michael@0: /*
michael@0: *******************************************************************************
michael@0: *
michael@0: *   Copyright (C) 1999-2004, International Business Machines
michael@0: *   Corporation and others.  All Rights Reserved.
michael@0: *
michael@0: *******************************************************************************
michael@0: *   file name:  utf.h
michael@0: *   encoding:   US-ASCII
michael@0: *   tab size:   8 (not used)
michael@0: *   indentation:4
michael@0: *
michael@0: *   created on: 1999sep09
michael@0: *   created by: Markus W. Scherer
michael@0: */
michael@0: 
michael@0: #ifndef BASE_THIRD_PARTY_ICU_ICU_UTF_H_
michael@0: #define BASE_THIRD_PARTY_ICU_ICU_UTF_H_
michael@0: 
michael@0: #include "base/basictypes.h"
michael@0: 
michael@0: namespace base_icu {
michael@0: 
michael@0: typedef uint32 UChar32;
michael@0: typedef int8 UBool;
michael@0: 
michael@0: // General ---------------------------------------------------------------------
michael@0: // from utf.h
michael@0: 
michael@0: /**
michael@0:  * This value is intended for sentinel values for APIs that
michael@0:  * (take or) return single code points (UChar32).
michael@0:  * It is outside of the Unicode code point range 0..0x10ffff.
michael@0:  *
michael@0:  * For example, a "done" or "error" value in a new API
michael@0:  * could be indicated with CBU_SENTINEL.
michael@0:  *
michael@0:  * ICU APIs designed before ICU 2.4 usually define service-specific "done"
michael@0:  * values, mostly 0xffff.
michael@0:  * Those may need to be distinguished from
michael@0:  * actual U+ffff text contents by calling functions like
michael@0:  * CharacterIterator::hasNext() or UnicodeString::length().
michael@0:  *
michael@0:  * @return -1
michael@0:  * @see UChar32
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU_SENTINEL (-1)
michael@0: 
michael@0: /**
michael@0:  * Is this code point a Unicode noncharacter?
michael@0:  * @param c 32-bit code point
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU_IS_UNICODE_NONCHAR(c) \
michael@0:     ((c)>=0xfdd0 && \
michael@0:      ((uint32)(c)<=0xfdef || ((c)&0xfffe)==0xfffe) && \
michael@0:      (uint32)(c)<=0x10ffff)
michael@0: 
michael@0: /**
michael@0:  * Is c a Unicode code point value (0..U+10ffff)
michael@0:  * that can be assigned a character?
michael@0:  *
michael@0:  * Code points that are not characters include:
michael@0:  * - single surrogate code points (U+d800..U+dfff, 2048 code points)
michael@0:  * - the last two code points on each plane (U+__fffe and U+__ffff, 34 code points)
michael@0:  * - U+fdd0..U+fdef (new with Unicode 3.1, 32 code points)
michael@0:  * - the highest Unicode code point value is U+10ffff
michael@0:  *
michael@0:  * This means that all code points below U+d800 are character code points,
michael@0:  * and that boundary is tested first for performance.
michael@0:  *
michael@0:  * @param c 32-bit code point
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU_IS_UNICODE_CHAR(c) \
michael@0:     ((uint32)(c)<0xd800 || \
michael@0:         ((uint32)(c)>0xdfff && \
michael@0:          (uint32)(c)<=0x10ffff && \
michael@0:          !CBU_IS_UNICODE_NONCHAR(c)))
michael@0: 
michael@0: /**
michael@0:  * Is this code point a surrogate (U+d800..U+dfff)?
michael@0:  * @param c 32-bit code point
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU_IS_SURROGATE(c) (((c)&0xfffff800)==0xd800)
michael@0: 
michael@0: /**
michael@0:  * Assuming c is a surrogate code point (U_IS_SURROGATE(c)),
michael@0:  * is it a lead surrogate?
michael@0:  * @param c 32-bit code point
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU_IS_SURROGATE_LEAD(c) (((c)&0x400)==0)
michael@0: 
michael@0: 
michael@0: // UTF-8 macros ----------------------------------------------------------------
michael@0: // from utf8.h
michael@0: 
michael@0: extern const uint8 utf8_countTrailBytes[256];
michael@0: 
michael@0: /**
michael@0:  * Count the trail bytes for a UTF-8 lead byte.
michael@0:  * @internal
michael@0:  */
michael@0: #define CBU8_COUNT_TRAIL_BYTES(leadByte) (base_icu::utf8_countTrailBytes[(uint8)leadByte])
michael@0: 
michael@0: /**
michael@0:  * Mask a UTF-8 lead byte, leave only the lower bits that form part of the code point value.
michael@0:  * @internal
michael@0:  */
michael@0: #define CBU8_MASK_LEAD_BYTE(leadByte, countTrailBytes) ((leadByte)&=(1<<(6-(countTrailBytes)))-1)
michael@0: 
michael@0: /**
michael@0:  * Does this code unit (byte) encode a code point by itself (US-ASCII 0..0x7f)?
michael@0:  * @param c 8-bit code unit (byte)
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU8_IS_SINGLE(c) (((c)&0x80)==0)
michael@0: 
michael@0: /**
michael@0:  * Is this code unit (byte) a UTF-8 lead byte?
michael@0:  * @param c 8-bit code unit (byte)
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU8_IS_LEAD(c) ((uint8)((c)-0xc0)<0x3e)
michael@0: 
michael@0: /**
michael@0:  * Is this code unit (byte) a UTF-8 trail byte?
michael@0:  * @param c 8-bit code unit (byte)
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU8_IS_TRAIL(c) (((c)&0xc0)==0x80)
michael@0: 
michael@0: /**
michael@0:  * How many code units (bytes) are used for the UTF-8 encoding
michael@0:  * of this Unicode code point?
michael@0:  * @param c 32-bit code point
michael@0:  * @return 1..4, or 0 if c is a surrogate or not a Unicode code point
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU8_LENGTH(c) \
michael@0:     ((uint32)(c)<=0x7f ? 1 : \
michael@0:         ((uint32)(c)<=0x7ff ? 2 : \
michael@0:             ((uint32)(c)<=0xd7ff ? 3 : \
michael@0:                 ((uint32)(c)<=0xdfff || (uint32)(c)>0x10ffff ? 0 : \
michael@0:                     ((uint32)(c)<=0xffff ? 3 : 4)\
michael@0:                 ) \
michael@0:             ) \
michael@0:         ) \
michael@0:     )
michael@0: 
michael@0: /**
michael@0:  * The maximum number of UTF-8 code units (bytes) per Unicode code point (U+0000..U+10ffff).
michael@0:  * @return 4
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU8_MAX_LENGTH 4
michael@0: 
michael@0: /**
michael@0:  * Function for handling "next code point" with error-checking.
michael@0:  * @internal
michael@0:  */
michael@0: UChar32 utf8_nextCharSafeBody(const uint8 *s, int32 *pi, int32 length, UChar32 c, UBool strict);
michael@0: 
michael@0: /**
michael@0:  * Get a code point from a string at a code point boundary offset,
michael@0:  * and advance the offset to the next code point boundary.
michael@0:  * (Post-incrementing forward iteration.)
michael@0:  * "Safe" macro, checks for illegal sequences and for string boundaries.
michael@0:  *
michael@0:  * The offset may point to the lead byte of a multi-byte sequence,
michael@0:  * in which case the macro will read the whole sequence.
michael@0:  * If the offset points to a trail byte or an illegal UTF-8 sequence, then
michael@0:  * c is set to a negative value.
michael@0:  *
michael@0:  * @param s const uint8 * string
michael@0:  * @param i string offset, i<length
michael@0:  * @param length string length
michael@0:  * @param c output UChar32 variable, set to <0 in case of an error
michael@0:  * @see CBU8_NEXT_UNSAFE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU8_NEXT(s, i, length, c) { \
michael@0:     (c)=(s)[(i)++]; \
michael@0:     if(((uint8)(c))>=0x80) { \
michael@0:         if(CBU8_IS_LEAD(c)) { \
michael@0:             (c)=base_icu::utf8_nextCharSafeBody((const uint8 *)s, &(i), (int32)(length), c, -1); \
michael@0:         } else { \
michael@0:             (c)=CBU_SENTINEL; \
michael@0:         } \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Append a code point to a string, overwriting 1 to 4 bytes.
michael@0:  * The offset points to the current end of the string contents
michael@0:  * and is advanced (post-increment).
michael@0:  * "Unsafe" macro, assumes a valid code point and sufficient space in the string.
michael@0:  * Otherwise, the result is undefined.
michael@0:  *
michael@0:  * @param s const uint8 * string buffer
michael@0:  * @param i string offset
michael@0:  * @param c code point to append
michael@0:  * @see CBU8_APPEND
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU8_APPEND_UNSAFE(s, i, c) { \
michael@0:     if((uint32)(c)<=0x7f) { \
michael@0:         (s)[(i)++]=(uint8)(c); \
michael@0:     } else { \
michael@0:         if((uint32)(c)<=0x7ff) { \
michael@0:             (s)[(i)++]=(uint8)(((c)>>6)|0xc0); \
michael@0:         } else { \
michael@0:             if((uint32)(c)<=0xffff) { \
michael@0:                 (s)[(i)++]=(uint8)(((c)>>12)|0xe0); \
michael@0:             } else { \
michael@0:                 (s)[(i)++]=(uint8)(((c)>>18)|0xf0); \
michael@0:                 (s)[(i)++]=(uint8)((((c)>>12)&0x3f)|0x80); \
michael@0:             } \
michael@0:             (s)[(i)++]=(uint8)((((c)>>6)&0x3f)|0x80); \
michael@0:         } \
michael@0:         (s)[(i)++]=(uint8)(((c)&0x3f)|0x80); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: // UTF-16 macros ---------------------------------------------------------------
michael@0: // from utf16.h
michael@0: 
michael@0: /**
michael@0:  * Does this code unit alone encode a code point (BMP, not a surrogate)?
michael@0:  * @param c 16-bit code unit
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_IS_SINGLE(c) !CBU_IS_SURROGATE(c)
michael@0: 
michael@0: /**
michael@0:  * Is this code unit a lead surrogate (U+d800..U+dbff)?
michael@0:  * @param c 16-bit code unit
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_IS_LEAD(c) (((c)&0xfffffc00)==0xd800)
michael@0: 
michael@0: /**
michael@0:  * Is this code unit a trail surrogate (U+dc00..U+dfff)?
michael@0:  * @param c 16-bit code unit
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_IS_TRAIL(c) (((c)&0xfffffc00)==0xdc00)
michael@0: 
michael@0: /**
michael@0:  * Is this code unit a surrogate (U+d800..U+dfff)?
michael@0:  * @param c 16-bit code unit
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_IS_SURROGATE(c) CBU_IS_SURROGATE(c)
michael@0: 
michael@0: /**
michael@0:  * Assuming c is a surrogate code point (U16_IS_SURROGATE(c)),
michael@0:  * is it a lead surrogate?
michael@0:  * @param c 16-bit code unit
michael@0:  * @return TRUE or FALSE
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_IS_SURROGATE_LEAD(c) (((c)&0x400)==0)
michael@0: 
michael@0: /**
michael@0:  * Helper constant for CBU16_GET_SUPPLEMENTARY.
michael@0:  * @internal
michael@0:  */
michael@0: #define CBU16_SURROGATE_OFFSET ((0xd800<<10UL)+0xdc00-0x10000)
michael@0: 
michael@0: /**
michael@0:  * Get a supplementary code point value (U+10000..U+10ffff)
michael@0:  * from its lead and trail surrogates.
michael@0:  * The result is undefined if the input values are not
michael@0:  * lead and trail surrogates.
michael@0:  *
michael@0:  * @param lead lead surrogate (U+d800..U+dbff)
michael@0:  * @param trail trail surrogate (U+dc00..U+dfff)
michael@0:  * @return supplementary code point (U+10000..U+10ffff)
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_GET_SUPPLEMENTARY(lead, trail) \
michael@0:     (((base_icu::UChar32)(lead)<<10UL)+(base_icu::UChar32)(trail)-CBU16_SURROGATE_OFFSET)
michael@0: 
michael@0: 
michael@0: /**
michael@0:  * Get the lead surrogate (0xd800..0xdbff) for a
michael@0:  * supplementary code point (0x10000..0x10ffff).
michael@0:  * @param supplementary 32-bit code point (U+10000..U+10ffff)
michael@0:  * @return lead surrogate (U+d800..U+dbff) for supplementary
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_LEAD(supplementary) (UChar)(((supplementary)>>10)+0xd7c0)
michael@0: 
michael@0: /**
michael@0:  * Get the trail surrogate (0xdc00..0xdfff) for a
michael@0:  * supplementary code point (0x10000..0x10ffff).
michael@0:  * @param supplementary 32-bit code point (U+10000..U+10ffff)
michael@0:  * @return trail surrogate (U+dc00..U+dfff) for supplementary
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_TRAIL(supplementary) (UChar)(((supplementary)&0x3ff)|0xdc00)
michael@0: 
michael@0: /**
michael@0:  * How many 16-bit code units are used to encode this Unicode code point? (1 or 2)
michael@0:  * The result is not defined if c is not a Unicode code point (U+0000..U+10ffff).
michael@0:  * @param c 32-bit code point
michael@0:  * @return 1 or 2
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_LENGTH(c) ((uint32)(c)<=0xffff ? 1 : 2)
michael@0: 
michael@0: /**
michael@0:  * The maximum number of 16-bit code units per Unicode code point (U+0000..U+10ffff).
michael@0:  * @return 2
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_MAX_LENGTH 2
michael@0: 
michael@0: /**
michael@0:  * Get a code point from a string at a code point boundary offset,
michael@0:  * and advance the offset to the next code point boundary.
michael@0:  * (Post-incrementing forward iteration.)
michael@0:  * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
michael@0:  *
michael@0:  * The offset may point to the lead surrogate unit
michael@0:  * for a supplementary code point, in which case the macro will read
michael@0:  * the following trail surrogate as well.
michael@0:  * If the offset points to a trail surrogate or
michael@0:  * to a single, unpaired lead surrogate, then that itself
michael@0:  * will be returned as the code point.
michael@0:  *
michael@0:  * @param s const UChar * string
michael@0:  * @param i string offset, i<length
michael@0:  * @param length string length
michael@0:  * @param c output UChar32 variable
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_NEXT(s, i, length, c) { \
michael@0:     (c)=(s)[(i)++]; \
michael@0:     if(CBU16_IS_LEAD(c)) { \
michael@0:         uint16 __c2; \
michael@0:         if((i)<(length) && CBU16_IS_TRAIL(__c2=(s)[(i)])) { \
michael@0:             ++(i); \
michael@0:             (c)=CBU16_GET_SUPPLEMENTARY((c), __c2); \
michael@0:         } \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Append a code point to a string, overwriting 1 or 2 code units.
michael@0:  * The offset points to the current end of the string contents
michael@0:  * and is advanced (post-increment).
michael@0:  * "Unsafe" macro, assumes a valid code point and sufficient space in the string.
michael@0:  * Otherwise, the result is undefined.
michael@0:  *
michael@0:  * @param s const UChar * string buffer
michael@0:  * @param i string offset
michael@0:  * @param c code point to append
michael@0:  * @see CBU16_APPEND
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define CBU16_APPEND_UNSAFE(s, i, c) { \
michael@0:     if((uint32)(c)<=0xffff) { \
michael@0:         (s)[(i)++]=(uint16)(c); \
michael@0:     } else { \
michael@0:         (s)[(i)++]=(uint16)(((c)>>10)+0xd7c0); \
michael@0:         (s)[(i)++]=(uint16)(((c)&0x3ff)|0xdc00); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: }  // namesapce base_icu
michael@0: 
michael@0: #endif  // BASE_THIRD_PARTY_ICU_ICU_UTF_H_