locale/FreeBSD/utf2.c

   1 /*-
   2  * Copyright (c) 2002-2004 Tim J. Robbins
   3  * All rights reserved.
   4  *
   5  * Redistribution and use in source and binary forms, with or without
   6  * modification, are permitted provided that the following conditions
   7  * are met:
   8  * 1. Redistributions of source code must retain the above copyright
   9  *    notice, this list of conditions and the following disclaimer.
  10  * 2. Redistributions in binary form must reproduce the above copyright
  11  *    notice, this list of conditions and the following disclaimer in the
  12  *    documentation and/or other materials provided with the distribution.
  13  *
  14  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  15  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  16  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  17  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
  18  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  19  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  20  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  21  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  22  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  23  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  24  * SUCH DAMAGE.
  25  */
  26
  27 // MWW: Generated by applying utf2.c.patch to utf8.c in the FreeBSD patch sets.
  28
  29 #include <sys/param.h>
  30 __FBSDID("$FreeBSD: src/lib/libc/locale/utf8.c,v 1.16 2007/10/15 09:51:30 ache Exp $");
  31
  32 #include "xlocale_private.h"
  33
  34 #include <errno.h>
  35 #include <limits.h>
  36 #include <runetype.h>
  37 #include <stdlib.h>
  38 #include <string.h>
  39 #include <wchar.h>
  40 #include "mblocal.h"
  41
  42 #define UTF2_MB_CUR_MAX         3
  43
  44 static size_t   _UTF2_mbrtowc(wchar_t * __restrict, const char * __restrict,
  45                     size_t, mbstate_t * __restrict, locale_t);
  46 static int      _UTF2_mbsinit(const mbstate_t *, locale_t);
  47 static size_t   _UTF2_mbsnrtowcs(wchar_t * __restrict,
  48                     const char ** __restrict, size_t, size_t,
  49                     mbstate_t * __restrict, locale_t);
  50 static size_t   _UTF2_wcrtomb(char * __restrict, wchar_t,
  51                     mbstate_t * __restrict, locale_t);
  52 static size_t   _UTF2_wcsnrtombs(char * __restrict, const wchar_t ** __restrict,
  53                     size_t, size_t, mbstate_t * __restrict, locale_t);
  54
  55 typedef struct {
  56         wchar_t ch;
  57         int     want;
  58         wchar_t lbound;
  59 } _UTF2State;
  60
  61 int
  62 _UTF2_init(struct __xlocale_st_runelocale *xrl)
  63 {
  64
  65         xrl->__mbrtowc = _UTF2_mbrtowc;
  66         xrl->__wcrtomb = _UTF2_wcrtomb;
  67         xrl->__mbsinit = _UTF2_mbsinit;
  68         xrl->__mbsnrtowcs = _UTF2_mbsnrtowcs;
  69         xrl->__wcsnrtombs = _UTF2_wcsnrtombs;
  70         xrl->__mb_cur_max = UTF2_MB_CUR_MAX;
  71         /*
  72          * UCS-4 encoding used as the internal representation, so
  73          * slots 0x0080-0x00FF are occuped and must be excluded
  74          * from the single byte ctype by setting the limit.
  75          */
  76         xrl->__mb_sb_limit = 128;
  77
  78         return (0);
  79 }
  80
  81 static int
  82 _UTF2_mbsinit(const mbstate_t *ps, locale_t loc)
  83 {
  84
  85         return (ps == NULL || ((const _UTF2State *)ps)->want == 0);
  86 }
  87
  88 static size_t
  89 _UTF2_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
  90     mbstate_t * __restrict ps, locale_t loc)
  91 {
  92         _UTF2State *us;
  93         int ch, i, mask, want;
  94         wchar_t lbound, wch;
  95
  96         us = (_UTF2State *)ps;
  97
  98         if (us->want < 0 || us->want > 6) {
  99                 errno = EINVAL;
 100                 return ((size_t)-1);
 101         }
 102
 103         if (s == NULL) {
 104                 s = "";
 105                 n = 1;
 106                 pwc = NULL;
 107         }
 108
 109         if (n == 0)
 110                 /* Incomplete multibyte sequence */
 111                 return ((size_t)-2);
 112
 113         if (us->want == 0 && ((ch = (unsigned char)*s) & ~0x7f) == 0) {
 114                 /* Fast path for plain ASCII characters. */
 115                 if (pwc != NULL)
 116                         *pwc = ch;
 117                 return (ch != '\0' ? 1 : 0);
 118         }
 119
 120         if (us->want == 0) {
 121                 /*
 122                  * Determine the number of octets that make up this character
 123                  * from the first octet, and a mask that extracts the
 124                  * interesting bits of the first octet. We already know
 125                  * the character is at least two bytes long.
 126                  *
 127                  * We also specify a lower bound for the character code to
 128                  * detect redundant, non-"shortest form" encodings. For
 129                  * example, the sequence C0 80 is _not_ a legal representation
 130                  * of the null character. This enforces a 1-to-1 mapping
 131                  * between character codes and their multibyte representations.
 132                  */
 133                 ch = (unsigned char)*s;
 134                 if ((ch & 0x80) == 0) {
 135                         mask = 0x7f;
 136                         want = 1;
 137                         lbound = 0;
 138                 } else if ((ch & 0xe0) == 0xc0) {
 139                         mask = 0x1f;
 140                         want = 2;
 141                         lbound = 0x80;
 142                 } else if ((ch & 0xf0) == 0xe0) {
 143                         mask = 0x0f;
 144                         want = 3;
 145                         lbound = 0x800;
 146                 } else {
 147                         /*
 148                          * Malformed input; input is not UTF2.
 149                          */
 150                         errno = EILSEQ;
 151                         return ((size_t)-1);
 152                 }
 153         } else {
 154                 want = us->want;
 155                 lbound = us->lbound;
 156         }
 157
 158         /*
 159          * Decode the octet sequence representing the character in chunks
 160          * of 6 bits, most significant first.
 161          */
 162         if (us->want == 0)
 163                 wch = (unsigned char)*s++ & mask;
 164         else
 165                 wch = us->ch;
 166         for (i = (us->want == 0) ? 1 : 0; i < MIN(want, n); i++) {
 167                 if ((*s & 0xc0) != 0x80) {
 168                         /*
 169                          * Malformed input; bad characters in the middle
 170                          * of a character.
 171                          */
 172                         errno = EILSEQ;
 173                         return ((size_t)-1);
 174                 }
 175                 wch <<= 6;
 176                 wch |= *s++ & 0x3f;
 177         }
 178         if (i < want) {
 179                 /* Incomplete multibyte sequence. */
 180                 us->want = want - i;
 181                 us->lbound = lbound;
 182                 us->ch = wch;
 183                 return ((size_t)-2);
 184         }
 185         if (wch < lbound) {
 186                 /*
 187                  * Malformed input; redundant encoding.
 188                  */
 189                 errno = EILSEQ;
 190                 return ((size_t)-1);
 191         }
 192         if (pwc != NULL)
 193                 *pwc = wch;
 194         us->want = 0;
 195         return (wch == L'\0' ? 0 : want);
 196 }
 197
 198 static size_t
 199 _UTF2_mbsnrtowcs(wchar_t * __restrict dst, const char ** __restrict src,
 200     size_t nms, size_t len, mbstate_t * __restrict ps, locale_t loc)
 201 {
 202         _UTF2State *us;
 203         const char *s;
 204         size_t nchr;
 205         wchar_t wc;
 206         size_t nb;
 207
 208         us = (_UTF2State *)ps;
 209
 210         s = *src;
 211         nchr = 0;
 212
 213         if (dst == NULL) {
 214                 /*
 215                  * The fast path in the loop below is not safe if an ASCII
 216                  * character appears as anything but the first byte of a
 217                  * multibyte sequence. Check now to avoid doing it in the loop.
 218                  */
 219                 if (nms > 0 && us->want > 0 && (signed char)*s > 0) {
 220                         errno = EILSEQ;
 221                         return ((size_t)-1);
 222                 }
 223                 for (;;) {
 224                         if (nms > 0 && (signed char)*s > 0)
 225                                 /*
 226                                  * Fast path for plain ASCII characters
 227                                  * excluding NUL.
 228                                  */
 229                                 nb = 1;
 230                         else if ((nb = _UTF2_mbrtowc(&wc, s, nms, ps, loc)) ==
 231                             (size_t)-1)
 232                                 /* Invalid sequence - mbrtowc() sets errno. */
 233                                 return ((size_t)-1);
 234                         else if (nb == 0 || nb == (size_t)-2)
 235                                 return (nchr);
 236                         s += nb;
 237                         nms -= nb;
 238                         nchr++;
 239                 }
 240                 /*NOTREACHED*/
 241         }
 242
 243         /*
 244          * The fast path in the loop below is not safe if an ASCII
 245          * character appears as anything but the first byte of a
 246          * multibyte sequence. Check now to avoid doing it in the loop.
 247          */
 248         if (nms > 0 && len > 0 && us->want > 0 && (signed char)*s > 0) {
 249                 errno = EILSEQ;
 250                 return ((size_t)-1);
 251         }
 252         while (len-- > 0) {
 253                 if (nms > 0 && (signed char)*s > 0) {
 254                         /*
 255                          * Fast path for plain ASCII characters
 256                          * excluding NUL.
 257                          */
 258                         *dst = (wchar_t)*s;
 259                         nb = 1;
 260                 } else if ((nb = _UTF2_mbrtowc(dst, s, nms, ps, loc)) ==
 261                     (size_t)-1) {
 262                         *src = s;
 263                         return ((size_t)-1);
 264                 } else if (nb == (size_t)-2) {
 265                         *src = s + nms;
 266                         return (nchr);
 267                 } else if (nb == 0) {
 268                         *src = NULL;
 269                         return (nchr);
 270                 }
 271                 s += nb;
 272                 nms -= nb;
 273                 nchr++;
 274                 dst++;
 275         }
 276         *src = s;
 277         return (nchr);
 278 }
 279
 280 static size_t
 281 _UTF2_wcrtomb(char * __restrict s, wchar_t wc, mbstate_t * __restrict ps, locale_t loc)
 282 {
 283         _UTF2State *us;
 284         unsigned char lead;
 285         int i, len;
 286
 287         us = (_UTF2State *)ps;
 288
 289         if (us->want != 0) {
 290                 errno = EINVAL;
 291                 return ((size_t)-1);
 292         }
 293
 294         if (s == NULL)
 295                 /* Reset to initial shift state (no-op) */
 296                 return (1);
 297
 298         if ((wc & ~0x7f) == 0) {
 299                 /* Fast path for plain ASCII characters. */
 300                 *s = (char)wc;
 301                 return (1);
 302         }
 303
 304         /*
 305          * Determine the number of octets needed to represent this character.
 306          * We always output the shortest sequence possible. Also specify the
 307          * first few bits of the first octet, which contains the information
 308          * about the sequence length.
 309          */
 310         if ((wc & ~0x7f) == 0) {
 311                 lead = 0;
 312                 len = 1;
 313         } else if ((wc & ~0x7ff) == 0) {
 314                 lead = 0xc0;
 315                 len = 2;
 316         } else if ((wc & ~0xffff) == 0) {
 317                 lead = 0xe0;
 318                 len = 3;
 319         } else {
 320                 errno = EILSEQ;
 321                 return ((size_t)-1);
 322         }
 323
 324         /*
 325          * Output the octets representing the character in chunks
 326          * of 6 bits, least significant last. The first octet is
 327          * a special case because it contains the sequence length
 328          * information.
 329          */
 330         for (i = len - 1; i > 0; i--) {
 331                 s[i] = (wc & 0x3f) | 0x80;
 332                 wc >>= 6;
 333         }
 334         *s = (wc & 0xff) | lead;
 335
 336         return (len);
 337 }
 338
 339 static size_t
 340 _UTF2_wcsnrtombs(char * __restrict dst, const wchar_t ** __restrict src,
 341     size_t nwc, size_t len, mbstate_t * __restrict ps, locale_t loc)
 342 {
 343         _UTF2State *us;
 344         char buf[MB_LEN_MAX];
 345         const wchar_t *s;
 346         size_t nbytes;
 347         size_t nb;
 348
 349         us = (_UTF2State *)ps;
 350
 351         if (us->want != 0) {
 352                 errno = EINVAL;
 353                 return ((size_t)-1);
 354         }
 355
 356         s = *src;
 357         nbytes = 0;
 358
 359         if (dst == NULL) {
 360                 while (nwc-- > 0) {
 361                         if (0 <= *s && *s < 0x80)
 362                                 /* Fast path for plain ASCII characters. */
 363                                 nb = 1;
 364                         else if ((nb = _UTF2_wcrtomb(buf, *s, ps, loc)) ==
 365                             (size_t)-1)
 366                                 /* Invalid character - wcrtomb() sets errno. */
 367                                 return ((size_t)-1);
 368                         if (*s == L'\0')
 369                                 return (nbytes + nb - 1);
 370                         s++;
 371                         nbytes += nb;
 372                 }
 373                 return (nbytes);
 374         }
 375
 376         while (len > 0 && nwc-- > 0) {
 377                 if (0 <= *s && *s < 0x80) {
 378                         /* Fast path for plain ASCII characters. */
 379                         nb = 1;
 380                         *dst = *s;
 381                 } else if (len > (size_t)UTF2_MB_CUR_MAX) {
 382                         /* Enough space to translate in-place. */
 383                         if ((nb = _UTF2_wcrtomb(dst, *s, ps, loc)) == (size_t)-1) {
 384                                 *src = s;
 385                                 return ((size_t)-1);
 386                         }
 387                 } else {
 388                         /*
 389                          * May not be enough space; use temp. buffer.
 390                          */
 391                         if ((nb = _UTF2_wcrtomb(buf, *s, ps, loc)) == (size_t)-1) {
 392                                 *src = s;
 393                                 return ((size_t)-1);
 394                         }
 395                         if (nb > (int)len)
 396                                 /* MB sequence for character won't fit. */
 397                                 break;
 398                         memcpy(dst, buf, nb);
 399                 }
 400                 if (*s == L'\0') {
 401                         *src = NULL;
 402                         return (nbytes + nb - 1);
 403                 }
 404                 s++;
 405                 dst += nb;
 406                 len -= nb;
 407                 nbytes += nb;
 408         }
 409         *src = s;
 410         return (nbytes);
 411 }