string_split.h revision 4a5e2dc747d50c653511c68ccb2cfbfb740bd5a7 
1// Copyright (c) 2010 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.
4
5#ifndef BASE_STRING_SPLIT_H_
6#define BASE_STRING_SPLIT_H_
7#pragma once
8
9#include <string>
10#include <utility>
11#include <vector>
12
13#include "base/string16.h"
14
15namespace base {
16
17// Splits |str| into a vector of strings delimited by |s|. Append the results
18// into |r| as they appear. If several instances of |s| are contiguous, or if
19// |str| begins with or ends with |s|, then an empty string is inserted.
20//
21// Every substring is trimmed of any leading or trailing white space.
22// Where wchar_t is char16 (i.e. Windows), |c| must be in BMP
23// (Basic Multilingual Plane). Elsewhere (Linux/Mac), wchar_t
24// should be a valid Unicode code point (32-bit).
25void SplitString(const std::wstring& str,
26                 wchar_t c,
27                 std::vector<std::wstring>* r);
28// NOTE: |c| must be in BMP (Basic Multilingual Plane)
29void SplitString(const string16& str,
30                 char16 c,
31                 std::vector<string16>* r);
32// |str| should not be in a multi-byte encoding like Shift-JIS or GBK in which
33// the trailing byte of a multi-byte character can be in the ASCII range.
34// UTF-8, and other single/multi-byte ASCII-compatible encodings are OK.
35// Note: |c| must be in the ASCII range.
36void SplitString(const std::string& str,
37                 char c,
38                 std::vector<std::string>* r);
39
40bool SplitStringIntoKeyValues(
41    const std::string& line,
42    char key_value_delimiter,
43    std::string* key, std::vector<std::string>* values);
44
45bool SplitStringIntoKeyValuePairs(
46    const std::string& line,
47    char key_value_delimiter,
48    char key_value_pair_delimiter,
49    std::vector<std::pair<std::string, std::string> >* kv_pairs);
50
51// The same as SplitString, but use a substring delimiter instead of a char.
52void SplitStringUsingSubstr(const string16& str,
53                            const string16& s,
54                            std::vector<string16>* r);
55void SplitStringUsingSubstr(const std::string& str,
56                            const std::string& s,
57                            std::vector<std::string>* r);
58
59// The same as SplitString, but don't trim white space.
60// NOTE: |c| must be in BMP (Basic Multilingual Plane)
61void SplitStringDontTrim(const string16& str,
62                         char16 c,
63                         std::vector<string16>* r);
64// |str| should not be in a multi-byte encoding like Shift-JIS or GBK in which
65// the trailing byte of a multi-byte character can be in the ASCII range.
66// UTF-8, and other single/multi-byte ASCII-compatible encodings are OK.
67// Note: |c| must be in the ASCII range.
68void SplitStringDontTrim(const std::string& str,
69                         char c,
70                         std::vector<std::string>* r);
71
72// WARNING: this uses whitespace as defined by the HTML5 spec. If you need
73// a function similar to this but want to trim all types of whitespace, then
74// factor this out into a function that takes a string containing the characters
75// that are treated as whitespace.
76//
77// Splits the string along whitespace (where whitespace is the five space
78// characters defined by HTML 5). Each contiguous block of non-whitespace
79// characters is added to result.
80void SplitStringAlongWhitespace(const std::wstring& str,
81                                std::vector<std::wstring>* result);
82void SplitStringAlongWhitespace(const string16& str,
83                                std::vector<string16>* result);
84void SplitStringAlongWhitespace(const std::string& str,
85                                std::vector<std::string>* result);
86
87}  // namespace base
88
89#endif  // BASE_STRING_SPLIT_H
90