StringRef.h revision c78c0c99a0fe1703ae72fc51e440aaa8e4e19e91 
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
18namespace llvm {
19  template<typename T>
20  class SmallVectorImpl;
21
22  /// StringRef - Represent a constant reference to a string, i.e. a character
23  /// array and a length, which need not be null terminated.
24  ///
25  /// This class does not own the string data, it is expected to be used in
26  /// situations where the character data resides in some other buffer, whose
27  /// lifetime extends past that of the StringRef. For this reason, it is not in
28  /// general safe to store a StringRef.
29  class StringRef {
30  public:
31    typedef const char *iterator;
32    static const size_t npos = ~size_t(0);
33    typedef size_t size_type;
34
35  private:
36    /// The start of the string, in an external buffer.
37    const char *Data;
38
39    /// The length of the string.
40    size_t Length;
41
42  public:
43    /// @name Constructors
44    /// @{
45
46    /// Construct an empty string ref.
47    /*implicit*/ StringRef() : Data(0), Length(0) {}
48
49    /// Construct a string ref from a cstring.
50    /*implicit*/ StringRef(const char *Str)
51      : Data(Str), Length(::strlen(Str)) {}
52
53    /// Construct a string ref from a pointer and length.
54    /*implicit*/ StringRef(const char *data, size_t length)
55      : Data(data), Length(length) {}
56
57    /// Construct a string ref from an std::string.
58    /*implicit*/ StringRef(const std::string &Str)
59      : Data(Str.c_str()), Length(Str.length()) {}
60
61    /// @}
62    /// @name Iterators
63    /// @{
64
65    iterator begin() const { return Data; }
66
67    iterator end() const { return Data + Length; }
68
69    /// @}
70    /// @name String Operations
71    /// @{
72
73    /// data - Get a pointer to the start of the string (which may not be null
74    /// terminated).
75    const char *data() const { return Data; }
76
77    /// empty - Check if the string is empty.
78    bool empty() const { return Length == 0; }
79
80    /// size - Get the string size.
81    size_t size() const { return Length; }
82
83    /// front - Get the first character in the string.
84    char front() const {
85      assert(!empty());
86      return Data[0];
87    }
88
89    /// back - Get the last character in the string.
90    char back() const {
91      assert(!empty());
92      return Data[Length-1];
93    }
94
95    /// equals - Check for string equality, this is more efficient than
96    /// compare() when the relative ordering of inequal strings isn't needed.
97    bool equals(StringRef RHS) const {
98      return (Length == RHS.Length &&
99              memcmp(Data, RHS.Data, RHS.Length) == 0);
100    }
101
102    /// equals_lower - Check for string equality, ignoring case.
103    bool equals_lower(StringRef RHS) const {
104      return Length == RHS.Length && compare_lower(RHS) == 0;
105    }
106
107    /// compare - Compare two strings; the result is -1, 0, or 1 if this string
108    /// is lexicographically less than, equal to, or greater than the \arg RHS.
109    int compare(StringRef RHS) const {
110      // Check the prefix for a mismatch.
111      if (int Res = memcmp(Data, RHS.Data, std::min(Length, RHS.Length)))
112        return Res < 0 ? -1 : 1;
113
114      // Otherwise the prefixes match, so we only need to check the lengths.
115      if (Length == RHS.Length)
116        return 0;
117      return Length < RHS.Length ? -1 : 1;
118    }
119
120    /// compare_lower - Compare two strings, ignoring case.
121    int compare_lower(StringRef RHS) const;
122
123    /// str - Get the contents as an std::string.
124    std::string str() const { return std::string(Data, Length); }
125
126    /// @}
127    /// @name Operator Overloads
128    /// @{
129
130    char operator[](size_t Index) const {
131      assert(Index < Length && "Invalid index!");
132      return Data[Index];
133    }
134
135    /// @}
136    /// @name Type Conversions
137    /// @{
138
139    operator std::string() const {
140      return str();
141    }
142
143    /// @}
144    /// @name String Predicates
145    /// @{
146
147    /// startswith - Check if this string starts with the given \arg Prefix.
148    bool startswith(StringRef Prefix) const {
149      return substr(0, Prefix.Length).equals(Prefix);
150    }
151
152    /// endswith - Check if this string ends with the given \arg Suffix.
153    bool endswith(StringRef Suffix) const {
154      return slice(size() - Suffix.Length, size()).equals(Suffix);
155    }
156
157    /// @}
158    /// @name String Searching
159    /// @{
160
161    /// find - Search for the first character \arg C in the string.
162    ///
163    /// \return - The index of the first occurence of \arg C, or npos if not
164    /// found.
165    size_t find(char C, size_t From = 0) const {
166      for (size_t i = std::min(From, Length), e = Length; i != e; ++i)
167        if (Data[i] == C)
168          return i;
169      return npos;
170    }
171
172    /// find - Search for the first string \arg Str in the string.
173    ///
174    /// \return - The index of the first occurence of \arg Str, or npos if not
175    /// found.
176    size_t find(StringRef Str, size_t From = 0) const;
177
178    /// rfind - Search for the last character \arg C in the string.
179    ///
180    /// \return - The index of the last occurence of \arg C, or npos if not
181    /// found.
182    size_t rfind(char C, size_t From = npos) const {
183      From = std::min(From, Length);
184      size_t i = From;
185      while (i != 0) {
186        --i;
187        if (Data[i] == C)
188          return i;
189      }
190      return npos;
191    }
192
193    /// rfind - Search for the last string \arg Str in the string.
194    ///
195    /// \return - The index of the last occurence of \arg Str, or npos if not
196    /// found.
197    size_t rfind(StringRef Str) const;
198
199    /// find_first_of - Find the first character in the string that is \arg C,
200    /// or npos if not found. Same as find.
201    size_type find_first_of(char C, size_t From = 0) const { return find(C); }
202
203    /// find_first_of - Find the first character in the string that is in \arg
204    /// Chars, or npos if not found.
205    ///
206    /// Note: O(size() * Chars.size())
207    size_type find_first_of(StringRef Chars, size_t From = 0) const;
208
209    /// find_first_not_of - Find the first character in the string that is not
210    /// \arg C or npos if not found.
211    size_type find_first_not_of(char C, size_t From = 0) const;
212
213    /// find_first_not_of - Find the first character in the string that is not
214    /// in the string \arg Chars, or npos if not found.
215    ///
216    /// Note: O(size() * Chars.size())
217    size_type find_first_not_of(StringRef Chars, size_t From = 0) const;
218
219    /// @}
220    /// @name Helpful Algorithms
221    /// @{
222
223    /// count - Return the number of occurrences of \arg C in the string.
224    size_t count(char C) const {
225      size_t Count = 0;
226      for (size_t i = 0, e = Length; i != e; ++i)
227        if (Data[i] == C)
228          ++Count;
229      return Count;
230    }
231
232    /// count - Return the number of non-overlapped occurrences of \arg Str in
233    /// the string.
234    size_t count(StringRef Str) const;
235
236    /// getAsInteger - Parse the current string as an integer of the specified
237    /// radix.  If Radix is specified as zero, this does radix autosensing using
238    /// extended C rules: 0 is octal, 0x is hex, 0b is binary.
239    ///
240    /// If the string is invalid or if only a subset of the string is valid,
241    /// this returns true to signify the error.  The string is considered
242    /// erroneous if empty.
243    ///
244    bool getAsInteger(unsigned Radix, long long &Result) const;
245    bool getAsInteger(unsigned Radix, unsigned long long &Result) const;
246    bool getAsInteger(unsigned Radix, int &Result) const;
247    bool getAsInteger(unsigned Radix, unsigned &Result) const;
248
249    // TODO: Provide overloads for int/unsigned that check for overflow.
250
251    /// @}
252    /// @name Substring Operations
253    /// @{
254
255    /// substr - Return a reference to the substring from [Start, Start + N).
256    ///
257    /// \param Start - The index of the starting character in the substring; if
258    /// the index is npos or greater than the length of the string then the
259    /// empty substring will be returned.
260    ///
261    /// \param N - The number of characters to included in the substring. If N
262    /// exceeds the number of characters remaining in the string, the string
263    /// suffix (starting with \arg Start) will be returned.
264    StringRef substr(size_t Start, size_t N = npos) const {
265      Start = std::min(Start, Length);
266      return StringRef(Data + Start, std::min(N, Length - Start));
267    }
268
269    /// slice - Return a reference to the substring from [Start, End).
270    ///
271    /// \param Start - The index of the starting character in the substring; if
272    /// the index is npos or greater than the length of the string then the
273    /// empty substring will be returned.
274    ///
275    /// \param End - The index following the last character to include in the
276    /// substring. If this is npos, or less than \arg Start, or exceeds the
277    /// number of characters remaining in the string, the string suffix
278    /// (starting with \arg Start) will be returned.
279    StringRef slice(size_t Start, size_t End) const {
280      Start = std::min(Start, Length);
281      End = std::min(std::max(Start, End), Length);
282      return StringRef(Data + Start, End - Start);
283    }
284
285    /// split - Split into two substrings around the first occurence of a
286    /// separator character.
287    ///
288    /// If \arg Separator is in the string, then the result is a pair (LHS, RHS)
289    /// such that (*this == LHS + Separator + RHS) is true and RHS is
290    /// maximal. If \arg Separator is not in the string, then the result is a
291    /// pair (LHS, RHS) where (*this == LHS) and (RHS == "").
292    ///
293    /// \param Separator - The character to split on.
294    /// \return - The split substrings.
295    std::pair<StringRef, StringRef> split(char Separator) const {
296      size_t Idx = find(Separator);
297      if (Idx == npos)
298        return std::make_pair(*this, StringRef());
299      return std::make_pair(slice(0, Idx), slice(Idx+1, npos));
300    }
301
302    /// split - Split into two substrings around the first occurence of a
303    /// separator string.
304    ///
305    /// If \arg Separator is in the string, then the result is a pair (LHS, RHS)
306    /// such that (*this == LHS + Separator + RHS) is true and RHS is
307    /// maximal. If \arg Separator is not in the string, then the result is a
308    /// pair (LHS, RHS) where (*this == LHS) and (RHS == "").
309    ///
310    /// \param Separator - The string to split on.
311    /// \return - The split substrings.
312    std::pair<StringRef, StringRef> split(StringRef Separator) const {
313      size_t Idx = find(Separator);
314      if (Idx == npos)
315        return std::make_pair(*this, StringRef());
316      return std::make_pair(slice(0, Idx), slice(Idx + Separator.size(), npos));
317    }
318
319    /// split - Split into substrings around the occurences of a separator
320    /// string.
321    ///
322    /// Each substring is stored in \arg A. If \arg MaxSplit is >= 0, at most
323    /// \arg MaxSplit splits are done and consequently <= \arg MaxSplit
324    /// elements are added to A.
325    /// If \arg KeepEmpty is false, empty strings are not added to \arg A. They
326    /// still count when considering \arg MaxSplit
327    /// An useful invariant is that
328    /// Separator.join(A) == *this if MaxSplit == -1 and KeepEmpty == true
329    ///
330    /// \param A - Where to put the substrings.
331    /// \param Separator - The string to split on.
332    /// \param MaxSplit - The maximum number of times the string is split.
333    /// \parm KeepEmpty - True if empty substring should be added.
334    void split(SmallVectorImpl<StringRef> &A,
335               StringRef Separator, int MaxSplit = -1,
336               bool KeepEmpty = true) const;
337
338    /// rsplit - Split into two substrings around the last occurence of a
339    /// separator character.
340    ///
341    /// If \arg Separator is in the string, then the result is a pair (LHS, RHS)
342    /// such that (*this == LHS + Separator + RHS) is true and RHS is
343    /// minimal. If \arg Separator is not in the string, then the result is a
344    /// pair (LHS, RHS) where (*this == LHS) and (RHS == "").
345    ///
346    /// \param Separator - The character to split on.
347    /// \return - The split substrings.
348    std::pair<StringRef, StringRef> rsplit(char Separator) const {
349      size_t Idx = rfind(Separator);
350      if (Idx == npos)
351        return std::make_pair(*this, StringRef());
352      return std::make_pair(slice(0, Idx), slice(Idx+1, npos));
353    }
354
355    /// @}
356  };
357
358  /// @name StringRef Comparison Operators
359  /// @{
360
361  inline bool operator==(StringRef LHS, StringRef RHS) {
362    return LHS.equals(RHS);
363  }
364
365  inline bool operator!=(StringRef LHS, StringRef RHS) {
366    return !(LHS == RHS);
367  }
368
369  inline bool operator<(StringRef LHS, StringRef RHS) {
370    return LHS.compare(RHS) == -1;
371  }
372
373  inline bool operator<=(StringRef LHS, StringRef RHS) {
374    return LHS.compare(RHS) != 1;
375  }
376
377  inline bool operator>(StringRef LHS, StringRef RHS) {
378    return LHS.compare(RHS) == 1;
379  }
380
381  inline bool operator>=(StringRef LHS, StringRef RHS) {
382    return LHS.compare(RHS) != -1;
383  }
384
385  /// @}
386
387}
388
389#endif
390