1 //===--- FormatToken.h - Format C++ code ------------------------*- C++ -*-===// 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 contains the declaration of the FormatToken, a wrapper 12 /// around Token with additional information related to formatting. 13 /// 14 //===----------------------------------------------------------------------===// 15 16 #ifndef LLVM_CLANG_LIB_FORMAT_FORMATTOKEN_H 17 #define LLVM_CLANG_LIB_FORMAT_FORMATTOKEN_H 18 19 #include "clang/Basic/IdentifierTable.h" 20 #include "clang/Basic/OperatorPrecedence.h" 21 #include "clang/Format/Format.h" 22 #include "clang/Lex/Lexer.h" 23 #include <memory> 24 #include <unordered_set> 25 26 namespace clang { 27 namespace format { 28 29 #define LIST_TOKEN_TYPES \ 30 TYPE(ArrayInitializerLSquare) \ 31 TYPE(ArraySubscriptLSquare) \ 32 TYPE(AttributeParen) \ 33 TYPE(BinaryOperator) \ 34 TYPE(BitFieldColon) \ 35 TYPE(BlockComment) \ 36 TYPE(CastRParen) \ 37 TYPE(ConditionalExpr) \ 38 TYPE(ConflictAlternative) \ 39 TYPE(ConflictEnd) \ 40 TYPE(ConflictStart) \ 41 TYPE(CtorInitializerColon) \ 42 TYPE(CtorInitializerComma) \ 43 TYPE(DesignatedInitializerLSquare) \ 44 TYPE(DesignatedInitializerPeriod) \ 45 TYPE(DictLiteral) \ 46 TYPE(ForEachMacro) \ 47 TYPE(FunctionAnnotationRParen) \ 48 TYPE(FunctionDeclarationName) \ 49 TYPE(FunctionLBrace) \ 50 TYPE(FunctionTypeLParen) \ 51 TYPE(ImplicitStringLiteral) \ 52 TYPE(InheritanceColon) \ 53 TYPE(InheritanceComma) \ 54 TYPE(InlineASMBrace) \ 55 TYPE(InlineASMColon) \ 56 TYPE(JavaAnnotation) \ 57 TYPE(JsComputedPropertyName) \ 58 TYPE(JsExponentiation) \ 59 TYPE(JsExponentiationEqual) \ 60 TYPE(JsFatArrow) \ 61 TYPE(JsNonNullAssertion) \ 62 TYPE(JsTypeColon) \ 63 TYPE(JsTypeOperator) \ 64 TYPE(JsTypeOptionalQuestion) \ 65 TYPE(LambdaArrow) \ 66 TYPE(LambdaLSquare) \ 67 TYPE(LeadingJavaAnnotation) \ 68 TYPE(LineComment) \ 69 TYPE(MacroBlockBegin) \ 70 TYPE(MacroBlockEnd) \ 71 TYPE(ObjCBlockLBrace) \ 72 TYPE(ObjCBlockLParen) \ 73 TYPE(ObjCDecl) \ 74 TYPE(ObjCForIn) \ 75 TYPE(ObjCMethodExpr) \ 76 TYPE(ObjCMethodSpecifier) \ 77 TYPE(ObjCProperty) \ 78 TYPE(ObjCStringLiteral) \ 79 TYPE(OverloadedOperator) \ 80 TYPE(OverloadedOperatorLParen) \ 81 TYPE(PointerOrReference) \ 82 TYPE(PureVirtualSpecifier) \ 83 TYPE(RangeBasedForLoopColon) \ 84 TYPE(RegexLiteral) \ 85 TYPE(SelectorName) \ 86 TYPE(StartOfName) \ 87 TYPE(StructuredBindingLSquare) \ 88 TYPE(TemplateCloser) \ 89 TYPE(TemplateOpener) \ 90 TYPE(TemplateString) \ 91 TYPE(ProtoExtensionLSquare) \ 92 TYPE(TrailingAnnotation) \ 93 TYPE(TrailingReturnArrow) \ 94 TYPE(TrailingUnaryOperator) \ 95 TYPE(UnaryOperator) \ 96 TYPE(Unknown) 97 98 enum TokenType { 99 #define TYPE(X) TT_##X, 100 LIST_TOKEN_TYPES 101 #undef TYPE 102 NUM_TOKEN_TYPES 103 }; 104 105 /// \brief Determines the name of a token type. 106 const char *getTokenTypeName(TokenType Type); 107 108 // Represents what type of block a set of braces open. 109 enum BraceBlockKind { BK_Unknown, BK_Block, BK_BracedInit }; 110 111 // The packing kind of a function's parameters. 112 enum ParameterPackingKind { PPK_BinPacked, PPK_OnePerLine, PPK_Inconclusive }; 113 114 enum FormatDecision { FD_Unformatted, FD_Continue, FD_Break }; 115 116 class TokenRole; 117 class AnnotatedLine; 118 119 /// \brief A wrapper around a \c Token storing information about the 120 /// whitespace characters preceding it. 121 struct FormatToken { 122 FormatToken() {} 123 124 /// \brief The \c Token. 125 Token Tok; 126 127 /// \brief The number of newlines immediately before the \c Token. 128 /// 129 /// This can be used to determine what the user wrote in the original code 130 /// and thereby e.g. leave an empty line between two function definitions. 131 unsigned NewlinesBefore = 0; 132 133 /// \brief Whether there is at least one unescaped newline before the \c 134 /// Token. 135 bool HasUnescapedNewline = false; 136 137 /// \brief The range of the whitespace immediately preceding the \c Token. 138 SourceRange WhitespaceRange; 139 140 /// \brief The offset just past the last '\n' in this token's leading 141 /// whitespace (relative to \c WhiteSpaceStart). 0 if there is no '\n'. 142 unsigned LastNewlineOffset = 0; 143 144 /// \brief The width of the non-whitespace parts of the token (or its first 145 /// line for multi-line tokens) in columns. 146 /// We need this to correctly measure number of columns a token spans. 147 unsigned ColumnWidth = 0; 148 149 /// \brief Contains the width in columns of the last line of a multi-line 150 /// token. 151 unsigned LastLineColumnWidth = 0; 152 153 /// \brief Whether the token text contains newlines (escaped or not). 154 bool IsMultiline = false; 155 156 /// \brief Indicates that this is the first token of the file. 157 bool IsFirst = false; 158 159 /// \brief Whether there must be a line break before this token. 160 /// 161 /// This happens for example when a preprocessor directive ended directly 162 /// before the token. 163 bool MustBreakBefore = false; 164 165 /// \brief The raw text of the token. 166 /// 167 /// Contains the raw token text without leading whitespace and without leading 168 /// escaped newlines. 169 StringRef TokenText; 170 171 /// \brief Set to \c true if this token is an unterminated literal. 172 bool IsUnterminatedLiteral = 0; 173 174 /// \brief Contains the kind of block if this token is a brace. 175 BraceBlockKind BlockKind = BK_Unknown; 176 177 TokenType Type = TT_Unknown; 178 179 /// \brief The number of spaces that should be inserted before this token. 180 unsigned SpacesRequiredBefore = 0; 181 182 /// \brief \c true if it is allowed to break before this token. 183 bool CanBreakBefore = false; 184 185 /// \brief \c true if this is the ">" of "template<..>". 186 bool ClosesTemplateDeclaration = false; 187 188 /// \brief Number of parameters, if this is "(", "[" or "<". 189 /// 190 /// This is initialized to 1 as we don't need to distinguish functions with 191 /// 0 parameters from functions with 1 parameter. Thus, we can simply count 192 /// the number of commas. 193 unsigned ParameterCount = 0; 194 195 /// \brief Number of parameters that are nested blocks, 196 /// if this is "(", "[" or "<". 197 unsigned BlockParameterCount = 0; 198 199 /// \brief If this is a bracket ("<", "(", "[" or "{"), contains the kind of 200 /// the surrounding bracket. 201 tok::TokenKind ParentBracket = tok::unknown; 202 203 /// \brief A token can have a special role that can carry extra information 204 /// about the token's formatting. 205 std::unique_ptr<TokenRole> Role; 206 207 /// \brief If this is an opening parenthesis, how are the parameters packed? 208 ParameterPackingKind PackingKind = PPK_Inconclusive; 209 210 /// \brief The total length of the unwrapped line up to and including this 211 /// token. 212 unsigned TotalLength = 0; 213 214 /// \brief The original 0-based column of this token, including expanded tabs. 215 /// The configured TabWidth is used as tab width. 216 unsigned OriginalColumn = 0; 217 218 /// \brief The length of following tokens until the next natural split point, 219 /// or the next token that can be broken. 220 unsigned UnbreakableTailLength = 0; 221 222 // FIXME: Come up with a 'cleaner' concept. 223 /// \brief The binding strength of a token. This is a combined value of 224 /// operator precedence, parenthesis nesting, etc. 225 unsigned BindingStrength = 0; 226 227 /// \brief The nesting level of this token, i.e. the number of surrounding (), 228 /// [], {} or <>. 229 unsigned NestingLevel = 0; 230 231 /// \brief The indent level of this token. Copied from the surrounding line. 232 unsigned IndentLevel = 0; 233 234 /// \brief Penalty for inserting a line break before this token. 235 unsigned SplitPenalty = 0; 236 237 /// \brief If this is the first ObjC selector name in an ObjC method 238 /// definition or call, this contains the length of the longest name. 239 /// 240 /// This being set to 0 means that the selectors should not be colon-aligned, 241 /// e.g. because several of them are block-type. 242 unsigned LongestObjCSelectorName = 0; 243 244 /// \brief How many parts ObjC selector have (i.e. how many parameters method 245 /// has). 246 unsigned ObjCSelectorNameParts = 0; 247 248 /// \brief Stores the number of required fake parentheses and the 249 /// corresponding operator precedence. 250 /// 251 /// If multiple fake parentheses start at a token, this vector stores them in 252 /// reverse order, i.e. inner fake parenthesis first. 253 SmallVector<prec::Level, 4> FakeLParens; 254 /// \brief Insert this many fake ) after this token for correct indentation. 255 unsigned FakeRParens = 0; 256 257 /// \brief \c true if this token starts a binary expression, i.e. has at least 258 /// one fake l_paren with a precedence greater than prec::Unknown. 259 bool StartsBinaryExpression = false; 260 /// \brief \c true if this token ends a binary expression. 261 bool EndsBinaryExpression = false; 262 263 /// \brief Is this is an operator (or "."/"->") in a sequence of operators 264 /// with the same precedence, contains the 0-based operator index. 265 unsigned OperatorIndex = 0; 266 267 /// \brief If this is an operator (or "."/"->") in a sequence of operators 268 /// with the same precedence, points to the next operator. 269 FormatToken *NextOperator = nullptr; 270 271 /// \brief Is this token part of a \c DeclStmt defining multiple variables? 272 /// 273 /// Only set if \c Type == \c TT_StartOfName. 274 bool PartOfMultiVariableDeclStmt = false; 275 276 /// \brief Does this line comment continue a line comment section? 277 /// 278 /// Only set to true if \c Type == \c TT_LineComment. 279 bool ContinuesLineCommentSection = false; 280 281 /// \brief If this is a bracket, this points to the matching one. 282 FormatToken *MatchingParen = nullptr; 283 284 /// \brief The previous token in the unwrapped line. 285 FormatToken *Previous = nullptr; 286 287 /// \brief The next token in the unwrapped line. 288 FormatToken *Next = nullptr; 289 290 /// \brief If this token starts a block, this contains all the unwrapped lines 291 /// in it. 292 SmallVector<AnnotatedLine *, 1> Children; 293 294 /// \brief Stores the formatting decision for the token once it was made. 295 FormatDecision Decision = FD_Unformatted; 296 297 /// \brief If \c true, this token has been fully formatted (indented and 298 /// potentially re-formatted inside), and we do not allow further formatting 299 /// changes. 300 bool Finalized = false; 301 302 bool is(tok::TokenKind Kind) const { return Tok.is(Kind); } 303 bool is(TokenType TT) const { return Type == TT; } 304 bool is(const IdentifierInfo *II) const { 305 return II && II == Tok.getIdentifierInfo(); 306 } 307 bool is(tok::PPKeywordKind Kind) const { 308 return Tok.getIdentifierInfo() && 309 Tok.getIdentifierInfo()->getPPKeywordID() == Kind; 310 } 311 template <typename A, typename B> bool isOneOf(A K1, B K2) const { 312 return is(K1) || is(K2); 313 } 314 template <typename A, typename B, typename... Ts> 315 bool isOneOf(A K1, B K2, Ts... Ks) const { 316 return is(K1) || isOneOf(K2, Ks...); 317 } 318 template <typename T> bool isNot(T Kind) const { return !is(Kind); } 319 320 /// \c true if this token starts a sequence with the given tokens in order, 321 /// following the ``Next`` pointers, ignoring comments. 322 template <typename A, typename... Ts> 323 bool startsSequence(A K1, Ts... Tokens) const { 324 return startsSequenceInternal(K1, Tokens...); 325 } 326 327 /// \c true if this token ends a sequence with the given tokens in order, 328 /// following the ``Previous`` pointers, ignoring comments. 329 template <typename A, typename... Ts> 330 bool endsSequence(A K1, Ts... Tokens) const { 331 return endsSequenceInternal(K1, Tokens...); 332 } 333 334 bool isStringLiteral() const { return tok::isStringLiteral(Tok.getKind()); } 335 336 bool isObjCAtKeyword(tok::ObjCKeywordKind Kind) const { 337 return Tok.isObjCAtKeyword(Kind); 338 } 339 340 bool isAccessSpecifier(bool ColonRequired = true) const { 341 return isOneOf(tok::kw_public, tok::kw_protected, tok::kw_private) && 342 (!ColonRequired || (Next && Next->is(tok::colon))); 343 } 344 345 /// \brief Determine whether the token is a simple-type-specifier. 346 bool isSimpleTypeSpecifier() const; 347 348 bool isObjCAccessSpecifier() const { 349 return is(tok::at) && Next && 350 (Next->isObjCAtKeyword(tok::objc_public) || 351 Next->isObjCAtKeyword(tok::objc_protected) || 352 Next->isObjCAtKeyword(tok::objc_package) || 353 Next->isObjCAtKeyword(tok::objc_private)); 354 } 355 356 /// \brief Returns whether \p Tok is ([{ or an opening < of a template or in 357 /// protos. 358 bool opensScope() const { 359 if (is(TT_TemplateString) && TokenText.endswith("${")) 360 return true; 361 if (is(TT_DictLiteral) && is(tok::less)) 362 return true; 363 return isOneOf(tok::l_paren, tok::l_brace, tok::l_square, 364 TT_TemplateOpener); 365 } 366 /// \brief Returns whether \p Tok is )]} or a closing > of a template or in 367 /// protos. 368 bool closesScope() const { 369 if (is(TT_TemplateString) && TokenText.startswith("}")) 370 return true; 371 if (is(TT_DictLiteral) && is(tok::greater)) 372 return true; 373 return isOneOf(tok::r_paren, tok::r_brace, tok::r_square, 374 TT_TemplateCloser); 375 } 376 377 /// \brief Returns \c true if this is a "." or "->" accessing a member. 378 bool isMemberAccess() const { 379 return isOneOf(tok::arrow, tok::period, tok::arrowstar) && 380 !isOneOf(TT_DesignatedInitializerPeriod, TT_TrailingReturnArrow, 381 TT_LambdaArrow); 382 } 383 384 bool isUnaryOperator() const { 385 switch (Tok.getKind()) { 386 case tok::plus: 387 case tok::plusplus: 388 case tok::minus: 389 case tok::minusminus: 390 case tok::exclaim: 391 case tok::tilde: 392 case tok::kw_sizeof: 393 case tok::kw_alignof: 394 return true; 395 default: 396 return false; 397 } 398 } 399 400 bool isBinaryOperator() const { 401 // Comma is a binary operator, but does not behave as such wrt. formatting. 402 return getPrecedence() > prec::Comma; 403 } 404 405 bool isTrailingComment() const { 406 return is(tok::comment) && 407 (is(TT_LineComment) || !Next || Next->NewlinesBefore > 0); 408 } 409 410 /// \brief Returns \c true if this is a keyword that can be used 411 /// like a function call (e.g. sizeof, typeid, ...). 412 bool isFunctionLikeKeyword() const { 413 switch (Tok.getKind()) { 414 case tok::kw_throw: 415 case tok::kw_typeid: 416 case tok::kw_return: 417 case tok::kw_sizeof: 418 case tok::kw_alignof: 419 case tok::kw_alignas: 420 case tok::kw_decltype: 421 case tok::kw_noexcept: 422 case tok::kw_static_assert: 423 case tok::kw___attribute: 424 return true; 425 default: 426 return false; 427 } 428 } 429 430 /// \brief Returns \c true if this is a string literal that's like a label, 431 /// e.g. ends with "=" or ":". 432 bool isLabelString() const { 433 if (!is(tok::string_literal)) 434 return false; 435 StringRef Content = TokenText; 436 if (Content.startswith("\"") || Content.startswith("'")) 437 Content = Content.drop_front(1); 438 if (Content.endswith("\"") || Content.endswith("'")) 439 Content = Content.drop_back(1); 440 Content = Content.trim(); 441 return Content.size() > 1 && 442 (Content.back() == ':' || Content.back() == '='); 443 } 444 445 /// \brief Returns actual token start location without leading escaped 446 /// newlines and whitespace. 447 /// 448 /// This can be different to Tok.getLocation(), which includes leading escaped 449 /// newlines. 450 SourceLocation getStartOfNonWhitespace() const { 451 return WhitespaceRange.getEnd(); 452 } 453 454 prec::Level getPrecedence() const { 455 return getBinOpPrecedence(Tok.getKind(), /*GreaterThanIsOperator=*/true, 456 /*CPlusPlus11=*/true); 457 } 458 459 /// \brief Returns the previous token ignoring comments. 460 FormatToken *getPreviousNonComment() const { 461 FormatToken *Tok = Previous; 462 while (Tok && Tok->is(tok::comment)) 463 Tok = Tok->Previous; 464 return Tok; 465 } 466 467 /// \brief Returns the next token ignoring comments. 468 const FormatToken *getNextNonComment() const { 469 const FormatToken *Tok = Next; 470 while (Tok && Tok->is(tok::comment)) 471 Tok = Tok->Next; 472 return Tok; 473 } 474 475 /// \brief Returns \c true if this tokens starts a block-type list, i.e. a 476 /// list that should be indented with a block indent. 477 bool opensBlockOrBlockTypeList(const FormatStyle &Style) const { 478 if (is(TT_TemplateString) && opensScope()) 479 return true; 480 return is(TT_ArrayInitializerLSquare) || 481 is(TT_ProtoExtensionLSquare) || 482 (is(tok::l_brace) && 483 (BlockKind == BK_Block || is(TT_DictLiteral) || 484 (!Style.Cpp11BracedListStyle && NestingLevel == 0))) || 485 (is(tok::less) && (Style.Language == FormatStyle::LK_Proto || 486 Style.Language == FormatStyle::LK_TextProto)); 487 } 488 489 /// \brief Returns whether the token is the left square bracket of a C++ 490 /// structured binding declaration. 491 bool isCppStructuredBinding(const FormatStyle &Style) const { 492 if (!Style.isCpp() || isNot(tok::l_square)) 493 return false; 494 const FormatToken *T = this; 495 do { 496 T = T->getPreviousNonComment(); 497 } while (T && T->isOneOf(tok::kw_const, tok::kw_volatile, tok::amp, 498 tok::ampamp)); 499 return T && T->is(tok::kw_auto); 500 } 501 502 /// \brief Same as opensBlockOrBlockTypeList, but for the closing token. 503 bool closesBlockOrBlockTypeList(const FormatStyle &Style) const { 504 if (is(TT_TemplateString) && closesScope()) 505 return true; 506 return MatchingParen && MatchingParen->opensBlockOrBlockTypeList(Style); 507 } 508 509 /// \brief Return the actual namespace token, if this token starts a namespace 510 /// block. 511 const FormatToken *getNamespaceToken() const { 512 const FormatToken *NamespaceTok = this; 513 if (is(tok::comment)) 514 NamespaceTok = NamespaceTok->getNextNonComment(); 515 // Detect "(inline)? namespace" in the beginning of a line. 516 if (NamespaceTok && NamespaceTok->is(tok::kw_inline)) 517 NamespaceTok = NamespaceTok->getNextNonComment(); 518 return NamespaceTok && NamespaceTok->is(tok::kw_namespace) ? NamespaceTok 519 : nullptr; 520 } 521 522 private: 523 // Disallow copying. 524 FormatToken(const FormatToken &) = delete; 525 void operator=(const FormatToken &) = delete; 526 527 template <typename A, typename... Ts> 528 bool startsSequenceInternal(A K1, Ts... Tokens) const { 529 if (is(tok::comment) && Next) 530 return Next->startsSequenceInternal(K1, Tokens...); 531 return is(K1) && Next && Next->startsSequenceInternal(Tokens...); 532 } 533 534 template <typename A> bool startsSequenceInternal(A K1) const { 535 if (is(tok::comment) && Next) 536 return Next->startsSequenceInternal(K1); 537 return is(K1); 538 } 539 540 template <typename A, typename... Ts> bool endsSequenceInternal(A K1) const { 541 if (is(tok::comment) && Previous) 542 return Previous->endsSequenceInternal(K1); 543 return is(K1); 544 } 545 546 template <typename A, typename... Ts> 547 bool endsSequenceInternal(A K1, Ts... Tokens) const { 548 if (is(tok::comment) && Previous) 549 return Previous->endsSequenceInternal(K1, Tokens...); 550 return is(K1) && Previous && Previous->endsSequenceInternal(Tokens...); 551 } 552 }; 553 554 class ContinuationIndenter; 555 struct LineState; 556 557 class TokenRole { 558 public: 559 TokenRole(const FormatStyle &Style) : Style(Style) {} 560 virtual ~TokenRole(); 561 562 /// \brief After the \c TokenAnnotator has finished annotating all the tokens, 563 /// this function precomputes required information for formatting. 564 virtual void precomputeFormattingInfos(const FormatToken *Token); 565 566 /// \brief Apply the special formatting that the given role demands. 567 /// 568 /// Assumes that the token having this role is already formatted. 569 /// 570 /// Continues formatting from \p State leaving indentation to \p Indenter and 571 /// returns the total penalty that this formatting incurs. 572 virtual unsigned formatFromToken(LineState &State, 573 ContinuationIndenter *Indenter, 574 bool DryRun) { 575 return 0; 576 } 577 578 /// \brief Same as \c formatFromToken, but assumes that the first token has 579 /// already been set thereby deciding on the first line break. 580 virtual unsigned formatAfterToken(LineState &State, 581 ContinuationIndenter *Indenter, 582 bool DryRun) { 583 return 0; 584 } 585 586 /// \brief Notifies the \c Role that a comma was found. 587 virtual void CommaFound(const FormatToken *Token) {} 588 589 protected: 590 const FormatStyle &Style; 591 }; 592 593 class CommaSeparatedList : public TokenRole { 594 public: 595 CommaSeparatedList(const FormatStyle &Style) 596 : TokenRole(Style), HasNestedBracedList(false) {} 597 598 void precomputeFormattingInfos(const FormatToken *Token) override; 599 600 unsigned formatAfterToken(LineState &State, ContinuationIndenter *Indenter, 601 bool DryRun) override; 602 603 unsigned formatFromToken(LineState &State, ContinuationIndenter *Indenter, 604 bool DryRun) override; 605 606 /// \brief Adds \p Token as the next comma to the \c CommaSeparated list. 607 void CommaFound(const FormatToken *Token) override { 608 Commas.push_back(Token); 609 } 610 611 private: 612 /// \brief A struct that holds information on how to format a given list with 613 /// a specific number of columns. 614 struct ColumnFormat { 615 /// \brief The number of columns to use. 616 unsigned Columns; 617 618 /// \brief The total width in characters. 619 unsigned TotalWidth; 620 621 /// \brief The number of lines required for this format. 622 unsigned LineCount; 623 624 /// \brief The size of each column in characters. 625 SmallVector<unsigned, 8> ColumnSizes; 626 }; 627 628 /// \brief Calculate which \c ColumnFormat fits best into 629 /// \p RemainingCharacters. 630 const ColumnFormat *getColumnFormat(unsigned RemainingCharacters) const; 631 632 /// \brief The ordered \c FormatTokens making up the commas of this list. 633 SmallVector<const FormatToken *, 8> Commas; 634 635 /// \brief The length of each of the list's items in characters including the 636 /// trailing comma. 637 SmallVector<unsigned, 8> ItemLengths; 638 639 /// \brief Precomputed formats that can be used for this list. 640 SmallVector<ColumnFormat, 4> Formats; 641 642 bool HasNestedBracedList; 643 }; 644 645 /// \brief Encapsulates keywords that are context sensitive or for languages not 646 /// properly supported by Clang's lexer. 647 struct AdditionalKeywords { 648 AdditionalKeywords(IdentifierTable &IdentTable) { 649 kw_final = &IdentTable.get("final"); 650 kw_override = &IdentTable.get("override"); 651 kw_in = &IdentTable.get("in"); 652 kw_of = &IdentTable.get("of"); 653 kw_CF_ENUM = &IdentTable.get("CF_ENUM"); 654 kw_CF_OPTIONS = &IdentTable.get("CF_OPTIONS"); 655 kw_NS_ENUM = &IdentTable.get("NS_ENUM"); 656 kw_NS_OPTIONS = &IdentTable.get("NS_OPTIONS"); 657 658 kw_as = &IdentTable.get("as"); 659 kw_async = &IdentTable.get("async"); 660 kw_await = &IdentTable.get("await"); 661 kw_declare = &IdentTable.get("declare"); 662 kw_finally = &IdentTable.get("finally"); 663 kw_from = &IdentTable.get("from"); 664 kw_function = &IdentTable.get("function"); 665 kw_get = &IdentTable.get("get"); 666 kw_import = &IdentTable.get("import"); 667 kw_is = &IdentTable.get("is"); 668 kw_let = &IdentTable.get("let"); 669 kw_module = &IdentTable.get("module"); 670 kw_readonly = &IdentTable.get("readonly"); 671 kw_set = &IdentTable.get("set"); 672 kw_type = &IdentTable.get("type"); 673 kw_typeof = &IdentTable.get("typeof"); 674 kw_var = &IdentTable.get("var"); 675 kw_yield = &IdentTable.get("yield"); 676 677 kw_abstract = &IdentTable.get("abstract"); 678 kw_assert = &IdentTable.get("assert"); 679 kw_extends = &IdentTable.get("extends"); 680 kw_implements = &IdentTable.get("implements"); 681 kw_instanceof = &IdentTable.get("instanceof"); 682 kw_interface = &IdentTable.get("interface"); 683 kw_native = &IdentTable.get("native"); 684 kw_package = &IdentTable.get("package"); 685 kw_synchronized = &IdentTable.get("synchronized"); 686 kw_throws = &IdentTable.get("throws"); 687 kw___except = &IdentTable.get("__except"); 688 kw___has_include = &IdentTable.get("__has_include"); 689 kw___has_include_next = &IdentTable.get("__has_include_next"); 690 691 kw_mark = &IdentTable.get("mark"); 692 693 kw_extend = &IdentTable.get("extend"); 694 kw_option = &IdentTable.get("option"); 695 kw_optional = &IdentTable.get("optional"); 696 kw_repeated = &IdentTable.get("repeated"); 697 kw_required = &IdentTable.get("required"); 698 kw_returns = &IdentTable.get("returns"); 699 700 kw_signals = &IdentTable.get("signals"); 701 kw_qsignals = &IdentTable.get("Q_SIGNALS"); 702 kw_slots = &IdentTable.get("slots"); 703 kw_qslots = &IdentTable.get("Q_SLOTS"); 704 705 // Keep this at the end of the constructor to make sure everything here is 706 // already initialized. 707 JsExtraKeywords = std::unordered_set<IdentifierInfo *>( 708 {kw_as, kw_async, kw_await, kw_declare, kw_finally, kw_from, 709 kw_function, kw_get, kw_import, kw_is, kw_let, kw_module, kw_readonly, 710 kw_set, kw_type, kw_typeof, kw_var, kw_yield, 711 // Keywords from the Java section. 712 kw_abstract, kw_extends, kw_implements, kw_instanceof, kw_interface}); 713 } 714 715 // Context sensitive keywords. 716 IdentifierInfo *kw_final; 717 IdentifierInfo *kw_override; 718 IdentifierInfo *kw_in; 719 IdentifierInfo *kw_of; 720 IdentifierInfo *kw_CF_ENUM; 721 IdentifierInfo *kw_CF_OPTIONS; 722 IdentifierInfo *kw_NS_ENUM; 723 IdentifierInfo *kw_NS_OPTIONS; 724 IdentifierInfo *kw___except; 725 IdentifierInfo *kw___has_include; 726 IdentifierInfo *kw___has_include_next; 727 728 // JavaScript keywords. 729 IdentifierInfo *kw_as; 730 IdentifierInfo *kw_async; 731 IdentifierInfo *kw_await; 732 IdentifierInfo *kw_declare; 733 IdentifierInfo *kw_finally; 734 IdentifierInfo *kw_from; 735 IdentifierInfo *kw_function; 736 IdentifierInfo *kw_get; 737 IdentifierInfo *kw_import; 738 IdentifierInfo *kw_is; 739 IdentifierInfo *kw_let; 740 IdentifierInfo *kw_module; 741 IdentifierInfo *kw_readonly; 742 IdentifierInfo *kw_set; 743 IdentifierInfo *kw_type; 744 IdentifierInfo *kw_typeof; 745 IdentifierInfo *kw_var; 746 IdentifierInfo *kw_yield; 747 748 // Java keywords. 749 IdentifierInfo *kw_abstract; 750 IdentifierInfo *kw_assert; 751 IdentifierInfo *kw_extends; 752 IdentifierInfo *kw_implements; 753 IdentifierInfo *kw_instanceof; 754 IdentifierInfo *kw_interface; 755 IdentifierInfo *kw_native; 756 IdentifierInfo *kw_package; 757 IdentifierInfo *kw_synchronized; 758 IdentifierInfo *kw_throws; 759 760 // Pragma keywords. 761 IdentifierInfo *kw_mark; 762 763 // Proto keywords. 764 IdentifierInfo *kw_extend; 765 IdentifierInfo *kw_option; 766 IdentifierInfo *kw_optional; 767 IdentifierInfo *kw_repeated; 768 IdentifierInfo *kw_required; 769 IdentifierInfo *kw_returns; 770 771 // QT keywords. 772 IdentifierInfo *kw_signals; 773 IdentifierInfo *kw_qsignals; 774 IdentifierInfo *kw_slots; 775 IdentifierInfo *kw_qslots; 776 777 /// \brief Returns \c true if \p Tok is a true JavaScript identifier, returns 778 /// \c false if it is a keyword or a pseudo keyword. 779 bool IsJavaScriptIdentifier(const FormatToken &Tok) const { 780 return Tok.is(tok::identifier) && 781 JsExtraKeywords.find(Tok.Tok.getIdentifierInfo()) == 782 JsExtraKeywords.end(); 783 } 784 785 private: 786 /// \brief The JavaScript keywords beyond the C++ keyword set. 787 std::unordered_set<IdentifierInfo *> JsExtraKeywords; 788 }; 789 790 } // namespace format 791 } // namespace clang 792 793 #endif 794