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/SourceLocation.h"
18#include "clang/Basic/TokenKinds.h"
19#include "llvm/ADT/StringRef.h"
20#include <cassert>
21
22namespace clang {
23
24class IdentifierInfo;
25
26/// Token - This structure provides full information about a lexed token.
27/// It is not intended to be space efficient, it is intended to return as much
28/// information as possible about each returned token.  This is expected to be
29/// compressed into a smaller form if memory footprint is important.
30///
31/// The parser can create a special "annotation token" representing a stream of
32/// tokens that were parsed and semantically resolved, e.g.: "foo::MyClass<int>"
33/// can be represented by a single typename annotation token that carries
34/// information about the SourceRange of the tokens and the type object.
35class Token {
36  /// The location of the token. This is actually a SourceLocation.
37  unsigned Loc;
38
39  // Conceptually these next two fields could be in a union.  However, this
40  // causes gcc 4.2 to pessimize LexTokenInternal, a very performance critical
41  // routine. Keeping as separate members with casts until a more beautiful fix
42  // presents itself.
43
44  /// UintData - This holds either the length of the token text, when
45  /// a normal token, or the end of the SourceRange when an annotation
46  /// token.
47  unsigned UintData;
48
49  /// PtrData - This is a union of four different pointer types, which depends
50  /// on what type of token this is:
51  ///  Identifiers, keywords, etc:
52  ///    This is an IdentifierInfo*, which contains the uniqued identifier
53  ///    spelling.
54  ///  Literals:  isLiteral() returns true.
55  ///    This is a pointer to the start of the token in a text buffer, which
56  ///    may be dirty (have trigraphs / escaped newlines).
57  ///  Annotations (resolved type names, C++ scopes, etc): isAnnotation().
58  ///    This is a pointer to sema-specific data for the annotation token.
59  ///  Eof:
60  //     This is a pointer to a Decl.
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 short Flags;
70
71public:
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    StringifiedInMacro = 0x100, // This string or character literal is formed by
85                                // macro stringizing or charizing operator.
86    CommaAfterElided = 0x200, // The comma following this token was elided (MS).
87    IsEditorPlaceholder = 0x400, // This identifier is a placeholder.
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 Get the specified flag.
239  bool getFlag(TokenFlags Flag) const {
240    return (Flags & Flag) != 0;
241  }
242
243  /// \brief Unset the specified flag.
244  void clearFlag(TokenFlags Flag) {
245    Flags &= ~Flag;
246  }
247
248  /// \brief Return the internal represtation of the flags.
249  ///
250  /// This is only intended for low-level operations such as writing tokens to
251  /// disk.
252  unsigned getFlags() const {
253    return Flags;
254  }
255
256  /// \brief Set a flag to either true or false.
257  void setFlagValue(TokenFlags Flag, bool Val) {
258    if (Val)
259      setFlag(Flag);
260    else
261      clearFlag(Flag);
262  }
263
264  /// isAtStartOfLine - Return true if this token is at the start of a line.
265  ///
266  bool isAtStartOfLine() const { return getFlag(StartOfLine); }
267
268  /// \brief Return true if this token has whitespace before it.
269  ///
270  bool hasLeadingSpace() const { return getFlag(LeadingSpace); }
271
272  /// \brief Return true if this identifier token should never
273  /// be expanded in the future, due to C99 6.10.3.4p2.
274  bool isExpandDisabled() const { return getFlag(DisableExpand); }
275
276  /// \brief Return true if we have an ObjC keyword identifier.
277  bool isObjCAtKeyword(tok::ObjCKeywordKind objcKey) const;
278
279  /// \brief Return the ObjC keyword kind.
280  tok::ObjCKeywordKind getObjCKeywordID() const;
281
282  /// \brief Return true if this token has trigraphs or escaped newlines in it.
283  bool needsCleaning() const { return getFlag(NeedsCleaning); }
284
285  /// \brief Return true if this token has an empty macro before it.
286  ///
287  bool hasLeadingEmptyMacro() const { return getFlag(LeadingEmptyMacro); }
288
289  /// \brief Return true if this token is a string or character literal which
290  /// has a ud-suffix.
291  bool hasUDSuffix() const { return getFlag(HasUDSuffix); }
292
293  /// Returns true if this token contains a universal character name.
294  bool hasUCN() const { return getFlag(HasUCN); }
295
296  /// Returns true if this token is formed by macro by stringizing or charizing
297  /// operator.
298  bool stringifiedInMacro() const { return getFlag(StringifiedInMacro); }
299
300  /// Returns true if the comma after this token was elided.
301  bool commaAfterElided() const { return getFlag(CommaAfterElided); }
302
303  /// Returns true if this token is an editor placeholder.
304  ///
305  /// Editor placeholders are produced by the code-completion engine and are
306  /// represented as characters between '<#' and '#>' in the source code. The
307  /// lexer uses identifier tokens to represent placeholders.
308  bool isEditorPlaceholder() const { return getFlag(IsEditorPlaceholder); }
309};
310
311/// \brief Information about the conditional stack (\#if directives)
312/// currently active.
313struct PPConditionalInfo {
314  /// \brief Location where the conditional started.
315  SourceLocation IfLoc;
316
317  /// \brief True if this was contained in a skipping directive, e.g.,
318  /// in a "\#if 0" block.
319  bool WasSkipping;
320
321  /// \brief True if we have emitted tokens already, and now we're in
322  /// an \#else block or something.  Only useful in Skipping blocks.
323  bool FoundNonSkip;
324
325  /// \brief True if we've seen a \#else in this block.  If so,
326  /// \#elif/\#else directives are not allowed.
327  bool FoundElse;
328};
329
330} // end namespace clang
331
332namespace llvm {
333  template <>
334  struct isPodLike<clang::Token> { static const bool value = true; };
335} // end namespace llvm
336
337#endif // LLVM_CLANG_LEX_TOKEN_H
338