utf8.h

linux下开源浏览器WebKit的源码,市面上的很多商用浏览器都是移植自WebKit

    } else { \
        (i)=utf8_appendCharSafeBody(s, (int32_t)(i), (int32_t)(length), c, &(isError)); \
    } \
}

/**
 * Advance the string offset from one code point boundary to the next.
 * (Post-incrementing iteration.)
 * "Unsafe" macro, assumes well-formed UTF-8.
 *
 * @param s const uint8_t * string
 * @param i string offset
 * @see U8_FWD_1
 * @stable ICU 2.4
 */
#define U8_FWD_1_UNSAFE(s, i) { \
    (i)+=1+U8_COUNT_TRAIL_BYTES((s)[i]); \
}

/**
 * Advance the string offset from one code point boundary to the next.
 * (Post-incrementing iteration.)
 * "Safe" macro, checks for illegal sequences and for string boundaries.
 *
 * @param s const uint8_t * string
 * @param i string offset, i<length
 * @param length string length
 * @see U8_FWD_1_UNSAFE
 * @stable ICU 2.4
 */
#define U8_FWD_1(s, i, length) { \
    uint8_t __b=(s)[(i)++]; \
    if(U8_IS_LEAD(__b)) { \
        uint8_t __count=U8_COUNT_TRAIL_BYTES(__b); \
        if((i)+__count>(length)) { \
            __count=(uint8_t)((length)-(i)); \
        } \
        while(__count>0 && U8_IS_TRAIL((s)[i])) { \
            ++(i); \
            --__count; \
        } \
    } \
}

/**
 * Advance the string offset from one code point boundary to the n-th next one,
 * i.e., move forward by n code points.
 * (Post-incrementing iteration.)
 * "Unsafe" macro, assumes well-formed UTF-8.
 *
 * @param s const uint8_t * string
 * @param i string offset
 * @param n number of code points to skip
 * @see U8_FWD_N
 * @stable ICU 2.4
 */
#define U8_FWD_N_UNSAFE(s, i, n) { \
    int32_t __N=(n); \
    while(__N>0) { \
        U8_FWD_1_UNSAFE(s, i); \
        --__N; \
    } \
}

/**
 * Advance the string offset from one code point boundary to the n-th next one,
 * i.e., move forward by n code points.
 * (Post-incrementing iteration.)
 * "Safe" macro, checks for illegal sequences and for string boundaries.
 *
 * @param s const uint8_t * string
 * @param i string offset, i<length
 * @param length string length
 * @param n number of code points to skip
 * @see U8_FWD_N_UNSAFE
 * @stable ICU 2.4
 */
#define U8_FWD_N(s, i, length, n) { \
    int32_t __N=(n); \
    while(__N>0 && (i)<(length)) { \
        U8_FWD_1(s, i, length); \
        --__N; \
    } \
}

/**
 * Adjust a random-access offset to a code point boundary
 * at the start of a code point.
 * If the offset points to a UTF-8 trail byte,
 * then the offset is moved backward to the corresponding lead byte.
 * Otherwise, it is not modified.
 * "Unsafe" macro, assumes well-formed UTF-8.
 *
 * @param s const uint8_t * string
 * @param i string offset
 * @see U8_SET_CP_START
 * @stable ICU 2.4
 */
#define U8_SET_CP_START_UNSAFE(s, i) { \
    while(U8_IS_TRAIL((s)[i])) { --(i); } \
}

/**
 * Adjust a random-access offset to a code point boundary
 * at the start of a code point.
 * If the offset points to a UTF-8 trail byte,
 * then the offset is moved backward to the corresponding lead byte.
 * Otherwise, it is not modified.
 * "Safe" macro, checks for illegal sequences and for string boundaries.
 *
 * @param s const uint8_t * string
 * @param start starting string offset (usually 0)
 * @param i string offset, start<=i
 * @see U8_SET_CP_START_UNSAFE
 * @stable ICU 2.4
 */
#define U8_SET_CP_START(s, start, i) { \
    if(U8_IS_TRAIL((s)[(i)])) { \
        (i)=utf8_back1SafeBody(s, start, (int32_t)(i)); \
    } \
}

/* definitions with backward iteration -------------------------------------- */

/**
 * Move the string offset from one code point boundary to the previous one
 * and get the code point between them.
 * (Pre-decrementing backward iteration.)
 * "Unsafe" macro, assumes well-formed UTF-8.
 *
 * The input offset may be the same as the string length.
 * If the offset is behind a multi-byte sequence, then the macro will read
 * the whole sequence.
 * If the offset is behind a lead byte, then that itself
 * will be returned as the code point.
 * The result is undefined if the offset is behind an illegal UTF-8 sequence.
 *
 * @param s const uint8_t * string
 * @param i string offset
 * @param c output UChar32 variable
 * @see U8_PREV
 * @stable ICU 2.4
 */
#define U8_PREV_UNSAFE(s, i, c) { \
    (c)=(s)[--(i)]; \
    if(U8_IS_TRAIL(c)) { \
        uint8_t __b, __count=1, __shift=6; \
\
        /* c is a trail byte */ \
        (c)&=0x3f; \
        for(;;) { \
            __b=(s)[--(i)]; \
            if(__b>=0xc0) { \
                U8_MASK_LEAD_BYTE(__b, __count); \
                (c)|=(UChar32)__b<<__shift; \
                break; \
            } else { \
                (c)|=(UChar32)(__b&0x3f)<<__shift; \
                ++__count; \
                __shift+=6; \
            } \
        } \
    } \
}

/**
 * Move the string offset from one code point boundary to the previous one
 * and get the code point between them.
 * (Pre-decrementing backward iteration.)
 * "Safe" macro, checks for illegal sequences and for string boundaries.
 *
 * The input offset may be the same as the string length.
 * If the offset is behind a multi-byte sequence, then the macro will read
 * the whole sequence.
 * If the offset is behind a lead byte, then that itself
 * will be returned as the code point.
 * If the offset is behind an illegal UTF-8 sequence, then c is set to a negative value.
 *
 * @param s const uint8_t * string
 * @param start starting string offset (usually 0)
 * @param i string offset, start<=i
 * @param c output UChar32 variable, set to <0 in case of an error
 * @see U8_PREV_UNSAFE
 * @stable ICU 2.4
 */
#define U8_PREV(s, start, i, c) { \
    (c)=(s)[--(i)]; \
    if((c)>=0x80) { \
        if((c)<=0xbf) { \
            (c)=utf8_prevCharSafeBody(s, start, &(i), c, -1); \
        } else { \
            (c)=U_SENTINEL; \
        } \
    } \
}

/**
 * Move the string offset from one code point boundary to the previous one.
 * (Pre-decrementing backward iteration.)
 * The input offset may be the same as the string length.
 * "Unsafe" macro, assumes well-formed UTF-8.
 *
 * @param s const uint8_t * string
 * @param i string offset
 * @see U8_BACK_1
 * @stable ICU 2.4
 */
#define U8_BACK_1_UNSAFE(s, i) { \
    while(U8_IS_TRAIL((s)[--(i)])) {} \
}

/**
 * Move the string offset from one code point boundary to the previous one.
 * (Pre-decrementing backward iteration.)
 * The input offset may be the same as the string length.
 * "Safe" macro, checks for illegal sequences and for string boundaries.
 *
 * @param s const uint8_t * string
 * @param start starting string offset (usually 0)
 * @param i string offset, start<=i
 * @see U8_BACK_1_UNSAFE
 * @stable ICU 2.4
 */
#define U8_BACK_1(s, start, i) { \
    if(U8_IS_TRAIL((s)[--(i)])) { \
        (i)=utf8_back1SafeBody(s, start, (int32_t)(i)); \
    } \
}

/**
 * Move the string offset from one code point boundary to the n-th one before it,
 * i.e., move backward by n code points.
 * (Pre-decrementing backward iteration.)
 * The input offset may be the same as the string length.
 * "Unsafe" macro, assumes well-formed UTF-8.
 *
 * @param s const uint8_t * string
 * @param i string offset
 * @param n number of code points to skip
 * @see U8_BACK_N
 * @stable ICU 2.4
 */
#define U8_BACK_N_UNSAFE(s, i, n) { \
    int32_t __N=(n); \
    while(__N>0) { \
        U8_BACK_1_UNSAFE(s, i); \
        --__N; \
    } \
}

/**
 * Move the string offset from one code point boundary to the n-th one before it,
 * i.e., move backward by n code points.
 * (Pre-decrementing backward iteration.)
 * The input offset may be the same as the string length.
 * "Safe" macro, checks for illegal sequences and for string boundaries.
 *
 * @param s const uint8_t * string
 * @param start index of the start of the string
 * @param i string offset, i<length
 * @param n number of code points to skip
 * @see U8_BACK_N_UNSAFE
 * @stable ICU 2.4
 */
#define U8_BACK_N(s, start, i, n) { \
    int32_t __N=(n); \
    while(__N>0 && (i)>(start)) { \
        U8_BACK_1(s, start, i); \
        --__N; \
    } \
}

/**
 * Adjust a random-access offset to a code point boundary after a code point.
 * If the offset is behind a partial multi-byte sequence,
 * then the offset is incremented to behind the whole sequence.
 * Otherwise, it is not modified.
 * The input offset may be the same as the string length.
 * "Unsafe" macro, assumes well-formed UTF-8.
 *
 * @param s const uint8_t * string
 * @param i string offset
 * @see U8_SET_CP_LIMIT
 * @stable ICU 2.4
 */
#define U8_SET_CP_LIMIT_UNSAFE(s, i) { \
    U8_BACK_1_UNSAFE(s, i); \
    U8_FWD_1_UNSAFE(s, i); \
}

/**
 * Adjust a random-access offset to a code point boundary after a code point.
 * If the offset is behind a partial multi-byte sequence,
 * then the offset is incremented to behind the whole sequence.
 * Otherwise, it is not modified.
 * The input offset may be the same as the string length.
 * "Safe" macro, checks for illegal sequences and for string boundaries.
 *
 * @param s const uint8_t * string
 * @param start starting string offset (usually 0)
 * @param i string offset, start<=i<=length
 * @param length string length
 * @see U8_SET_CP_LIMIT_UNSAFE
 * @stable ICU 2.4
 */
#define U8_SET_CP_LIMIT(s, start, i, length) { \
    if((start)<(i) && (i)<(length)) { \
        U8_BACK_1(s, start, i); \
        U8_FWD_1(s, i, length); \
    } \
}

#endif