1//===- TokenLexer.h - Lex from a token buffer -------------------*- 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// This file defines the TokenLexer interface.
10//
11//===----------------------------------------------------------------------===//
12
13#ifndef LLVM_CLANG_LEX_TOKENLEXER_H
14#define LLVM_CLANG_LEX_TOKENLEXER_H
15
16#include "clang/Basic/SourceLocation.h"
17#include "llvm/ADT/ArrayRef.h"
18
19namespace clang {
20
21class MacroArgs;
22class MacroInfo;
23class Preprocessor;
24class Token;
25class VAOptExpansionContext;
26
27/// TokenLexer - This implements a lexer that returns tokens from a macro body
28/// or token stream instead of lexing from a character buffer. This is used for
29/// macro expansion and _Pragma handling, for example.
30class TokenLexer {
31 friend class Preprocessor;
32
33 /// The macro we are expanding from. This is null if expanding a token stream.
34 MacroInfo *Macro = nullptr;
35
36 /// The actual arguments specified for a function-like macro, or null. The
37 /// TokenLexer owns the pointed-to object.
38 MacroArgs *ActualArgs = nullptr;
39
40 /// The current preprocessor object we are expanding for.
41 Preprocessor &PP;
42
43 /// This is the pointer to an array of tokens that the macro is
44 /// defined to, with arguments expanded for function-like macros. If this is
45 /// a token stream, these are the tokens we are returning. This points into
46 /// the macro definition we are lexing from, a cache buffer that is owned by
47 /// the preprocessor, or some other buffer that we may or may not own
48 /// (depending on OwnsTokens).
49 /// Note that if it points into Preprocessor's cache buffer, the Preprocessor
50 /// may update the pointer as needed.
51 const Token *Tokens;
52
53 /// This is the length of the Tokens array.
54 unsigned NumTokens;
55
56 /// This is the index of the next token that Lex will return.
57 unsigned CurTokenIdx;
58
59 /// The source location range where this macro was expanded.
60 SourceLocation ExpandLocStart, ExpandLocEnd;
61
62 /// Source location pointing at the source location entry chunk that
63 /// was reserved for the current macro expansion.
64 SourceLocation MacroExpansionStart;
65
66 /// The offset of the macro expansion in the
67 /// "source location address space".
68 unsigned MacroStartSLocOffset;
69
70 /// Location of the macro definition.
71 SourceLocation MacroDefStart;
72
73 /// Length of the macro definition.
74 unsigned MacroDefLength;
75
76 /// Lexical information about the expansion point of the macro: the identifier
77 /// that the macro expanded from had these properties.
78 bool AtStartOfLine : 1;
79 bool HasLeadingSpace : 1;
80
81 // When this is true, the next token appended to the
82 // output list during function argument expansion will get a leading space,
83 // regardless of whether it had one to begin with or not. This is used for
84 // placemarker support. If still true after function argument expansion, the
85 // leading space will be applied to the first token following the macro
86 // expansion.
87 bool NextTokGetsSpace : 1;
88
89 /// This is true if this TokenLexer allocated the Tokens
90 /// array, and thus needs to free it when destroyed. For simple object-like
91 /// macros (for example) we just point into the token buffer of the macro
92 /// definition, we don't make a copy of it.
93 bool OwnsTokens : 1;
94
95 /// This is true when tokens lexed from the TokenLexer
96 /// should not be subject to further macro expansion.
97 bool DisableMacroExpansion : 1;
98
99public:
100 /// Create a TokenLexer for the specified macro with the specified actual
101 /// arguments. Note that this ctor takes ownership of the ActualArgs pointer.
102 /// ILEnd specifies the location of the ')' for a function-like macro or the
103 /// identifier for an object-like macro.
104 TokenLexer(Token &Tok, SourceLocation ILEnd, MacroInfo *MI,
105 MacroArgs *ActualArgs, Preprocessor &pp)
106 : PP(pp), OwnsTokens(false) {
107 Init(Tok, ILEnd, MI, ActualArgs);
108 }
109
110 /// Create a TokenLexer for the specified token stream. If 'OwnsTokens' is
111 /// specified, this takes ownership of the tokens and delete[]'s them when
112 /// the token lexer is empty.
113 TokenLexer(const Token *TokArray, unsigned NumToks, bool DisableExpansion,
114 bool ownsTokens, Preprocessor &pp)
115 : PP(pp), OwnsTokens(false) {
116 Init(TokArray, NumToks, DisableExpansion, ownsTokens);
117 }
118
119 TokenLexer(const TokenLexer &) = delete;
120 TokenLexer &operator=(const TokenLexer &) = delete;
121 ~TokenLexer() { destroy(); }
122
123 /// Initialize this TokenLexer to expand from the specified macro
124 /// with the specified argument information. Note that this ctor takes
125 /// ownership of the ActualArgs pointer. ILEnd specifies the location of the
126 /// ')' for a function-like macro or the identifier for an object-like macro.
127 void Init(Token &Tok, SourceLocation ELEnd, MacroInfo *MI,
128 MacroArgs *Actuals);
129
130 /// Initialize this TokenLexer with the specified token stream.
131 /// This does not take ownership of the specified token vector.
132 ///
133 /// DisableExpansion is true when macro expansion of tokens lexed from this
134 /// stream should be disabled.
135 void Init(const Token *TokArray, unsigned NumToks,
136 bool DisableMacroExpansion, bool OwnsTokens);
137
138 /// If the next token lexed will pop this macro off the
139 /// expansion stack, return 2. If the next unexpanded token is a '(', return
140 /// 1, otherwise return 0.
141 unsigned isNextTokenLParen() const;
142
143 /// Lex and return a token from this macro stream.
144 bool Lex(Token &Tok);
145
146 /// isParsingPreprocessorDirective - Return true if we are in the middle of a
147 /// preprocessor directive.
148 bool isParsingPreprocessorDirective() const;
149
150private:
151 void destroy();
152
153 /// Return true if the next lex call will pop this macro off the include
154 /// stack.
155 bool isAtEnd() const {
156 return CurTokenIdx == NumTokens;
157 }
158
159 /// Concatenates the next (sub-)sequence of \p Tokens separated by '##'
160 /// starting with LHSTok - stopping when we encounter a token that is neither
161 /// '##' nor preceded by '##'. Places the result back into \p LHSTok and sets
162 /// \p CurIdx to point to the token following the last one that was pasted.
163 ///
164 /// Also performs the MSVC extension wide-literal token pasting involved with:
165 /// \code L #macro-arg. \endcode
166 ///
167 /// \param[in,out] LHSTok - Contains the token to the left of '##' in \p
168 /// Tokens upon entry and will contain the resulting concatenated Token upon
169 /// exit.
170 ///
171 /// \param[in] TokenStream - The stream of Tokens we are lexing from.
172 ///
173 /// \param[in,out] CurIdx - Upon entry, \pTokens[\pCurIdx] must equal '##'
174 /// (with the exception of the MSVC extension mentioned above). Upon exit, it
175 /// is set to the index of the token following the last token that was
176 /// concatenated together.
177 ///
178 /// \returns If this returns true, the caller should immediately return the
179 /// token.
180 bool pasteTokens(Token &LHSTok, ArrayRef<Token> TokenStream,
181 unsigned int &CurIdx);
182
183 /// Calls pasteTokens above, passing in the '*this' object's Tokens and
184 /// CurTokenIdx data members.
185 bool pasteTokens(Token &Tok);
186
187
188 /// Takes the tail sequence of tokens within ReplacementToks that represent
189 /// the just expanded __VA_OPT__ tokens (possibly zero tokens) and transforms
190 /// them into a string. \p VCtx is used to determine which token represents
191 /// the first __VA_OPT__ replacement token.
192 ///
193 /// \param[in,out] ResultToks - Contains the current Replacement Tokens
194 /// (prior to rescanning and token pasting), the tail end of which represents
195 /// the tokens just expanded through __VA_OPT__ processing. These (sub)
196 /// sequence of tokens are folded into one stringified token.
197 ///
198 /// \param[in] VCtx - contains relevant contextual information about the
199 /// state of the tokens around and including the __VA_OPT__ token, necessary
200 /// for stringification.
201 void stringifyVAOPTContents(SmallVectorImpl<Token> &ResultToks,
202 const VAOptExpansionContext &VCtx,
203 SourceLocation VAOPTClosingParenLoc);
204
205 /// Expand the arguments of a function-like macro so that we can quickly
206 /// return preexpanded tokens from Tokens.
207 void ExpandFunctionArguments();
208
209 /// In microsoft compatibility mode, /##/ pastes
210 /// together to form a comment that comments out everything in the current
211 /// macro, other active macros, and anything left on the current physical
212 /// source line of the expanded buffer. Handle this by returning the
213 /// first token on the next line.
214 void HandleMicrosoftCommentPaste(Token &Tok, SourceLocation OpLoc);
215
216 /// If \p loc is a FileID and points inside the current macro
217 /// definition, returns the appropriate source location pointing at the
218 /// macro expansion source location entry.
219 SourceLocation getExpansionLocForMacroDefLoc(SourceLocation loc) const;
220
221 /// Creates SLocEntries and updates the locations of macro argument
222 /// tokens to their new expanded locations.
223 ///
224 /// \param ArgIdSpellLoc the location of the macro argument id inside the
225 /// macro definition.
226 void updateLocForMacroArgTokens(SourceLocation ArgIdSpellLoc,
227 Token *begin_tokens, Token *end_tokens);
228
229 /// Remove comma ahead of __VA_ARGS__, if present, according to compiler
230 /// dialect settings. Returns true if the comma is removed.
231 bool MaybeRemoveCommaBeforeVaArgs(SmallVectorImpl<Token> &ResultToks,
232 bool HasPasteOperator,
233 MacroInfo *Macro, unsigned MacroArgNo,
234 Preprocessor &PP);
235
236 void PropagateLineStartLeadingSpaceInfo(Token &Result);
237};
238
239} // namespace clang
240
241#endif // LLVM_CLANG_LEX_TOKENLEXER_H
242