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_LEX_TOKEN_H
     15 #define LLVM_CLANG_LEX_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 "llvm/ADT/StringRef.h"
     22 #include <cstdlib>
     23 
     24 namespace clang {
     25 
     26 class IdentifierInfo;
     27 
     28 /// Token - This structure provides full information about a lexed token.
     29 /// It is not intended to be space efficient, it is intended to return as much
     30 /// information as possible about each returned token.  This is expected to be
     31 /// compressed into a smaller form if memory footprint is important.
     32 ///
     33 /// The parser can create a special "annotation token" representing a stream of
     34 /// tokens that were parsed and semantically resolved, e.g.: "foo::MyClass<int>"
     35 /// can be represented by a single typename annotation token that carries
     36 /// information about the SourceRange of the tokens and the type object.
     37 class Token {
     38   /// The location of the token. This is actually a SourceLocation.
     39   unsigned Loc;
     40 
     41   // Conceptually these next two fields could be in a union.  However, this
     42   // causes gcc 4.2 to pessimize LexTokenInternal, a very performance critical
     43   // routine. Keeping as separate members with casts until a more beautiful fix
     44   // presents itself.
     45 
     46   /// UintData - This holds either the length of the token text, when
     47   /// a normal token, or the end of the SourceRange when an annotation
     48   /// token.
     49   unsigned UintData;
     50 
     51   /// PtrData - This is a union of four different pointer types, which depends
     52   /// on what type of token this is:
     53   ///  Identifiers, keywords, etc:
     54   ///    This is an IdentifierInfo*, which contains the uniqued identifier
     55   ///    spelling.
     56   ///  Literals:  isLiteral() returns true.
     57   ///    This is a pointer to the start of the token in a text buffer, which
     58   ///    may be dirty (have trigraphs / escaped newlines).
     59   ///  Annotations (resolved type names, C++ scopes, etc): isAnnotation().
     60   ///    This is a pointer to sema-specific data for the annotation token.
     61   ///  Eof:
     62   //     This is a pointer to a Decl.
     63   ///  Other:
     64   ///    This is null.
     65   void *PtrData;
     66 
     67   /// Kind - The actual flavor of token this is.
     68   tok::TokenKind Kind;
     69 
     70   /// Flags - Bits we track about this token, members of the TokenFlags enum.
     71   unsigned short Flags;
     72 public:
     73 
     74   // Various flags set per token:
     75   enum TokenFlags {
     76     StartOfLine   = 0x01,  // At start of line or only after whitespace
     77                            // (considering the line after macro expansion).
     78     LeadingSpace  = 0x02,  // Whitespace exists before this token (considering
     79                            // whitespace after macro expansion).
     80     DisableExpand = 0x04,  // This identifier may never be macro expanded.
     81     NeedsCleaning = 0x08,  // Contained an escaped newline or trigraph.
     82     LeadingEmptyMacro = 0x10, // Empty macro exists before this token.
     83     HasUDSuffix = 0x20,    // This string or character literal has a ud-suffix.
     84     HasUCN = 0x40,         // This identifier contains a UCN.
     85     IgnoredComma = 0x80,   // This comma is not a macro argument separator (MS).
     86     StringifiedInMacro = 0x100, // This string or character literal is formed by
     87                                 // macro stringizing or charizing operator.
     88   };
     89 
     90   tok::TokenKind getKind() const { return Kind; }
     91   void setKind(tok::TokenKind K) { Kind = K; }
     92 
     93   /// is/isNot - Predicates to check if this token is a specific kind, as in
     94   /// "if (Tok.is(tok::l_brace)) {...}".
     95   bool is(tok::TokenKind K) const { return Kind == K; }
     96   bool isNot(tok::TokenKind K) const { return Kind != K; }
     97   bool isOneOf(tok::TokenKind K1, tok::TokenKind K2) const {
     98     return is(K1) || is(K2);
     99   }
    100   template <typename... Ts>
    101   bool isOneOf(tok::TokenKind K1, tok::TokenKind K2, Ts... Ks) const {
    102     return is(K1) || isOneOf(K2, Ks...);
    103   }
    104 
    105   /// \brief Return true if this is a raw identifier (when lexing
    106   /// in raw mode) or a non-keyword identifier (when lexing in non-raw mode).
    107   bool isAnyIdentifier() const {
    108     return tok::isAnyIdentifier(getKind());
    109   }
    110 
    111   /// \brief Return true if this is a "literal", like a numeric
    112   /// constant, string, etc.
    113   bool isLiteral() const {
    114     return tok::isLiteral(getKind());
    115   }
    116 
    117   /// \brief Return true if this is any of tok::annot_* kind tokens.
    118   bool isAnnotation() const {
    119     return tok::isAnnotation(getKind());
    120   }
    121 
    122   /// \brief Return a source location identifier for the specified
    123   /// offset in the current file.
    124   SourceLocation getLocation() const {
    125     return SourceLocation::getFromRawEncoding(Loc);
    126   }
    127   unsigned getLength() const {
    128     assert(!isAnnotation() && "Annotation tokens have no length field");
    129     return UintData;
    130   }
    131 
    132   void setLocation(SourceLocation L) { Loc = L.getRawEncoding(); }
    133   void setLength(unsigned Len) {
    134     assert(!isAnnotation() && "Annotation tokens have no length field");
    135     UintData = Len;
    136   }
    137 
    138   SourceLocation getAnnotationEndLoc() const {
    139     assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token");
    140     return SourceLocation::getFromRawEncoding(UintData ? UintData : Loc);
    141   }
    142   void setAnnotationEndLoc(SourceLocation L) {
    143     assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token");
    144     UintData = L.getRawEncoding();
    145   }
    146 
    147   SourceLocation getLastLoc() const {
    148     return isAnnotation() ? getAnnotationEndLoc() : getLocation();
    149   }
    150 
    151   SourceLocation getEndLoc() const {
    152     return isAnnotation() ? getAnnotationEndLoc()
    153                           : getLocation().getLocWithOffset(getLength());
    154   }
    155 
    156   /// \brief SourceRange of the group of tokens that this annotation token
    157   /// represents.
    158   SourceRange getAnnotationRange() const {
    159     return SourceRange(getLocation(), getAnnotationEndLoc());
    160   }
    161   void setAnnotationRange(SourceRange R) {
    162     setLocation(R.getBegin());
    163     setAnnotationEndLoc(R.getEnd());
    164   }
    165 
    166   const char *getName() const { return tok::getTokenName(Kind); }
    167 
    168   /// \brief Reset all flags to cleared.
    169   void startToken() {
    170     Kind = tok::unknown;
    171     Flags = 0;
    172     PtrData = nullptr;
    173     UintData = 0;
    174     Loc = SourceLocation().getRawEncoding();
    175   }
    176 
    177   IdentifierInfo *getIdentifierInfo() const {
    178     assert(isNot(tok::raw_identifier) &&
    179            "getIdentifierInfo() on a tok::raw_identifier token!");
    180     assert(!isAnnotation() &&
    181            "getIdentifierInfo() on an annotation token!");
    182     if (isLiteral()) return nullptr;
    183     if (is(tok::eof)) return nullptr;
    184     return (IdentifierInfo*) PtrData;
    185   }
    186   void setIdentifierInfo(IdentifierInfo *II) {
    187     PtrData = (void*) II;
    188   }
    189 
    190   const void *getEofData() const {
    191     assert(is(tok::eof));
    192     return reinterpret_cast<const void *>(PtrData);
    193   }
    194   void setEofData(const void *D) {
    195     assert(is(tok::eof));
    196     assert(!PtrData);
    197     PtrData = const_cast<void *>(D);
    198   }
    199 
    200   /// getRawIdentifier - For a raw identifier token (i.e., an identifier
    201   /// lexed in raw mode), returns a reference to the text substring in the
    202   /// buffer if known.
    203   StringRef getRawIdentifier() const {
    204     assert(is(tok::raw_identifier));
    205     return StringRef(reinterpret_cast<const char *>(PtrData), getLength());
    206   }
    207   void setRawIdentifierData(const char *Ptr) {
    208     assert(is(tok::raw_identifier));
    209     PtrData = const_cast<char*>(Ptr);
    210   }
    211 
    212   /// getLiteralData - For a literal token (numeric constant, string, etc), this
    213   /// returns a pointer to the start of it in the text buffer if known, null
    214   /// otherwise.
    215   const char *getLiteralData() const {
    216     assert(isLiteral() && "Cannot get literal data of non-literal");
    217     return reinterpret_cast<const char*>(PtrData);
    218   }
    219   void setLiteralData(const char *Ptr) {
    220     assert(isLiteral() && "Cannot set literal data of non-literal");
    221     PtrData = const_cast<char*>(Ptr);
    222   }
    223 
    224   void *getAnnotationValue() const {
    225     assert(isAnnotation() && "Used AnnotVal on non-annotation token");
    226     return PtrData;
    227   }
    228   void setAnnotationValue(void *val) {
    229     assert(isAnnotation() && "Used AnnotVal on non-annotation token");
    230     PtrData = val;
    231   }
    232 
    233   /// \brief Set the specified flag.
    234   void setFlag(TokenFlags Flag) {
    235     Flags |= Flag;
    236   }
    237 
    238   /// \brief Unset the specified flag.
    239   void clearFlag(TokenFlags Flag) {
    240     Flags &= ~Flag;
    241   }
    242 
    243   /// \brief Return the internal represtation of the flags.
    244   ///
    245   /// This is only intended for low-level operations such as writing tokens to
    246   /// disk.
    247   unsigned getFlags() const {
    248     return Flags;
    249   }
    250 
    251   /// \brief Set a flag to either true or false.
    252   void setFlagValue(TokenFlags Flag, bool Val) {
    253     if (Val)
    254       setFlag(Flag);
    255     else
    256       clearFlag(Flag);
    257   }
    258 
    259   /// isAtStartOfLine - Return true if this token is at the start of a line.
    260   ///
    261   bool isAtStartOfLine() const { return (Flags & StartOfLine) ? true : false; }
    262 
    263   /// \brief Return true if this token has whitespace before it.
    264   ///
    265   bool hasLeadingSpace() const { return (Flags & LeadingSpace) ? true : false; }
    266 
    267   /// \brief Return true if this identifier token should never
    268   /// be expanded in the future, due to C99 6.10.3.4p2.
    269   bool isExpandDisabled() const {
    270     return (Flags & DisableExpand) ? true : false;
    271   }
    272 
    273   /// \brief Return true if we have an ObjC keyword identifier.
    274   bool isObjCAtKeyword(tok::ObjCKeywordKind objcKey) const;
    275 
    276   /// \brief Return the ObjC keyword kind.
    277   tok::ObjCKeywordKind getObjCKeywordID() const;
    278 
    279   /// \brief Return true if this token has trigraphs or escaped newlines in it.
    280   bool needsCleaning() const { return (Flags & NeedsCleaning) ? true : false; }
    281 
    282   /// \brief Return true if this token has an empty macro before it.
    283   ///
    284   bool hasLeadingEmptyMacro() const {
    285     return (Flags & LeadingEmptyMacro) ? true : false;
    286   }
    287 
    288   /// \brief Return true if this token is a string or character literal which
    289   /// has a ud-suffix.
    290   bool hasUDSuffix() const { return (Flags & HasUDSuffix) ? true : false; }
    291 
    292   /// Returns true if this token contains a universal character name.
    293   bool hasUCN() const { return (Flags & HasUCN) ? true : false; }
    294 
    295   /// Returns true if this token is formed by macro by stringizing or charizing
    296   /// operator.
    297   bool stringifiedInMacro() const {
    298     return (Flags & StringifiedInMacro) ? true : false;
    299   }
    300 };
    301 
    302 /// \brief Information about the conditional stack (\#if directives)
    303 /// currently active.
    304 struct PPConditionalInfo {
    305   /// \brief Location where the conditional started.
    306   SourceLocation IfLoc;
    307 
    308   /// \brief True if this was contained in a skipping directive, e.g.,
    309   /// in a "\#if 0" block.
    310   bool WasSkipping;
    311 
    312   /// \brief True if we have emitted tokens already, and now we're in
    313   /// an \#else block or something.  Only useful in Skipping blocks.
    314   bool FoundNonSkip;
    315 
    316   /// \brief True if we've seen a \#else in this block.  If so,
    317   /// \#elif/\#else directives are not allowed.
    318   bool FoundElse;
    319 };
    320 
    321 }  // end namespace clang
    322 
    323 namespace llvm {
    324   template <>
    325   struct isPodLike<clang::Token> { static const bool value = true; };
    326 }  // end namespace llvm
    327 
    328 #endif
    329