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 "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.
     39   SourceLocation 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   ///  Other:
     62   ///    This is null.
     63   void *PtrData;
     64 
     65   /// Kind - The actual flavor of token this is.
     66   tok::TokenKind 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                            // (considering the line after macro expansion).
     76     LeadingSpace  = 0x02,  // Whitespace exists before this token (considering
     77                            // whitespace after macro expansion).
     78     DisableExpand = 0x04,  // This identifier may never be macro expanded.
     79     NeedsCleaning = 0x08,  // Contained an escaped newline or trigraph.
     80     LeadingEmptyMacro = 0x10, // Empty macro exists before this token.
     81     HasUDSuffix = 0x20,    // This string or character literal has a ud-suffix.
     82     HasUCN = 0x40,         // This identifier contains a UCN.
     83     IgnoredComma = 0x80    // This comma is not a macro argument separator (MS).
     84   };
     85 
     86   tok::TokenKind getKind() const { return Kind; }
     87   void setKind(tok::TokenKind K) { Kind = K; }
     88 
     89   /// is/isNot - Predicates to check if this token is a specific kind, as in
     90   /// "if (Tok.is(tok::l_brace)) {...}".
     91   bool is(tok::TokenKind K) const { return Kind == K; }
     92   bool isNot(tok::TokenKind K) const { return Kind != K; }
     93 
     94   /// \brief Return true if this is a raw identifier (when lexing
     95   /// in raw mode) or a non-keyword identifier (when lexing in non-raw mode).
     96   bool isAnyIdentifier() const {
     97     return tok::isAnyIdentifier(getKind());
     98   }
     99 
    100   /// \brief Return true if this is a "literal", like a numeric
    101   /// constant, string, etc.
    102   bool isLiteral() const {
    103     return tok::isLiteral(getKind());
    104   }
    105 
    106   /// \brief Return true if this is any of tok::annot_* kind tokens.
    107   bool isAnnotation() const {
    108     return tok::isAnnotation(getKind());
    109   }
    110 
    111   /// \brief Return a source location identifier for the specified
    112   /// offset in the current file.
    113   SourceLocation getLocation() const { return Loc; }
    114   unsigned getLength() const {
    115     assert(!isAnnotation() && "Annotation tokens have no length field");
    116     return UintData;
    117   }
    118 
    119   void setLocation(SourceLocation L) { Loc = L; }
    120   void setLength(unsigned Len) {
    121     assert(!isAnnotation() && "Annotation tokens have no length field");
    122     UintData = Len;
    123   }
    124 
    125   SourceLocation getAnnotationEndLoc() const {
    126     assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token");
    127     return SourceLocation::getFromRawEncoding(UintData);
    128   }
    129   void setAnnotationEndLoc(SourceLocation L) {
    130     assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token");
    131     UintData = L.getRawEncoding();
    132   }
    133 
    134   SourceLocation getLastLoc() const {
    135     return isAnnotation() ? getAnnotationEndLoc() : getLocation();
    136   }
    137 
    138   /// \brief SourceRange of the group of tokens that this annotation token
    139   /// represents.
    140   SourceRange getAnnotationRange() const {
    141     return SourceRange(getLocation(), getAnnotationEndLoc());
    142   }
    143   void setAnnotationRange(SourceRange R) {
    144     setLocation(R.getBegin());
    145     setAnnotationEndLoc(R.getEnd());
    146   }
    147 
    148   const char *getName() const { return tok::getTokenName(Kind); }
    149 
    150   /// \brief Reset all flags to cleared.
    151   void startToken() {
    152     Kind = tok::unknown;
    153     Flags = 0;
    154     PtrData = nullptr;
    155     UintData = 0;
    156     Loc = SourceLocation();
    157   }
    158 
    159   IdentifierInfo *getIdentifierInfo() const {
    160     assert(isNot(tok::raw_identifier) &&
    161            "getIdentifierInfo() on a tok::raw_identifier token!");
    162     assert(!isAnnotation() &&
    163            "getIdentifierInfo() on an annotation token!");
    164     if (isLiteral()) return nullptr;
    165     return (IdentifierInfo*) PtrData;
    166   }
    167   void setIdentifierInfo(IdentifierInfo *II) {
    168     PtrData = (void*) II;
    169   }
    170 
    171   /// getRawIdentifier - For a raw identifier token (i.e., an identifier
    172   /// lexed in raw mode), returns a reference to the text substring in the
    173   /// buffer if known.
    174   StringRef getRawIdentifier() const {
    175     assert(is(tok::raw_identifier));
    176     return StringRef(reinterpret_cast<const char *>(PtrData), getLength());
    177   }
    178   void setRawIdentifierData(const char *Ptr) {
    179     assert(is(tok::raw_identifier));
    180     PtrData = const_cast<char*>(Ptr);
    181   }
    182 
    183   /// getLiteralData - For a literal token (numeric constant, string, etc), this
    184   /// returns a pointer to the start of it in the text buffer if known, null
    185   /// otherwise.
    186   const char *getLiteralData() const {
    187     assert(isLiteral() && "Cannot get literal data of non-literal");
    188     return reinterpret_cast<const char*>(PtrData);
    189   }
    190   void setLiteralData(const char *Ptr) {
    191     assert(isLiteral() && "Cannot set literal data of non-literal");
    192     PtrData = const_cast<char*>(Ptr);
    193   }
    194 
    195   void *getAnnotationValue() const {
    196     assert(isAnnotation() && "Used AnnotVal on non-annotation token");
    197     return PtrData;
    198   }
    199   void setAnnotationValue(void *val) {
    200     assert(isAnnotation() && "Used AnnotVal on non-annotation token");
    201     PtrData = val;
    202   }
    203 
    204   /// \brief Set the specified flag.
    205   void setFlag(TokenFlags Flag) {
    206     Flags |= Flag;
    207   }
    208 
    209   /// \brief Unset the specified flag.
    210   void clearFlag(TokenFlags Flag) {
    211     Flags &= ~Flag;
    212   }
    213 
    214   /// \brief Return the internal represtation of the flags.
    215   ///
    216   /// This is only intended for low-level operations such as writing tokens to
    217   /// disk.
    218   unsigned getFlags() const {
    219     return Flags;
    220   }
    221 
    222   /// \brief Set a flag to either true or false.
    223   void setFlagValue(TokenFlags Flag, bool Val) {
    224     if (Val)
    225       setFlag(Flag);
    226     else
    227       clearFlag(Flag);
    228   }
    229 
    230   /// isAtStartOfLine - Return true if this token is at the start of a line.
    231   ///
    232   bool isAtStartOfLine() const { return (Flags & StartOfLine) ? true : false; }
    233 
    234   /// \brief Return true if this token has whitespace before it.
    235   ///
    236   bool hasLeadingSpace() const { return (Flags & LeadingSpace) ? true : false; }
    237 
    238   /// \brief Return true if this identifier token should never
    239   /// be expanded in the future, due to C99 6.10.3.4p2.
    240   bool isExpandDisabled() const {
    241     return (Flags & DisableExpand) ? true : false;
    242   }
    243 
    244   /// \brief Return true if we have an ObjC keyword identifier.
    245   bool isObjCAtKeyword(tok::ObjCKeywordKind objcKey) const;
    246 
    247   /// \brief Return the ObjC keyword kind.
    248   tok::ObjCKeywordKind getObjCKeywordID() const;
    249 
    250   /// \brief Return true if this token has trigraphs or escaped newlines in it.
    251   bool needsCleaning() const { return (Flags & NeedsCleaning) ? true : false; }
    252 
    253   /// \brief Return true if this token has an empty macro before it.
    254   ///
    255   bool hasLeadingEmptyMacro() const {
    256     return (Flags & LeadingEmptyMacro) ? true : false;
    257   }
    258 
    259   /// \brief Return true if this token is a string or character literal which
    260   /// has a ud-suffix.
    261   bool hasUDSuffix() const { return (Flags & HasUDSuffix) ? true : false; }
    262 
    263   /// Returns true if this token contains a universal character name.
    264   bool hasUCN() const { return (Flags & HasUCN) ? true : false; }
    265 };
    266 
    267 /// \brief Information about the conditional stack (\#if directives)
    268 /// currently active.
    269 struct PPConditionalInfo {
    270   /// \brief Location where the conditional started.
    271   SourceLocation IfLoc;
    272 
    273   /// \brief True if this was contained in a skipping directive, e.g.,
    274   /// in a "\#if 0" block.
    275   bool WasSkipping;
    276 
    277   /// \brief True if we have emitted tokens already, and now we're in
    278   /// an \#else block or something.  Only useful in Skipping blocks.
    279   bool FoundNonSkip;
    280 
    281   /// \brief True if we've seen a \#else in this block.  If so,
    282   /// \#elif/\#else directives are not allowed.
    283   bool FoundElse;
    284 };
    285 
    286 }  // end namespace clang
    287 
    288 namespace llvm {
    289   template <>
    290   struct isPodLike<clang::Token> { static const bool value = true; };
    291 }  // end namespace llvm
    292 
    293 #endif
    294