Add PopupZoomer view to hierarchy only when necessary.
[chromium-blink-merge.git] / base / string_tokenizer.h
blobc2307a5209d1a496ce60745af11e4bf37d6f2266
1 // Copyright (c) 2011 The Chromium Authors. All rights reserved.
2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file.
5 #ifndef BASE_STRING_TOKENIZER_H_
6 #define BASE_STRING_TOKENIZER_H_
8 #include <algorithm>
9 #include <string>
11 #include "base/string_piece.h"
13 // StringTokenizerT is a simple string tokenizer class. It works like an
14 // iterator that with each step (see the Advance method) updates members that
15 // refer to the next token in the input string. The user may optionally
16 // configure the tokenizer to return delimiters.
18 // Warning: be careful not to pass a C string into the 2-arg constructor:
19 // StringTokenizer t("this is a test", " "); // WRONG
20 // This will create a temporary std::string, save the begin() and end()
21 // iterators, and then the string will be freed before we actually start
22 // tokenizing it.
23 // Instead, use a std::string or use the 3 arg constructor of CStringTokenizer.
26 // EXAMPLE 1:
28 // char input[] = "this is a test";
29 // CStringTokenizer t(input, input + strlen(input), " ");
30 // while (t.GetNext()) {
31 // printf("%s\n", t.token().c_str());
32 // }
34 // Output:
36 // this
37 // is
38 // a
39 // test
42 // EXAMPLE 2:
44 // std::string input = "no-cache=\"foo, bar\", private";
45 // StringTokenizer t(input, ", ");
46 // t.set_quote_chars("\"");
47 // while (t.GetNext()) {
48 // printf("%s\n", t.token().c_str());
49 // }
51 // Output:
53 // no-cache="foo, bar"
54 // private
57 // EXAMPLE 3:
59 // bool next_is_option = false, next_is_value = false;
60 // std::string input = "text/html; charset=UTF-8; foo=bar";
61 // StringTokenizer t(input, "; =");
62 // t.set_options(StringTokenizer::RETURN_DELIMS);
63 // while (t.GetNext()) {
64 // if (t.token_is_delim()) {
65 // switch (*t.token_begin()) {
66 // case ';':
67 // next_is_option = true;
68 // break;
69 // case '=':
70 // next_is_value = true;
71 // break;
72 // }
73 // } else {
74 // const char* label;
75 // if (next_is_option) {
76 // label = "option-name";
77 // next_is_option = false;
78 // } else if (next_is_value) {
79 // label = "option-value";
80 // next_is_value = false;
81 // } else {
82 // label = "mime-type";
83 // }
84 // printf("%s: %s\n", label, t.token().c_str());
85 // }
86 // }
89 template <class str, class const_iterator>
90 class StringTokenizerT {
91 public:
92 typedef typename str::value_type char_type;
94 // Options that may be pass to set_options()
95 enum {
96 // Specifies the delimiters should be returned as tokens
97 RETURN_DELIMS = 1 << 0,
100 // The string object must live longer than the tokenizer. (In particular this
101 // should not be constructed with a temporary.)
102 StringTokenizerT(const str& string,
103 const str& delims) {
104 Init(string.begin(), string.end(), delims);
107 StringTokenizerT(const_iterator string_begin,
108 const_iterator string_end,
109 const str& delims) {
110 Init(string_begin, string_end, delims);
113 // Set the options for this tokenizer. By default, this is 0.
114 void set_options(int options) { options_ = options; }
116 // Set the characters to regard as quotes. By default, this is empty. When
117 // a quote char is encountered, the tokenizer will switch into a mode where
118 // it ignores delimiters that it finds. It switches out of this mode once it
119 // finds another instance of the quote char. If a backslash is encountered
120 // within a quoted string, then the next character is skipped.
121 void set_quote_chars(const str& quotes) { quotes_ = quotes; }
123 // Call this method to advance the tokenizer to the next delimiter. This
124 // returns false if the tokenizer is complete. This method must be called
125 // before calling any of the token* methods.
126 bool GetNext() {
127 if (quotes_.empty() && options_ == 0)
128 return QuickGetNext();
129 else
130 return FullGetNext();
133 // Start iterating through tokens from the beginning of the string.
134 void Reset() {
135 token_end_ = start_pos_;
138 // Returns true if token is a delimiter. When the tokenizer is constructed
139 // with the RETURN_DELIMS option, this method can be used to check if the
140 // returned token is actually a delimiter.
141 bool token_is_delim() const { return token_is_delim_; }
143 // If GetNext() returned true, then these methods may be used to read the
144 // value of the token.
145 const_iterator token_begin() const { return token_begin_; }
146 const_iterator token_end() const { return token_end_; }
147 str token() const { return str(token_begin_, token_end_); }
148 base::StringPiece token_piece() const {
149 return base::StringPiece(&*token_begin_,
150 std::distance(token_begin_, token_end_));
153 private:
154 void Init(const_iterator string_begin,
155 const_iterator string_end,
156 const str& delims) {
157 start_pos_ = string_begin;
158 token_begin_ = string_begin;
159 token_end_ = string_begin;
160 end_ = string_end;
161 delims_ = delims;
162 options_ = 0;
163 token_is_delim_ = false;
166 // Implementation of GetNext() for when we have no quote characters. We have
167 // two separate implementations because AdvanceOne() is a hot spot in large
168 // text files with large tokens.
169 bool QuickGetNext() {
170 token_is_delim_ = false;
171 for (;;) {
172 token_begin_ = token_end_;
173 if (token_end_ == end_)
174 return false;
175 ++token_end_;
176 if (delims_.find(*token_begin_) == str::npos)
177 break;
178 // else skip over delimiter.
180 while (token_end_ != end_ && delims_.find(*token_end_) == str::npos)
181 ++token_end_;
182 return true;
185 // Implementation of GetNext() for when we have to take quotes into account.
186 bool FullGetNext() {
187 AdvanceState state;
188 token_is_delim_ = false;
189 for (;;) {
190 token_begin_ = token_end_;
191 if (token_end_ == end_)
192 return false;
193 ++token_end_;
194 if (AdvanceOne(&state, *token_begin_))
195 break;
196 if (options_ & RETURN_DELIMS) {
197 token_is_delim_ = true;
198 return true;
200 // else skip over delimiter.
202 while (token_end_ != end_ && AdvanceOne(&state, *token_end_))
203 ++token_end_;
204 return true;
207 bool IsDelim(char_type c) const {
208 return delims_.find(c) != str::npos;
211 bool IsQuote(char_type c) const {
212 return quotes_.find(c) != str::npos;
215 struct AdvanceState {
216 bool in_quote;
217 bool in_escape;
218 char_type quote_char;
219 AdvanceState() : in_quote(false), in_escape(false), quote_char('\0') {}
222 // Returns true if a delimiter was not hit.
223 bool AdvanceOne(AdvanceState* state, char_type c) {
224 if (state->in_quote) {
225 if (state->in_escape) {
226 state->in_escape = false;
227 } else if (c == '\\') {
228 state->in_escape = true;
229 } else if (c == state->quote_char) {
230 state->in_quote = false;
232 } else {
233 if (IsDelim(c))
234 return false;
235 state->in_quote = IsQuote(state->quote_char = c);
237 return true;
240 const_iterator start_pos_;
241 const_iterator token_begin_;
242 const_iterator token_end_;
243 const_iterator end_;
244 str delims_;
245 str quotes_;
246 int options_;
247 bool token_is_delim_;
250 typedef StringTokenizerT<std::string, std::string::const_iterator>
251 StringTokenizer;
252 typedef StringTokenizerT<std::wstring, std::wstring::const_iterator>
253 WStringTokenizer;
254 typedef StringTokenizerT<std::string, const char*> CStringTokenizer;
256 #endif // BASE_STRING_TOKENIZER_H_