1 // © 2016 and later: Unicode, Inc. and others. 2 // License & terms of use: http://www.unicode.org/copyright.html 3 /* 4 ****************************************************************************** 5 * 6 * Copyright (C) 1999-2015, International Business Machines 7 * Corporation and others. All Rights Reserved. 8 * 9 ****************************************************************************** 10 * file name: umachine.h 11 * encoding: UTF-8 12 * tab size: 8 (not used) 13 * indentation:4 14 * 15 * created on: 1999sep13 16 * created by: Markus W. Scherer 17 * 18 * This file defines basic types and constants for ICU to be 19 * platform-independent. umachine.h and utf.h are included into 20 * utypes.h to provide all the general definitions for ICU. 21 * All of these definitions used to be in utypes.h before 22 * the UTF-handling macros made this unmaintainable. 23 */ 24 25 #ifndef __UMACHINE_H__ 26 #define __UMACHINE_H__ 27 28 29 /** 30 * @addtogroup ICU4C 31 * @{ 32 * \file 33 * \brief Basic types and constants for UTF 34 * 35 * <h2> Basic types and constants for UTF </h2> 36 * This file defines basic types and constants for utf.h to be 37 * platform-independent. umachine.h and utf.h are included into 38 * utypes.h to provide all the general definitions for ICU. 39 * All of these definitions used to be in utypes.h before 40 * the UTF-handling macros made this unmaintainable. 41 * 42 */ 43 /*==========================================================================*/ 44 /* Include platform-dependent definitions */ 45 /* which are contained in the platform-specific file platform.h */ 46 /*==========================================================================*/ 47 48 #include "unicode/ptypes.h" /* platform.h is included in ptypes.h */ 49 50 /* 51 * ANSI C headers: 52 * stddef.h defines wchar_t 53 */ 54 #include <stdbool.h> 55 #include <stddef.h> 56 57 /*==========================================================================*/ 58 /* For C wrappers, we use the symbol U_CAPI. */ 59 /* This works properly if the includer is C or C++. */ 60 /* Functions are declared U_CAPI return-type U_EXPORT2 function-name()... */ 61 /*==========================================================================*/ 62 63 /** 64 * \def U_CFUNC 65 * This is used in a declaration of a library private ICU C function. 66 * \xrefitem stable "Stable" "Stable List" ICU 2.4 67 */ 68 69 /** 70 * \def U_CDECL_BEGIN 71 * This is used to begin a declaration of a library private ICU C API. 72 * \xrefitem stable "Stable" "Stable List" ICU 2.4 73 */ 74 75 /** 76 * \def U_CDECL_END 77 * This is used to end a declaration of a library private ICU C API 78 * \xrefitem stable "Stable" "Stable List" ICU 2.4 79 */ 80 81 #ifdef __cplusplus 82 # define U_CFUNC extern "C" 83 # define U_CDECL_BEGIN extern "C" { 84 # define U_CDECL_END } 85 #else 86 # define U_CFUNC extern 87 # define U_CDECL_BEGIN 88 # define U_CDECL_END 89 #endif 90 91 #ifndef U_ATTRIBUTE_DEPRECATED 92 /** 93 * \def U_ATTRIBUTE_DEPRECATED 94 * This is used for GCC specific attributes 95 * \xrefitem internal "Internal" "Internal List" Do not use. This API is for internal use only. 96 */ 97 #if U_GCC_MAJOR_MINOR >= 302 98 # define U_ATTRIBUTE_DEPRECATED __attribute__ ((deprecated)) 99 /** 100 * \def U_ATTRIBUTE_DEPRECATED 101 * This is used for Visual C++ specific attributes 102 * \xrefitem internal "Internal" "Internal List" Do not use. This API is for internal use only. 103 */ 104 #elif defined(_MSC_VER) && (_MSC_VER >= 1400) 105 # define U_ATTRIBUTE_DEPRECATED __declspec(deprecated) 106 #else 107 # define U_ATTRIBUTE_DEPRECATED 108 #endif 109 #endif 110 111 /** This is used to declare a function as a public ICU C API \xrefitem stable "Stable" "Stable List" ICU 2.0*/ 112 #define U_CAPI U_CFUNC U_EXPORT 113 /** Obsolete/same as U_CAPI; was used to declare a function as a stable public ICU C API*/ 114 #define U_STABLE U_CAPI 115 /** Obsolete/same as U_CAPI; was used to declare a function as a draft public ICU C API */ 116 #define U_DRAFT U_CAPI 117 /** This is used to declare a function as a deprecated public ICU C API */ 118 #define U_DEPRECATED U_CAPI U_ATTRIBUTE_DEPRECATED 119 /** Obsolete/same as U_CAPI; was used to declare a function as an obsolete public ICU C API */ 120 #define U_OBSOLETE U_CAPI 121 /** Obsolete/same as U_CAPI; was used to declare a function as an internal ICU C API */ 122 #define U_INTERNAL U_CAPI 123 124 /** 125 * \def U_OVERRIDE 126 * Defined to the C++11 "override" keyword if available. 127 * Denotes a class or member which is an override of the base class. 128 * May result in an error if it applied to something not an override. 129 * \xrefitem internal "Internal" "Internal List" Do not use. This API is for internal use only. 130 */ 131 #ifndef U_OVERRIDE 132 #define U_OVERRIDE override 133 #endif 134 135 /** 136 * \def U_FINAL 137 * Defined to the C++11 "final" keyword if available. 138 * Denotes a class or member which may not be overridden in subclasses. 139 * May result in an error if subclasses attempt to override. 140 * \xrefitem internal "Internal" "Internal List" Do not use. This API is for internal use only. 141 */ 142 #if !defined(U_FINAL) || defined(U_IN_DOXYGEN) 143 #define U_FINAL final 144 #endif 145 146 // Before ICU 65, function-like, multi-statement ICU macros were just defined as 147 // series of statements wrapped in { } blocks and the caller could choose to 148 // either treat them as if they were actual functions and end the invocation 149 // with a trailing ; creating an empty statement after the block or else omit 150 // this trailing ; using the knowledge that the macro would expand to { }. 151 // 152 // But doing so doesn't work well with macros that look like functions and 153 // compiler warnings about empty statements (ICU-20601) and ICU 65 therefore 154 // switches to the standard solution of wrapping such macros in do { } while. 155 // 156 // This will however break existing code that depends on being able to invoke 157 // these macros without a trailing ; so to be able to remain compatible with 158 // such code the wrapper is itself defined as macros so that it's possible to 159 // build ICU 65 and later with the old macro behaviour, like this: 160 // 161 // export CPPFLAGS='-DUPRV_BLOCK_MACRO_BEGIN="" -DUPRV_BLOCK_MACRO_END=""' 162 // runConfigureICU ... 163 // 164 165 /** 166 * \def UPRV_BLOCK_MACRO_BEGIN 167 * Defined as the "do" keyword by default. 168 * \xrefitem internal "Internal" "Internal List" Do not use. This API is for internal use only. 169 */ 170 #ifndef UPRV_BLOCK_MACRO_BEGIN 171 #define UPRV_BLOCK_MACRO_BEGIN do 172 #endif 173 174 /** 175 * \def UPRV_BLOCK_MACRO_END 176 * Defined as "while (false)" by default. 177 * \xrefitem internal "Internal" "Internal List" Do not use. This API is for internal use only. 178 */ 179 #ifndef UPRV_BLOCK_MACRO_END 180 #define UPRV_BLOCK_MACRO_END while (false) 181 #endif 182 183 /*==========================================================================*/ 184 /* limits for int32_t etc., like in POSIX inttypes.h */ 185 /*==========================================================================*/ 186 187 #ifndef INT8_MIN 188 /** The smallest value an 8 bit signed integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 189 # define INT8_MIN ((int8_t)(-128)) 190 #endif 191 #ifndef INT16_MIN 192 /** The smallest value a 16 bit signed integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 193 # define INT16_MIN ((int16_t)(-32767-1)) 194 #endif 195 #ifndef INT32_MIN 196 /** The smallest value a 32 bit signed integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 197 # define INT32_MIN ((int32_t)(-2147483647-1)) 198 #endif 199 200 #ifndef INT8_MAX 201 /** The largest value an 8 bit signed integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 202 # define INT8_MAX ((int8_t)(127)) 203 #endif 204 #ifndef INT16_MAX 205 /** The largest value a 16 bit signed integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 206 # define INT16_MAX ((int16_t)(32767)) 207 #endif 208 #ifndef INT32_MAX 209 /** The largest value a 32 bit signed integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 210 # define INT32_MAX ((int32_t)(2147483647)) 211 #endif 212 213 #ifndef UINT8_MAX 214 /** The largest value an 8 bit unsigned integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 215 # define UINT8_MAX ((uint8_t)(255U)) 216 #endif 217 #ifndef UINT16_MAX 218 /** The largest value a 16 bit unsigned integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 219 # define UINT16_MAX ((uint16_t)(65535U)) 220 #endif 221 #ifndef UINT32_MAX 222 /** The largest value a 32 bit unsigned integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 223 # define UINT32_MAX ((uint32_t)(4294967295U)) 224 #endif 225 226 #if defined(U_INT64_T_UNAVAILABLE) 227 # error int64_t is required for decimal format and rule-based number format. 228 #else 229 # ifndef INT64_C 230 /** 231 * Provides a platform independent way to specify a signed 64-bit integer constant. 232 * note: may be wrong for some 64 bit platforms - ensure your compiler provides INT64_C 233 * \xrefitem stable "Stable" "Stable List" ICU 2.8 234 */ 235 # define INT64_C(c) c ## LL 236 # endif 237 # ifndef UINT64_C 238 /** 239 * Provides a platform independent way to specify an unsigned 64-bit integer constant. 240 * note: may be wrong for some 64 bit platforms - ensure your compiler provides UINT64_C 241 * \xrefitem stable "Stable" "Stable List" ICU 2.8 242 */ 243 # define UINT64_C(c) c ## ULL 244 # endif 245 # ifndef U_INT64_MIN 246 /** The smallest value a 64 bit signed integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.8 */ 247 # define U_INT64_MIN ((int64_t)(INT64_C(-9223372036854775807)-1)) 248 # endif 249 # ifndef U_INT64_MAX 250 /** The largest value a 64 bit signed integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.8 */ 251 # define U_INT64_MAX ((int64_t)(INT64_C(9223372036854775807))) 252 # endif 253 # ifndef U_UINT64_MAX 254 /** The largest value a 64 bit unsigned integer can hold \xrefitem stable "Stable" "Stable List" ICU 2.8 */ 255 # define U_UINT64_MAX ((uint64_t)(UINT64_C(18446744073709551615))) 256 # endif 257 #endif 258 259 /*==========================================================================*/ 260 /* Boolean data type */ 261 /*==========================================================================*/ 262 263 /** 264 * The ICU boolean type, a signed-byte integer. 265 * ICU-specific for historical reasons: The C and C++ standards used to not define type bool. 266 * Also provides a fixed type definition, as opposed to 267 * type bool whose details (e.g., sizeof) may vary by compiler and between C and C++. 268 * 269 * \xrefitem stable "Stable" "Stable List" ICU 2.0 270 */ 271 typedef int8_t UBool; 272 273 /** 274 * \def U_DEFINE_FALSE_AND_TRUE 275 * Normally turns off defining macros FALSE=0 & TRUE=1 in public ICU headers. 276 * These obsolete macros sometimes break compilation of other code that 277 * defines enum constants or similar with these names. 278 * C++ has long defined bool/false/true. 279 * C99 also added definitions for these, although as macros; see stdbool.h. 280 * 281 * You may transitionally define U_DEFINE_FALSE_AND_TRUE=1 if you need time to migrate code. 282 * 283 * \xrefitem internal "Internal" "Internal List" Do not use. This API is for internal use only. ICU 68 284 */ 285 #ifdef U_DEFINE_FALSE_AND_TRUE 286 // Use the predefined value. 287 #elif defined(U_COMBINED_IMPLEMENTATION) || \ 288 defined(U_COMMON_IMPLEMENTATION) || defined(U_I18N_IMPLEMENTATION) || \ 289 defined(U_IO_IMPLEMENTATION) || defined(U_LAYOUTEX_IMPLEMENTATION) || \ 290 defined(U_TOOLUTIL_IMPLEMENTATION) 291 // Inside ICU: Keep FALSE & TRUE available. 292 # define U_DEFINE_FALSE_AND_TRUE 1 293 #else 294 // Outside ICU: Avoid collision with non-macro definitions of FALSE & TRUE. 295 # define U_DEFINE_FALSE_AND_TRUE 0 296 #endif 297 298 #if U_DEFINE_FALSE_AND_TRUE || defined(U_IN_DOXYGEN) 299 #ifndef TRUE 300 /** 301 * The TRUE value of a UBool. 302 * 303 * \xrefitem deprecated "Deprecated" "Deprecated List" ICU 68 Use standard "true" instead. 304 */ 305 # define TRUE 1 306 #endif 307 #ifndef FALSE 308 /** 309 * The FALSE value of a UBool. 310 * 311 * \xrefitem deprecated "Deprecated" "Deprecated List" ICU 68 Use standard "false" instead. 312 */ 313 # define FALSE 0 314 #endif 315 #endif // U_DEFINE_FALSE_AND_TRUE 316 317 /*==========================================================================*/ 318 /* Unicode data types */ 319 /*==========================================================================*/ 320 321 /* wchar_t-related definitions -------------------------------------------- */ 322 323 /* 324 * \def U_WCHAR_IS_UTF16 325 * Defined if wchar_t uses UTF-16. 326 * 327 * \xrefitem stable "Stable" "Stable List" ICU 2.0 328 */ 329 /* 330 * \def U_WCHAR_IS_UTF32 331 * Defined if wchar_t uses UTF-32. 332 * 333 * \xrefitem stable "Stable" "Stable List" ICU 2.0 334 */ 335 #if !defined(U_WCHAR_IS_UTF16) && !defined(U_WCHAR_IS_UTF32) 336 # ifdef __STDC_ISO_10646__ 337 # if (U_SIZEOF_WCHAR_T==2) 338 # define U_WCHAR_IS_UTF16 339 # elif (U_SIZEOF_WCHAR_T==4) 340 # define U_WCHAR_IS_UTF32 341 # endif 342 # elif defined __UCS2__ 343 # if (U_PF_OS390 <= U_PLATFORM && U_PLATFORM <= U_PF_OS400) && (U_SIZEOF_WCHAR_T==2) 344 # define U_WCHAR_IS_UTF16 345 # endif 346 # elif defined(__UCS4__) || (U_PLATFORM == U_PF_OS400 && defined(__UTF32__)) 347 # if (U_SIZEOF_WCHAR_T==4) 348 # define U_WCHAR_IS_UTF32 349 # endif 350 # elif U_PLATFORM_IS_DARWIN_BASED || (U_SIZEOF_WCHAR_T==4 && U_PLATFORM_IS_LINUX_BASED) 351 # define U_WCHAR_IS_UTF32 352 # elif U_PLATFORM_HAS_WIN32_API 353 # define U_WCHAR_IS_UTF16 354 # endif 355 #endif 356 357 /* UChar and UChar32 definitions -------------------------------------------- */ 358 359 /** Number of bytes in a UChar. \xrefitem stable "Stable" "Stable List" ICU 2.0 */ 360 #define U_SIZEOF_UCHAR 2 361 362 /** 363 * \def U_CHAR16_IS_TYPEDEF 364 * If 1, then char16_t is a typedef and not a real type (yet) 365 * \xrefitem internal "Internal" "Internal List" Do not use. This API is for internal use only. 366 */ 367 #if (U_PLATFORM == U_PF_AIX) && defined(__cplusplus) &&(U_CPLUSPLUS_VERSION < 11) 368 // for AIX, uchar.h needs to be included 369 # include <uchar.h> 370 # define U_CHAR16_IS_TYPEDEF 1 371 #elif defined(_MSC_VER) && (_MSC_VER < 1900) 372 // Versions of Visual Studio/MSVC below 2015 do not support char16_t as a real type, 373 // and instead use a typedef. https://msdn.microsoft.com/library/bb531344.aspx 374 # define U_CHAR16_IS_TYPEDEF 1 375 #else 376 # define U_CHAR16_IS_TYPEDEF 0 377 #endif 378 379 380 /** 381 * \var UChar 382 * 383 * The base type for UTF-16 code units and pointers. 384 * Unsigned 16-bit integer. 385 * Starting with ICU 59, C++ API uses char16_t directly, while C API continues to use UChar. 386 * 387 * UChar is configurable by defining the macro UCHAR_TYPE 388 * on the preprocessor or compiler command line: 389 * -DUCHAR_TYPE=uint16_t or -DUCHAR_TYPE=wchar_t (if U_SIZEOF_WCHAR_T==2) etc. 390 * (The UCHAR_TYPE can also be \#defined earlier in this file, for outside the ICU library code.) 391 * This is for transitional use from application code that uses uint16_t or wchar_t for UTF-16. 392 * 393 * The default is UChar=char16_t. 394 * 395 * C++11 defines char16_t as bit-compatible with uint16_t, but as a distinct type. 396 * 397 * In C, char16_t is a simple typedef of uint_least16_t. 398 * ICU requires uint_least16_t=uint16_t for data memory mapping. 399 * On macOS, char16_t is not available because the uchar.h standard header is missing. 400 * 401 * \xrefitem stable "Stable" "Stable List" ICU 4.4 402 */ 403 404 #if 1 405 // #if 1 is normal. UChar defaults to char16_t in C++. 406 // For configuration testing of UChar=uint16_t temporarily change this to #if 0. 407 // The intltest Makefile #defines UCHAR_TYPE=char16_t, 408 // so we only #define it to uint16_t if it is undefined so far. 409 #elif !defined(UCHAR_TYPE) 410 # define UCHAR_TYPE uint16_t 411 #endif 412 413 #if defined(U_COMBINED_IMPLEMENTATION) || defined(U_COMMON_IMPLEMENTATION) || \ 414 defined(U_I18N_IMPLEMENTATION) || defined(U_IO_IMPLEMENTATION) 415 // Inside the ICU library code, never configurable. 416 typedef char16_t UChar; 417 #elif defined(UCHAR_TYPE) 418 typedef UCHAR_TYPE UChar; 419 #elif (U_CPLUSPLUS_VERSION >= 11) 420 typedef char16_t UChar; 421 #else 422 typedef uint16_t UChar; 423 #endif 424 425 /** 426 * \var OldUChar 427 * Default ICU 58 definition of UChar. 428 * A base type for UTF-16 code units and pointers. 429 * Unsigned 16-bit integer. 430 * 431 * Define OldUChar to be wchar_t if that is 16 bits wide. 432 * If wchar_t is not 16 bits wide, then define UChar to be uint16_t. 433 * 434 * This makes the definition of OldUChar platform-dependent 435 * but allows direct string type compatibility with platforms with 436 * 16-bit wchar_t types. 437 * 438 * This is how UChar was defined in ICU 58, for transition convenience. 439 * Exception: ICU 58 UChar was defined to UCHAR_TYPE if that macro was defined. 440 * The current UChar responds to UCHAR_TYPE but OldUChar does not. 441 * 442 * \xrefitem stable "Stable" "Stable List" ICU 59 443 */ 444 #if U_SIZEOF_WCHAR_T==2 445 typedef wchar_t OldUChar; 446 #elif defined(__CHAR16_TYPE__) 447 typedef __CHAR16_TYPE__ OldUChar; 448 #else 449 typedef uint16_t OldUChar; 450 #endif 451 452 /** 453 * Define UChar32 as a type for single Unicode code points. 454 * UChar32 is a signed 32-bit integer (same as int32_t). 455 * 456 * The Unicode code point range is 0..0x10ffff. 457 * All other values (negative or >=0x110000) are illegal as Unicode code points. 458 * They may be used as sentinel values to indicate "done", "error" 459 * or similar non-code point conditions. 460 * 461 * Before ICU 2.4 (Jitterbug 2146), UChar32 was defined 462 * to be wchar_t if that is 32 bits wide (wchar_t may be signed or unsigned) 463 * or else to be uint32_t. 464 * That is, the definition of UChar32 was platform-dependent. 465 * 466 * @see U_SENTINEL 467 * \xrefitem stable "Stable" "Stable List" ICU 2.4 468 */ 469 typedef int32_t UChar32; 470 471 /** 472 * This value is intended for sentinel values for APIs that 473 * (take or) return single code points (UChar32). 474 * It is outside of the Unicode code point range 0..0x10ffff. 475 * 476 * For example, a "done" or "error" value in a new API 477 * could be indicated with U_SENTINEL. 478 * 479 * ICU APIs designed before ICU 2.4 usually define service-specific "done" 480 * values, mostly 0xffff. 481 * Those may need to be distinguished from 482 * actual U+ffff text contents by calling functions like 483 * CharacterIterator::hasNext() or UnicodeString::length(). 484 * 485 * @return -1 486 * @see UChar32 487 * \xrefitem stable "Stable" "Stable List" ICU 2.4 488 */ 489 #define U_SENTINEL (-1) 490 491 #include "unicode/urename.h" 492 493 #endif 494 495 /** @} */ // addtogroup 496