FM / unicode

/opt/alt/libicu/usr/include/unicode

Name	Size	Mode	Actions
alphaindex.h	27056	0644	edit dl rm
appendable.h	8632	0644	edit dl rm
basictz.h	9055	0644	edit dl rm
brkiter.h	28195	0644	edit dl rm
bytestream.h	9767	0644	edit dl rm
bytestrie.h	19718	0644	edit dl rm
bytestriebuilder.h	7186	0644	edit dl rm
calendar.h	107581	0644	edit dl rm
caniter.h	7550	0644	edit dl rm
casemap.h	25871	0644	edit dl rm
char16ptr.h	7331	0644	edit dl rm
chariter.h	24573	0644	edit dl rm
choicfmt.h	24418	0644	edit dl rm
coleitr.h	14033	0644	edit dl rm
coll.h	56928	0644	edit dl rm
compactdecimalformat.h	6971	0644	edit dl rm
curramt.h	3782	0644	edit dl rm
currpinf.h	7412	0644	edit dl rm
currunit.h	3606	0644	edit dl rm
datefmt.h	41334	0644	edit dl rm
dbbi.h	1133	0644	edit dl rm
dcfmtsym.h	20584	0644	edit dl rm
decimfmt.h	88478	0644	edit dl rm
docmain.h	6878	0644	edit dl rm
dtfmtsym.h	38548	0644	edit dl rm
dtintrv.h	3848	0644	edit dl rm
dtitvfmt.h	43898	0644	edit dl rm
dtitvinf.h	18903	0644	edit dl rm
dtptngen.h	25626	0644	edit dl rm
dtrule.h	8831	0644	edit dl rm
edits.h	21169	0644	edit dl rm
enumset.h	2130	0644	edit dl rm
errorcode.h	4894	0644	edit dl rm
fieldpos.h	8838	0644	edit dl rm
filteredbrk.h	5364	0644	edit dl rm
fmtable.h	24946	0644	edit dl rm
format.h	12741	0644	edit dl rm
fpositer.h	3047	0644	edit dl rm
gender.h	3347	0644	edit dl rm
gregocal.h	32331	0644	edit dl rm
icudataver.h	1051	0644	edit dl rm
icuplug.h	12166	0644	edit dl rm
idna.h	12938	0644	edit dl rm
listformatter.h	6331	0644	edit dl rm
localpointer.h	17284	0644	edit dl rm
locdspnm.h	7224	0644	edit dl rm
locid.h	46001	0644	edit dl rm
measfmt.h	12298	0644	edit dl rm
measunit.h	39427	0644	edit dl rm
measure.h	4361	0644	edit dl rm
messagepattern.h	34460	0644	edit dl rm
msgfmt.h	44398	0644	edit dl rm
normalizer2.h	34789	0644	edit dl rm
normlzr.h	31476	0644	edit dl rm
nounit.h	2693	0644	edit dl rm
numberformatter.h	89164	0644	edit dl rm
numberrangeformatter.h	29430	0644	edit dl rm
numfmt.h	50322	0644	edit dl rm
numsys.h	6895	0644	edit dl rm
parseerr.h	3155	0644	edit dl rm
parsepos.h	5581	0644	edit dl rm
platform.h	27754	0644	edit dl rm
plurfmt.h	25786	0644	edit dl rm
plurrule.h	17834	0644	edit dl rm
ptypes.h	3577	0644	edit dl rm
putil.h	6487	0644	edit dl rm
rbbi.h	27047	0644	edit dl rm
rbnf.h	50984	0644	edit dl rm
rbtz.h	15838	0644	edit dl rm
regex.h	86943	0644	edit dl rm
region.h	9342	0644	edit dl rm
reldatefmt.h	14812	0644	edit dl rm
rep.h	9726	0644	edit dl rm
resbund.h	18440	0644	edit dl rm
schriter.h	6410	0644	edit dl rm
scientificnumberformatter.h	6474	0644	edit dl rm
search.h	22742	0644	edit dl rm
selfmt.h	14579	0644	edit dl rm
simpleformatter.h	11701	0644	edit dl rm
simpletz.h	46389	0644	edit dl rm
smpdtfmt.h	72442	0644	edit dl rm
sortkey.h	11383	0644	edit dl rm
std_string.h	1015	0644	edit dl rm
strenum.h	10279	0644	edit dl rm
stringoptions.h	5926	0644	edit dl rm
stringpiece.h	6610	0644	edit dl rm
stringtriebuilder.h	15636	0644	edit dl rm
stsearch.h	21801	0644	edit dl rm
symtable.h	4312	0644	edit dl rm
tblcoll.h	37272	0644	edit dl rm
timezone.h	41862	0644	edit dl rm
tmunit.h	3371	0644	edit dl rm
tmutamt.h	4918	0644	edit dl rm
tmutfmt.h	7977	0644	edit dl rm
translit.h	67339	0644	edit dl rm
tzfmt.h	43846	0644	edit dl rm
tznames.h	17190	0644	edit dl rm
tzrule.h	36170	0644	edit dl rm
tztrans.h	6214	0644	edit dl rm
ubidi.h	91711	0644	edit dl rm
ubiditransform.h	12950	0644	edit dl rm
ubrk.h	24547	0644	edit dl rm
ucal.h	56526	0644	edit dl rm
ucasemap.h	15546	0644	edit dl rm
ucat.h	5484	0644	edit dl rm
uchar.h	143314	0644	edit dl rm
ucharstrie.h	21569	0644	edit dl rm
ucharstriebuilder.h	7316	0644	edit dl rm
uchriter.h	13457	0644	edit dl rm
uclean.h	11474	0644	edit dl rm
ucnv.h	85085	0644	edit dl rm
ucnvsel.h	6283	0644	edit dl rm
ucnv_cb.h	6748	0644	edit dl rm
ucnv_err.h	21492	0644	edit dl rm
ucol.h	62864	0644	edit dl rm
ucoleitr.h	9684	0644	edit dl rm
uconfig.h	11949	0644	edit dl rm
ucpmap.h	5709	0644	edit dl rm
ucptrie.h	22824	0644	edit dl rm
ucsdet.h	15018	0644	edit dl rm
ucurr.h	15947	0644	edit dl rm
udat.h	61572	0644	edit dl rm
udata.h	15905	0644	edit dl rm
udateintervalformat.h	6950	0644	edit dl rm
udatpg.h	26664	0644	edit dl rm
udisplaycontext.h	6029	0644	edit dl rm
uenum.h	7970	0644	edit dl rm
ufieldpositer.h	4461	0644	edit dl rm
uformattable.h	11198	0644	edit dl rm
ugender.h	2052	0644	edit dl rm
uidna.h	34169	0644	edit dl rm
uiter.h	23319	0644	edit dl rm
uldnames.h	10702	0644	edit dl rm
ulistformatter.h	5148	0644	edit dl rm
uloc.h	51971	0644	edit dl rm
ulocdata.h	11533	0644	edit dl rm
umachine.h	13509	0644	edit dl rm
umisc.h	1365	0644	edit dl rm
umsg.h	24812	0644	edit dl rm
umutablecptrie.h	8477	0644	edit dl rm
unifilt.h	3716	0644	edit dl rm
unifunct.h	4079	0644	edit dl rm
unimatch.h	6184	0644	edit dl rm
unirepl.h	3403	0644	edit dl rm
uniset.h	65796	0644	edit dl rm
unistr.h	178587	0644	edit dl rm
unorm.h	20938	0644	edit dl rm
unorm2.h	25254	0644	edit dl rm
unum.h	54666	0644	edit dl rm
unumberformatter.h	24335	0644	edit dl rm
unumsys.h	7314	0644	edit dl rm
uobject.h	10958	0644	edit dl rm
upluralrules.h	6880	0644	edit dl rm
uregex.h	73784	0644	edit dl rm
uregion.h	10073	0644	edit dl rm
ureldatefmt.h	12518	0644	edit dl rm
urename.h	130791	0644	edit dl rm
urep.h	5507	0644	edit dl rm
ures.h	37404	0644	edit dl rm
uscript.h	27153	0644	edit dl rm
usearch.h	39038	0644	edit dl rm
uset.h	40958	0644	edit dl rm
usetiter.h	9719	0644	edit dl rm
ushape.h	18432	0644	edit dl rm
uspoof.h	67613	0644	edit dl rm
usprep.h	8331	0644	edit dl rm
ustdio.h	39469	0644	edit dl rm
ustream.h	1844	0644	edit dl rm
ustring.h	74211	0644	edit dl rm
ustringtrie.h	3224	0644	edit dl rm
utext.h	59503	0644	edit dl rm
utf.h	8046	0644	edit dl rm
utf8.h	30734	0644	edit dl rm
utf16.h	22866	0644	edit dl rm
utf32.h	763	0644	edit dl rm
utf_old.h	44827	0644	edit dl rm
utmscale.h	14113	0644	edit dl rm
utrace.h	14219	0644	edit dl rm
utrans.h	26157	0644	edit dl rm
utypes.h	31212	0644	edit dl rm
uvernum.h	6206	0644	edit dl rm
uversion.h	6865	0644	edit dl rm
vtzone.h	20651	0644	edit dl rm

Edit: /opt/alt/libicu/usr/include/unicode/utf8.h (30734B)

// © 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.
*
*******************************************************************************
*   file name:  utf8.h
*   encoding:   UTF-8
*   tab size:   8 (not used)
*   indentation:4
*
*   created on: 1999sep13
*   created by: Markus W. Scherer
*/

/**
 * \file
 * \brief C API: 8-bit Unicode handling macros
 * 
 * This file defines macros to deal with 8-bit Unicode (UTF-8) code units (bytes) and strings.
 *
 * For more information see utf.h and the ICU User Guide Strings chapter
 * (http://userguide.icu-project.org/strings).
 *
 * <em>Usage:</em>
 * ICU coding guidelines for if() statements should be followed when using these macros.
 * Compound statements (curly braces {}) must be used  for if-else-while... 
 * bodies and all macro statements should be terminated with semicolon.
 */

#ifndef __UTF8_H__
#define __UTF8_H__

#include "unicode/umachine.h"
#ifndef __UTF_H__
#   include "unicode/utf.h"
#endif

/* internal definitions ----------------------------------------------------- */

/**
 * Counts the trail bytes for a UTF-8 lead byte.
 * Returns 0 for 0..0xc1 as well as for 0xf5..0xff.
 * leadByte might be evaluated multiple times.
 *
 * This is internal since it is not meant to be called directly by external clients;
 * however it is called by public macros in this file and thus must remain stable.
 *
 * @param leadByte The first byte of a UTF-8 sequence. Must be 0..0xff.
 * @internal
 */
#define U8_COUNT_TRAIL_BYTES(leadByte) \
    (U8_IS_LEAD(leadByte) ? \
        ((uint8_t)(leadByte)>=0xe0)+((uint8_t)(leadByte)>=0xf0)+1 : 0)

/**
 * Counts the trail bytes for a UTF-8 lead byte of a valid UTF-8 sequence.
 * Returns 0 for 0..0xc1. Undefined for 0xf5..0xff.
 * leadByte might be evaluated multiple times.
 *
 * This is internal since it is not meant to be called directly by external clients;
 * however it is called by public macros in this file and thus must remain stable.
 *
 * @param leadByte The first byte of a UTF-8 sequence. Must be 0..0xff.
 * @internal
 */
#define U8_COUNT_TRAIL_BYTES_UNSAFE(leadByte) \
    (((uint8_t)(leadByte)>=0xc2)+((uint8_t)(leadByte)>=0xe0)+((uint8_t)(leadByte)>=0xf0))

/**
 * Mask a UTF-8 lead byte, leave only the lower bits that form part of the code point value.
 *
 * This is internal since it is not meant to be called directly by external clients;
 * however it is called by public macros in this file and thus must remain stable.
 * @internal
 */
#define U8_MASK_LEAD_BYTE(leadByte, countTrailBytes) ((leadByte)&=(1<<(6-(countTrailBytes)))-1)

/**
 * 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 U8_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 U8_IS_VALID_LEAD3_AND_T1(lead, t1) (U8_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 U8_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 U8_IS_VALID_LEAD4_AND_T1(lead, t1) (U8_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 clients;
 * however it is U_STABLE (not U_INTERNAL) since it is called by public macros in this
 * file and thus must remain stable, and should not be hidden when other internal
 * functions are hidden (otherwise public macros would fail to compile).
 * @internal
 */
U_STABLE UChar32 U_EXPORT2
utf8_nextCharSafeBody(const uint8_t *s, int32_t *pi, int32_t length, UChar32 c, UBool strict);

/**
 * Function for handling "append code point" with error-checking.
 *
 * This is internal since it is not meant to be called directly by external clients;
 * however it is U_STABLE (not U_INTERNAL) since it is called by public macros in this
 * file and thus must remain stable, and should not be hidden when other internal
 * functions are hidden (otherwise public macros would fail to compile).
 * @internal
 */
U_STABLE int32_t U_EXPORT2
utf8_appendCharSafeBody(uint8_t *s, int32_t i, int32_t length, UChar32 c, UBool *pIsError);

/**
 * Function for handling "previous code point" with error-checking.
 *
 * This is internal since it is not meant to be called directly by external clients;
 * however it is U_STABLE (not U_INTERNAL) since it is called by public macros in this
 * file and thus must remain stable, and should not be hidden when other internal
 * functions are hidden (otherwise public macros would fail to compile).
 * @internal
 */
U_STABLE UChar32 U_EXPORT2
utf8_prevCharSafeBody(const uint8_t *s, int32_t start, int32_t *pi, UChar32 c, UBool strict);

/**
 * Function for handling "skip backward one code point" with error-checking.
 *
 * This is internal since it is not meant to be called directly by external clients;
 * however it is U_STABLE (not U_INTERNAL) since it is called by public macros in this
 * file and thus must remain stable, and should not be hidden when other internal
 * functions are hidden (otherwise public macros would fail to compile).
 * @internal
 */
U_STABLE int32_t U_EXPORT2
utf8_back1SafeBody(const uint8_t *s, int32_t start, int32_t i);

/* single-code point definitions -------------------------------------------- */

/**
 * 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 U8_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 U8_IS_LEAD(c) ((uint8_t)((c)-0xc2)<=0x32)
// 0x32=0xf4-0xc2

/**
 * 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 U8_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 U8_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 U8_MAX_LENGTH 4

/**
 * Get a code point from a string at a random-access offset,
 * without changing the offset.
 * The offset may point to either the lead byte or one of the trail bytes
 * for a code point, in which case the macro will read all of the bytes
 * for the code point.
 *
 * The length can be negative for a NUL-terminated string.
 *
 * If the offset points to an illegal UTF-8 byte sequence, then
 * c is set to a negative value.
 * Iteration through a string is more efficient with U8_NEXT_UNSAFE or U8_NEXT.
 *
 * @param s const uint8_t * string
 * @param start int32_t starting string offset
 * @param i int32_t string offset, must be start<=i<length
 * @param length int32_t string length
 * @param c output UChar32 variable, set to <0 in case of an error
 * @see U8_GET_UNSAFE
 * @stable ICU 2.4
 */
#define U8_GET(s, start, i, length, c) { \
    int32_t _u8_get_index=(i); \
    U8_SET_CP_START(s, start, _u8_get_index); \
    U8_NEXT(s, _u8_get_index, length, c); \
}

/**
 * Get a code point from a string at a random-access offset,
 * without changing the offset.
 * The offset may point to either the lead byte or one of the trail bytes
 * for a code point, in which case the macro will read all of the bytes
 * for the code point.
 *
 * The length can be negative for a NUL-terminated string.
 *
 * If the offset points to an illegal UTF-8 byte sequence, then
 * c is set to U+FFFD.
 * Iteration through a string is more efficient with U8_NEXT_UNSAFE or U8_NEXT_OR_FFFD.
 *
 * This macro does not distinguish between a real U+FFFD in the text
 * and U+FFFD returned for an ill-formed sequence.
 * Use U8_GET() if that distinction is important.
 *
 * @param s const uint8_t * string
 * @param start int32_t starting string offset
 * @param i int32_t string offset, must be start<=i<length
 * @param length int32_t string length
 * @param c output UChar32 variable, set to U+FFFD in case of an error
 * @see U8_GET
 * @stable ICU 51
 */
#define U8_GET_OR_FFFD(s, start, i, length, c) { \
    int32_t _u8_get_index=(i); \
    U8_SET_CP_START(s, start, _u8_get_index); \
    U8_NEXT_OR_FFFD(s, _u8_get_index, length, c); \
}

/* definitions with forward iteration --------------------------------------- */

/**
 * 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.)
 * "Unsafe" macro, assumes well-formed UTF-8.
 *
 * The offset may point to the lead byte of a multi-byte sequence,
 * in which case the macro will read the whole sequence.
 * The result is undefined if the offset points to a trail byte
 * or an illegal UTF-8 sequence.
 *
 * @param s const uint8_t * string
 * @param i string offset
 * @param c output UChar32 variable
 * @see U8_NEXT
 * @stable ICU 2.4
 */
#define U8_NEXT_UNSAFE(s, i, c) { \
    (c)=(uint8_t)(s)[(i)++]; \
    if(!U8_IS_SINGLE(c)) { \
        if((c)<0xe0) { \
            (c)=(((c)&0x1f)<<6)|((s)[(i)++]&0x3f); \
        } else if((c)<0xf0) { \
            /* no need for (c&0xf) because the upper bits are truncated after <<12 in the cast to (UChar) */ \
            (c)=(UChar)(((c)<<12)|(((s)[i]&0x3f)<<6)|((s)[(i)+1]&0x3f)); \
            (i)+=2; \
        } else { \
            (c)=(((c)&7)<<18)|(((s)[i]&0x3f)<<12)|(((s)[(i)+1]&0x3f)<<6)|((s)[(i)+2]&0x3f); \
            (i)+=3; \
        } \
    } \
}

/**
 * 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<length
 * @param length int32_t string length
 * @param c output UChar32 variable, set to <0 in case of an error
 * @see U8_NEXT_UNSAFE
 * @stable ICU 2.4
 */
#define U8_NEXT(s, i, length, c) U8_INTERNAL_NEXT_OR_SUB(s, i, length, c, U_SENTINEL)

/**
 * 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 U+FFFD.
 *
 * This macro does not distinguish between a real U+FFFD in the text
 * and U+FFFD returned for an ill-formed sequence.
 * Use U8_NEXT() if that distinction is important.
 *
 * @param s const uint8_t * string
 * @param i int32_t string offset, must be i<length
 * @param length int32_t string length
 * @param c output UChar32 variable, set to U+FFFD in case of an error
 * @see U8_NEXT
 * @stable ICU 51
 */
#define U8_NEXT_OR_FFFD(s, i, length, c) U8_INTERNAL_NEXT_OR_SUB(s, i, length, c, 0xfffd)

/** @internal */
#define U8_INTERNAL_NEXT_OR_SUB(s, i, length, c, sub) { \
    (c)=(uint8_t)(s)[(i)++]; \
    if(!U8_IS_SINGLE(c)) { \
        uint8_t __t = 0; \
        if((i)!=(length) && \
            /* fetch/validate/assemble all but last trail byte */ \
            ((c)>=0xe0 ? \
                ((c)<0xf0 ?  /* U+0800..U+FFFF except surrogates */ \
                    U8_LEAD3_T1_BITS[(c)&=0xf]&(1<<((__t=(s)[i])>>5)) && \
                    (__t&=0x3f, 1) \
                :  /* U+10000..U+10FFFF */ \
                    ((c)-=0xf0)<=4 && \
                    U8_LEAD4_T1_BITS[(__t=(s)[i])>>4]&(1<<(c)) && \
                    ((c)=((c)<<6)|(__t&0x3f), ++(i)!=(length)) && \
                    (__t=(s)[i]-0x80)<=0x3f) && \
                /* valid second-to-last trail byte */ \
                ((c)=((c)<<6)|__t, ++(i)!=(length)) \
            :  /* U+0080..U+07FF */ \
                (c)>=0xc2 && ((c)&=0x1f, 1)) && \
            /* last trail byte */ \
            (__t=(s)[i]-0x80)<=0x3f && \
            ((c)=((c)<<6)|__t, ++(i), 1)) { \
        } else { \
            (c)=(sub);  /* ill-formed*/ \
        } \
    } \
}

/**
 * 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 U8_APPEND_UNSAFE(s, i, c) { \
    uint32_t __uc=(c); \
    if(__uc<=0x7f) { \
        (s)[(i)++]=(uint8_t)__uc; \
    } else { \
        if(__uc<=0x7ff) { \
            (s)[(i)++]=(uint8_t)((__uc>>6)|0xc0); \
        } else { \
            if(__uc<=0xffff) { \
                (s)[(i)++]=(uint8_t)((__uc>>12)|0xe0); \
            } else { \
                (s)[(i)++]=(uint8_t)((__uc>>18)|0xf0); \
                (s)[(i)++]=(uint8_t)(((__uc>>12)&0x3f)|0x80); \
            } \
            (s)[(i)++]=(uint8_t)(((__uc>>6)&0x3f)|0x80); \
        } \
        (s)[(i)++]=(uint8_t)((__uc&0x3f)|0x80); \
    } \
}

/**
 * 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).
 * "Safe" macro, checks for a valid code point.
 * If a non-ASCII code point is written, checks for sufficient space in the string.
 * If the code point is not valid or trail bytes do not fit,
 * then isError is set to TRUE.
 *
 * @param s const uint8_t * string buffer
 * @param i int32_t string offset, must be i<capacity
 * @param capacity int32_t size of the string buffer
 * @param c UChar32 code point to append
 * @param isError output UBool set to TRUE if an error occurs, otherwise not modified
 * @see U8_APPEND_UNSAFE
 * @stable ICU 2.4
 */
#define U8_APPEND(s, i, capacity, c, isError) { \
    uint32_t __uc=(c); \
    if(__uc<=0x7f) { \
        (s)[(i)++]=(uint8_t)__uc; \
    } else if(__uc<=0x7ff && (i)+1<(capacity)) { \
        (s)[(i)++]=(uint8_t)((__uc>>6)|0xc0); \
        (s)[(i)++]=(uint8_t)((__uc&0x3f)|0x80); \
    } else if((__uc<=0xd7ff || (0xe000<=__uc && __uc<=0xffff)) && (i)+2<(capacity)) { \
        (s)[(i)++]=(uint8_t)((__uc>>12)|0xe0); \
        (s)[(i)++]=(uint8_t)(((__uc>>6)&0x3f)|0x80); \
        (s)[(i)++]=(uint8_t)((__uc&0x3f)|0x80); \
    } else if(0xffff<__uc && __uc<=0x10ffff && (i)+3<(capacity)) { \
        (s)[(i)++]=(uint8_t)((__uc>>18)|0xf0); \
        (s)[(i)++]=(uint8_t)(((__uc>>12)&0x3f)|0x80); \
        (s)[(i)++]=(uint8_t)(((__uc>>6)&0x3f)|0x80); \
        (s)[(i)++]=(uint8_t)((__uc&0x3f)|0x80); \
    } else { \
        (isError)=TRUE; \
    } \
}

/**
 * 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_UNSAFE((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.
 *
 * The length can be negative for a NUL-terminated string.
 *
 * @param s const uint8_t * string
 * @param i int32_t string offset, must be i<length
 * @param length int32_t 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) && (i)!=(length)) { \
        uint8_t __t1=(s)[i]; \
        if((0xe0<=__b && __b<0xf0)) { \
            if(U8_IS_VALID_LEAD3_AND_T1(__b, __t1) && \
                    ++(i)!=(length) && U8_IS_TRAIL((s)[i])) { \
                ++(i); \
            } \
        } else if(__b<0xe0) { \
            if(U8_IS_TRAIL(__t1)) { \
                ++(i); \
            } \
        } else /* c>=0xf0 */ { \
            if(U8_IS_VALID_LEAD4_AND_T1(__b, __t1) && \
                    ++(i)!=(length) && U8_IS_TRAIL((s)[i]) && \
                    ++(i)!=(length) && U8_IS_TRAIL((s)[i])) { \
                ++(i); \
            } \
        } \
    } \
}

/**
 * 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.
 *
 * The length can be negative for a NUL-terminated string.
 *
 * @param s const uint8_t * string
 * @param i int32_t string offset, must be i<length
 * @param length int32_t 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) || ((length)<0 && (s)[i]!=0))) { \
        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.
 * Unlike U8_TRUNCATE_IF_INCOMPLETE(), this macro always reads s[i].
 *
 * @param s const uint8_t * string
 * @param start int32_t starting string offset (usually 0)
 * @param i int32_t string offset, must be start<=i
 * @see U8_SET_CP_START_UNSAFE
 * @see U8_TRUNCATE_IF_INCOMPLETE
 * @stable ICU 2.4
 */
#define U8_SET_CP_START(s, start, i) { \
    if(U8_IS_TRAIL((s)[(i)])) { \
        (i)=utf8_back1SafeBody(s, start, (i)); \
    } \
}

#ifndef U_HIDE_DRAFT_API
/**
 * If the string ends with a UTF-8 byte sequence that is valid so far
 * but incomplete, then reduce the length of the string to end before
 * the lead byte of that incomplete sequence.
 * For example, if the string ends with E1 80, the length is reduced by 2.
 *
 * In all other cases (the string ends with a complete sequence, or it is not
 * possible for any further trail byte to extend the trailing sequence)
 * the length remains unchanged.
 *
 * Useful for processing text split across multiple buffers
 * (save the incomplete sequence for later)
 * and for optimizing iteration
 * (check for string length only once per character).
 *
 * "Safe" macro, checks for illegal sequences and for string boundaries.
 * Unlike U8_SET_CP_START(), this macro never reads s[length].
 *
 * (In UTF-16, simply check for U16_IS_LEAD(last code unit).)
 *
 * @param s const uint8_t * string
 * @param start int32_t starting string offset (usually 0)
 * @param length int32_t string length (usually start<=length)
 * @see U8_SET_CP_START
 * @draft ICU 61
 */
#define U8_TRUNCATE_IF_INCOMPLETE(s, start, length) \
    if((length)>(start)) { \
        uint8_t __b1=s[(length)-1]; \
        if(U8_IS_SINGLE(__b1)) { \
            /* common ASCII character */ \
        } else if(U8_IS_LEAD(__b1)) { \
            --(length); \
        } else if(U8_IS_TRAIL(__b1) && ((length)-2)>=(start)) { \
            uint8_t __b2=s[(length)-2]; \
            if(0xe0<=__b2 && __b2<=0xf4) { \
                if(__b2<0xf0 ? U8_IS_VALID_LEAD3_AND_T1(__b2, __b1) : \
                        U8_IS_VALID_LEAD4_AND_T1(__b2, __b1)) { \
                    (length)-=2; \
                } \
            } else if(U8_IS_TRAIL(__b2) && ((length)-3)>=(start)) { \
                uint8_t __b3=s[(length)-3]; \
                if(0xf0<=__b3 && __b3<=0xf4 && U8_IS_VALID_LEAD4_AND_T1(__b3, __b2)) { \
                    (length)-=3; \
                } \
            } \
        } \
    }
#endif  // U_HIDE_DRAFT_API

/* 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)=(uint8_t)(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 int32_t starting string offset (usually 0)
 * @param i int32_t string offset, must be 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)=(uint8_t)(s)[--(i)]; \
    if(!U8_IS_SINGLE(c)) { \
        (c)=utf8_prevCharSafeBody((const uint8_t *)s, start, &(i), c, -1); \
    } \
}

/**
 * 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 U+FFFD.
 *
 * This macro does not distinguish between a real U+FFFD in the text
 * and U+FFFD returned for an ill-formed sequence.
 * Use U8_PREV() if that distinction is important.
 *
 * @param s const uint8_t * string
 * @param start int32_t starting string offset (usually 0)
 * @param i int32_t string offset, must be start<i
 * @param c output UChar32 variable, set to U+FFFD in case of an error
 * @see U8_PREV
 * @stable ICU 51
 */
#define U8_PREV_OR_FFFD(s, start, i, c) { \
    (c)=(uint8_t)(s)[--(i)]; \
    if(!U8_IS_SINGLE(c)) { \
        (c)=utf8_prevCharSafeBody((const uint8_t *)s, start, &(i), c, -3); \
    } \
}

/**
 * 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 int32_t starting string offset (usually 0)
 * @param i int32_t string offset, must be 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, (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 int32_t index of the start of the string
 * @param i int32_t string offset, must be start<i
 * @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.
 * "Safe" macro, checks for illegal sequences and for string boundaries.
 *
 * The length can be negative for a NUL-terminated string.
 *
 * @param s const uint8_t * string
 * @param start int32_t starting string offset (usually 0)
 * @param i int32_t string offset, must be start<=i<=length
 * @param length int32_t 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) || (length)<0)) { \
        U8_BACK_1(s, start, i); \
        U8_FWD_1(s, i, length); \
    } \
}

#endif

cmd: