]> git.saurik.com Git - wxWidgets.git/blob - interface/wx/convauto.h
Fix for caret positioning bug when selecting content
[wxWidgets.git] / interface / wx / convauto.h
1 /////////////////////////////////////////////////////////////////////////////
2 // Name: convauto.h
3 // Purpose: interface of wxConvAuto
4 // Author: wxWidgets team
5 // RCS-ID: $Id$
6 // Licence: wxWindows licence
7 /////////////////////////////////////////////////////////////////////////////
8
9 /**
10 Constants representing various BOM types.
11
12 BOM is an abbreviation for "Byte Order Mark", a special Unicode character
13 which may be inserted into the beginning of a text stream to indicate its
14 encoding.
15
16 @since 2.9.3
17 */
18 enum wxBOM
19 {
20 /**
21 Unknown BOM.
22
23 This is returned if BOM presence couldn't be determined and normally
24 happens because not enough bytes of input have been analysed.
25 */
26 wxBOM_Unknown = -1,
27
28 /**
29 No BOM.
30
31 The stream doesn't contain BOM character at all.
32 */
33 wxBOM_None,
34
35 /**
36 UTF-32 big endian BOM.
37
38 The stream is encoded in big endian variant of UTF-32.
39 */
40 wxBOM_UTF32BE,
41
42 /**
43 UTF-32 little endian BOM.
44
45 The stream is encoded in little endian variant of UTF-32.
46 */
47 wxBOM_UTF32LE,
48
49 /**
50 UTF-16 big endian BOM.
51
52 The stream is encoded in big endian variant of UTF-16.
53 */
54 wxBOM_UTF16BE,
55
56 /**
57 UTF-16 little endian BOM.
58
59 The stream is encoded in little endian variant of UTF-16.
60 */
61 wxBOM_UTF16LE,
62
63 /**
64 UTF-8 BOM.
65
66 The stream is encoded in UTF-8.
67
68 Notice that contrary to a popular belief, it's perfectly possible and,
69 n fact, common under Microsoft Windows systems, to have a BOM in an
70 UTF-8 stream: while it's not used to indicate the endianness of UTF-8
71 stream (as it's byte-oriented), the BOM can still be useful just as an
72 unambiguous indicator of UTF-8 being used.
73 */
74 wxBOM_UTF8
75 };
76
77 /**
78 @class wxConvAuto
79
80 This class implements a Unicode to/from multibyte converter capable of
81 automatically recognizing the encoding of the multibyte text on input. The
82 logic used is very simple: the class uses the BOM (byte order mark) if it's
83 present and tries to interpret the input as UTF-8 otherwise. If this fails,
84 the input is interpreted as being in the default multibyte encoding which
85 can be specified in the constructor of a wxConvAuto instance and, in turn,
86 defaults to the value of GetFallbackEncoding() if not explicitly given.
87
88 For the conversion from Unicode to multibyte, the same encoding as was
89 previously used for multibyte to Unicode conversion is reused. If there had
90 been no previous multibyte to Unicode conversion, UTF-8 is used by default.
91 Notice that once the multibyte encoding is automatically detected, it
92 doesn't change any more, i.e. it is entirely determined by the first use of
93 wxConvAuto object in the multibyte-to-Unicode direction. However creating a
94 copy of wxConvAuto object, either via the usual copy constructor or
95 assignment operator, or using wxMBConv::Clone(), resets the automatically
96 detected encoding so that the new copy will try to detect the encoding of
97 the input on first use.
98
99 This class is used by default in wxWidgets classes and functions reading
100 text from files such as wxFile, wxFFile, wxTextFile, wxFileConfig and
101 various stream classes so the encoding set with its SetFallbackEncoding()
102 method will affect how these classes treat input files. In particular, use
103 this method to change the fall-back multibyte encoding used to interpret
104 the contents of the files whose contents isn't valid UTF-8 or to disallow
105 it completely.
106
107 @library{wxbase}
108 @category{data}
109
110 @see @ref overview_mbconv
111 */
112 class wxConvAuto : public wxMBConv
113 {
114 public:
115 /**
116 Constructs a new wxConvAuto instance. The object will try to detect the
117 input of the multibyte text given to its wxMBConv::ToWChar() method
118 automatically but if the automatic detection of Unicode encodings
119 fails, the fall-back encoding @a enc will be used to interpret it as
120 multibyte text.
121
122 The default value of @a enc, @c wxFONTENCODING_DEFAULT, means that the
123 global default value (which can be set using SetFallbackEncoding())
124 should be used. As with that method, passing @c wxFONTENCODING_MAX
125 inhibits using this encoding completely so the input multibyte text
126 will always be interpreted as UTF-8 in the absence of BOM and the
127 conversion will fail if the input doesn't form valid UTF-8 sequence.
128
129 Another special value is @c wxFONTENCODING_SYSTEM which means to use
130 the encoding currently used on the user system, i.e. the encoding
131 returned by wxLocale::GetSystemEncoding(). Any other encoding will be
132 used as is, e.g. passing @c wxFONTENCODING_ISO8859_1 ensures that
133 non-UTF-8 input will be treated as latin1.
134 */
135 wxConvAuto(wxFontEncoding enc = wxFONTENCODING_DEFAULT);
136
137
138 /**
139 Return the detected BOM type.
140
141 The BOM type is detected after sufficiently many initial bytes have
142 passed through this conversion object so it will always return
143 wxBOM_Unknown immediately after the object creation but may return a
144 different value later.
145
146 @since 2.9.3
147 */
148 wxBOM GetBOM() const;
149
150 /**
151 Return a pointer to the characters that makes up this BOM.
152
153 The returned character count is 2, 3 or 4, or undefined if the return
154 value is NULL.
155
156 @param bom
157 A valid BOM type, i.e. not wxBOM_Unknown or wxBOM_None.
158 @param count
159 A non-@NULL pointer receiving the number of characters in this BOM.
160 @return
161 Pointer to characters composing the BOM or @NULL if BOM is unknown
162 or invalid. Notice that the returned string is not NUL-terminated
163 and may contain embedded NULs so @a count must be used to handle it
164 correctly.
165
166 @since 2.9.3
167 */
168 const char* GetBOMChars(wxBOM bom, size_t* count);
169
170 /**
171 Disable the use of the fall back encoding: if the input doesn't have a
172 BOM and is not valid UTF-8, the conversion will fail.
173 */
174 static void DisableFallbackEncoding();
175
176 /**
177 Returns the encoding used by default by wxConvAuto if no other encoding
178 is explicitly specified in constructor. By default, returns
179 @c wxFONTENCODING_ISO8859_1 but can be changed using
180 SetFallbackEncoding().
181 */
182 static wxFontEncoding GetFallbackEncoding();
183
184 /**
185 Changes the encoding used by default by wxConvAuto if no other encoding
186 is explicitly specified in constructor. The default value, which can be
187 retrieved using GetFallbackEncoding(), is @c wxFONTENCODING_ISO8859_1.
188
189 Special values of @c wxFONTENCODING_SYSTEM or @c wxFONTENCODING_MAX can
190 be used for the @a enc parameter to use the encoding of the current
191 user locale as fall back or not use any encoding for fall back at all,
192 respectively (just as with the similar constructor parameter). However,
193 @c wxFONTENCODING_DEFAULT can't be used here.
194 */
195 static void SetFallbackEncoding(wxFontEncoding enc);
196
197 /**
198 Return the BOM type of this buffer.
199
200 This is a helper function which is normally only used internally by
201 wxConvAuto but provided for convenience of the code that wants to
202 detect the encoding of a stream by checking it for BOM presence on its
203 own.
204
205 @since 2.9.3
206 */
207 static wxBOM DetectBOM(const char *src, size_t srcLen);
208 };