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