1 //===--- FormatToken.h - Format C++ code ------------------------*- C++ -*-===// 2 // 3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 4 // See https://llvm.org/LICENSE.txt for license information. 5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 6 // 7 //===----------------------------------------------------------------------===// 8 /// 9 /// \file 10 /// This file contains the declaration of the FormatToken, a wrapper 11 /// around Token with additional information related to formatting. 12 /// 13 //===----------------------------------------------------------------------===// 14 15 #ifndef LLVM_CLANG_LIB_FORMAT_FORMATTOKEN_H 16 #define LLVM_CLANG_LIB_FORMAT_FORMATTOKEN_H 17 18 #include "clang/Basic/IdentifierTable.h" 19 #include "clang/Basic/OperatorPrecedence.h" 20 #include "clang/Format/Format.h" 21 #include "clang/Lex/Lexer.h" 22 #include <memory> 23 #include <unordered_set> 24 25 namespace clang { 26 namespace format { 27 28 #define LIST_TOKEN_TYPES \ 29 TYPE(ArrayInitializerLSquare) \ 30 TYPE(ArraySubscriptLSquare) \ 31 TYPE(AttributeColon) \ 32 TYPE(AttributeParen) \ 33 TYPE(AttributeSquare) \ 34 TYPE(BinaryOperator) \ 35 TYPE(BitFieldColon) \ 36 TYPE(BlockComment) \ 37 TYPE(CastRParen) \ 38 TYPE(ConditionalExpr) \ 39 TYPE(ConflictAlternative) \ 40 TYPE(ConflictEnd) \ 41 TYPE(ConflictStart) \ 42 TYPE(CtorInitializerColon) \ 43 TYPE(CtorInitializerComma) \ 44 TYPE(DesignatedInitializerLSquare) \ 45 TYPE(DesignatedInitializerPeriod) \ 46 TYPE(DictLiteral) \ 47 TYPE(ForEachMacro) \ 48 TYPE(FunctionAnnotationRParen) \ 49 TYPE(FunctionDeclarationName) \ 50 TYPE(FunctionLBrace) \ 51 TYPE(FunctionTypeLParen) \ 52 TYPE(ImplicitStringLiteral) \ 53 TYPE(InheritanceColon) \ 54 TYPE(InheritanceComma) \ 55 TYPE(InlineASMBrace) \ 56 TYPE(InlineASMColon) \ 57 TYPE(InlineASMSymbolicNameLSquare) \ 58 TYPE(JavaAnnotation) \ 59 TYPE(JsComputedPropertyName) \ 60 TYPE(JsExponentiation) \ 61 TYPE(JsExponentiationEqual) \ 62 TYPE(JsFatArrow) \ 63 TYPE(JsNonNullAssertion) \ 64 TYPE(JsNullishCoalescingOperator) \ 65 TYPE(JsNullPropagatingOperator) \ 66 TYPE(JsPrivateIdentifier) \ 67 TYPE(JsTypeColon) \ 68 TYPE(JsTypeOperator) \ 69 TYPE(JsTypeOptionalQuestion) \ 70 TYPE(LambdaArrow) \ 71 TYPE(LambdaLBrace) \ 72 TYPE(LambdaLSquare) \ 73 TYPE(LeadingJavaAnnotation) \ 74 TYPE(LineComment) \ 75 TYPE(MacroBlockBegin) \ 76 TYPE(MacroBlockEnd) \ 77 TYPE(NamespaceMacro) \ 78 TYPE(ObjCBlockLBrace) \ 79 TYPE(ObjCBlockLParen) \ 80 TYPE(ObjCDecl) \ 81 TYPE(ObjCForIn) \ 82 TYPE(ObjCMethodExpr) \ 83 TYPE(ObjCMethodSpecifier) \ 84 TYPE(ObjCProperty) \ 85 TYPE(ObjCStringLiteral) \ 86 TYPE(OverloadedOperator) \ 87 TYPE(OverloadedOperatorLParen) \ 88 TYPE(PointerOrReference) \ 89 TYPE(PureVirtualSpecifier) \ 90 TYPE(RangeBasedForLoopColon) \ 91 TYPE(RegexLiteral) \ 92 TYPE(SelectorName) \ 93 TYPE(StartOfName) \ 94 TYPE(StatementMacro) \ 95 TYPE(StructuredBindingLSquare) \ 96 TYPE(TemplateCloser) \ 97 TYPE(TemplateOpener) \ 98 TYPE(TemplateString) \ 99 TYPE(ProtoExtensionLSquare) \ 100 TYPE(TrailingAnnotation) \ 101 TYPE(TrailingReturnArrow) \ 102 TYPE(TrailingUnaryOperator) \ 103 TYPE(TypenameMacro) \ 104 TYPE(UnaryOperator) \ 105 TYPE(CSharpStringLiteral) \ 106 TYPE(CSharpNamedArgumentColon) \ 107 TYPE(CSharpNullable) \ 108 TYPE(CSharpNullCoalescing) \ 109 TYPE(CSharpNullConditional) \ 110 TYPE(CSharpNullConditionalLSquare) \ 111 TYPE(Unknown) 112 113 enum TokenType { 114 #define TYPE(X) TT_##X, 115 LIST_TOKEN_TYPES 116 #undef TYPE 117 NUM_TOKEN_TYPES 118 }; 119 120 /// Determines the name of a token type. 121 const char *getTokenTypeName(TokenType Type); 122 123 // Represents what type of block a set of braces open. 124 enum BraceBlockKind { BK_Unknown, BK_Block, BK_BracedInit }; 125 126 // The packing kind of a function's parameters. 127 enum ParameterPackingKind { PPK_BinPacked, PPK_OnePerLine, PPK_Inconclusive }; 128 129 enum FormatDecision { FD_Unformatted, FD_Continue, FD_Break }; 130 131 class TokenRole; 132 class AnnotatedLine; 133 134 /// A wrapper around a \c Token storing information about the 135 /// whitespace characters preceding it. 136 struct FormatToken { 137 FormatToken() {} 138 139 /// The \c Token. 140 Token Tok; 141 142 /// The number of newlines immediately before the \c Token. 143 /// 144 /// This can be used to determine what the user wrote in the original code 145 /// and thereby e.g. leave an empty line between two function definitions. 146 unsigned NewlinesBefore = 0; 147 148 /// Whether there is at least one unescaped newline before the \c 149 /// Token. 150 bool HasUnescapedNewline = false; 151 152 /// The range of the whitespace immediately preceding the \c Token. 153 SourceRange WhitespaceRange; 154 155 /// The offset just past the last '\n' in this token's leading 156 /// whitespace (relative to \c WhiteSpaceStart). 0 if there is no '\n'. 157 unsigned LastNewlineOffset = 0; 158 159 /// The width of the non-whitespace parts of the token (or its first 160 /// line for multi-line tokens) in columns. 161 /// We need this to correctly measure number of columns a token spans. 162 unsigned ColumnWidth = 0; 163 164 /// Contains the width in columns of the last line of a multi-line 165 /// token. 166 unsigned LastLineColumnWidth = 0; 167 168 /// Whether the token text contains newlines (escaped or not). 169 bool IsMultiline = false; 170 171 /// Indicates that this is the first token of the file. 172 bool IsFirst = false; 173 174 /// Whether there must be a line break before this token. 175 /// 176 /// This happens for example when a preprocessor directive ended directly 177 /// before the token. 178 bool MustBreakBefore = false; 179 180 /// The raw text of the token. 181 /// 182 /// Contains the raw token text without leading whitespace and without leading 183 /// escaped newlines. 184 StringRef TokenText; 185 186 /// Set to \c true if this token is an unterminated literal. 187 bool IsUnterminatedLiteral = 0; 188 189 /// Contains the kind of block if this token is a brace. 190 BraceBlockKind BlockKind = BK_Unknown; 191 192 TokenType Type = TT_Unknown; 193 194 /// The number of spaces that should be inserted before this token. 195 unsigned SpacesRequiredBefore = 0; 196 197 /// \c true if it is allowed to break before this token. 198 bool CanBreakBefore = false; 199 200 /// \c true if this is the ">" of "template<..>". 201 bool ClosesTemplateDeclaration = false; 202 203 /// Number of parameters, if this is "(", "[" or "<". 204 unsigned ParameterCount = 0; 205 206 /// Number of parameters that are nested blocks, 207 /// if this is "(", "[" or "<". 208 unsigned BlockParameterCount = 0; 209 210 /// If this is a bracket ("<", "(", "[" or "{"), contains the kind of 211 /// the surrounding bracket. 212 tok::TokenKind ParentBracket = tok::unknown; 213 214 /// A token can have a special role that can carry extra information 215 /// about the token's formatting. 216 std::unique_ptr<TokenRole> Role; 217 218 /// If this is an opening parenthesis, how are the parameters packed? 219 ParameterPackingKind PackingKind = PPK_Inconclusive; 220 221 /// The total length of the unwrapped line up to and including this 222 /// token. 223 unsigned TotalLength = 0; 224 225 /// The original 0-based column of this token, including expanded tabs. 226 /// The configured TabWidth is used as tab width. 227 unsigned OriginalColumn = 0; 228 229 /// The length of following tokens until the next natural split point, 230 /// or the next token that can be broken. 231 unsigned UnbreakableTailLength = 0; 232 233 // FIXME: Come up with a 'cleaner' concept. 234 /// The binding strength of a token. This is a combined value of 235 /// operator precedence, parenthesis nesting, etc. 236 unsigned BindingStrength = 0; 237 238 /// The nesting level of this token, i.e. the number of surrounding (), 239 /// [], {} or <>. 240 unsigned NestingLevel = 0; 241 242 /// The indent level of this token. Copied from the surrounding line. 243 unsigned IndentLevel = 0; 244 245 /// Penalty for inserting a line break before this token. 246 unsigned SplitPenalty = 0; 247 248 /// If this is the first ObjC selector name in an ObjC method 249 /// definition or call, this contains the length of the longest name. 250 /// 251 /// This being set to 0 means that the selectors should not be colon-aligned, 252 /// e.g. because several of them are block-type. 253 unsigned LongestObjCSelectorName = 0; 254 255 /// If this is the first ObjC selector name in an ObjC method 256 /// definition or call, this contains the number of parts that the whole 257 /// selector consist of. 258 unsigned ObjCSelectorNameParts = 0; 259 260 /// The 0-based index of the parameter/argument. For ObjC it is set 261 /// for the selector name token. 262 /// For now calculated only for ObjC. 263 unsigned ParameterIndex = 0; 264 265 /// Stores the number of required fake parentheses and the 266 /// corresponding operator precedence. 267 /// 268 /// If multiple fake parentheses start at a token, this vector stores them in 269 /// reverse order, i.e. inner fake parenthesis first. 270 SmallVector<prec::Level, 4> FakeLParens; 271 /// Insert this many fake ) after this token for correct indentation. 272 unsigned FakeRParens = 0; 273 274 /// \c true if this token starts a binary expression, i.e. has at least 275 /// one fake l_paren with a precedence greater than prec::Unknown. 276 bool StartsBinaryExpression = false; 277 /// \c true if this token ends a binary expression. 278 bool EndsBinaryExpression = false; 279 280 /// If this is an operator (or "."/"->") in a sequence of operators 281 /// with the same precedence, contains the 0-based operator index. 282 unsigned OperatorIndex = 0; 283 284 /// If this is an operator (or "."/"->") in a sequence of operators 285 /// with the same precedence, points to the next operator. 286 FormatToken *NextOperator = nullptr; 287 288 /// Is this token part of a \c DeclStmt defining multiple variables? 289 /// 290 /// Only set if \c Type == \c TT_StartOfName. 291 bool PartOfMultiVariableDeclStmt = false; 292 293 /// Does this line comment continue a line comment section? 294 /// 295 /// Only set to true if \c Type == \c TT_LineComment. 296 bool ContinuesLineCommentSection = false; 297 298 /// If this is a bracket, this points to the matching one. 299 FormatToken *MatchingParen = nullptr; 300 301 /// The previous token in the unwrapped line. 302 FormatToken *Previous = nullptr; 303 304 /// The next token in the unwrapped line. 305 FormatToken *Next = nullptr; 306 307 /// If this token starts a block, this contains all the unwrapped lines 308 /// in it. 309 SmallVector<AnnotatedLine *, 1> Children; 310 311 /// Stores the formatting decision for the token once it was made. 312 FormatDecision Decision = FD_Unformatted; 313 314 /// If \c true, this token has been fully formatted (indented and 315 /// potentially re-formatted inside), and we do not allow further formatting 316 /// changes. 317 bool Finalized = false; 318 319 bool is(tok::TokenKind Kind) const { return Tok.is(Kind); } 320 bool is(TokenType TT) const { return Type == TT; } 321 bool is(const IdentifierInfo *II) const { 322 return II && II == Tok.getIdentifierInfo(); 323 } 324 bool is(tok::PPKeywordKind Kind) const { 325 return Tok.getIdentifierInfo() && 326 Tok.getIdentifierInfo()->getPPKeywordID() == Kind; 327 } 328 template <typename A, typename B> bool isOneOf(A K1, B K2) const { 329 return is(K1) || is(K2); 330 } 331 template <typename A, typename B, typename... Ts> 332 bool isOneOf(A K1, B K2, Ts... Ks) const { 333 return is(K1) || isOneOf(K2, Ks...); 334 } 335 template <typename T> bool isNot(T Kind) const { return !is(Kind); } 336 337 bool isIf(bool AllowConstexprMacro = true) const { 338 return is(tok::kw_if) || endsSequence(tok::kw_constexpr, tok::kw_if) || 339 (endsSequence(tok::identifier, tok::kw_if) && AllowConstexprMacro); 340 } 341 342 bool closesScopeAfterBlock() const { 343 if (BlockKind == BK_Block) 344 return true; 345 if (closesScope()) 346 return Previous->closesScopeAfterBlock(); 347 return false; 348 } 349 350 /// \c true if this token starts a sequence with the given tokens in order, 351 /// following the ``Next`` pointers, ignoring comments. 352 template <typename A, typename... Ts> 353 bool startsSequence(A K1, Ts... Tokens) const { 354 return startsSequenceInternal(K1, Tokens...); 355 } 356 357 /// \c true if this token ends a sequence with the given tokens in order, 358 /// following the ``Previous`` pointers, ignoring comments. 359 /// For example, given tokens [T1, T2, T3], the function returns true if 360 /// 3 tokens ending at this (ignoring comments) are [T3, T2, T1]. In other 361 /// words, the tokens passed to this function need to the reverse of the 362 /// order the tokens appear in code. 363 template <typename A, typename... Ts> 364 bool endsSequence(A K1, Ts... Tokens) const { 365 return endsSequenceInternal(K1, Tokens...); 366 } 367 368 bool isStringLiteral() const { return tok::isStringLiteral(Tok.getKind()); } 369 370 bool isObjCAtKeyword(tok::ObjCKeywordKind Kind) const { 371 return Tok.isObjCAtKeyword(Kind); 372 } 373 374 bool isAccessSpecifier(bool ColonRequired = true) const { 375 return isOneOf(tok::kw_public, tok::kw_protected, tok::kw_private) && 376 (!ColonRequired || (Next && Next->is(tok::colon))); 377 } 378 379 /// Determine whether the token is a simple-type-specifier. 380 bool isSimpleTypeSpecifier() const; 381 382 bool isObjCAccessSpecifier() const { 383 return is(tok::at) && Next && 384 (Next->isObjCAtKeyword(tok::objc_public) || 385 Next->isObjCAtKeyword(tok::objc_protected) || 386 Next->isObjCAtKeyword(tok::objc_package) || 387 Next->isObjCAtKeyword(tok::objc_private)); 388 } 389 390 /// Returns whether \p Tok is ([{ or an opening < of a template or in 391 /// protos. 392 bool opensScope() const { 393 if (is(TT_TemplateString) && TokenText.endswith("${")) 394 return true; 395 if (is(TT_DictLiteral) && is(tok::less)) 396 return true; 397 return isOneOf(tok::l_paren, tok::l_brace, tok::l_square, 398 TT_TemplateOpener); 399 } 400 /// Returns whether \p Tok is )]} or a closing > of a template or in 401 /// protos. 402 bool closesScope() const { 403 if (is(TT_TemplateString) && TokenText.startswith("}")) 404 return true; 405 if (is(TT_DictLiteral) && is(tok::greater)) 406 return true; 407 return isOneOf(tok::r_paren, tok::r_brace, tok::r_square, 408 TT_TemplateCloser); 409 } 410 411 /// Returns \c true if this is a "." or "->" accessing a member. 412 bool isMemberAccess() const { 413 return isOneOf(tok::arrow, tok::period, tok::arrowstar) && 414 !isOneOf(TT_DesignatedInitializerPeriod, TT_TrailingReturnArrow, 415 TT_LambdaArrow, TT_LeadingJavaAnnotation); 416 } 417 418 bool isUnaryOperator() const { 419 switch (Tok.getKind()) { 420 case tok::plus: 421 case tok::plusplus: 422 case tok::minus: 423 case tok::minusminus: 424 case tok::exclaim: 425 case tok::tilde: 426 case tok::kw_sizeof: 427 case tok::kw_alignof: 428 return true; 429 default: 430 return false; 431 } 432 } 433 434 bool isBinaryOperator() const { 435 // Comma is a binary operator, but does not behave as such wrt. formatting. 436 return getPrecedence() > prec::Comma; 437 } 438 439 bool isTrailingComment() const { 440 return is(tok::comment) && 441 (is(TT_LineComment) || !Next || Next->NewlinesBefore > 0); 442 } 443 444 /// Returns \c true if this is a keyword that can be used 445 /// like a function call (e.g. sizeof, typeid, ...). 446 bool isFunctionLikeKeyword() const { 447 switch (Tok.getKind()) { 448 case tok::kw_throw: 449 case tok::kw_typeid: 450 case tok::kw_return: 451 case tok::kw_sizeof: 452 case tok::kw_alignof: 453 case tok::kw_alignas: 454 case tok::kw_decltype: 455 case tok::kw_noexcept: 456 case tok::kw_static_assert: 457 case tok::kw___attribute: 458 return true; 459 default: 460 return false; 461 } 462 } 463 464 /// Returns \c true if this is a string literal that's like a label, 465 /// e.g. ends with "=" or ":". 466 bool isLabelString() const { 467 if (!is(tok::string_literal)) 468 return false; 469 StringRef Content = TokenText; 470 if (Content.startswith("\"") || Content.startswith("'")) 471 Content = Content.drop_front(1); 472 if (Content.endswith("\"") || Content.endswith("'")) 473 Content = Content.drop_back(1); 474 Content = Content.trim(); 475 return Content.size() > 1 && 476 (Content.back() == ':' || Content.back() == '='); 477 } 478 479 /// Returns actual token start location without leading escaped 480 /// newlines and whitespace. 481 /// 482 /// This can be different to Tok.getLocation(), which includes leading escaped 483 /// newlines. 484 SourceLocation getStartOfNonWhitespace() const { 485 return WhitespaceRange.getEnd(); 486 } 487 488 prec::Level getPrecedence() const { 489 return getBinOpPrecedence(Tok.getKind(), /*GreaterThanIsOperator=*/true, 490 /*CPlusPlus11=*/true); 491 } 492 493 /// Returns the previous token ignoring comments. 494 FormatToken *getPreviousNonComment() const { 495 FormatToken *Tok = Previous; 496 while (Tok && Tok->is(tok::comment)) 497 Tok = Tok->Previous; 498 return Tok; 499 } 500 501 /// Returns the next token ignoring comments. 502 const FormatToken *getNextNonComment() const { 503 const FormatToken *Tok = Next; 504 while (Tok && Tok->is(tok::comment)) 505 Tok = Tok->Next; 506 return Tok; 507 } 508 509 /// Returns \c true if this tokens starts a block-type list, i.e. a 510 /// list that should be indented with a block indent. 511 bool opensBlockOrBlockTypeList(const FormatStyle &Style) const { 512 // C# Does not indent object initialisers as continuations. 513 if (is(tok::l_brace) && BlockKind == BK_BracedInit && Style.isCSharp()) 514 return true; 515 if (is(TT_TemplateString) && opensScope()) 516 return true; 517 return is(TT_ArrayInitializerLSquare) || is(TT_ProtoExtensionLSquare) || 518 (is(tok::l_brace) && 519 (BlockKind == BK_Block || is(TT_DictLiteral) || 520 (!Style.Cpp11BracedListStyle && NestingLevel == 0))) || 521 (is(tok::less) && (Style.Language == FormatStyle::LK_Proto || 522 Style.Language == FormatStyle::LK_TextProto)); 523 } 524 525 /// Returns whether the token is the left square bracket of a C++ 526 /// structured binding declaration. 527 bool isCppStructuredBinding(const FormatStyle &Style) const { 528 if (!Style.isCpp() || isNot(tok::l_square)) 529 return false; 530 const FormatToken *T = this; 531 do { 532 T = T->getPreviousNonComment(); 533 } while (T && T->isOneOf(tok::kw_const, tok::kw_volatile, tok::amp, 534 tok::ampamp)); 535 return T && T->is(tok::kw_auto); 536 } 537 538 /// Same as opensBlockOrBlockTypeList, but for the closing token. 539 bool closesBlockOrBlockTypeList(const FormatStyle &Style) const { 540 if (is(TT_TemplateString) && closesScope()) 541 return true; 542 return MatchingParen && MatchingParen->opensBlockOrBlockTypeList(Style); 543 } 544 545 /// Return the actual namespace token, if this token starts a namespace 546 /// block. 547 const FormatToken *getNamespaceToken() const { 548 const FormatToken *NamespaceTok = this; 549 if (is(tok::comment)) 550 NamespaceTok = NamespaceTok->getNextNonComment(); 551 // Detect "(inline|export)? namespace" in the beginning of a line. 552 if (NamespaceTok && NamespaceTok->isOneOf(tok::kw_inline, tok::kw_export)) 553 NamespaceTok = NamespaceTok->getNextNonComment(); 554 return NamespaceTok && 555 NamespaceTok->isOneOf(tok::kw_namespace, TT_NamespaceMacro) 556 ? NamespaceTok 557 : nullptr; 558 } 559 560 private: 561 // Disallow copying. 562 FormatToken(const FormatToken &) = delete; 563 void operator=(const FormatToken &) = delete; 564 565 template <typename A, typename... Ts> 566 bool startsSequenceInternal(A K1, Ts... Tokens) const { 567 if (is(tok::comment) && Next) 568 return Next->startsSequenceInternal(K1, Tokens...); 569 return is(K1) && Next && Next->startsSequenceInternal(Tokens...); 570 } 571 572 template <typename A> bool startsSequenceInternal(A K1) const { 573 if (is(tok::comment) && Next) 574 return Next->startsSequenceInternal(K1); 575 return is(K1); 576 } 577 578 template <typename A, typename... Ts> bool endsSequenceInternal(A K1) const { 579 if (is(tok::comment) && Previous) 580 return Previous->endsSequenceInternal(K1); 581 return is(K1); 582 } 583 584 template <typename A, typename... Ts> 585 bool endsSequenceInternal(A K1, Ts... Tokens) const { 586 if (is(tok::comment) && Previous) 587 return Previous->endsSequenceInternal(K1, Tokens...); 588 return is(K1) && Previous && Previous->endsSequenceInternal(Tokens...); 589 } 590 }; 591 592 class ContinuationIndenter; 593 struct LineState; 594 595 class TokenRole { 596 public: 597 TokenRole(const FormatStyle &Style) : Style(Style) {} 598 virtual ~TokenRole(); 599 600 /// After the \c TokenAnnotator has finished annotating all the tokens, 601 /// this function precomputes required information for formatting. 602 virtual void precomputeFormattingInfos(const FormatToken *Token); 603 604 /// Apply the special formatting that the given role demands. 605 /// 606 /// Assumes that the token having this role is already formatted. 607 /// 608 /// Continues formatting from \p State leaving indentation to \p Indenter and 609 /// returns the total penalty that this formatting incurs. 610 virtual unsigned formatFromToken(LineState &State, 611 ContinuationIndenter *Indenter, 612 bool DryRun) { 613 return 0; 614 } 615 616 /// Same as \c formatFromToken, but assumes that the first token has 617 /// already been set thereby deciding on the first line break. 618 virtual unsigned formatAfterToken(LineState &State, 619 ContinuationIndenter *Indenter, 620 bool DryRun) { 621 return 0; 622 } 623 624 /// Notifies the \c Role that a comma was found. 625 virtual void CommaFound(const FormatToken *Token) {} 626 627 virtual const FormatToken *lastComma() { return nullptr; } 628 629 protected: 630 const FormatStyle &Style; 631 }; 632 633 class CommaSeparatedList : public TokenRole { 634 public: 635 CommaSeparatedList(const FormatStyle &Style) 636 : TokenRole(Style), HasNestedBracedList(false) {} 637 638 void precomputeFormattingInfos(const FormatToken *Token) override; 639 640 unsigned formatAfterToken(LineState &State, ContinuationIndenter *Indenter, 641 bool DryRun) override; 642 643 unsigned formatFromToken(LineState &State, ContinuationIndenter *Indenter, 644 bool DryRun) override; 645 646 /// Adds \p Token as the next comma to the \c CommaSeparated list. 647 void CommaFound(const FormatToken *Token) override { 648 Commas.push_back(Token); 649 } 650 651 const FormatToken *lastComma() override { 652 if (Commas.empty()) 653 return nullptr; 654 return Commas.back(); 655 } 656 657 private: 658 /// A struct that holds information on how to format a given list with 659 /// a specific number of columns. 660 struct ColumnFormat { 661 /// The number of columns to use. 662 unsigned Columns; 663 664 /// The total width in characters. 665 unsigned TotalWidth; 666 667 /// The number of lines required for this format. 668 unsigned LineCount; 669 670 /// The size of each column in characters. 671 SmallVector<unsigned, 8> ColumnSizes; 672 }; 673 674 /// Calculate which \c ColumnFormat fits best into 675 /// \p RemainingCharacters. 676 const ColumnFormat *getColumnFormat(unsigned RemainingCharacters) const; 677 678 /// The ordered \c FormatTokens making up the commas of this list. 679 SmallVector<const FormatToken *, 8> Commas; 680 681 /// The length of each of the list's items in characters including the 682 /// trailing comma. 683 SmallVector<unsigned, 8> ItemLengths; 684 685 /// Precomputed formats that can be used for this list. 686 SmallVector<ColumnFormat, 4> Formats; 687 688 bool HasNestedBracedList; 689 }; 690 691 /// Encapsulates keywords that are context sensitive or for languages not 692 /// properly supported by Clang's lexer. 693 struct AdditionalKeywords { 694 AdditionalKeywords(IdentifierTable &IdentTable) { 695 kw_final = &IdentTable.get("final"); 696 kw_override = &IdentTable.get("override"); 697 kw_in = &IdentTable.get("in"); 698 kw_of = &IdentTable.get("of"); 699 kw_CF_CLOSED_ENUM = &IdentTable.get("CF_CLOSED_ENUM"); 700 kw_CF_ENUM = &IdentTable.get("CF_ENUM"); 701 kw_CF_OPTIONS = &IdentTable.get("CF_OPTIONS"); 702 kw_NS_CLOSED_ENUM = &IdentTable.get("NS_CLOSED_ENUM"); 703 kw_NS_ENUM = &IdentTable.get("NS_ENUM"); 704 kw_NS_OPTIONS = &IdentTable.get("NS_OPTIONS"); 705 706 kw_as = &IdentTable.get("as"); 707 kw_async = &IdentTable.get("async"); 708 kw_await = &IdentTable.get("await"); 709 kw_declare = &IdentTable.get("declare"); 710 kw_finally = &IdentTable.get("finally"); 711 kw_from = &IdentTable.get("from"); 712 kw_function = &IdentTable.get("function"); 713 kw_get = &IdentTable.get("get"); 714 kw_import = &IdentTable.get("import"); 715 kw_infer = &IdentTable.get("infer"); 716 kw_is = &IdentTable.get("is"); 717 kw_let = &IdentTable.get("let"); 718 kw_module = &IdentTable.get("module"); 719 kw_readonly = &IdentTable.get("readonly"); 720 kw_set = &IdentTable.get("set"); 721 kw_type = &IdentTable.get("type"); 722 kw_typeof = &IdentTable.get("typeof"); 723 kw_var = &IdentTable.get("var"); 724 kw_yield = &IdentTable.get("yield"); 725 726 kw_abstract = &IdentTable.get("abstract"); 727 kw_assert = &IdentTable.get("assert"); 728 kw_extends = &IdentTable.get("extends"); 729 kw_implements = &IdentTable.get("implements"); 730 kw_instanceof = &IdentTable.get("instanceof"); 731 kw_interface = &IdentTable.get("interface"); 732 kw_native = &IdentTable.get("native"); 733 kw_package = &IdentTable.get("package"); 734 kw_synchronized = &IdentTable.get("synchronized"); 735 kw_throws = &IdentTable.get("throws"); 736 kw___except = &IdentTable.get("__except"); 737 kw___has_include = &IdentTable.get("__has_include"); 738 kw___has_include_next = &IdentTable.get("__has_include_next"); 739 740 kw_mark = &IdentTable.get("mark"); 741 742 kw_extend = &IdentTable.get("extend"); 743 kw_option = &IdentTable.get("option"); 744 kw_optional = &IdentTable.get("optional"); 745 kw_repeated = &IdentTable.get("repeated"); 746 kw_required = &IdentTable.get("required"); 747 kw_returns = &IdentTable.get("returns"); 748 749 kw_signals = &IdentTable.get("signals"); 750 kw_qsignals = &IdentTable.get("Q_SIGNALS"); 751 kw_slots = &IdentTable.get("slots"); 752 kw_qslots = &IdentTable.get("Q_SLOTS"); 753 754 // C# keywords 755 kw_dollar = &IdentTable.get("dollar"); 756 kw_base = &IdentTable.get("base"); 757 kw_byte = &IdentTable.get("byte"); 758 kw_checked = &IdentTable.get("checked"); 759 kw_decimal = &IdentTable.get("decimal"); 760 kw_delegate = &IdentTable.get("delegate"); 761 kw_event = &IdentTable.get("event"); 762 kw_fixed = &IdentTable.get("fixed"); 763 kw_foreach = &IdentTable.get("foreach"); 764 kw_implicit = &IdentTable.get("implicit"); 765 kw_internal = &IdentTable.get("internal"); 766 kw_lock = &IdentTable.get("lock"); 767 kw_null = &IdentTable.get("null"); 768 kw_object = &IdentTable.get("object"); 769 kw_out = &IdentTable.get("out"); 770 kw_params = &IdentTable.get("params"); 771 kw_ref = &IdentTable.get("ref"); 772 kw_string = &IdentTable.get("string"); 773 kw_stackalloc = &IdentTable.get("stackalloc"); 774 kw_sbyte = &IdentTable.get("sbyte"); 775 kw_sealed = &IdentTable.get("sealed"); 776 kw_uint = &IdentTable.get("uint"); 777 kw_ulong = &IdentTable.get("ulong"); 778 kw_unchecked = &IdentTable.get("unchecked"); 779 kw_unsafe = &IdentTable.get("unsafe"); 780 kw_ushort = &IdentTable.get("ushort"); 781 kw_when = &IdentTable.get("when"); 782 783 // Keep this at the end of the constructor to make sure everything here 784 // is 785 // already initialized. 786 JsExtraKeywords = std::unordered_set<IdentifierInfo *>( 787 {kw_as, kw_async, kw_await, kw_declare, kw_finally, kw_from, 788 kw_function, kw_get, kw_import, kw_is, kw_let, kw_module, kw_readonly, 789 kw_set, kw_type, kw_typeof, kw_var, kw_yield, 790 // Keywords from the Java section. 791 kw_abstract, kw_extends, kw_implements, kw_instanceof, kw_interface}); 792 793 CSharpExtraKeywords = std::unordered_set<IdentifierInfo *>( 794 {kw_base, kw_byte, kw_checked, kw_decimal, kw_delegate, kw_event, 795 kw_fixed, kw_foreach, kw_implicit, kw_in, kw_interface, kw_internal, 796 kw_is, kw_lock, kw_null, kw_object, kw_out, kw_override, kw_params, 797 kw_readonly, kw_ref, kw_string, kw_stackalloc, kw_sbyte, kw_sealed, 798 kw_uint, kw_ulong, kw_unchecked, kw_unsafe, kw_ushort, kw_when, 799 // Keywords from the JavaScript section. 800 kw_as, kw_async, kw_await, kw_declare, kw_finally, kw_from, 801 kw_function, kw_get, kw_import, kw_is, kw_let, kw_module, kw_readonly, 802 kw_set, kw_type, kw_typeof, kw_var, kw_yield, 803 // Keywords from the Java section. 804 kw_abstract, kw_extends, kw_implements, kw_instanceof, kw_interface}); 805 } 806 807 // Context sensitive keywords. 808 IdentifierInfo *kw_final; 809 IdentifierInfo *kw_override; 810 IdentifierInfo *kw_in; 811 IdentifierInfo *kw_of; 812 IdentifierInfo *kw_CF_CLOSED_ENUM; 813 IdentifierInfo *kw_CF_ENUM; 814 IdentifierInfo *kw_CF_OPTIONS; 815 IdentifierInfo *kw_NS_CLOSED_ENUM; 816 IdentifierInfo *kw_NS_ENUM; 817 IdentifierInfo *kw_NS_OPTIONS; 818 IdentifierInfo *kw___except; 819 IdentifierInfo *kw___has_include; 820 IdentifierInfo *kw___has_include_next; 821 822 // JavaScript keywords. 823 IdentifierInfo *kw_as; 824 IdentifierInfo *kw_async; 825 IdentifierInfo *kw_await; 826 IdentifierInfo *kw_declare; 827 IdentifierInfo *kw_finally; 828 IdentifierInfo *kw_from; 829 IdentifierInfo *kw_function; 830 IdentifierInfo *kw_get; 831 IdentifierInfo *kw_import; 832 IdentifierInfo *kw_infer; 833 IdentifierInfo *kw_is; 834 IdentifierInfo *kw_let; 835 IdentifierInfo *kw_module; 836 IdentifierInfo *kw_readonly; 837 IdentifierInfo *kw_set; 838 IdentifierInfo *kw_type; 839 IdentifierInfo *kw_typeof; 840 IdentifierInfo *kw_var; 841 IdentifierInfo *kw_yield; 842 843 // Java keywords. 844 IdentifierInfo *kw_abstract; 845 IdentifierInfo *kw_assert; 846 IdentifierInfo *kw_extends; 847 IdentifierInfo *kw_implements; 848 IdentifierInfo *kw_instanceof; 849 IdentifierInfo *kw_interface; 850 IdentifierInfo *kw_native; 851 IdentifierInfo *kw_package; 852 IdentifierInfo *kw_synchronized; 853 IdentifierInfo *kw_throws; 854 855 // Pragma keywords. 856 IdentifierInfo *kw_mark; 857 858 // Proto keywords. 859 IdentifierInfo *kw_extend; 860 IdentifierInfo *kw_option; 861 IdentifierInfo *kw_optional; 862 IdentifierInfo *kw_repeated; 863 IdentifierInfo *kw_required; 864 IdentifierInfo *kw_returns; 865 866 // QT keywords. 867 IdentifierInfo *kw_signals; 868 IdentifierInfo *kw_qsignals; 869 IdentifierInfo *kw_slots; 870 IdentifierInfo *kw_qslots; 871 872 // C# keywords 873 IdentifierInfo *kw_dollar; 874 IdentifierInfo *kw_base; 875 IdentifierInfo *kw_byte; 876 IdentifierInfo *kw_checked; 877 IdentifierInfo *kw_decimal; 878 IdentifierInfo *kw_delegate; 879 IdentifierInfo *kw_event; 880 IdentifierInfo *kw_fixed; 881 IdentifierInfo *kw_foreach; 882 IdentifierInfo *kw_implicit; 883 IdentifierInfo *kw_internal; 884 885 IdentifierInfo *kw_lock; 886 IdentifierInfo *kw_null; 887 IdentifierInfo *kw_object; 888 IdentifierInfo *kw_out; 889 890 IdentifierInfo *kw_params; 891 892 IdentifierInfo *kw_ref; 893 IdentifierInfo *kw_string; 894 IdentifierInfo *kw_stackalloc; 895 IdentifierInfo *kw_sbyte; 896 IdentifierInfo *kw_sealed; 897 IdentifierInfo *kw_uint; 898 IdentifierInfo *kw_ulong; 899 IdentifierInfo *kw_unchecked; 900 IdentifierInfo *kw_unsafe; 901 IdentifierInfo *kw_ushort; 902 IdentifierInfo *kw_when; 903 904 /// Returns \c true if \p Tok is a true JavaScript identifier, returns 905 /// \c false if it is a keyword or a pseudo keyword. 906 bool IsJavaScriptIdentifier(const FormatToken &Tok) const { 907 return Tok.is(tok::identifier) && 908 JsExtraKeywords.find(Tok.Tok.getIdentifierInfo()) == 909 JsExtraKeywords.end(); 910 } 911 912 /// Returns \c true if \p Tok is a C# keyword, returns 913 /// \c false if it is a anything else. 914 bool isCSharpKeyword(const FormatToken &Tok) const { 915 switch (Tok.Tok.getKind()) { 916 case tok::kw_bool: 917 case tok::kw_break: 918 case tok::kw_case: 919 case tok::kw_catch: 920 case tok::kw_char: 921 case tok::kw_class: 922 case tok::kw_const: 923 case tok::kw_continue: 924 case tok::kw_default: 925 case tok::kw_do: 926 case tok::kw_double: 927 case tok::kw_else: 928 case tok::kw_enum: 929 case tok::kw_explicit: 930 case tok::kw_extern: 931 case tok::kw_false: 932 case tok::kw_float: 933 case tok::kw_for: 934 case tok::kw_goto: 935 case tok::kw_if: 936 case tok::kw_int: 937 case tok::kw_long: 938 case tok::kw_namespace: 939 case tok::kw_new: 940 case tok::kw_operator: 941 case tok::kw_private: 942 case tok::kw_protected: 943 case tok::kw_public: 944 case tok::kw_return: 945 case tok::kw_short: 946 case tok::kw_sizeof: 947 case tok::kw_static: 948 case tok::kw_struct: 949 case tok::kw_switch: 950 case tok::kw_this: 951 case tok::kw_throw: 952 case tok::kw_true: 953 case tok::kw_try: 954 case tok::kw_typeof: 955 case tok::kw_using: 956 case tok::kw_virtual: 957 case tok::kw_void: 958 case tok::kw_volatile: 959 case tok::kw_while: 960 return true; 961 default: 962 return Tok.is(tok::identifier) && 963 CSharpExtraKeywords.find(Tok.Tok.getIdentifierInfo()) == 964 CSharpExtraKeywords.end(); 965 } 966 } 967 968 private: 969 /// The JavaScript keywords beyond the C++ keyword set. 970 std::unordered_set<IdentifierInfo *> JsExtraKeywords; 971 972 /// The C# keywords beyond the C++ keyword set 973 std::unordered_set<IdentifierInfo *> CSharpExtraKeywords; 974 }; 975 976 } // namespace format 977 } // namespace clang 978 979 #endif 980