1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/unicode/utf16.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,623 @@
1.4 +/*
1.5 +*******************************************************************************
1.6 +*
1.7 +* Copyright (C) 1999-2012, International Business Machines
1.8 +* Corporation and others. All Rights Reserved.
1.9 +*
1.10 +*******************************************************************************
1.11 +* file name: utf16.h
1.12 +* encoding: US-ASCII
1.13 +* tab size: 8 (not used)
1.14 +* indentation:4
1.15 +*
1.16 +* created on: 1999sep09
1.17 +* created by: Markus W. Scherer
1.18 +*/
1.19 +
1.20 +/**
1.21 + * \file
1.22 + * \brief C API: 16-bit Unicode handling macros
1.23 + *
1.24 + * This file defines macros to deal with 16-bit Unicode (UTF-16) code units and strings.
1.25 + *
1.26 + * For more information see utf.h and the ICU User Guide Strings chapter
1.27 + * (http://userguide.icu-project.org/strings).
1.28 + *
1.29 + * <em>Usage:</em>
1.30 + * ICU coding guidelines for if() statements should be followed when using these macros.
1.31 + * Compound statements (curly braces {}) must be used for if-else-while...
1.32 + * bodies and all macro statements should be terminated with semicolon.
1.33 + */
1.34 +
1.35 +#ifndef __UTF16_H__
1.36 +#define __UTF16_H__
1.37 +
1.38 +#include "unicode/umachine.h"
1.39 +#ifndef __UTF_H__
1.40 +# include "unicode/utf.h"
1.41 +#endif
1.42 +
1.43 +/* single-code point definitions -------------------------------------------- */
1.44 +
1.45 +/**
1.46 + * Does this code unit alone encode a code point (BMP, not a surrogate)?
1.47 + * @param c 16-bit code unit
1.48 + * @return TRUE or FALSE
1.49 + * @stable ICU 2.4
1.50 + */
1.51 +#define U16_IS_SINGLE(c) !U_IS_SURROGATE(c)
1.52 +
1.53 +/**
1.54 + * Is this code unit a lead surrogate (U+d800..U+dbff)?
1.55 + * @param c 16-bit code unit
1.56 + * @return TRUE or FALSE
1.57 + * @stable ICU 2.4
1.58 + */
1.59 +#define U16_IS_LEAD(c) (((c)&0xfffffc00)==0xd800)
1.60 +
1.61 +/**
1.62 + * Is this code unit a trail surrogate (U+dc00..U+dfff)?
1.63 + * @param c 16-bit code unit
1.64 + * @return TRUE or FALSE
1.65 + * @stable ICU 2.4
1.66 + */
1.67 +#define U16_IS_TRAIL(c) (((c)&0xfffffc00)==0xdc00)
1.68 +
1.69 +/**
1.70 + * Is this code unit a surrogate (U+d800..U+dfff)?
1.71 + * @param c 16-bit code unit
1.72 + * @return TRUE or FALSE
1.73 + * @stable ICU 2.4
1.74 + */
1.75 +#define U16_IS_SURROGATE(c) U_IS_SURROGATE(c)
1.76 +
1.77 +/**
1.78 + * Assuming c is a surrogate code point (U16_IS_SURROGATE(c)),
1.79 + * is it a lead surrogate?
1.80 + * @param c 16-bit code unit
1.81 + * @return TRUE or FALSE
1.82 + * @stable ICU 2.4
1.83 + */
1.84 +#define U16_IS_SURROGATE_LEAD(c) (((c)&0x400)==0)
1.85 +
1.86 +/**
1.87 + * Assuming c is a surrogate code point (U16_IS_SURROGATE(c)),
1.88 + * is it a trail surrogate?
1.89 + * @param c 16-bit code unit
1.90 + * @return TRUE or FALSE
1.91 + * @stable ICU 4.2
1.92 + */
1.93 +#define U16_IS_SURROGATE_TRAIL(c) (((c)&0x400)!=0)
1.94 +
1.95 +/**
1.96 + * Helper constant for U16_GET_SUPPLEMENTARY.
1.97 + * @internal
1.98 + */
1.99 +#define U16_SURROGATE_OFFSET ((0xd800<<10UL)+0xdc00-0x10000)
1.100 +
1.101 +/**
1.102 + * Get a supplementary code point value (U+10000..U+10ffff)
1.103 + * from its lead and trail surrogates.
1.104 + * The result is undefined if the input values are not
1.105 + * lead and trail surrogates.
1.106 + *
1.107 + * @param lead lead surrogate (U+d800..U+dbff)
1.108 + * @param trail trail surrogate (U+dc00..U+dfff)
1.109 + * @return supplementary code point (U+10000..U+10ffff)
1.110 + * @stable ICU 2.4
1.111 + */
1.112 +#define U16_GET_SUPPLEMENTARY(lead, trail) \
1.113 + (((UChar32)(lead)<<10UL)+(UChar32)(trail)-U16_SURROGATE_OFFSET)
1.114 +
1.115 +
1.116 +/**
1.117 + * Get the lead surrogate (0xd800..0xdbff) for a
1.118 + * supplementary code point (0x10000..0x10ffff).
1.119 + * @param supplementary 32-bit code point (U+10000..U+10ffff)
1.120 + * @return lead surrogate (U+d800..U+dbff) for supplementary
1.121 + * @stable ICU 2.4
1.122 + */
1.123 +#define U16_LEAD(supplementary) (UChar)(((supplementary)>>10)+0xd7c0)
1.124 +
1.125 +/**
1.126 + * Get the trail surrogate (0xdc00..0xdfff) for a
1.127 + * supplementary code point (0x10000..0x10ffff).
1.128 + * @param supplementary 32-bit code point (U+10000..U+10ffff)
1.129 + * @return trail surrogate (U+dc00..U+dfff) for supplementary
1.130 + * @stable ICU 2.4
1.131 + */
1.132 +#define U16_TRAIL(supplementary) (UChar)(((supplementary)&0x3ff)|0xdc00)
1.133 +
1.134 +/**
1.135 + * How many 16-bit code units are used to encode this Unicode code point? (1 or 2)
1.136 + * The result is not defined if c is not a Unicode code point (U+0000..U+10ffff).
1.137 + * @param c 32-bit code point
1.138 + * @return 1 or 2
1.139 + * @stable ICU 2.4
1.140 + */
1.141 +#define U16_LENGTH(c) ((uint32_t)(c)<=0xffff ? 1 : 2)
1.142 +
1.143 +/**
1.144 + * The maximum number of 16-bit code units per Unicode code point (U+0000..U+10ffff).
1.145 + * @return 2
1.146 + * @stable ICU 2.4
1.147 + */
1.148 +#define U16_MAX_LENGTH 2
1.149 +
1.150 +/**
1.151 + * Get a code point from a string at a random-access offset,
1.152 + * without changing the offset.
1.153 + * "Unsafe" macro, assumes well-formed UTF-16.
1.154 + *
1.155 + * The offset may point to either the lead or trail surrogate unit
1.156 + * for a supplementary code point, in which case the macro will read
1.157 + * the adjacent matching surrogate as well.
1.158 + * The result is undefined if the offset points to a single, unpaired surrogate.
1.159 + * Iteration through a string is more efficient with U16_NEXT_UNSAFE or U16_NEXT.
1.160 + *
1.161 + * @param s const UChar * string
1.162 + * @param i string offset
1.163 + * @param c output UChar32 variable
1.164 + * @see U16_GET
1.165 + * @stable ICU 2.4
1.166 + */
1.167 +#define U16_GET_UNSAFE(s, i, c) { \
1.168 + (c)=(s)[i]; \
1.169 + if(U16_IS_SURROGATE(c)) { \
1.170 + if(U16_IS_SURROGATE_LEAD(c)) { \
1.171 + (c)=U16_GET_SUPPLEMENTARY((c), (s)[(i)+1]); \
1.172 + } else { \
1.173 + (c)=U16_GET_SUPPLEMENTARY((s)[(i)-1], (c)); \
1.174 + } \
1.175 + } \
1.176 +}
1.177 +
1.178 +/**
1.179 + * Get a code point from a string at a random-access offset,
1.180 + * without changing the offset.
1.181 + * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
1.182 + *
1.183 + * The offset may point to either the lead or trail surrogate unit
1.184 + * for a supplementary code point, in which case the macro will read
1.185 + * the adjacent matching surrogate as well.
1.186 + *
1.187 + * The length can be negative for a NUL-terminated string.
1.188 + *
1.189 + * If the offset points to a single, unpaired surrogate, then that itself
1.190 + * will be returned as the code point.
1.191 + * Iteration through a string is more efficient with U16_NEXT_UNSAFE or U16_NEXT.
1.192 + *
1.193 + * @param s const UChar * string
1.194 + * @param start starting string offset (usually 0)
1.195 + * @param i string offset, must be start<=i<length
1.196 + * @param length string length
1.197 + * @param c output UChar32 variable
1.198 + * @see U16_GET_UNSAFE
1.199 + * @stable ICU 2.4
1.200 + */
1.201 +#define U16_GET(s, start, i, length, c) { \
1.202 + (c)=(s)[i]; \
1.203 + if(U16_IS_SURROGATE(c)) { \
1.204 + uint16_t __c2; \
1.205 + if(U16_IS_SURROGATE_LEAD(c)) { \
1.206 + if((i)+1!=(length) && U16_IS_TRAIL(__c2=(s)[(i)+1])) { \
1.207 + (c)=U16_GET_SUPPLEMENTARY((c), __c2); \
1.208 + } \
1.209 + } else { \
1.210 + if((i)>(start) && U16_IS_LEAD(__c2=(s)[(i)-1])) { \
1.211 + (c)=U16_GET_SUPPLEMENTARY(__c2, (c)); \
1.212 + } \
1.213 + } \
1.214 + } \
1.215 +}
1.216 +
1.217 +/* definitions with forward iteration --------------------------------------- */
1.218 +
1.219 +/**
1.220 + * Get a code point from a string at a code point boundary offset,
1.221 + * and advance the offset to the next code point boundary.
1.222 + * (Post-incrementing forward iteration.)
1.223 + * "Unsafe" macro, assumes well-formed UTF-16.
1.224 + *
1.225 + * The offset may point to the lead surrogate unit
1.226 + * for a supplementary code point, in which case the macro will read
1.227 + * the following trail surrogate as well.
1.228 + * If the offset points to a trail surrogate, then that itself
1.229 + * will be returned as the code point.
1.230 + * The result is undefined if the offset points to a single, unpaired lead surrogate.
1.231 + *
1.232 + * @param s const UChar * string
1.233 + * @param i string offset
1.234 + * @param c output UChar32 variable
1.235 + * @see U16_NEXT
1.236 + * @stable ICU 2.4
1.237 + */
1.238 +#define U16_NEXT_UNSAFE(s, i, c) { \
1.239 + (c)=(s)[(i)++]; \
1.240 + if(U16_IS_LEAD(c)) { \
1.241 + (c)=U16_GET_SUPPLEMENTARY((c), (s)[(i)++]); \
1.242 + } \
1.243 +}
1.244 +
1.245 +/**
1.246 + * Get a code point from a string at a code point boundary offset,
1.247 + * and advance the offset to the next code point boundary.
1.248 + * (Post-incrementing forward iteration.)
1.249 + * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
1.250 + *
1.251 + * The length can be negative for a NUL-terminated string.
1.252 + *
1.253 + * The offset may point to the lead surrogate unit
1.254 + * for a supplementary code point, in which case the macro will read
1.255 + * the following trail surrogate as well.
1.256 + * If the offset points to a trail surrogate or
1.257 + * to a single, unpaired lead surrogate, then that itself
1.258 + * will be returned as the code point.
1.259 + *
1.260 + * @param s const UChar * string
1.261 + * @param i string offset, must be i<length
1.262 + * @param length string length
1.263 + * @param c output UChar32 variable
1.264 + * @see U16_NEXT_UNSAFE
1.265 + * @stable ICU 2.4
1.266 + */
1.267 +#define U16_NEXT(s, i, length, c) { \
1.268 + (c)=(s)[(i)++]; \
1.269 + if(U16_IS_LEAD(c)) { \
1.270 + uint16_t __c2; \
1.271 + if((i)!=(length) && U16_IS_TRAIL(__c2=(s)[(i)])) { \
1.272 + ++(i); \
1.273 + (c)=U16_GET_SUPPLEMENTARY((c), __c2); \
1.274 + } \
1.275 + } \
1.276 +}
1.277 +
1.278 +/**
1.279 + * Append a code point to a string, overwriting 1 or 2 code units.
1.280 + * The offset points to the current end of the string contents
1.281 + * and is advanced (post-increment).
1.282 + * "Unsafe" macro, assumes a valid code point and sufficient space in the string.
1.283 + * Otherwise, the result is undefined.
1.284 + *
1.285 + * @param s const UChar * string buffer
1.286 + * @param i string offset
1.287 + * @param c code point to append
1.288 + * @see U16_APPEND
1.289 + * @stable ICU 2.4
1.290 + */
1.291 +#define U16_APPEND_UNSAFE(s, i, c) { \
1.292 + if((uint32_t)(c)<=0xffff) { \
1.293 + (s)[(i)++]=(uint16_t)(c); \
1.294 + } else { \
1.295 + (s)[(i)++]=(uint16_t)(((c)>>10)+0xd7c0); \
1.296 + (s)[(i)++]=(uint16_t)(((c)&0x3ff)|0xdc00); \
1.297 + } \
1.298 +}
1.299 +
1.300 +/**
1.301 + * Append a code point to a string, overwriting 1 or 2 code units.
1.302 + * The offset points to the current end of the string contents
1.303 + * and is advanced (post-increment).
1.304 + * "Safe" macro, checks for a valid code point.
1.305 + * If a surrogate pair is written, checks for sufficient space in the string.
1.306 + * If the code point is not valid or a trail surrogate does not fit,
1.307 + * then isError is set to TRUE.
1.308 + *
1.309 + * @param s const UChar * string buffer
1.310 + * @param i string offset, must be i<capacity
1.311 + * @param capacity size of the string buffer
1.312 + * @param c code point to append
1.313 + * @param isError output UBool set to TRUE if an error occurs, otherwise not modified
1.314 + * @see U16_APPEND_UNSAFE
1.315 + * @stable ICU 2.4
1.316 + */
1.317 +#define U16_APPEND(s, i, capacity, c, isError) { \
1.318 + if((uint32_t)(c)<=0xffff) { \
1.319 + (s)[(i)++]=(uint16_t)(c); \
1.320 + } else if((uint32_t)(c)<=0x10ffff && (i)+1<(capacity)) { \
1.321 + (s)[(i)++]=(uint16_t)(((c)>>10)+0xd7c0); \
1.322 + (s)[(i)++]=(uint16_t)(((c)&0x3ff)|0xdc00); \
1.323 + } else /* c>0x10ffff or not enough space */ { \
1.324 + (isError)=TRUE; \
1.325 + } \
1.326 +}
1.327 +
1.328 +/**
1.329 + * Advance the string offset from one code point boundary to the next.
1.330 + * (Post-incrementing iteration.)
1.331 + * "Unsafe" macro, assumes well-formed UTF-16.
1.332 + *
1.333 + * @param s const UChar * string
1.334 + * @param i string offset
1.335 + * @see U16_FWD_1
1.336 + * @stable ICU 2.4
1.337 + */
1.338 +#define U16_FWD_1_UNSAFE(s, i) { \
1.339 + if(U16_IS_LEAD((s)[(i)++])) { \
1.340 + ++(i); \
1.341 + } \
1.342 +}
1.343 +
1.344 +/**
1.345 + * Advance the string offset from one code point boundary to the next.
1.346 + * (Post-incrementing iteration.)
1.347 + * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
1.348 + *
1.349 + * The length can be negative for a NUL-terminated string.
1.350 + *
1.351 + * @param s const UChar * string
1.352 + * @param i string offset, must be i<length
1.353 + * @param length string length
1.354 + * @see U16_FWD_1_UNSAFE
1.355 + * @stable ICU 2.4
1.356 + */
1.357 +#define U16_FWD_1(s, i, length) { \
1.358 + if(U16_IS_LEAD((s)[(i)++]) && (i)!=(length) && U16_IS_TRAIL((s)[i])) { \
1.359 + ++(i); \
1.360 + } \
1.361 +}
1.362 +
1.363 +/**
1.364 + * Advance the string offset from one code point boundary to the n-th next one,
1.365 + * i.e., move forward by n code points.
1.366 + * (Post-incrementing iteration.)
1.367 + * "Unsafe" macro, assumes well-formed UTF-16.
1.368 + *
1.369 + * @param s const UChar * string
1.370 + * @param i string offset
1.371 + * @param n number of code points to skip
1.372 + * @see U16_FWD_N
1.373 + * @stable ICU 2.4
1.374 + */
1.375 +#define U16_FWD_N_UNSAFE(s, i, n) { \
1.376 + int32_t __N=(n); \
1.377 + while(__N>0) { \
1.378 + U16_FWD_1_UNSAFE(s, i); \
1.379 + --__N; \
1.380 + } \
1.381 +}
1.382 +
1.383 +/**
1.384 + * Advance the string offset from one code point boundary to the n-th next one,
1.385 + * i.e., move forward by n code points.
1.386 + * (Post-incrementing iteration.)
1.387 + * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
1.388 + *
1.389 + * The length can be negative for a NUL-terminated string.
1.390 + *
1.391 + * @param s const UChar * string
1.392 + * @param i int32_t string offset, must be i<length
1.393 + * @param length int32_t string length
1.394 + * @param n number of code points to skip
1.395 + * @see U16_FWD_N_UNSAFE
1.396 + * @stable ICU 2.4
1.397 + */
1.398 +#define U16_FWD_N(s, i, length, n) { \
1.399 + int32_t __N=(n); \
1.400 + while(__N>0 && ((i)<(length) || ((length)<0 && (s)[i]!=0))) { \
1.401 + U16_FWD_1(s, i, length); \
1.402 + --__N; \
1.403 + } \
1.404 +}
1.405 +
1.406 +/**
1.407 + * Adjust a random-access offset to a code point boundary
1.408 + * at the start of a code point.
1.409 + * If the offset points to the trail surrogate of a surrogate pair,
1.410 + * then the offset is decremented.
1.411 + * Otherwise, it is not modified.
1.412 + * "Unsafe" macro, assumes well-formed UTF-16.
1.413 + *
1.414 + * @param s const UChar * string
1.415 + * @param i string offset
1.416 + * @see U16_SET_CP_START
1.417 + * @stable ICU 2.4
1.418 + */
1.419 +#define U16_SET_CP_START_UNSAFE(s, i) { \
1.420 + if(U16_IS_TRAIL((s)[i])) { \
1.421 + --(i); \
1.422 + } \
1.423 +}
1.424 +
1.425 +/**
1.426 + * Adjust a random-access offset to a code point boundary
1.427 + * at the start of a code point.
1.428 + * If the offset points to the trail surrogate of a surrogate pair,
1.429 + * then the offset is decremented.
1.430 + * Otherwise, it is not modified.
1.431 + * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
1.432 + *
1.433 + * @param s const UChar * string
1.434 + * @param start starting string offset (usually 0)
1.435 + * @param i string offset, must be start<=i
1.436 + * @see U16_SET_CP_START_UNSAFE
1.437 + * @stable ICU 2.4
1.438 + */
1.439 +#define U16_SET_CP_START(s, start, i) { \
1.440 + if(U16_IS_TRAIL((s)[i]) && (i)>(start) && U16_IS_LEAD((s)[(i)-1])) { \
1.441 + --(i); \
1.442 + } \
1.443 +}
1.444 +
1.445 +/* definitions with backward iteration -------------------------------------- */
1.446 +
1.447 +/**
1.448 + * Move the string offset from one code point boundary to the previous one
1.449 + * and get the code point between them.
1.450 + * (Pre-decrementing backward iteration.)
1.451 + * "Unsafe" macro, assumes well-formed UTF-16.
1.452 + *
1.453 + * The input offset may be the same as the string length.
1.454 + * If the offset is behind a trail surrogate unit
1.455 + * for a supplementary code point, then the macro will read
1.456 + * the preceding lead surrogate as well.
1.457 + * If the offset is behind a lead surrogate, then that itself
1.458 + * will be returned as the code point.
1.459 + * The result is undefined if the offset is behind a single, unpaired trail surrogate.
1.460 + *
1.461 + * @param s const UChar * string
1.462 + * @param i string offset
1.463 + * @param c output UChar32 variable
1.464 + * @see U16_PREV
1.465 + * @stable ICU 2.4
1.466 + */
1.467 +#define U16_PREV_UNSAFE(s, i, c) { \
1.468 + (c)=(s)[--(i)]; \
1.469 + if(U16_IS_TRAIL(c)) { \
1.470 + (c)=U16_GET_SUPPLEMENTARY((s)[--(i)], (c)); \
1.471 + } \
1.472 +}
1.473 +
1.474 +/**
1.475 + * Move the string offset from one code point boundary to the previous one
1.476 + * and get the code point between them.
1.477 + * (Pre-decrementing backward iteration.)
1.478 + * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
1.479 + *
1.480 + * The input offset may be the same as the string length.
1.481 + * If the offset is behind a trail surrogate unit
1.482 + * for a supplementary code point, then the macro will read
1.483 + * the preceding lead surrogate as well.
1.484 + * If the offset is behind a lead surrogate or behind a single, unpaired
1.485 + * trail surrogate, then that itself
1.486 + * will be returned as the code point.
1.487 + *
1.488 + * @param s const UChar * string
1.489 + * @param start starting string offset (usually 0)
1.490 + * @param i string offset, must be start<i
1.491 + * @param c output UChar32 variable
1.492 + * @see U16_PREV_UNSAFE
1.493 + * @stable ICU 2.4
1.494 + */
1.495 +#define U16_PREV(s, start, i, c) { \
1.496 + (c)=(s)[--(i)]; \
1.497 + if(U16_IS_TRAIL(c)) { \
1.498 + uint16_t __c2; \
1.499 + if((i)>(start) && U16_IS_LEAD(__c2=(s)[(i)-1])) { \
1.500 + --(i); \
1.501 + (c)=U16_GET_SUPPLEMENTARY(__c2, (c)); \
1.502 + } \
1.503 + } \
1.504 +}
1.505 +
1.506 +/**
1.507 + * Move the string offset from one code point boundary to the previous one.
1.508 + * (Pre-decrementing backward iteration.)
1.509 + * The input offset may be the same as the string length.
1.510 + * "Unsafe" macro, assumes well-formed UTF-16.
1.511 + *
1.512 + * @param s const UChar * string
1.513 + * @param i string offset
1.514 + * @see U16_BACK_1
1.515 + * @stable ICU 2.4
1.516 + */
1.517 +#define U16_BACK_1_UNSAFE(s, i) { \
1.518 + if(U16_IS_TRAIL((s)[--(i)])) { \
1.519 + --(i); \
1.520 + } \
1.521 +}
1.522 +
1.523 +/**
1.524 + * Move the string offset from one code point boundary to the previous one.
1.525 + * (Pre-decrementing backward iteration.)
1.526 + * The input offset may be the same as the string length.
1.527 + * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
1.528 + *
1.529 + * @param s const UChar * string
1.530 + * @param start starting string offset (usually 0)
1.531 + * @param i string offset, must be start<i
1.532 + * @see U16_BACK_1_UNSAFE
1.533 + * @stable ICU 2.4
1.534 + */
1.535 +#define U16_BACK_1(s, start, i) { \
1.536 + if(U16_IS_TRAIL((s)[--(i)]) && (i)>(start) && U16_IS_LEAD((s)[(i)-1])) { \
1.537 + --(i); \
1.538 + } \
1.539 +}
1.540 +
1.541 +/**
1.542 + * Move the string offset from one code point boundary to the n-th one before it,
1.543 + * i.e., move backward by n code points.
1.544 + * (Pre-decrementing backward iteration.)
1.545 + * The input offset may be the same as the string length.
1.546 + * "Unsafe" macro, assumes well-formed UTF-16.
1.547 + *
1.548 + * @param s const UChar * string
1.549 + * @param i string offset
1.550 + * @param n number of code points to skip
1.551 + * @see U16_BACK_N
1.552 + * @stable ICU 2.4
1.553 + */
1.554 +#define U16_BACK_N_UNSAFE(s, i, n) { \
1.555 + int32_t __N=(n); \
1.556 + while(__N>0) { \
1.557 + U16_BACK_1_UNSAFE(s, i); \
1.558 + --__N; \
1.559 + } \
1.560 +}
1.561 +
1.562 +/**
1.563 + * Move the string offset from one code point boundary to the n-th one before it,
1.564 + * i.e., move backward by n code points.
1.565 + * (Pre-decrementing backward iteration.)
1.566 + * The input offset may be the same as the string length.
1.567 + * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
1.568 + *
1.569 + * @param s const UChar * string
1.570 + * @param start start of string
1.571 + * @param i string offset, must be start<i
1.572 + * @param n number of code points to skip
1.573 + * @see U16_BACK_N_UNSAFE
1.574 + * @stable ICU 2.4
1.575 + */
1.576 +#define U16_BACK_N(s, start, i, n) { \
1.577 + int32_t __N=(n); \
1.578 + while(__N>0 && (i)>(start)) { \
1.579 + U16_BACK_1(s, start, i); \
1.580 + --__N; \
1.581 + } \
1.582 +}
1.583 +
1.584 +/**
1.585 + * Adjust a random-access offset to a code point boundary after a code point.
1.586 + * If the offset is behind the lead surrogate of a surrogate pair,
1.587 + * then the offset is incremented.
1.588 + * Otherwise, it is not modified.
1.589 + * The input offset may be the same as the string length.
1.590 + * "Unsafe" macro, assumes well-formed UTF-16.
1.591 + *
1.592 + * @param s const UChar * string
1.593 + * @param i string offset
1.594 + * @see U16_SET_CP_LIMIT
1.595 + * @stable ICU 2.4
1.596 + */
1.597 +#define U16_SET_CP_LIMIT_UNSAFE(s, i) { \
1.598 + if(U16_IS_LEAD((s)[(i)-1])) { \
1.599 + ++(i); \
1.600 + } \
1.601 +}
1.602 +
1.603 +/**
1.604 + * Adjust a random-access offset to a code point boundary after a code point.
1.605 + * If the offset is behind the lead surrogate of a surrogate pair,
1.606 + * then the offset is incremented.
1.607 + * Otherwise, it is not modified.
1.608 + * The input offset may be the same as the string length.
1.609 + * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
1.610 + *
1.611 + * The length can be negative for a NUL-terminated string.
1.612 + *
1.613 + * @param s const UChar * string
1.614 + * @param start int32_t starting string offset (usually 0)
1.615 + * @param i int32_t string offset, start<=i<=length
1.616 + * @param length int32_t string length
1.617 + * @see U16_SET_CP_LIMIT_UNSAFE
1.618 + * @stable ICU 2.4
1.619 + */
1.620 +#define U16_SET_CP_LIMIT(s, start, i, length) { \
1.621 + if((start)<(i) && ((i)<(length) || (length)<0) && U16_IS_LEAD((s)[(i)-1]) && U16_IS_TRAIL((s)[i])) { \
1.622 + ++(i); \
1.623 + } \
1.624 +}
1.625 +
1.626 +#endif
