1 //===--- StringRef.h - Constant String Reference Wrapper --------*- C++ -*-===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 #ifndef LLVM_ADT_STRINGREF_H
11 #define LLVM_ADT_STRINGREF_H
13 #include "llvm/Support/type_traits.h"
24 class SmallVectorImpl;
29 /// Helper functions for StringRef::getAsInteger.
30 bool getAsUnsignedInteger(StringRef Str, unsigned Radix,
31 unsigned long long &Result);
33 bool getAsSignedInteger(StringRef Str, unsigned Radix, long long &Result);
35 /// StringRef - Represent a constant reference to a string, i.e. a character
36 /// array and a length, which need not be null terminated.
38 /// This class does not own the string data, it is expected to be used in
39 /// situations where the character data resides in some other buffer, whose
40 /// lifetime extends past that of the StringRef. For this reason, it is not in
41 /// general safe to store a StringRef.
44 typedef const char *iterator;
45 typedef const char *const_iterator;
46 static const size_t npos = ~size_t(0);
47 typedef size_t size_type;
50 /// The start of the string, in an external buffer.
53 /// The length of the string.
56 // Workaround PR5482: nearly all gcc 4.x miscompile StringRef and std::min()
57 // Changing the arg of min to be an integer, instead of a reference to an
58 // integer works around this bug.
59 static size_t min(size_t a, size_t b) { return a < b ? a : b; }
60 static size_t max(size_t a, size_t b) { return a > b ? a : b; }
62 // Workaround memcmp issue with null pointers (undefined behavior)
63 // by providing a specialized version
64 static int compareMemory(const char *Lhs, const char *Rhs, size_t Length) {
65 if (Length == 0) { return 0; }
66 return ::memcmp(Lhs,Rhs,Length);
70 /// @name Constructors
73 /// Construct an empty string ref.
74 /*implicit*/ StringRef() : Data(0), Length(0) {}
76 /// Construct a string ref from a cstring.
77 /*implicit*/ StringRef(const char *Str)
79 assert(Str && "StringRef cannot be built from a NULL argument");
80 Length = ::strlen(Str); // invoking strlen(NULL) is undefined behavior
83 /// Construct a string ref from a pointer and length.
84 /*implicit*/ StringRef(const char *data, size_t length)
85 : Data(data), Length(length) {
86 assert((data || length == 0) &&
87 "StringRef cannot be built from a NULL argument with non-null length");
90 /// Construct a string ref from an std::string.
91 /*implicit*/ StringRef(const std::string &Str)
92 : Data(Str.data()), Length(Str.length()) {}
98 iterator begin() const { return Data; }
100 iterator end() const { return Data + Length; }
103 /// @name String Operations
106 /// data - Get a pointer to the start of the string (which may not be null
108 const char *data() const { return Data; }
110 /// empty - Check if the string is empty.
111 bool empty() const { return Length == 0; }
113 /// size - Get the string size.
114 size_t size() const { return Length; }
116 /// front - Get the first character in the string.
122 /// back - Get the last character in the string.
125 return Data[Length-1];
128 /// equals - Check for string equality, this is more efficient than
129 /// compare() when the relative ordering of inequal strings isn't needed.
130 bool equals(StringRef RHS) const {
131 return (Length == RHS.Length &&
132 compareMemory(Data, RHS.Data, RHS.Length) == 0);
135 /// equals_lower - Check for string equality, ignoring case.
136 bool equals_lower(StringRef RHS) const {
137 return Length == RHS.Length && compare_lower(RHS) == 0;
140 /// compare - Compare two strings; the result is -1, 0, or 1 if this string
141 /// is lexicographically less than, equal to, or greater than the \arg RHS.
142 int compare(StringRef RHS) const {
143 // Check the prefix for a mismatch.
144 if (int Res = compareMemory(Data, RHS.Data, min(Length, RHS.Length)))
145 return Res < 0 ? -1 : 1;
147 // Otherwise the prefixes match, so we only need to check the lengths.
148 if (Length == RHS.Length)
150 return Length < RHS.Length ? -1 : 1;
153 /// compare_lower - Compare two strings, ignoring case.
154 int compare_lower(StringRef RHS) const;
156 /// compare_numeric - Compare two strings, treating sequences of digits as
158 int compare_numeric(StringRef RHS) const;
160 /// \brief Determine the edit distance between this string and another
163 /// \param Other the string to compare this string against.
165 /// \param AllowReplacements whether to allow character
166 /// replacements (change one character into another) as a single
167 /// operation, rather than as two operations (an insertion and a
170 /// \param MaxEditDistance If non-zero, the maximum edit distance that
171 /// this routine is allowed to compute. If the edit distance will exceed
172 /// that maximum, returns \c MaxEditDistance+1.
174 /// \returns the minimum number of character insertions, removals,
175 /// or (if \p AllowReplacements is \c true) replacements needed to
176 /// transform one of the given strings into the other. If zero,
177 /// the strings are identical.
178 unsigned edit_distance(StringRef Other, bool AllowReplacements = true,
179 unsigned MaxEditDistance = 0);
181 /// str - Get the contents as an std::string.
182 std::string str() const {
183 if (Data == 0) return std::string();
184 return std::string(Data, Length);
188 /// @name Operator Overloads
191 char operator[](size_t Index) const {
192 assert(Index < Length && "Invalid index!");
197 /// @name Type Conversions
200 operator std::string() const {
205 /// @name String Predicates
208 /// startswith - Check if this string starts with the given \arg Prefix.
209 bool startswith(StringRef Prefix) const {
210 return Length >= Prefix.Length &&
211 compareMemory(Data, Prefix.Data, Prefix.Length) == 0;
214 /// endswith - Check if this string ends with the given \arg Suffix.
215 bool endswith(StringRef Suffix) const {
216 return Length >= Suffix.Length &&
217 compareMemory(end() - Suffix.Length, Suffix.Data, Suffix.Length) == 0;
221 /// @name String Searching
224 /// find - Search for the first character \arg C in the string.
226 /// \return - The index of the first occurrence of \arg C, or npos if not
228 size_t find(char C, size_t From = 0) const {
229 for (size_t i = min(From, Length), e = Length; i != e; ++i)
235 /// find - Search for the first string \arg Str in the string.
237 /// \return - The index of the first occurrence of \arg Str, or npos if not
239 size_t find(StringRef Str, size_t From = 0) const;
241 /// rfind - Search for the last character \arg C in the string.
243 /// \return - The index of the last occurrence of \arg C, or npos if not
245 size_t rfind(char C, size_t From = npos) const {
246 From = min(From, Length);
256 /// rfind - Search for the last string \arg Str in the string.
258 /// \return - The index of the last occurrence of \arg Str, or npos if not
260 size_t rfind(StringRef Str) const;
262 /// find_first_of - Find the first character in the string that is \arg C,
263 /// or npos if not found. Same as find.
264 size_type find_first_of(char C, size_t From = 0) const {
265 return find(C, From);
268 /// find_first_of - Find the first character in the string that is in \arg
269 /// Chars, or npos if not found.
271 /// Note: O(size() + Chars.size())
272 size_type find_first_of(StringRef Chars, size_t From = 0) const;
274 /// find_first_not_of - Find the first character in the string that is not
275 /// \arg C or npos if not found.
276 size_type find_first_not_of(char C, size_t From = 0) const;
278 /// find_first_not_of - Find the first character in the string that is not
279 /// in the string \arg Chars, or npos if not found.
281 /// Note: O(size() + Chars.size())
282 size_type find_first_not_of(StringRef Chars, size_t From = 0) const;
284 /// find_last_of - Find the last character in the string that is \arg C, or
285 /// npos if not found.
286 size_type find_last_of(char C, size_t From = npos) const {
287 return rfind(C, From);
290 /// find_last_of - Find the last character in the string that is in \arg C,
291 /// or npos if not found.
293 /// Note: O(size() + Chars.size())
294 size_type find_last_of(StringRef Chars, size_t From = npos) const;
296 /// find_last_not_of - Find the last character in the string that is not
297 /// \arg C, or npos if not found.
298 size_type find_last_not_of(char C, size_t From = npos) const;
300 /// find_last_not_of - Find the last character in the string that is not in
301 /// \arg Chars, or npos if not found.
303 /// Note: O(size() + Chars.size())
304 size_type find_last_not_of(StringRef Chars, size_t From = npos) const;
307 /// @name Helpful Algorithms
310 /// count - Return the number of occurrences of \arg C in the string.
311 size_t count(char C) const {
313 for (size_t i = 0, e = Length; i != e; ++i)
319 /// count - Return the number of non-overlapped occurrences of \arg Str in
321 size_t count(StringRef Str) const;
323 /// getAsInteger - Parse the current string as an integer of the specified
324 /// radix. If Radix is specified as zero, this does radix autosensing using
325 /// extended C rules: 0 is octal, 0x is hex, 0b is binary.
327 /// If the string is invalid or if only a subset of the string is valid,
328 /// this returns true to signify the error. The string is considered
329 /// erroneous if empty or if it overflows T.
331 template <typename T>
332 typename enable_if_c<std::numeric_limits<T>::is_signed, bool>::type
333 getAsInteger(unsigned Radix, T &Result) const {
335 if (getAsSignedInteger(*this, Radix, LLVal) ||
336 static_cast<T>(LLVal) != LLVal)
342 template <typename T>
343 typename enable_if_c<!std::numeric_limits<T>::is_signed, bool>::type
344 getAsInteger(unsigned Radix, T &Result) const {
345 unsigned long long ULLVal;
346 if (getAsUnsignedInteger(*this, Radix, ULLVal) ||
347 static_cast<T>(ULLVal) != ULLVal)
353 /// getAsInteger - Parse the current string as an integer of the
354 /// specified radix, or of an autosensed radix if the radix given
355 /// is 0. The current value in Result is discarded, and the
356 /// storage is changed to be wide enough to store the parsed
359 /// Returns true if the string does not solely consist of a valid
360 /// non-empty number in the appropriate base.
362 /// APInt::fromString is superficially similar but assumes the
363 /// string is well-formed in the given radix.
364 bool getAsInteger(unsigned Radix, APInt &Result) const;
367 /// @name String Operations
370 // lower - Convert the given ASCII string to lowercase.
371 std::string lower() const;
373 /// upper - Convert the given ASCII string to uppercase.
374 std::string upper() const;
377 /// @name Substring Operations
380 /// substr - Return a reference to the substring from [Start, Start + N).
382 /// \param Start - The index of the starting character in the substring; if
383 /// the index is npos or greater than the length of the string then the
384 /// empty substring will be returned.
386 /// \param N - The number of characters to included in the substring. If N
387 /// exceeds the number of characters remaining in the string, the string
388 /// suffix (starting with \arg Start) will be returned.
389 StringRef substr(size_t Start, size_t N = npos) const {
390 Start = min(Start, Length);
391 return StringRef(Data + Start, min(N, Length - Start));
394 /// drop_front - Return a StringRef equal to 'this' but with the first
395 /// elements dropped.
396 StringRef drop_front(unsigned N = 1) const {
397 assert(size() >= N && "Dropping more elements than exist");
401 /// drop_back - Return a StringRef equal to 'this' but with the last
402 /// elements dropped.
403 StringRef drop_back(unsigned N = 1) const {
404 assert(size() >= N && "Dropping more elements than exist");
405 return substr(0, size()-N);
408 /// slice - Return a reference to the substring from [Start, End).
410 /// \param Start - The index of the starting character in the substring; if
411 /// the index is npos or greater than the length of the string then the
412 /// empty substring will be returned.
414 /// \param End - The index following the last character to include in the
415 /// substring. If this is npos, or less than \arg Start, or exceeds the
416 /// number of characters remaining in the string, the string suffix
417 /// (starting with \arg Start) will be returned.
418 StringRef slice(size_t Start, size_t End) const {
419 Start = min(Start, Length);
420 End = min(max(Start, End), Length);
421 return StringRef(Data + Start, End - Start);
424 /// split - Split into two substrings around the first occurrence of a
425 /// separator character.
427 /// If \arg Separator is in the string, then the result is a pair (LHS, RHS)
428 /// such that (*this == LHS + Separator + RHS) is true and RHS is
429 /// maximal. If \arg Separator is not in the string, then the result is a
430 /// pair (LHS, RHS) where (*this == LHS) and (RHS == "").
432 /// \param Separator - The character to split on.
433 /// \return - The split substrings.
434 std::pair<StringRef, StringRef> split(char Separator) const {
435 size_t Idx = find(Separator);
437 return std::make_pair(*this, StringRef());
438 return std::make_pair(slice(0, Idx), slice(Idx+1, npos));
441 /// split - Split into two substrings around the first occurrence of a
442 /// separator string.
444 /// If \arg Separator is in the string, then the result is a pair (LHS, RHS)
445 /// such that (*this == LHS + Separator + RHS) is true and RHS is
446 /// maximal. If \arg Separator is not in the string, then the result is a
447 /// pair (LHS, RHS) where (*this == LHS) and (RHS == "").
449 /// \param Separator - The string to split on.
450 /// \return - The split substrings.
451 std::pair<StringRef, StringRef> split(StringRef Separator) const {
452 size_t Idx = find(Separator);
454 return std::make_pair(*this, StringRef());
455 return std::make_pair(slice(0, Idx), slice(Idx + Separator.size(), npos));
458 /// split - Split into substrings around the occurrences of a separator
461 /// Each substring is stored in \arg A. If \arg MaxSplit is >= 0, at most
462 /// \arg MaxSplit splits are done and consequently <= \arg MaxSplit
463 /// elements are added to A.
464 /// If \arg KeepEmpty is false, empty strings are not added to \arg A. They
465 /// still count when considering \arg MaxSplit
466 /// An useful invariant is that
467 /// Separator.join(A) == *this if MaxSplit == -1 and KeepEmpty == true
469 /// \param A - Where to put the substrings.
470 /// \param Separator - The string to split on.
471 /// \param MaxSplit - The maximum number of times the string is split.
472 /// \param KeepEmpty - True if empty substring should be added.
473 void split(SmallVectorImpl<StringRef> &A,
474 StringRef Separator, int MaxSplit = -1,
475 bool KeepEmpty = true) const;
477 /// rsplit - Split into two substrings around the last occurrence of a
478 /// separator character.
480 /// If \arg Separator is in the string, then the result is a pair (LHS, RHS)
481 /// such that (*this == LHS + Separator + RHS) is true and RHS is
482 /// minimal. If \arg Separator is not in the string, then the result is a
483 /// pair (LHS, RHS) where (*this == LHS) and (RHS == "").
485 /// \param Separator - The character to split on.
486 /// \return - The split substrings.
487 std::pair<StringRef, StringRef> rsplit(char Separator) const {
488 size_t Idx = rfind(Separator);
490 return std::make_pair(*this, StringRef());
491 return std::make_pair(slice(0, Idx), slice(Idx+1, npos));
494 /// ltrim - Return string with consecutive characters in \arg Chars starting
495 /// from the left removed.
496 StringRef ltrim(StringRef Chars = " \t\n\v\f\r") const {
497 return drop_front(std::min(Length, find_first_not_of(Chars)));
500 /// rtrim - Return string with consecutive characters in \arg Chars starting
501 /// from the right removed.
502 StringRef rtrim(StringRef Chars = " \t\n\v\f\r") const {
503 return drop_back(Length - std::min(Length, find_last_not_of(Chars) + 1));
506 /// trim - Return string with consecutive characters in \arg Chars starting
507 /// from the left and right removed.
508 StringRef trim(StringRef Chars = " \t\n\v\f\r") const {
509 return ltrim(Chars).rtrim(Chars);
515 /// @name StringRef Comparison Operators
518 inline bool operator==(StringRef LHS, StringRef RHS) {
519 return LHS.equals(RHS);
522 inline bool operator!=(StringRef LHS, StringRef RHS) {
523 return !(LHS == RHS);
526 inline bool operator<(StringRef LHS, StringRef RHS) {
527 return LHS.compare(RHS) == -1;
530 inline bool operator<=(StringRef LHS, StringRef RHS) {
531 return LHS.compare(RHS) != 1;
534 inline bool operator>(StringRef LHS, StringRef RHS) {
535 return LHS.compare(RHS) == 1;
538 inline bool operator>=(StringRef LHS, StringRef RHS) {
539 return LHS.compare(RHS) != -1;
542 inline std::string &operator+=(std::string &buffer, llvm::StringRef string) {
543 return buffer.append(string.data(), string.size());
548 /// \brief Compute a hash_code for a StringRef.
549 hash_code hash_value(StringRef S);
551 // StringRefs can be treated like a POD type.
552 template <typename T> struct isPodLike;
553 template <> struct isPodLike<StringRef> { static const bool value = true; };