ICU 54.1  54.1
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Groups Pages
Macros | Functions | Variables
utf8.h File Reference

C API: 8-bit Unicode handling macros. More...

#include "unicode/umachine.h"
#include "unicode/utf.h"

Go to the source code of this file.

Macros

#define U8_COUNT_TRAIL_BYTES(leadByte)
 Counts the trail bytes for a UTF-8 lead byte.
#define U8_COUNT_TRAIL_BYTES_UNSAFE(leadByte)   (((leadByte)>=0xc0)+((leadByte)>=0xe0)+((leadByte)>=0xf0))
 Counts the trail bytes for a UTF-8 lead byte of a valid UTF-8 sequence.
#define U8_MASK_LEAD_BYTE(leadByte, countTrailBytes)   ((leadByte)&=(1<<(6-(countTrailBytes)))-1)
 Mask a UTF-8 lead byte, leave only the lower bits that form part of the code point value.
#define U8_IS_SINGLE(c)   (((c)&0x80)==0)
 Does this code unit (byte) encode a code point by itself (US-ASCII 0..0x7f)?
#define U8_IS_LEAD(c)   ((uint8_t)((c)-0xc0)<0x3e)
 Is this code unit (byte) a UTF-8 lead byte?
#define U8_IS_TRAIL(c)   (((c)&0xc0)==0x80)
 Is this code unit (byte) a UTF-8 trail byte?
#define U8_LENGTH(c)
 How many code units (bytes) are used for the UTF-8 encoding of this Unicode code point?
#define U8_MAX_LENGTH   4
 The maximum number of UTF-8 code units (bytes) per Unicode code point (U+0000..U+10ffff).
#define U8_GET_UNSAFE(s, i, c)
 Get a code point from a string at a random-access offset, without changing the offset.
#define U8_GET(s, start, i, length, c)
 Get a code point from a string at a random-access offset, without changing the offset.
#define U8_GET_OR_FFFD(s, start, i, length, c)
 Get a code point from a string at a random-access offset, without changing the offset.
#define U8_NEXT_UNSAFE(s, i, c)
 Get a code point from a string at a code point boundary offset, and advance the offset to the next code point boundary.
#define U8_NEXT(s, i, length, c)
 Get a code point from a string at a code point boundary offset, and advance the offset to the next code point boundary.
#define U8_NEXT_OR_FFFD(s, i, length, c)
 Get a code point from a string at a code point boundary offset, and advance the offset to the next code point boundary.
#define U8_APPEND_UNSAFE(s, i, c)
 Append a code point to a string, overwriting 1 to 4 bytes.
#define U8_APPEND(s, i, capacity, c, isError)
 Append a code point to a string, overwriting 1 to 4 bytes.
#define U8_FWD_1_UNSAFE(s, i)
 Advance the string offset from one code point boundary to the next.
#define U8_FWD_1(s, i, length)
 Advance the string offset from one code point boundary to the next.
#define U8_FWD_N_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.
#define U8_FWD_N(s, i, length, n)
 Advance the string offset from one code point boundary to the n-th next one, i.e., move forward by n code points.
#define U8_SET_CP_START_UNSAFE(s, i)
 Adjust a random-access offset to a code point boundary at the start of a code point.
#define U8_SET_CP_START(s, start, i)
 Adjust a random-access offset to a code point boundary at the start of a code point.
#define U8_PREV_UNSAFE(s, i, c)
 Move the string offset from one code point boundary to the previous one and get the code point between them.
#define U8_PREV(s, start, i, c)
 Move the string offset from one code point boundary to the previous one and get the code point between them.
#define U8_PREV_OR_FFFD(s, start, i, c)
 Move the string offset from one code point boundary to the previous one and get the code point between them.
#define U8_BACK_1_UNSAFE(s, i)
 Move the string offset from one code point boundary to the previous one.
#define U8_BACK_1(s, start, i)
 Move the string offset from one code point boundary to the previous one.
#define U8_BACK_N_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.
#define U8_BACK_N(s, start, 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.
#define U8_SET_CP_LIMIT_UNSAFE(s, i)
 Adjust a random-access offset to a code point boundary after a code point.
#define U8_SET_CP_LIMIT(s, start, i, length)
 Adjust a random-access offset to a code point boundary after a code point.

Functions

UChar32 utf8_nextCharSafeBody (const uint8_t *s, int32_t *pi, int32_t length, UChar32 c, UBool strict)
 Function for handling "next code point" with error-checking.
int32_t utf8_appendCharSafeBody (uint8_t *s, int32_t i, int32_t length, UChar32 c, UBool *pIsError)
 Function for handling "append code point" with error-checking.
UChar32 utf8_prevCharSafeBody (const uint8_t *s, int32_t start, int32_t *pi, UChar32 c, UBool strict)
 Function for handling "previous code point" with error-checking.
int32_t utf8_back1SafeBody (const uint8_t *s, int32_t start, int32_t i)
 Function for handling "skip backward one code point" with error-checking.

Variables

U_CFUNC U_IMPORT const uint8_t utf8_countTrailBytes [256]
 Internal array with numbers of trail bytes for any given byte used in lead byte position.

Detailed Description

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).

Usage: 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.

Definition in file utf8.h.

Macro Definition Documentation

#define U8_APPEND (   s,
  i,
  capacity,
  c,
  isError 
)
Value:
{ \
if((uint32_t)(c)<=0x7f) { \
(s)[(i)++]=(uint8_t)(c); \
} else if((uint32_t)(c)<=0x7ff && (i)+1<(capacity)) { \
(s)[(i)++]=(uint8_t)(((c)>>6)|0xc0); \
(s)[(i)++]=(uint8_t)(((c)&0x3f)|0x80); \
} else if((uint32_t)(c)<=0xd7ff && (i)+2<(capacity)) { \
(s)[(i)++]=(uint8_t)(((c)>>12)|0xe0); \
(s)[(i)++]=(uint8_t)((((c)>>6)&0x3f)|0x80); \
(s)[(i)++]=(uint8_t)(((c)&0x3f)|0x80); \
} else { \
(i)=utf8_appendCharSafeBody(s, (i), (capacity), c, &(isError)); \
} \
}

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.

Parameters
sconst uint8_t * string buffer
iint32_t string offset, must be i<capacity
capacityint32_t size of the string buffer
cUChar32 code point to append
isErroroutput UBool set to TRUE if an error occurs, otherwise not modified
See Also
U8_APPEND_UNSAFE
Stable:
ICU 2.4

Definition at line 469 of file utf8.h.

#define U8_APPEND_UNSAFE (   s,
  i,
 
)
Value:
{ \
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); \
} \
}

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.

Parameters
sconst uint8_t * string buffer
istring offset
ccode point to append
See Also
U8_APPEND
Stable:
ICU 2.4

Definition at line 433 of file utf8.h.

#define U8_BACK_1 (   s,
  start,
 
)
Value:
{ \
if(U8_IS_TRAIL((s)[--(i)])) { \
(i)=utf8_back1SafeBody(s, start, (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.

Parameters
sconst uint8_t * string
startint32_t starting string offset (usually 0)
iint32_t string offset, must be start<i
See Also
U8_BACK_1_UNSAFE
Stable:
ICU 2.4

Definition at line 733 of file utf8.h.

#define U8_BACK_1_UNSAFE (   s,
 
)
Value:
{ \
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. "Unsafe" macro, assumes well-formed UTF-8.

Parameters
sconst uint8_t * string
istring offset
See Also
U8_BACK_1
Stable:
ICU 2.4

Definition at line 717 of file utf8.h.

#define U8_BACK_N (   s,
  start,
  i,
 
)
Value:
{ \
int32_t __N=(n); \
while(__N>0 && (i)>(start)) { \
U8_BACK_1(s, start, 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.

Parameters
sconst uint8_t * string
startint32_t index of the start of the string
iint32_t string offset, must be start<i
nnumber of code points to skip
See Also
U8_BACK_N_UNSAFE
Stable:
ICU 2.4

Definition at line 774 of file utf8.h.

#define U8_BACK_N_UNSAFE (   s,
  i,
 
)
Value:
{ \
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. "Unsafe" macro, assumes well-formed UTF-8.

Parameters
sconst uint8_t * string
istring offset
nnumber of code points to skip
See Also
U8_BACK_N
Stable:
ICU 2.4

Definition at line 752 of file utf8.h.

#define U8_COUNT_TRAIL_BYTES (   leadByte)
Value:
((uint8_t)(leadByte)<0xf0 ? \
((uint8_t)(leadByte)>=0xc0)+((uint8_t)(leadByte)>=0xe0) : \
(uint8_t)(leadByte)<0xfe ? 3+((uint8_t)(leadByte)>=0xf8)+((uint8_t)(leadByte)>=0xfc) : 0)

Counts the trail bytes for a UTF-8 lead byte.

Returns 0 for 0..0xbf as well as for 0xfe and 0xff.

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.

Note: Beginning with ICU 50, the implementation uses a multi-condition expression which was shown in 2012 (on x86-64) to compile to fast, branch-free code. leadByte is evaluated multiple times.

The pre-ICU 50 implementation used the exported array utf8_countTrailBytes: #define U8_COUNT_TRAIL_BYTES(leadByte) (utf8_countTrailBytes[leadByte]) leadByte was evaluated exactly once.

Parameters
leadByteThe first byte of a UTF-8 sequence. Must be 0..0xff.
Internal:
Do not use. This API is for internal use only.

Definition at line 80 of file utf8.h.

#define U8_COUNT_TRAIL_BYTES_UNSAFE (   leadByte)    (((leadByte)>=0xc0)+((leadByte)>=0xe0)+((leadByte)>=0xf0))

Counts the trail bytes for a UTF-8 lead byte of a valid UTF-8 sequence.

The maximum supported lead byte is 0xf4 corresponding to U+10FFFF. 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.

Parameters
leadByteThe first byte of a UTF-8 sequence. Must be 0..0xff.
Internal:
Do not use. This API is for internal use only.

Definition at line 96 of file utf8.h.

#define U8_FWD_1 (   s,
  i,
  length 
)
Value:
{ \
uint8_t __b=(uint8_t)(s)[(i)++]; \
if(U8_IS_LEAD(__b)) { \
uint8_t __count=U8_COUNT_TRAIL_BYTES(__b); \
if((i)+__count>(length) && (length)>=0) { \
__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 next.

(Post-incrementing iteration.) "Safe" macro, checks for illegal sequences and for string boundaries.

The length can be negative for a NUL-terminated string.

Parameters
sconst uint8_t * string
iint32_t string offset, must be i<length
lengthint32_t string length
See Also
U8_FWD_1_UNSAFE
Stable:
ICU 2.4

Definition at line 511 of file utf8.h.

#define U8_FWD_1_UNSAFE (   s,
 
)
Value:
{ \
(i)+=1+U8_COUNT_TRAIL_BYTES_UNSAFE((uint8_t)(s)[i]); \
}

Advance the string offset from one code point boundary to the next.

(Post-incrementing iteration.) "Unsafe" macro, assumes well-formed UTF-8.

Parameters
sconst uint8_t * string
istring offset
See Also
U8_FWD_1
Stable:
ICU 2.4

Definition at line 494 of file utf8.h.

#define U8_FWD_N (   s,
  i,
  length,
 
)
Value:
{ \
int32_t __N=(n); \
while(__N>0 && ((i)<(length) || ((length)<0 && (s)[i]!=0))) { \
U8_FWD_1(s, i, length); \
--__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.

Parameters
sconst uint8_t * string
iint32_t string offset, must be i<length
lengthint32_t string length
nnumber of code points to skip
See Also
U8_FWD_N_UNSAFE
Stable:
ICU 2.4

Definition at line 560 of file utf8.h.

#define U8_FWD_N_UNSAFE (   s,
  i,
 
)
Value:
{ \
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.) "Unsafe" macro, assumes well-formed UTF-8.

Parameters
sconst uint8_t * string
istring offset
nnumber of code points to skip
See Also
U8_FWD_N
Stable:
ICU 2.4

Definition at line 537 of file utf8.h.

#define U8_GET (   s,
  start,
  i,
  length,
 
)
Value:
{ \
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 a negative value. Iteration through a string is more efficient with U8_NEXT_UNSAFE or U8_NEXT.

Parameters
sconst uint8_t * string
startint32_t starting string offset
iint32_t string offset, must be start<=i<length
lengthint32_t string length
coutput UChar32 variable, set to <0 in case of an error
See Also
U8_GET_UNSAFE
Stable:
ICU 2.4

Definition at line 250 of file utf8.h.

#define U8_GET_OR_FFFD (   s,
  start,
  i,
  length,
 
)
Value:
{ \
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); \
}

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.

Parameters
sconst uint8_t * string
startint32_t starting string offset
iint32_t string offset, must be start<=i<length
lengthint32_t string length
coutput UChar32 variable, set to U+FFFD in case of an error
See Also
U8_GET
Stable:
ICU 51

Definition at line 281 of file utf8.h.

#define U8_GET_UNSAFE (   s,
  i,
 
)
Value:
{ \
int32_t _u8_get_unsafe_index=(int32_t)(i); \
U8_SET_CP_START_UNSAFE(s, _u8_get_unsafe_index); \
U8_NEXT_UNSAFE(s, _u8_get_unsafe_index, 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 result is undefined if the offset points to an illegal UTF-8 byte sequence. Iteration through a string is more efficient with U8_NEXT_UNSAFE or U8_NEXT.

Parameters
sconst uint8_t * string
istring offset
coutput UChar32 variable
See Also
U8_GET
Stable:
ICU 2.4

Definition at line 223 of file utf8.h.

#define U8_IS_LEAD (   c)    ((uint8_t)((c)-0xc0)<0x3e)

Is this code unit (byte) a UTF-8 lead byte?

Parameters
c8-bit code unit (byte)
Returns
TRUE or FALSE
Stable:
ICU 2.4

Definition at line 172 of file utf8.h.

#define U8_IS_SINGLE (   c)    (((c)&0x80)==0)

Does this code unit (byte) encode a code point by itself (US-ASCII 0..0x7f)?

Parameters
c8-bit code unit (byte)
Returns
TRUE or FALSE
Stable:
ICU 2.4

Definition at line 164 of file utf8.h.

#define U8_IS_TRAIL (   c)    (((c)&0xc0)==0x80)

Is this code unit (byte) a UTF-8 trail byte?

Parameters
c8-bit code unit (byte)
Returns
TRUE or FALSE
Stable:
ICU 2.4

Definition at line 180 of file utf8.h.

#define U8_LENGTH (   c)
Value:
((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)\
) \
) \
) \
)

How many code units (bytes) are used for the UTF-8 encoding of this Unicode code point?

Parameters
c32-bit code point
Returns
1..4, or 0 if c is a surrogate or not a Unicode code point
Stable:
ICU 2.4

Definition at line 189 of file utf8.h.

#define U8_MASK_LEAD_BYTE (   leadByte,
  countTrailBytes 
)    ((leadByte)&=(1<<(6-(countTrailBytes)))-1)

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:
Do not use. This API is for internal use only.

Definition at line 106 of file utf8.h.

#define U8_MAX_LENGTH   4

The maximum number of UTF-8 code units (bytes) per Unicode code point (U+0000..U+10ffff).

Returns
4
Stable:
ICU 2.4

Definition at line 205 of file utf8.h.

#define U8_NEXT (   s,
  i,
  length,
 
)
Value:
{ \
(c)=(uint8_t)(s)[(i)++]; \
if((c)>=0x80) { \
uint8_t __t1, __t2; \
if( /* handle U+1000..U+CFFF inline */ \
(0xe0<(c) && (c)<=0xec) && \
(((i)+1)<(length) || (length)<0) && \
(__t1=(uint8_t)((s)[i]-0x80))<=0x3f && \
(__t2=(uint8_t)((s)[(i)+1]-0x80))<= 0x3f \
) { \
/* no need for (c&0xf) because the upper bits are truncated after <<12 in the cast to (UChar) */ \
(c)=(UChar)(((c)<<12)|(__t1<<6)|__t2); \
(i)+=2; \
} else if( /* handle U+0080..U+07FF inline */ \
((c)<0xe0 && (c)>=0xc2) && \
((i)!=(length)) && \
(__t1=(uint8_t)((s)[i]-0x80))<=0x3f \
) { \
(c)=(((c)&0x1f)<<6)|__t1; \
++(i); \
} else { \
/* function call for "complicated" and error cases */ \
(c)=utf8_nextCharSafeBody((const uint8_t *)s, &(i), (length), c, -1); \
} \
} \
}

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.

Parameters
sconst uint8_t * string
iint32_t string offset, must be i<length
lengthint32_t string length
coutput UChar32 variable, set to <0 in case of an error
See Also
U8_NEXT_UNSAFE
Stable:
ICU 2.4

Definition at line 342 of file utf8.h.

#define U8_NEXT_OR_FFFD (   s,
  i,
  length,
 
)
Value:
{ \
(c)=(uint8_t)(s)[(i)++]; \
if((c)>=0x80) { \
uint8_t __t1, __t2; \
if( /* handle U+1000..U+CFFF inline */ \
(0xe0<(c) && (c)<=0xec) && \
(((i)+1)<(length) || (length)<0) && \
(__t1=(uint8_t)((s)[i]-0x80))<=0x3f && \
(__t2=(uint8_t)((s)[(i)+1]-0x80))<= 0x3f \
) { \
/* no need for (c&0xf) because the upper bits are truncated after <<12 in the cast to (UChar) */ \
(c)=(UChar)(((c)<<12)|(__t1<<6)|__t2); \
(i)+=2; \
} else if( /* handle U+0080..U+07FF inline */ \
((c)<0xe0 && (c)>=0xc2) && \
((i)!=(length)) && \
(__t1=(uint8_t)((s)[i]-0x80))<=0x3f \
) { \
(c)=(((c)&0x1f)<<6)|__t1; \
++(i); \
} else { \
/* function call for "complicated" and error cases */ \
(c)=utf8_nextCharSafeBody((const uint8_t *)s, &(i), (length), c, -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 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.

Parameters
sconst uint8_t * string
iint32_t string offset, must be i<length
lengthint32_t string length
coutput UChar32 variable, set to U+FFFD in case of an error
See Also
U8_NEXT
Stable:
ICU 51

Definition at line 393 of file utf8.h.

#define U8_NEXT_UNSAFE (   s,
  i,
 
)
Value:
{ \
(c)=(uint8_t)(s)[(i)++]; \
if((c)>=0x80) { \
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.) "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.

Parameters
sconst uint8_t * string
istring offset
coutput UChar32 variable
See Also
U8_NEXT
Stable:
ICU 2.4

Definition at line 306 of file utf8.h.

#define U8_PREV (   s,
  start,
  i,
 
)
Value:
{ \
(c)=(uint8_t)(s)[--(i)]; \
if((c)>=0x80) { \
(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 a negative value.

Parameters
sconst uint8_t * string
startint32_t starting string offset (usually 0)
iint32_t string offset, must be start<i
coutput UChar32 variable, set to <0 in case of an error
See Also
U8_PREV_UNSAFE
Stable:
ICU 2.4

Definition at line 668 of file utf8.h.

#define U8_PREV_OR_FFFD (   s,
  start,
  i,
 
)
Value:
{ \
(c)=(uint8_t)(s)[--(i)]; \
if((c)>=0x80) { \
(c)=utf8_prevCharSafeBody((const uint8_t *)s, start, &(i), c, -3); \
} \
}

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.

Parameters
sconst uint8_t * string
startint32_t starting string offset (usually 0)
iint32_t string offset, must be start<i
coutput UChar32 variable, set to U+FFFD in case of an error
See Also
U8_PREV
Stable:
ICU 51

Definition at line 699 of file utf8.h.

#define U8_PREV_UNSAFE (   s,
  i,
 
)
Value:
{ \
(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=(uint8_t)(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.) "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.

Parameters
sconst uint8_t * string
istring offset
coutput UChar32 variable
See Also
U8_PREV
Stable:
ICU 2.4

Definition at line 626 of file utf8.h.

#define U8_SET_CP_LIMIT (   s,
  start,
  i,
  length 
)
Value:
{ \
if((start)<(i) && ((i)<(length) || ((length)<0 && (s)[i]!=0))) { \
U8_BACK_1(s, start, i); \
U8_FWD_1(s, i, length); \
} \
}

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.

Parameters
sconst uint8_t * string
startint32_t starting string offset (usually 0)
iint32_t string offset, must be start<=i<=length
lengthint32_t string length
See Also
U8_SET_CP_LIMIT_UNSAFE
Stable:
ICU 2.4

Definition at line 817 of file utf8.h.

#define U8_SET_CP_LIMIT_UNSAFE (   s,
 
)
Value:
{ \
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. "Unsafe" macro, assumes well-formed UTF-8.

Parameters
sconst uint8_t * string
istring offset
See Also
U8_SET_CP_LIMIT
Stable:
ICU 2.4

Definition at line 795 of file utf8.h.

#define U8_SET_CP_START (   s,
  start,
 
)
Value:
{ \
if(U8_IS_TRAIL((s)[(i)])) { \
(i)=utf8_back1SafeBody(s, start, (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.

Parameters
sconst uint8_t * string
startint32_t starting string offset (usually 0)
iint32_t string offset, must be start<=i
See Also
U8_SET_CP_START_UNSAFE
Stable:
ICU 2.4

Definition at line 599 of file utf8.h.

#define U8_SET_CP_START_UNSAFE (   s,
 
)
Value:
{ \
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. "Unsafe" macro, assumes well-formed UTF-8.

Parameters
sconst uint8_t * string
istring offset
See Also
U8_SET_CP_START
Stable:
ICU 2.4

Definition at line 581 of file utf8.h.

Function Documentation

int32_t utf8_appendCharSafeBody ( uint8_t *  s,
int32_t  i,
int32_t  length,
UChar32  c,
UBool pIsError 
)

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:
Do not use. This API is for internal use only.
int32_t utf8_back1SafeBody ( const uint8_t *  s,
int32_t  start,
int32_t  i 
)

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:
Do not use. This API is for internal use only.
UChar32 utf8_nextCharSafeBody ( const uint8_t *  s,
int32_t *  pi,
int32_t  length,
UChar32  c,
UBool  strict 
)

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:
Do not use. This API is for internal use only.
UChar32 utf8_prevCharSafeBody ( const uint8_t *  s,
int32_t  start,
int32_t *  pi,
UChar32  c,
UBool  strict 
)

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:
Do not use. This API is for internal use only.

Variable Documentation

utf8_countTrailBytes

Internal array with numbers of trail bytes for any given byte used in lead byte position.

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, and should not be hidden when other internal functions are hidden (otherwise public macros would fail to compile).

Internal:
Do not use. This API is for internal use only.

Definition at line 60 of file utf8.h.