15821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// Copyright 2001-2010 The RE2 Authors.  All Rights Reserved.
25821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// Use of this source code is governed by a BSD-style
35821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// license that can be found in the LICENSE file.
45821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
55821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// A string-like object that points to a sized piece of memory.
65821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)//
75821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// Functions or methods may use const StringPiece& parameters to accept either
85821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// a "const char*" or a "string" value that will be implicitly converted to
95821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// a StringPiece.  The implicit conversion means that it is often appropriate
105821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// to include this .h file in other files rather than forward-declaring
115821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// StringPiece as would be appropriate for most other Google classes.
125821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)//
135821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// Systematic usage of StringPiece is encouraged as it will reduce unnecessary
145821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// conversions from "const char*" to "string" and back again.
155821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)//
165821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)//
175821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// Arghh!  I wish C++ literals were "string".
185821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
195821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#ifndef STRINGS_STRINGPIECE_H__
205821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#define STRINGS_STRINGPIECE_H__
215821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
225821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#include <string.h>
235821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#include <cstddef>
245821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#include <iosfwd>
255821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#include <string>
265821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#ifdef WIN32
275821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#include <algorithm>
285821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#endif
295821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
305821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)namespace re2 {
315821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
325821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)class StringPiece {
335821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles) private:
345821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  const char*   ptr_;
355821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int           length_;
365821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
375821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles) public:
385821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // We provide non-explicit singleton constructors so users can pass
395821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // in a "const char*" or a "string" wherever a "StringPiece" is
405821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // expected.
415821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  StringPiece() : ptr_(NULL), length_(0) { }
425821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  StringPiece(const char* str)
435821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    : ptr_(str), length_((str == NULL) ? 0 : static_cast<int>(strlen(str))) { }
445821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  StringPiece(const std::string& str)
455821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    : ptr_(str.data()), length_(static_cast<int>(str.size())) { }
465821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  StringPiece(const char* offset, int len) : ptr_(offset), length_(len) { }
475821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
485821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // data() may return a pointer to a buffer with embedded NULs, and the
495821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // returned buffer may or may not be null terminated.  Therefore it is
505821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // typically a mistake to pass data() to a routine that expects a NUL
515821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // terminated string.
525821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  const char* data() const { return ptr_; }
535821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int size() const { return length_; }
545821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int length() const { return length_; }
555821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  bool empty() const { return length_ == 0; }
565821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
575821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  void clear() { ptr_ = NULL; length_ = 0; }
585821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  void set(const char* data, int len) { ptr_ = data; length_ = len; }
595821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  void set(const char* str) {
605821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    ptr_ = str;
615821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    if (str != NULL)
625821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)      length_ = static_cast<int>(strlen(str));
635821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    else
645821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)      length_ = 0;
655821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
665821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  void set(const void* data, int len) {
675821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    ptr_ = reinterpret_cast<const char*>(data);
685821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    length_ = len;
695821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
705821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
715821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  char operator[](int i) const { return ptr_[i]; }
725821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
735821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  void remove_prefix(int n) {
745821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    ptr_ += n;
755821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    length_ -= n;
765821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
775821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
785821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  void remove_suffix(int n) {
795821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    length_ -= n;
805821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
815821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
825821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int compare(const StringPiece& x) const {
835821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    int r = memcmp(ptr_, x.ptr_, std::min(length_, x.length_));
845821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    if (r == 0) {
855821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)      if (length_ < x.length_) r = -1;
865821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)      else if (length_ > x.length_) r = +1;
875821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    }
885821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    return r;
895821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
905821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
915821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  std::string as_string() const {
925821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    return std::string(data(), size());
935821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
945821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // We also define ToString() here, since many other string-like
955821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // interfaces name the routine that converts to a C++ string
965821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // "ToString", and it's confusing to have the method that does that
975821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // for a StringPiece be called "as_string()".  We also leave the
985821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // "as_string()" method defined here for existing code.
995821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  std::string ToString() const {
1005821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    return std::string(data(), size());
1015821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
1025821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1035821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  void CopyToString(std::string* target) const;
1045821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  void AppendToString(std::string* target) const;
1055821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1065821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // Does "this" start with "x"
1075821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  bool starts_with(const StringPiece& x) const {
1085821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    return ((length_ >= x.length_) &&
1095821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)            (memcmp(ptr_, x.ptr_, x.length_) == 0));
1105821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
1115821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1125821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // Does "this" end with "x"
1135821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  bool ends_with(const StringPiece& x) const {
1145821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    return ((length_ >= x.length_) &&
1155821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)            (memcmp(ptr_ + (length_-x.length_), x.ptr_, x.length_) == 0));
1165821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
1175821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1185821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // standard STL container boilerplate
1195821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef char value_type;
1205821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef const char* pointer;
1215821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef const char& reference;
1225821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef const char& const_reference;
1235821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef size_t size_type;
1245821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef ptrdiff_t difference_type;
1255821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  static const size_type npos;
1265821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef const char* const_iterator;
1275821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef const char* iterator;
1285821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef std::reverse_iterator<const_iterator> const_reverse_iterator;
1295821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  typedef std::reverse_iterator<iterator> reverse_iterator;
1305821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  iterator begin() const { return ptr_; }
1315821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  iterator end() const { return ptr_ + length_; }
1325821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  const_reverse_iterator rbegin() const {
1335821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    return const_reverse_iterator(ptr_ + length_);
1345821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
1355821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  const_reverse_iterator rend() const {
1365821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)    return const_reverse_iterator(ptr_);
1375821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  }
1385821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  // STLS says return size_type, but Google says return int
1395821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int max_size() const { return length_; }
1405821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int capacity() const { return length_; }
1415821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1425821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int copy(char* buf, size_type n, size_type pos = 0) const;
1435821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1445821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int find(const StringPiece& s, size_type pos = 0) const;
1455821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int find(char c, size_type pos = 0) const;
1465821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int rfind(const StringPiece& s, size_type pos = npos) const;
1475821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  int rfind(char c, size_type pos = npos) const;
1485821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1495821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  StringPiece substr(size_type pos, size_type n = npos) const;
1505821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1515821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  static bool _equal(const StringPiece&, const StringPiece&);
1525821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)};
1535821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1545821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)inline bool operator==(const StringPiece& x, const StringPiece& y) {
1555821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  return StringPiece::_equal(x, y);
1565821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)}
1575821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1585821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)inline bool operator!=(const StringPiece& x, const StringPiece& y) {
1595821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  return !(x == y);
1605821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)}
1615821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1625821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)inline bool operator<(const StringPiece& x, const StringPiece& y) {
1635821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  const int r = memcmp(x.data(), y.data(),
1645821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)                       std::min(x.size(), y.size()));
1655821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  return ((r < 0) || ((r == 0) && (x.size() < y.size())));
1665821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)}
1675821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1685821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)inline bool operator>(const StringPiece& x, const StringPiece& y) {
1695821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  return y < x;
1705821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)}
1715821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1725821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)inline bool operator<=(const StringPiece& x, const StringPiece& y) {
1735821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  return !(x > y);
1745821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)}
1755821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1765821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)inline bool operator>=(const StringPiece& x, const StringPiece& y) {
1775821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)  return !(x < y);
1785821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)}
1795821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1805821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)}  // namespace re2
1815821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1825821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)// allow StringPiece to be logged
1835821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)extern std::ostream& operator<<(std::ostream& o, const re2::StringPiece& piece);
1845821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)
1855821806d5e7f356e8fa4b058a389a808ea183019Torne (Richard Coles)#endif  // STRINGS_STRINGPIECE_H__
186