// Copyright (c) 2010 The Chromium Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#ifndef BASE_STRING_TOKENIZER_H_
#define BASE_STRING_TOKENIZER_H_
#pragma once
#include <algorithm>
#include <string>
#include "base/string_piece.h"
// StringTokenizerT is a simple string tokenizer class. It works like an
// iterator that with each step (see the Advance method) updates members that
// refer to the next token in the input string. The user may optionally
// configure the tokenizer to return delimiters.
//
// Warning: be careful not to pass a C string into the 2-arg constructor:
// StringTokenizer t("this is a test", " "); // WRONG
// This will create a temporary std::string, save the begin() and end()
// iterators, and then the string will be freed before we actually start
// tokenizing it.
// Instead, use a std::string or use the 3 arg constructor of CStringTokenizer.
//
//
// EXAMPLE 1:
//
// char input[] = "this is a test";
// CStringTokenizer t(input, input + strlen(input), " ");
// while (t.GetNext()) {
// printf("%s\n", t.token().c_str());
// }
//
// Output:
//
// this
// is
// a
// test
//
//
// EXAMPLE 2:
//
// std::string input = "no-cache=\"foo, bar\", private";
// StringTokenizer t(input, ", ");
// t.set_quote_chars("\"");
// while (t.GetNext()) {
// printf("%s\n", t.token().c_str());
// }
//
// Output:
//
// no-cache="foo, bar"
// private
//
//
// EXAMPLE 3:
//
// bool next_is_option = false, next_is_value = false;
// std::string input = "text/html; charset=UTF-8; foo=bar";
// StringTokenizer t(input, "; =");
// t.set_options(StringTokenizer::RETURN_DELIMS);
// while (t.GetNext()) {
// if (t.token_is_delim()) {
// switch (*t.token_begin()) {
// case ';':
// next_is_option = true;
// break;
// case '=':
// next_is_value = true;
// break;
// }
// } else {
// const char* label;
// if (next_is_option) {
// label = "option-name";
// next_is_option = false;
// } else if (next_is_value) {
// label = "option-value";
// next_is_value = false;
// } else {
// label = "mime-type";
// }
// printf("%s: %s\n", label, t.token().c_str());
// }
// }
//
//
template <class str, class const_iterator>
class StringTokenizerT {
public:
typedef typename str::value_type char_type;
// Options that may be pass to set_options()
enum {
// Specifies the delimiters should be returned as tokens
RETURN_DELIMS = 1 << 0,
};
// The string object must live longer than the tokenizer. (In particular this
// should not be constructed with a temporary.)
StringTokenizerT(const str& string,
const str& delims) {
Init(string.begin(), string.end(), delims);
}
StringTokenizerT(const_iterator string_begin,
const_iterator string_end,
const str& delims) {
Init(string_begin, string_end, delims);
}
// Set the options for this tokenizer. By default, this is 0.
void set_options(int options) { options_ = options; }
// Set the characters to regard as quotes. By default, this is empty. When
// a quote char is encountered, the tokenizer will switch into a mode where
// it ignores delimiters that it finds. It switches out of this mode once it
// finds another instance of the quote char. If a backslash is encountered
// within a quoted string, then the next character is skipped.
void set_quote_chars(const str& quotes) { quotes_ = quotes; }
// Call this method to advance the tokenizer to the next delimiter. This
// returns false if the tokenizer is complete. This method must be called
// before calling any of the token* methods.
bool GetNext() {
if (quotes_.empty() && options_ == 0)
return QuickGetNext();
else
return FullGetNext();
}
// Start iterating through tokens from the beginning of the string.
void Reset() {
token_end_ = start_pos_;
}
// Returns true if token is a delimiter. When the tokenizer is constructed
// with the RETURN_DELIMS option, this method can be used to check if the
// returned token is actually a delimiter.
bool token_is_delim() const { return token_is_delim_; }
// If GetNext() returned true, then these methods may be used to read the
// value of the token.
const_iterator token_begin() const { return token_begin_; }
const_iterator token_end() const { return token_end_; }
str token() const { return str(token_begin_, token_end_); }
base::StringPiece token_piece() const {
return base::StringPiece(&*token_begin_,
std::distance(token_begin_, token_end_));
}
private:
void Init(const_iterator string_begin,
const_iterator string_end,
const str& delims) {
start_pos_ = string_begin;
token_begin_ = string_begin;
token_end_ = string_begin;
end_ = string_end;
delims_ = delims;
options_ = 0;
token_is_delim_ = false;
}
// Implementation of GetNext() for when we have no quote characters. We have
// two separate implementations because AdvanceOne() is a hot spot in large
// text files with large tokens.
bool QuickGetNext() {
token_is_delim_ = false;
for (;;) {
token_begin_ = token_end_;
if (token_end_ == end_)
return false;
++token_end_;
if (delims_.find(*token_begin_) == str::npos)
break;
// else skip over delimiter.
}
while (token_end_ != end_ && delims_.find(*token_end_) == str::npos)
++token_end_;
return true;
}
// Implementation of GetNext() for when we have to take quotes into account.
bool FullGetNext() {
AdvanceState state;
token_is_delim_ = false;
for (;;) {
token_begin_ = token_end_;
if (token_end_ == end_)
return false;
++token_end_;
if (AdvanceOne(&state, *token_begin_))
break;
if (options_ & RETURN_DELIMS) {
token_is_delim_ = true;
return true;
}
// else skip over delimiter.
}
while (token_end_ != end_ && AdvanceOne(&state, *token_end_))
++token_end_;
return true;
}
bool IsDelim(char_type c) const {
return delims_.find(c) != str::npos;
}
bool IsQuote(char_type c) const {
return quotes_.find(c) != str::npos;
}
struct AdvanceState {
bool in_quote;
bool in_escape;
char_type quote_char;
AdvanceState() : in_quote(false), in_escape(false) {}
};
// Returns true if a delimiter was not hit.
bool AdvanceOne(AdvanceState* state, char_type c) {
if (state->in_quote) {
if (state->in_escape) {
state->in_escape = false;
} else if (c == '\\') {
state->in_escape = true;
} else if (c == state->quote_char) {
state->in_quote = false;
}
} else {
if (IsDelim(c))
return false;
state->in_quote = IsQuote(state->quote_char = c);
}
return true;
}
const_iterator start_pos_;
const_iterator token_begin_;
const_iterator token_end_;
const_iterator end_;
str delims_;
str quotes_;
int options_;
bool token_is_delim_;
};
typedef StringTokenizerT<std::string, std::string::const_iterator>
StringTokenizer;
typedef StringTokenizerT<std::wstring, std::wstring::const_iterator>
WStringTokenizer;
typedef StringTokenizerT<std::string, const char*> CStringTokenizer;
#endif // BASE_STRING_TOKENIZER_H_