1 /*===--- ConvertUTF.h - Universal Character Names conversions ---------------===
2  *
3  * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4  * See https://llvm.org/LICENSE.txt for license information.
5  * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6  *
7  *==------------------------------------------------------------------------==*/
8 /*
9  * Copyright 2001-2004 Unicode, Inc.
10  *
11  * Disclaimer
12  *
13  * This source code is provided as is by Unicode, Inc. No claims are
14  * made as to fitness for any particular purpose. No warranties of any
15  * kind are expressed or implied. The recipient agrees to determine
16  * applicability of information provided. If this file has been
17  * purchased on magnetic or optical media from Unicode, Inc., the
18  * sole remedy for any claim will be exchange of defective media
19  * within 90 days of receipt.
20  *
21  * Limitations on Rights to Redistribute This Code
22  *
23  * Unicode, Inc. hereby grants the right to freely use the information
24  * supplied in this file in the creation of products supporting the
25  * Unicode Standard, and to make copies of this file in any form
26  * for internal or external distribution as long as this notice
27  * remains attached.
28  */
29 
30 /* ---------------------------------------------------------------------
31 
32     Conversions between UTF32, UTF-16, and UTF-8.  Header file.
33 
34     Several funtions are included here, forming a complete set of
35     conversions between the three formats.  UTF-7 is not included
36     here, but is handled in a separate source file.
37 
38     Each of these routines takes pointers to input buffers and output
39     buffers.  The input buffers are const.
40 
41     Each routine converts the text between *sourceStart and sourceEnd,
42     putting the result into the buffer between *targetStart and
43     targetEnd. Note: the end pointers are *after* the last item: e.g.
44     *(sourceEnd - 1) is the last item.
45 
46     The return result indicates whether the conversion was successful,
47     and if not, whether the problem was in the source or target buffers.
48     (Only the first encountered problem is indicated.)
49 
50     After the conversion, *sourceStart and *targetStart are both
51     updated to point to the end of last text successfully converted in
52     the respective buffers.
53 
54     Input parameters:
55         sourceStart - pointer to a pointer to the source buffer.
56                 The contents of this are modified on return so that
57                 it points at the next thing to be converted.
58         targetStart - similarly, pointer to pointer to the target buffer.
59         sourceEnd, targetEnd - respectively pointers to the ends of the
60                 two buffers, for overflow checking only.
61 
62     These conversion functions take a ConversionFlags argument. When this
63     flag is set to strict, both irregular sequences and isolated surrogates
64     will cause an error.  When the flag is set to lenient, both irregular
65     sequences and isolated surrogates are converted.
66 
67     Whether the flag is strict or lenient, all illegal sequences will cause
68     an error return. This includes sequences such as: <F4 90 80 80>, <C0 80>,
69     or <A0> in UTF-8, and values above 0x10FFFF in UTF-32. Conformant code
70     must check for illegal sequences.
71 
72     When the flag is set to lenient, characters over 0x10FFFF are converted
73     to the replacement character; otherwise (when the flag is set to strict)
74     they constitute an error.
75 
76     Output parameters:
77         The value "sourceIllegal" is returned from some routines if the input
78         sequence is malformed.  When "sourceIllegal" is returned, the source
79         value will point to the illegal value that caused the problem. E.g.,
80         in UTF-8 when a sequence is malformed, it points to the start of the
81         malformed sequence.
82 
83     Author: Mark E. Davis, 1994.
84     Rev History: Rick McGowan, fixes & updates May 2001.
85          Fixes & updates, Sept 2001.
86 
87 ------------------------------------------------------------------------ */
88 
89 #ifndef LLVM_SUPPORT_CONVERTUTF_H
90 #define LLVM_SUPPORT_CONVERTUTF_H
91 
92 #include <cstddef>
93 #include <string>
94 
95 #if defined(_WIN32)
96 #include <system_error>
97 #endif
98 
99 // Wrap everything in namespace llvm so that programs can link with llvm and
100 // their own version of the unicode libraries.
101 
102 namespace llvm {
103 
104 /* ---------------------------------------------------------------------
105     The following 4 definitions are compiler-specific.
106     The C standard does not guarantee that wchar_t has at least
107     16 bits, so wchar_t is no less portable than unsigned short!
108     All should be unsigned values to avoid sign extension during
109     bit mask & shift operations.
110 ------------------------------------------------------------------------ */
111 
112 typedef unsigned int    UTF32;  /* at least 32 bits */
113 typedef unsigned short  UTF16;  /* at least 16 bits */
114 typedef unsigned char   UTF8;   /* typically 8 bits */
115 typedef unsigned char   Boolean; /* 0 or 1 */
116 
117 /* Some fundamental constants */
118 #define UNI_REPLACEMENT_CHAR (UTF32)0x0000FFFD
119 #define UNI_MAX_BMP (UTF32)0x0000FFFF
120 #define UNI_MAX_UTF16 (UTF32)0x0010FFFF
121 #define UNI_MAX_UTF32 (UTF32)0x7FFFFFFF
122 #define UNI_MAX_LEGAL_UTF32 (UTF32)0x0010FFFF
123 
124 #define UNI_MAX_UTF8_BYTES_PER_CODE_POINT 4
125 
126 #define UNI_UTF16_BYTE_ORDER_MARK_NATIVE  0xFEFF
127 #define UNI_UTF16_BYTE_ORDER_MARK_SWAPPED 0xFFFE
128 
129 #define UNI_UTF32_BYTE_ORDER_MARK_NATIVE 0x0000FEFF
130 #define UNI_UTF32_BYTE_ORDER_MARK_SWAPPED 0xFFFE0000
131 
132 typedef enum {
133   conversionOK,           /* conversion successful */
134   sourceExhausted,        /* partial character in source, but hit end */
135   targetExhausted,        /* insuff. room in target for conversion */
136   sourceIllegal           /* source sequence is illegal/malformed */
137 } ConversionResult;
138 
139 typedef enum {
140   strictConversion = 0,
141   lenientConversion
142 } ConversionFlags;
143 
144 ConversionResult ConvertUTF8toUTF16 (
145   const UTF8** sourceStart, const UTF8* sourceEnd,
146   UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags);
147 
148 /**
149  * Convert a partial UTF8 sequence to UTF32.  If the sequence ends in an
150  * incomplete code unit sequence, returns \c sourceExhausted.
151  */
152 ConversionResult ConvertUTF8toUTF32Partial(
153   const UTF8** sourceStart, const UTF8* sourceEnd,
154   UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags);
155 
156 /**
157  * Convert a partial UTF8 sequence to UTF32.  If the sequence ends in an
158  * incomplete code unit sequence, returns \c sourceIllegal.
159  */
160 ConversionResult ConvertUTF8toUTF32(
161   const UTF8** sourceStart, const UTF8* sourceEnd,
162   UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags);
163 
164 ConversionResult ConvertUTF16toUTF8 (
165   const UTF16** sourceStart, const UTF16* sourceEnd,
166   UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags);
167 
168 ConversionResult ConvertUTF32toUTF8 (
169   const UTF32** sourceStart, const UTF32* sourceEnd,
170   UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags);
171 
172 ConversionResult ConvertUTF16toUTF32 (
173   const UTF16** sourceStart, const UTF16* sourceEnd,
174   UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags);
175 
176 ConversionResult ConvertUTF32toUTF16 (
177   const UTF32** sourceStart, const UTF32* sourceEnd,
178   UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags);
179 
180 Boolean isLegalUTF8Sequence(const UTF8 *source, const UTF8 *sourceEnd);
181 
182 Boolean isLegalUTF8String(const UTF8 **source, const UTF8 *sourceEnd);
183 
184 unsigned getUTF8SequenceSize(const UTF8 *source, const UTF8 *sourceEnd);
185 
186 unsigned getNumBytesForUTF8(UTF8 firstByte);
187 
188 /*************************************************************************/
189 /* Below are LLVM-specific wrappers of the functions above. */
190 
191 template <typename T> class ArrayRef;
192 template <typename T> class SmallVectorImpl;
193 class StringRef;
194 
195 /**
196  * Convert an UTF8 StringRef to UTF8, UTF16, or UTF32 depending on
197  * WideCharWidth. The converted data is written to ResultPtr, which needs to
198  * point to at least WideCharWidth * (Source.Size() + 1) bytes. On success,
199  * ResultPtr will point one after the end of the copied string. On failure,
200  * ResultPtr will not be changed, and ErrorPtr will be set to the location of
201  * the first character which could not be converted.
202  * \return true on success.
203  */
204 bool ConvertUTF8toWide(unsigned WideCharWidth, llvm::StringRef Source,
205                        char *&ResultPtr, const UTF8 *&ErrorPtr);
206 
207 /**
208 * Converts a UTF-8 StringRef to a std::wstring.
209 * \return true on success.
210 */
211 bool ConvertUTF8toWide(llvm::StringRef Source, std::wstring &Result);
212 
213 /**
214 * Converts a UTF-8 C-string to a std::wstring.
215 * \return true on success.
216 */
217 bool ConvertUTF8toWide(const char *Source, std::wstring &Result);
218 
219 /**
220 * Converts a std::wstring to a UTF-8 encoded std::string.
221 * \return true on success.
222 */
223 bool convertWideToUTF8(const std::wstring &Source, std::string &Result);
224 
225 
226 /**
227  * Convert an Unicode code point to UTF8 sequence.
228  *
229  * \param Source a Unicode code point.
230  * \param [in,out] ResultPtr pointer to the output buffer, needs to be at least
231  * \c UNI_MAX_UTF8_BYTES_PER_CODE_POINT bytes.  On success \c ResultPtr is
232  * updated one past end of the converted sequence.
233  *
234  * \returns true on success.
235  */
236 bool ConvertCodePointToUTF8(unsigned Source, char *&ResultPtr);
237 
238 /**
239  * Convert the first UTF8 sequence in the given source buffer to a UTF32
240  * code point.
241  *
242  * \param [in,out] source A pointer to the source buffer. If the conversion
243  * succeeds, this pointer will be updated to point to the byte just past the
244  * end of the converted sequence.
245  * \param sourceEnd A pointer just past the end of the source buffer.
246  * \param [out] target The converted code
247  * \param flags Whether the conversion is strict or lenient.
248  *
249  * \returns conversionOK on success
250  *
251  * \sa ConvertUTF8toUTF32
252  */
253 inline ConversionResult convertUTF8Sequence(const UTF8 **source,
254                                             const UTF8 *sourceEnd,
255                                             UTF32 *target,
256                                             ConversionFlags flags) {
257   if (*source == sourceEnd)
258     return sourceExhausted;
259   unsigned size = getNumBytesForUTF8(**source);
260   if ((ptrdiff_t)size > sourceEnd - *source)
261     return sourceExhausted;
262   return ConvertUTF8toUTF32(source, *source + size, &target, target + 1, flags);
263 }
264 
265 /**
266  * Returns true if a blob of text starts with a UTF-16 big or little endian byte
267  * order mark.
268  */
269 bool hasUTF16ByteOrderMark(ArrayRef<char> SrcBytes);
270 
271 /**
272  * Converts a stream of raw bytes assumed to be UTF16 into a UTF8 std::string.
273  *
274  * \param [in] SrcBytes A buffer of what is assumed to be UTF-16 encoded text.
275  * \param [out] Out Converted UTF-8 is stored here on success.
276  * \returns true on success
277  */
278 bool convertUTF16ToUTF8String(ArrayRef<char> SrcBytes, std::string &Out);
279 
280 /**
281 * Converts a UTF16 string into a UTF8 std::string.
282 *
283 * \param [in] Src A buffer of UTF-16 encoded text.
284 * \param [out] Out Converted UTF-8 is stored here on success.
285 * \returns true on success
286 */
287 bool convertUTF16ToUTF8String(ArrayRef<UTF16> Src, std::string &Out);
288 
289 /**
290  * Converts a stream of raw bytes assumed to be UTF32 into a UTF8 std::string.
291  *
292  * \param [in] SrcBytes A buffer of what is assumed to be UTF-32 encoded text.
293  * \param [out] Out Converted UTF-8 is stored here on success.
294  * \returns true on success
295  */
296 bool convertUTF32ToUTF8String(ArrayRef<char> SrcBytes, std::string &Out);
297 
298 /**
299  * Converts a UTF32 string into a UTF8 std::string.
300  *
301  * \param [in] Src A buffer of UTF-32 encoded text.
302  * \param [out] Out Converted UTF-8 is stored here on success.
303  * \returns true on success
304  */
305 bool convertUTF32ToUTF8String(ArrayRef<UTF32> Src, std::string &Out);
306 
307 /**
308  * Converts a UTF-8 string into a UTF-16 string with native endianness.
309  *
310  * \returns true on success
311  */
312 bool convertUTF8ToUTF16String(StringRef SrcUTF8,
313                               SmallVectorImpl<UTF16> &DstUTF16);
314 
315 #if defined(_WIN32)
316 namespace sys {
317 namespace windows {
318 std::error_code UTF8ToUTF16(StringRef utf8, SmallVectorImpl<wchar_t> &utf16);
319 /// Convert to UTF16 from the current code page used in the system
320 std::error_code CurCPToUTF16(StringRef utf8, SmallVectorImpl<wchar_t> &utf16);
321 std::error_code UTF16ToUTF8(const wchar_t *utf16, size_t utf16_len,
322                             SmallVectorImpl<char> &utf8);
323 /// Convert from UTF16 to the current code page used in the system
324 std::error_code UTF16ToCurCP(const wchar_t *utf16, size_t utf16_len,
325                              SmallVectorImpl<char> &utf8);
326 } // namespace windows
327 } // namespace sys
328 #endif
329 
330 } /* end namespace llvm */
331 
332 #endif
333