Token.h revision be74740cc246ce08d42804a684385a42eb814edb
1//===--- Token.h - Token interface ------------------------------*- 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// This file defines the Token interface. 11// 12//===----------------------------------------------------------------------===// 13 14#ifndef LLVM_CLANG_TOKEN_H 15#define LLVM_CLANG_TOKEN_H 16 17#include "clang/Basic/TemplateKinds.h" 18#include "clang/Basic/TokenKinds.h" 19#include "clang/Basic/SourceLocation.h" 20#include "clang/Basic/OperatorKinds.h" 21#include <cstdlib> 22 23namespace clang { 24 25class IdentifierInfo; 26 27/// Token - This structure provides full information about a lexed token. 28/// It is not intended to be space efficient, it is intended to return as much 29/// information as possible about each returned token. This is expected to be 30/// compressed into a smaller form if memory footprint is important. 31/// 32/// The parser can create a special "annotation token" representing a stream of 33/// tokens that were parsed and semantically resolved, e.g.: "foo::MyClass<int>" 34/// can be represented by a single typename annotation token that carries 35/// information about the SourceRange of the tokens and the type object. 36class Token { 37 /// The location of the token. 38 SourceLocation Loc; 39 40 // Conceptually these next two fields could be in a union. However, this 41 // causes gcc 4.2 to pessimize LexTokenInternal, a very performance critical 42 // routine. Keeping as separate members with casts until a more beautiful fix 43 // presents itself. 44 45 /// UintData - This holds either the length of the token text, when 46 /// a normal token, or the end of the SourceRange when an annotation 47 /// token. 48 unsigned UintData; 49 50 /// PtrData - This is a union of four different pointer types, which depends 51 /// on what type of token this is: 52 /// Identifiers, keywords, etc: 53 /// This is an IdentifierInfo*, which contains the uniqued identifier 54 /// spelling. 55 /// Literals: isLiteral() returns true. 56 /// This is a pointer to the start of the token in a text buffer, which 57 /// may be dirty (have trigraphs / escaped newlines). 58 /// Annotations (resolved type names, C++ scopes, etc): isAnnotation(). 59 /// This is a pointer to sema-specific data for the annotation token. 60 /// Other: 61 /// This is null. 62 void *PtrData; 63 64 /// Kind - The actual flavor of token this is. 65 /// 66 unsigned char Kind; // DON'T make Kind a 'tok::TokenKind'; 67 // MSVC will treat it as a signed char and 68 // TokenKinds > 127 won't be handled correctly. 69 70 /// Flags - Bits we track about this token, members of the TokenFlags enum. 71 unsigned char Flags; 72public: 73 74 // Various flags set per token: 75 enum TokenFlags { 76 StartOfLine = 0x01, // At start of line or only after whitespace. 77 LeadingSpace = 0x02, // Whitespace exists before this token. 78 DisableExpand = 0x04, // This identifier may never be macro expanded. 79 NeedsCleaning = 0x08, // Contained an escaped newline or trigraph. 80 CPlusPlusOpKeyword = 0x10 // alternative representation of 81 // a C++ operator in objc selectors. 82 }; 83 84 tok::TokenKind getKind() const { return (tok::TokenKind)Kind; } 85 void setKind(tok::TokenKind K) { Kind = K; } 86 87 /// is/isNot - Predicates to check if this token is a specific kind, as in 88 /// "if (Tok.is(tok::l_brace)) {...}". 89 bool is(tok::TokenKind K) const { return Kind == (unsigned) K; } 90 bool isNot(tok::TokenKind K) const { return Kind != (unsigned) K; } 91 92 /// isLiteral - Return true if this is a "literal", like a numeric 93 /// constant, string, etc. 94 bool isLiteral() const { 95 return is(tok::numeric_constant) || is(tok::char_constant) || 96 is(tok::string_literal) || is(tok::wide_string_literal) || 97 is(tok::angle_string_literal); 98 } 99 100 bool isAnnotation() const { 101 return is(tok::annot_typename) || 102 is(tok::annot_cxxscope) || 103 is(tok::annot_template_id); 104 } 105 106 /// getLocation - Return a source location identifier for the specified 107 /// offset in the current file. 108 SourceLocation getLocation() const { return Loc; } 109 unsigned getLength() const { 110 assert(!isAnnotation() && "Annotation tokens have no length field"); 111 return UintData; 112 } 113 114 void setLocation(SourceLocation L) { Loc = L; } 115 void setLength(unsigned Len) { 116 assert(!isAnnotation() && "Annotation tokens have no length field"); 117 UintData = Len; 118 } 119 120 SourceLocation getAnnotationEndLoc() const { 121 assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token"); 122 return SourceLocation::getFromRawEncoding(UintData); 123 } 124 void setAnnotationEndLoc(SourceLocation L) { 125 assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token"); 126 UintData = L.getRawEncoding(); 127 } 128 129 SourceLocation getLastLoc() const { 130 return isAnnotation() ? getAnnotationEndLoc() : getLocation(); 131 } 132 133 /// getAnnotationRange - SourceRange of the group of tokens that this 134 /// annotation token represents. 135 SourceRange getAnnotationRange() const { 136 return SourceRange(getLocation(), getAnnotationEndLoc()); 137 } 138 void setAnnotationRange(SourceRange R) { 139 setLocation(R.getBegin()); 140 setAnnotationEndLoc(R.getEnd()); 141 } 142 143 const char *getName() const { 144 return tok::getTokenName( (tok::TokenKind) Kind); 145 } 146 147 /// startToken - Reset all flags to cleared. 148 /// 149 void startToken() { 150 Kind = tok::unknown; 151 Flags = 0; 152 PtrData = 0; 153 UintData = 0; 154 Loc = SourceLocation(); 155 } 156 157 IdentifierInfo *getIdentifierInfo() const { 158 assert(!isAnnotation() && "Used IdentInfo on annotation token!"); 159 if (isLiteral()) return 0; 160 return (IdentifierInfo*) PtrData; 161 } 162 void setIdentifierInfo(IdentifierInfo *II) { 163 PtrData = (void*) II; 164 } 165 166 /// getLiteralData - For a literal token (numeric constant, string, etc), this 167 /// returns a pointer to the start of it in the text buffer if known, null 168 /// otherwise. 169 const char *getLiteralData() const { 170 assert(isLiteral() && "Cannot get literal data of non-literal"); 171 return reinterpret_cast<const char*>(PtrData); 172 } 173 void setLiteralData(const char *Ptr) { 174 assert(isLiteral() && "Cannot set literal data of non-literal"); 175 PtrData = const_cast<char*>(Ptr); 176 } 177 178 void *getAnnotationValue() const { 179 assert(isAnnotation() && "Used AnnotVal on non-annotation token"); 180 return PtrData; 181 } 182 void setAnnotationValue(void *val) { 183 assert(isAnnotation() && "Used AnnotVal on non-annotation token"); 184 PtrData = val; 185 } 186 187 /// setFlag - Set the specified flag. 188 void setFlag(TokenFlags Flag) { 189 Flags |= Flag; 190 } 191 192 /// clearFlag - Unset the specified flag. 193 void clearFlag(TokenFlags Flag) { 194 Flags &= ~Flag; 195 } 196 197 /// getFlags - Return the internal represtation of the flags. 198 /// Only intended for low-level operations such as writing tokens to 199 // disk. 200 unsigned getFlags() const { 201 return Flags; 202 } 203 204 /// setFlagValue - Set a flag to either true or false. 205 void setFlagValue(TokenFlags Flag, bool Val) { 206 if (Val) 207 setFlag(Flag); 208 else 209 clearFlag(Flag); 210 } 211 212 /// isAtStartOfLine - Return true if this token is at the start of a line. 213 /// 214 bool isAtStartOfLine() const { return (Flags & StartOfLine) ? true : false; } 215 216 /// hasLeadingSpace - Return true if this token has whitespace before it. 217 /// 218 bool hasLeadingSpace() const { return (Flags & LeadingSpace) ? true : false; } 219 220 /// isExpandDisabled - Return true if this identifier token should never 221 /// be expanded in the future, due to C99 6.10.3.4p2. 222 bool isExpandDisabled() const { 223 return (Flags & DisableExpand) ? true : false; 224 } 225 226 /// isObjCAtKeyword - Return true if we have an ObjC keyword identifier. 227 bool isObjCAtKeyword(tok::ObjCKeywordKind objcKey) const; 228 229 /// getObjCKeywordID - Return the ObjC keyword kind. 230 tok::ObjCKeywordKind getObjCKeywordID() const; 231 232 /// needsCleaning - Return true if this token has trigraphs or escaped 233 /// newlines in it. 234 /// 235 bool needsCleaning() const { return (Flags & NeedsCleaning) ? true : false; } 236 237 /// isCPlusPlusOpKeyword - Return true if this token is an operator 238 /// for C++ operator keywords. 239 bool isCPlusPlusOpKeyword() const 240 { return (Flags & CPlusPlusOpKeyword) ? true : false; } 241 242}; 243 244/// PPConditionalInfo - Information about the conditional stack (#if directives) 245/// currently active. 246struct PPConditionalInfo { 247 /// IfLoc - Location where the conditional started. 248 /// 249 SourceLocation IfLoc; 250 251 /// WasSkipping - True if this was contained in a skipping directive, e.g. 252 /// in a "#if 0" block. 253 bool WasSkipping; 254 255 /// FoundNonSkip - True if we have emitted tokens already, and now we're in 256 /// an #else block or something. Only useful in Skipping blocks. 257 bool FoundNonSkip; 258 259 /// FoundElse - True if we've seen a #else in this block. If so, 260 /// #elif/#else directives are not allowed. 261 bool FoundElse; 262}; 263 264} // end namespace clang 265 266namespace llvm { 267 template <> 268 struct isPodLike<clang::Token> { static const bool value = true; }; 269} // end namespace llvm 270 271#endif 272