]> git.saurik.com Git - wxWidgets.git/blob - interface/wx/convauto.h
Applied #15226 with modifications: wxRichTextCtrl: Implement setting properties with...
[wxWidgets.git] / interface / wx / convauto.h
1 /////////////////////////////////////////////////////////////////////////////
2 // Name: convauto.h
3 // Purpose: interface of wxConvAuto
4 // Author: wxWidgets team
5 // Licence: wxWindows licence
6 /////////////////////////////////////////////////////////////////////////////
7
8 /**
9 Constants representing various BOM types.
10
11 BOM is an abbreviation for "Byte Order Mark", a special Unicode character
12 which may be inserted into the beginning of a text stream to indicate its
13 encoding.
14
15 @since 2.9.3
16 */
17 enum wxBOM
18 {
19 /**
20 Unknown BOM.
21
22 This is returned if BOM presence couldn't be determined and normally
23 happens because not enough bytes of input have been analysed.
24 */
25 wxBOM_Unknown = -1,
26
27 /**
28 No BOM.
29
30 The stream doesn't contain BOM character at all.
31 */
32 wxBOM_None,
33
34 /**
35 UTF-32 big endian BOM.
36
37 The stream is encoded in big endian variant of UTF-32.
38 */
39 wxBOM_UTF32BE,
40
41 /**
42 UTF-32 little endian BOM.
43
44 The stream is encoded in little endian variant of UTF-32.
45 */
46 wxBOM_UTF32LE,
47
48 /**
49 UTF-16 big endian BOM.
50
51 The stream is encoded in big endian variant of UTF-16.
52 */
53 wxBOM_UTF16BE,
54
55 /**
56 UTF-16 little endian BOM.
57
58 The stream is encoded in little endian variant of UTF-16.
59 */
60 wxBOM_UTF16LE,
61
62 /**
63 UTF-8 BOM.
64
65 The stream is encoded in UTF-8.
66
67 Notice that contrary to a popular belief, it's perfectly possible and,
68 n fact, common under Microsoft Windows systems, to have a BOM in an
69 UTF-8 stream: while it's not used to indicate the endianness of UTF-8
70 stream (as it's byte-oriented), the BOM can still be useful just as an
71 unambiguous indicator of UTF-8 being used.
72 */
73 wxBOM_UTF8
74 };
75
76 /**
77 @class wxConvAuto
78
79 This class implements a Unicode to/from multibyte converter capable of
80 automatically recognizing the encoding of the multibyte text on input. The
81 logic used is very simple: the class uses the BOM (byte order mark) if it's
82 present and tries to interpret the input as UTF-8 otherwise. If this fails,
83 the input is interpreted as being in the default multibyte encoding which
84 can be specified in the constructor of a wxConvAuto instance and, in turn,
85 defaults to the value of GetFallbackEncoding() if not explicitly given.
86
87 For the conversion from Unicode to multibyte, the same encoding as was
88 previously used for multibyte to Unicode conversion is reused. If there had
89 been no previous multibyte to Unicode conversion, UTF-8 is used by default.
90 Notice that once the multibyte encoding is automatically detected, it
91 doesn't change any more, i.e. it is entirely determined by the first use of
92 wxConvAuto object in the multibyte-to-Unicode direction. However creating a
93 copy of wxConvAuto object, either via the usual copy constructor or
94 assignment operator, or using wxMBConv::Clone(), resets the automatically
95 detected encoding so that the new copy will try to detect the encoding of
96 the input on first use.
97
98 This class is used by default in wxWidgets classes and functions reading
99 text from files such as wxFile, wxFFile, wxTextFile, wxFileConfig and
100 various stream classes so the encoding set with its SetFallbackEncoding()
101 method will affect how these classes treat input files. In particular, use
102 this method to change the fall-back multibyte encoding used to interpret
103 the contents of the files whose contents isn't valid UTF-8 or to disallow
104 it completely.
105
106 @library{wxbase}
107 @category{data}
108
109 @see @ref overview_mbconv
110 */
111 class wxConvAuto : public wxMBConv
112 {
113 public:
114 /**
115 Constructs a new wxConvAuto instance. The object will try to detect the
116 input of the multibyte text given to its wxMBConv::ToWChar() method
117 automatically but if the automatic detection of Unicode encodings
118 fails, the fall-back encoding @a enc will be used to interpret it as
119 multibyte text.
120
121 The default value of @a enc, @c wxFONTENCODING_DEFAULT, means that the
122 global default value (which can be set using SetFallbackEncoding())
123 should be used. As with that method, passing @c wxFONTENCODING_MAX
124 inhibits using this encoding completely so the input multibyte text
125 will always be interpreted as UTF-8 in the absence of BOM and the
126 conversion will fail if the input doesn't form valid UTF-8 sequence.
127
128 Another special value is @c wxFONTENCODING_SYSTEM which means to use
129 the encoding currently used on the user system, i.e. the encoding
130 returned by wxLocale::GetSystemEncoding(). Any other encoding will be
131 used as is, e.g. passing @c wxFONTENCODING_ISO8859_1 ensures that
132 non-UTF-8 input will be treated as latin1.
133 */
134 wxConvAuto(wxFontEncoding enc = wxFONTENCODING_DEFAULT);
135
136
137 /**
138 Return the detected BOM type.
139
140 The BOM type is detected after sufficiently many initial bytes have
141 passed through this conversion object so it will always return
142 wxBOM_Unknown immediately after the object creation but may return a
143 different value later.
144
145 @since 2.9.3
146 */
147 wxBOM GetBOM() const;
148
149 /**
150 Return a pointer to the characters that makes up this BOM.
151
152 The returned character count is 2, 3 or 4, or undefined if the return
153 value is NULL.
154
155 @param bom
156 A valid BOM type, i.e. not wxBOM_Unknown or wxBOM_None.
157 @param count
158 A non-@NULL pointer receiving the number of characters in this BOM.
159 @return
160 Pointer to characters composing the BOM or @NULL if BOM is unknown
161 or invalid. Notice that the returned string is not NUL-terminated
162 and may contain embedded NULs so @a count must be used to handle it
163 correctly.
164
165 @since 2.9.3
166 */
167 const char* GetBOMChars(wxBOM bom, size_t* count);
168
169 /**
170 Disable the use of the fall back encoding: if the input doesn't have a
171 BOM and is not valid UTF-8, the conversion will fail.
172 */
173 static void DisableFallbackEncoding();
174
175 /**
176 Returns the encoding used by default by wxConvAuto if no other encoding
177 is explicitly specified in constructor. By default, returns
178 @c wxFONTENCODING_ISO8859_1 but can be changed using
179 SetFallbackEncoding().
180 */
181 static wxFontEncoding GetFallbackEncoding();
182
183 /**
184 Changes the encoding used by default by wxConvAuto if no other encoding
185 is explicitly specified in constructor. The default value, which can be
186 retrieved using GetFallbackEncoding(), is @c wxFONTENCODING_ISO8859_1.
187
188 Special values of @c wxFONTENCODING_SYSTEM or @c wxFONTENCODING_MAX can
189 be used for the @a enc parameter to use the encoding of the current
190 user locale as fall back or not use any encoding for fall back at all,
191 respectively (just as with the similar constructor parameter). However,
192 @c wxFONTENCODING_DEFAULT can't be used here.
193 */
194 static void SetFallbackEncoding(wxFontEncoding enc);
195
196 /**
197 Return the BOM type of this buffer.
198
199 This is a helper function which is normally only used internally by
200 wxConvAuto but provided for convenience of the code that wants to
201 detect the encoding of a stream by checking it for BOM presence on its
202 own.
203
204 @since 2.9.3
205 */
206 static wxBOM DetectBOM(const char *src, size_t srcLen);
207 };