Home | History | Annotate | Download | only in Lex
      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/OperatorKinds.h"
     18 #include "clang/Basic/SourceLocation.h"
     19 #include "clang/Basic/TemplateKinds.h"
     20 #include "clang/Basic/TokenKinds.h"
     21 #include <cstdlib>
     22 
     23 namespace clang {
     24 
     25 class 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.
     36 class 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 short Kind;
     67 
     68   /// Flags - Bits we track about this token, members of the TokenFlags enum.
     69   unsigned char Flags;
     70 public:
     71 
     72   // Various flags set per token:
     73   enum TokenFlags {
     74     StartOfLine   = 0x01,  // At start of line or only after whitespace.
     75     LeadingSpace  = 0x02,  // Whitespace exists before this token.
     76     DisableExpand = 0x04,  // This identifier may never be macro expanded.
     77     NeedsCleaning = 0x08,  // Contained an escaped newline or trigraph.
     78     LeadingEmptyMacro = 0x10, // Empty macro exists before this token.
     79     HasUDSuffix = 0x20,    // This string or character literal has a ud-suffix.
     80     HasUCN = 0x40          // This identifier contains a UCN.
     81   };
     82 
     83   tok::TokenKind getKind() const { return (tok::TokenKind)Kind; }
     84   void setKind(tok::TokenKind K) { Kind = K; }
     85 
     86   /// is/isNot - Predicates to check if this token is a specific kind, as in
     87   /// "if (Tok.is(tok::l_brace)) {...}".
     88   bool is(tok::TokenKind K) const { return Kind == (unsigned) K; }
     89   bool isNot(tok::TokenKind K) const { return Kind != (unsigned) K; }
     90 
     91   /// \brief Return true if this is a raw identifier (when lexing
     92   /// in raw mode) or a non-keyword identifier (when lexing in non-raw mode).
     93   bool isAnyIdentifier() const {
     94     return tok::isAnyIdentifier(getKind());
     95   }
     96 
     97   /// \brief Return true if this is a "literal", like a numeric
     98   /// constant, string, etc.
     99   bool isLiteral() const {
    100     return tok::isLiteral(getKind());
    101   }
    102 
    103   /// \brief Return true if this is any of tok::annot_* kind tokens.
    104   bool isAnnotation() const {
    105     return tok::isAnnotation(getKind());
    106   }
    107 
    108   /// \brief Return a source location identifier for the specified
    109   /// offset in the current file.
    110   SourceLocation getLocation() const { return Loc; }
    111   unsigned getLength() const {
    112     assert(!isAnnotation() && "Annotation tokens have no length field");
    113     return UintData;
    114   }
    115 
    116   void setLocation(SourceLocation L) { Loc = L; }
    117   void setLength(unsigned Len) {
    118     assert(!isAnnotation() && "Annotation tokens have no length field");
    119     UintData = Len;
    120   }
    121 
    122   SourceLocation getAnnotationEndLoc() const {
    123     assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token");
    124     return SourceLocation::getFromRawEncoding(UintData);
    125   }
    126   void setAnnotationEndLoc(SourceLocation L) {
    127     assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token");
    128     UintData = L.getRawEncoding();
    129   }
    130 
    131   SourceLocation getLastLoc() const {
    132     return isAnnotation() ? getAnnotationEndLoc() : getLocation();
    133   }
    134 
    135   /// \brief SourceRange of the group of tokens that this annotation token
    136   /// represents.
    137   SourceRange getAnnotationRange() const {
    138     return SourceRange(getLocation(), getAnnotationEndLoc());
    139   }
    140   void setAnnotationRange(SourceRange R) {
    141     setLocation(R.getBegin());
    142     setAnnotationEndLoc(R.getEnd());
    143   }
    144 
    145   const char *getName() const {
    146     return tok::getTokenName( (tok::TokenKind) Kind);
    147   }
    148 
    149   /// \brief Reset all flags to cleared.
    150   void startToken() {
    151     Kind = tok::unknown;
    152     Flags = 0;
    153     PtrData = 0;
    154     UintData = 0;
    155     Loc = SourceLocation();
    156   }
    157 
    158   IdentifierInfo *getIdentifierInfo() const {
    159     assert(isNot(tok::raw_identifier) &&
    160            "getIdentifierInfo() on a tok::raw_identifier token!");
    161     assert(!isAnnotation() &&
    162            "getIdentifierInfo() on an annotation token!");
    163     if (isLiteral()) return 0;
    164     return (IdentifierInfo*) PtrData;
    165   }
    166   void setIdentifierInfo(IdentifierInfo *II) {
    167     PtrData = (void*) II;
    168   }
    169 
    170   /// getRawIdentifierData - For a raw identifier token (i.e., an identifier
    171   /// lexed in raw mode), returns a pointer to the start of it in the text
    172   /// buffer if known, null otherwise.
    173   const char *getRawIdentifierData() const {
    174     assert(is(tok::raw_identifier));
    175     return reinterpret_cast<const char*>(PtrData);
    176   }
    177   void setRawIdentifierData(const char *Ptr) {
    178     assert(is(tok::raw_identifier));
    179     PtrData = const_cast<char*>(Ptr);
    180   }
    181 
    182   /// getLiteralData - For a literal token (numeric constant, string, etc), this
    183   /// returns a pointer to the start of it in the text buffer if known, null
    184   /// otherwise.
    185   const char *getLiteralData() const {
    186     assert(isLiteral() && "Cannot get literal data of non-literal");
    187     return reinterpret_cast<const char*>(PtrData);
    188   }
    189   void setLiteralData(const char *Ptr) {
    190     assert(isLiteral() && "Cannot set literal data of non-literal");
    191     PtrData = const_cast<char*>(Ptr);
    192   }
    193 
    194   void *getAnnotationValue() const {
    195     assert(isAnnotation() && "Used AnnotVal on non-annotation token");
    196     return PtrData;
    197   }
    198   void setAnnotationValue(void *val) {
    199     assert(isAnnotation() && "Used AnnotVal on non-annotation token");
    200     PtrData = val;
    201   }
    202 
    203   /// \brief Set the specified flag.
    204   void setFlag(TokenFlags Flag) {
    205     Flags |= Flag;
    206   }
    207 
    208   /// \brief Unset the specified flag.
    209   void clearFlag(TokenFlags Flag) {
    210     Flags &= ~Flag;
    211   }
    212 
    213   /// \brief Return the internal represtation of the flags.
    214   ///
    215   /// This is only intended for low-level operations such as writing tokens to
    216   /// disk.
    217   unsigned getFlags() const {
    218     return Flags;
    219   }
    220 
    221   /// \brief Set a flag to either true or false.
    222   void setFlagValue(TokenFlags Flag, bool Val) {
    223     if (Val)
    224       setFlag(Flag);
    225     else
    226       clearFlag(Flag);
    227   }
    228 
    229   /// isAtStartOfLine - Return true if this token is at the start of a line.
    230   ///
    231   bool isAtStartOfLine() const { return (Flags & StartOfLine) ? true : false; }
    232 
    233   /// \brief Return true if this token has whitespace before it.
    234   ///
    235   bool hasLeadingSpace() const { return (Flags & LeadingSpace) ? true : false; }
    236 
    237   /// \brief Return true if this identifier token should never
    238   /// be expanded in the future, due to C99 6.10.3.4p2.
    239   bool isExpandDisabled() const {
    240     return (Flags & DisableExpand) ? true : false;
    241   }
    242 
    243   /// \brief Return true if we have an ObjC keyword identifier.
    244   bool isObjCAtKeyword(tok::ObjCKeywordKind objcKey) const;
    245 
    246   /// \brief Return the ObjC keyword kind.
    247   tok::ObjCKeywordKind getObjCKeywordID() const;
    248 
    249   /// \brief Return true if this token has trigraphs or escaped newlines in it.
    250   bool needsCleaning() const { return (Flags & NeedsCleaning) ? true : false; }
    251 
    252   /// \brief Return true if this token has an empty macro before it.
    253   ///
    254   bool hasLeadingEmptyMacro() const {
    255     return (Flags & LeadingEmptyMacro) ? true : false;
    256   }
    257 
    258   /// \brief Return true if this token is a string or character literal which
    259   /// has a ud-suffix.
    260   bool hasUDSuffix() const { return (Flags & HasUDSuffix) ? true : false; }
    261 
    262   /// Returns true if this token contains a universal character name.
    263   bool hasUCN() const { return (Flags & HasUCN) ? true : false; }
    264 };
    265 
    266 /// \brief Information about the conditional stack (\#if directives)
    267 /// currently active.
    268 struct PPConditionalInfo {
    269   /// \brief Location where the conditional started.
    270   SourceLocation IfLoc;
    271 
    272   /// \brief True if this was contained in a skipping directive, e.g.,
    273   /// in a "\#if 0" block.
    274   bool WasSkipping;
    275 
    276   /// \brief True if we have emitted tokens already, and now we're in
    277   /// an \#else block or something.  Only useful in Skipping blocks.
    278   bool FoundNonSkip;
    279 
    280   /// \brief True if we've seen a \#else in this block.  If so,
    281   /// \#elif/\#else directives are not allowed.
    282   bool FoundElse;
    283 };
    284 
    285 }  // end namespace clang
    286 
    287 namespace llvm {
    288   template <>
    289   struct isPodLike<clang::Token> { static const bool value = true; };
    290 }  // end namespace llvm
    291 
    292 #endif
    293