include/llvm/ADT/StringRef.h

   1 //===--- StringRef.h - Constant String Reference Wrapper --------*- 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 #ifndef LLVM_ADT_STRINGREF_H
  11 #define LLVM_ADT_STRINGREF_H
  12
  13 #include <algorithm>
  14 #include <cassert>
  15 #include <cstring>
  16 #include <string>
  17
  18 namespace llvm {
  19
  20   /// StringRef - Represent a constant reference to a string, i.e. a character
  21   /// array and a length, which need not be null terminated.
  22   ///
  23   /// This class does not own the string data, it is expected to be used in
  24   /// situations where the character data resides in some other buffer, whose
  25   /// lifetime extends past that of the StringRef. For this reason, it is not in
  26   /// general safe to store a StringRef.
  27   class StringRef {
  28   public:
  29     typedef const char *iterator;
  30     static const size_t npos = ~size_t(0);
  31
  32   private:
  33     /// The start of the string, in an external buffer.
  34     const char *Data;
  35
  36     /// The length of the string.
  37     size_t Length;
  38
  39   public:
  40     /// @name Constructors
  41     /// @{
  42
  43     /// Construct an empty string ref.
  44     /*implicit*/ StringRef() : Data(0), Length(0) {}
  45
  46     /// Construct a string ref from a cstring.
  47     /*implicit*/ StringRef(const char *Str)
  48       : Data(Str), Length(::strlen(Str)) {}
  49
  50     /// Construct a string ref from a pointer and length.
  51     /*implicit*/ StringRef(const char *_Data, unsigned _Length)
  52       : Data(_Data), Length(_Length) {}
  53
  54     /// Construct a string ref from an std::string.
  55     /*implicit*/ StringRef(const std::string &Str)
  56       : Data(Str.c_str()), Length(Str.length()) {}
  57
  58     /// @}
  59     /// @name Iterators
  60     /// @{
  61
  62     iterator begin() const { return Data; }
  63
  64     iterator end() const { return Data + Length; }
  65
  66     /// @}
  67     /// @name String Operations
  68     /// @{
  69
  70     /// data - Get a pointer to the start of the string (which may not be null
  71     /// terminated).
  72     const char *data() const { return Data; }
  73
  74     /// empty - Check if the string is empty.
  75     bool empty() const { return Length == 0; }
  76
  77     /// size - Get the string size.
  78     size_t size() const { return Length; }
  79
  80     /// front - Get the first character in the string.
  81     char front() const {
  82       assert(!empty());
  83       return Data[0];
  84     }
  85
  86     /// back - Get the last character in the string.
  87     char back() const {
  88       assert(!empty());
  89       return Data[Length-1];
  90     }
  91
  92     /// equals - Check for string equality, this is more efficient than
  93     /// compare() when the relative ordering of inequal strings isn't needed.
  94     bool equals(const StringRef &RHS) const {
  95       return (Length == RHS.Length &&
  96               memcmp(Data, RHS.Data, RHS.Length) == 0);
  97     }
  98
  99     /// compare - Compare two strings; the result is -1, 0, or 1 if this string
 100     /// is lexicographically less than, equal to, or greater than the \arg RHS.
 101     int compare(const StringRef &RHS) const {
 102       // Check the prefix for a mismatch.
 103       if (int Res = memcmp(Data, RHS.Data, std::min(Length, RHS.Length)))
 104         return Res < 0 ? -1 : 1;
 105
 106       // Otherwise the prefixes match, so we only need to check the lengths.
 107       if (Length == RHS.Length)
 108         return 0;
 109       return Length < RHS.Length ? -1 : 1;
 110     }
 111
 112     /// str - Get the contents as an std::string.
 113     std::string str() const { return std::string(Data, Length); }
 114
 115     /// @}
 116     /// @name Operator Overloads
 117     /// @{
 118
 119     char operator[](size_t Index) const {
 120       assert(Index < Length && "Invalid index!");
 121       return Data[Index];
 122     }
 123
 124     /// @}
 125     /// @name Type Conversions
 126     /// @{
 127
 128     operator std::string() const {
 129       return str();
 130     }
 131
 132     /// @}
 133     /// @name String Predicates
 134     /// @{
 135
 136     /// startswith - Check if this string starts with the given \arg Prefix.
 137     bool startswith(const StringRef &Prefix) const {
 138       return substr(0, Prefix.Length).equals(Prefix);
 139     }
 140
 141     /// endswith - Check if this string ends with the given \arg Suffix.
 142     bool endswith(const StringRef &Suffix) const {
 143       return slice(size() - Suffix.Length, size()).equals(Suffix);
 144     }
 145
 146     /// @}
 147     /// @name String Searching
 148     /// @{
 149
 150     /// find - Search for the first character \arg C in the string.
 151     ///
 152     /// \return - The index of the first occurence of \arg C, or npos if not
 153     /// found.
 154     size_t find(char C) const {
 155       for (size_t i = 0, e = Length; i != e; ++i)
 156         if (Data[i] == C)
 157           return i;
 158       return npos;
 159     }
 160
 161     /// find - Search for the first string \arg Str in the string.
 162     ///
 163     /// \return - The index of the first occurence of \arg Str, or npos if not
 164     /// found.
 165     size_t find(const StringRef &Str) const {
 166       size_t N = Str.size();
 167       if (N > Length)
 168         return npos;
 169       for (size_t i = 0, e = Length - N + 1; i != e; ++i)
 170         if (substr(i, N).equals(Str))
 171           return i;
 172       return npos;
 173     }
 174
 175     /// rfind - Search for the last character \arg C in the string.
 176     ///
 177     /// \return - The index of the last occurence of \arg C, or npos if not
 178     /// found.
 179     size_t rfind(char C) const {
 180       for (size_t i = Length, e = 0; i != e;) {
 181         --i;
 182         if (Data[i] == C)
 183           return i;
 184       }
 185       return npos;
 186     }
 187
 188     /// rfind - Search for the last string \arg Str in the string.
 189     ///
 190     /// \return - The index of the last occurence of \arg Str, or npos if not
 191     /// found.
 192     size_t rfind(const StringRef &Str) const {
 193       size_t N = Str.size();
 194       if (N > Length)
 195         return npos;
 196       for (size_t i = Length - N + 1, e = 0; i != e;) {
 197         --i;
 198         if (substr(i, N).equals(Str))
 199           return i;
 200       }
 201       return npos;
 202     }
 203
 204     /// count - Return the number of occurrences of \arg C in the string.
 205     size_t count(char C) const {
 206       size_t Count = 0;
 207       for (size_t i = 0, e = Length; i != e; ++i)
 208         if (Data[i] == C)
 209           ++Count;
 210       return Count;
 211     }
 212
 213     /// count - Return the number of non-overlapped occurrences of \arg Str in
 214     /// the string.
 215     size_t count(const StringRef &Str) const {
 216       size_t Count = 0;
 217       size_t N = Str.size();
 218       if (N > Length)
 219         return 0;
 220       for (size_t i = 0, e = Length - N + 1; i != e; ++i)
 221         if (substr(i, N).equals(Str))
 222           ++Count;
 223       return Count;
 224     }
 225
 226     /// @}
 227     /// @name Substring Operations
 228     /// @{
 229
 230     /// substr - Return a reference to the substring from [Start, Start + N).
 231     ///
 232     /// \param Start - The index of the starting character in the substring; if
 233     /// the index is npos or greater than the length of the string then the
 234     /// empty substring will be returned.
 235     ///
 236     /// \param N - The number of characters to included in the substring. If N
 237     /// exceeds the number of characters remaining in the string, the string
 238     /// suffix (starting with \arg Start) will be returned.
 239     StringRef substr(size_t Start, size_t N = npos) const {
 240       Start = std::min(Start, Length);
 241       return StringRef(Data + Start, std::min(N, Length - Start));
 242     }
 243
 244     /// slice - Return a reference to the substring from [Start, End).
 245     ///
 246     /// \param Start - The index of the starting character in the substring; if
 247     /// the index is npos or greater than the length of the string then the
 248     /// empty substring will be returned.
 249     ///
 250     /// \param End - The index following the last character to include in the
 251     /// substring. If this is npos, or less than \arg Start, or exceeds the
 252     /// number of characters remaining in the string, the string suffix
 253     /// (starting with \arg Start) will be returned.
 254     StringRef slice(size_t Start, size_t End) const {
 255       Start = std::min(Start, Length);
 256       End = std::min(std::max(Start, End), Length);
 257       return StringRef(Data + Start, End - Start);
 258     }
 259
 260     /// split - Split into two substrings around the first occurence of a
 261     /// separator character.
 262     ///
 263     /// If \arg Separator is in the string, then the result is a pair (LHS, RHS)
 264     /// such that (*this == LHS + Separator + RHS) is true and RHS is
 265     /// maximal. If \arg Separator is not in the string, then the result is a
 266     /// pair (LHS, RHS) where (*this == LHS) and (RHS == "").
 267     ///
 268     /// \param Separator - The character to split on.
 269     /// \return - The split substrings.
 270     std::pair<StringRef, StringRef> split(char Separator) const {
 271       size_t Idx = find(Separator);
 272       if (Idx == npos)
 273         return std::make_pair(*this, StringRef());
 274       return std::make_pair(slice(0, Idx), slice(Idx+1, npos));
 275     }
 276
 277     /// rsplit - Split into two substrings around the last occurence of a
 278     /// separator character.
 279     ///
 280     /// If \arg Separator is in the string, then the result is a pair (LHS, RHS)
 281     /// such that (*this == LHS + Separator + RHS) is true and RHS is
 282     /// minimal. If \arg Separator is not in the string, then the result is a
 283     /// pair (LHS, RHS) where (*this == LHS) and (RHS == "").
 284     ///
 285     /// \param Separator - The character to split on.
 286     /// \return - The split substrings.
 287     std::pair<StringRef, StringRef> rsplit(char Separator) const {
 288       size_t Idx = rfind(Separator);
 289       if (Idx == npos)
 290         return std::make_pair(*this, StringRef());
 291       return std::make_pair(slice(0, Idx), slice(Idx+1, npos));
 292     }
 293
 294     /// @}
 295   };
 296
 297   /// @name StringRef Comparison Operators
 298   /// @{
 299
 300   inline bool operator==(const StringRef &LHS, const StringRef &RHS) {
 301     return LHS.equals(RHS);
 302   }
 303
 304   inline bool operator!=(const StringRef &LHS, const StringRef &RHS) {
 305     return !(LHS == RHS);
 306   }
 307
 308   inline bool operator<(const StringRef &LHS, const StringRef &RHS) {
 309     return LHS.compare(RHS) == -1;
 310   }
 311
 312   inline bool operator<=(const StringRef &LHS, const StringRef &RHS) {
 313     return LHS.compare(RHS) != 1;
 314   }
 315
 316   inline bool operator>(const StringRef &LHS, const StringRef &RHS) {
 317     return LHS.compare(RHS) == 1;
 318   }
 319
 320   inline bool operator>=(const StringRef &LHS, const StringRef &RHS) {
 321     return LHS.compare(RHS) != -1;
 322   }
 323
 324   /// @}
 325
 326 }
 327
 328 #endif