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 /// \brief 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 /// \brief A parser that gathers additional information about tokens. 29 /// 30 /// The \c TokenAnnotator tries to match parenthesis and square brakets and 31 /// store a parenthesis levels. It also tries to resolve matching "<" and ">" 32 /// into template parameter lists. 33 class AnnotatingParser { 34 public: 35 AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line, 36 const AdditionalKeywords &Keywords) 37 : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false), 38 Keywords(Keywords) { 39 Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false)); 40 resetTokenMetadata(CurrentToken); 41 } 42 43 private: 44 bool parseAngle() { 45 if (!CurrentToken || !CurrentToken->Previous) 46 return false; 47 if (NonTemplateLess.count(CurrentToken->Previous)) 48 return false; 49 50 const FormatToken &Previous = *CurrentToken->Previous; // The '<'. 51 if (Previous.Previous) { 52 if (Previous.Previous->Tok.isLiteral()) 53 return false; 54 if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 && 55 (!Previous.Previous->MatchingParen || 56 !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen))) 57 return false; 58 } 59 60 FormatToken *Left = CurrentToken->Previous; 61 Left->ParentBracket = Contexts.back().ContextKind; 62 ScopedContextCreator ContextCreator(*this, tok::less, 12); 63 64 // If this angle is in the context of an expression, we need to be more 65 // hesitant to detect it as opening template parameters. 66 bool InExprContext = Contexts.back().IsExpression; 67 68 Contexts.back().IsExpression = false; 69 // If there's a template keyword before the opening angle bracket, this is a 70 // template parameter, not an argument. 71 Contexts.back().InTemplateArgument = 72 Left->Previous && Left->Previous->Tok.isNot(tok::kw_template); 73 74 if (Style.Language == FormatStyle::LK_Java && 75 CurrentToken->is(tok::question)) 76 next(); 77 78 while (CurrentToken) { 79 if (CurrentToken->is(tok::greater)) { 80 Left->MatchingParen = CurrentToken; 81 CurrentToken->MatchingParen = Left; 82 CurrentToken->Type = TT_TemplateCloser; 83 next(); 84 return true; 85 } 86 if (CurrentToken->is(tok::question) && 87 Style.Language == FormatStyle::LK_Java) { 88 next(); 89 continue; 90 } 91 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) || 92 (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext && 93 Style.Language != FormatStyle::LK_Proto && 94 Style.Language != FormatStyle::LK_TextProto)) 95 return false; 96 // If a && or || is found and interpreted as a binary operator, this set 97 // of angles is likely part of something like "a < b && c > d". If the 98 // angles are inside an expression, the ||/&& might also be a binary 99 // operator that was misinterpreted because we are parsing template 100 // parameters. 101 // FIXME: This is getting out of hand, write a decent parser. 102 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) && 103 CurrentToken->Previous->is(TT_BinaryOperator) && 104 Contexts[Contexts.size() - 2].IsExpression && 105 !Line.startsWith(tok::kw_template)) 106 return false; 107 updateParameterCount(Left, CurrentToken); 108 if (Style.Language == FormatStyle::LK_Proto) { 109 if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) { 110 if (CurrentToken->is(tok::colon) || 111 (CurrentToken->isOneOf(tok::l_brace, tok::less) && 112 Previous->isNot(tok::colon))) 113 Previous->Type = TT_SelectorName; 114 } 115 } 116 if (!consumeToken()) 117 return false; 118 } 119 return false; 120 } 121 122 bool parseParens(bool LookForDecls = false) { 123 if (!CurrentToken) 124 return false; 125 FormatToken *Left = CurrentToken->Previous; 126 Left->ParentBracket = Contexts.back().ContextKind; 127 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1); 128 129 // FIXME: This is a bit of a hack. Do better. 130 Contexts.back().ColonIsForRangeExpr = 131 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr; 132 133 bool StartsObjCMethodExpr = false; 134 if (CurrentToken->is(tok::caret)) { 135 // (^ can start a block type. 136 Left->Type = TT_ObjCBlockLParen; 137 } else if (FormatToken *MaybeSel = Left->Previous) { 138 // @selector( starts a selector. 139 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous && 140 MaybeSel->Previous->is(tok::at)) { 141 StartsObjCMethodExpr = true; 142 } 143 } 144 145 if (Left->is(TT_OverloadedOperatorLParen)) { 146 Contexts.back().IsExpression = false; 147 } else if (Style.Language == FormatStyle::LK_JavaScript && 148 (Line.startsWith(Keywords.kw_type, tok::identifier) || 149 Line.startsWith(tok::kw_export, Keywords.kw_type, 150 tok::identifier))) { 151 // type X = (...); 152 // export type X = (...); 153 Contexts.back().IsExpression = false; 154 } else if (Left->Previous && 155 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype, 156 tok::kw_if, tok::kw_while, tok::l_paren, 157 tok::comma) || 158 Left->Previous->endsSequence(tok::kw_constexpr, tok::kw_if) || 159 Left->Previous->is(TT_BinaryOperator))) { 160 // static_assert, if and while usually contain expressions. 161 Contexts.back().IsExpression = true; 162 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 163 (Left->Previous->is(Keywords.kw_function) || 164 (Left->Previous->endsSequence(tok::identifier, 165 Keywords.kw_function)))) { 166 // function(...) or function f(...) 167 Contexts.back().IsExpression = false; 168 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 169 Left->Previous->is(TT_JsTypeColon)) { 170 // let x: (SomeType); 171 Contexts.back().IsExpression = false; 172 } else if (Left->Previous && Left->Previous->is(tok::r_square) && 173 Left->Previous->MatchingParen && 174 Left->Previous->MatchingParen->is(TT_LambdaLSquare)) { 175 // This is a parameter list of a lambda expression. 176 Contexts.back().IsExpression = false; 177 } else if (Line.InPPDirective && 178 (!Left->Previous || !Left->Previous->is(tok::identifier))) { 179 Contexts.back().IsExpression = true; 180 } else if (Contexts[Contexts.size() - 2].CaretFound) { 181 // This is the parameter list of an ObjC block. 182 Contexts.back().IsExpression = false; 183 } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) { 184 Left->Type = TT_AttributeParen; 185 } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) { 186 // The first argument to a foreach macro is a declaration. 187 Contexts.back().IsForEachMacro = true; 188 Contexts.back().IsExpression = false; 189 } else if (Left->Previous && Left->Previous->MatchingParen && 190 Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) { 191 Contexts.back().IsExpression = false; 192 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) { 193 bool IsForOrCatch = 194 Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch); 195 Contexts.back().IsExpression = !IsForOrCatch; 196 } 197 198 if (StartsObjCMethodExpr) { 199 Contexts.back().ColonIsObjCMethodExpr = true; 200 Left->Type = TT_ObjCMethodExpr; 201 } 202 203 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression; 204 bool ProbablyFunctionType = CurrentToken->isOneOf(tok::star, tok::amp); 205 bool HasMultipleLines = false; 206 bool HasMultipleParametersOnALine = false; 207 bool MightBeObjCForRangeLoop = 208 Left->Previous && Left->Previous->is(tok::kw_for); 209 while (CurrentToken) { 210 // LookForDecls is set when "if (" has been seen. Check for 211 // 'identifier' '*' 'identifier' followed by not '=' -- this 212 // '*' has to be a binary operator but determineStarAmpUsage() will 213 // categorize it as an unary operator, so set the right type here. 214 if (LookForDecls && CurrentToken->Next) { 215 FormatToken *Prev = CurrentToken->getPreviousNonComment(); 216 if (Prev) { 217 FormatToken *PrevPrev = Prev->getPreviousNonComment(); 218 FormatToken *Next = CurrentToken->Next; 219 if (PrevPrev && PrevPrev->is(tok::identifier) && 220 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) && 221 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) { 222 Prev->Type = TT_BinaryOperator; 223 LookForDecls = false; 224 } 225 } 226 } 227 228 if (CurrentToken->Previous->is(TT_PointerOrReference) && 229 CurrentToken->Previous->Previous->isOneOf(tok::l_paren, 230 tok::coloncolon)) 231 ProbablyFunctionType = true; 232 if (CurrentToken->is(tok::comma)) 233 MightBeFunctionType = false; 234 if (CurrentToken->Previous->is(TT_BinaryOperator)) 235 Contexts.back().IsExpression = true; 236 if (CurrentToken->is(tok::r_paren)) { 237 if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next && 238 (CurrentToken->Next->is(tok::l_paren) || 239 (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration))) 240 Left->Type = TT_FunctionTypeLParen; 241 Left->MatchingParen = CurrentToken; 242 CurrentToken->MatchingParen = Left; 243 244 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) && 245 Left->Previous && Left->Previous->is(tok::l_paren)) { 246 // Detect the case where macros are used to generate lambdas or 247 // function bodies, e.g.: 248 // auto my_lambda = MARCO((Type *type, int i) { .. body .. }); 249 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) { 250 if (Tok->is(TT_BinaryOperator) && 251 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) 252 Tok->Type = TT_PointerOrReference; 253 } 254 } 255 256 if (StartsObjCMethodExpr) { 257 CurrentToken->Type = TT_ObjCMethodExpr; 258 if (Contexts.back().FirstObjCSelectorName) { 259 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 260 Contexts.back().LongestObjCSelectorName; 261 } 262 } 263 264 if (Left->is(TT_AttributeParen)) 265 CurrentToken->Type = TT_AttributeParen; 266 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation)) 267 CurrentToken->Type = TT_JavaAnnotation; 268 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation)) 269 CurrentToken->Type = TT_LeadingJavaAnnotation; 270 271 if (!HasMultipleLines) 272 Left->PackingKind = PPK_Inconclusive; 273 else if (HasMultipleParametersOnALine) 274 Left->PackingKind = PPK_BinPacked; 275 else 276 Left->PackingKind = PPK_OnePerLine; 277 278 next(); 279 return true; 280 } 281 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace)) 282 return false; 283 284 if (CurrentToken->is(tok::l_brace)) 285 Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen 286 if (CurrentToken->is(tok::comma) && CurrentToken->Next && 287 !CurrentToken->Next->HasUnescapedNewline && 288 !CurrentToken->Next->isTrailingComment()) 289 HasMultipleParametersOnALine = true; 290 if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) || 291 CurrentToken->Previous->isSimpleTypeSpecifier()) && 292 !CurrentToken->is(tok::l_brace)) 293 Contexts.back().IsExpression = false; 294 if (CurrentToken->isOneOf(tok::semi, tok::colon)) 295 MightBeObjCForRangeLoop = false; 296 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) 297 CurrentToken->Type = TT_ObjCForIn; 298 // When we discover a 'new', we set CanBeExpression to 'false' in order to 299 // parse the type correctly. Reset that after a comma. 300 if (CurrentToken->is(tok::comma)) 301 Contexts.back().CanBeExpression = true; 302 303 FormatToken *Tok = CurrentToken; 304 if (!consumeToken()) 305 return false; 306 updateParameterCount(Left, Tok); 307 if (CurrentToken && CurrentToken->HasUnescapedNewline) 308 HasMultipleLines = true; 309 } 310 return false; 311 } 312 313 bool parseSquare() { 314 if (!CurrentToken) 315 return false; 316 317 // A '[' could be an index subscript (after an identifier or after 318 // ')' or ']'), it could be the start of an Objective-C method 319 // expression, or it could the start of an Objective-C array literal. 320 FormatToken *Left = CurrentToken->Previous; 321 Left->ParentBracket = Contexts.back().ContextKind; 322 FormatToken *Parent = Left->getPreviousNonComment(); 323 324 // Cases where '>' is followed by '['. 325 // In C++, this can happen either in array of templates (foo<int>[10]) 326 // or when array is a nested template type (unique_ptr<type1<type2>[]>). 327 bool CppArrayTemplates = 328 Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) && 329 (Contexts.back().CanBeExpression || Contexts.back().IsExpression || 330 Contexts.back().InTemplateArgument); 331 332 bool StartsObjCMethodExpr = 333 !CppArrayTemplates && Style.isCpp() && 334 Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) && 335 CurrentToken->isNot(tok::l_brace) && 336 (!Parent || 337 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren, 338 tok::kw_return, tok::kw_throw) || 339 Parent->isUnaryOperator() || 340 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) || 341 getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown); 342 bool ColonFound = false; 343 344 unsigned BindingIncrease = 1; 345 if (Left->isCppStructuredBinding(Style)) { 346 Left->Type = TT_StructuredBindingLSquare; 347 } else if (Left->is(TT_Unknown)) { 348 if (StartsObjCMethodExpr) { 349 Left->Type = TT_ObjCMethodExpr; 350 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent && 351 Contexts.back().ContextKind == tok::l_brace && 352 Parent->isOneOf(tok::l_brace, tok::comma)) { 353 Left->Type = TT_JsComputedPropertyName; 354 } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace && 355 Parent && Parent->isOneOf(tok::l_brace, tok::comma)) { 356 Left->Type = TT_DesignatedInitializerLSquare; 357 } else if (CurrentToken->is(tok::r_square) && Parent && 358 Parent->is(TT_TemplateCloser)) { 359 Left->Type = TT_ArraySubscriptLSquare; 360 } else if (Style.Language == FormatStyle::LK_Proto || 361 (!CppArrayTemplates && Parent && 362 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at, 363 tok::comma, tok::l_paren, tok::l_square, 364 tok::question, tok::colon, tok::kw_return, 365 // Should only be relevant to JavaScript: 366 tok::kw_default))) { 367 Left->Type = TT_ArrayInitializerLSquare; 368 } else { 369 BindingIncrease = 10; 370 Left->Type = TT_ArraySubscriptLSquare; 371 } 372 } 373 374 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease); 375 Contexts.back().IsExpression = true; 376 if (Style.Language == FormatStyle::LK_JavaScript && Parent && 377 Parent->is(TT_JsTypeColon)) 378 Contexts.back().IsExpression = false; 379 380 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr; 381 382 while (CurrentToken) { 383 if (CurrentToken->is(tok::r_square)) { 384 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_paren) && 385 Left->is(TT_ObjCMethodExpr)) { 386 // An ObjC method call is rarely followed by an open parenthesis. 387 // FIXME: Do we incorrectly label ":" with this? 388 StartsObjCMethodExpr = false; 389 Left->Type = TT_Unknown; 390 } 391 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) { 392 CurrentToken->Type = TT_ObjCMethodExpr; 393 // determineStarAmpUsage() thinks that '*' '[' is allocating an 394 // array of pointers, but if '[' starts a selector then '*' is a 395 // binary operator. 396 if (Parent && Parent->is(TT_PointerOrReference)) 397 Parent->Type = TT_BinaryOperator; 398 } 399 Left->MatchingParen = CurrentToken; 400 CurrentToken->MatchingParen = Left; 401 if (Contexts.back().FirstObjCSelectorName) { 402 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 403 Contexts.back().LongestObjCSelectorName; 404 if (Left->BlockParameterCount > 1) 405 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0; 406 } 407 next(); 408 return true; 409 } 410 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace)) 411 return false; 412 if (CurrentToken->is(tok::colon)) { 413 if (Left->isOneOf(TT_ArraySubscriptLSquare, 414 TT_DesignatedInitializerLSquare)) { 415 Left->Type = TT_ObjCMethodExpr; 416 StartsObjCMethodExpr = true; 417 Contexts.back().ColonIsObjCMethodExpr = true; 418 if (Parent && Parent->is(tok::r_paren)) 419 Parent->Type = TT_CastRParen; 420 } 421 ColonFound = true; 422 } 423 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) && 424 !ColonFound) 425 Left->Type = TT_ArrayInitializerLSquare; 426 FormatToken *Tok = CurrentToken; 427 if (!consumeToken()) 428 return false; 429 updateParameterCount(Left, Tok); 430 } 431 return false; 432 } 433 434 bool parseBrace() { 435 if (CurrentToken) { 436 FormatToken *Left = CurrentToken->Previous; 437 Left->ParentBracket = Contexts.back().ContextKind; 438 439 if (Contexts.back().CaretFound) 440 Left->Type = TT_ObjCBlockLBrace; 441 Contexts.back().CaretFound = false; 442 443 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1); 444 Contexts.back().ColonIsDictLiteral = true; 445 if (Left->BlockKind == BK_BracedInit) 446 Contexts.back().IsExpression = true; 447 if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous && 448 Left->Previous->is(TT_JsTypeColon)) 449 Contexts.back().IsExpression = false; 450 451 while (CurrentToken) { 452 if (CurrentToken->is(tok::r_brace)) { 453 Left->MatchingParen = CurrentToken; 454 CurrentToken->MatchingParen = Left; 455 next(); 456 return true; 457 } 458 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square)) 459 return false; 460 updateParameterCount(Left, CurrentToken); 461 if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) { 462 FormatToken *Previous = CurrentToken->getPreviousNonComment(); 463 if (Previous->is(TT_JsTypeOptionalQuestion)) 464 Previous = Previous->getPreviousNonComment(); 465 if ((CurrentToken->is(tok::colon) && 466 (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) || 467 Style.Language == FormatStyle::LK_Proto || 468 Style.Language == FormatStyle::LK_TextProto) { 469 Left->Type = TT_DictLiteral; 470 if (Previous->Tok.getIdentifierInfo() || 471 Previous->is(tok::string_literal)) 472 Previous->Type = TT_SelectorName; 473 } 474 if (CurrentToken->is(tok::colon) || 475 Style.Language == FormatStyle::LK_JavaScript) 476 Left->Type = TT_DictLiteral; 477 } 478 if (CurrentToken->is(tok::comma) && 479 Style.Language == FormatStyle::LK_JavaScript) 480 Left->Type = TT_DictLiteral; 481 if (!consumeToken()) 482 return false; 483 } 484 } 485 return true; 486 } 487 488 void updateParameterCount(FormatToken *Left, FormatToken *Current) { 489 if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block) 490 ++Left->BlockParameterCount; 491 if (Current->is(tok::comma)) { 492 ++Left->ParameterCount; 493 if (!Left->Role) 494 Left->Role.reset(new CommaSeparatedList(Style)); 495 Left->Role->CommaFound(Current); 496 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) { 497 Left->ParameterCount = 1; 498 } 499 } 500 501 bool parseConditional() { 502 while (CurrentToken) { 503 if (CurrentToken->is(tok::colon)) { 504 CurrentToken->Type = TT_ConditionalExpr; 505 next(); 506 return true; 507 } 508 if (!consumeToken()) 509 return false; 510 } 511 return false; 512 } 513 514 bool parseTemplateDeclaration() { 515 if (CurrentToken && CurrentToken->is(tok::less)) { 516 CurrentToken->Type = TT_TemplateOpener; 517 next(); 518 if (!parseAngle()) 519 return false; 520 if (CurrentToken) 521 CurrentToken->Previous->ClosesTemplateDeclaration = true; 522 return true; 523 } 524 return false; 525 } 526 527 bool consumeToken() { 528 FormatToken *Tok = CurrentToken; 529 next(); 530 switch (Tok->Tok.getKind()) { 531 case tok::plus: 532 case tok::minus: 533 if (!Tok->Previous && Line.MustBeDeclaration) 534 Tok->Type = TT_ObjCMethodSpecifier; 535 break; 536 case tok::colon: 537 if (!Tok->Previous) 538 return false; 539 // Colons from ?: are handled in parseConditional(). 540 if (Style.Language == FormatStyle::LK_JavaScript) { 541 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop 542 (Contexts.size() == 1 && // switch/case labels 543 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) || 544 Contexts.back().ContextKind == tok::l_paren || // function params 545 Contexts.back().ContextKind == tok::l_square || // array type 546 (!Contexts.back().IsExpression && 547 Contexts.back().ContextKind == tok::l_brace) || // object type 548 (Contexts.size() == 1 && 549 Line.MustBeDeclaration)) { // method/property declaration 550 Contexts.back().IsExpression = false; 551 Tok->Type = TT_JsTypeColon; 552 break; 553 } 554 } 555 if (Contexts.back().ColonIsDictLiteral || 556 Style.Language == FormatStyle::LK_Proto || 557 Style.Language == FormatStyle::LK_TextProto) { 558 Tok->Type = TT_DictLiteral; 559 if (Style.Language == FormatStyle::LK_TextProto) { 560 if (FormatToken *Previous = Tok->getPreviousNonComment()) 561 Previous->Type = TT_SelectorName; 562 } 563 } else if (Contexts.back().ColonIsObjCMethodExpr || 564 Line.startsWith(TT_ObjCMethodSpecifier)) { 565 Tok->Type = TT_ObjCMethodExpr; 566 const FormatToken *BeforePrevious = Tok->Previous->Previous; 567 if (!BeforePrevious || 568 !(BeforePrevious->is(TT_CastRParen) || 569 (BeforePrevious->is(TT_ObjCMethodExpr) && 570 BeforePrevious->is(tok::colon))) || 571 BeforePrevious->is(tok::r_square) || 572 Contexts.back().LongestObjCSelectorName == 0) { 573 Tok->Previous->Type = TT_SelectorName; 574 if (Tok->Previous->ColumnWidth > 575 Contexts.back().LongestObjCSelectorName) 576 Contexts.back().LongestObjCSelectorName = 577 Tok->Previous->ColumnWidth; 578 if (!Contexts.back().FirstObjCSelectorName) 579 Contexts.back().FirstObjCSelectorName = Tok->Previous; 580 } 581 } else if (Contexts.back().ColonIsForRangeExpr) { 582 Tok->Type = TT_RangeBasedForLoopColon; 583 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) { 584 Tok->Type = TT_BitFieldColon; 585 } else if (Contexts.size() == 1 && 586 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) { 587 if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren, 588 tok::kw_noexcept)) 589 Tok->Type = TT_CtorInitializerColon; 590 else 591 Tok->Type = TT_InheritanceColon; 592 } else if (Tok->Previous->is(tok::identifier) && Tok->Next && 593 Tok->Next->isOneOf(tok::r_paren, tok::comma)) { 594 // This handles a special macro in ObjC code where selectors including 595 // the colon are passed as macro arguments. 596 Tok->Type = TT_ObjCMethodExpr; 597 } else if (Contexts.back().ContextKind == tok::l_paren) { 598 Tok->Type = TT_InlineASMColon; 599 } 600 break; 601 case tok::pipe: 602 case tok::amp: 603 // | and & in declarations/type expressions represent union and 604 // intersection types, respectively. 605 if (Style.Language == FormatStyle::LK_JavaScript && 606 !Contexts.back().IsExpression) 607 Tok->Type = TT_JsTypeOperator; 608 break; 609 case tok::kw_if: 610 case tok::kw_while: 611 if (Tok->is(tok::kw_if) && CurrentToken && 612 CurrentToken->is(tok::kw_constexpr)) 613 next(); 614 if (CurrentToken && CurrentToken->is(tok::l_paren)) { 615 next(); 616 if (!parseParens(/*LookForDecls=*/true)) 617 return false; 618 } 619 break; 620 case tok::kw_for: 621 if (Style.Language == FormatStyle::LK_JavaScript) { 622 // x.for and {for: ...} 623 if ((Tok->Previous && Tok->Previous->is(tok::period)) || 624 (Tok->Next && Tok->Next->is(tok::colon))) 625 break; 626 // JS' for await ( ... 627 if (CurrentToken && CurrentToken->is(Keywords.kw_await)) 628 next(); 629 } 630 Contexts.back().ColonIsForRangeExpr = true; 631 next(); 632 if (!parseParens()) 633 return false; 634 break; 635 case tok::l_paren: 636 // When faced with 'operator()()', the kw_operator handler incorrectly 637 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make 638 // the first two parens OverloadedOperators and the second l_paren an 639 // OverloadedOperatorLParen. 640 if (Tok->Previous && Tok->Previous->is(tok::r_paren) && 641 Tok->Previous->MatchingParen && 642 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) { 643 Tok->Previous->Type = TT_OverloadedOperator; 644 Tok->Previous->MatchingParen->Type = TT_OverloadedOperator; 645 Tok->Type = TT_OverloadedOperatorLParen; 646 } 647 648 if (!parseParens()) 649 return false; 650 if (Line.MustBeDeclaration && Contexts.size() == 1 && 651 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) && 652 (!Tok->Previous || 653 !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute, 654 TT_LeadingJavaAnnotation))) 655 Line.MightBeFunctionDecl = true; 656 break; 657 case tok::l_square: 658 if (!parseSquare()) 659 return false; 660 break; 661 case tok::l_brace: 662 if (Style.Language == FormatStyle::LK_TextProto) { 663 FormatToken *Previous = Tok->getPreviousNonComment(); 664 if (Previous && Previous->Type != TT_DictLiteral) 665 Previous->Type = TT_SelectorName; 666 } 667 if (!parseBrace()) 668 return false; 669 break; 670 case tok::less: 671 if (parseAngle()) { 672 Tok->Type = TT_TemplateOpener; 673 if (Style.Language == FormatStyle::LK_TextProto) { 674 FormatToken *Previous = Tok->getPreviousNonComment(); 675 if (Previous && Previous->Type != TT_DictLiteral) 676 Previous->Type = TT_SelectorName; 677 } 678 } else { 679 Tok->Type = TT_BinaryOperator; 680 NonTemplateLess.insert(Tok); 681 CurrentToken = Tok; 682 next(); 683 } 684 break; 685 case tok::r_paren: 686 case tok::r_square: 687 return false; 688 case tok::r_brace: 689 // Lines can start with '}'. 690 if (Tok->Previous) 691 return false; 692 break; 693 case tok::greater: 694 Tok->Type = TT_BinaryOperator; 695 break; 696 case tok::kw_operator: 697 while (CurrentToken && 698 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) { 699 if (CurrentToken->isOneOf(tok::star, tok::amp)) 700 CurrentToken->Type = TT_PointerOrReference; 701 consumeToken(); 702 if (CurrentToken && 703 CurrentToken->Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator, 704 tok::comma)) 705 CurrentToken->Previous->Type = TT_OverloadedOperator; 706 } 707 if (CurrentToken) { 708 CurrentToken->Type = TT_OverloadedOperatorLParen; 709 if (CurrentToken->Previous->is(TT_BinaryOperator)) 710 CurrentToken->Previous->Type = TT_OverloadedOperator; 711 } 712 break; 713 case tok::question: 714 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next && 715 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren, 716 tok::r_brace)) { 717 // Question marks before semicolons, colons, etc. indicate optional 718 // types (fields, parameters), e.g. 719 // function(x?: string, y?) {...} 720 // class X { y?; } 721 Tok->Type = TT_JsTypeOptionalQuestion; 722 break; 723 } 724 // Declarations cannot be conditional expressions, this can only be part 725 // of a type declaration. 726 if (Line.MustBeDeclaration && !Contexts.back().IsExpression && 727 Style.Language == FormatStyle::LK_JavaScript) 728 break; 729 parseConditional(); 730 break; 731 case tok::kw_template: 732 parseTemplateDeclaration(); 733 break; 734 case tok::comma: 735 if (Contexts.back().InCtorInitializer) 736 Tok->Type = TT_CtorInitializerComma; 737 else if (Contexts.back().InInheritanceList) 738 Tok->Type = TT_InheritanceComma; 739 else if (Contexts.back().FirstStartOfName && 740 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) { 741 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true; 742 Line.IsMultiVariableDeclStmt = true; 743 } 744 if (Contexts.back().IsForEachMacro) 745 Contexts.back().IsExpression = true; 746 break; 747 case tok::identifier: 748 if (Tok->isOneOf(Keywords.kw___has_include, 749 Keywords.kw___has_include_next)) { 750 parseHasInclude(); 751 } 752 break; 753 default: 754 break; 755 } 756 return true; 757 } 758 759 void parseIncludeDirective() { 760 if (CurrentToken && CurrentToken->is(tok::less)) { 761 next(); 762 while (CurrentToken) { 763 // Mark tokens up to the trailing line comments as implicit string 764 // literals. 765 if (CurrentToken->isNot(tok::comment) && 766 !CurrentToken->TokenText.startswith("//")) 767 CurrentToken->Type = TT_ImplicitStringLiteral; 768 next(); 769 } 770 } 771 } 772 773 void parseWarningOrError() { 774 next(); 775 // We still want to format the whitespace left of the first token of the 776 // warning or error. 777 next(); 778 while (CurrentToken) { 779 CurrentToken->Type = TT_ImplicitStringLiteral; 780 next(); 781 } 782 } 783 784 void parsePragma() { 785 next(); // Consume "pragma". 786 if (CurrentToken && 787 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) { 788 bool IsMark = CurrentToken->is(Keywords.kw_mark); 789 next(); // Consume "mark". 790 next(); // Consume first token (so we fix leading whitespace). 791 while (CurrentToken) { 792 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator)) 793 CurrentToken->Type = TT_ImplicitStringLiteral; 794 next(); 795 } 796 } 797 } 798 799 void parseHasInclude() { 800 if (!CurrentToken || !CurrentToken->is(tok::l_paren)) 801 return; 802 next(); // '(' 803 parseIncludeDirective(); 804 next(); // ')' 805 } 806 807 LineType parsePreprocessorDirective() { 808 bool IsFirstToken = CurrentToken->IsFirst; 809 LineType Type = LT_PreprocessorDirective; 810 next(); 811 if (!CurrentToken) 812 return Type; 813 814 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) { 815 // JavaScript files can contain shebang lines of the form: 816 // #!/usr/bin/env node 817 // Treat these like C++ #include directives. 818 while (CurrentToken) { 819 // Tokens cannot be comments here. 820 CurrentToken->Type = TT_ImplicitStringLiteral; 821 next(); 822 } 823 return LT_ImportStatement; 824 } 825 826 if (CurrentToken->Tok.is(tok::numeric_constant)) { 827 CurrentToken->SpacesRequiredBefore = 1; 828 return Type; 829 } 830 // Hashes in the middle of a line can lead to any strange token 831 // sequence. 832 if (!CurrentToken->Tok.getIdentifierInfo()) 833 return Type; 834 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) { 835 case tok::pp_include: 836 case tok::pp_include_next: 837 case tok::pp_import: 838 next(); 839 parseIncludeDirective(); 840 Type = LT_ImportStatement; 841 break; 842 case tok::pp_error: 843 case tok::pp_warning: 844 parseWarningOrError(); 845 break; 846 case tok::pp_pragma: 847 parsePragma(); 848 break; 849 case tok::pp_if: 850 case tok::pp_elif: 851 Contexts.back().IsExpression = true; 852 parseLine(); 853 break; 854 default: 855 break; 856 } 857 while (CurrentToken) { 858 FormatToken *Tok = CurrentToken; 859 next(); 860 if (Tok->is(tok::l_paren)) 861 parseParens(); 862 else if (Tok->isOneOf(Keywords.kw___has_include, 863 Keywords.kw___has_include_next)) 864 parseHasInclude(); 865 } 866 return Type; 867 } 868 869 public: 870 LineType parseLine() { 871 NonTemplateLess.clear(); 872 if (CurrentToken->is(tok::hash)) 873 return parsePreprocessorDirective(); 874 875 // Directly allow to 'import <string-literal>' to support protocol buffer 876 // definitions (github.com/google/protobuf) or missing "#" (either way we 877 // should not break the line). 878 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo(); 879 if ((Style.Language == FormatStyle::LK_Java && 880 CurrentToken->is(Keywords.kw_package)) || 881 (Info && Info->getPPKeywordID() == tok::pp_import && 882 CurrentToken->Next && 883 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier, 884 tok::kw_static))) { 885 next(); 886 parseIncludeDirective(); 887 return LT_ImportStatement; 888 } 889 890 // If this line starts and ends in '<' and '>', respectively, it is likely 891 // part of "#define <a/b.h>". 892 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) { 893 parseIncludeDirective(); 894 return LT_ImportStatement; 895 } 896 897 // In .proto files, top-level options are very similar to import statements 898 // and should not be line-wrapped. 899 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 && 900 CurrentToken->is(Keywords.kw_option)) { 901 next(); 902 if (CurrentToken && CurrentToken->is(tok::identifier)) 903 return LT_ImportStatement; 904 } 905 906 bool KeywordVirtualFound = false; 907 bool ImportStatement = false; 908 909 // import {...} from '...'; 910 if (Style.Language == FormatStyle::LK_JavaScript && 911 CurrentToken->is(Keywords.kw_import)) 912 ImportStatement = true; 913 914 while (CurrentToken) { 915 if (CurrentToken->is(tok::kw_virtual)) 916 KeywordVirtualFound = true; 917 if (Style.Language == FormatStyle::LK_JavaScript) { 918 // export {...} from '...'; 919 // An export followed by "from 'some string';" is a re-export from 920 // another module identified by a URI and is treated as a 921 // LT_ImportStatement (i.e. prevent wraps on it for long URIs). 922 // Just "export {...};" or "export class ..." should not be treated as 923 // an import in this sense. 924 if (Line.First->is(tok::kw_export) && 925 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next && 926 CurrentToken->Next->isStringLiteral()) 927 ImportStatement = true; 928 if (isClosureImportStatement(*CurrentToken)) 929 ImportStatement = true; 930 } 931 if (!consumeToken()) 932 return LT_Invalid; 933 } 934 if (KeywordVirtualFound) 935 return LT_VirtualFunctionDecl; 936 if (ImportStatement) 937 return LT_ImportStatement; 938 939 if (Line.startsWith(TT_ObjCMethodSpecifier)) { 940 if (Contexts.back().FirstObjCSelectorName) 941 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 942 Contexts.back().LongestObjCSelectorName; 943 return LT_ObjCMethodDecl; 944 } 945 946 return LT_Other; 947 } 948 949 private: 950 bool isClosureImportStatement(const FormatToken &Tok) { 951 // FIXME: Closure-library specific stuff should not be hard-coded but be 952 // configurable. 953 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) && 954 Tok.Next->Next && 955 (Tok.Next->Next->TokenText == "module" || 956 Tok.Next->Next->TokenText == "provide" || 957 Tok.Next->Next->TokenText == "require" || 958 Tok.Next->Next->TokenText == "forwardDeclare") && 959 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren); 960 } 961 962 void resetTokenMetadata(FormatToken *Token) { 963 if (!Token) 964 return; 965 966 // Reset token type in case we have already looked at it and then 967 // recovered from an error (e.g. failure to find the matching >). 968 if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro, 969 TT_FunctionLBrace, TT_ImplicitStringLiteral, 970 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow, 971 TT_OverloadedOperator, TT_RegexLiteral, 972 TT_TemplateString, TT_ObjCStringLiteral)) 973 CurrentToken->Type = TT_Unknown; 974 CurrentToken->Role.reset(); 975 CurrentToken->MatchingParen = nullptr; 976 CurrentToken->FakeLParens.clear(); 977 CurrentToken->FakeRParens = 0; 978 } 979 980 void next() { 981 if (CurrentToken) { 982 CurrentToken->NestingLevel = Contexts.size() - 1; 983 CurrentToken->BindingStrength = Contexts.back().BindingStrength; 984 modifyContext(*CurrentToken); 985 determineTokenType(*CurrentToken); 986 CurrentToken = CurrentToken->Next; 987 } 988 989 resetTokenMetadata(CurrentToken); 990 } 991 992 /// \brief A struct to hold information valid in a specific context, e.g. 993 /// a pair of parenthesis. 994 struct Context { 995 Context(tok::TokenKind ContextKind, unsigned BindingStrength, 996 bool IsExpression) 997 : ContextKind(ContextKind), BindingStrength(BindingStrength), 998 IsExpression(IsExpression) {} 999 1000 tok::TokenKind ContextKind; 1001 unsigned BindingStrength; 1002 bool IsExpression; 1003 unsigned LongestObjCSelectorName = 0; 1004 bool ColonIsForRangeExpr = false; 1005 bool ColonIsDictLiteral = false; 1006 bool ColonIsObjCMethodExpr = false; 1007 FormatToken *FirstObjCSelectorName = nullptr; 1008 FormatToken *FirstStartOfName = nullptr; 1009 bool CanBeExpression = true; 1010 bool InTemplateArgument = false; 1011 bool InCtorInitializer = false; 1012 bool InInheritanceList = false; 1013 bool CaretFound = false; 1014 bool IsForEachMacro = false; 1015 }; 1016 1017 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime 1018 /// of each instance. 1019 struct ScopedContextCreator { 1020 AnnotatingParser &P; 1021 1022 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind, 1023 unsigned Increase) 1024 : P(P) { 1025 P.Contexts.push_back(Context(ContextKind, 1026 P.Contexts.back().BindingStrength + Increase, 1027 P.Contexts.back().IsExpression)); 1028 } 1029 1030 ~ScopedContextCreator() { P.Contexts.pop_back(); } 1031 }; 1032 1033 void modifyContext(const FormatToken &Current) { 1034 if (Current.getPrecedence() == prec::Assignment && 1035 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) && 1036 // Type aliases use `type X = ...;` in TypeScript and can be exported 1037 // using `export type ...`. 1038 !(Style.Language == FormatStyle::LK_JavaScript && 1039 (Line.startsWith(Keywords.kw_type, tok::identifier) || 1040 Line.startsWith(tok::kw_export, Keywords.kw_type, 1041 tok::identifier))) && 1042 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) { 1043 Contexts.back().IsExpression = true; 1044 if (!Line.startsWith(TT_UnaryOperator)) { 1045 for (FormatToken *Previous = Current.Previous; 1046 Previous && Previous->Previous && 1047 !Previous->Previous->isOneOf(tok::comma, tok::semi); 1048 Previous = Previous->Previous) { 1049 if (Previous->isOneOf(tok::r_square, tok::r_paren)) { 1050 Previous = Previous->MatchingParen; 1051 if (!Previous) 1052 break; 1053 } 1054 if (Previous->opensScope()) 1055 break; 1056 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) && 1057 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) && 1058 Previous->Previous && Previous->Previous->isNot(tok::equal)) 1059 Previous->Type = TT_PointerOrReference; 1060 } 1061 } 1062 } else if (Current.is(tok::lessless) && 1063 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) { 1064 Contexts.back().IsExpression = true; 1065 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) { 1066 Contexts.back().IsExpression = true; 1067 } else if (Current.is(TT_TrailingReturnArrow)) { 1068 Contexts.back().IsExpression = false; 1069 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) { 1070 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java; 1071 } else if (Current.Previous && 1072 Current.Previous->is(TT_CtorInitializerColon)) { 1073 Contexts.back().IsExpression = true; 1074 Contexts.back().InCtorInitializer = true; 1075 } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) { 1076 Contexts.back().InInheritanceList = true; 1077 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) { 1078 for (FormatToken *Previous = Current.Previous; 1079 Previous && Previous->isOneOf(tok::star, tok::amp); 1080 Previous = Previous->Previous) 1081 Previous->Type = TT_PointerOrReference; 1082 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer) 1083 Contexts.back().IsExpression = false; 1084 } else if (Current.is(tok::kw_new)) { 1085 Contexts.back().CanBeExpression = false; 1086 } else if (Current.isOneOf(tok::semi, tok::exclaim)) { 1087 // This should be the condition or increment in a for-loop. 1088 Contexts.back().IsExpression = true; 1089 } 1090 } 1091 1092 void determineTokenType(FormatToken &Current) { 1093 if (!Current.is(TT_Unknown)) 1094 // The token type is already known. 1095 return; 1096 1097 if (Style.Language == FormatStyle::LK_JavaScript) { 1098 if (Current.is(tok::exclaim)) { 1099 if (Current.Previous && 1100 (Current.Previous->isOneOf(tok::identifier, tok::kw_namespace, 1101 tok::r_paren, tok::r_square, 1102 tok::r_brace) || 1103 Current.Previous->Tok.isLiteral())) { 1104 Current.Type = TT_JsNonNullAssertion; 1105 return; 1106 } 1107 if (Current.Next && 1108 Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) { 1109 Current.Type = TT_JsNonNullAssertion; 1110 return; 1111 } 1112 } 1113 } 1114 1115 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1116 // function declaration have been found. In this case, 'Current' is a 1117 // trailing token of this declaration and thus cannot be a name. 1118 if (Current.is(Keywords.kw_instanceof)) { 1119 Current.Type = TT_BinaryOperator; 1120 } else if (isStartOfName(Current) && 1121 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) { 1122 Contexts.back().FirstStartOfName = &Current; 1123 Current.Type = TT_StartOfName; 1124 } else if (Current.is(tok::semi)) { 1125 // Reset FirstStartOfName after finding a semicolon so that a for loop 1126 // with multiple increment statements is not confused with a for loop 1127 // having multiple variable declarations. 1128 Contexts.back().FirstStartOfName = nullptr; 1129 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) { 1130 AutoFound = true; 1131 } else if (Current.is(tok::arrow) && 1132 Style.Language == FormatStyle::LK_Java) { 1133 Current.Type = TT_LambdaArrow; 1134 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration && 1135 Current.NestingLevel == 0) { 1136 Current.Type = TT_TrailingReturnArrow; 1137 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) { 1138 Current.Type = determineStarAmpUsage(Current, 1139 Contexts.back().CanBeExpression && 1140 Contexts.back().IsExpression, 1141 Contexts.back().InTemplateArgument); 1142 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) { 1143 Current.Type = determinePlusMinusCaretUsage(Current); 1144 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret)) 1145 Contexts.back().CaretFound = true; 1146 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) { 1147 Current.Type = determineIncrementUsage(Current); 1148 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) { 1149 Current.Type = TT_UnaryOperator; 1150 } else if (Current.is(tok::question)) { 1151 if (Style.Language == FormatStyle::LK_JavaScript && 1152 Line.MustBeDeclaration && !Contexts.back().IsExpression) { 1153 // In JavaScript, `interface X { foo?(): bar; }` is an optional method 1154 // on the interface, not a ternary expression. 1155 Current.Type = TT_JsTypeOptionalQuestion; 1156 } else { 1157 Current.Type = TT_ConditionalExpr; 1158 } 1159 } else if (Current.isBinaryOperator() && 1160 (!Current.Previous || Current.Previous->isNot(tok::l_square))) { 1161 Current.Type = TT_BinaryOperator; 1162 } else if (Current.is(tok::comment)) { 1163 if (Current.TokenText.startswith("/*")) { 1164 if (Current.TokenText.endswith("*/")) 1165 Current.Type = TT_BlockComment; 1166 else 1167 // The lexer has for some reason determined a comment here. But we 1168 // cannot really handle it, if it isn't properly terminated. 1169 Current.Tok.setKind(tok::unknown); 1170 } else { 1171 Current.Type = TT_LineComment; 1172 } 1173 } else if (Current.is(tok::r_paren)) { 1174 if (rParenEndsCast(Current)) 1175 Current.Type = TT_CastRParen; 1176 if (Current.MatchingParen && Current.Next && 1177 !Current.Next->isBinaryOperator() && 1178 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace, 1179 tok::comma, tok::period, tok::arrow, 1180 tok::coloncolon)) 1181 if (FormatToken *AfterParen = Current.MatchingParen->Next) { 1182 // Make sure this isn't the return type of an Obj-C block declaration 1183 if (AfterParen->Tok.isNot(tok::caret)) { 1184 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) 1185 if (BeforeParen->is(tok::identifier) && 1186 BeforeParen->TokenText == BeforeParen->TokenText.upper() && 1187 (!BeforeParen->Previous || 1188 BeforeParen->Previous->ClosesTemplateDeclaration)) 1189 Current.Type = TT_FunctionAnnotationRParen; 1190 } 1191 } 1192 } else if (Current.is(tok::at) && Current.Next && 1193 Style.Language != FormatStyle::LK_JavaScript && 1194 Style.Language != FormatStyle::LK_Java) { 1195 // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it 1196 // marks declarations and properties that need special formatting. 1197 switch (Current.Next->Tok.getObjCKeywordID()) { 1198 case tok::objc_interface: 1199 case tok::objc_implementation: 1200 case tok::objc_protocol: 1201 Current.Type = TT_ObjCDecl; 1202 break; 1203 case tok::objc_property: 1204 Current.Type = TT_ObjCProperty; 1205 break; 1206 default: 1207 break; 1208 } 1209 } else if (Current.is(tok::period)) { 1210 FormatToken *PreviousNoComment = Current.getPreviousNonComment(); 1211 if (PreviousNoComment && 1212 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) 1213 Current.Type = TT_DesignatedInitializerPeriod; 1214 else if (Style.Language == FormatStyle::LK_Java && Current.Previous && 1215 Current.Previous->isOneOf(TT_JavaAnnotation, 1216 TT_LeadingJavaAnnotation)) { 1217 Current.Type = Current.Previous->Type; 1218 } 1219 } else if (Current.isOneOf(tok::identifier, tok::kw_const) && 1220 Current.Previous && 1221 !Current.Previous->isOneOf(tok::equal, tok::at) && 1222 Line.MightBeFunctionDecl && Contexts.size() == 1) { 1223 // Line.MightBeFunctionDecl can only be true after the parentheses of a 1224 // function declaration have been found. 1225 Current.Type = TT_TrailingAnnotation; 1226 } else if ((Style.Language == FormatStyle::LK_Java || 1227 Style.Language == FormatStyle::LK_JavaScript) && 1228 Current.Previous) { 1229 if (Current.Previous->is(tok::at) && 1230 Current.isNot(Keywords.kw_interface)) { 1231 const FormatToken &AtToken = *Current.Previous; 1232 const FormatToken *Previous = AtToken.getPreviousNonComment(); 1233 if (!Previous || Previous->is(TT_LeadingJavaAnnotation)) 1234 Current.Type = TT_LeadingJavaAnnotation; 1235 else 1236 Current.Type = TT_JavaAnnotation; 1237 } else if (Current.Previous->is(tok::period) && 1238 Current.Previous->isOneOf(TT_JavaAnnotation, 1239 TT_LeadingJavaAnnotation)) { 1240 Current.Type = Current.Previous->Type; 1241 } 1242 } 1243 } 1244 1245 /// \brief Take a guess at whether \p Tok starts a name of a function or 1246 /// variable declaration. 1247 /// 1248 /// This is a heuristic based on whether \p Tok is an identifier following 1249 /// something that is likely a type. 1250 bool isStartOfName(const FormatToken &Tok) { 1251 if (Tok.isNot(tok::identifier) || !Tok.Previous) 1252 return false; 1253 1254 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof, 1255 Keywords.kw_as)) 1256 return false; 1257 if (Style.Language == FormatStyle::LK_JavaScript && 1258 Tok.Previous->is(Keywords.kw_in)) 1259 return false; 1260 1261 // Skip "const" as it does not have an influence on whether this is a name. 1262 FormatToken *PreviousNotConst = Tok.getPreviousNonComment(); 1263 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const)) 1264 PreviousNotConst = PreviousNotConst->getPreviousNonComment(); 1265 1266 if (!PreviousNotConst) 1267 return false; 1268 1269 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) && 1270 PreviousNotConst->Previous && 1271 PreviousNotConst->Previous->is(tok::hash); 1272 1273 if (PreviousNotConst->is(TT_TemplateCloser)) 1274 return PreviousNotConst && PreviousNotConst->MatchingParen && 1275 PreviousNotConst->MatchingParen->Previous && 1276 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) && 1277 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template); 1278 1279 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen && 1280 PreviousNotConst->MatchingParen->Previous && 1281 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype)) 1282 return true; 1283 1284 return (!IsPPKeyword && 1285 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) || 1286 PreviousNotConst->is(TT_PointerOrReference) || 1287 PreviousNotConst->isSimpleTypeSpecifier(); 1288 } 1289 1290 /// \brief Determine whether ')' is ending a cast. 1291 bool rParenEndsCast(const FormatToken &Tok) { 1292 // C-style casts are only used in C++ and Java. 1293 if (!Style.isCpp() && Style.Language != FormatStyle::LK_Java) 1294 return false; 1295 1296 // Empty parens aren't casts and there are no casts at the end of the line. 1297 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen) 1298 return false; 1299 1300 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment(); 1301 if (LeftOfParens) { 1302 // If there is a closing parenthesis left of the current parentheses, 1303 // look past it as these might be chained casts. 1304 if (LeftOfParens->is(tok::r_paren)) { 1305 if (!LeftOfParens->MatchingParen || 1306 !LeftOfParens->MatchingParen->Previous) 1307 return false; 1308 LeftOfParens = LeftOfParens->MatchingParen->Previous; 1309 } 1310 1311 // If there is an identifier (or with a few exceptions a keyword) right 1312 // before the parentheses, this is unlikely to be a cast. 1313 if (LeftOfParens->Tok.getIdentifierInfo() && 1314 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case, 1315 tok::kw_delete)) 1316 return false; 1317 1318 // Certain other tokens right before the parentheses are also signals that 1319 // this cannot be a cast. 1320 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator, 1321 TT_TemplateCloser, tok::ellipsis)) 1322 return false; 1323 } 1324 1325 if (Tok.Next->is(tok::question)) 1326 return false; 1327 1328 // As Java has no function types, a "(" after the ")" likely means that this 1329 // is a cast. 1330 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren)) 1331 return true; 1332 1333 // If a (non-string) literal follows, this is likely a cast. 1334 if (Tok.Next->isNot(tok::string_literal) && 1335 (Tok.Next->Tok.isLiteral() || 1336 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) 1337 return true; 1338 1339 // Heuristically try to determine whether the parentheses contain a type. 1340 bool ParensAreType = 1341 !Tok.Previous || 1342 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) || 1343 Tok.Previous->isSimpleTypeSpecifier(); 1344 bool ParensCouldEndDecl = 1345 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater); 1346 if (ParensAreType && !ParensCouldEndDecl) 1347 return true; 1348 1349 // At this point, we heuristically assume that there are no casts at the 1350 // start of the line. We assume that we have found most cases where there 1351 // are by the logic above, e.g. "(void)x;". 1352 if (!LeftOfParens) 1353 return false; 1354 1355 // Certain token types inside the parentheses mean that this can't be a 1356 // cast. 1357 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok; 1358 Token = Token->Next) 1359 if (Token->is(TT_BinaryOperator)) 1360 return false; 1361 1362 // If the following token is an identifier or 'this', this is a cast. All 1363 // cases where this can be something else are handled above. 1364 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this)) 1365 return true; 1366 1367 if (!Tok.Next->Next) 1368 return false; 1369 1370 // If the next token after the parenthesis is a unary operator, assume 1371 // that this is cast, unless there are unexpected tokens inside the 1372 // parenthesis. 1373 bool NextIsUnary = 1374 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star); 1375 if (!NextIsUnary || Tok.Next->is(tok::plus) || 1376 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) 1377 return false; 1378 // Search for unexpected tokens. 1379 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen; 1380 Prev = Prev->Previous) { 1381 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon)) 1382 return false; 1383 } 1384 return true; 1385 } 1386 1387 /// \brief Return the type of the given token assuming it is * or &. 1388 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression, 1389 bool InTemplateArgument) { 1390 if (Style.Language == FormatStyle::LK_JavaScript) 1391 return TT_BinaryOperator; 1392 1393 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1394 if (!PrevToken) 1395 return TT_UnaryOperator; 1396 1397 const FormatToken *NextToken = Tok.getNextNonComment(); 1398 if (!NextToken || 1399 NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_const) || 1400 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) 1401 return TT_PointerOrReference; 1402 1403 if (PrevToken->is(tok::coloncolon)) 1404 return TT_PointerOrReference; 1405 1406 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace, 1407 tok::comma, tok::semi, tok::kw_return, tok::colon, 1408 tok::equal, tok::kw_delete, tok::kw_sizeof, 1409 tok::kw_throw) || 1410 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr, 1411 TT_UnaryOperator, TT_CastRParen)) 1412 return TT_UnaryOperator; 1413 1414 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare)) 1415 return TT_PointerOrReference; 1416 if (NextToken->is(tok::kw_operator) && !IsExpression) 1417 return TT_PointerOrReference; 1418 if (NextToken->isOneOf(tok::comma, tok::semi)) 1419 return TT_PointerOrReference; 1420 1421 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen) { 1422 FormatToken *TokenBeforeMatchingParen = 1423 PrevToken->MatchingParen->getPreviousNonComment(); 1424 if (TokenBeforeMatchingParen && 1425 TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype)) 1426 return TT_PointerOrReference; 1427 } 1428 1429 if (PrevToken->Tok.isLiteral() || 1430 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true, 1431 tok::kw_false, tok::r_brace) || 1432 NextToken->Tok.isLiteral() || 1433 NextToken->isOneOf(tok::kw_true, tok::kw_false) || 1434 NextToken->isUnaryOperator() || 1435 // If we know we're in a template argument, there are no named 1436 // declarations. Thus, having an identifier on the right-hand side 1437 // indicates a binary operator. 1438 (InTemplateArgument && NextToken->Tok.isAnyIdentifier())) 1439 return TT_BinaryOperator; 1440 1441 // "&&(" is quite unlikely to be two successive unary "&". 1442 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren)) 1443 return TT_BinaryOperator; 1444 1445 // This catches some cases where evaluation order is used as control flow: 1446 // aaa && aaa->f(); 1447 const FormatToken *NextNextToken = NextToken->getNextNonComment(); 1448 if (NextNextToken && NextNextToken->is(tok::arrow)) 1449 return TT_BinaryOperator; 1450 1451 // It is very unlikely that we are going to find a pointer or reference type 1452 // definition on the RHS of an assignment. 1453 if (IsExpression && !Contexts.back().CaretFound) 1454 return TT_BinaryOperator; 1455 1456 return TT_PointerOrReference; 1457 } 1458 1459 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) { 1460 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1461 if (!PrevToken) 1462 return TT_UnaryOperator; 1463 1464 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator) && 1465 !PrevToken->is(tok::exclaim)) 1466 // There aren't any trailing unary operators except for TypeScript's 1467 // non-null operator (!). Thus, this must be squence of leading operators. 1468 return TT_UnaryOperator; 1469 1470 // Use heuristics to recognize unary operators. 1471 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square, 1472 tok::question, tok::colon, tok::kw_return, 1473 tok::kw_case, tok::at, tok::l_brace)) 1474 return TT_UnaryOperator; 1475 1476 // There can't be two consecutive binary operators. 1477 if (PrevToken->is(TT_BinaryOperator)) 1478 return TT_UnaryOperator; 1479 1480 // Fall back to marking the token as binary operator. 1481 return TT_BinaryOperator; 1482 } 1483 1484 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements. 1485 TokenType determineIncrementUsage(const FormatToken &Tok) { 1486 const FormatToken *PrevToken = Tok.getPreviousNonComment(); 1487 if (!PrevToken || PrevToken->is(TT_CastRParen)) 1488 return TT_UnaryOperator; 1489 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier)) 1490 return TT_TrailingUnaryOperator; 1491 1492 return TT_UnaryOperator; 1493 } 1494 1495 SmallVector<Context, 8> Contexts; 1496 1497 const FormatStyle &Style; 1498 AnnotatedLine &Line; 1499 FormatToken *CurrentToken; 1500 bool AutoFound; 1501 const AdditionalKeywords &Keywords; 1502 1503 // Set of "<" tokens that do not open a template parameter list. If parseAngle 1504 // determines that a specific token can't be a template opener, it will make 1505 // same decision irrespective of the decisions for tokens leading up to it. 1506 // Store this information to prevent this from causing exponential runtime. 1507 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess; 1508 }; 1509 1510 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1; 1511 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2; 1512 1513 /// \brief Parses binary expressions by inserting fake parenthesis based on 1514 /// operator precedence. 1515 class ExpressionParser { 1516 public: 1517 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords, 1518 AnnotatedLine &Line) 1519 : Style(Style), Keywords(Keywords), Current(Line.First) {} 1520 1521 /// \brief Parse expressions with the given operator precedence. 1522 void parse(int Precedence = 0) { 1523 // Skip 'return' and ObjC selector colons as they are not part of a binary 1524 // expression. 1525 while (Current && (Current->is(tok::kw_return) || 1526 (Current->is(tok::colon) && 1527 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) 1528 next(); 1529 1530 if (!Current || Precedence > PrecedenceArrowAndPeriod) 1531 return; 1532 1533 // Conditional expressions need to be parsed separately for proper nesting. 1534 if (Precedence == prec::Conditional) { 1535 parseConditionalExpr(); 1536 return; 1537 } 1538 1539 // Parse unary operators, which all have a higher precedence than binary 1540 // operators. 1541 if (Precedence == PrecedenceUnaryOperator) { 1542 parseUnaryOperator(); 1543 return; 1544 } 1545 1546 FormatToken *Start = Current; 1547 FormatToken *LatestOperator = nullptr; 1548 unsigned OperatorIndex = 0; 1549 1550 while (Current) { 1551 // Consume operators with higher precedence. 1552 parse(Precedence + 1); 1553 1554 int CurrentPrecedence = getCurrentPrecedence(); 1555 1556 if (Current && Current->is(TT_SelectorName) && 1557 Precedence == CurrentPrecedence) { 1558 if (LatestOperator) 1559 addFakeParenthesis(Start, prec::Level(Precedence)); 1560 Start = Current; 1561 } 1562 1563 // At the end of the line or when an operator with higher precedence is 1564 // found, insert fake parenthesis and return. 1565 if (!Current || 1566 (Current->closesScope() && 1567 (Current->MatchingParen || Current->is(TT_TemplateString))) || 1568 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) || 1569 (CurrentPrecedence == prec::Conditional && 1570 Precedence == prec::Assignment && Current->is(tok::colon))) { 1571 break; 1572 } 1573 1574 // Consume scopes: (), [], <> and {} 1575 if (Current->opensScope()) { 1576 // In fragment of a JavaScript template string can look like '}..${' and 1577 // thus close a scope and open a new one at the same time. 1578 while (Current && (!Current->closesScope() || Current->opensScope())) { 1579 next(); 1580 parse(); 1581 } 1582 next(); 1583 } else { 1584 // Operator found. 1585 if (CurrentPrecedence == Precedence) { 1586 if (LatestOperator) 1587 LatestOperator->NextOperator = Current; 1588 LatestOperator = Current; 1589 Current->OperatorIndex = OperatorIndex; 1590 ++OperatorIndex; 1591 } 1592 next(/*SkipPastLeadingComments=*/Precedence > 0); 1593 } 1594 } 1595 1596 if (LatestOperator && (Current || Precedence > 0)) { 1597 // LatestOperator->LastOperator = true; 1598 if (Precedence == PrecedenceArrowAndPeriod) { 1599 // Call expressions don't have a binary operator precedence. 1600 addFakeParenthesis(Start, prec::Unknown); 1601 } else { 1602 addFakeParenthesis(Start, prec::Level(Precedence)); 1603 } 1604 } 1605 } 1606 1607 private: 1608 /// \brief Gets the precedence (+1) of the given token for binary operators 1609 /// and other tokens that we treat like binary operators. 1610 int getCurrentPrecedence() { 1611 if (Current) { 1612 const FormatToken *NextNonComment = Current->getNextNonComment(); 1613 if (Current->is(TT_ConditionalExpr)) 1614 return prec::Conditional; 1615 if (NextNonComment && Current->is(TT_SelectorName) && 1616 (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) || 1617 ((Style.Language == FormatStyle::LK_Proto || 1618 Style.Language == FormatStyle::LK_TextProto) && 1619 NextNonComment->is(tok::less)))) 1620 return prec::Assignment; 1621 if (Current->is(TT_JsComputedPropertyName)) 1622 return prec::Assignment; 1623 if (Current->is(TT_LambdaArrow)) 1624 return prec::Comma; 1625 if (Current->is(TT_JsFatArrow)) 1626 return prec::Assignment; 1627 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) || 1628 (Current->is(tok::comment) && NextNonComment && 1629 NextNonComment->is(TT_SelectorName))) 1630 return 0; 1631 if (Current->is(TT_RangeBasedForLoopColon)) 1632 return prec::Comma; 1633 if ((Style.Language == FormatStyle::LK_Java || 1634 Style.Language == FormatStyle::LK_JavaScript) && 1635 Current->is(Keywords.kw_instanceof)) 1636 return prec::Relational; 1637 if (Style.Language == FormatStyle::LK_JavaScript && 1638 Current->isOneOf(Keywords.kw_in, Keywords.kw_as)) 1639 return prec::Relational; 1640 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma)) 1641 return Current->getPrecedence(); 1642 if (Current->isOneOf(tok::period, tok::arrow)) 1643 return PrecedenceArrowAndPeriod; 1644 if ((Style.Language == FormatStyle::LK_Java || 1645 Style.Language == FormatStyle::LK_JavaScript) && 1646 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements, 1647 Keywords.kw_throws)) 1648 return 0; 1649 } 1650 return -1; 1651 } 1652 1653 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) { 1654 Start->FakeLParens.push_back(Precedence); 1655 if (Precedence > prec::Unknown) 1656 Start->StartsBinaryExpression = true; 1657 if (Current) { 1658 FormatToken *Previous = Current->Previous; 1659 while (Previous->is(tok::comment) && Previous->Previous) 1660 Previous = Previous->Previous; 1661 ++Previous->FakeRParens; 1662 if (Precedence > prec::Unknown) 1663 Previous->EndsBinaryExpression = true; 1664 } 1665 } 1666 1667 /// \brief Parse unary operator expressions and surround them with fake 1668 /// parentheses if appropriate. 1669 void parseUnaryOperator() { 1670 llvm::SmallVector<FormatToken *, 2> Tokens; 1671 while (Current && Current->is(TT_UnaryOperator)) { 1672 Tokens.push_back(Current); 1673 next(); 1674 } 1675 parse(PrecedenceArrowAndPeriod); 1676 for (FormatToken *Token : llvm::reverse(Tokens)) 1677 // The actual precedence doesn't matter. 1678 addFakeParenthesis(Token, prec::Unknown); 1679 } 1680 1681 void parseConditionalExpr() { 1682 while (Current && Current->isTrailingComment()) { 1683 next(); 1684 } 1685 FormatToken *Start = Current; 1686 parse(prec::LogicalOr); 1687 if (!Current || !Current->is(tok::question)) 1688 return; 1689 next(); 1690 parse(prec::Assignment); 1691 if (!Current || Current->isNot(TT_ConditionalExpr)) 1692 return; 1693 next(); 1694 parse(prec::Assignment); 1695 addFakeParenthesis(Start, prec::Conditional); 1696 } 1697 1698 void next(bool SkipPastLeadingComments = true) { 1699 if (Current) 1700 Current = Current->Next; 1701 while (Current && 1702 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) && 1703 Current->isTrailingComment()) 1704 Current = Current->Next; 1705 } 1706 1707 const FormatStyle &Style; 1708 const AdditionalKeywords &Keywords; 1709 FormatToken *Current; 1710 }; 1711 1712 } // end anonymous namespace 1713 1714 void TokenAnnotator::setCommentLineLevels( 1715 SmallVectorImpl<AnnotatedLine *> &Lines) { 1716 const AnnotatedLine *NextNonCommentLine = nullptr; 1717 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(), 1718 E = Lines.rend(); 1719 I != E; ++I) { 1720 bool CommentLine = true; 1721 for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) { 1722 if (!Tok->is(tok::comment)) { 1723 CommentLine = false; 1724 break; 1725 } 1726 } 1727 1728 // If the comment is currently aligned with the line immediately following 1729 // it, that's probably intentional and we should keep it. 1730 if (NextNonCommentLine && CommentLine && 1731 NextNonCommentLine->First->NewlinesBefore <= 1 && 1732 NextNonCommentLine->First->OriginalColumn == 1733 (*I)->First->OriginalColumn) { 1734 // Align comments for preprocessor lines with the # in column 0. 1735 // Otherwise, align with the next line. 1736 (*I)->Level = (NextNonCommentLine->Type == LT_PreprocessorDirective || 1737 NextNonCommentLine->Type == LT_ImportStatement) 1738 ? 0 1739 : NextNonCommentLine->Level; 1740 } else { 1741 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr; 1742 } 1743 1744 setCommentLineLevels((*I)->Children); 1745 } 1746 } 1747 1748 static unsigned maxNestingDepth(const AnnotatedLine &Line) { 1749 unsigned Result = 0; 1750 for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next) 1751 Result = std::max(Result, Tok->NestingLevel); 1752 return Result; 1753 } 1754 1755 void TokenAnnotator::annotate(AnnotatedLine &Line) { 1756 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1757 E = Line.Children.end(); 1758 I != E; ++I) { 1759 annotate(**I); 1760 } 1761 AnnotatingParser Parser(Style, Line, Keywords); 1762 Line.Type = Parser.parseLine(); 1763 1764 // With very deep nesting, ExpressionParser uses lots of stack and the 1765 // formatting algorithm is very slow. We're not going to do a good job here 1766 // anyway - it's probably generated code being formatted by mistake. 1767 // Just skip the whole line. 1768 if (maxNestingDepth(Line) > 50) 1769 Line.Type = LT_Invalid; 1770 1771 if (Line.Type == LT_Invalid) 1772 return; 1773 1774 ExpressionParser ExprParser(Style, Keywords, Line); 1775 ExprParser.parse(); 1776 1777 if (Line.startsWith(TT_ObjCMethodSpecifier)) 1778 Line.Type = LT_ObjCMethodDecl; 1779 else if (Line.startsWith(TT_ObjCDecl)) 1780 Line.Type = LT_ObjCDecl; 1781 else if (Line.startsWith(TT_ObjCProperty)) 1782 Line.Type = LT_ObjCProperty; 1783 1784 Line.First->SpacesRequiredBefore = 1; 1785 Line.First->CanBreakBefore = Line.First->MustBreakBefore; 1786 } 1787 1788 // This function heuristically determines whether 'Current' starts the name of a 1789 // function declaration. 1790 static bool isFunctionDeclarationName(const FormatToken &Current, 1791 const AnnotatedLine &Line) { 1792 auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * { 1793 for (; Next; Next = Next->Next) { 1794 if (Next->is(TT_OverloadedOperatorLParen)) 1795 return Next; 1796 if (Next->is(TT_OverloadedOperator)) 1797 continue; 1798 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) { 1799 // For 'new[]' and 'delete[]'. 1800 if (Next->Next && Next->Next->is(tok::l_square) && Next->Next->Next && 1801 Next->Next->Next->is(tok::r_square)) 1802 Next = Next->Next->Next; 1803 continue; 1804 } 1805 1806 break; 1807 } 1808 return nullptr; 1809 }; 1810 1811 // Find parentheses of parameter list. 1812 const FormatToken *Next = Current.Next; 1813 if (Current.is(tok::kw_operator)) { 1814 if (Current.Previous && Current.Previous->is(tok::coloncolon)) 1815 return false; 1816 Next = skipOperatorName(Next); 1817 } else { 1818 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0) 1819 return false; 1820 for (; Next; Next = Next->Next) { 1821 if (Next->is(TT_TemplateOpener)) { 1822 Next = Next->MatchingParen; 1823 } else if (Next->is(tok::coloncolon)) { 1824 Next = Next->Next; 1825 if (!Next) 1826 return false; 1827 if (Next->is(tok::kw_operator)) { 1828 Next = skipOperatorName(Next->Next); 1829 break; 1830 } 1831 if (!Next->is(tok::identifier)) 1832 return false; 1833 } else if (Next->is(tok::l_paren)) { 1834 break; 1835 } else { 1836 return false; 1837 } 1838 } 1839 } 1840 1841 // Check whether parameter list can belong to a function declaration. 1842 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen) 1843 return false; 1844 // If the lines ends with "{", this is likely an function definition. 1845 if (Line.Last->is(tok::l_brace)) 1846 return true; 1847 if (Next->Next == Next->MatchingParen) 1848 return true; // Empty parentheses. 1849 // If there is an &/&& after the r_paren, this is likely a function. 1850 if (Next->MatchingParen->Next && 1851 Next->MatchingParen->Next->is(TT_PointerOrReference)) 1852 return true; 1853 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen; 1854 Tok = Tok->Next) { 1855 if (Tok->is(tok::l_paren) && Tok->MatchingParen) { 1856 Tok = Tok->MatchingParen; 1857 continue; 1858 } 1859 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() || 1860 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis)) 1861 return true; 1862 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) || 1863 Tok->Tok.isLiteral()) 1864 return false; 1865 } 1866 return false; 1867 } 1868 1869 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const { 1870 assert(Line.MightBeFunctionDecl); 1871 1872 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel || 1873 Style.AlwaysBreakAfterReturnType == 1874 FormatStyle::RTBS_TopLevelDefinitions) && 1875 Line.Level > 0) 1876 return false; 1877 1878 switch (Style.AlwaysBreakAfterReturnType) { 1879 case FormatStyle::RTBS_None: 1880 return false; 1881 case FormatStyle::RTBS_All: 1882 case FormatStyle::RTBS_TopLevel: 1883 return true; 1884 case FormatStyle::RTBS_AllDefinitions: 1885 case FormatStyle::RTBS_TopLevelDefinitions: 1886 return Line.mightBeFunctionDefinition(); 1887 } 1888 1889 return false; 1890 } 1891 1892 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) { 1893 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(), 1894 E = Line.Children.end(); 1895 I != E; ++I) { 1896 calculateFormattingInformation(**I); 1897 } 1898 1899 Line.First->TotalLength = 1900 Line.First->IsMultiline ? Style.ColumnLimit 1901 : Line.FirstStartColumn + Line.First->ColumnWidth; 1902 FormatToken *Current = Line.First->Next; 1903 bool InFunctionDecl = Line.MightBeFunctionDecl; 1904 while (Current) { 1905 if (isFunctionDeclarationName(*Current, Line)) 1906 Current->Type = TT_FunctionDeclarationName; 1907 if (Current->is(TT_LineComment)) { 1908 if (Current->Previous->BlockKind == BK_BracedInit && 1909 Current->Previous->opensScope()) 1910 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1; 1911 else 1912 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments; 1913 1914 // If we find a trailing comment, iterate backwards to determine whether 1915 // it seems to relate to a specific parameter. If so, break before that 1916 // parameter to avoid changing the comment's meaning. E.g. don't move 'b' 1917 // to the previous line in: 1918 // SomeFunction(a, 1919 // b, // comment 1920 // c); 1921 if (!Current->HasUnescapedNewline) { 1922 for (FormatToken *Parameter = Current->Previous; Parameter; 1923 Parameter = Parameter->Previous) { 1924 if (Parameter->isOneOf(tok::comment, tok::r_brace)) 1925 break; 1926 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) { 1927 if (!Parameter->Previous->is(TT_CtorInitializerComma) && 1928 Parameter->HasUnescapedNewline) 1929 Parameter->MustBreakBefore = true; 1930 break; 1931 } 1932 } 1933 } 1934 } else if (Current->SpacesRequiredBefore == 0 && 1935 spaceRequiredBefore(Line, *Current)) { 1936 Current->SpacesRequiredBefore = 1; 1937 } 1938 1939 Current->MustBreakBefore = 1940 Current->MustBreakBefore || mustBreakBefore(Line, *Current); 1941 1942 if (!Current->MustBreakBefore && InFunctionDecl && 1943 Current->is(TT_FunctionDeclarationName)) 1944 Current->MustBreakBefore = mustBreakForReturnType(Line); 1945 1946 Current->CanBreakBefore = 1947 Current->MustBreakBefore || canBreakBefore(Line, *Current); 1948 unsigned ChildSize = 0; 1949 if (Current->Previous->Children.size() == 1) { 1950 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last; 1951 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit 1952 : LastOfChild.TotalLength + 1; 1953 } 1954 const FormatToken *Prev = Current->Previous; 1955 if (Current->MustBreakBefore || Prev->Children.size() > 1 || 1956 (Prev->Children.size() == 1 && 1957 Prev->Children[0]->First->MustBreakBefore) || 1958 Current->IsMultiline) 1959 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit; 1960 else 1961 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth + 1962 ChildSize + Current->SpacesRequiredBefore; 1963 1964 if (Current->is(TT_CtorInitializerColon)) 1965 InFunctionDecl = false; 1966 1967 // FIXME: Only calculate this if CanBreakBefore is true once static 1968 // initializers etc. are sorted out. 1969 // FIXME: Move magic numbers to a better place. 1970 Current->SplitPenalty = 20 * Current->BindingStrength + 1971 splitPenalty(Line, *Current, InFunctionDecl); 1972 1973 Current = Current->Next; 1974 } 1975 1976 calculateUnbreakableTailLengths(Line); 1977 unsigned IndentLevel = Line.Level; 1978 for (Current = Line.First; Current != nullptr; Current = Current->Next) { 1979 if (Current->Role) 1980 Current->Role->precomputeFormattingInfos(Current); 1981 if (Current->MatchingParen && 1982 Current->MatchingParen->opensBlockOrBlockTypeList(Style)) { 1983 assert(IndentLevel > 0); 1984 --IndentLevel; 1985 } 1986 Current->IndentLevel = IndentLevel; 1987 if (Current->opensBlockOrBlockTypeList(Style)) 1988 ++IndentLevel; 1989 } 1990 1991 DEBUG({ printDebugInfo(Line); }); 1992 } 1993 1994 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) { 1995 unsigned UnbreakableTailLength = 0; 1996 FormatToken *Current = Line.Last; 1997 while (Current) { 1998 Current->UnbreakableTailLength = UnbreakableTailLength; 1999 if (Current->CanBreakBefore || 2000 Current->isOneOf(tok::comment, tok::string_literal)) { 2001 UnbreakableTailLength = 0; 2002 } else { 2003 UnbreakableTailLength += 2004 Current->ColumnWidth + Current->SpacesRequiredBefore; 2005 } 2006 Current = Current->Previous; 2007 } 2008 } 2009 2010 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line, 2011 const FormatToken &Tok, 2012 bool InFunctionDecl) { 2013 const FormatToken &Left = *Tok.Previous; 2014 const FormatToken &Right = Tok; 2015 2016 if (Left.is(tok::semi)) 2017 return 0; 2018 2019 if (Style.Language == FormatStyle::LK_Java) { 2020 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws)) 2021 return 1; 2022 if (Right.is(Keywords.kw_implements)) 2023 return 2; 2024 if (Left.is(tok::comma) && Left.NestingLevel == 0) 2025 return 3; 2026 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2027 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma)) 2028 return 100; 2029 if (Left.is(TT_JsTypeColon)) 2030 return 35; 2031 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2032 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2033 return 100; 2034 // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()". 2035 if (Left.opensScope() && Right.closesScope()) 2036 return 200; 2037 } 2038 2039 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2040 return 1; 2041 if (Right.is(tok::l_square)) { 2042 if (Style.Language == FormatStyle::LK_Proto) 2043 return 1; 2044 if (Left.is(tok::r_square)) 2045 return 200; 2046 // Slightly prefer formatting local lambda definitions like functions. 2047 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal)) 2048 return 35; 2049 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2050 TT_ArrayInitializerLSquare, 2051 TT_DesignatedInitializerLSquare)) 2052 return 500; 2053 } 2054 2055 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2056 Right.is(tok::kw_operator)) { 2057 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt) 2058 return 3; 2059 if (Left.is(TT_StartOfName)) 2060 return 110; 2061 if (InFunctionDecl && Right.NestingLevel == 0) 2062 return Style.PenaltyReturnTypeOnItsOwnLine; 2063 return 200; 2064 } 2065 if (Right.is(TT_PointerOrReference)) 2066 return 190; 2067 if (Right.is(TT_LambdaArrow)) 2068 return 110; 2069 if (Left.is(tok::equal) && Right.is(tok::l_brace)) 2070 return 160; 2071 if (Left.is(TT_CastRParen)) 2072 return 100; 2073 if (Left.is(tok::coloncolon) || 2074 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) 2075 return 500; 2076 if (Left.isOneOf(tok::kw_class, tok::kw_struct)) 2077 return 5000; 2078 if (Left.is(tok::comment)) 2079 return 1000; 2080 2081 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon, 2082 TT_CtorInitializerColon)) 2083 return 2; 2084 2085 if (Right.isMemberAccess()) { 2086 // Breaking before the "./->" of a chained call/member access is reasonably 2087 // cheap, as formatting those with one call per line is generally 2088 // desirable. In particular, it should be cheaper to break before the call 2089 // than it is to break inside a call's parameters, which could lead to weird 2090 // "hanging" indents. The exception is the very last "./->" to support this 2091 // frequent pattern: 2092 // 2093 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc( 2094 // dddddddd); 2095 // 2096 // which might otherwise be blown up onto many lines. Here, clang-format 2097 // won't produce "hanging" indents anyway as there is no other trailing 2098 // call. 2099 // 2100 // Also apply higher penalty is not a call as that might lead to a wrapping 2101 // like: 2102 // 2103 // aaaaaaa 2104 // .aaaaaaaaa.bbbbbbbb(cccccccc); 2105 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope() 2106 ? 150 2107 : 35; 2108 } 2109 2110 if (Right.is(TT_TrailingAnnotation) && 2111 (!Right.Next || Right.Next->isNot(tok::l_paren))) { 2112 // Moving trailing annotations to the next line is fine for ObjC method 2113 // declarations. 2114 if (Line.startsWith(TT_ObjCMethodSpecifier)) 2115 return 10; 2116 // Generally, breaking before a trailing annotation is bad unless it is 2117 // function-like. It seems to be especially preferable to keep standard 2118 // annotations (i.e. "const", "final" and "override") on the same line. 2119 // Use a slightly higher penalty after ")" so that annotations like 2120 // "const override" are kept together. 2121 bool is_short_annotation = Right.TokenText.size() < 10; 2122 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0); 2123 } 2124 2125 // In for-loops, prefer breaking at ',' and ';'. 2126 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal)) 2127 return 4; 2128 2129 // In Objective-C method expressions, prefer breaking before "param:" over 2130 // breaking after it. 2131 if (Right.is(TT_SelectorName)) 2132 return 0; 2133 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr)) 2134 return Line.MightBeFunctionDecl ? 50 : 500; 2135 2136 if (Left.is(tok::l_paren) && InFunctionDecl && 2137 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) 2138 return 100; 2139 if (Left.is(tok::l_paren) && Left.Previous && 2140 (Left.Previous->isOneOf(tok::kw_if, tok::kw_for) || 2141 Left.Previous->endsSequence(tok::kw_constexpr, tok::kw_if))) 2142 return 1000; 2143 if (Left.is(tok::equal) && InFunctionDecl) 2144 return 110; 2145 if (Right.is(tok::r_brace)) 2146 return 1; 2147 if (Left.is(TT_TemplateOpener)) 2148 return 100; 2149 if (Left.opensScope()) { 2150 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign) 2151 return 0; 2152 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter 2153 : 19; 2154 } 2155 if (Left.is(TT_JavaAnnotation)) 2156 return 50; 2157 2158 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous && 2159 Left.Previous->isLabelString() && 2160 (Left.NextOperator || Left.OperatorIndex != 0)) 2161 return 50; 2162 if (Right.is(tok::plus) && Left.isLabelString() && 2163 (Right.NextOperator || Right.OperatorIndex != 0)) 2164 return 25; 2165 if (Left.is(tok::comma)) 2166 return 1; 2167 if (Right.is(tok::lessless) && Left.isLabelString() && 2168 (Right.NextOperator || Right.OperatorIndex != 1)) 2169 return 25; 2170 if (Right.is(tok::lessless)) { 2171 // Breaking at a << is really cheap. 2172 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0) 2173 // Slightly prefer to break before the first one in log-like statements. 2174 return 2; 2175 return 1; 2176 } 2177 if (Left.is(TT_ConditionalExpr)) 2178 return prec::Conditional; 2179 prec::Level Level = Left.getPrecedence(); 2180 if (Level == prec::Unknown) 2181 Level = Right.getPrecedence(); 2182 if (Level == prec::Assignment) 2183 return Style.PenaltyBreakAssignment; 2184 if (Level != prec::Unknown) 2185 return Level; 2186 2187 return 3; 2188 } 2189 2190 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line, 2191 const FormatToken &Left, 2192 const FormatToken &Right) { 2193 if (Left.is(tok::kw_return) && Right.isNot(tok::semi)) 2194 return true; 2195 if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java) 2196 return true; 2197 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty && 2198 Left.Tok.getObjCKeywordID() == tok::objc_property) 2199 return true; 2200 if (Right.is(tok::hashhash)) 2201 return Left.is(tok::hash); 2202 if (Left.isOneOf(tok::hashhash, tok::hash)) 2203 return Right.is(tok::hash); 2204 if (Left.is(tok::l_paren) && Right.is(tok::r_paren)) 2205 return Style.SpaceInEmptyParentheses; 2206 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) 2207 return (Right.is(TT_CastRParen) || 2208 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen))) 2209 ? Style.SpacesInCStyleCastParentheses 2210 : Style.SpacesInParentheses; 2211 if (Right.isOneOf(tok::semi, tok::comma)) 2212 return false; 2213 if (Right.is(tok::less) && Line.Type == LT_ObjCDecl && 2214 Style.ObjCSpaceBeforeProtocolList) 2215 return true; 2216 if (Right.is(tok::less) && Left.is(tok::kw_template)) 2217 return Style.SpaceAfterTemplateKeyword; 2218 if (Left.isOneOf(tok::exclaim, tok::tilde)) 2219 return false; 2220 if (Left.is(tok::at) && 2221 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant, 2222 tok::numeric_constant, tok::l_paren, tok::l_brace, 2223 tok::kw_true, tok::kw_false)) 2224 return false; 2225 if (Left.is(tok::colon)) 2226 return !Left.is(TT_ObjCMethodExpr); 2227 if (Left.is(tok::coloncolon)) 2228 return false; 2229 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) 2230 return false; 2231 if (Right.is(tok::ellipsis)) 2232 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous && 2233 Left.Previous->is(tok::kw_case)); 2234 if (Left.is(tok::l_square) && Right.is(tok::amp)) 2235 return false; 2236 if (Right.is(TT_PointerOrReference)) { 2237 if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) { 2238 if (!Left.MatchingParen) 2239 return true; 2240 FormatToken *TokenBeforeMatchingParen = 2241 Left.MatchingParen->getPreviousNonComment(); 2242 if (!TokenBeforeMatchingParen || 2243 !TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype)) 2244 return true; 2245 } 2246 return (Left.Tok.isLiteral() || 2247 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) && 2248 (Style.PointerAlignment != FormatStyle::PAS_Left || 2249 (Line.IsMultiVariableDeclStmt && 2250 (Left.NestingLevel == 0 || 2251 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for))))))); 2252 } 2253 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) && 2254 (!Left.is(TT_PointerOrReference) || 2255 (Style.PointerAlignment != FormatStyle::PAS_Right && 2256 !Line.IsMultiVariableDeclStmt))) 2257 return true; 2258 if (Left.is(TT_PointerOrReference)) 2259 return Right.Tok.isLiteral() || Right.is(TT_BlockComment) || 2260 (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) && 2261 !Right.is(TT_StartOfName)) || 2262 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) || 2263 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare, 2264 tok::l_paren) && 2265 (Style.PointerAlignment != FormatStyle::PAS_Right && 2266 !Line.IsMultiVariableDeclStmt) && 2267 Left.Previous && 2268 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon)); 2269 if (Right.is(tok::star) && Left.is(tok::l_paren)) 2270 return false; 2271 if (Left.is(tok::l_square)) 2272 return (Left.is(TT_ArrayInitializerLSquare) && 2273 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square)) || 2274 (Left.isOneOf(TT_ArraySubscriptLSquare, 2275 TT_StructuredBindingLSquare) && 2276 Style.SpacesInSquareBrackets && Right.isNot(tok::r_square)); 2277 if (Right.is(tok::r_square)) 2278 return Right.MatchingParen && 2279 ((Style.SpacesInContainerLiterals && 2280 Right.MatchingParen->is(TT_ArrayInitializerLSquare)) || 2281 (Style.SpacesInSquareBrackets && 2282 Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare, 2283 TT_StructuredBindingLSquare))); 2284 if (Right.is(tok::l_square) && 2285 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare, 2286 TT_DesignatedInitializerLSquare, 2287 TT_StructuredBindingLSquare) && 2288 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral)) 2289 return false; 2290 if (Left.is(tok::l_brace) && Right.is(tok::r_brace)) 2291 return !Left.Children.empty(); // No spaces in "{}". 2292 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) || 2293 (Right.is(tok::r_brace) && Right.MatchingParen && 2294 Right.MatchingParen->BlockKind != BK_Block)) 2295 return !Style.Cpp11BracedListStyle; 2296 if (Left.is(TT_BlockComment)) 2297 return !Left.TokenText.endswith("=*/"); 2298 if (Right.is(tok::l_paren)) { 2299 if (Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) 2300 return true; 2301 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) || 2302 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never && 2303 (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while, 2304 tok::kw_switch, tok::kw_case, TT_ForEachMacro, 2305 TT_ObjCForIn) || 2306 Left.endsSequence(tok::kw_constexpr, tok::kw_if) || 2307 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch, 2308 tok::kw_new, tok::kw_delete) && 2309 (!Left.Previous || Left.Previous->isNot(tok::period))))) || 2310 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always && 2311 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() || 2312 Left.is(tok::r_paren)) && 2313 Line.Type != LT_PreprocessorDirective); 2314 } 2315 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword) 2316 return false; 2317 if (Right.is(TT_UnaryOperator)) 2318 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) && 2319 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr)); 2320 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square, 2321 tok::r_paren) || 2322 Left.isSimpleTypeSpecifier()) && 2323 Right.is(tok::l_brace) && Right.getNextNonComment() && 2324 Right.BlockKind != BK_Block) 2325 return false; 2326 if (Left.is(tok::period) || Right.is(tok::period)) 2327 return false; 2328 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L") 2329 return false; 2330 if (Left.is(TT_TemplateCloser) && Left.MatchingParen && 2331 Left.MatchingParen->Previous && 2332 Left.MatchingParen->Previous->is(tok::period)) 2333 // A.<B<C<...>>>DoSomething(); 2334 return false; 2335 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square)) 2336 return false; 2337 return true; 2338 } 2339 2340 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line, 2341 const FormatToken &Right) { 2342 const FormatToken &Left = *Right.Previous; 2343 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo()) 2344 return true; // Never ever merge two identifiers. 2345 if (Style.isCpp()) { 2346 if (Left.is(tok::kw_operator)) 2347 return Right.is(tok::coloncolon); 2348 } else if (Style.Language == FormatStyle::LK_Proto || 2349 Style.Language == FormatStyle::LK_TextProto) { 2350 if (Right.is(tok::period) && 2351 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required, 2352 Keywords.kw_repeated, Keywords.kw_extend)) 2353 return true; 2354 if (Right.is(tok::l_paren) && 2355 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) 2356 return true; 2357 if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName)) 2358 return true; 2359 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2360 if (Left.is(TT_JsFatArrow)) 2361 return true; 2362 // for await ( ... 2363 if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous && 2364 Left.Previous->is(tok::kw_for)) 2365 return true; 2366 if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) && 2367 Right.MatchingParen) { 2368 const FormatToken *Next = Right.MatchingParen->getNextNonComment(); 2369 // An async arrow function, for example: `x = async () => foo();`, 2370 // as opposed to calling a function called async: `x = async();` 2371 if (Next && Next->is(TT_JsFatArrow)) 2372 return true; 2373 } 2374 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) || 2375 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) 2376 return false; 2377 // In tagged template literals ("html`bar baz`"), there is no space between 2378 // the tag identifier and the template string. getIdentifierInfo makes sure 2379 // that the identifier is not a pseudo keyword like `yield`, either. 2380 if (Left.is(tok::identifier) && Keywords.IsJavaScriptIdentifier(Left) && 2381 Right.is(TT_TemplateString)) 2382 return false; 2383 if (Right.is(tok::star) && 2384 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) 2385 return false; 2386 if (Right.isOneOf(tok::l_brace, tok::l_square) && 2387 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield, 2388 Keywords.kw_extends, Keywords.kw_implements)) 2389 return true; 2390 if (Right.is(tok::l_paren)) { 2391 // JS methods can use some keywords as names (e.g. `delete()`). 2392 if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo()) 2393 return false; 2394 // Valid JS method names can include keywords, e.g. `foo.delete()` or 2395 // `bar.instanceof()`. Recognize call positions by preceding period. 2396 if (Left.Previous && Left.Previous->is(tok::period) && 2397 Left.Tok.getIdentifierInfo()) 2398 return false; 2399 // Additional unary JavaScript operators that need a space after. 2400 if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof, 2401 tok::kw_void)) 2402 return true; 2403 } 2404 if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in, 2405 tok::kw_const) || 2406 // "of" is only a keyword if it appears after another identifier 2407 // (e.g. as "const x of y" in a for loop), or after a destructuring 2408 // operation (const [x, y] of z, const {a, b} of c). 2409 (Left.is(Keywords.kw_of) && Left.Previous && 2410 (Left.Previous->Tok.getIdentifierInfo() || 2411 Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) && 2412 (!Left.Previous || !Left.Previous->is(tok::period))) 2413 return true; 2414 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous && 2415 Left.Previous->is(tok::period) && Right.is(tok::l_paren)) 2416 return false; 2417 if (Left.is(Keywords.kw_as) && 2418 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren)) 2419 return true; 2420 if (Left.is(tok::kw_default) && Left.Previous && 2421 Left.Previous->is(tok::kw_export)) 2422 return true; 2423 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace)) 2424 return true; 2425 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion)) 2426 return false; 2427 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator)) 2428 return false; 2429 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) && 2430 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) 2431 return false; 2432 if (Left.is(tok::ellipsis)) 2433 return false; 2434 if (Left.is(TT_TemplateCloser) && 2435 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square, 2436 Keywords.kw_implements, Keywords.kw_extends)) 2437 // Type assertions ('<type>expr') are not followed by whitespace. Other 2438 // locations that should have whitespace following are identified by the 2439 // above set of follower tokens. 2440 return false; 2441 if (Right.is(TT_JsNonNullAssertion)) 2442 return false; 2443 if (Left.is(TT_JsNonNullAssertion) && 2444 Right.isOneOf(Keywords.kw_as, Keywords.kw_in)) 2445 return true; // "x! as string", "x! in y" 2446 } else if (Style.Language == FormatStyle::LK_Java) { 2447 if (Left.is(tok::r_square) && Right.is(tok::l_brace)) 2448 return true; 2449 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) 2450 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never; 2451 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private, 2452 tok::kw_protected) || 2453 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract, 2454 Keywords.kw_native)) && 2455 Right.is(TT_TemplateOpener)) 2456 return true; 2457 } 2458 if (Left.is(TT_ImplicitStringLiteral)) 2459 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2460 if (Line.Type == LT_ObjCMethodDecl) { 2461 if (Left.is(TT_ObjCMethodSpecifier)) 2462 return true; 2463 if (Left.is(tok::r_paren) && Right.is(tok::identifier)) 2464 // Don't space between ')' and <id> 2465 return false; 2466 } 2467 if (Line.Type == LT_ObjCProperty && 2468 (Right.is(tok::equal) || Left.is(tok::equal))) 2469 return false; 2470 2471 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) || 2472 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) 2473 return true; 2474 if (Right.is(TT_OverloadedOperatorLParen)) 2475 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2476 if (Left.is(tok::comma)) 2477 return true; 2478 if (Right.is(tok::comma)) 2479 return false; 2480 if (Right.isOneOf(TT_CtorInitializerColon, TT_ObjCBlockLParen)) 2481 return true; 2482 if (Right.is(tok::colon)) { 2483 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) || 2484 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi)) 2485 return false; 2486 if (Right.is(TT_ObjCMethodExpr)) 2487 return false; 2488 if (Left.is(tok::question)) 2489 return false; 2490 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon)) 2491 return false; 2492 if (Right.is(TT_DictLiteral)) 2493 return Style.SpacesInContainerLiterals; 2494 return true; 2495 } 2496 if (Left.is(TT_UnaryOperator)) 2497 return Right.is(TT_BinaryOperator); 2498 2499 // If the next token is a binary operator or a selector name, we have 2500 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly. 2501 if (Left.is(TT_CastRParen)) 2502 return Style.SpaceAfterCStyleCast || 2503 Right.isOneOf(TT_BinaryOperator, TT_SelectorName); 2504 2505 if (Left.is(tok::greater) && Right.is(tok::greater)) 2506 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) && 2507 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles); 2508 if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) || 2509 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) || 2510 (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod))) 2511 return false; 2512 if (!Style.SpaceBeforeAssignmentOperators && 2513 Right.getPrecedence() == prec::Assignment) 2514 return false; 2515 if (Right.is(tok::coloncolon) && Left.is(tok::identifier)) 2516 // Generally don't remove existing spaces between an identifier and "::". 2517 // The identifier might actually be a macro name such as ALWAYS_INLINE. If 2518 // this turns out to be too lenient, add analysis of the identifier itself. 2519 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd(); 2520 if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment)) 2521 return (Left.is(TT_TemplateOpener) && 2522 Style.Standard == FormatStyle::LS_Cpp03) || 2523 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square, 2524 tok::kw___super, TT_TemplateCloser, 2525 TT_TemplateOpener)); 2526 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser))) 2527 return Style.SpacesInAngles; 2528 // Space before TT_StructuredBindingLSquare. 2529 if (Right.is(TT_StructuredBindingLSquare)) 2530 return !Left.isOneOf(tok::amp, tok::ampamp) || 2531 Style.PointerAlignment != FormatStyle::PAS_Right; 2532 // Space before & or && following a TT_StructuredBindingLSquare. 2533 if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) && 2534 Right.isOneOf(tok::amp, tok::ampamp)) 2535 return Style.PointerAlignment != FormatStyle::PAS_Left; 2536 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) || 2537 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && 2538 !Right.is(tok::r_paren))) 2539 return true; 2540 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) && 2541 Right.isNot(TT_FunctionTypeLParen)) 2542 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always; 2543 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) && 2544 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen)) 2545 return false; 2546 if (Right.is(tok::less) && Left.isNot(tok::l_paren) && 2547 Line.startsWith(tok::hash)) 2548 return true; 2549 if (Right.is(TT_TrailingUnaryOperator)) 2550 return false; 2551 if (Left.is(TT_RegexLiteral)) 2552 return false; 2553 return spaceRequiredBetween(Line, Left, Right); 2554 } 2555 2556 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style. 2557 static bool isAllmanBrace(const FormatToken &Tok) { 2558 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block && 2559 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral); 2560 } 2561 2562 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line, 2563 const FormatToken &Right) { 2564 const FormatToken &Left = *Right.Previous; 2565 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0) 2566 return true; 2567 2568 if (Style.Language == FormatStyle::LK_JavaScript) { 2569 // FIXME: This might apply to other languages and token kinds. 2570 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous && 2571 Left.Previous->is(tok::string_literal)) 2572 return true; 2573 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 && 2574 Left.Previous && Left.Previous->is(tok::equal) && 2575 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export, 2576 tok::kw_const) && 2577 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match 2578 // above. 2579 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) 2580 // Object literals on the top level of a file are treated as "enum-style". 2581 // Each key/value pair is put on a separate line, instead of bin-packing. 2582 return true; 2583 if (Left.is(tok::l_brace) && Line.Level == 0 && 2584 (Line.startsWith(tok::kw_enum) || 2585 Line.startsWith(tok::kw_const, tok::kw_enum) || 2586 Line.startsWith(tok::kw_export, tok::kw_enum) || 2587 Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum))) 2588 // JavaScript top-level enum key/value pairs are put on separate lines 2589 // instead of bin-packing. 2590 return true; 2591 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && 2592 !Left.Children.empty()) 2593 // Support AllowShortFunctionsOnASingleLine for JavaScript. 2594 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None || 2595 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty || 2596 (Left.NestingLevel == 0 && Line.Level == 0 && 2597 Style.AllowShortFunctionsOnASingleLine & 2598 FormatStyle::SFS_InlineOnly); 2599 } else if (Style.Language == FormatStyle::LK_Java) { 2600 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next && 2601 Right.Next->is(tok::string_literal)) 2602 return true; 2603 } else if (Style.Language == FormatStyle::LK_Cpp || 2604 Style.Language == FormatStyle::LK_ObjC || 2605 Style.Language == FormatStyle::LK_Proto) { 2606 if (Left.isStringLiteral() && Right.isStringLiteral()) 2607 return true; 2608 } 2609 2610 // If the last token before a '}', ']', or ')' is a comma or a trailing 2611 // comment, the intention is to insert a line break after it in order to make 2612 // shuffling around entries easier. Import statements, especially in 2613 // JavaScript, can be an exception to this rule. 2614 if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) { 2615 const FormatToken *BeforeClosingBrace = nullptr; 2616 if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) || 2617 (Style.Language == FormatStyle::LK_JavaScript && 2618 Left.is(tok::l_paren))) && 2619 Left.BlockKind != BK_Block && Left.MatchingParen) 2620 BeforeClosingBrace = Left.MatchingParen->Previous; 2621 else if (Right.MatchingParen && 2622 (Right.MatchingParen->isOneOf(tok::l_brace, 2623 TT_ArrayInitializerLSquare) || 2624 (Style.Language == FormatStyle::LK_JavaScript && 2625 Right.MatchingParen->is(tok::l_paren)))) 2626 BeforeClosingBrace = &Left; 2627 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) || 2628 BeforeClosingBrace->isTrailingComment())) 2629 return true; 2630 } 2631 2632 if (Right.is(tok::comment)) 2633 return Left.BlockKind != BK_BracedInit && 2634 Left.isNot(TT_CtorInitializerColon) && 2635 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline); 2636 if (Left.isTrailingComment()) 2637 return true; 2638 if (Right.Previous->IsUnterminatedLiteral) 2639 return true; 2640 if (Right.is(tok::lessless) && Right.Next && 2641 Right.Previous->is(tok::string_literal) && 2642 Right.Next->is(tok::string_literal)) 2643 return true; 2644 if (Right.Previous->ClosesTemplateDeclaration && 2645 Right.Previous->MatchingParen && 2646 Right.Previous->MatchingParen->NestingLevel == 0 && 2647 Style.AlwaysBreakTemplateDeclarations) 2648 return true; 2649 if (Right.is(TT_CtorInitializerComma) && 2650 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 2651 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 2652 return true; 2653 if (Right.is(TT_CtorInitializerColon) && 2654 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma && 2655 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 2656 return true; 2657 // Break only if we have multiple inheritance. 2658 if (Style.BreakBeforeInheritanceComma && Right.is(TT_InheritanceComma)) 2659 return true; 2660 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) 2661 // Raw string literals are special wrt. line breaks. The author has made a 2662 // deliberate choice and might have aligned the contents of the string 2663 // literal accordingly. Thus, we try keep existing line breaks. 2664 return Right.NewlinesBefore > 0; 2665 if ((Right.Previous->is(tok::l_brace) || 2666 (Right.Previous->is(tok::less) && Right.Previous->Previous && 2667 Right.Previous->Previous->is(tok::equal))) && 2668 Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) { 2669 // Don't put enums or option definitions onto single lines in protocol 2670 // buffers. 2671 return true; 2672 } 2673 if (Right.is(TT_InlineASMBrace)) 2674 return Right.HasUnescapedNewline; 2675 if (isAllmanBrace(Left) || isAllmanBrace(Right)) 2676 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) || 2677 (Line.startsWith(tok::kw_typedef, tok::kw_enum) && 2678 Style.BraceWrapping.AfterEnum) || 2679 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) || 2680 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct); 2681 if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine) 2682 return true; 2683 2684 if ((Style.Language == FormatStyle::LK_Java || 2685 Style.Language == FormatStyle::LK_JavaScript) && 2686 Left.is(TT_LeadingJavaAnnotation) && 2687 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) && 2688 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) 2689 return true; 2690 2691 return false; 2692 } 2693 2694 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line, 2695 const FormatToken &Right) { 2696 const FormatToken &Left = *Right.Previous; 2697 2698 // Language-specific stuff. 2699 if (Style.Language == FormatStyle::LK_Java) { 2700 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2701 Keywords.kw_implements)) 2702 return false; 2703 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends, 2704 Keywords.kw_implements)) 2705 return true; 2706 } else if (Style.Language == FormatStyle::LK_JavaScript) { 2707 const FormatToken *NonComment = Right.getPreviousNonComment(); 2708 if (NonComment && 2709 NonComment->isOneOf( 2710 tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break, 2711 tok::kw_throw, Keywords.kw_interface, Keywords.kw_type, 2712 tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected, 2713 Keywords.kw_readonly, Keywords.kw_abstract, Keywords.kw_get, 2714 Keywords.kw_set, Keywords.kw_async, Keywords.kw_await)) 2715 return false; // Otherwise automatic semicolon insertion would trigger. 2716 if (Right.NestingLevel == 0 && 2717 (Left.Tok.getIdentifierInfo() || 2718 Left.isOneOf(tok::r_square, tok::r_paren)) && 2719 Right.isOneOf(tok::l_square, tok::l_paren)) 2720 return false; // Otherwise automatic semicolon insertion would trigger. 2721 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace)) 2722 return false; 2723 if (Left.is(TT_JsTypeColon)) 2724 return true; 2725 if (Right.NestingLevel == 0 && Right.is(Keywords.kw_is)) 2726 return false; 2727 if (Left.is(Keywords.kw_in)) 2728 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None; 2729 if (Right.is(Keywords.kw_in)) 2730 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None; 2731 if (Right.is(Keywords.kw_as)) 2732 return false; // must not break before as in 'x as type' casts 2733 if (Left.is(Keywords.kw_as)) 2734 return true; 2735 if (Left.is(TT_JsNonNullAssertion)) 2736 return true; 2737 if (Left.is(Keywords.kw_declare) && 2738 Right.isOneOf(Keywords.kw_module, tok::kw_namespace, 2739 Keywords.kw_function, tok::kw_class, tok::kw_enum, 2740 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var, 2741 Keywords.kw_let, tok::kw_const)) 2742 // See grammar for 'declare' statements at: 2743 // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10 2744 return false; 2745 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) && 2746 Right.isOneOf(tok::identifier, tok::string_literal)) 2747 return false; // must not break in "module foo { ...}" 2748 if (Right.is(TT_TemplateString) && Right.closesScope()) 2749 return false; 2750 if (Left.is(TT_TemplateString) && Left.opensScope()) 2751 return true; 2752 } 2753 2754 if (Left.is(tok::at)) 2755 return false; 2756 if (Left.Tok.getObjCKeywordID() == tok::objc_interface) 2757 return false; 2758 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation)) 2759 return !Right.is(tok::l_paren); 2760 if (Right.is(TT_PointerOrReference)) 2761 return Line.IsMultiVariableDeclStmt || 2762 (Style.PointerAlignment == FormatStyle::PAS_Right && 2763 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName))); 2764 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) || 2765 Right.is(tok::kw_operator)) 2766 return true; 2767 if (Left.is(TT_PointerOrReference)) 2768 return false; 2769 if (Right.isTrailingComment()) 2770 // We rely on MustBreakBefore being set correctly here as we should not 2771 // change the "binding" behavior of a comment. 2772 // The first comment in a braced lists is always interpreted as belonging to 2773 // the first list element. Otherwise, it should be placed outside of the 2774 // list. 2775 return Left.BlockKind == BK_BracedInit || 2776 (Left.is(TT_CtorInitializerColon) && 2777 Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon); 2778 if (Left.is(tok::question) && Right.is(tok::colon)) 2779 return false; 2780 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question)) 2781 return Style.BreakBeforeTernaryOperators; 2782 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question)) 2783 return !Style.BreakBeforeTernaryOperators; 2784 if (Right.is(TT_InheritanceColon)) 2785 return true; 2786 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) && 2787 Left.isNot(TT_SelectorName)) 2788 return true; 2789 if (Right.is(tok::colon) && 2790 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) 2791 return false; 2792 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) 2793 return true; 2794 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next && 2795 Right.Next->is(TT_ObjCMethodExpr))) 2796 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls. 2797 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty) 2798 return true; 2799 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen)) 2800 return true; 2801 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen, 2802 TT_OverloadedOperator)) 2803 return false; 2804 if (Left.is(TT_RangeBasedForLoopColon)) 2805 return true; 2806 if (Right.is(TT_RangeBasedForLoopColon)) 2807 return false; 2808 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener)) 2809 return true; 2810 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) || 2811 Left.is(tok::kw_operator)) 2812 return false; 2813 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) && 2814 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) 2815 return false; 2816 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen)) 2817 return false; 2818 if (Left.is(tok::l_paren) && Left.Previous && 2819 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) 2820 return false; 2821 if (Right.is(TT_ImplicitStringLiteral)) 2822 return false; 2823 2824 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser)) 2825 return false; 2826 if (Right.is(tok::r_square) && Right.MatchingParen && 2827 Right.MatchingParen->is(TT_LambdaLSquare)) 2828 return false; 2829 2830 // We only break before r_brace if there was a corresponding break before 2831 // the l_brace, which is tracked by BreakBeforeClosingBrace. 2832 if (Right.is(tok::r_brace)) 2833 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block; 2834 2835 // Allow breaking after a trailing annotation, e.g. after a method 2836 // declaration. 2837 if (Left.is(TT_TrailingAnnotation)) 2838 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren, 2839 tok::less, tok::coloncolon); 2840 2841 if (Right.is(tok::kw___attribute)) 2842 return true; 2843 2844 if (Left.is(tok::identifier) && Right.is(tok::string_literal)) 2845 return true; 2846 2847 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral)) 2848 return true; 2849 2850 if (Left.is(TT_CtorInitializerColon)) 2851 return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon; 2852 if (Right.is(TT_CtorInitializerColon)) 2853 return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon; 2854 if (Left.is(TT_CtorInitializerComma) && 2855 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 2856 return false; 2857 if (Right.is(TT_CtorInitializerComma) && 2858 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) 2859 return true; 2860 if (Left.is(TT_InheritanceComma) && Style.BreakBeforeInheritanceComma) 2861 return false; 2862 if (Right.is(TT_InheritanceComma) && Style.BreakBeforeInheritanceComma) 2863 return true; 2864 if ((Left.is(tok::greater) && Right.is(tok::greater)) || 2865 (Left.is(tok::less) && Right.is(tok::less))) 2866 return false; 2867 if (Right.is(TT_BinaryOperator) && 2868 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None && 2869 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All || 2870 Right.getPrecedence() != prec::Assignment)) 2871 return true; 2872 if (Left.is(TT_ArrayInitializerLSquare)) 2873 return true; 2874 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const)) 2875 return true; 2876 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) && 2877 !Left.isOneOf(tok::arrowstar, tok::lessless) && 2878 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All && 2879 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None || 2880 Left.getPrecedence() == prec::Assignment)) 2881 return true; 2882 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace, 2883 tok::kw_class, tok::kw_struct, tok::comment) || 2884 Right.isMemberAccess() || 2885 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless, 2886 tok::colon, tok::l_square, tok::at) || 2887 (Left.is(tok::r_paren) && 2888 Right.isOneOf(tok::identifier, tok::kw_const)) || 2889 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) || 2890 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser)); 2891 } 2892 2893 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) { 2894 llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n"; 2895 const FormatToken *Tok = Line.First; 2896 while (Tok) { 2897 llvm::errs() << " M=" << Tok->MustBreakBefore 2898 << " C=" << Tok->CanBreakBefore 2899 << " T=" << getTokenTypeName(Tok->Type) 2900 << " S=" << Tok->SpacesRequiredBefore 2901 << " B=" << Tok->BlockParameterCount 2902 << " BK=" << Tok->BlockKind << " P=" << Tok->SplitPenalty 2903 << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength 2904 << " PPK=" << Tok->PackingKind << " FakeLParens="; 2905 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i) 2906 llvm::errs() << Tok->FakeLParens[i] << "/"; 2907 llvm::errs() << " FakeRParens=" << Tok->FakeRParens; 2908 llvm::errs() << " Text='" << Tok->TokenText << "'\n"; 2909 if (!Tok->Next) 2910 assert(Tok == Line.Last); 2911 Tok = Tok->Next; 2912 } 2913 llvm::errs() << "----\n"; 2914 } 2915 2916 } // namespace format 2917 } // namespace clang 2918