Fix mouse warp with 2x displays
[chromium-blink-merge.git] / base / strings / string_split.cc
blobcc62a6f4b47a1a03d7fce6947c70590f59fc55eb
1 // Copyright (c) 2012 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 #include "base/strings/string_split.h"
7 #include "base/logging.h"
8 #include "base/strings/string_util.h"
9 #include "base/third_party/icu/icu_utf.h"
11 namespace base {
13 namespace {
15 // PieceToOutputType converts a StringPiece as needed to a given output type,
16 // which is either the same type of StringPiece (a NOP) or the corresponding
17 // non-piece string type.
19 // The default converter is a NOP, it works when the OutputType is the
20 // correct StringPiece.
21 template<typename Str, typename OutputType>
22 OutputType PieceToOutputType(BasicStringPiece<Str> piece) {
23 return piece;
25 template<> // Convert StringPiece to std::string
26 std::string PieceToOutputType<std::string, std::string>(StringPiece piece) {
27 return piece.as_string();
29 template<> // Convert StringPiece16 to string16.
30 string16 PieceToOutputType<string16, string16>(StringPiece16 piece) {
31 return piece.as_string();
34 // Returns either the ASCII or UTF-16 whitespace.
35 template<typename Str> BasicStringPiece<Str> WhitespaceForType();
36 template<> StringPiece16 WhitespaceForType<string16>() {
37 return kWhitespaceUTF16;
39 template<> StringPiece WhitespaceForType<std::string>() {
40 return kWhitespaceASCII;
43 // Optimize the single-character case to call find() on the string instead,
44 // since this is the common case and can be made faster. This could have been
45 // done with template specialization too, but would have been less clear.
47 // There is no corresponding FindFirstNotOf because StringPiece already
48 // implements these different versions that do the optimized searching.
49 size_t FindFirstOf(StringPiece piece, char c, size_t pos) {
50 return piece.find(c, pos);
52 size_t FindFirstOf(StringPiece16 piece, char16 c, size_t pos) {
53 return piece.find(c, pos);
55 size_t FindFirstOf(StringPiece piece, StringPiece one_of, size_t pos) {
56 return piece.find_first_of(one_of, pos);
58 size_t FindFirstOf(StringPiece16 piece, StringPiece16 one_of, size_t pos) {
59 return piece.find_first_of(one_of, pos);
62 // General string splitter template. Can take 8- or 16-bit input, can produce
63 // the corresponding string or StringPiece output, and can take single- or
64 // multiple-character delimiters.
66 // DelimiterType is either a character (Str::value_type) or a string piece of
67 // multiple characters (BasicStringPiece<Str>). StringPiece has a version of
68 // find for both of these cases, and the single-character version is the most
69 // common and can be implemented faster, which is why this is a template.
70 template<typename Str, typename OutputStringType, typename DelimiterType>
71 static std::vector<OutputStringType> SplitStringT(
72 BasicStringPiece<Str> str,
73 DelimiterType delimiter,
74 WhitespaceHandling whitespace,
75 SplitResult result_type) {
76 std::vector<OutputStringType> result;
77 if (str.empty())
78 return result;
80 size_t start = 0;
81 while (start != Str::npos) {
82 size_t end = FindFirstOf(str, delimiter, start);
84 BasicStringPiece<Str> piece;
85 if (end == Str::npos) {
86 piece = str.substr(start);
87 start = Str::npos;
88 } else {
89 piece = str.substr(start, end - start);
90 start = end + 1;
93 if (whitespace == TRIM_WHITESPACE)
94 piece = TrimString(piece, WhitespaceForType<Str>(), TRIM_ALL);
96 if (result_type == SPLIT_WANT_ALL || !piece.empty())
97 result.push_back(PieceToOutputType<Str, OutputStringType>(piece));
99 return result;
102 bool SplitStringIntoKeyValue(const std::string& line,
103 char key_value_delimiter,
104 std::string* key,
105 std::string* value) {
106 key->clear();
107 value->clear();
109 // Find the delimiter.
110 size_t end_key_pos = line.find_first_of(key_value_delimiter);
111 if (end_key_pos == std::string::npos) {
112 DVLOG(1) << "cannot find delimiter in: " << line;
113 return false; // no delimiter
115 key->assign(line, 0, end_key_pos);
117 // Find the value string.
118 std::string remains(line, end_key_pos, line.size() - end_key_pos);
119 size_t begin_value_pos = remains.find_first_not_of(key_value_delimiter);
120 if (begin_value_pos == std::string::npos) {
121 DVLOG(1) << "cannot parse value from line: " << line;
122 return false; // no value
124 value->assign(remains, begin_value_pos, remains.size() - begin_value_pos);
125 return true;
128 template <typename STR>
129 void SplitStringUsingSubstrT(const STR& str,
130 const STR& s,
131 std::vector<STR>* r) {
132 r->clear();
133 typename STR::size_type begin_index = 0;
134 while (true) {
135 const typename STR::size_type end_index = str.find(s, begin_index);
136 if (end_index == STR::npos) {
137 const STR term = str.substr(begin_index);
138 STR tmp;
139 TrimWhitespace(term, TRIM_ALL, &tmp);
140 r->push_back(tmp);
141 return;
143 const STR term = str.substr(begin_index, end_index - begin_index);
144 STR tmp;
145 TrimWhitespace(term, TRIM_ALL, &tmp);
146 r->push_back(tmp);
147 begin_index = end_index + s.size();
151 } // namespace
153 std::vector<std::string> SplitString(StringPiece input,
154 StringPiece separators,
155 WhitespaceHandling whitespace,
156 SplitResult result_type) {
157 if (separators.size() == 1) {
158 return SplitStringT<std::string, std::string, char>(
159 input, separators[0], whitespace, result_type);
161 return SplitStringT<std::string, std::string, StringPiece>(
162 input, separators, whitespace, result_type);
165 std::vector<string16> SplitString(StringPiece16 input,
166 StringPiece16 separators,
167 WhitespaceHandling whitespace,
168 SplitResult result_type) {
169 if (separators.size() == 1) {
170 return SplitStringT<string16, string16, char16>(
171 input, separators[0], whitespace, result_type);
173 return SplitStringT<string16, string16, StringPiece16>(
174 input, separators, whitespace, result_type);
177 std::vector<StringPiece> SplitStringPiece(StringPiece input,
178 StringPiece separators,
179 WhitespaceHandling whitespace,
180 SplitResult result_type) {
181 if (separators.size() == 1) {
182 return SplitStringT<std::string, StringPiece, char>(
183 input, separators[0], whitespace, result_type);
185 return SplitStringT<std::string, StringPiece, StringPiece>(
186 input, separators, whitespace, result_type);
189 std::vector<StringPiece16> SplitStringPiece(StringPiece16 input,
190 StringPiece16 separators,
191 WhitespaceHandling whitespace,
192 SplitResult result_type) {
193 if (separators.size() == 1) {
194 return SplitStringT<string16, StringPiece16, char16>(
195 input, separators[0], whitespace, result_type);
197 return SplitStringT<string16, StringPiece16, StringPiece16>(
198 input, separators, whitespace, result_type);
201 void SplitString(const string16& str,
202 char16 c,
203 std::vector<string16>* result) {
204 DCHECK(CBU16_IS_SINGLE(c));
205 *result = SplitStringT<string16, string16, char16>(
206 str, c, TRIM_WHITESPACE, SPLIT_WANT_ALL);
208 // Backward-compat hack: The old SplitString implementation would keep
209 // empty substrings, for example:
210 // "a,,b" -> ["a", "", "b"]
211 // "a, ,b" -> ["a", "", "b"]
212 // which the current code also does. But the old one would discard them when
213 // the only result was that empty string:
214 // " " -> []
215 // In the latter case, our new code will give [""]
216 if (result->size() == 1 && (*result)[0].empty())
217 result->clear();
220 void SplitString(const std::string& str,
221 char c,
222 std::vector<std::string>* result) {
223 #if CHAR_MIN < 0
224 DCHECK_GE(c, 0);
225 #endif
226 DCHECK_LT(c, 0x7F);
227 *result = SplitStringT<std::string, std::string, char>(
228 str, c, TRIM_WHITESPACE, SPLIT_WANT_ALL);
230 // Backward-compat hack, see above.
231 if (result->size() == 1 && (*result)[0].empty())
232 result->clear();
235 bool SplitStringIntoKeyValuePairs(const std::string& line,
236 char key_value_delimiter,
237 char key_value_pair_delimiter,
238 StringPairs* key_value_pairs) {
239 key_value_pairs->clear();
241 std::vector<std::string> pairs;
242 SplitString(line, key_value_pair_delimiter, &pairs);
244 bool success = true;
245 for (size_t i = 0; i < pairs.size(); ++i) {
246 // Don't add empty pairs into the result.
247 if (pairs[i].empty())
248 continue;
250 std::string key;
251 std::string value;
252 if (!SplitStringIntoKeyValue(pairs[i], key_value_delimiter, &key, &value)) {
253 // Don't return here, to allow for pairs without associated
254 // value or key; just record that the split failed.
255 success = false;
257 key_value_pairs->push_back(make_pair(key, value));
259 return success;
262 void SplitStringUsingSubstr(const string16& str,
263 const string16& s,
264 std::vector<string16>* r) {
265 SplitStringUsingSubstrT(str, s, r);
268 void SplitStringUsingSubstr(const std::string& str,
269 const std::string& s,
270 std::vector<std::string>* r) {
271 SplitStringUsingSubstrT(str, s, r);
274 void SplitStringDontTrim(StringPiece16 str,
275 char16 c,
276 std::vector<string16>* result) {
277 DCHECK(CBU16_IS_SINGLE(c));
278 *result = SplitStringT<string16, string16, char16>(
279 str, c, KEEP_WHITESPACE, SPLIT_WANT_ALL);
282 void SplitStringDontTrim(StringPiece str,
283 char c,
284 std::vector<std::string>* result) {
285 #if CHAR_MIN < 0
286 DCHECK_GE(c, 0);
287 #endif
288 DCHECK_LT(c, 0x7F);
289 *result = SplitStringT<std::string, std::string, char>(
290 str, c, KEEP_WHITESPACE, SPLIT_WANT_ALL);
293 void SplitStringAlongWhitespace(const string16& str,
294 std::vector<string16>* result) {
295 *result = SplitStringT<string16, string16, StringPiece16>(
296 str, StringPiece16(kWhitespaceASCIIAs16),
297 TRIM_WHITESPACE, SPLIT_WANT_NONEMPTY);
300 void SplitStringAlongWhitespace(const std::string& str,
301 std::vector<std::string>* result) {
302 *result = SplitStringT<std::string, std::string, StringPiece>(
303 str, StringPiece(kWhitespaceASCII),
304 TRIM_WHITESPACE, SPLIT_WANT_NONEMPTY);
307 } // namespace base