Token.h revision 6cf750298d3621d8a10a6dd07fcee8e274b9d94d 
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  };
81
82  tok::TokenKind getKind() const { return (tok::TokenKind)Kind; }
83  void setKind(tok::TokenKind K) { Kind = K; }
84
85  /// is/isNot - Predicates to check if this token is a specific kind, as in
86  /// "if (Tok.is(tok::l_brace)) {...}".
87  bool is(tok::TokenKind K) const { return Kind == (unsigned) K; }
88  bool isNot(tok::TokenKind K) const { return Kind != (unsigned) K; }
89
90  /// isLiteral - Return true if this is a "literal", like a numeric
91  /// constant, string, etc.
92  bool isLiteral() const {
93    return is(tok::numeric_constant) || is(tok::char_constant) ||
94           is(tok::string_literal) || is(tok::wide_string_literal) ||
95           is(tok::angle_string_literal);
96  }
97
98  bool isAnnotation() const {
99    return is(tok::annot_typename) ||
100           is(tok::annot_cxxscope) ||
101           is(tok::annot_template_id);
102  }
103
104  /// getLocation - Return a source location identifier for the specified
105  /// offset in the current file.
106  SourceLocation getLocation() const { return Loc; }
107  unsigned getLength() const {
108    assert(!isAnnotation() && "Annotation tokens have no length field");
109    return UintData;
110  }
111
112  void setLocation(SourceLocation L) { Loc = L; }
113  void setLength(unsigned Len) {
114    assert(!isAnnotation() && "Annotation tokens have no length field");
115    UintData = Len;
116  }
117
118  SourceLocation getAnnotationEndLoc() const {
119    assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token");
120    return SourceLocation::getFromRawEncoding(UintData);
121  }
122  void setAnnotationEndLoc(SourceLocation L) {
123    assert(isAnnotation() && "Used AnnotEndLocID on non-annotation token");
124    UintData = L.getRawEncoding();
125  }
126
127  SourceLocation getLastLoc() const {
128    return isAnnotation() ? getAnnotationEndLoc() : getLocation();
129  }
130
131  /// getAnnotationRange - SourceRange of the group of tokens that this
132  /// annotation token represents.
133  SourceRange getAnnotationRange() const {
134    return SourceRange(getLocation(), getAnnotationEndLoc());
135  }
136  void setAnnotationRange(SourceRange R) {
137    setLocation(R.getBegin());
138    setAnnotationEndLoc(R.getEnd());
139  }
140
141  const char *getName() const {
142    return tok::getTokenName( (tok::TokenKind) Kind);
143  }
144
145  /// startToken - Reset all flags to cleared.
146  ///
147  void startToken() {
148    Kind = tok::unknown;
149    Flags = 0;
150    PtrData = 0;
151    UintData = 0;
152    Loc = SourceLocation();
153  }
154
155  IdentifierInfo *getIdentifierInfo() const {
156    assert(!isAnnotation() && "Used IdentInfo on annotation token!");
157    if (isLiteral()) return 0;
158    return (IdentifierInfo*) PtrData;
159  }
160  void setIdentifierInfo(IdentifierInfo *II) {
161    PtrData = (void*) II;
162  }
163
164  /// getLiteralData - For a literal token (numeric constant, string, etc), this
165  /// returns a pointer to the start of it in the text buffer if known, null
166  /// otherwise.
167  const char *getLiteralData() const {
168    assert(isLiteral() && "Cannot get literal data of non-literal");
169    return reinterpret_cast<const char*>(PtrData);
170  }
171  void setLiteralData(const char *Ptr) {
172    assert(isLiteral() && "Cannot set literal data of non-literal");
173    PtrData = const_cast<char*>(Ptr);
174  }
175
176  void *getAnnotationValue() const {
177    assert(isAnnotation() && "Used AnnotVal on non-annotation token");
178    return PtrData;
179  }
180  void setAnnotationValue(void *val) {
181    assert(isAnnotation() && "Used AnnotVal on non-annotation token");
182    PtrData = val;
183  }
184
185  /// setFlag - Set the specified flag.
186  void setFlag(TokenFlags Flag) {
187    Flags |= Flag;
188  }
189
190  /// clearFlag - Unset the specified flag.
191  void clearFlag(TokenFlags Flag) {
192    Flags &= ~Flag;
193  }
194
195  /// getFlags - Return the internal represtation of the flags.
196  ///  Only intended for low-level operations such as writing tokens to
197  //   disk.
198  unsigned getFlags() const {
199    return Flags;
200  }
201
202  /// setFlagValue - Set a flag to either true or false.
203  void setFlagValue(TokenFlags Flag, bool Val) {
204    if (Val)
205      setFlag(Flag);
206    else
207      clearFlag(Flag);
208  }
209
210  /// isAtStartOfLine - Return true if this token is at the start of a line.
211  ///
212  bool isAtStartOfLine() const { return (Flags & StartOfLine) ? true : false; }
213
214  /// hasLeadingSpace - Return true if this token has whitespace before it.
215  ///
216  bool hasLeadingSpace() const { return (Flags & LeadingSpace) ? true : false; }
217
218  /// isExpandDisabled - Return true if this identifier token should never
219  /// be expanded in the future, due to C99 6.10.3.4p2.
220  bool isExpandDisabled() const {
221    return (Flags & DisableExpand) ? true : false;
222  }
223
224  /// isObjCAtKeyword - Return true if we have an ObjC keyword identifier.
225  bool isObjCAtKeyword(tok::ObjCKeywordKind objcKey) const;
226
227  /// getObjCKeywordID - Return the ObjC keyword kind.
228  tok::ObjCKeywordKind getObjCKeywordID() const;
229
230  /// needsCleaning - Return true if this token has trigraphs or escaped
231  /// newlines in it.
232  ///
233  bool needsCleaning() const { return (Flags & NeedsCleaning) ? true : false; }
234};
235
236/// PPConditionalInfo - Information about the conditional stack (#if directives)
237/// currently active.
238struct PPConditionalInfo {
239  /// IfLoc - Location where the conditional started.
240  ///
241  SourceLocation IfLoc;
242
243  /// WasSkipping - True if this was contained in a skipping directive, e.g.
244  /// in a "#if 0" block.
245  bool WasSkipping;
246
247  /// FoundNonSkip - True if we have emitted tokens already, and now we're in
248  /// an #else block or something.  Only useful in Skipping blocks.
249  bool FoundNonSkip;
250
251  /// FoundElse - True if we've seen a #else in this block.  If so,
252  /// #elif/#else directives are not allowed.
253  bool FoundElse;
254};
255
256}  // end namespace clang
257
258namespace llvm {
259  template <>
260  struct isPodLike<clang::Token> { static const bool value = true; };
261}  // end namespace llvm
262
263#endif
264