// © 2016 and later: Unicode, Inc. and others. // License & terms of use: http://www.unicode.org/copyright.html /* ****************************************************************************** * * Copyright (C) 1999-2015, International Business Machines * Corporation and others. All Rights Reserved. * ****************************************************************************** */ #ifndef BASE_THIRD_PARTY_ICU_ICU_UTF_H_ #define BASE_THIRD_PARTY_ICU_ICU_UTF_H_ #include namespace base_icu { // source/common/unicode/umachine.h /** The ICU boolean type @stable ICU 2.0 */ typedef int8_t UBool; /** * Define UChar32 as a type for single Unicode code points. * UChar32 is a signed 32-bit integer (same as int32_t). * * The Unicode code point range is 0..0x10ffff. * All other values (negative or >=0x110000) are illegal as Unicode code points. * They may be used as sentinel values to indicate "done", "error" * or similar non-code point conditions. * * Before ICU 2.4 (Jitterbug 2146), UChar32 was defined * to be wchar_t if that is 32 bits wide (wchar_t may be signed or unsigned) * or else to be uint32_t. * That is, the definition of UChar32 was platform-dependent. * * @see U_SENTINEL * @stable ICU 2.4 */ typedef int32_t UChar32; /** * This value is intended for sentinel values for APIs that * (take or) return single code points (UChar32). * It is outside of the Unicode code point range 0..0x10ffff. * * For example, a "done" or "error" value in a new API * could be indicated with U_SENTINEL. * * ICU APIs designed before ICU 2.4 usually define service-specific "done" * values, mostly 0xffff. * Those may need to be distinguished from * actual U+ffff text contents by calling functions like * CharacterIterator::hasNext() or UnicodeString::length(). * * @return -1 * @see UChar32 * @stable ICU 2.4 */ #define CBU_SENTINEL (-1) // source/common/unicode/utf.h /** * Is this code point a Unicode noncharacter? * @param c 32-bit code point * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU_IS_UNICODE_NONCHAR(c) \ ((c) >= 0xfdd0 && ((c) <= 0xfdef || ((c)&0xfffe) == 0xfffe) && \ (c) <= 0x10ffff) /** * Is c a Unicode code point value (0..U+10ffff) * that can be assigned a character? * * Code points that are not characters include: * - single surrogate code points (U+d800..U+dfff, 2048 code points) * - the last two code points on each plane (U+__fffe and U+__ffff, 34 code * points) - U+fdd0..U+fdef (new with Unicode 3.1, 32 code points) - the highest * Unicode code point value is U+10ffff * * This means that all code points below U+d800 are character code points, * and that boundary is tested first for performance. * * @param c 32-bit code point * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU_IS_UNICODE_CHAR(c) \ ((uint32_t)(c) < 0xd800 || \ (0xdfff < (c) && (c) <= 0x10ffff && !CBU_IS_UNICODE_NONCHAR(c))) /** * Is this code point a surrogate (U+d800..U+dfff)? * @param c 32-bit code point * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU_IS_SURROGATE(c) (((c)&0xfffff800) == 0xd800) /** * Assuming c is a surrogate code point (U_IS_SURROGATE(c)), * is it a lead surrogate? * @param c 32-bit code point * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU_IS_SURROGATE_LEAD(c) (((c)&0x400) == 0) // source/common/unicode/utf8.h /** * Internal bit vector for 3-byte UTF-8 validity check, for use in * U8_IS_VALID_LEAD3_AND_T1. Each bit indicates whether one lead byte + first * trail byte pair starts a valid sequence. Lead byte E0..EF bits 3..0 are used * as byte index, first trail byte bits 7..5 are used as bit index into that * byte. * @see U8_IS_VALID_LEAD3_AND_T1 * @internal */ #define CBU8_LEAD3_T1_BITS \ "\x20\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x10\x30\x30" /** * Internal 3-byte UTF-8 validity check. * Non-zero if lead byte E0..EF and first trail byte 00..FF start a valid * sequence. * @internal */ #define CBU8_IS_VALID_LEAD3_AND_T1(lead, t1) \ (CBU8_LEAD3_T1_BITS[(lead)&0xf] & (1 << ((uint8_t)(t1) >> 5))) /** * Internal bit vector for 4-byte UTF-8 validity check, for use in * U8_IS_VALID_LEAD4_AND_T1. Each bit indicates whether one lead byte + first * trail byte pair starts a valid sequence. First trail byte bits 7..4 are used * as byte index, lead byte F0..F4 bits 2..0 are used as bit index into that * byte. * @see U8_IS_VALID_LEAD4_AND_T1 * @internal */ #define CBU8_LEAD4_T1_BITS \ "\x00\x00\x00\x00\x00\x00\x00\x00\x1E\x0F\x0F\x0F\x00\x00\x00\x00" /** * Internal 4-byte UTF-8 validity check. * Non-zero if lead byte F0..F4 and first trail byte 00..FF start a valid * sequence. * @internal */ #define CBU8_IS_VALID_LEAD4_AND_T1(lead, t1) \ (CBU8_LEAD4_T1_BITS[(uint8_t)(t1) >> 4] & (1 << ((lead)&7))) /** * Function for handling "next code point" with error-checking. * * This is internal since it is not meant to be called directly by external clie nts; * however it is U_STABLE (not U_INTERNAL) since it is called by public macros i n this * file and thus must remain stable, and should not be hidden when other interna l * functions are hidden (otherwise public macros would fail to compile). * @internal */ UChar32 utf8_nextCharSafeBody(const uint8_t* s, int32_t* pi, int32_t length, ::base_icu::UChar32 c, ::base_icu::UBool strict); /** * Does this code unit (byte) encode a code point by itself (US-ASCII 0..0x7f)? * @param c 8-bit code unit (byte) * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU8_IS_SINGLE(c) (((c)&0x80) == 0) /** * Is this code unit (byte) a UTF-8 lead byte? (0xC2..0xF4) * @param c 8-bit code unit (byte) * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU8_IS_LEAD(c) ((uint8_t)((c)-0xc2) <= 0x32) /** * Is this code unit (byte) a UTF-8 trail byte? (0x80..0xBF) * @param c 8-bit code unit (byte) * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU8_IS_TRAIL(c) ((int8_t)(c) < -0x40) /** * How many code units (bytes) are used for the UTF-8 encoding * of this Unicode code point? * @param c 32-bit code point * @return 1..4, or 0 if c is a surrogate or not a Unicode code point * @stable ICU 2.4 */ #define CBU8_LENGTH(c) \ ((uint32_t)(c) <= 0x7f \ ? 1 \ : ((uint32_t)(c) <= 0x7ff \ ? 2 \ : ((uint32_t)(c) <= 0xd7ff \ ? 3 \ : ((uint32_t)(c) <= 0xdfff || (uint32_t)(c) > 0x10ffff \ ? 0 \ : ((uint32_t)(c) <= 0xffff ? 3 : 4))))) /** * The maximum number of UTF-8 code units (bytes) per Unicode code point * (U+0000..U+10ffff). * @return 4 * @stable ICU 2.4 */ #define CBU8_MAX_LENGTH 4 /** * Get a code point from a string at a code point boundary offset, * and advance the offset to the next code point boundary. * (Post-incrementing forward iteration.) * "Safe" macro, checks for illegal sequences and for string boundaries. * * The length can be negative for a NUL-terminated string. * * The offset may point to the lead byte of a multi-byte sequence, * in which case the macro will read the whole sequence. * If the offset points to a trail byte or an illegal UTF-8 sequence, then * c is set to a negative value. * * @param s const uint8_t * string * @param i int32_t string offset, must be i= 0xc2) && ((i) != (length)) && \ (__t1 = (s)[i] - 0x80) <= 0x3f) { \ (c) = (((c)&0x1f) << 6) | __t1; \ ++(i); \ } else { \ /* function call for "complicated" and error cases */ \ (c) = ::base_icu::utf8_nextCharSafeBody((const uint8_t*)s, &(i), \ (length), c, -1); \ } \ } \ } /** * Append a code point to a string, overwriting 1 to 4 bytes. * The offset points to the current end of the string contents * and is advanced (post-increment). * "Unsafe" macro, assumes a valid code point and sufficient space in the * string. Otherwise, the result is undefined. * * @param s const uint8_t * string buffer * @param i string offset * @param c code point to append * @see U8_APPEND * @stable ICU 2.4 */ #define CBU8_APPEND_UNSAFE(s, i, c) \ { \ if ((uint32_t)(c) <= 0x7f) { \ (s)[(i)++] = (uint8_t)(c); \ } else { \ if ((uint32_t)(c) <= 0x7ff) { \ (s)[(i)++] = (uint8_t)(((c) >> 6) | 0xc0); \ } else { \ if ((uint32_t)(c) <= 0xffff) { \ (s)[(i)++] = (uint8_t)(((c) >> 12) | 0xe0); \ } else { \ (s)[(i)++] = (uint8_t)(((c) >> 18) | 0xf0); \ (s)[(i)++] = (uint8_t)((((c) >> 12) & 0x3f) | 0x80); \ } \ (s)[(i)++] = (uint8_t)((((c) >> 6) & 0x3f) | 0x80); \ } \ (s)[(i)++] = (uint8_t)(((c)&0x3f) | 0x80); \ } \ } // source/common/unicode/utf16.h /** * Does this code unit alone encode a code point (BMP, not a surrogate)? * @param c 16-bit code unit * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU16_IS_SINGLE(c) !CBU_IS_SURROGATE(c) /** * Is this code unit a lead surrogate (U+d800..U+dbff)? * @param c 16-bit code unit * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU16_IS_LEAD(c) (((c)&0xfffffc00) == 0xd800) /** * Is this code unit a trail surrogate (U+dc00..U+dfff)? * @param c 16-bit code unit * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU16_IS_TRAIL(c) (((c)&0xfffffc00) == 0xdc00) /** * Is this code unit a surrogate (U+d800..U+dfff)? * @param c 16-bit code unit * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU16_IS_SURROGATE(c) CBU_IS_SURROGATE(c) /** * Assuming c is a surrogate code point (U16_IS_SURROGATE(c)), * is it a lead surrogate? * @param c 16-bit code unit * @return TRUE or FALSE * @stable ICU 2.4 */ #define CBU16_IS_SURROGATE_LEAD(c) (((c)&0x400) == 0) /** * Helper constant for U16_GET_SUPPLEMENTARY. * @internal */ #define CBU16_SURROGATE_OFFSET ((0xd800 << 10UL) + 0xdc00 - 0x10000) /** * Get a supplementary code point value (U+10000..U+10ffff) * from its lead and trail surrogates. * The result is undefined if the input values are not * lead and trail surrogates. * * @param lead lead surrogate (U+d800..U+dbff) * @param trail trail surrogate (U+dc00..U+dfff) * @return supplementary code point (U+10000..U+10ffff) * @stable ICU 2.4 */ #define CBU16_GET_SUPPLEMENTARY(lead, trail) \ (((::base_icu::UChar32)(lead) << 10UL) + \ (::base_icu::UChar32)(trail)-CBU16_SURROGATE_OFFSET) /** * Get the lead surrogate (0xd800..0xdbff) for a * supplementary code point (0x10000..0x10ffff). * @param supplementary 32-bit code point (U+10000..U+10ffff) * @return lead surrogate (U+d800..U+dbff) for supplementary * @stable ICU 2.4 */ #define CBU16_LEAD(supplementary) \ (::base_icu::UChar)(((supplementary) >> 10) + 0xd7c0) /** * Get the trail surrogate (0xdc00..0xdfff) for a * supplementary code point (0x10000..0x10ffff). * @param supplementary 32-bit code point (U+10000..U+10ffff) * @return trail surrogate (U+dc00..U+dfff) for supplementary * @stable ICU 2.4 */ #define CBU16_TRAIL(supplementary) \ (::base_icu::UChar)(((supplementary)&0x3ff) | 0xdc00) /** * How many 16-bit code units are used to encode this Unicode code point? (1 or * 2) The result is not defined if c is not a Unicode code point * (U+0000..U+10ffff). * @param c 32-bit code point * @return 1 or 2 * @stable ICU 2.4 */ #define CBU16_LENGTH(c) ((uint32_t)(c) <= 0xffff ? 1 : 2) /** * The maximum number of 16-bit code units per Unicode code point * (U+0000..U+10ffff). * @return 2 * @stable ICU 2.4 */ #define CBU16_MAX_LENGTH 2 /** * Get a code point from a string at a code point boundary offset, * and advance the offset to the next code point boundary. * (Post-incrementing forward iteration.) * "Safe" macro, handles unpaired surrogates and checks for string boundaries. * * The length can be negative for a NUL-terminated string. * * The offset may point to the lead surrogate unit * for a supplementary code point, in which case the macro will read * the following trail surrogate as well. * If the offset points to a trail surrogate or * to a single, unpaired lead surrogate, then c is set to that unpaired * surrogate. * * @param s const UChar * string * @param i string offset, must be i> 10) + 0xd7c0); \ (s)[(i)++] = (uint16_t)(((c)&0x3ff) | 0xdc00); \ } \ } } // namespace base_icu #endif // BASE_THIRD_PARTY_ICU_ICU_UTF_H_