| 1 | /* $Id: utf-16-case.cpp 106061 2024-09-16 14:03:52Z vboxsync $ */
|
|---|
| 2 | /** @file
|
|---|
| 3 | * IPRT - UTF-16, Case Sensitivity.
|
|---|
| 4 | */
|
|---|
| 5 |
|
|---|
| 6 | /*
|
|---|
| 7 | * Copyright (C) 2006-2024 Oracle and/or its affiliates.
|
|---|
| 8 | *
|
|---|
| 9 | * This file is part of VirtualBox base platform packages, as
|
|---|
| 10 | * available from https://www.virtualbox.org.
|
|---|
| 11 | *
|
|---|
| 12 | * This program is free software; you can redistribute it and/or
|
|---|
| 13 | * modify it under the terms of the GNU General Public License
|
|---|
| 14 | * as published by the Free Software Foundation, in version 3 of the
|
|---|
| 15 | * License.
|
|---|
| 16 | *
|
|---|
| 17 | * This program is distributed in the hope that it will be useful, but
|
|---|
| 18 | * WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 19 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|---|
| 20 | * General Public License for more details.
|
|---|
| 21 | *
|
|---|
| 22 | * You should have received a copy of the GNU General Public License
|
|---|
| 23 | * along with this program; if not, see <https://www.gnu.org/licenses>.
|
|---|
| 24 | *
|
|---|
| 25 | * The contents of this file may alternatively be used under the terms
|
|---|
| 26 | * of the Common Development and Distribution License Version 1.0
|
|---|
| 27 | * (CDDL), a copy of it is provided in the "COPYING.CDDL" file included
|
|---|
| 28 | * in the VirtualBox distribution, in which case the provisions of the
|
|---|
| 29 | * CDDL are applicable instead of those of the GPL.
|
|---|
| 30 | *
|
|---|
| 31 | * You may elect to license modified versions of this file under the
|
|---|
| 32 | * terms and conditions of either the GPL or the CDDL or both.
|
|---|
| 33 | *
|
|---|
| 34 | * SPDX-License-Identifier: GPL-3.0-only OR CDDL-1.0
|
|---|
| 35 | */
|
|---|
| 36 |
|
|---|
| 37 |
|
|---|
| 38 | /*********************************************************************************************************************************
|
|---|
| 39 | * Header Files *
|
|---|
| 40 | *********************************************************************************************************************************/
|
|---|
| 41 | #include <iprt/utf16.h>
|
|---|
| 42 | #include "internal/iprt.h"
|
|---|
| 43 |
|
|---|
| 44 | #include <iprt/uni.h>
|
|---|
| 45 | #include <iprt/alloc.h>
|
|---|
| 46 | #include <iprt/asm.h>
|
|---|
| 47 | #include <iprt/assert.h>
|
|---|
| 48 | #include <iprt/errcore.h>
|
|---|
| 49 | #include "internal/string.h"
|
|---|
| 50 |
|
|---|
| 51 |
|
|---|
| 52 | RTDECL(int) RTUtf16ICmp(PCRTUTF16 pwsz1, PCRTUTF16 pwsz2)
|
|---|
| 53 | {
|
|---|
| 54 | if (pwsz1 == pwsz2)
|
|---|
| 55 | return 0;
|
|---|
| 56 | if (!pwsz1)
|
|---|
| 57 | return -1;
|
|---|
| 58 | if (!pwsz2)
|
|---|
| 59 | return 1;
|
|---|
| 60 |
|
|---|
| 61 | PCRTUTF16 pwsz1Start = pwsz1; /* keep it around in case we have to backtrack on a surrogate pair */
|
|---|
| 62 | for (;;)
|
|---|
| 63 | {
|
|---|
| 64 | RTUTF16 wc1 = *pwsz1;
|
|---|
| 65 | RTUTF16 wc2 = *pwsz2;
|
|---|
| 66 | int iDiff = wc1 - wc2;
|
|---|
| 67 | if (iDiff)
|
|---|
| 68 | {
|
|---|
| 69 | /* unless they are *both* surrogate pairs, there is no chance they'll be identical. */
|
|---|
| 70 | if ( wc1 < 0xd800
|
|---|
| 71 | || wc2 < 0xd800
|
|---|
| 72 | || wc1 > 0xdfff
|
|---|
| 73 | || wc2 > 0xdfff)
|
|---|
| 74 | {
|
|---|
| 75 | /* simple UCS-2 char */
|
|---|
| 76 | iDiff = RTUniCpToUpper(wc1) - RTUniCpToUpper(wc2);
|
|---|
| 77 | if (iDiff)
|
|---|
| 78 | iDiff = RTUniCpToLower(wc1) - RTUniCpToLower(wc2);
|
|---|
| 79 | }
|
|---|
| 80 | else
|
|---|
| 81 | {
|
|---|
| 82 | /* a damned pair */
|
|---|
| 83 | RTUNICP uc1;
|
|---|
| 84 | RTUNICP uc2;
|
|---|
| 85 | if (wc1 >= 0xdc00)
|
|---|
| 86 | {
|
|---|
| 87 | if (pwsz1Start == pwsz1)
|
|---|
| 88 | return iDiff;
|
|---|
| 89 | uc1 = pwsz1[-1];
|
|---|
| 90 | if (uc1 < 0xd800 || uc1 >= 0xdc00)
|
|---|
| 91 | return iDiff;
|
|---|
| 92 | uc1 = 0x10000 + (((uc1 & 0x3ff) << 10) | (wc1 & 0x3ff));
|
|---|
| 93 | uc2 = 0x10000 + (((pwsz2[-1] & 0x3ff) << 10) | (wc2 & 0x3ff));
|
|---|
| 94 | }
|
|---|
| 95 | else
|
|---|
| 96 | {
|
|---|
| 97 | uc1 = *++pwsz1;
|
|---|
| 98 | if (uc1 < 0xdc00 || uc1 >= 0xe000)
|
|---|
| 99 | return iDiff;
|
|---|
| 100 | uc1 = 0x10000 + (((wc1 & 0x3ff) << 10) | (uc1 & 0x3ff));
|
|---|
| 101 | uc2 = 0x10000 + (((wc2 & 0x3ff) << 10) | (*++pwsz2 & 0x3ff));
|
|---|
| 102 | }
|
|---|
| 103 | iDiff = RTUniCpToUpper(uc1) - RTUniCpToUpper(uc2);
|
|---|
| 104 | if (iDiff)
|
|---|
| 105 | iDiff = RTUniCpToLower(uc1) - RTUniCpToLower(uc2); /* serious paranoia! */
|
|---|
| 106 | }
|
|---|
| 107 | if (iDiff)
|
|---|
| 108 | return iDiff;
|
|---|
| 109 | }
|
|---|
| 110 | if (!wc1)
|
|---|
| 111 | return 0;
|
|---|
| 112 | pwsz1++;
|
|---|
| 113 | pwsz2++;
|
|---|
| 114 | }
|
|---|
| 115 | }
|
|---|
| 116 | RT_EXPORT_SYMBOL(RTUtf16ICmp);
|
|---|
| 117 |
|
|---|
| 118 |
|
|---|
| 119 | RTDECL(int) RTUtf16BigICmp(PCRTUTF16 pwsz1, PCRTUTF16 pwsz2)
|
|---|
| 120 | {
|
|---|
| 121 | if (pwsz1 == pwsz2)
|
|---|
| 122 | return 0;
|
|---|
| 123 | if (!pwsz1)
|
|---|
| 124 | return -1;
|
|---|
| 125 | if (!pwsz2)
|
|---|
| 126 | return 1;
|
|---|
| 127 |
|
|---|
| 128 | PCRTUTF16 pwsz1Start = pwsz1; /* keep it around in case we have to backtrack on a surrogate pair */
|
|---|
| 129 | for (;;)
|
|---|
| 130 | {
|
|---|
| 131 | RTUTF16 wc1 = *pwsz1;
|
|---|
| 132 | RTUTF16 wc2 = *pwsz2;
|
|---|
| 133 | int iDiff = wc1 - wc2;
|
|---|
| 134 | if (iDiff)
|
|---|
| 135 | {
|
|---|
| 136 | /* unless they are *both* surrogate pairs, there is no chance they'll be identical. */
|
|---|
| 137 | wc1 = RT_BE2H_U16(wc1);
|
|---|
| 138 | wc2 = RT_BE2H_U16(wc2);
|
|---|
| 139 | if ( wc1 < 0xd800
|
|---|
| 140 | || wc2 < 0xd800
|
|---|
| 141 | || wc1 > 0xdfff
|
|---|
| 142 | || wc2 > 0xdfff)
|
|---|
| 143 | {
|
|---|
| 144 | /* simple UCS-2 char */
|
|---|
| 145 | iDiff = RTUniCpToUpper(wc1) - RTUniCpToUpper(wc2);
|
|---|
| 146 | if (iDiff)
|
|---|
| 147 | iDiff = RTUniCpToLower(wc1) - RTUniCpToLower(wc2);
|
|---|
| 148 | }
|
|---|
| 149 | else
|
|---|
| 150 | {
|
|---|
| 151 | /* a damned pair */
|
|---|
| 152 | RTUNICP uc1;
|
|---|
| 153 | RTUNICP uc2;
|
|---|
| 154 | if (wc1 >= 0xdc00)
|
|---|
| 155 | {
|
|---|
| 156 | if (pwsz1Start == pwsz1)
|
|---|
| 157 | return iDiff;
|
|---|
| 158 | uc1 = RT_BE2H_U16(pwsz1[-1]);
|
|---|
| 159 | if (uc1 < 0xd800 || uc1 >= 0xdc00)
|
|---|
| 160 | return iDiff;
|
|---|
| 161 | uc1 = 0x10000 + (((uc1 & 0x3ff) << 10) | (wc1 & 0x3ff));
|
|---|
| 162 | uc2 = 0x10000 + (((RT_BE2H_U16(pwsz2[-1]) & 0x3ff) << 10) | (wc2 & 0x3ff));
|
|---|
| 163 | }
|
|---|
| 164 | else
|
|---|
| 165 | {
|
|---|
| 166 | RTUTF16 wcTmp = *++pwsz1;
|
|---|
| 167 | uc1 = RT_BE2H_U16(wcTmp);
|
|---|
| 168 | if (uc1 < 0xdc00 || uc1 >= 0xe000)
|
|---|
| 169 | return iDiff;
|
|---|
| 170 | uc1 = 0x10000 + (((wc1 & 0x3ff) << 10) | (uc1 & 0x3ff));
|
|---|
| 171 | wcTmp = *++pwsz2;
|
|---|
| 172 | uc2 = 0x10000 + (((wc2 & 0x3ff) << 10) | (RT_BE2H_U16(wcTmp) & 0x3ff));
|
|---|
| 173 | }
|
|---|
| 174 | iDiff = RTUniCpToUpper(uc1) - RTUniCpToUpper(uc2);
|
|---|
| 175 | if (iDiff)
|
|---|
| 176 | iDiff = RTUniCpToLower(uc1) - RTUniCpToLower(uc2); /* serious paranoia! */
|
|---|
| 177 | }
|
|---|
| 178 | if (iDiff)
|
|---|
| 179 | return iDiff;
|
|---|
| 180 | }
|
|---|
| 181 | if (!wc1)
|
|---|
| 182 | return 0;
|
|---|
| 183 | pwsz1++;
|
|---|
| 184 | pwsz2++;
|
|---|
| 185 | }
|
|---|
| 186 | }
|
|---|
| 187 | RT_EXPORT_SYMBOL(RTUtf16BigICmp);
|
|---|
| 188 |
|
|---|
| 189 |
|
|---|
| 190 | RTDECL(int) RTUtf16ICmpUtf8(PCRTUTF16 pwsz1, const char *psz2)
|
|---|
| 191 | {
|
|---|
| 192 | /*
|
|---|
| 193 | * NULL and empty strings are all the same.
|
|---|
| 194 | */
|
|---|
| 195 | if (!pwsz1)
|
|---|
| 196 | return !psz2 || !*psz2 ? 0 : -1;
|
|---|
| 197 | if (!psz2)
|
|---|
| 198 | return !*pwsz1 ? 0 : 1;
|
|---|
| 199 |
|
|---|
| 200 | /*
|
|---|
| 201 | * Compare with a UTF-8 string by enumerating them char by char.
|
|---|
| 202 | */
|
|---|
| 203 | for (;;)
|
|---|
| 204 | {
|
|---|
| 205 | RTUNICP uc1;
|
|---|
| 206 | int rc = RTUtf16GetCpEx(&pwsz1, &uc1);
|
|---|
| 207 | AssertRCReturn(rc, 1);
|
|---|
| 208 |
|
|---|
| 209 | RTUNICP uc2;
|
|---|
| 210 | rc = RTStrGetCpEx(&psz2, &uc2);
|
|---|
| 211 | AssertRCReturn(rc, -1);
|
|---|
| 212 | if (uc1 == uc2)
|
|---|
| 213 | {
|
|---|
| 214 | if (uc1)
|
|---|
| 215 | continue;
|
|---|
| 216 | return 0;
|
|---|
| 217 | }
|
|---|
| 218 |
|
|---|
| 219 | if (RTUniCpToUpper(uc1) == RTUniCpToUpper(uc2))
|
|---|
| 220 | continue;
|
|---|
| 221 | if (RTUniCpToLower(uc1) == RTUniCpToLower(uc2))
|
|---|
| 222 | continue;
|
|---|
| 223 | return uc1 < uc2 ? -1 : 1;
|
|---|
| 224 | }
|
|---|
| 225 | }
|
|---|
| 226 | RT_EXPORT_SYMBOL(RTUtf16CmpIUtf8);
|
|---|
| 227 |
|
|---|
| 228 |
|
|---|
| 229 | RTDECL(int) RTUtf16NICmp(PCRTUTF16 pwsz1, PCRTUTF16 pwsz2, size_t cwcMax)
|
|---|
| 230 | {
|
|---|
| 231 | if (pwsz1 == pwsz2)
|
|---|
| 232 | return 0;
|
|---|
| 233 | if (!pwsz1)
|
|---|
| 234 | return -1;
|
|---|
| 235 | if (!pwsz2)
|
|---|
| 236 | return 1;
|
|---|
| 237 |
|
|---|
| 238 | PCRTUTF16 pwsz1Start = pwsz1; /* keep it around in case we have to backtrack on a surrogate pair */
|
|---|
| 239 | while (cwcMax-- > 0)
|
|---|
| 240 | {
|
|---|
| 241 | RTUTF16 wc1 = *pwsz1;
|
|---|
| 242 | RTUTF16 wc2 = *pwsz2;
|
|---|
| 243 | int iDiff = wc1 - wc2;
|
|---|
| 244 | if (iDiff)
|
|---|
| 245 | {
|
|---|
| 246 | /* unless they are *both* surrogate pairs, there is no chance they'll be identical. */
|
|---|
| 247 | if ( wc1 < 0xd800
|
|---|
| 248 | || wc2 < 0xd800
|
|---|
| 249 | || wc1 > 0xdfff
|
|---|
| 250 | || wc2 > 0xdfff)
|
|---|
| 251 | {
|
|---|
| 252 | /* simple UCS-2 char */
|
|---|
| 253 | iDiff = RTUniCpToUpper(wc1) - RTUniCpToUpper(wc2);
|
|---|
| 254 | if (iDiff)
|
|---|
| 255 | iDiff = RTUniCpToLower(wc1) - RTUniCpToLower(wc2);
|
|---|
| 256 | }
|
|---|
| 257 | else
|
|---|
| 258 | {
|
|---|
| 259 | /* a damned pair */
|
|---|
| 260 | RTUNICP uc1;
|
|---|
| 261 | RTUNICP uc2;
|
|---|
| 262 | if (wc1 >= 0xdc00)
|
|---|
| 263 | {
|
|---|
| 264 | if (pwsz1Start == pwsz1)
|
|---|
| 265 | return iDiff;
|
|---|
| 266 | uc1 = pwsz1[-1];
|
|---|
| 267 | if (uc1 < 0xd800 || uc1 >= 0xdc00)
|
|---|
| 268 | return iDiff;
|
|---|
| 269 | uc1 = 0x10000 + (((uc1 & 0x3ff) << 10) | (wc1 & 0x3ff));
|
|---|
| 270 | uc2 = 0x10000 + (((pwsz2[-1] & 0x3ff) << 10) | (wc2 & 0x3ff));
|
|---|
| 271 | }
|
|---|
| 272 | else if (cwcMax-- > 0)
|
|---|
| 273 | {
|
|---|
| 274 | uc1 = *++pwsz1;
|
|---|
| 275 | if (uc1 < 0xdc00 || uc1 >= 0xe000)
|
|---|
| 276 | return iDiff;
|
|---|
| 277 | uc1 = 0x10000 + (((wc1 & 0x3ff) << 10) | (uc1 & 0x3ff));
|
|---|
| 278 | uc2 = 0x10000 + (((wc2 & 0x3ff) << 10) | (*++pwsz2 & 0x3ff));
|
|---|
| 279 | }
|
|---|
| 280 | else
|
|---|
| 281 | {
|
|---|
| 282 | iDiff = wc1 - wc2;
|
|---|
| 283 | return iDiff;
|
|---|
| 284 | }
|
|---|
| 285 | iDiff = RTUniCpToUpper(uc1) - RTUniCpToUpper(uc2);
|
|---|
| 286 | if (iDiff)
|
|---|
| 287 | iDiff = RTUniCpToLower(uc1) - RTUniCpToLower(uc2); /* serious paranoia! */
|
|---|
| 288 | }
|
|---|
| 289 | if (iDiff)
|
|---|
| 290 | return iDiff;
|
|---|
| 291 | }
|
|---|
| 292 | if (!wc1)
|
|---|
| 293 | return 0;
|
|---|
| 294 | pwsz1++;
|
|---|
| 295 | pwsz2++;
|
|---|
| 296 | }
|
|---|
| 297 | return 0;
|
|---|
| 298 | }
|
|---|
| 299 | RT_EXPORT_SYMBOL(RTUtf16NICmp);
|
|---|
| 300 |
|
|---|
| 301 |
|
|---|
| 302 | RTDECL(int) RTUtf16BigNICmp(PCRTUTF16 pwsz1, PCRTUTF16 pwsz2, size_t cwcMax)
|
|---|
| 303 | {
|
|---|
| 304 | if (pwsz1 == pwsz2)
|
|---|
| 305 | return 0;
|
|---|
| 306 | if (!pwsz1)
|
|---|
| 307 | return -1;
|
|---|
| 308 | if (!pwsz2)
|
|---|
| 309 | return 1;
|
|---|
| 310 |
|
|---|
| 311 | PCRTUTF16 pwsz1Start = pwsz1; /* keep it around in case we have to backtrack on a surrogate pair */
|
|---|
| 312 | while (cwcMax-- > 0)
|
|---|
| 313 | {
|
|---|
| 314 | RTUTF16 wc1 = *pwsz1;
|
|---|
| 315 | RTUTF16 wc2 = *pwsz2;
|
|---|
| 316 | int iDiff = wc1 - wc2;
|
|---|
| 317 | if (iDiff)
|
|---|
| 318 | {
|
|---|
| 319 | /* unless they are *both* surrogate pairs, there is no chance they'll be identical. */
|
|---|
| 320 | wc1 = RT_BE2H_U16(wc1);
|
|---|
| 321 | wc2 = RT_BE2H_U16(wc2);
|
|---|
| 322 | if ( wc1 < 0xd800
|
|---|
| 323 | || wc2 < 0xd800
|
|---|
| 324 | || wc1 > 0xdfff
|
|---|
| 325 | || wc2 > 0xdfff)
|
|---|
| 326 | {
|
|---|
| 327 | /* simple UCS-2 char */
|
|---|
| 328 | iDiff = RTUniCpToUpper(wc1) - RTUniCpToUpper(wc2);
|
|---|
| 329 | if (iDiff)
|
|---|
| 330 | iDiff = RTUniCpToLower(wc1) - RTUniCpToLower(wc2);
|
|---|
| 331 | }
|
|---|
| 332 | else
|
|---|
| 333 | {
|
|---|
| 334 | /* a damned pair */
|
|---|
| 335 | RTUNICP uc1;
|
|---|
| 336 | RTUNICP uc2;
|
|---|
| 337 | if (wc1 >= 0xdc00)
|
|---|
| 338 | {
|
|---|
| 339 | if (pwsz1Start == pwsz1)
|
|---|
| 340 | return iDiff;
|
|---|
| 341 | uc1 = RT_BE2H_U16(pwsz1[-1]);
|
|---|
| 342 | if (uc1 < 0xd800 || uc1 >= 0xdc00)
|
|---|
| 343 | return iDiff;
|
|---|
| 344 | uc1 = 0x10000 + (((uc1 & 0x3ff) << 10) | (wc1 & 0x3ff));
|
|---|
| 345 | uc2 = 0x10000 + (((RT_BE2H_U16(pwsz2[-1]) & 0x3ff) << 10) | (wc2 & 0x3ff));
|
|---|
| 346 | }
|
|---|
| 347 | else if (cwcMax > 0)
|
|---|
| 348 | {
|
|---|
| 349 | RTUTF16 wcTmp = *++pwsz1;
|
|---|
| 350 | uc1 = RT_BE2H_U16(wcTmp);
|
|---|
| 351 | if (uc1 < 0xdc00 || uc1 >= 0xe000)
|
|---|
| 352 | return iDiff;
|
|---|
| 353 | uc1 = 0x10000 + (((wc1 & 0x3ff) << 10) | (uc1 & 0x3ff));
|
|---|
| 354 | wcTmp = *++pwsz2;
|
|---|
| 355 | uc2 = 0x10000 + (((wc2 & 0x3ff) << 10) | (RT_BE2H_U16(wcTmp) & 0x3ff));
|
|---|
| 356 | }
|
|---|
| 357 | else
|
|---|
| 358 | {
|
|---|
| 359 | iDiff = wc1 - wc2;
|
|---|
| 360 | return iDiff;
|
|---|
| 361 | }
|
|---|
| 362 | iDiff = RTUniCpToUpper(uc1) - RTUniCpToUpper(uc2);
|
|---|
| 363 | if (iDiff)
|
|---|
| 364 | iDiff = RTUniCpToLower(uc1) - RTUniCpToLower(uc2); /* serious paranoia! */
|
|---|
| 365 | }
|
|---|
| 366 | if (iDiff)
|
|---|
| 367 | return iDiff;
|
|---|
| 368 | }
|
|---|
| 369 | if (!wc1)
|
|---|
| 370 | return 0;
|
|---|
| 371 | pwsz1++;
|
|---|
| 372 | pwsz2++;
|
|---|
| 373 | }
|
|---|
| 374 | return 0;
|
|---|
| 375 | }
|
|---|
| 376 | RT_EXPORT_SYMBOL(RTUtf16BigNICmp);
|
|---|
| 377 |
|
|---|
| 378 |
|
|---|
| 379 | RTDECL(PRTUTF16) RTUtf16ToLower(PRTUTF16 pwsz)
|
|---|
| 380 | {
|
|---|
| 381 | PRTUTF16 pwc = pwsz;
|
|---|
| 382 | for (;;)
|
|---|
| 383 | {
|
|---|
| 384 | RTUTF16 wc = *pwc;
|
|---|
| 385 | if (!wc)
|
|---|
| 386 | break;
|
|---|
| 387 | if (wc < 0xd800 || wc >= 0xdc00)
|
|---|
| 388 | {
|
|---|
| 389 | RTUNICP ucFolded = RTUniCpToLower(wc);
|
|---|
| 390 | if (ucFolded < 0x10000)
|
|---|
| 391 | *pwc++ = RTUniCpToLower(wc);
|
|---|
| 392 | }
|
|---|
| 393 | else
|
|---|
| 394 | {
|
|---|
| 395 | /* surrogate */
|
|---|
| 396 | RTUTF16 wc2 = pwc[1];
|
|---|
| 397 | if (wc2 >= 0xdc00 && wc2 <= 0xdfff)
|
|---|
| 398 | {
|
|---|
| 399 | RTUNICP uc = 0x10000 + (((wc & 0x3ff) << 10) | (wc2 & 0x3ff));
|
|---|
| 400 | RTUNICP ucFolded = RTUniCpToLower(uc);
|
|---|
| 401 | if (uc != ucFolded && ucFolded >= 0x10000) /* we don't support shrinking the string */
|
|---|
| 402 | {
|
|---|
| 403 | uc -= 0x10000;
|
|---|
| 404 | *pwc++ = 0xd800 | (uc >> 10);
|
|---|
| 405 | *pwc++ = 0xdc00 | (uc & 0x3ff);
|
|---|
| 406 | }
|
|---|
| 407 | }
|
|---|
| 408 | else /* invalid encoding. */
|
|---|
| 409 | pwc++;
|
|---|
| 410 | }
|
|---|
| 411 | }
|
|---|
| 412 | return pwsz;
|
|---|
| 413 | }
|
|---|
| 414 | RT_EXPORT_SYMBOL(RTUtf16ToLower);
|
|---|
| 415 |
|
|---|
| 416 |
|
|---|
| 417 | RTDECL(PRTUTF16) RTUtf16ToUpper(PRTUTF16 pwsz)
|
|---|
| 418 | {
|
|---|
| 419 | PRTUTF16 pwc = pwsz;
|
|---|
| 420 | for (;;)
|
|---|
| 421 | {
|
|---|
| 422 | RTUTF16 wc = *pwc;
|
|---|
| 423 | if (!wc)
|
|---|
| 424 | break;
|
|---|
| 425 | if (wc < 0xd800 || wc >= 0xdc00)
|
|---|
| 426 | *pwc++ = RTUniCpToUpper(wc);
|
|---|
| 427 | else
|
|---|
| 428 | {
|
|---|
| 429 | /* surrogate */
|
|---|
| 430 | RTUTF16 wc2 = pwc[1];
|
|---|
| 431 | if (wc2 >= 0xdc00 && wc2 <= 0xdfff)
|
|---|
| 432 | {
|
|---|
| 433 | RTUNICP uc = 0x10000 + (((wc & 0x3ff) << 10) | (wc2 & 0x3ff));
|
|---|
| 434 | RTUNICP ucFolded = RTUniCpToUpper(uc);
|
|---|
| 435 | if (uc != ucFolded && ucFolded >= 0x10000) /* we don't support shrinking the string */
|
|---|
| 436 | {
|
|---|
| 437 | uc -= 0x10000;
|
|---|
| 438 | *pwc++ = 0xd800 | (uc >> 10);
|
|---|
| 439 | *pwc++ = 0xdc00 | (uc & 0x3ff);
|
|---|
| 440 | }
|
|---|
| 441 | }
|
|---|
| 442 | else /* invalid encoding. */
|
|---|
| 443 | pwc++;
|
|---|
| 444 | }
|
|---|
| 445 | }
|
|---|
| 446 | return pwsz;
|
|---|
| 447 | }
|
|---|
| 448 | RT_EXPORT_SYMBOL(RTUtf16ToUpper);
|
|---|
| 449 |
|
|---|
