break; 774 case tok::kw_for: 775 if (Style.Language == FormatStyle::LK_JavaScript) { 776 // x.for and {for: ...} 777 if ((Tok->Previous && Tok->Previous->is(tok::period)) || 778 (Tok->Next && Tok->Next->is(tok::colon))) 779 break; 780 // JS' for await ( ... 781 if (CurrentToken && CurrentToken->is(Keywords.kw_await)) 782 next(); 783 } 784 Contexts.back().ColonIsForRangeExpr = true; 785 next(); 786 if (!parseParens()) 787 return false; 788 break; 789 case tok::l_paren: 790 // When faced with 'operator()()', the kw_operator handler incorrectly 791 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make 792 // the first two parens OverloadedOperators and the second l_paren an 793 // OverloadedOperatorLParen. 794 if (Tok->Previous && Tok->Previous->is(tok::r_paren) && 795 Tok->Previous->MatchingParen && 796 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) { 797 Tok->Previous->Type = TT_OverloadedOperator; 798 Tok->Previous->MatchingParen->Type = TT_OverloadedOperator; 799 Tok->Type = TT_OverloadedOperatorLParen; 800 } 801 802 if (!parseParens()) 803 return false; 804 if (Line.MustBeDeclaration && Contexts.size() == 1 && 805 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) && 806 (!Tok->Previous || 807 !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute, 808 TT_LeadingJavaAnnotation))) 809 Line.MightBeFunctionDecl = true; 810 break; 811 case tok::l_square: 812 if (!parseSquare()) 813 return false; 814 break; 815 case tok::l_brace: 816 if (Style.Language == FormatStyle::LK_TextProto) { 817 FormatToken *Previous = Tok->getPreviousNonComment(); 818 if (Previous && Previous->Type != TT_DictLiteral) 819 Previous->Type = TT_SelectorName; 820 } 821 if (!parseBrace()) 822 return false; 823 break; 824 case tok::less: 825 if (parseAngle()) { 826 Tok->Type = TT_TemplateOpener; 827 // In TT_Proto, we must distignuish between: 828 // map<key, value> 829 // msg < item: data > 830 // msg: < item: data > 831 // In TT_TextProto, map<key, value> does not occur. 832 if (Style.Language == FormatStyle::LK_TextProto || 833 (Style.Language == FormatStyle::LK_Proto && Tok->Previous && 834 Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) { 835 Tok->Type = TT_DictLiteral; 836 FormatToken *Previous = Tok->getPreviousNonComment(); 837 if (Previous && Previous->Type != TT_DictLiteral) 838 Previous->Type = TT_SelectorName; 839 } 840 } else { 841 Tok->Type = TT_BinaryOperator; 842 NonTemplateLess.insert(Tok); 843 CurrentToken = Tok; 844 next(); 845 } 846 break; 847 case tok::r_paren: 848 case tok::r_square: 849 return false; 850 case tok::r_brace: 851 // Lines can start with '}'. 852 if (Tok->Previous) 853 return false; 854 break; 855 case tok::greater: 856 if (Style.Language != FormatStyle::LK_TextProto) 857 Tok->Type = TT_BinaryOperator; 858 break; 859 case tok::kw_operator: 860 if (Style.Language == FormatStyle::LK_TextProto || 861 Style.Language == FormatStyle::LK_Proto) 862 break; 863 while (CurrentToken && 864 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 865 if (CurrentToken->isOneOf(tok::star, tok::amp)) 866 CurrentToken->Type = TT_PointerOrReference; 867 consumeToken(); 868 if (CurrentToken && 869 CurrentToken->Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator, 870 tok::comma)) 871 CurrentToken->Previous->Type = TT_OverloadedOperator; 872 } 873 if (CurrentToken) { 874 CurrentToken->Type = TT_OverloadedOperatorLParen; 875 if (CurrentToken->Previous->is(TT_BinaryOperator)) 876 CurrentToken->Previous->Type = TT_OverloadedOperator; 877 } 878 break; 879 case tok::question: 880 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next && 881 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren, 882 tok::r_brace)) { 883 // Question marks before semicolons, colons, etc. indicate optional 884 // types (fields, parameters), e.g. 885 // function(x?: string, y?) {...} 886 // class X { y?; } 887 Tok->Type = TT_JsTypeOptionalQuestion; 888 break; 889 } 890 // Declarations cannot be conditional expressions, this can only be part 891 // of a type declaration. 892 if (Line.MustBeDeclaration && !Contexts.back().IsExpression && 893 Style.Language == FormatStyle::LK_JavaScript) 894 break; 895 parseConditional(); 896 break; 897 case tok::kw_template: 898 parseTemplateDeclaration(); 899 break; 900 case tok::comma: 901 if (Contexts.back().InCtorInitializer) 902 Tok->Type = TT_CtorInitializerComma; 903 else if (Contexts.back().InInheritanceList) 904 Tok->Type = TT_InheritanceComma; 905 else if (Contexts.back().FirstStartOfName && 906 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) { 907 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 908 Line.IsMultiVariableDeclStmt = true; 909 } 910 if (Contexts.back().IsForEachMacro) 911 Contexts.back().IsExpression = true; 912 break; 913 case tok::identifier: 914 if (Tok->isOneOf(Keywords.kw___has_include, 915 Keywords.kw___has_include_next)) { 916 parseHasInclude(); 917 } 918 break; 919 default: 920 break; 921 } 922 return true; 923 } 924 925 void parseIncludeDirective() { 926 if (CurrentToken && CurrentToken->is(tok::less)) { 927 next(); 928 while (CurrentToken) { 929 // Mark tokens up to the trailing line comments as implicit string 930 // literals. 931 if (CurrentToken->isNot(tok::comment) && 932 !CurrentToken->TokenText.startswith("//")) 933 CurrentToken->Type = TT_ImplicitStringLiteral; 934 next(); 935 } 936 } 937 } 938 939 void parseWarningOrError() { 940 next(); 941 // We still want to format the whitespace left of the first token of the 942 // warning or error. 943 next(); 944 while (CurrentToken) { 945 CurrentToken->Type = TT_ImplicitStringLiteral; 946 next(); 947 } 948 } 949 950 void parsePragma() { 951 next(); // Consume "pragma". 952 if (CurrentToken && 953 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) { 954 bool IsMark = CurrentToken->is(Keywords.kw_mark); 955 next(); // Consume "mark". 956 next(); // Consume first token (so we fix leading whitespace). 957 while (CurrentToken) { 958 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator)) 959 CurrentToken->Type = TT_ImplicitStringLiteral; 960 next(); 961 } 962 } 963 } 964 965 void parseHasInclude() { 966 if (!CurrentToken || !CurrentToken->is(tok::l_paren)) 967 return; 968 next(); // '(' 969 parseIncludeDirective(); 970 next(); // ')' 971 } 972 973 LineType parsePreprocessorDirective() { 974 bool IsFirstToken = CurrentToken->IsFirst; 975 LineType Type = LT_PreprocessorDirective; 976 next(); 977 if (!CurrentToken) 978 return Type; 979 980 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) { 981 // JavaScript files can contain shebang lines of the form: 982 // #!/usr/bin/env node 983 // Treat these like C++ #include directives. 984 while (CurrentToken) { 985 // Tokens cannot be comments here. 986 CurrentToken->Type = TT_ImplicitStringLiteral; 987 next(); 988 } 989 return LT_ImportStatement; 990 } 991 992 if (CurrentToken->Tok.is(tok::numeric_constant)) { 993 CurrentToken->SpacesRequiredBefore = 1; 994 return Type; 995 } 996 // Hashes in the middle of a line can lead to any strange token 997 // sequence. 998 if (!CurrentToken->Tok.getIdentifierInfo()) 999 return Type; 1000 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 1001 case tok::pp_include: 1002 case tok::pp_include_next: 1003 case tok::pp_import: 1004 next(); 1005 parseIncludeDirective(); 1006 Type = LT_ImportStatement; 1007 break; 1008 case tok::pp_error: 1009 case tok::pp_warning: 1010 parseWarningOrError(); 1011 break; 1012 case tok::pp_pragma: 1013 parsePragma(); 1014 break; 1015 case tok::pp_if: 1016 case tok::pp_elif: 1017 Contexts.back().IsExpression = true; 1018 parseLine(); 1019 break; 1020 default: 1021 break; 1022 } 1023 while (CurrentToken) { 1024 FormatToken *Tok = CurrentToken; 1025 next(); 1026 if (Tok->is(tok::l_paren)) 1027 parseParens(); 1028 else if (Tok->isOneOf(Keywords.kw___has_include, 1029 Keywords.kw___has_include_next)) 1030 parseHasInclude(); 1031 } 1032 return Type; 1033 } 1034 1035 public: 1036 LineType parseLine() { 1037 NonTemplateLess.clear(); 1038 if (CurrentToken->is(tok::hash)) 1039 return parsePreprocessorDirective(); 1040 1041 // Directly allow to 'import <string-literal>' to support protocol buffer 1042 // definitions (github.com/google/protobuf) or missing "#" (either way we 1043 // should not break the line). 1044 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 1045 if ((Style.Language == FormatStyle::LK_Java && 1046 CurrentToken->is(Keywords.kw_package)) || 1047 (Info && Info->getPPKeywordID() == tok::pp_import && 1048 CurrentToken->Next && 1049 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier, 1050 tok::kw_static))) { 1051 next(); 1052 parseIncludeDirective(); 1053 return LT_ImportStatement; 1054 } 1055 1056 // If this line starts and ends in '<' and '>', respectively, it is likely 1057 // part of "#define <a/b.h>". 1058 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 1059 parseIncludeDirective(); 1060 return LT_ImportStatement; 1061 } 1062 1063 // In .proto files, top-level options are very similar to import statements 1064 // and should not be line-wrapped. 1065 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 && 1066 CurrentToken->is(Keywords.kw_option)) { 1067 next(); 1068 if (CurrentToken && CurrentToken->is(tok::identifier)) 1069 return LT_ImportStatement; 1070 } 1071 1072 bool KeywordVirtualFound = false; 1073 bool ImportStatement = false; 1074 1075 // import {...} from '...'; 1076 if (Style.Language == FormatStyle::LK_JavaScript && 1077 CurrentToken->is(Keywords.kw_import)) 1078 ImportStatement = true; 1079 1080 while (CurrentToken) { 1081 if (CurrentToken->is(tok::kw_virtual)) 1082 KeywordVirtualFound = true; 1083 if (Style.Language == FormatStyle::LK_JavaScript) { 1084 // export {...} from '...'; 1085 // An export followed by "from 'some string';" is a re-export from 1086 // another module identified by a URI and is treated as a 1087 // LT_ImportStatement (i.e. prevent wraps on it for long URIs). 1088 // Just "export {...};" or "export class ..." should not be treated as 1089 // an import in this sense. 1090 if (Line.First->is(tok::kw_export) && 1091 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next && 1092 CurrentToken->Next->isStringLiteral()) 1093 ImportStatement = true; 1094 if (isClosureImportStatement(*CurrentToken)) 1095 ImportStatement = true; 1096 } 1097 if (!consumeToken()) 1098 return LT_Invalid; 1099 } 1100 if (KeywordVirtualFound) 1101 return LT_VirtualFunctionDecl; 1102 if (ImportStatement) 1103 return LT_ImportStatement; 1104 1105 if (Line.startsWith(TT_ObjCMethodSpecifier)) { 1106 if (Contexts.back().FirstObjCSelectorName) 1107 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 1108 Contexts.back().LongestObjCSelectorName; 1109 return LT_ObjCMethodDecl; 1110 } 1111 1112 return LT_Other; 1113 } 1114 1115 private: 1116 bool isClosureImportStatement(const FormatToken &Tok) { 1117 // FIXME: Closure-library specific stuff should not be hard-coded but be 1118 // configurable. 1119 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) && 1120 Tok.Next->Next && 1121 (Tok.Next->Next->TokenText == "module" || 1122 Tok.Next->Next->TokenText == "provide" || 1123 Tok.Next->Next->TokenText == "require" || 1124 Tok.Next->Next->TokenText == "forwardDeclare") && 1125 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren); 1126 } 1127 1128 void resetTokenMetadata(FormatToken *Token) { 1129 if (!Token) 1130 return; 1131 1132 // Reset token type in case we have already looked at it and then 1133 // recovered from an error (e.g. failure to find the matching >). 1134 if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro, 1135 TT_FunctionLBrace, TT_ImplicitStringLiteral, 1136 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow, 1137 TT_OverloadedOperator, TT_RegexLiteral, 1138 TT_TemplateString, TT_ObjCStringLiteral)) 1139 CurrentToken->Type = TT_Unknown; 1140 CurrentToken->Role.reset(); 1141 CurrentToken->MatchingParen = nullptr; 1142 CurrentToken->FakeLParens.clear(); 1143 CurrentToken->FakeRParens = 0; 1144 } 1145 1146 void next() { 1147 if (CurrentToken) { 1148 CurrentToken->NestingLevel = Contexts.size() - 1; 1149 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 1150 modifyContext(*CurrentToken); 1151 determineTokenType(*CurrentToken); 1152 CurrentToken = CurrentToken->Next; 1153 } 1154 1155 resetTokenMetadata(CurrentToken); 1156 } 1157 1158 /// A struct to hold information valid in a specific context, e.g. 1159 /// a pair of parenthesis. 1160 struct Context { 1161 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 1162 bool IsExpression) 1163 : ContextKind(ContextKind), BindingStrength(BindingStrength), 1164 IsExpression(IsExpression) {} 1165 1166 tok::TokenKind ContextKind; 1167 unsigned BindingStrength; 1168 bool IsExpression; 1169 unsigned LongestObjCSelectorName = 0; 1170 bool ColonIsForRangeExpr = false; 1171 bool ColonIsDictLiteral = false; 1172 bool ColonIsObjCMethodExpr = false; 1173 FormatToken *FirstObjCSelectorName = nullptr; 1174 FormatToken *FirstStartOfName = nullptr; 1175 bool CanBeExpression = true; 1176 bool InTemplateArgument = false; 1177 bool InCtorInitializer = false; 1178 bool InInheritanceList = false; 1179 bool CaretFound = false; 1180 bool IsForEachMacro = false; 1181 bool InCpp11AttributeSpecifier = false; 1182 }; 1183 1184 /// Puts a new \c Context onto the stack \c Contexts for the lifetime 1185 /// of each instance. 1186 struct ScopedContextCreator { 1187 AnnotatingParser &P; 1188 1189 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 1190 unsigned Increase) 1191 : P(P) { 1192 P.Contexts.push_back(Context(ContextKind, 1193 P.Contexts.back().BindingStrength + Increase, 1194 P.Contexts.back().IsExpression)); 1195 } 1196 1197 ~ScopedContextCreator() { P.Contexts.pop_back(); } 1198 }; 1199 1200 void modifyContext(const FormatToken &Current) { 1201 if (Current.getPrecedence() == prec::Assignment && 1202 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) && 1203 // Type aliases use `type X = ...;` in TypeScript and can be exported 1204 // using `export type ...`. 1205 !(Style.Language == FormatStyle::LK_JavaScript && 1206 (Line.startsWith(Keywords.kw_type, tok::identifier) || 1207 Line.startsWith(tok::kw_export, Keywords.kw_type, 1208 tok::identifier))) && 1209 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 1210 Contexts.back().IsExpression = true; 1211 if (!Line.startsWith(TT_UnaryOperator)) { 1212 for (FormatToken *Previous = Current.Previous; 1213 Previous && Previous->Previous && 1214 !Previous->Previous->isOneOf(tok::comma, tok::semi); 1215 Previous = Previous->Previous) { 1216 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 1217 Previous = Previous->MatchingParen; 1218 if (!Previous) 1219 break; 1220 } 1221 if (Previous->opensScope()) 1222 break; 1223 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) && 1224 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) && 1225 Previous->Previous && Previous->Previous->isNot(tok::equal)) 1226 Previous->Type = TT_PointerOrReference; 1227 } 1228 } 1229 } else if (Current.is(tok::lessless) && 1230 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) { 1231 Contexts.back().IsExpression = true; 1232 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 1233 Contexts.back().IsExpression = true; 1234 } else if (Current.is(TT_TrailingReturnArrow)) { 1235 Contexts.back().IsExpression = false; 1236 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) { 1237 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java; 1238 } else if (Current.Previous && 1239 Current.Previous->is(TT_CtorInitializerColon)) { 1240 Contexts.back().IsExpression = true; 1241 Contexts.back().InCtorInitializer = true; 1242 } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) { 1243 Contexts.back().InInheritanceList = true; 1244 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 1245 for (FormatToken *Previous = Current.Previous; 1246 Previous && Previous->isOneOf(tok::star, tok::amp); 1247 Previous = Previous->Previous) 1248 Previous->Type = TT_PointerOrReference; 1249 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer) 1250 Contexts.back().IsExpression = false; 1251 } else if (Current.is(tok::kw_new)) { 1252 Contexts.back().CanBeExpression = false; 1253 } else if (Current.isOneOf(tok::semi, tok::exclaim)) { 1254 // This should be the condition or increment in a for-loop. 1255 Contexts.back().IsExpression = true; 1256 } 1257 } 1258 1259 void determineTokenType(FormatToken &Current) { 1260 if (!Current.is(TT_Unknown)) 1261 // The token type is already known. 1262 return; 1263 1264 if (Style.Language == FormatStyle::LK_JavaScript) { 1265 if (Current.is(tok::exclaim)) { 1266 if (Current.Previous && 1267 (Current.Previous->isOneOf(tok::identifier, tok::kw_namespace, 1268 tok::r_paren, tok::r_square, 1269 tok::r_brace) || 1270 Current.Previous->Tok.isLiteral())) { 1271 Current.Type = TT_JsNonNullAssertion; 1272 return; 1273 } 1274 if (Current.Next && 1275 Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) { 1276 Current.Type = TT_JsNonNullAssertion; 1277 return; 1278 } 1279 } 1280 } 1281 1282 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1283 // function declaration have been found. In this case, 'Current' is a 1284 // trailing token of this declaration and thus cannot be a name. 1285 if (Current.is(Keywords.kw_instanceof)) { 1286 Current.Type = TT_BinaryOperator; 1287 } else if (isStartOfName(Current) && 1288 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 1289 Contexts.back().FirstStartOfName = &Current; 1290 Current.Type = TT_StartOfName; 1291 } else if (Current.is(tok::semi)) { 1292 // Reset FirstStartOfName after finding a semicolon so that a for loop 1293 // with multiple increment statements is not confused with a for loop 1294 // having multiple variable declarations. 1295 Contexts.back().FirstStartOfName = nullptr; 1296 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) { 1297 AutoFound = true; 1298 } else if (Current.is(tok::arrow) && 1299 Style.Language == FormatStyle::LK_Java) { 1300 Current.Type = TT_LambdaArrow; 1301 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration && 1302 Current.NestingLevel == 0) { 1303 Current.Type = TT_TrailingReturnArrow; 1304 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 1305 Current.Type = determineStarAmpUsage(Current, 1306 Contexts.back().CanBeExpression && 1307 Contexts.back().IsExpression, 1308 Contexts.back().InTemplateArgument); 1309 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 1310 Current.Type = determinePlusMinusCaretUsage(Current); 1311 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret)) 1312 Contexts.back().CaretFound = true; 1313 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 1314 Current.Type = determineIncrementUsage(Current); 1315 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 1316 Current.Type = TT_UnaryOperator; 1317 } else if (Current.is(tok::question)) { 1318 if (Style.Language == FormatStyle::LK_JavaScript && 1319 Line.MustBeDeclaration && !Contexts.back().IsExpression) { 1320 // In JavaScript, `interface X { foo?(): bar; }` is an optional method 1321 // on the interface, not a ternary expression. 1322 Current.Type = TT_JsTypeOptionalQuestion; 1323 } else { 1324 Current.Type = TT_ConditionalExpr; 1325 } 1326 } else if (Current.isBinaryOperator() && 1327 (!Current.Previous || Current.Previous->isNot(tok::l_square)) && 1328 (!Current.is(tok::greater) && 1329 Style.Language != FormatStyle::LK_TextProto)) { 1330 Current.Type = TT_BinaryOperator; 1331 } else if (Current.is(tok::comment)) { 1332 if (Current.TokenText.startswith("/*")) { 1333 if (Current.TokenText.endswith("*/")) 1334 Current.Type = TT_BlockComment; 1335 else 1336 // The lexer has for some reason determined a comment here. But we 1337 // cannot really handle it, if it isn't properly terminated. 1338 Current.Tok.setKind(tok::unknown); 1339 } else { 1340 Current.Type = TT_LineComment; 1341 } 1342 } else if (Current.is(tok::r_paren)) { 1343 if (rParenEndsCast(Current)) 1344 Current.Type = TT_CastRParen; 1345 if (Current.MatchingParen && Current.Next && 1346 !Current.Next->isBinaryOperator() && 1347 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace, 1348 tok::comma, tok::period, tok::arrow, 1349 tok::coloncolon)) 1350 if (FormatToken *AfterParen = Current.MatchingParen->Next) { 1351 // Make sure this isn't the return type of an Obj-C block declaration 1352 if (AfterParen->Tok.isNot(tok::caret)) { 1353 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) 1354 if (BeforeParen->is(tok::identifier) && 1355 BeforeParen->TokenText == BeforeParen->TokenText.upper() && 1356 (!BeforeParen->Previous || 1357 BeforeParen->Previous->ClosesTemplateDeclaration)) 1358 Current.Type = TT_FunctionAnnotationRParen; 1359 } 1360 } 1361 } else if (Current.is(tok::at) && Current.Next && 1362 Style.Language != FormatStyle::LK_JavaScript && 1363 Style.Language != FormatStyle::LK_Java) { 1364 // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it 1365 // marks declarations and properties that need special formatting. 1366 switch (Current.Next->Tok.getObjCKeywordID()) { 1367 case tok::objc_interface: 1368 case tok::objc_implementation: 1369 case tok::objc_protocol: 1370 Current.Type = TT_ObjCDecl; 1371 break; 1372 case tok::objc_property: 1373 Current.Type = TT_ObjCProperty; 1374 break; 1375 default: 1376 break; 1377 } 1378 } else if (Current.is(tok::period)) { 1379 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 1380 if (PreviousNoComment && 1381 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 1382 Current.Type = TT_DesignatedInitializerPeriod; 1383 else if (Style.Language == FormatStyle::LK_Java && Current.Previous && 1384 Current.Previous->isOneOf(TT_JavaAnnotation, 1385 TT_LeadingJavaAnnotation)) { 1386 Current.Type = Current.Previous->Type; 1387 } 1388 } else if (canBeObjCSelectorComponent(Current) && 1389 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen. 1390 Current.Previous && Current.Previous->is(TT_CastRParen) && 1391 Current.Previous->MatchingParen && 1392 Current.Previous->MatchingParen->Previous && 1393 Current.Previous->MatchingParen->Previous->is( 1394 TT_ObjCMethodSpecifier)) { 1395 // This is the first part of an Objective-C selector name. (If there's no 1396 // colon after this, this is the only place which annotates the identifier 1397 // as a selector.) 1398 Current.Type = TT_SelectorName; 1399 } else if (Current.isOneOf(tok::identifier, tok::kw_const) && 1400 Current.Previous && 1401 !Current.Previous->isOneOf(tok::equal, tok::at) && 1402 Line.MightBeFunctionDecl && Contexts.size() == 1) { 1403 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1404 // function declaration have been found. 1405 Current.Type = TT_TrailingAnnotation; 1406 } else if ((Style.Language == FormatStyle::LK_Java || 1407 Style.Language == FormatStyle::LK_JavaScript) && 1408 Current.Previous) { 1409 if (Current.Previous->is(tok::at) && 1410 Current.isNot(Keywords.kw_interface)) { 1411 const FormatToken &AtToken = *Current.Previous; 1412 const FormatToken *Previous = AtToken.getPreviousNonComment(); 1413 if (!Previous || Previous->is(TT_LeadingJavaAnnotation)) 1414 Current.Type = TT_LeadingJavaAnnotation; 1415 else 1416 Current.Type = TT_JavaAnnotation; 1417 } else if (Current.Previous->is(tok::period) && 1418 Current.Previous->isOneOf(TT_JavaAnnotation, 1419 TT_LeadingJavaAnnotation)) { 1420 Current.Type = Current.Previous->Type; 1421 } 1422 } 1423 } 1424 1425 /// Take a guess at whether \p Tok starts a name of a function or 1426 /// variable declaration. 1427 /// 1428 /// This is a heuristic based on whether \p Tok is an identifier following 1429 /// something that is likely a type. 1430 bool isStartOfName(const FormatToken &Tok) { 1431 if (Tok.isNot(tok::identifier) || !Tok.Previous) 1432 return false; 1433 1434 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof, 1435 Keywords.kw_as)) 1436 return false; 1437 if (Style.Language == FormatStyle::LK_JavaScript && 1438 Tok.Previous->is(Keywords.kw_in)) 1439 return false; 1440 1441 // Skip "const" as it does not have an influence on whether this is a name. 1442 FormatToken *PreviousNotConst = Tok.getPreviousNonComment(); 1443 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 1444 PreviousNotConst = PreviousNotConst->getPreviousNonComment(); 1445 1446 if (!PreviousNotConst) 1447 return false; 1448 1449 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 1450 PreviousNotConst->Previous && 1451 PreviousNotConst->Previous->is(tok::hash); 1452 1453 if (PreviousNotConst->is(TT_TemplateCloser)) 1454 return PreviousNotConst && PreviousNotConst->MatchingParen && 1455 PreviousNotConst->MatchingParen->Previous && 1456 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) && 1457 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 1458 1459 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen && 1460 PreviousNotConst->MatchingParen->Previous && 1461 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype)) 1462 return true; 1463 1464 return (!IsPPKeyword && 1465 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) || 1466 PreviousNotConst->is(TT_PointerOrReference) || 1467 PreviousNotConst->isSimpleTypeSpecifier(); 1468 } 1469 1470 /// Determine whether ')' is ending a cast. 1471 bool rParenEndsCast(const FormatToken &Tok) { 1472 // C-style casts are only used in C++ and Java. 1473 if (!Style.isCpp() && Style.Language != FormatStyle::LK_Java) 1474 return false; 1475 1476 // Empty parens aren't casts and there are no casts at the end of the line. 1477 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen) 1478 return false; 1479 1480 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 1481 if (LeftOfParens) { 1482 // If there is a closing parenthesis left of the current parentheses, 1483 // look past it as these might be chained casts. 1484 if (LeftOfParens->is(tok::r_paren)) { 1485 if (!LeftOfParens->MatchingParen || 1486 !LeftOfParens->MatchingParen->Previous) 1487 return false; 1488 LeftOfParens = LeftOfParens->MatchingParen->Previous; 1489 } 1490 1491 // If there is an identifier (or with a few exceptions a keyword) right 1492 // before the parentheses, this is unlikely to be a cast. 1493 if (LeftOfParens->Tok.getIdentifierInfo() && 1494 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case, 1495 tok::kw_delete)) 1496 return false; 1497 1498 // Certain other tokens right before the parentheses are also signals that 1499 // this cannot be a cast. 1500 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator, 1501 TT_TemplateCloser, tok::ellipsis)) 1502 return false; 1503 } 1504 1505 if (Tok.Next->is(tok::question)) 1506 return false; 1507 1508 // As Java has no function types, a "(" after the ")" likely means that this 1509 // is a cast. 1510 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren)) 1511 return true; 1512 1513 // If a (non-string) literal follows, this is likely a cast. 1514 if (Tok.Next->isNot(tok::string_literal) && 1515 (Tok.Next->Tok.isLiteral() || 1516 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 1517 return true; 1518 1519 // Heuristically try to determine whether the parentheses contain a type. 1520 bool ParensAreType = 1521 !Tok.Previous || 1522 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) || 1523 Tok.Previous->isSimpleTypeSpecifier(); 1524 bool ParensCouldEndDecl = 1525 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater); 1526 if (ParensAreType && !ParensCouldEndDecl) 1527 return true; 1528 1529 // At this point, we heuristically assume that there are no casts at the 1530 // start of the line. We assume that we have found most cases where there 1531 // are by the logic above, e.g. "(void)x;". 1532 if (!LeftOfParens) 1533 return false; 1534 1535 // Certain token types inside the parentheses mean that this can't be a 1536 // cast. 1537 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok; 1538 Token = Token->Next) 1539 if (Token->is(TT_BinaryOperator)) 1540 return false; 1541 1542 // If the following token is an identifier or 'this', this is a cast. All 1543 // cases where this can be something else are handled above. 1544 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this)) 1545 return true; 1546 1547 if (!Tok.Next->Next) 1548 return false; 1549 1550 // If the next token after the parenthesis is a unary operator, assume 1551 // that this is cast, unless there are unexpected tokens inside the 1552 // parenthesis. 1553 bool NextIsUnary = 1554 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star); 1555 if (!NextIsUnary || Tok.Next->is(tok::plus) || 1556 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) 1557 return false; 1558 // Search for unexpected tokens. 1559 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen; 1560 Prev = Prev->Previous) { 1561 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon)) 1562 return false; 1563 } 1564 return true; 1565 } 1566 1567 /// Return the type of the given token assuming it is * or &. 1568 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 1569 bool InTemplateArgument) { 1570 if (Style.Language == FormatStyle::LK_JavaScript) 1571 return TT_BinaryOperator; 1572 1573 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1574 if (!PrevToken) 1575 return TT_UnaryOperator; 1576 1577 const FormatToken *NextToken = Tok.getNextNonComment(); 1578 if (!NextToken || 1579 NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_const) || 1580 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) 1581 return TT_PointerOrReference; 1582 1583 if (PrevToken->is(tok::coloncolon)) 1584 return TT_PointerOrReference; 1585 1586 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 1587 tok::comma, tok::semi, tok::kw_return, tok::colon, 1588 tok::equal, tok::kw_delete, tok::kw_sizeof, 1589 tok::kw_throw) || 1590 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr, 1591 TT_UnaryOperator, TT_CastRParen)) 1592 return TT_UnaryOperator; 1593 1594 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare)) 1595 return TT_PointerOrReference; 1596 if (NextToken->is(tok::kw_operator) && !IsExpression) 1597 return TT_PointerOrReference; 1598 if (NextToken->isOneOf(tok::comma, tok::semi)) 1599 return TT_PointerOrReference; 1600 1601 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen) { 1602 FormatToken *TokenBeforeMatchingParen = 1603 PrevToken->MatchingParen->getPreviousNonComment(); 1604 if (TokenBeforeMatchingParen && 1605 TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype)) 1606 return TT_PointerOrReference; 1607 } 1608 1609 if (PrevToken->Tok.isLiteral() || 1610 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 1611 tok::kw_false, tok::r_brace) || 1612 NextToken->Tok.isLiteral() || 1613 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 1614 NextToken->isUnaryOperator() || 1615 // If we know we're in a template argument, there are no named 1616 // declarations. Thus, having an identifier on the right-hand side 1617 // indicates a binary operator. 1618 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 1619 return TT_BinaryOperator; 1620 1621 // "&&(" is quite unlikely to be two successive unary "&". 1622 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren)) 1623 return TT_BinaryOperator; 1624 1625 // This catches some cases where evaluation order is used as control flow: 1626 // aaa && aaa->f(); 1627 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 1628 if (NextNextToken && NextNextToken->is(tok::arrow)) 1629 return TT_BinaryOperator; 1630 1631 // It is very unlikely that we are going to find a pointer or reference type 1632 // definition on the RHS of an assignment. 1633 if (IsExpression && !Contexts.back().CaretFound) 1634 return TT_BinaryOperator; 1635 1636 return TT_PointerOrReference; 1637 } 1638 1639 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 1640 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1641 if (!PrevToken) 1642 return TT_UnaryOperator; 1643 1644 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator)) 1645 // This must be a sequence of leading unary operators. 1646 return TT_UnaryOperator; 1647 1648 // Use heuristics to recognize unary operators. 1649 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 1650 tok::question, tok::colon, tok::kw_return, 1651 tok::kw_case, tok::at, tok::l_brace)) 1652 return TT_UnaryOperator; 1653 1654 // There can't be two consecutive binary operators. 1655 if (PrevToken->is(TT_BinaryOperator)) 1656 return TT_UnaryOperator; 1657 1658 // Fall back to marking the token as binary operator. 1659 return TT_BinaryOperator; 1660 } 1661 1662 /// Determine whether ++/-- are pre- or post-increments/-decrements. 1663 TokenType determineIncrementUsage(const FormatToken &Tok) { 1664 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1665 if (!PrevToken || PrevToken->is(TT_CastRParen)) 1666 return TT_UnaryOperator; 1667 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 1668 return TT_TrailingUnaryOperator; 1669 1670 return TT_UnaryOperator; 1671 } 1672 1673 SmallVector<Context, 8> Contexts; 1674 1675 const FormatStyle &Style; 1676 AnnotatedLine &Line; 1677 FormatToken *CurrentToken; 1678 bool AutoFound; 1679 const AdditionalKeywords &Keywords; 1680 1681 // Set of "<" tokens that do not open a template parameter list. If parseAngle 1682 // determines that a specific token can't be a template opener, it will make 1683 // same decision irrespective of the decisions for tokens leading up to it. 1684 // Store this information to prevent this from causing exponential runtime. 1685 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess; 1686 }; 1687 1688 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1; 1689 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 1690 1691 /// Parses binary expressions by inserting fake parenthesis based on 1692 /// operator precedence. 1693 class ExpressionParser { 1694 public: 1695 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords, 1696 AnnotatedLine &Line) 1697 : Style(Style), Keywords(Keywords), Current(Line.First) {} 1698 1699 /// Parse expressions with the given operator precedence. 1700 void parse(int Precedence = 0) { 1701 // Skip 'return' and ObjC selector colons as they are not part of a binary 1702 // expression. 1703 while (Current && (Current->is(tok::kw_return) || 1704 (Current->is(tok::colon) && 1705 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) 1706 next(); 1707 1708 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1709 return; 1710 1711 // Conditional expressions need to be parsed separately for proper nesting. 1712 if (Precedence == prec::Conditional) { 1713 parseConditionalExpr(); 1714 return; 1715 } 1716 1717 // Parse unary operators, which all have a higher precedence than binary 1718 // operators. 1719 if (Precedence == PrecedenceUnaryOperator) { 1720 parseUnaryOperator(); 1721 return; 1722 } 1723 1724 FormatToken *Start = Current; 1725 FormatToken *LatestOperator = nullptr; 1726 unsigned OperatorIndex = 0; 1727 1728 while (Current) { 1729 // Consume operators with higher precedence. 1730 parse(Precedence + 1); 1731 1732 int CurrentPrecedence = getCurrentPrecedence(); 1733 1734 if (Current && Current->is(TT_SelectorName) && 1735 Precedence == CurrentPrecedence) { 1736 if (LatestOperator) 1737 addFakeParenthesis(Start, prec::Level(Precedence)); 1738 Start = Current; 1739 } 1740 1741 // At the end of the line or when an operator with higher precedence is 1742 // found, insert fake parenthesis and return. 1743 if (!Current || 1744 (Current->closesScope() && 1745 (Current->MatchingParen || Current->is(TT_TemplateString))) || 1746 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) || 1747 (CurrentPrecedence == prec::Conditional && 1748 Precedence == prec::Assignment && Current->is(tok::colon))) { 1749 break; 1750 } 1751 1752 // Consume scopes: (), [], <> and {} 1753 if (Current->opensScope()) { 1754 // In fragment of a JavaScript template string can look like '}..${' and 1755 // thus close a scope and open a new one at the same time. 1756 while (Current && (!Current->closesScope() || Current->opensScope())) { 1757 next(); 1758 parse(); 1759 } 1760 next(); 1761 } else { 1762 // Operator found. 1763 if (CurrentPrecedence == Precedence) { 1764 if (LatestOperator) 1765 LatestOperator->NextOperator = Current; 1766 LatestOperator = Current; 1767 Current->OperatorIndex = OperatorIndex; 1768 ++OperatorIndex; 1769 } 1770 next(/*SkipPastLeadingComments=*/Precedence > 0); 1771 } 1772 } 1773 1774 if (LatestOperator && (Current || Precedence > 0)) { 1775 // LatestOperator->LastOperator = true; 1776 if (Precedence == PrecedenceArrowAndPeriod) { 1777 // Call expressions don't have a binary operator precedence. 1778 addFakeParenthesis(Start, prec::Unknown); 1779 } else { 1780 addFakeParenthesis(Start, prec::Level(Precedence)); 1781 } 1782 } 1783 } 1784 1785 private: 1786 /// Gets the precedence (+1) of the given token for binary operators 1787 /// and other tokens that we treat like binary operators. 1788 int getCurrentPrecedence() { 1789 if (Current) { 1790 const FormatToken *NextNonComment = Current->getNextNonComment(); 1791 if (Current->is(TT_ConditionalExpr)) 1792 return prec::Conditional; 1793 if (NextNonComment && Current->is(TT_SelectorName) && 1794 (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) || 1795 ((Style.Language == FormatStyle::LK_Proto || 1796 Style.Language == FormatStyle::LK_TextProto) && 1797 NextNonComment->is(tok::less)))) 1798 return prec::Assignment; 1799 if (Current->is(TT_JsComputedPropertyName)) 1800 return prec::Assignment; 1801 if (Current->is(TT_LambdaArrow)) 1802 return prec::Comma; 1803 if (Current->is(TT_JsFatArrow)) 1804 return prec::Assignment; 1805 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) || 1806 (Current->is(tok::comment) && NextNonComment && 1807 NextNonComment->is(TT_SelectorName))) 1808 return 0; 1809 if (Current->is(TT_RangeBasedForLoopColon)) 1810 return prec::Comma; 1811 if ((Style.Language == FormatStyle::LK_Java || 1812 Style.Language == FormatStyle::LK_JavaScript) && 1813 Current->is(Keywords.kw_instanceof)) 1814 return prec::Relational; 1815 if (Style.Language == FormatStyle::LK_JavaScript && 1816 Current->isOneOf(Keywords.kw_in, Keywords.kw_as)) 1817 return prec::Relational; 1818 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma)) 1819 return Current->getPrecedence(); 1820 if (Current->isOneOf(tok::period, tok::arrow)) 1821 return PrecedenceArrowAndPeriod; 1822 if ((Style.Language == FormatStyle::LK_Java || 1823 Style.Language == FormatStyle::LK_JavaScript) && 1824 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements, 1825 Keywords.kw_throws)) 1826 return 0; 1827 } 1828 return -1; 1829 } 1830 1831 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1832 Start->FakeLParens.push_back(Precedence); 1833 if (Precedence > prec::Unknown) 1834 Start->StartsBinaryExpression = true; 1835 if (Current) { 1836 FormatToken *Previous = Current->Previous; 1837 while (Previous->is(tok::comment) && Previous->Previous) 1838 Previous = Previous->Previous; 1839 ++Previous->FakeRParens; 1840 if (Precedence > prec::Unknown) 1841 Previous->EndsBinaryExpression = true; 1842 } 1843 } 1844 1845 /// Parse unary operator expressions and surround them with fake 1846 /// parentheses if appropriate. 1847 void parseUnaryOperator() { 1848 llvm::SmallVector<FormatToken *, 2> Tokens; 1849 while (Current && Current->is(TT_UnaryOperator)) { 1850 Tokens.push_back(Current); 1851 next(); 1852 } 1853 parse(PrecedenceArrowAndPeriod); 1854 for (FormatToken *Token : llvm::reverse(Tokens)) 1855 // The actual precedence doesn't matter. 1856 addFakeParenthesis(Token, prec::Unknown); 1857 } 1858 1859 void parseConditionalExpr() { 1860 while (Current && Current->isTrailingComment()) { 1861 next(); 1862 } 1863 FormatToken *Start = Current; 1864 parse(prec::LogicalOr); 1865 if (!Current || !Current->is(tok::question)) 1866 return; 1867 next(); 1868 parse(prec::Assignment); 1869 if (!Current || Current->isNot(TT_ConditionalExpr)) 1870 return; 1871 next(); 1872 parse(prec::Assignment); 1873 addFakeParenthesis(Start, prec::Conditional); 1874 } 1875 1876 void next(bool SkipPastLeadingComments = true) { 1877 if (Current) 1878 Current = Current->Next; 1879 while (Current && 1880 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 1881 Current->isTrailingComment()) 1882 Current = Current->Next; 1883 } 1884 1885 const FormatStyle &Style; 1886 const AdditionalKeywords &Keywords; 1887 FormatToken *Current; 1888 }; 1889 1890 } // end anonymous namespace 1891 1892 void TokenAnnotator::setCommentLineLevels( 1893 SmallVectorImpl<AnnotatedLine *> &Lines) { 1894 const AnnotatedLine *NextNonCommentLine = nullptr; 1895 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 1896 E = Lines.rend(); 1897 I != E; ++I) { 1898 bool CommentLine = true; 1899 for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) { 1900 if (!Tok->is(tok::comment)) { 1901 CommentLine = false; 1902 break; 1903 } 1904 } 1905 1906 // If the comment is currently aligned with the line immediately following 1907 // it, that's probably intentional and we should keep it. 1908 if (NextNonCommentLine && CommentLine && 1909 NextNonCommentLine->First->NewlinesBefore <= 1 && 1910 NextNonCommentLine->First->OriginalColumn == 1911 (*I)->First->OriginalColumn) { 1912 // Align comments for preprocessor lines with the # in column 0. 1913 // Otherwise, align with the next line. 1914 (*I)->Level = (NextNonCommentLine->Type == LT_PreprocessorDirective || 1915 NextNonCommentLine->Type == LT_ImportStatement) 1916 ? 0 1917 : NextNonCommentLine->Level; 1918 } else { 1919 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 1920 } 1921 1922 setCommentLineLevels((*I)->Children); 1923 } 1924 } 1925 1926 static unsigned maxNestingDepth(const AnnotatedLine &Line) { 1927 unsigned Result = 0; 1928 for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next) 1929 Result = std::max(Result, Tok->NestingLevel); 1930 return Result; 1931 } 1932 1933 void TokenAnnotator::annotate(AnnotatedLine &Line) { 1934 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1935 E = Line.Children.end(); 1936 I != E; ++I) { 1937 annotate(**I); 1938 } 1939 AnnotatingParser Parser(Style, Line, Keywords); 1940 Line.Type = Parser.parseLine(); 1941 1942 // With very deep nesting, ExpressionParser uses lots of stack and the 1943 // formatting algorithm is very slow. We're not going to do a good job here 1944 // anyway - it's probably generated code being formatted by mistake. 1945 // Just skip the whole line. 1946 if (maxNestingDepth(Line) > 50) 1947 Line.Type = LT_Invalid; 1948 1949 if (Line.Type == LT_Invalid) 1950 return; 1951 1952 ExpressionParser ExprParser(Style, Keywords, Line); 1953 ExprParser.parse(); 1954 1955 if (Line.startsWith(TT_ObjCMethodSpecifier)) 1956 Line.Type = LT_ObjCMethodDecl; 1957 else if (Line.startsWith(TT_ObjCDecl)) 1958 Line.Type = LT_ObjCDecl; 1959 else if (Line.startsWith(TT_ObjCProperty)) 1960 Line.Type = LT_ObjCProperty; 1961 1962 Line.First->SpacesRequiredBefore = 1; 1963 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 1964 } 1965 1966 // This function heuristically determines whether 'Current' starts the name of a 1967 // function declaration. 1968 static bool isFunctionDeclarationName(const FormatToken &Current, 1969 const AnnotatedLine &Line) { 1970 auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * { 1971 for (; Next; Next = Next->Next) { 1972 if (Next->is(TT_OverloadedOperatorLParen)) 1973 return Next; 1974 if (Next->is(TT_OverloadedOperator)) 1975 continue; 1976 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) { 1977 // For 'new[]' and 'delete[]'. 1978 if (Next->Next && Next->Next->is(tok::l_square) && Next->Next->Next && 1979 Next->Next->Next->is(tok::r_square)) 1980 Next = Next->Next->Next; 1981 continue; 1982 } 1983 1984 break; 1985 } 1986 return nullptr; 1987 }; 1988 1989 // Find parentheses of parameter list. 1990 const FormatToken *Next = Current.Next; 1991 if (Current.is(tok::kw_operator)) { 1992 if (Current.Previous && Current.Previous->is(tok::coloncolon)) 1993 return false; 1994 Next = skipOperatorName(Next); 1995 } else { 1996 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0) 1997 return false; 1998 for (; Next; Next = Next->Next) { 1999 if (Next->is(TT_TemplateOpener)) { 2000 Next = Next->MatchingParen; 2001 } else if (Next->is(tok::coloncolon)) { 2002 Next = Next->Next; 2003 if (!Next) 2004 return false; 2005 if (Next->is(tok::kw_operator)) { 2006 Next = skipOperatorName(Next->Next); 2007 break; 2008 } 2009 if (!Next->is(tok::identifier)) 2010 return false; 2011 } else if (Next->is(tok::l_paren)) { 2012 break; 2013 } else { 2014 return false; 2015 } 2016 } 2017 } 2018 2019 // Check whether parameter list can belong to a function declaration. 2020 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen) 2021 return false; 2022 // If the lines ends with "{", this is likely an function definition. 2023 if (Line.Last->is(tok::l_brace)) 2024 return true; 2025 if (Next->Next == Next->MatchingParen) 2026 return true; // Empty parentheses. 2027 // If there is an &/&& after the r_paren, this is likely a function. 2028 if (Next->MatchingParen->Next && 2029 Next->MatchingParen->Next->is(TT_PointerOrReference)) 2030 return true; 2031 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen; 2032 Tok = Tok->Next) { 2033 if (Tok->is(tok::l_paren) && Tok->MatchingParen) { 2034 Tok = Tok->MatchingParen; 2035 continue; 2036 } 2037 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 2038 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis)) 2039 return true; 2040 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) || 2041 Tok->Tok.isLiteral()) 2042 return false; 2043 } 2044 return false; 2045 } 2046 2047 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const { 2048 assert(Line.MightBeFunctionDecl); 2049 2050 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel || 2051 Style.AlwaysBreakAfterReturnType == 2052 FormatStyle::RTBS_TopLevelDefinitions) && 2053 Line.Level > 0) 2054 return false; 2055 2056 switch (Style.AlwaysBreakAfterReturnType) { 2057 case FormatStyle::RTBS_None: 2058 return false; 2059 case FormatStyle::RTBS_All: 2060 case FormatStyle::RTBS_TopLevel: 2061 return true; 2062 case FormatStyle::RTBS_AllDefinitions: 2063 case FormatStyle::RTBS_TopLevelDefinitions: 2064 return Line.mightBeFunctionDefinition(); 2065 } 2066 2067 return false; 2068 } 2069 2070 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 2071 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 2072 E = Line.Children.end(); 2073 I != E; ++I) { 2074 calculateFormattingInformation(**I); 2075 } 2076 2077 Line.First->TotalLength = 2078 Line.First->IsMultiline ? Style.ColumnLimit 2079 : Line.FirstStartColumn + Line.First->ColumnWidth; 2080 FormatToken *Current = Line.First->Next; 2081 bool InFunctionDecl = Line.MightBeFunctionDecl; 2082 while (Current) { 2083 if (isFunctionDeclarationName(*Current, Line)) 2084 Current->Type = TT_FunctionDeclarationName; 2085 if (Current->is(TT_LineComment)) { 2086 if (Current->Previous->BlockKind == BK_BracedInit && 2087 Current->Previous->opensScope()) 2088 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1; 2089 else 2090 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 2091 2092 // If we find a trailing comment, iterate backwards to determine whether 2093 // it seems to relate to a specific parameter. If so, break before that 2094 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 2095 // to the previous line in: 2096 // SomeFunction(a, 2097 // b, // comment 2098 // c); 2099 if (!Current->HasUnescapedNewline) { 2100 for (FormatToken *Parameter = Current->Previous; Parameter; 2101 Parameter = Parameter->Previous) { 2102 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 2103 break; 2104 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 2105 if (!Parameter->Previous->is(TT_CtorInitializerComma) && 2106 Parameter->HasUnescapedNewline) 2107 Parameter->MustBreakBefore = true; 2108 break; 2109 } 2110 } 2111 } 2112 } else if (Current->SpacesRequiredBefore == 0 && 2113 spaceRequiredBefore(Line, *Current)) { 2114 Current->SpacesRequiredBefore = 1; 2115 } 2116 2117 Current->MustBreakBefore = 2118 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 2119 2120 if (!Current->MustBreakBefore && InFunctionDecl && 2121 Current->is(TT_FunctionDeclarationName)) 2122 Current->MustBreakBefore = mustBreakForReturnType(Line); 2123 2124 Current->CanBreakBefore = 2125 Current->MustBreakBefore || canBreakBefore(Line, *Current); 2126 unsigned ChildSize = 0; 2127 if (Current->Previous->Children.size() == 1) { 2128 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 2129 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 2130 : LastOfChild.TotalLength + 1; 2131 } 2132 const FormatToken *Prev = Current->Previous; 2133 if (Current->MustBreakBefore || Prev->Children.size() > 1 || 2134 (Prev->Children.size() == 1 && 2135 Prev->Children[0]->First->MustBreakBefore) || 2136 Current->IsMultiline) 2137 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit; 2138 else 2139 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth + 2140 ChildSize + Current->SpacesRequiredBefore; 2141 2142 if (Current->is(TT_CtorInitializerColon)) 2143 InFunctionDecl = false; 2144 2145 // FIXME: Only calculate this if CanBreakBefore is true once static 2146 // initializers etc. are sorted out. 2147 // FIXME: Move magic numbers to a better place. 2148 2149 // Reduce penalty for aligning ObjC method arguments using the colon 2150 // alignment as this is the canonical way (still prefer fitting everything 2151 // into one line if possible). Trying to fit a whole expression into one 2152 // line should not force other line breaks (e.g. when ObjC method 2153 // expression is a part of other expression). 2154 Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl); 2155 if (Style.Language == FormatStyle::LK_ObjC && 2156 Current->is(TT_SelectorName) && Current->ParameterIndex > 0) { 2157 if (Current->ParameterIndex == 1) 2158 Current->SplitPenalty += 5 * Current->BindingStrength; 2159 } else { 2160 Current->SplitPenalty += 20 * Current->BindingStrength; 2161 } 2162 2163 Current = Current->Next; 2164 } 2165 2166 calculateUnbreakableTailLengths(Line); 2167 unsigned IndentLevel = Line.Level; 2168 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 2169 if (Current->Role) 2170 Current->Role->precomputeFormattingInfos(Current); 2171 if (Current->MatchingParen && 2172 Current->MatchingParen->opensBlockOrBlockTypeList(Style)) { 2173 assert(IndentLevel > 0); 2174 --IndentLevel; 2175 } 2176 Current->IndentLevel = IndentLevel; 2177 if (Current->opensBlockOrBlockTypeList(Style)) 2178 ++IndentLevel; 2179 } 2180 2181 LLVM_DEBUG({ printDebugInfo(Line); }); 2182 } 2183 2184 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 2185 unsigned UnbreakableTailLength = 0; 2186 FormatToken *Current = Line.Last; 2187 while (Current) { 2188 Current->UnbreakableTailLength = UnbreakableTailLength; 2189 if (Current->CanBreakBefore || 2190 Current->isOneOf(tok::comment, tok::string_literal)) { 2191 UnbreakableTailLength = 0; 2192 } else { 2193 UnbreakableTailLength += 2194 Current->ColumnWidth + Current->SpacesRequiredBefore; 2195 } 2196 Current = Current->Previous; 2197 } 2198 } 2199 2200 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 2201 const FormatToken &Tok, 2202 bool InFunctionDecl) { 2203 const FormatToken &Left = *Tok.Previous; 2204 const FormatToken &Right = Tok; 2205 2206 if (Left.is(tok::semi)) 2207 return 0; 2208 2209 if (Style.Language == FormatStyle::LK_Java) { 2210 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws)) 2211 return 1; 2212 if (Right.is(Keywords.kw_implements)) 2213 return 2; 2214 if (Left.is(tok::comma) && Left.NestingLevel == 0) 2215 return 3; 2216 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2217 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma)) 2218 return 100; 2219 if (Left.is(TT_JsTypeColon)) 2220 return 35; 2221 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2222 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2223 return 100; 2224 // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()". 2225 if (Left.opensScope() && Right.closesScope()) 2226 return 200; 2227 } 2228 2229 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2230 return 1; 2231 if (Right.is(tok::l_square)) { 2232 if (Style.Language == FormatStyle::LK_Proto) 2233 return 1; 2234 if (Left.is(tok::r_square)) 2235 return 200; 2236 // Slightly prefer formatting local lambda definitions like functions. 2237 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal)) 2238 return 35; 2239 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2240 TT_ArrayInitializerLSquare, 2241 TT_DesignatedInitializerLSquare, TT_AttributeSquare)) 2242 return 500; 2243 } 2244 2245 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2246 Right.is(tok::kw_operator)) { 2247 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 2248 return 3; 2249 if (Left.is(TT_StartOfName)) 2250 return 110; 2251 if (InFunctionDecl && Right.NestingLevel == 0) 2252 return Style.PenaltyReturnTypeOnItsOwnLine; 2253 return 200; 2254 } 2255 if (Right.is(TT_PointerOrReference)) 2256 return 190; 2257 if (Right.is(TT_LambdaArrow)) 2258 return 110; 2259 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 2260 return 160; 2261 if (Left.is(TT_CastRParen)) 2262 return 100; 2263 if (Left.is(tok::coloncolon) || 2264 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 2265 return 500; 2266 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 2267 return 5000; 2268 if (Left.is(tok::comment)) 2269 return 1000; 2270 2271 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon, 2272 TT_CtorInitializerColon)) 2273 return 2; 2274 2275 if (Right.isMemberAccess()) { 2276 // Breaking before the "./->" of a chained call/member access is reasonably 2277 // cheap, as formatting those with one call per line is generally 2278 // desirable. In particular, it should be cheaper to break before the call 2279 // than it is to break inside a call's parameters, which could lead to weird 2280 // "hanging" indents. The exception is the very last "./->" to support this 2281 // frequent pattern: 2282 // 2283 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc( 2284 // dddddddd); 2285 // 2286 // which might otherwise be blown up onto many lines. Here, clang-format 2287 // won't produce "hanging" indents anyway as there is no other trailing 2288 // call. 2289 // 2290 // Also apply higher penalty is not a call as that might lead to a wrapping 2291 // like: 2292 // 2293 // aaaaaaa 2294 // .aaaaaaaaa.bbbbbbbb(cccccccc); 2295 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope() 2296 ? 150 2297 : 35; 2298 } 2299 2300 if (Right.is(TT_TrailingAnnotation) && 2301 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 2302 // Moving trailing annotations to the next line is fine for ObjC method 2303 // declarations. 2304 if (Line.startsWith(TT_ObjCMethodSpecifier)) 2305 return 10; 2306 // Generally, breaking before a trailing annotation is bad unless it is 2307 // function-like. It seems to be especially preferable to keep standard 2308 // annotations (i.e. "const", "final" and "override") on the same line. 2309 // Use a slightly higher penalty after ")" so that annotations like 2310 // "const override" are kept together. 2311 bool is_short_annotation = Right.TokenText.size() < 10; 2312 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 2313 } 2314 2315 // In for-loops, prefer breaking at ',' and ';'. 2316 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal)) 2317 return 4; 2318 2319 // In Objective-C method expressions, prefer breaking before "param:" over 2320 // breaking after it. 2321 if (Right.is(TT_SelectorName)) 2322 return 0; 2323 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr)) 2324 return Line.MightBeFunctionDecl ? 50 : 500; 2325 2326 // In Objective-C type declarations, avoid breaking after the category's 2327 // open paren (we'll prefer breaking after the protocol list's opening 2328 // angle bracket, if present). 2329 if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous && 2330 Left.Previous->isOneOf(tok::identifier, tok::greater)) 2331 return 500; 2332 2333 if (Left.is(tok::l_paren) && InFunctionDecl && 2334 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) 2335 return 100; 2336 if (Left.is(tok::l_paren) && Left.Previous && 2337 (Left.Previous->isOneOf(tok::kw_if, tok::kw_for) || 2338 Left.Previous->endsSequence(tok::kw_constexpr, tok::kw_if))) 2339 return 1000; 2340 if (Left.is(tok::equal) && InFunctionDecl) 2341 return 110; 2342 if (Right.is(tok::r_brace)) 2343 return 1; 2344 if (Left.is(TT_TemplateOpener)) 2345 return 100; 2346 if (Left.opensScope()) { 2347 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign) 2348 return 0; 2349 if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle) 2350 return 19; 2351 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 2352 : 19; 2353 } 2354 if (Left.is(TT_JavaAnnotation)) 2355 return 50; 2356 2357 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous && 2358 Left.Previous->isLabelString() && 2359 (Left.NextOperator || Left.OperatorIndex != 0)) 2360 return 50; 2361 if (Right.is(tok::plus) && Left.isLabelString() && 2362 (Right.NextOperator || Right.OperatorIndex != 0)) 2363 return 25; 2364 if (Left.is(tok::comma)) 2365 return 1; 2366 if (Right.is(tok::lessless) && Left.isLabelString() && 2367 (Right.NextOperator || Right.OperatorIndex != 1)) 2368 return 25; 2369 if (Right.is(tok::lessless)) { 2370 // Breaking at a << is really cheap. 2371 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0) 2372 // Slightly prefer to break before the first one in log-like statements. 2373 return 2; 2374 return 1; 2375 } 2376 if (Left.ClosesTemplateDeclaration) 2377 return Style.PenaltyBreakTemplateDeclaration; 2378 if (Left.is(TT_ConditionalExpr)) 2379 return prec::Conditional; 2380 prec::Level Level = Left.getPrecedence(); 2381 if (Level == prec::Unknown) 2382 Level = Right.getPrecedence(); 2383 if (Level == prec::Assignment) 2384 return Style.PenaltyBreakAssignment; 2385 if (Level != prec::Unknown) 2386 return Level; 2387 2388 return 3; 2389 } 2390 2391 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 2392 const FormatToken &Left, 2393 const FormatToken &Right) { 2394 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 2395 return true; 2396 if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java) 2397 return true; 2398 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 2399 Left.Tok.getObjCKeywordID() == tok::objc_property) 2400 return true; 2401 if (Right.is(tok::hashhash)) 2402 return Left.is(tok::hash); 2403 if (Left.isOneOf(tok::hashhash, tok::hash)) 2404 return Right.is(tok::hash); 2405 if (Left.is(tok::l_paren) && Right.is(tok::r_paren)) 2406 return Style.SpaceInEmptyParentheses; 2407 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 2408 return (Right.is(TT_CastRParen) || 2409 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen))) 2410 ? Style.SpacesInCStyleCastParentheses 2411 : Style.SpacesInParentheses; 2412 if (Right.isOneOf(tok::semi, tok::comma)) 2413 return false; 2414 if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) { 2415 bool IsLightweightGeneric = 2416 Right.MatchingParen && Right.MatchingParen->Next && 2417 Right.MatchingParen->Next->is(tok::colon); 2418 return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList; 2419 } 2420 if (Right.is(tok::less) && Left.is(tok::kw_template)) 2421 return Style.SpaceAfterTemplateKeyword; 2422 if (Left.isOneOf(tok::exclaim, tok::tilde)) 2423 return false; 2424 if (Left.is(tok::at) && 2425 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 2426 tok::numeric_constant, tok::l_paren, tok::l_brace, 2427 tok::kw_true, tok::kw_false)) 2428 return false; 2429 if (Left.is(tok::colon)) 2430 return !Left.is(TT_ObjCMethodExpr); 2431 if (Left.is(tok::coloncolon)) 2432 return false; 2433 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) { 2434 if (Style.Language == FormatStyle::LK_TextProto || 2435 (Style.Language == FormatStyle::LK_Proto && 2436 (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) { 2437 // Format empty list as `<>`. 2438 if (Left.is(tok::less) && Right.is(tok::greater)) 2439 return false; 2440 return !Style.Cpp11BracedListStyle; 2441 } 2442 return false; 2443 } 2444 if (Right.is(tok::ellipsis)) 2445 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous && 2446 Left.Previous->is(tok::kw_case)); 2447 if (Left.is(tok::l_square) && Right.is(tok::amp)) 2448 return false; 2449 if (Right.is(TT_PointerOrReference)) { 2450 if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) { 2451 if (!Left.MatchingParen) 2452 return true; 2453 FormatToken *TokenBeforeMatchingParen = 2454 Left.MatchingParen->getPreviousNonComment(); 2455 if (!TokenBeforeMatchingParen || 2456 !TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype)) 2457 return true; 2458 } 2459 return (Left.Tok.isLiteral() || 2460 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) && 2461 (Style.PointerAlignment != FormatStyle::PAS_Left || 2462 (Line.IsMultiVariableDeclStmt && 2463 (Left.NestingLevel == 0 || 2464 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for))))))); 2465 } 2466 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) && 2467 (!Left.is(TT_PointerOrReference) || 2468 (Style.PointerAlignment != FormatStyle::PAS_Right && 2469 !Line.IsMultiVariableDeclStmt))) 2470 return true; 2471 if (Left.is(TT_PointerOrReference)) 2472 return Right.Tok.isLiteral() || Right.is(TT_BlockComment) || 2473 (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) && 2474 !Right.is(TT_StartOfName)) || 2475 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) || 2476 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare, 2477 tok::l_paren) && 2478 (Style.PointerAlignment != FormatStyle::PAS_Right && 2479 !Line.IsMultiVariableDeclStmt) && 2480 Left.Previous && 2481 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon)); 2482 if (Right.is(tok::star) && Left.is(tok::l_paren)) 2483 return false; 2484 const auto SpaceRequiredForArrayInitializerLSquare = 2485 [](const FormatToken &LSquareTok, const FormatStyle &Style) { 2486 return Style.SpacesInContainerLiterals || 2487 ((Style.Language == FormatStyle::LK_Proto || 2488 Style.Language == FormatStyle::LK_TextProto) && 2489 !Style.Cpp11BracedListStyle && 2490 LSquareTok.endsSequence(tok::l_square, tok::colon, 2491 TT_SelectorName)); 2492 }; 2493 if (Left.is(tok::l_square)) 2494 return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) && 2495 SpaceRequiredForArrayInitializerLSquare(Left, Style)) || 2496 (Left.isOneOf(TT_ArraySubscriptLSquare, 2497 TT_StructuredBindingLSquare) && 2498 Style.SpacesInSquareBrackets && Right.isNot(tok::r_square)); 2499 if (Right.is(tok::r_square)) 2500 return Right.MatchingParen && 2501 ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) && 2502 SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen, 2503 Style)) || 2504 (Style.SpacesInSquareBrackets && 2505 Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare, 2506 TT_StructuredBindingLSquare)) || 2507 Right.MatchingParen->is(TT_AttributeParen)); 2508 if (Right.is(tok::l_square) && 2509 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2510 TT_DesignatedInitializerLSquare, 2511 TT_StructuredBindingLSquare, TT_AttributeSquare) && 2512 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral)) 2513 return false; 2514 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 2515 return !Left.Children.empty(); // No spaces in "{}". 2516 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 2517 (Right.is(tok::r_brace) && Right.MatchingParen && 2518 Right.MatchingParen->BlockKind != BK_Block)) 2519 return !Style.Cpp11BracedListStyle; 2520 if (Left.is(TT_BlockComment)) 2521 // No whitespace in x(/*foo=*/1), except for JavaScript. 2522 return Style.Language == FormatStyle::LK_JavaScript || 2523 !Left.TokenText.endswith("=*/"); 2524 if (Right.is(tok::l_paren)) { 2525 if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) || 2526 (Left.is(tok::r_square) && Left.is(TT_AttributeSquare))) 2527 return true; 2528 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) || 2529 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 2530 (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while, 2531 tok::kw_switch, tok::kw_case, TT_ForEachMacro, 2532 TT_ObjCForIn) || 2533 Left.endsSequence(tok::kw_constexpr, tok::kw_if) || 2534 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch, 2535 tok::kw_new, tok::kw_delete) && 2536 (!Left.Previous || Left.Previous->isNot(tok::period))))) || 2537 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always && 2538 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() || 2539 Left.is(tok::r_paren)) && 2540 Line.Type != LT_PreprocessorDirective); 2541 } 2542 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 2543 return false; 2544 if (Right.is(TT_UnaryOperator)) 2545 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 2546 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr)); 2547 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 2548 tok::r_paren) || 2549 Left.isSimpleTypeSpecifier()) && 2550 Right.is(tok::l_brace) && Right.getNextNonComment() && 2551 Right.BlockKind != BK_Block) 2552 return false; 2553 if (Left.is(tok::period) || Right.is(tok::period)) 2554 return false; 2555 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 2556 return false; 2557 if (Left.is(TT_TemplateCloser) && Left.MatchingParen && 2558 Left.MatchingParen->Previous && 2559 (Left.MatchingParen->Previous->is(tok::period) || 2560 Left.MatchingParen->Previous->is(tok::coloncolon))) 2561 // Java call to generic function with explicit type: 2562 // A.<B<C<...>>>DoSomething(); 2563 // A::<B<C<...>>>DoSomething(); // With a Java 8 method reference. 2564 return false; 2565 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square)) 2566 return false; 2567 if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at)) 2568 // Objective-C dictionary literal -> no space after opening brace. 2569 return false; 2570 if (Right.is(tok::r_brace) && Right.MatchingParen && 2571 Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at)) 2572 // Objective-C dictionary literal -> no space before closing brace. 2573 return false; 2574 return true; 2575 } 2576 2577 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 2578 const FormatToken &Right) { 2579 const FormatToken &Left = *Right.Previous; 2580 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo()) 2581 return true; // Never ever merge two identifiers. 2582 if (Style.isCpp()) { 2583 if (Left.is(tok::kw_operator)) 2584 return Right.is(tok::coloncolon); 2585 if (Right.is(tok::l_brace) && Right.BlockKind == BK_BracedInit && 2586 !Left.opensScope() && Style.SpaceBeforeCpp11BracedList) 2587 return true; 2588 } else if (Style.Language == FormatStyle::LK_Proto || 2589 Style.Language == FormatStyle::LK_TextProto) { 2590 if (Right.is(tok::period) && 2591 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required, 2592 Keywords.kw_repeated, Keywords.kw_extend)) 2593 return true; 2594 if (Right.is(tok::l_paren) && 2595 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) 2596 return true; 2597 if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName)) 2598 return true; 2599 // Slashes occur in text protocol extension syntax: [type/type] { ... }. 2600 if (Left.is(tok::slash) || Right.is(tok::slash)) 2601 return false; 2602 if (Left.MatchingParen && Left.MatchingParen->is(TT_ProtoExtensionLSquare) && 2603 Right.isOneOf(tok::l_brace, tok::less)) 2604 return !Style.Cpp11BracedListStyle; 2605 // A percent is probably part of a formatting specification, such as %lld. 2606 if (Left.is(tok::percent)) 2607 return false; 2608 // Preserve the existence of a space before a percent for cases like 0x%04x 2609 // and "%d %d" 2610 if (Left.is(tok::numeric_constant) && Right.is(tok::percent)) 2611 return Right.WhitespaceRange.getEnd() != Right.WhitespaceRange.getBegin(); 2612 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2613 if (Left.is(TT_JsFatArrow)) 2614 return true; 2615 // for await ( ... 2616 if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous && 2617 Left.Previous->is(tok::kw_for)) 2618 return true; 2619 if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) && 2620 Right.MatchingParen) { 2621 const FormatToken *Next = Right.MatchingParen->getNextNonComment(); 2622 // An async arrow function, for example: `x = async () => foo();`, 2623 // as opposed to calling a function called async: `x = async();` 2624 if (Next && Next->is(TT_JsFatArrow)) 2625 return true; 2626 } 2627 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2628 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2629 return false; 2630 // In tagged template literals ("html`bar baz`"), there is no space between 2631 // the tag identifier and the template string. getIdentifierInfo makes sure 2632 // that the identifier is not a pseudo keyword like `yield`, either. 2633 if (Left.is(tok::identifier) && Keywords.IsJavaScriptIdentifier(Left) && 2634 Right.is(TT_TemplateString)) 2635 return false; 2636 if (Right.is(tok::star) && 2637 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) 2638 return false; 2639 if (Right.isOneOf(tok::l_brace, tok::l_square) && 2640 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield, 2641 Keywords.kw_extends, Keywords.kw_implements)) 2642 return true; 2643 if (Right.is(tok::l_paren)) { 2644 // JS methods can use some keywords as names (e.g. `delete()`). 2645 if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo()) 2646 return false; 2647 // Valid JS method names can include keywords, e.g. `foo.delete()` or 2648 // `bar.instanceof()`. Recognize call positions by preceding period. 2649 if (Left.Previous && Left.Previous->is(tok::period) && 2650 Left.Tok.getIdentifierInfo()) 2651 return false; 2652 // Additional unary JavaScript operators that need a space after. 2653 if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof, 2654 tok::kw_void)) 2655 return true; 2656 } 2657 if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in, 2658 tok::kw_const) || 2659 // "of" is only a keyword if it appears after another identifier 2660 // (e.g. as "const x of y" in a for loop), or after a destructuring 2661 // operation (const [x, y] of z, const {a, b} of c). 2662 (Left.is(Keywords.kw_of) && Left.Previous && 2663 (Left.Previous->Tok.is(tok::identifier) || 2664 Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) && 2665 (!Left.Previous || !Left.Previous->is(tok::period))) 2666 return true; 2667 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous && 2668 Left.Previous->is(tok::period) && Right.is(tok::l_paren)) 2669 return false; 2670 if (Left.is(Keywords.kw_as) && 2671 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren)) 2672 return true; 2673 if (Left.is(tok::kw_default) && Left.Previous && 2674 Left.Previous->is(tok::kw_export)) 2675 return true; 2676 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace)) 2677 return true; 2678 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion)) 2679 return false; 2680 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator)) 2681 return false; 2682 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) && 2683 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) 2684 return false; 2685 if (Left.is(tok::ellipsis)) 2686 return false; 2687 if (Left.is(TT_TemplateCloser) && 2688 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square, 2689 Keywords.kw_implements, Keywords.kw_extends)) 2690 // Type assertions ('<type>expr') are not followed by whitespace. Other 2691 // locations that should have whitespace following are identified by the 2692 // above set of follower tokens. 2693 return false; 2694 if (Right.is(TT_JsNonNullAssertion)) 2695 return false; 2696 if (Left.is(TT_JsNonNullAssertion) && 2697 Right.isOneOf(Keywords.kw_as, Keywords.kw_in)) 2698 return true; // "x! as string", "x! in y" 2699 } else if (Style.Language == FormatStyle::LK_Java) { 2700 if (Left.is(tok::r_square) && Right.is(tok::l_brace)) 2701 return true; 2702 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) 2703 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never; 2704 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private, 2705 tok::kw_protected) || 2706 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract, 2707 Keywords.kw_native)) && 2708 Right.is(TT_TemplateOpener)) 2709 return true; 2710 } 2711 if (Left.is(TT_ImplicitStringLiteral)) 2712 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2713 if (Line.Type == LT_ObjCMethodDecl) { 2714 if (Left.is(TT_ObjCMethodSpecifier)) 2715 return true; 2716 if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right)) 2717 // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a 2718 // keyword in Objective-C, and '+ (instancetype)new;' is a standard class 2719 // method declaration. 2720 return false; 2721 } 2722 if (Line.Type == LT_ObjCProperty && 2723 (Right.is(tok::equal) || Left.is(tok::equal))) 2724 return false; 2725 2726 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) || 2727 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) 2728 return true; 2729 if (Right.is(TT_OverloadedOperatorLParen)) 2730 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2731 if (Left.is(tok::comma)) 2732 return true; 2733 if (Right.is(tok::comma)) 2734 return false; 2735 if (Right.is(TT_ObjCBlockLParen)) 2736 return true; 2737 if (Right.is(TT_CtorInitializerColon)) 2738 return Style.SpaceBeforeCtorInitializerColon; 2739 if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon) 2740 return false; 2741 if (Right.is(TT_RangeBasedForLoopColon) && 2742 !Style.SpaceBeforeRangeBasedForLoopColon) 2743 return false; 2744 if (Right.is(tok::colon)) { 2745 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) || 2746 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi)) 2747 return false; 2748 if (Right.is(TT_ObjCMethodExpr)) 2749 return false; 2750 if (Left.is(tok::question)) 2751 return false; 2752 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon)) 2753 return false; 2754 if (Right.is(TT_DictLiteral)) 2755 return Style.SpacesInContainerLiterals; 2756 if (Right.is(TT_AttributeColon)) 2757 return false; 2758 return true; 2759 } 2760 if (Left.is(TT_UnaryOperator)) 2761 return Right.is(TT_BinaryOperator); 2762 2763 // If the next token is a binary operator or a selector name, we have 2764 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly. 2765 if (Left.is(TT_CastRParen)) 2766 return Style.SpaceAfterCStyleCast || 2767 Right.isOneOf(TT_BinaryOperator, TT_SelectorName); 2768 2769 if (Left.is(tok::greater) && Right.is(tok::greater)) { 2770 if (Style.Language == FormatStyle::LK_TextProto || 2771 (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral))) 2772 return !Style.Cpp11BracedListStyle; 2773 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) && 2774 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 2775 } 2776 if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) || 2777 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) || 2778 (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod))) 2779 return false; 2780 if (!Style.SpaceBeforeAssignmentOperators && 2781 Right.getPrecedence() == prec::Assignment) 2782 return false; 2783 if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) && 2784 (Left.is(tok::identifier) || Left.is(tok::kw_this))) 2785 return false; 2786 if (Right.is(tok::coloncolon) && Left.is(tok::identifier)) 2787 // Generally don't remove existing spaces between an identifier and "::". 2788 // The identifier might actually be a macro name such as ALWAYS_INLINE. If 2789 // this turns out to be too lenient, add analysis of the identifier itself. 2790 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2791 if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment)) 2792 return (Left.is(TT_TemplateOpener) && 2793 Style.Standard == FormatStyle::LS_Cpp03) || 2794 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square, 2795 tok::kw___super, TT_TemplateCloser, 2796 TT_TemplateOpener)) || 2797 (Left.is(tok ::l_paren) && Style.SpacesInParentheses); 2798 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser))) 2799 return Style.SpacesInAngles; 2800 // Space before TT_StructuredBindingLSquare. 2801 if (Right.is(TT_StructuredBindingLSquare)) 2802 return !Left.isOneOf(tok::amp, tok::ampamp) || 2803 Style.PointerAlignment != FormatStyle::PAS_Right; 2804 // Space before & or && following a TT_StructuredBindingLSquare. 2805 if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) && 2806 Right.isOneOf(tok::amp, tok::ampamp)) 2807 return Style.PointerAlignment != FormatStyle::PAS_Left; 2808 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) || 2809 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && 2810 !Right.is(tok::r_paren))) 2811 return true; 2812 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) && 2813 Right.isNot(TT_FunctionTypeLParen)) 2814 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2815 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) && 2816 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen)) 2817 return false; 2818 if (Right.is(tok::less) && Left.isNot(tok::l_paren) && 2819 Line.startsWith(tok::hash)) 2820 return true; 2821 if (Right.is(TT_TrailingUnaryOperator)) 2822 return false; 2823 if (Left.is(TT_RegexLiteral)) 2824 return false; 2825 return spaceRequiredBetween(Line, Left, Right); 2826 } 2827 2828 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 2829 static bool isAllmanBrace(const FormatToken &Tok) { 2830 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 2831 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral); 2832 } 2833 2834 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 2835 const FormatToken &Right) { 2836 const FormatToken &Left = *Right.Previous; 2837 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0) 2838 return true; 2839 2840 if (Style.Language == FormatStyle::LK_JavaScript) { 2841 // FIXME: This might apply to other languages and token kinds. 2842 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous && 2843 Left.Previous->is(tok::string_literal)) 2844 return true; 2845 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 && 2846 Left.Previous && Left.Previous->is(tok::equal) && 2847 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export, 2848 tok::kw_const) && 2849 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match 2850 // above. 2851 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) 2852 // Object literals on the top level of a file are treated as "enum-style". 2853 // Each key/value pair is put on a separate line, instead of bin-packing. 2854 return true; 2855 if (Left.is(tok::l_brace) && Line.Level == 0 && 2856 (Line.startsWith(tok::kw_enum) || 2857 Line.startsWith(tok::kw_const, tok::kw_enum) || 2858 Line.startsWith(tok::kw_export, tok::kw_enum) || 2859 Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum))) 2860 // JavaScript top-level enum key/value pairs are put on separate lines 2861 // instead of bin-packing. 2862 return true; 2863 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && 2864 !Left.Children.empty()) 2865 // Support AllowShortFunctionsOnASingleLine for JavaScript. 2866 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 2867 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty || 2868 (Left.NestingLevel == 0 && Line.Level == 0 && 2869 Style.AllowShortFunctionsOnASingleLine & 2870 FormatStyle::SFS_InlineOnly); 2871 } else if (Style.Language == FormatStyle::LK_Java) { 2872 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next && 2873 Right.Next->is(tok::string_literal)) 2874 return true; 2875 } else if (Style.Language == FormatStyle::LK_Cpp || 2876 Style.Language == FormatStyle::LK_ObjC || 2877 Style.Language == FormatStyle::LK_Proto || 2878 Style.Language == FormatStyle::LK_TextProto) { 2879 if (Left.isStringLiteral() && Right.isStringLiteral()) 2880 return true; 2881 } 2882 2883 // If the last token before a '}', ']', or ')' is a comma or a trailing 2884 // comment, the intention is to insert a line break after it in order to make 2885 // shuffling around entries easier. Import statements, especially in 2886 // JavaScript, can be an exception to this rule. 2887 if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) { 2888 const FormatToken *BeforeClosingBrace = nullptr; 2889 if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) || 2890 (Style.Language == FormatStyle::LK_JavaScript && 2891 Left.is(tok::l_paren))) && 2892 Left.BlockKind != BK_Block && Left.MatchingParen) 2893 BeforeClosingBrace = Left.MatchingParen->Previous; 2894 else if (Right.MatchingParen && 2895 (Right.MatchingParen->isOneOf(tok::l_brace, 2896 TT_ArrayInitializerLSquare) || 2897 (Style.Language == FormatStyle::LK_JavaScript && 2898 Right.MatchingParen->is(tok::l_paren)))) 2899 BeforeClosingBrace = &Left; 2900 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 2901 BeforeClosingBrace->isTrailingComment())) 2902 return true; 2903 } 2904 2905 if (Right.is(tok::comment)) 2906 return Left.BlockKind != BK_BracedInit && 2907 Left.isNot(TT_CtorInitializerColon) && 2908 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 2909 if (Left.isTrailingComment()) 2910 return true; 2911 if (Right.Previous->IsUnterminatedLiteral) 2912 return true; 2913 if (Right.is(tok::lessless) && Right.Next && 2914 Right.Previous->is(tok::string_literal) && 2915 Right.Next->is(tok::string_literal)) 2916 return true; 2917 if (Right.Previous->ClosesTemplateDeclaration && 2918 Right.Previous->MatchingParen && 2919 Right.Previous->MatchingParen->NestingLevel == 0 && 2920 Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes) 2921 return true; 2922 if (Right.is(TT_CtorInitializerComma) && 2923 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 2924 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 2925 return true; 2926 if (Right.is(TT_CtorInitializerColon) && 2927 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 2928 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 2929 return true; 2930 // Break only if we have multiple inheritance. 2931 if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma && 2932 Right.is(TT_InheritanceComma)) 2933 return true; 2934 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) 2935 // Multiline raw string literals are special wrt. line breaks. The author 2936 // has made a deliberate choice and might have aligned the contents of the 2937 // string literal accordingly. Thus, we try keep existing line breaks. 2938 return Right.IsMultiline && Right.NewlinesBefore > 0; 2939 if ((Right.Previous->is(tok::l_brace) || 2940 (Right.Previous->is(tok::less) && Right.Previous->Previous && 2941 Right.Previous->Previous->is(tok::equal))) && 2942 Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) { 2943 // Don't put enums or option definitions onto single lines in protocol 2944 // buffers. 2945 return true; 2946 } 2947 if (Right.is(TT_InlineASMBrace)) 2948 return Right.HasUnescapedNewline; 2949 if (isAllmanBrace(Left) || isAllmanBrace(Right)) 2950 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) || 2951 (Line.startsWith(tok::kw_typedef, tok::kw_enum) && 2952 Style.BraceWrapping.AfterEnum) || 2953 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) || 2954 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct); 2955 if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine) 2956 return true; 2957 2958 if ((Style.Language == FormatStyle::LK_Java || 2959 Style.Language == FormatStyle::LK_JavaScript) && 2960 Left.is(TT_LeadingJavaAnnotation) && 2961 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) && 2962 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) 2963 return true; 2964 2965 if (Right.is(TT_ProtoExtensionLSquare)) 2966 return true; 2967 2968 // In text proto instances if a submessage contains at least 2 entries and at 2969 // least one of them is a submessage, like A { ... B { ... } ... }, 2970 // put all of the entries of A on separate lines by forcing the selector of 2971 // the submessage B to be put on a newline. 2972 // 2973 // Example: these can stay on one line: 2974 // a { scalar_1: 1 scalar_2: 2 } 2975 // a { b { key: value } } 2976 // 2977 // and these entries need to be on a new line even if putting them all in one 2978 // line is under the column limit: 2979 // a { 2980 // scalar: 1 2981 // b { key: value } 2982 // } 2983 // 2984 // We enforce this by breaking before a submessage field that has previous 2985 // siblings, *and* breaking before a field that follows a submessage field. 2986 // 2987 // Be careful to exclude the case [proto.ext] { ... } since the `]` is 2988 // the TT_SelectorName there, but we don't want to break inside the brackets. 2989 // 2990 // Another edge case is @submessage { key: value }, which is a common 2991 // substitution placeholder. In this case we want to keep `@` and `submessage` 2992 // together. 2993 // 2994 // We ensure elsewhere that extensions are always on their own line. 2995 if ((Style.Language == FormatStyle::LK_Proto || 2996 Style.Language == FormatStyle::LK_TextProto) && 2997 Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) { 2998 // Keep `@submessage` together in: 2999 // @submessage { key: value } 3000 if (Right.Previous && Right.Previous->is(tok::at)) 3001 return false; 3002 // Look for the scope opener after selector in cases like: 3003 // selector { ... 3004 // selector: { ... 3005 // selector: @base { ... 3006 FormatToken *LBrace = Right.Next; 3007 if (LBrace && LBrace->is(tok::colon)) { 3008 LBrace = LBrace->Next; 3009 if (LBrace && LBrace->is(tok::at)) { 3010 LBrace = LBrace->Next; 3011 if (LBrace) 3012 LBrace = LBrace->Next; 3013 } 3014 } 3015 if (LBrace && 3016 // The scope opener is one of {, [, <: 3017 // selector { ... } 3018 // selector [ ... ] 3019 // selector < ... > 3020 // 3021 // In case of selector { ... }, the l_brace is TT_DictLiteral. 3022 // In case of an empty selector {}, the l_brace is not TT_DictLiteral, 3023 // so we check for immediately following r_brace. 3024 ((LBrace->is(tok::l_brace) && 3025 (LBrace->is(TT_DictLiteral) || 3026 (LBrace->Next && LBrace->Next->is(tok::r_brace)))) || 3027 LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) { 3028 // If Left.ParameterCount is 0, then this submessage entry is not the 3029 // first in its parent submessage, and we want to break before this entry. 3030 // If Left.ParameterCount is greater than 0, then its parent submessage 3031 // might contain 1 or more entries and we want to break before this entry 3032 // if it contains at least 2 entries. We deal with this case later by 3033 // detecting and breaking before the next entry in the parent submessage. 3034 if (Left.ParameterCount == 0) 3035 return true; 3036 // However, if this submessage is the first entry in its parent 3037 // submessage, Left.ParameterCount might be 1 in some cases. 3038 // We deal with this case later by detecting an entry 3039 // following a closing paren of this submessage. 3040 } 3041 3042 // If this is an entry immediately following a submessage, it will be 3043 // preceded by a closing paren of that submessage, like in: 3044 // left---. .---right 3045 // v v 3046 // sub: { ... } key: value 3047 // If there was a comment between `}` an `key` above, then `key` would be 3048 // put on a new line anyways. 3049 if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square)) 3050 return true; 3051 } 3052 3053 // Deal with lambda arguments in C++ - we want consistent line breaks whether 3054 // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced 3055 // as aggressive line breaks are placed when the lambda is not the last arg. 3056 if ((Style.Language == FormatStyle::LK_Cpp || 3057 Style.Language == FormatStyle::LK_ObjC) && 3058 Left.is(tok::l_paren) && Left.BlockParameterCount > 0 && 3059 !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) { 3060 // Multiple lambdas in the same function call force line breaks. 3061 if (Left.BlockParameterCount > 1) 3062 return true; 3063 3064 // A lambda followed by another arg forces a line break. 3065 if (!Left.Role) 3066 return false; 3067 auto Comma = Left.Role->lastComma(); 3068 if (!Comma) 3069 return false; 3070 auto Next = Comma->getNextNonComment(); 3071 if (!Next) 3072 return false; 3073 if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret)) 3074 return true; 3075 } 3076 3077 return false; 3078 } 3079 3080 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 3081 const FormatToken &Right) { 3082 const FormatToken &Left = *Right.Previous; 3083 3084 // Language-specific stuff. 3085 if (Style.Language == FormatStyle::LK_Java) { 3086 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 3087 Keywords.kw_implements)) 3088 return false; 3089 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 3090 Keywords.kw_implements)) 3091 return true; 3092 } else if (Style.Language == FormatStyle::LK_JavaScript) { 3093 const FormatToken *NonComment = Right.getPreviousNonComment(); 3094 if (NonComment && 3095 NonComment->isOneOf( 3096 tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break, 3097 tok::kw_throw, Keywords.kw_interface, Keywords.kw_type, 3098 tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected, 3099 Keywords.kw_readonly, Keywords.kw_abstract, Keywords.kw_get, 3100 Keywords.kw_set, Keywords.kw_async, Keywords.kw_await)) 3101 return false; // Otherwise automatic semicolon insertion would trigger. 3102 if (Right.NestingLevel == 0 && 3103 (Left.Tok.getIdentifierInfo() || 3104 Left.isOneOf(tok::r_square, tok::r_paren)) && 3105 Right.isOneOf(tok::l_square, tok::l_paren)) 3106 return false; // Otherwise automatic semicolon insertion would trigger. 3107 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace)) 3108 return false; 3109 if (Left.is(TT_JsTypeColon)) 3110 return true; 3111 // Don't wrap between ":" and "!" of a strict prop init ("field!: type;"). 3112 if (Left.is(tok::exclaim) && Right.is(tok::colon)) 3113 return false; 3114 if (Right.is(Keywords.kw_is)) 3115 return false; 3116 if (Left.is(Keywords.kw_in)) 3117 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None; 3118 if (Right.is(Keywords.kw_in)) 3119 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None; 3120 if (Right.is(Keywords.kw_as)) 3121 return false; // must not break before as in 'x as type' casts 3122 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) { 3123 // extends and infer can appear as keywords in conditional types: 3124 // https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types 3125 // do not break before them, as the expressions are subject to ASI. 3126 return false; 3127 } 3128 if (Left.is(Keywords.kw_as)) 3129 return true; 3130 if (Left.is(TT_JsNonNullAssertion)) 3131 return true; 3132 if (Left.is(Keywords.kw_declare) && 3133 Right.isOneOf(Keywords.kw_module, tok::kw_namespace, 3134 Keywords.kw_function, tok::kw_class, tok::kw_enum, 3135 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var, 3136 Keywords.kw_let, tok::kw_const)) 3137 // See grammar for 'declare' statements at: 3138 // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10 3139 return false; 3140 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) && 3141 Right.isOneOf(tok::identifier, tok::string_literal)) 3142 return false; // must not break in "module foo { ...}" 3143 if (Right.is(TT_TemplateString) && Right.closesScope()) 3144 return false; 3145 if (Left.is(TT_TemplateString) && Left.opensScope()) 3146 return true; 3147 } 3148 3149 if (Left.is(tok::at)) 3150 return false; 3151 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 3152 return false; 3153 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation)) 3154 return !Right.is(tok::l_paren); 3155 if (Right.is(TT_PointerOrReference)) 3156 return Line.IsMultiVariableDeclStmt || 3157 (Style.PointerAlignment == FormatStyle::PAS_Right && 3158 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName))); 3159 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 3160 Right.is(tok::kw_operator)) 3161 return true; 3162 if (Left.is(TT_PointerOrReference)) 3163 return false; 3164 if (Right.isTrailingComment()) 3165 // We rely on MustBreakBefore being set correctly here as we should not 3166 // change the "binding" behavior of a comment. 3167 // The first comment in a braced lists is always interpreted as belonging to 3168 // the first list element. Otherwise, it should be placed outside of the 3169 // list. 3170 return Left.BlockKind == BK_BracedInit || 3171 (Left.is(TT_CtorInitializerColon) && 3172 Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon); 3173 if (Left.is(tok::question) && Right.is(tok::colon)) 3174 return false; 3175 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question)) 3176 return Style.BreakBeforeTernaryOperators; 3177 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question)) 3178 return !Style.BreakBeforeTernaryOperators; 3179 if (Left.is(TT_InheritanceColon)) 3180 return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon; 3181 if (Right.is(TT_InheritanceColon)) 3182 return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon; 3183 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) && 3184 Left.isNot(TT_SelectorName)) 3185 return true; 3186 3187 if (Right.is(tok::colon) && 3188 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) 3189 return false; 3190 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) { 3191 if (Style.Language == FormatStyle::LK_Proto || 3192 Style.Language == FormatStyle::LK_TextProto) { 3193 if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral()) 3194 return false; 3195 // Prevent cases like: 3196 // 3197 // submessage: 3198 // { key: valueeeeeeeeeeee } 3199 // 3200 // when the snippet does not fit into one line. 3201 // Prefer: 3202 // 3203 // submessage: { 3204 // key: valueeeeeeeeeeee 3205 // } 3206 // 3207 // instead, even if it is longer by one line. 3208 // 3209 // Note that this allows allows the "{" to go over the column limit 3210 // when the column limit is just between ":" and "{", but that does 3211 // not happen too often and alternative formattings in this case are 3212 // not much better. 3213 // 3214 // The code covers the cases: 3215 // 3216 // submessage: { ... } 3217 // submessage: < ... > 3218 // repeated: [ ... ] 3219 if (((Right.is(tok::l_brace) || Right.is(tok::less)) && 3220 Right.is(TT_DictLiteral)) || 3221 Right.is(TT_ArrayInitializerLSquare)) 3222 return false; 3223 } 3224 return true; 3225 } 3226 if (Right.is(tok::r_square) && Right.MatchingParen && 3227 Right.MatchingParen->is(TT_ProtoExtensionLSquare)) 3228 return false; 3229 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next && 3230 Right.Next->is(TT_ObjCMethodExpr))) 3231 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls. 3232 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 3233 return true; 3234 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen)) 3235 return true; 3236 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen, 3237 TT_OverloadedOperator)) 3238 return false; 3239 if (Left.is(TT_RangeBasedForLoopColon)) 3240 return true; 3241 if (Right.is(TT_RangeBasedForLoopColon)) 3242 return false; 3243 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener)) 3244 return true; 3245 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) || 3246 Left.is(tok::kw_operator)) 3247 return false; 3248 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) && 3249 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) 3250 return false; 3251 if (Left.is(tok::equal) && Right.is(tok::l_brace) && 3252 !Style.Cpp11BracedListStyle) 3253 return false; 3254 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen)) 3255 return false; 3256 if (Left.is(tok::l_paren) && Left.Previous && 3257 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) 3258 return false; 3259 if (Right.is(TT_ImplicitStringLiteral)) 3260 return false; 3261 3262 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser)) 3263 return false; 3264 if (Right.is(tok::r_square) && Right.MatchingParen && 3265 Right.MatchingParen->is(TT_LambdaLSquare)) 3266 return false; 3267 3268 // We only break before r_brace if there was a corresponding break before 3269 // the l_brace, which is tracked by BreakBeforeClosingBrace. 3270 if (Right.is(tok::r_brace)) 3271 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 3272 3273 // Allow breaking after a trailing annotation, e.g. after a method 3274 // declaration. 3275 if (Left.is(TT_TrailingAnnotation)) 3276 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 3277 tok::less, tok::coloncolon); 3278 3279 if (Right.is(tok::kw___attribute) || 3280 (Right.is(tok::l_square) && Right.is(TT_AttributeSquare))) 3281 return true; 3282 3283 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 3284 return true; 3285 3286 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 3287 return true; 3288 3289 if (Left.is(TT_CtorInitializerColon)) 3290 return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon; 3291 if (Right.is(TT_CtorInitializerColon)) 3292 return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon; 3293 if (Left.is(TT_CtorInitializerComma) && 3294 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 3295 return false; 3296 if (Right.is(TT_CtorInitializerComma) && 3297 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 3298 return true; 3299 if (Left.is(TT_InheritanceComma) && 3300 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) 3301 return false; 3302 if (Right.is(TT_InheritanceComma) && 3303 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) 3304 return true; 3305 if ((Left.is(tok::greater) && Right.is(tok::greater)) || 3306 (Left.is(tok::less) && Right.is(tok::less))) 3307 return false; 3308 if (Right.is(TT_BinaryOperator) && 3309 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 3310 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 3311 Right.getPrecedence() != prec::Assignment)) 3312 return true; 3313 if (Left.is(TT_ArrayInitializerLSquare)) 3314 return true; 3315 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 3316 return true; 3317 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) && 3318 !Left.isOneOf(tok::arrowstar, tok::lessless) && 3319 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 3320 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 3321 Left.getPrecedence() == prec::Assignment)) 3322 return true; 3323 if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) || 3324 (Left.is(tok::r_square) && Right.is(TT_AttributeSquare))) 3325 return false; 3326 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 3327 tok::kw_class, tok::kw_struct, tok::comment) || 3328 Right.isMemberAccess() || 3329 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless, 3330 tok::colon, tok::l_square, tok::at) || 3331 (Left.is(tok::r_paren) && 3332 Right.isOneOf(tok::identifier, tok::kw_const)) || 3333 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) || 3334 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser)); 3335 } 3336 3337 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 3338 llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n"; 3339 const FormatToken *Tok = Line.First; 3340 while (Tok) { 3341 llvm::errs() << " M=" << Tok->MustBreakBefore 3342 << " C=" << Tok->CanBreakBefore 3343 << " T=" << getTokenTypeName(Tok->Type) 3344 << " S=" << Tok->SpacesRequiredBefore 3345 << " B=" << Tok->BlockParameterCount 3346 << " BK=" << Tok->BlockKind << " P=" << Tok->SplitPenalty 3347 << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength 3348 << " PPK=" << Tok->PackingKind << " FakeLParens="; 3349 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 3350 llvm::errs() << Tok->FakeLParens[i] << "/"; 3351 llvm::errs() << " FakeRParens=" << Tok->FakeRParens; 3352 llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo(); 3353 llvm::errs() << " Text='" << Tok->TokenText << "'\n"; 3354 if (!Tok->Next) 3355 assert(Tok == Line.Last); 3356 Tok = Tok->Next; 3357 } 3358 llvm::errs() << "----\n"; 3359 } 3360 3361 } // namespace format 3362 } // namespace clang 3363