1 //===- TokenLexer.h - Lex from a token buffer -------------------*- C++ -*-===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 // This file defines the TokenLexer interface.
12 //===----------------------------------------------------------------------===//
14 #ifndef LLVM_CLANG_LEX_TOKENLEXER_H
15 #define LLVM_CLANG_LEX_TOKENLEXER_H
17 #include "clang/Basic/SourceLocation.h"
18 #include "llvm/ADT/ArrayRef.h"
26 class VAOptExpansionContext;
28 /// TokenLexer - This implements a lexer that returns tokens from a macro body
29 /// or token stream instead of lexing from a character buffer. This is used for
30 /// macro expansion and _Pragma handling, for example.
32 friend class Preprocessor;
34 /// Macro - The macro we are expanding from. This is null if expanding a
36 MacroInfo *Macro = nullptr;
38 /// ActualArgs - The actual arguments specified for a function-like macro, or
39 /// null. The TokenLexer owns the pointed-to object.
40 MacroArgs *ActualArgs = nullptr;
42 /// PP - The current preprocessor object we are expanding for.
45 /// Tokens - This is the pointer to an array of tokens that the macro is
46 /// defined to, with arguments expanded for function-like macros. If this is
47 /// a token stream, these are the tokens we are returning. This points into
48 /// the macro definition we are lexing from, a cache buffer that is owned by
49 /// the preprocessor, or some other buffer that we may or may not own
50 /// (depending on OwnsTokens).
51 /// Note that if it points into Preprocessor's cache buffer, the Preprocessor
52 /// may update the pointer as needed.
55 /// NumTokens - This is the length of the Tokens array.
58 /// This is the index of the next token that Lex will return.
61 /// ExpandLocStart/End - The source location range where this macro was
63 SourceLocation ExpandLocStart, ExpandLocEnd;
65 /// \brief Source location pointing at the source location entry chunk that
66 /// was reserved for the current macro expansion.
67 SourceLocation MacroExpansionStart;
69 /// \brief The offset of the macro expansion in the
70 /// "source location address space".
71 unsigned MacroStartSLocOffset;
73 /// \brief Location of the macro definition.
74 SourceLocation MacroDefStart;
76 /// \brief Length of the macro definition.
77 unsigned MacroDefLength;
79 /// Lexical information about the expansion point of the macro: the identifier
80 /// that the macro expanded from had these properties.
81 bool AtStartOfLine : 1;
82 bool HasLeadingSpace : 1;
84 // NextTokGetsSpace - When this is true, the next token appended to the
85 // output list during function argument expansion will get a leading space,
86 // regardless of whether it had one to begin with or not. This is used for
87 // placemarker support. If still true after function argument expansion, the
88 // leading space will be applied to the first token following the macro
90 bool NextTokGetsSpace : 1;
92 /// OwnsTokens - This is true if this TokenLexer allocated the Tokens
93 /// array, and thus needs to free it when destroyed. For simple object-like
94 /// macros (for example) we just point into the token buffer of the macro
95 /// definition, we don't make a copy of it.
98 /// DisableMacroExpansion - This is true when tokens lexed from the TokenLexer
99 /// should not be subject to further macro expansion.
100 bool DisableMacroExpansion : 1;
103 /// Create a TokenLexer for the specified macro with the specified actual
104 /// arguments. Note that this ctor takes ownership of the ActualArgs pointer.
105 /// ILEnd specifies the location of the ')' for a function-like macro or the
106 /// identifier for an object-like macro.
107 TokenLexer(Token &Tok, SourceLocation ILEnd, MacroInfo *MI,
108 MacroArgs *ActualArgs, Preprocessor &pp)
109 : PP(pp), OwnsTokens(false) {
110 Init(Tok, ILEnd, MI, ActualArgs);
113 /// Create a TokenLexer for the specified token stream. If 'OwnsTokens' is
114 /// specified, this takes ownership of the tokens and delete[]'s them when
115 /// the token lexer is empty.
116 TokenLexer(const Token *TokArray, unsigned NumToks, bool DisableExpansion,
117 bool ownsTokens, Preprocessor &pp)
118 : PP(pp), OwnsTokens(false) {
119 Init(TokArray, NumToks, DisableExpansion, ownsTokens);
122 TokenLexer(const TokenLexer &) = delete;
123 TokenLexer &operator=(const TokenLexer &) = delete;
124 ~TokenLexer() { destroy(); }
126 /// Init - Initialize this TokenLexer to expand from the specified macro
127 /// with the specified argument information. Note that this ctor takes
128 /// ownership of the ActualArgs pointer. ILEnd specifies the location of the
129 /// ')' for a function-like macro or the identifier for an object-like macro.
130 void Init(Token &Tok, SourceLocation ILEnd, MacroInfo *MI,
131 MacroArgs *ActualArgs);
133 /// Init - Initialize this TokenLexer with the specified token stream.
134 /// This does not take ownership of the specified token vector.
136 /// DisableExpansion is true when macro expansion of tokens lexed from this
137 /// stream should be disabled.
138 void Init(const Token *TokArray, unsigned NumToks,
139 bool DisableMacroExpansion, bool OwnsTokens);
141 /// isNextTokenLParen - If the next token lexed will pop this macro off the
142 /// expansion stack, return 2. If the next unexpanded token is a '(', return
143 /// 1, otherwise return 0.
144 unsigned isNextTokenLParen() const;
146 /// Lex - Lex and return a token from this macro stream.
147 bool Lex(Token &Tok);
149 /// isParsingPreprocessorDirective - Return true if we are in the middle of a
150 /// preprocessor directive.
151 bool isParsingPreprocessorDirective() const;
156 /// isAtEnd - Return true if the next lex call will pop this macro off the
158 bool isAtEnd() const {
159 return CurTokenIdx == NumTokens;
162 /// Concatenates the next (sub-)sequence of \p Tokens separated by '##'
163 /// starting with LHSTok - stopping when we encounter a token that is neither
164 /// '##' nor preceded by '##'. Places the result back into \p LHSTok and sets
165 /// \p CurIdx to point to the token following the last one that was pasted.
167 /// Also performs the MSVC extension wide-literal token pasting involved with:
168 /// \code L #macro-arg. \endcode
170 /// \param[in,out] LHSTok - Contains the token to the left of '##' in \p
171 /// Tokens upon entry and will contain the resulting concatenated Token upon
174 /// \param[in] TokenStream - The stream of Tokens we are lexing from.
176 /// \param[in,out] CurIdx - Upon entry, \pTokens[\pCurIdx] must equal '##'
177 /// (with the exception of the MSVC extension mentioned above). Upon exit, it
178 /// is set to the index of the token following the last token that was
179 /// concatenated together.
181 /// \returns If this returns true, the caller should immediately return the
183 bool pasteTokens(Token &LHSTok, ArrayRef<Token> TokenStream,
184 unsigned int &CurIdx);
186 /// Calls pasteTokens above, passing in the '*this' object's Tokens and
187 /// CurTokenIdx data members.
188 bool pasteTokens(Token &Tok);
191 /// Takes the tail sequence of tokens within ReplacementToks that represent
192 /// the just expanded __VA_OPT__ tokens (possibly zero tokens) and transforms
193 /// them into a string. \p VCtx is used to determine which token represents
194 /// the first __VA_OPT__ replacement token.
196 /// \param[in,out] ReplacementToks - Contains the current Replacement Tokens
197 /// (prior to rescanning and token pasting), the tail end of which represents
198 /// the tokens just expanded through __VA_OPT__ processing. These (sub)
199 /// sequence of tokens are folded into one stringified token.
201 /// \param[in] VCtx - contains relevent contextual information about the
202 /// state of the tokens around and including the __VA_OPT__ token, necessary
203 /// for stringification.
204 void stringifyVAOPTContents(SmallVectorImpl<Token> &ReplacementToks,
205 const VAOptExpansionContext &VCtx,
206 SourceLocation VAOPTClosingParenLoc);
208 /// Expand the arguments of a function-like macro so that we can quickly
209 /// return preexpanded tokens from Tokens.
210 void ExpandFunctionArguments();
212 /// HandleMicrosoftCommentPaste - In microsoft compatibility mode, /##/ pastes
213 /// together to form a comment that comments out everything in the current
214 /// macro, other active macros, and anything left on the current physical
215 /// source line of the expanded buffer. Handle this by returning the
216 /// first token on the next line.
217 void HandleMicrosoftCommentPaste(Token &Tok, SourceLocation OpLoc);
219 /// \brief If \p loc is a FileID and points inside the current macro
220 /// definition, returns the appropriate source location pointing at the
221 /// macro expansion source location entry.
222 SourceLocation getExpansionLocForMacroDefLoc(SourceLocation loc) const;
224 /// \brief Creates SLocEntries and updates the locations of macro argument
225 /// tokens to their new expanded locations.
227 /// \param ArgIdSpellLoc the location of the macro argument id inside the
228 /// macro definition.
229 void updateLocForMacroArgTokens(SourceLocation ArgIdSpellLoc,
230 Token *begin_tokens, Token *end_tokens);
232 /// Remove comma ahead of __VA_ARGS__, if present, according to compiler
233 /// dialect settings. Returns true if the comma is removed.
234 bool MaybeRemoveCommaBeforeVaArgs(SmallVectorImpl<Token> &ResultToks,
235 bool HasPasteOperator,
236 MacroInfo *Macro, unsigned MacroArgNo,
239 void PropagateLineStartLeadingSpaceInfo(Token &Result);
244 #endif // LLVM_CLANG_LEX_TOKENLEXER_H