1 #ifndef UTF_CONVERSION_GROUP
2 #define UTF_CONVERSION_GROUP
4 /*****************************************************************************\
6 * Name : utf_conversion *
7 * Author : Unicode, Inc. (C conversion functions) *
8 * Author : Chris Koeritz (C++ conversion classes) *
10 *******************************************************************************
11 * Copyright (c) 2006-$now By Author. This program is free software; you can *
12 * redistribute it and/or modify it under the terms of the GNU General Public *
13 * License as published by the Free Software Foundation; either version 2 of *
14 * the License or (at your option) any later version. This is online at: *
15 * http://www.fsf.org/copyleft/gpl.html *
16 * Please send any updates to: fred@gruntose.com *
17 \*****************************************************************************/
20 #include "definitions.h"
22 //! @file utf_conversion.h Support for unicode builds.
23 /*! @file utf_conversion.h
24 This file provides conversions between UTF-8 and UTF-16 that are essential
25 to having the code compile in either the UNICODE or regular ANSI character
26 builds of win32-based code.
29 // original copyright notice still applies to low-level conversion code:
31 * Copyright 2001-$now Unicode, Inc.
35 * This source code is provided as is by Unicode, Inc. No claims are
36 * made as to fitness for any particular purpose. No warranties of any
37 * kind are expressed or implied. The recipient agrees to determine
38 * applicability of information provided. If this file has been
39 * purchased on magnetic or optical media from Unicode, Inc., the
40 * sole remedy for any claim will be exchange of defective media
41 * within 90 days of receipt.
43 * Limitations on Rights to Redistribute This Code
45 * Unicode, Inc. hereby grants the right to freely use the information
46 * supplied in this file in the creation of products supporting the
47 * Unicode Standard, and to make copies of this file in any form
48 * for internal or external distribution as long as this notice
52 /* ---------------------------------------------------------------------
54 Conversions between UTF32, UTF-16, and UTF-8. Header file.
56 Several funtions are included here, forming a complete set of
57 conversions between the three formats. UTF-7 is not included
58 here, but is handled in a separate source file.
60 Each of these routines takes pointers to input buffers and output
61 buffers. The input buffers are const.
63 Each routine converts the text between *sourceStart and sourceEnd,
64 putting the result into the buffer between *targetStart and
65 targetEnd. Note: the end pointers are *after* the last item: e.g.
66 *(sourceEnd - 1) is the last item.
68 The return result indicates whether the conversion was successful,
69 and if not, whether the problem was in the source or target buffers.
70 (Only the first encountered problem is indicated.)
72 After the conversion, *sourceStart and *targetStart are both
73 updated to point to the end of last text successfully converted in
74 the respective buffers.
77 sourceStart - pointer to a pointer to the source buffer.
78 The contents of this are modified on return so that
79 it points at the next thing to be converted.
80 targetStart - similarly, pointer to pointer to the target buffer.
81 sourceEnd, targetEnd - respectively pointers to the ends of the
82 two buffers, for overflow checking only.
84 These conversion functions take a ConversionFlags argument. When this
85 flag is set to strict, both irregular sequences and isolated surrogates
86 will cause an error. When the flag is set to lenient, both irregular
87 sequences and isolated surrogates are converted.
89 Whether the flag is strict or lenient, all illegal sequences will cause
90 an error return. This includes sequences such as: <F4 90 80 80>, <C0 80>,
91 or <A0> in UTF-8, and values above 0x10FFFF in UTF-32. Conformant code
92 must check for illegal sequences.
94 When the flag is set to lenient, characters over 0x10FFFF are converted
95 to the replacement character; otherwise (when the flag is set to strict)
96 they constitute an error.
99 The value "sourceIllegal" is returned from some routines if the input
100 sequence is malformed. When "sourceIllegal" is returned, the source
101 value will point to the illegal value that caused the problem. E.g.,
102 in UTF-8 when a sequence is malformed, it points to the start of the
105 Author: Mark E. Davis, 1994.
106 Rev History: Rick McGowan, fixes & updates May 2001.
107 Fixes & updates, Sept 2001.
109 ------------------------------------------------------------------------ */
113 /* ---------------------------------------------------------------------
114 The following 4 definitions are compiler-specific.
115 The C standard does not guarantee that wchar_t has at least
116 16 bits, so wchar_t is no less portable than unsigned short!
117 All should be unsigned values to avoid sign extension during
118 bit mask & shift operations.
119 ------------------------------------------------------------------------ */
120 typedef unsigned long UTF32; /* at least 32 bits */
121 typedef unsigned short UTF16; /* at least 16 bits */
122 typedef unsigned char UTF8; /* typically 8 bits */
123 typedef unsigned char Booleano; /* 0 or 1 */
125 //hmmm: this is some really gross stuff below, just because it's got
126 // so much of its guts hanging out. make a class out of the lower-
127 // level conversion stuff and hide all the details.
129 /* Some fundamental constants */
130 #define UNI_REPLACEMENT_CHAR (UTF32)0x0000FFFD
131 #define UNI_MAX_BMP (UTF32)0x0000FFFF
132 #define UNI_MAX_UTF16 (UTF32)0x0010FFFF
133 #define UNI_MAX_UTF32 (UTF32)0x7FFFFFFF
134 #define UNI_MAX_LEGAL_UTF32 (UTF32)0x0010FFFF
137 conversionOK, /* conversion successful */
138 sourceExhausted, /* partial character in source, but hit end */
139 targetExhausted, /* insuff. room in target for conversion */
140 sourceIllegal /* source sequence is illegal/malformed */
144 strictConversion = 0,
152 ConversionResult ConvertUTF8toUTF16 (const UTF8** sourceStart,
153 const UTF8* sourceEnd, UTF16** targetStart, UTF16* targetEnd,
154 ConversionFlags flags);
156 ConversionResult ConvertUTF16toUTF8 (const UTF16** sourceStart,
157 const UTF16* sourceEnd, UTF8** targetStart, UTF8* targetEnd,
158 ConversionFlags flags);
160 ConversionResult ConvertUTF8toUTF32 (const UTF8** sourceStart,
161 const UTF8* sourceEnd, UTF32** targetStart, UTF32* targetEnd,
162 ConversionFlags flags);
164 ConversionResult ConvertUTF32toUTF8 (const UTF32** sourceStart,
165 const UTF32* sourceEnd, UTF8** targetStart, UTF8* targetEnd,
166 ConversionFlags flags);
168 ConversionResult ConvertUTF16toUTF32 (const UTF16** sourceStart,
169 const UTF16* sourceEnd, UTF32** targetStart, UTF32* targetEnd,
170 ConversionFlags flags);
172 ConversionResult ConvertUTF32toUTF16 (const UTF32** sourceStart,
173 const UTF32* sourceEnd, UTF16** targetStart, UTF16* targetEnd,
174 ConversionFlags flags);
176 Booleano isLegalUTF8Sequence(const UTF8 *source, const UTF8 *sourceEnd);
186 // The following types and macros help to make it irrelevant what kind of
187 // win32 build is being done. They will adapt as needed to provide the
188 // types used in system calls. They are rendered harmless for other operating
189 // systems or for non-Unicode builds; this is especially useful for POSIX
190 // compliant functions that required Unicode in win32 but not in Unix systems.
193 #error should not be in here right now --cak
194 ///holding to test wx && defined(__WIN32__)
195 //! to_unicode_temp() converts to UTF-16 as needed for win32 system calls.
196 /*! this conversion is only appropriate to use within expressions. it does
197 not create a permanent object and will go out of scope when the full
199 #define to_unicode_temp(s) transcode_to_utf16(s)
200 //! the from_unicode_temp() macro converts the win32 unicode string to UTF-8.
201 /*! this has the same contraint as to_unicode_temp; it does not persist
202 beyond the lifetime of the expression containing it. */
203 #define from_unicode_temp(s) transcode_to_utf8(s)
204 //! to_unicode_persist creates a UTF-16 object with block scope.
205 /*! the object with "name" that is created by this macro will exist on
206 the stack until the block containing it is closed. this is needed when
207 storing unicode strings into windows structure, for example. */
208 #define to_unicode_persist(name, s) transcode_to_utf16 name(s)
209 //! from_unicode_persist creates a UTF-8 object with block scope.
210 /*! similar to to_unicode_persist, this is used for longer-lived objects. */
211 #define from_unicode_persist(name, s) transcode_to_utf8 name(s)
213 // these versions of the macros simply defang any conversions.
214 #define to_unicode_temp(s) null_transcoder(s, false)
215 #define from_unicode_temp(s) null_transcoder(s, false)
216 #define to_unicode_persist(name, s) null_transcoder name(s, true)
217 #define from_unicode_persist(name, s) null_transcoder name(s, true)
221 //! sends UTF-8 information to the diagnostic view in the IDE.
222 #define TRACE_PRINT(s) TRACE(_T("%s"), to_unicode_temp(s))
227 // The next two classes support converting a UTF-8 string into a UTF-16
228 // string and vice-versa. They hold onto the converted string and provide
229 // operators that return it.
231 //! handles the conversion of UTF-8 format strings into UTF-16.
233 class transcode_to_utf16 : public virtual root_object
236 transcode_to_utf16(const char *utf8_input);
237 //!< translates the "utf8_input" from UTF-8 into UTF-16 format.
238 /*!< the resulting memory is held in this object until destruction, which
239 makes this especially convenient for in-place conversions. */
241 transcode_to_utf16(const astring &utf8_input);
242 //!< this conversion converts the astring in "utf8_input" into UTF-16.
244 virtual ~transcode_to_utf16();
247 //!< reports the length of the double-byte-based converted string.
249 operator const UTF16 * () const { return _converted; }
250 //!< observes the converted version of the string (constant version).
251 operator UTF16 * () { return _converted; }
252 //!< accesses the converted version of the string (modifiable version).
253 operator const flexichar * () const { return (const flexichar *)_converted; }
254 //!< accesses the converted version of the string (modifiable version).
255 operator flexichar * () { return (flexichar *)_converted; }
256 //!< accesses the converted version of the string (modifiable version).
258 ConversionResult _result;
260 int _orig_length; //!< how long is the original argument?
261 UTF16 *_converted; //!< the double-byte version of the original string.
266 //! handles the conversion of UTF-16 format strings into UTF-8.
268 class transcode_to_utf8 : public virtual root_object
271 transcode_to_utf8(const UTF16 *utf16_input);
272 //!< translates the "utf16_input" from UTF-16 into UTF-8 format.
273 /*!< the resulting memory is held in this object until destruction, which
274 makes this especially convenient for in-place conversions. */
276 transcode_to_utf8(const wchar_t *utf16_input);
277 //!< converts win32 style wide characters to UTF-8.
279 virtual ~transcode_to_utf8();
282 //!< reports the length of the byte-based converted string.
284 operator const UTF8 * () const { return _converted; }
285 //!< observes the converted version of the string (constant version).
286 operator UTF8 * () { return _converted; }
287 //!< accesses the converted version of the string (modifiable version).
289 operator astring() const;
290 //!< converts the char pointer into an astring object.
292 ConversionResult _result;
294 int _orig_length; //!< how long is the original argument?
295 int _new_length; //!< what did we predictively allocate for the result?
296 UTF8 *_converted; //!< the double-byte version of the original string.
301 //! provides an interface like the transcoders above, but stays in UTF-8.
303 class null_transcoder : public virtual root_object
306 //! allocates memory for the converted form if "make_own_copy" is true.
307 null_transcoder(const char *utf8_input, bool make_own_copy);
308 //! similar, but supports astrings.
309 null_transcoder(const astring &utf8_input, bool make_own_copy);
311 if (_make_own_copy) delete [] _converted;
312 _converted = NULL_POINTER;
316 operator char * () { return (char *)_converted; }
317 operator const char * () const { return (const char *)_converted; }
321 const UTF8 *_converted;
328 #endif // outer guard.