| 1 | ///////////////////////////////////////////////////////////////////////////// |
| 2 | // Name: strconv.h |
| 3 | // Purpose: interface of wxMBConvUTF7 |
| 4 | // Author: wxWidgets team |
| 5 | // RCS-ID: $Id$ |
| 6 | // Licence: wxWindows licence |
| 7 | ///////////////////////////////////////////////////////////////////////////// |
| 8 | |
| 9 | /** |
| 10 | @class wxMBConv |
| 11 | |
| 12 | This class is the base class of a hierarchy of classes capable of |
| 13 | converting text strings between multibyte (SBCS or DBCS) encodings and |
| 14 | Unicode. |
| 15 | |
| 16 | This is an abstract base class which defines the operations implemented by |
| 17 | all different conversion classes. The derived classes don't add any new |
| 18 | operations of their own (except, possibly, some non-default constructors) |
| 19 | and so you should simply use this class ToWChar() and FromWChar() (or |
| 20 | cMB2WC() and cWC2MB()) methods with the objects of the derived class. |
| 21 | |
| 22 | In the documentation for this and related classes please notice that |
| 23 | length of the string refers to the number of characters in the string |
| 24 | not counting the terminating @c NUL, if any. While the size of the string |
| 25 | is the total number of bytes in the string, including any trailing @c NUL. |
| 26 | Thus, length of wide character string @c L"foo" is 3 while its size can |
| 27 | be either 8 or 16 depending on whether @c wchar_t is 2 bytes (as |
| 28 | under Windows) or 4 (Unix). |
| 29 | |
| 30 | @library{wxbase} |
| 31 | @category{conv} |
| 32 | |
| 33 | @see wxCSConv, wxEncodingConverter, @ref overview_mbconv |
| 34 | */ |
| 35 | class wxMBConv |
| 36 | { |
| 37 | public: |
| 38 | /** |
| 39 | Trivial default constructor. |
| 40 | */ |
| 41 | wxMBConv(); |
| 42 | |
| 43 | /** |
| 44 | This pure virtual function is overridden in each of the derived classes |
| 45 | to return a new copy of the object it is called on. |
| 46 | |
| 47 | It is used for copying the conversion objects while preserving their |
| 48 | dynamic type. |
| 49 | */ |
| 50 | virtual wxMBConv* Clone() const = 0; |
| 51 | |
| 52 | /** |
| 53 | This function returns 1 for most of the multibyte encodings in which the |
| 54 | string is terminated by a single @c NUL, 2 for UTF-16 and 4 for UTF-32 for |
| 55 | which the string is terminated with 2 and 4 @c NUL characters respectively. |
| 56 | The other cases are not currently supported and @c wxCONV_FAILED |
| 57 | (defined as -1) is returned for them. |
| 58 | */ |
| 59 | virtual size_t GetMBNulLen() const; |
| 60 | |
| 61 | /** |
| 62 | Returns the maximal value which can be returned by GetMBNulLen() for |
| 63 | any conversion object. |
| 64 | |
| 65 | Currently this value is 4. |
| 66 | |
| 67 | This method can be used to allocate the buffer with enough space for the |
| 68 | trailing @c NUL characters for any encoding. |
| 69 | */ |
| 70 | static size_t GetMaxMBNulLen(); |
| 71 | |
| 72 | /** |
| 73 | Convert multibyte string to a wide character one. |
| 74 | |
| 75 | This is the most general function for converting a multibyte string to |
| 76 | a wide string, cMB2WC() may be often more convenient, however this |
| 77 | function is the most efficient one as it allows to avoid any |
| 78 | unnecessary copying. |
| 79 | |
| 80 | The main case is when @a dst is not @NULL and @a srcLen is not |
| 81 | @c wxNO_LEN (which is defined as @c (size_t)-1): then the function |
| 82 | converts exactly @a srcLen bytes starting at @a src into wide string |
| 83 | which it output to @e dst. If the length of the resulting wide |
| 84 | string is greater than @e dstLen, an error is returned. Note that if |
| 85 | @a srcLen bytes don't include @c NUL characters, the resulting wide |
| 86 | string is not @c NUL-terminated neither. |
| 87 | |
| 88 | If @a srcLen is @c wxNO_LEN, the function supposes that the string is |
| 89 | properly (i.e. as necessary for the encoding handled by this |
| 90 | conversion) @c NUL-terminated and converts the entire string, including |
| 91 | any trailing @c NUL bytes. In this case the wide string is also @c |
| 92 | NUL-terminated. |
| 93 | |
| 94 | Finally, if @a dst is @NULL, the function returns the length of the |
| 95 | needed buffer. |
| 96 | |
| 97 | Example of use of this function: |
| 98 | @code |
| 99 | size_t dstLen = conv.ToWChar(NULL, 0, src); |
| 100 | if ( dstLen == wxCONV_FAILED ) |
| 101 | ... handle error ... |
| 102 | wchar_t *dst = new wchar_t[dstLen]; |
| 103 | if ( conv.ToWChar(dst, dstLen, src) == wxCONV_FAILED ) |
| 104 | ... handle error ... |
| 105 | @endcode |
| 106 | |
| 107 | Notice that when passing the explicit source length the output will |
| 108 | @e not be @c NUL terminated if you pass @c strlen(str) as parameter. |
| 109 | Either leave @a srcLen as default @c wxNO_LEN or add one to @c strlen |
| 110 | result if you want the output to be @c NUL terminated. |
| 111 | |
| 112 | @param dst |
| 113 | Pointer to output buffer of the size of at least @a dstLen or @NULL. |
| 114 | @param dstLen |
| 115 | Maximal number of characters to be written to the output buffer if |
| 116 | @a dst is non-@NULL, unused otherwise. |
| 117 | @param src |
| 118 | Point to the source string, must not be @NULL. |
| 119 | @param srcLen |
| 120 | The number of characters of the source string to convert or |
| 121 | @c wxNO_LEN (default parameter) to convert everything up to and |
| 122 | including the terminating @c NUL character(s). |
| 123 | |
| 124 | @return |
| 125 | The number of character written (or which would have been written |
| 126 | if it were non-@NULL) to @a dst or @c wxCONV_FAILED on error. |
| 127 | */ |
| 128 | virtual size_t ToWChar(wchar_t* dst, size_t dstLen, const char* src, |
| 129 | size_t srcLen = wxNO_LEN) const; |
| 130 | |
| 131 | /** |
| 132 | Converts wide character string to multibyte. |
| 133 | |
| 134 | This function has the same semantics as ToWChar() except that it |
| 135 | converts a wide string to multibyte one. As with ToWChar(), it may be |
| 136 | more convenient to use cWC2MB() when working with @c NUL terminated |
| 137 | strings. |
| 138 | |
| 139 | @param dst |
| 140 | Pointer to output buffer of the size of at least @a dstLen or @NULL. |
| 141 | @param dstLen |
| 142 | Maximal number of characters to be written to the output buffer if |
| 143 | @a dst is non-@NULL, unused otherwise. |
| 144 | @param src |
| 145 | Point to the source string, must not be @NULL. |
| 146 | @param srcLen |
| 147 | The number of characters of the source string to convert or |
| 148 | @c wxNO_LEN (default parameter) to convert everything up to and |
| 149 | including the terminating @c NUL character. |
| 150 | |
| 151 | @return |
| 152 | The number of character written (or which would have been written |
| 153 | if it were non-@NULL) to @a dst or @c wxCONV_FAILED on error. |
| 154 | */ |
| 155 | virtual size_t FromWChar(char* dst, size_t dstLen, const wchar_t* src, |
| 156 | size_t srcLen = wxNO_LEN) const; |
| 157 | |
| 158 | /** |
| 159 | Converts from multibyte encoding to Unicode by calling ToWChar() and |
| 160 | allocating a temporary wxWCharBuffer to hold the result. |
| 161 | |
| 162 | This function is a convenient wrapper around ToWChar() as it takes care |
| 163 | of allocating the buffer of the necessary size itself. Its parameters |
| 164 | have the same meaning as for ToWChar(), in particular @a inLen can be |
| 165 | specified explicitly in which case exactly that many characters are |
| 166 | converted and @a outLen receives (if non-@NULL) exactly the |
| 167 | corresponding number of wide characters, whether the last one of them |
| 168 | is @c NUL or not. However if @c inLen is @c wxNO_LEN, then @c outLen |
| 169 | doesn't count the trailing @c NUL even if it is always present in this |
| 170 | case. |
| 171 | |
| 172 | Finally notice that if the conversion fails, the returned buffer is |
| 173 | invalid and @a outLen is set to 0 (and not @c wxCONV_FAILED for |
| 174 | compatibility concerns). |
| 175 | */ |
| 176 | const wxWCharBuffer cMB2WC(const char* in, |
| 177 | size_t inLen, |
| 178 | size_t *outLen) const; |
| 179 | |
| 180 | /** |
| 181 | Converts a char buffer to wide char one. |
| 182 | |
| 183 | This is the most convenient and safest conversion function as you |
| 184 | don't have to deal with the buffer lengths directly. Use it if the |
| 185 | input buffer is known not to be empty or if you are sure that the |
| 186 | conversion is going to succeed -- otherwise, use the overload above to |
| 187 | be able to distinguish between empty input and conversion failure. |
| 188 | |
| 189 | @return |
| 190 | The buffer containing the converted text, empty if the input was |
| 191 | empty or if the conversion failed. |
| 192 | |
| 193 | @since 2.9.1 |
| 194 | */ |
| 195 | const wxWCharBuffer cMB2WC(const wxCharBuffer& buf) const; |
| 196 | |
| 197 | //@{ |
| 198 | /** |
| 199 | Converts from multibyte encoding to the current wxChar type (which |
| 200 | depends on whether wxUSE_UNICODE is set to 1). |
| 201 | |
| 202 | If wxChar is char, it returns the parameter unaltered. If wxChar is |
| 203 | wchar_t, it returns the result in a wxWCharBuffer. The macro wxMB2WXbuf |
| 204 | is defined as the correct return type (without const). |
| 205 | */ |
| 206 | const char* cMB2WX(const char* psz) const; |
| 207 | const wxWCharBuffer cMB2WX(const char* psz) const; |
| 208 | //@} |
| 209 | |
| 210 | /** |
| 211 | Converts from Unicode to multibyte encoding by calling FromWChar() and |
| 212 | allocating a temporary wxCharBuffer to hold the result. |
| 213 | |
| 214 | This function is a convenient wrapper around FromWChar() as it takes |
| 215 | care of allocating the buffer of necessary size itself. |
| 216 | |
| 217 | Its parameters have the same meaning as the corresponding parameters of |
| 218 | FromWChar(), please see the description of cMB2WC() for more details. |
| 219 | */ |
| 220 | const wxCharBuffer cWC2MB(const wchar_t* in, |
| 221 | size_t inLen, |
| 222 | size_t *outLen) const; |
| 223 | |
| 224 | /** |
| 225 | Converts a wide char buffer to char one. |
| 226 | |
| 227 | This is the most convenient and safest conversion function as you |
| 228 | don't have to deal with the buffer lengths directly. Use it if the |
| 229 | input buffer is known not to be empty or if you are sure that the |
| 230 | conversion is going to succeed -- otherwise, use the overload above to |
| 231 | be able to distinguish between empty input and conversion failure. |
| 232 | |
| 233 | @return |
| 234 | The buffer containing the converted text, empty if the input was |
| 235 | empty or if the conversion failed. |
| 236 | |
| 237 | @since 2.9.1 |
| 238 | */ |
| 239 | const wxCharBuffer cWC2MB(const wxWCharBuffer& buf) const; |
| 240 | |
| 241 | //@{ |
| 242 | /** |
| 243 | Converts from Unicode to the current wxChar type. |
| 244 | |
| 245 | If wxChar is wchar_t, it returns the parameter unaltered. If wxChar is |
| 246 | char, it returns the result in a wxCharBuffer. The macro wxWC2WXbuf is |
| 247 | defined as the correct return type (without const). |
| 248 | */ |
| 249 | const wchar_t* cWC2WX(const wchar_t* psz) const; |
| 250 | const wxCharBuffer cWC2WX(const wchar_t* psz) const; |
| 251 | //@} |
| 252 | |
| 253 | //@{ |
| 254 | /** |
| 255 | Converts from the current wxChar type to multibyte encoding. |
| 256 | |
| 257 | If wxChar is char, it returns the parameter unaltered. If wxChar is |
| 258 | wchar_t, it returns the result in a wxCharBuffer. The macro wxWX2MBbuf |
| 259 | is defined as the correct return type (without const). |
| 260 | */ |
| 261 | const char* cWX2MB(const wxChar* psz) const; |
| 262 | const wxCharBuffer cWX2MB(const wxChar* psz) const; |
| 263 | //@} |
| 264 | |
| 265 | //@{ |
| 266 | /** |
| 267 | Converts from the current wxChar type to Unicode. |
| 268 | |
| 269 | If wxChar is wchar_t, it returns the parameter unaltered. If wxChar is |
| 270 | char, it returns the result in a wxWCharBuffer. The macro wxWX2WCbuf is |
| 271 | defined as the correct return type (without const). |
| 272 | */ |
| 273 | const wchar_t* cWX2WC(const wxChar* psz) const; |
| 274 | const wxWCharBuffer cWX2WC(const wxChar* psz) const; |
| 275 | //@} |
| 276 | |
| 277 | /** |
| 278 | @deprecated This function is deprecated, please use ToWChar() instead. |
| 279 | |
| 280 | Converts from a string @a in in multibyte encoding to Unicode putting up to |
| 281 | @a outLen characters into the buffer @e out. |
| 282 | |
| 283 | If @a out is @NULL, only the length of the string which would result |
| 284 | from the conversion is calculated and returned. Note that this is the |
| 285 | length and not size, i.e. the returned value does not include the |
| 286 | trailing @c NUL. But when the function is called with a non-@NULL @a |
| 287 | out buffer, the @a outLen parameter should be one more to allow to |
| 288 | properly @c NUL-terminate the string. |
| 289 | |
| 290 | So to properly use this function you need to write: |
| 291 | @code |
| 292 | size_t lenConv = conv.MB2WC(NULL, in, 0); |
| 293 | if ( lenConv == wxCONV_FAILED ) |
| 294 | ... handle error ... |
| 295 | // allocate 1 more character for the trailing NUL and also pass |
| 296 | // the size of the buffer to the function now |
| 297 | wchar_t *out = new wchar_t[lenConv + 1]; |
| 298 | if ( conv.MB2WC(out, in, lenConv + 1) == wxCONV_FAILED ) |
| 299 | ... handle error ... |
| 300 | @endcode |
| 301 | For this and other reasons, ToWChar() is strongly recommended as a |
| 302 | replacement. |
| 303 | |
| 304 | @param out |
| 305 | The output buffer, may be @NULL if the caller is only |
| 306 | interested in the length of the resulting string |
| 307 | @param in |
| 308 | The NUL-terminated input string, cannot be @NULL |
| 309 | @param outLen |
| 310 | The length of the output buffer but including |
| 311 | NUL, ignored if out is @NULL |
| 312 | |
| 313 | @return The length of the converted string excluding the trailing NUL. |
| 314 | */ |
| 315 | virtual size_t MB2WC(wchar_t* out, const char* in, size_t outLen) const; |
| 316 | |
| 317 | /** |
| 318 | @deprecated This function is deprecated, please use FromWChar() instead. |
| 319 | |
| 320 | Converts from Unicode to multibyte encoding. |
| 321 | The semantics of this function (including the return value meaning) is |
| 322 | the same as for wxMBConv::MB2WC. Notice that when the function is |
| 323 | called with a non-@NULL buffer, the @a n parameter should be the size |
| 324 | of the buffer and so it should take into account the trailing @c NUL, |
| 325 | which might take two or four bytes for some encodings (UTF-16 and |
| 326 | UTF-32) and not one, i.e. GetMBNulLen(). |
| 327 | */ |
| 328 | virtual size_t WC2MB(char* buf, const wchar_t* psz, size_t n) const; |
| 329 | }; |
| 330 | |
| 331 | |
| 332 | /** |
| 333 | @class wxMBConvUTF7 |
| 334 | |
| 335 | This class converts between the UTF-7 encoding and Unicode. |
| 336 | It has one predefined instance, @b wxConvUTF7. |
| 337 | |
| 338 | Notice that, unlike all the other conversion objects, this converter is |
| 339 | stateful, i.e. it remembers its state from the last call to its ToWChar() |
| 340 | or FromWChar() and assumes it is called on the continuation of the same |
| 341 | string when the same method is called again. This assumption is only made |
| 342 | if an explicit length is specified as parameter to these functions as if an |
| 343 | entire @c NUL terminated string is processed the state doesn't need to be |
| 344 | remembered. |
| 345 | |
| 346 | This also means that, unlike the other predefined conversion objects, |
| 347 | @b wxConvUTF7 is @em not thread-safe. |
| 348 | |
| 349 | @library{wxbase} |
| 350 | @category{conv} |
| 351 | |
| 352 | @see wxMBConvUTF8, @ref overview_mbconv |
| 353 | */ |
| 354 | class wxMBConvUTF7 : public wxMBConv |
| 355 | { |
| 356 | }; |
| 357 | |
| 358 | |
| 359 | |
| 360 | /** |
| 361 | @class wxMBConvUTF8 |
| 362 | |
| 363 | This class converts between the UTF-8 encoding and Unicode. |
| 364 | It has one predefined instance, @b wxConvUTF8. |
| 365 | |
| 366 | @library{wxbase} |
| 367 | @category{conv} |
| 368 | |
| 369 | @see wxMBConvUTF7, @ref overview_mbconv |
| 370 | */ |
| 371 | class wxMBConvUTF8 : public wxMBConv |
| 372 | { |
| 373 | }; |
| 374 | |
| 375 | |
| 376 | |
| 377 | /** |
| 378 | @class wxMBConvUTF16 |
| 379 | |
| 380 | This class is used to convert between multibyte encodings and UTF-16 Unicode |
| 381 | encoding (also known as UCS-2). |
| 382 | |
| 383 | Unlike UTF-8 encoding, UTF-16 uses words and not bytes and hence depends |
| 384 | on the byte ordering: big or little endian. Hence this class is provided in |
| 385 | two versions: wxMBConvUTF16LE and wxMBConvUTF16BE and wxMBConvUTF16 itself |
| 386 | is just a typedef for one of them (native for the given platform, e.g. LE |
| 387 | under Windows and BE under Mac). |
| 388 | |
| 389 | @library{wxbase} |
| 390 | @category{conv} |
| 391 | |
| 392 | @see wxMBConvUTF8, wxMBConvUTF32, @ref overview_mbconv |
| 393 | */ |
| 394 | class wxMBConvUTF16 : public wxMBConv |
| 395 | { |
| 396 | }; |
| 397 | |
| 398 | |
| 399 | /** |
| 400 | @class wxMBConvUTF32 |
| 401 | |
| 402 | This class is used to convert between multibyte encodings and UTF-32 |
| 403 | Unicode encoding (also known as UCS-4). |
| 404 | Unlike UTF-8 encoding, UTF-32 uses (double) words and not bytes and hence |
| 405 | depends on the byte ordering: big or little endian. Hence this class is |
| 406 | provided in two versions: wxMBConvUTF32LE and wxMBConvUTF32BE and |
| 407 | wxMBConvUTF32 itself is just a typedef for one of them (native for the |
| 408 | given platform, e.g. LE under Windows and BE under Mac). |
| 409 | |
| 410 | @library{wxbase} |
| 411 | @category{conv} |
| 412 | |
| 413 | @see wxMBConvUTF8, wxMBConvUTF16, @ref overview_mbconv |
| 414 | */ |
| 415 | class wxMBConvUTF32 : public wxMBConv |
| 416 | { |
| 417 | }; |
| 418 | |
| 419 | |
| 420 | |
| 421 | |
| 422 | /** |
| 423 | @class wxCSConv |
| 424 | |
| 425 | This class converts between any character set supported by the system and |
| 426 | Unicode. |
| 427 | |
| 428 | Please notice that this class uses system-provided conversion functions, |
| 429 | e.g. @c MultiByteToWideChar() and @c WideCharToMultiByte() under MSW and @c |
| 430 | iconv(3) under Unix systems and as such may support different encodings and |
| 431 | different encoding names on different platforms (although all relatively |
| 432 | common encodings are supported should be supported everywhere). |
| 433 | |
| 434 | It has one predefined instance, @b wxConvLocal, for the default user |
| 435 | character set. |
| 436 | |
| 437 | @library{wxbase} |
| 438 | @category{conv} |
| 439 | |
| 440 | @see wxMBConv, wxEncodingConverter, @ref overview_mbconv |
| 441 | */ |
| 442 | class wxCSConv : public wxMBConv |
| 443 | { |
| 444 | public: |
| 445 | /** |
| 446 | Constructor. |
| 447 | |
| 448 | You can specify the name of the character set you want to convert |
| 449 | from/to. If the character set name is not recognized, ISO 8859-1 is |
| 450 | used as fall back, use IsOk() to test for this. |
| 451 | |
| 452 | @param charset The name of the encoding, shouldn't be empty. |
| 453 | */ |
| 454 | wxCSConv(const wxString& charset); |
| 455 | |
| 456 | /** |
| 457 | Constructor. |
| 458 | |
| 459 | You can specify an encoding constant for the character set you want to |
| 460 | convert from/to. Use IsOk() after construction to check whether the |
| 461 | encoding is supported by the current system. |
| 462 | |
| 463 | @param encoding Any valid (i.e. not wxFONTENCODING_MAX) font encoding. |
| 464 | */ |
| 465 | wxCSConv(wxFontEncoding encoding); |
| 466 | |
| 467 | /** |
| 468 | Returns @true if the charset (or the encoding) given at constructor is |
| 469 | really available to use. |
| 470 | |
| 471 | Returns @false if ISO 8859-1 will be used instead. |
| 472 | |
| 473 | Note this does not mean that a given string will be correctly |
| 474 | converted. A malformed string may still make conversion functions |
| 475 | return @c wxCONV_FAILED. |
| 476 | |
| 477 | @since 2.8.2 |
| 478 | */ |
| 479 | bool IsOk() const; |
| 480 | }; |
| 481 | |
| 482 | |
| 483 | |
| 484 | /** |
| 485 | @class wxMBConvFile |
| 486 | |
| 487 | This class used to define the class instance @b wxConvFileName, but |
| 488 | nowadays @b wxConvFileName is either of type wxConvLibc (on most platforms) |
| 489 | or wxConvUTF8 (on MacOS X). |
| 490 | |
| 491 | @b wxConvFileName converts filenames between filesystem multibyte encoding |
| 492 | and Unicode. @b wxConvFileName can also be set to a something else at |
| 493 | run-time which is used e.g. by wxGTK to use a class which checks the |
| 494 | environment variable @b G_FILESYSTEM_ENCODING indicating that filenames |
| 495 | should not be interpreted as UTF8 and also for converting invalid UTF8 |
| 496 | characters (e.g. if there is a filename in iso8859_1) to strings with octal |
| 497 | values. |
| 498 | |
| 499 | Since some platforms (such as Win32) use Unicode in the filenames, |
| 500 | and others (such as Unix) use multibyte encodings, this class should only |
| 501 | be used directly if wxMBFILES is defined to 1. A convenience macro, |
| 502 | @c wxFNCONV, is defined to @c wxConvFileName->cWX2MB in this case. You |
| 503 | could use it like this: |
| 504 | |
| 505 | @code |
| 506 | wxChar *name = "rawfile.doc"; |
| 507 | FILE *fil = fopen(wxFNCONV(name), "r"); |
| 508 | @endcode |
| 509 | |
| 510 | (although it would be better to just use wxFopen(name, "r") in this |
| 511 | particular case, you only need to use this class for functions taking file |
| 512 | names not wrapped by wxWidgets.) |
| 513 | |
| 514 | @library{wxbase} |
| 515 | @category{conv} |
| 516 | |
| 517 | @see @ref overview_mbconv |
| 518 | */ |
| 519 | class wxMBConvFile : public wxMBConv |
| 520 | { |
| 521 | public: |
| 522 | }; |