lib/loop_wchar.h

   1 /*
   2  * Copyright (C) 2000-2001 Free Software Foundation, Inc.
   3  * This file is part of the GNU LIBICONV Library.
   4  *
   5  * The GNU LIBICONV Library is free software; you can redistribute it
   6  * and/or modify it under the terms of the GNU Library General Public
   7  * License as published by the Free Software Foundation; either version 2
   8  * of the License, or (at your option) any later version.
   9  *
  10  * The GNU LIBICONV Library is distributed in the hope that it will be
  11  * useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * Library General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU Library General Public
  16  * License along with the GNU LIBICONV Library; see the file COPYING.LIB.
  17  * If not, write to the Free Software Foundation, Inc., 59 Temple Place -
  18  * Suite 330, Boston, MA 02111-1307, USA.
  19  */
  20
  21 /* This file defines three conversion loops:
  22      - from wchar_t to anything else,
  23      - from anything else to wchar_t,
  24      - from wchar_t to wchar_t.
  25  */
  26
  27 #if HAVE_WCRTOMB || HAVE_MBRTOWC
  28 # include <wchar.h>
  29 # define BUF_SIZE 64  /* assume MB_LEN_MAX <= 64 */
  30   /* Some systems, like BeOS, have multibyte encodings but lack mbstate_t.  */
  31   extern size_t mbrtowc ();
  32 # ifdef mbstate_t
  33 #  define mbrtowc(pwc, s, n, ps) (mbrtowc)(pwc, s, n, 0)
  34 #  define mbsinit(ps) 1
  35 # endif
  36 # ifndef mbsinit
  37 #  if !HAVE_MBSINIT
  38 #   define mbsinit(ps) 1
  39 #  endif
  40 # endif
  41 #else
  42 # ifndef mbstate_t
  43    typedef int mbstate_t;
  44 # endif
  45 #endif
  46
  47 /*
  48  * The first two conversion loops have an extended conversion descriptor.
  49  */
  50 struct wchar_conv_struct {
  51   struct conv_struct parent;
  52   mbstate_t state;
  53 };
  54
  55
  56 #if HAVE_WCRTOMB
  57
  58 /* From wchar_t to anything else. */
  59
  60 static size_t wchar_from_loop_convert (iconv_t icd,
  61                                        const char* * inbuf, size_t *inbytesleft,
  62                                        char* * outbuf, size_t *outbytesleft)
  63 {
  64   struct wchar_conv_struct * wcd = (struct wchar_conv_struct *) icd;
  65   size_t result = 0;
  66   while (*inbytesleft >= sizeof(wchar_t)) {
  67     const wchar_t * inptr = (const wchar_t *) *inbuf;
  68     size_t inleft = *inbytesleft;
  69     char buf[BUF_SIZE];
  70     mbstate_t state = wcd->state;
  71     size_t bufcount = 0;
  72     while (inleft >= sizeof(wchar_t)) {
  73       /* Convert one wchar_t to multibyte representation. */
  74       size_t count = wcrtomb(buf+bufcount,*inptr,&state);
  75       if (count == (size_t)(-1)) {
  76         /* Invalid input. */
  77         errno = EILSEQ;
  78         return -1;
  79       }
  80       inptr++;
  81       inleft -= sizeof(wchar_t);
  82       bufcount += count;
  83       if (count == 0) {
  84         /* Continue, append next wchar_t. */
  85       } else {
  86         /* Attempt to convert the accumulated multibyte representations
  87            to the target encoding. */
  88         const char* bufptr = buf;
  89         size_t bufleft = bufcount;
  90         char* outptr = *outbuf;
  91         size_t outleft = *outbytesleft;
  92         size_t res = unicode_loop_convert(&wcd->parent,
  93                                           &bufptr,&bufleft,
  94                                           &outptr,&outleft);
  95         if (res == (size_t)(-1)) {
  96           if (errno == EILSEQ)
  97             /* Invalid input. */
  98             return -1;
  99           else if (errno == E2BIG)
 100             /* Output buffer too small. */
 101             return -1;
 102           else if (errno == EINVAL) {
 103             /* Continue, append next wchar_t, but avoid buffer overrun. */
 104             if (bufcount + MB_CUR_MAX > BUF_SIZE)
 105               abort();
 106           } else
 107             abort();
 108         } else {
 109           /* Successful conversion. */
 110           wcd->state = state;
 111           *inbuf = (const char *) inptr;
 112           *inbytesleft = inleft;
 113           *outbuf = outptr;
 114           *outbytesleft = outleft;
 115           result += res;
 116           break;
 117         }
 118       }
 119     }
 120   }
 121   return result;
 122 }
 123
 124 static size_t wchar_from_loop_reset (iconv_t icd,
 125                                      char* * outbuf, size_t *outbytesleft)
 126 {
 127   struct wchar_conv_struct * wcd = (struct wchar_conv_struct *) icd;
 128   if (outbuf == NULL || *outbuf == NULL) {
 129     /* Reset the states. */
 130     memset(&wcd->state,'\0',sizeof(mbstate_t));
 131     return unicode_loop_reset(&wcd->parent,NULL,NULL);
 132   } else {
 133     if (!mbsinit(&wcd->state)) {
 134       mbstate_t state = wcd->state;
 135       char buf[BUF_SIZE];
 136       size_t bufcount = wcrtomb(buf,(wchar_t)0,&state);
 137       if (bufcount == (size_t)(-1) || bufcount == 0 || buf[bufcount-1] != '\0')
 138         abort();
 139       else {
 140         const char* bufptr = buf;
 141         size_t bufleft = bufcount-1;
 142         char* outptr = *outbuf;
 143         size_t outleft = *outbytesleft;
 144         size_t res = unicode_loop_convert(&wcd->parent,
 145                                           &bufptr,&bufleft,
 146                                           &outptr,&outleft);
 147         if (res == (size_t)(-1)) {
 148           if (errno == E2BIG)
 149             return -1;
 150           else
 151             abort();
 152         } else {
 153           res = unicode_loop_reset(&wcd->parent,&outptr,&outleft);
 154           if (res == (size_t)(-1))
 155             return res;
 156           else {
 157             /* Successful. */
 158             wcd->state = state;
 159             *outbuf = outptr;
 160             *outbytesleft = outleft;
 161             return 0;
 162           }
 163         }
 164       }
 165     } else
 166       return unicode_loop_reset(&wcd->parent,outbuf,outbytesleft);
 167   }
 168 }
 169
 170 #endif
 171
 172
 173 #if HAVE_MBRTOWC
 174
 175 /* From anything else to wchar_t. */
 176
 177 static size_t wchar_to_loop_convert (iconv_t icd,
 178                                      const char* * inbuf, size_t *inbytesleft,
 179                                      char* * outbuf, size_t *outbytesleft)
 180 {
 181   struct wchar_conv_struct * wcd = (struct wchar_conv_struct *) icd;
 182   size_t result = 0;
 183   while (*inbytesleft > 0) {
 184     size_t incount;
 185     for (incount = 1; incount <= *inbytesleft; incount++) {
 186       char buf[BUF_SIZE];
 187       const char* inptr = *inbuf;
 188       size_t inleft = incount;
 189       char* bufptr = buf;
 190       size_t bufleft = BUF_SIZE;
 191       size_t res = unicode_loop_convert(&wcd->parent,
 192                                         &inptr,&inleft,
 193                                         &bufptr,&bufleft);
 194       if (res == (size_t)(-1)) {
 195         if (errno == EILSEQ)
 196           /* Invalid input. */
 197           return -1;
 198         else if (errno == EINVAL) {
 199           /* Incomplete input. Next try with one more input byte. */
 200         } else
 201           /* E2BIG shouldn't occur. */
 202           abort();
 203       } else {
 204         /* Successful conversion. */
 205         size_t bufcount = bufptr-buf; /* = BUF_SIZE-bufleft */
 206         mbstate_t state = wcd->state;
 207         wchar_t wc;
 208         res = mbrtowc(&wc,buf,bufcount,&state);
 209         if (res == (size_t)(-2)) {
 210           /* Next try with one more input byte. */
 211         } else if (res == (size_t)(-1)) {
 212           /* Invalid input. */
 213           return -1;
 214         } else {
 215           if (*outbytesleft < sizeof(wchar_t)) {
 216             errno = E2BIG;
 217             return -1;
 218           }
 219           *(wchar_t*) *outbuf = wc;
 220           *outbuf += sizeof(wchar_t);
 221           *outbytesleft -= sizeof(wchar_t);
 222           *inbuf += incount;
 223           *inbytesleft -= incount;
 224           result += res;
 225           break;
 226         }
 227       }
 228     }
 229   }
 230   return result;
 231 }
 232
 233 static size_t wchar_to_loop_reset (iconv_t icd,
 234                                    char* * outbuf, size_t *outbytesleft)
 235 {
 236   struct wchar_conv_struct * wcd = (struct wchar_conv_struct *) icd;
 237   size_t res = unicode_loop_reset(&wcd->parent,outbuf,outbytesleft);
 238   if (res == (size_t)(-1))
 239     return res;
 240   memset(&wcd->state,0,sizeof(mbstate_t));
 241   return 0;
 242 }
 243
 244 #endif
 245
 246
 247 /* From wchar_t to wchar_t. */
 248
 249 static size_t wchar_id_loop_convert (iconv_t icd,
 250                                      const char* * inbuf, size_t *inbytesleft,
 251                                      char* * outbuf, size_t *outbytesleft)
 252 {
 253   const wchar_t* inptr = (const wchar_t*) *inbuf;
 254   size_t inleft = *inbytesleft / sizeof(wchar_t);
 255   wchar_t* outptr = (wchar_t*) *outbuf;
 256   size_t outleft = *outbytesleft / sizeof(wchar_t);
 257   size_t count = (inleft <= outleft ? inleft : outleft);
 258   if (count > 0) {
 259     *inbytesleft -= count * sizeof(wchar_t);
 260     *outbytesleft -= count * sizeof(wchar_t);
 261     do
 262       *outptr++ = *inptr++;
 263     while (--count > 0);
 264     *inbuf = (const char*) inptr;
 265     *outbuf = (char*) outptr;
 266   }
 267   return 0;
 268 }
 269
 270 static size_t wchar_id_loop_reset (iconv_t icd,
 271                                    char* * outbuf, size_t *outbytesleft)
 272 {
 273   return 0;
 274 }