diplodocus.org Git - nmh/blob - sbr/fmt_rfc2047.c

   1
   2 /*
   3  * fmt_rfc2047.c -- decode RFC-2047 header format
   4  *
   5  * This code is Copyright (c) 2002, by the authors of nmh.  See the
   6  * COPYRIGHT file in the root directory of the nmh distribution for
   7  * complete copyright information.
   8  */
   9
  10 #include <h/mh.h>
  11 #include <h/utils.h>
  12 #ifdef HAVE_ICONV
  13 #  include <iconv.h>
  14 #  include <errno.h>
  15 #endif
  16
  17 static signed char hexindex[] = {
  18     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  19     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  20     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  21      0, 1, 2, 3, 4, 5, 6, 7, 8, 9,-1,-1,-1,-1,-1,-1,
  22     -1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  23     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  24     -1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  25     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
  26 };
  27
  28 static signed char index_64[128] = {
  29     -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
  30     -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
  31     -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,62, -1,-1,-1,63,
  32     52,53,54,55, 56,57,58,59, 60,61,-1,-1, -1,-1,-1,-1,
  33     -1, 0, 1, 2,  3, 4, 5, 6,  7, 8, 9,10, 11,12,13,14,
  34     15,16,17,18, 19,20,21,22, 23,24,25,-1, -1,-1,-1,-1,
  35     -1,26,27,28, 29,30,31,32, 33,34,35,36, 37,38,39,40,
  36     41,42,43,44, 45,46,47,48, 49,50,51,-1, -1,-1,-1,-1
  37 };
  38
  39 #define char64(c) (((unsigned char) (c) > 127) ? -1 : index_64[(unsigned char) (c)])
  40
  41 static int
  42 unqp (unsigned char byte1, unsigned char byte2)
  43 {
  44     if (hexindex[byte1] == -1 || hexindex[byte2] == -1)
  45         return -1;
  46     return (hexindex[byte1] << 4 | hexindex[byte2]);
  47 }
  48
  49 /* Check if character is linear whitespace */
  50 #define is_lws(c)  ((c) == ' ' || (c) == '\t' || (c) == '\n')
  51
  52
  53 /*
  54  * Decode the string as a RFC-2047 header field
  55  */
  56
  57 /* Add character to the destination buffer, and bomb out if it fills up */
  58 #define ADDCHR(C) do { *q++ = (C); dstlen--; if (!dstlen) goto buffull; } while (0)
  59
  60 int
  61 decode_rfc2047 (char *str, char *dst, size_t dstlen)
  62 {
  63     char *p, *q, *pp;
  64     char *startofmime, *endofmime;
  65     int c, quoted_printable;
  66     int encoding_found = 0;     /* did we decode anything?                */
  67     int between_encodings = 0;  /* are we between two encodings?          */
  68     int equals_pending = 0;     /* is there a '=' pending?                */
  69     int whitespace = 0;         /* how much whitespace between encodings? */
  70 #ifdef HAVE_ICONV
  71     int use_iconv = 0;          /* are we converting encoding with iconv? */
  72     iconv_t cd = NULL;
  73     int fromutf8 = 0;
  74     char *saveq, *convbuf = NULL;
  75     size_t savedstlen;
  76 #endif
  77
  78     if (!str)
  79         return 0;
  80
  81     /*
  82      * Do a quick and dirty check for the '=' character.
  83      * This should quickly eliminate many cases.
  84      */
  85     if (!strchr (str, '='))
  86         return 0;
  87
  88     for (p = str, q = dst; *p; p++) {
  89
  90         /* reset iconv */
  91 #ifdef HAVE_ICONV
  92         if (use_iconv) {
  93             iconv_close(cd);
  94             use_iconv = 0;
  95         }
  96 #endif
  97         /*
  98          * If we had an '=' character pending from
  99          * last iteration, then add it first.
 100          */
 101         if (equals_pending) {
 102             ADDCHR('=');
 103             equals_pending = 0;
 104             between_encodings = 0;      /* we have added non-whitespace text */
 105         }
 106
 107         if (*p != '=') {
 108             /* count linear whitespace while between encodings */
 109             if (between_encodings && is_lws(*p))
 110                 whitespace++;
 111             else
 112                 between_encodings = 0;  /* we have added non-whitespace text */
 113             ADDCHR(*p);
 114             continue;
 115         }
 116
 117         equals_pending = 1;     /* we have a '=' pending */
 118
 119         /* Check for initial =? */
 120         if (*p == '=' && p[1] && p[1] == '?' && p[2]) {
 121             startofmime = p + 2;
 122
 123             /* Scan ahead for the next '?' character */
 124             for (pp = startofmime; *pp && *pp != '?'; pp++)
 125                 ;
 126
 127             if (!*pp)
 128                 continue;
 129
 130             /* Check if character set can be handled natively */
 131             if (!check_charset(startofmime, pp - startofmime)) {
 132 #ifdef HAVE_ICONV
 133                 /* .. it can't. We'll use iconv then. */
 134                 *pp = '\0';
 135                 cd = iconv_open(get_charset(), startofmime);
 136                 fromutf8 = !mh_strcasecmp(startofmime, "UTF-8");
 137                 *pp = '?';
 138                 if (cd == (iconv_t)-1) continue;
 139                 use_iconv = 1;
 140 #else
 141                 continue;
 142 #endif
 143             }
 144
 145             startofmime = pp + 1;
 146
 147             /* Check for valid encoding type */
 148             if (*startofmime != 'B' && *startofmime != 'b' &&
 149                 *startofmime != 'Q' && *startofmime != 'q')
 150                 continue;
 151
 152             /* Is encoding quoted printable or base64? */
 153             quoted_printable = (*startofmime == 'Q' || *startofmime == 'q');
 154             startofmime++;
 155
 156             /* Check for next '?' character */
 157             if (*startofmime != '?')
 158                 continue;
 159             startofmime++;
 160
 161             /*
 162              * Scan ahead for the ending ?=
 163              *
 164              * While doing this, we will also check if encoded
 165              * word has any embedded linear whitespace.
 166              */
 167             endofmime = NULL;
 168             for (pp = startofmime; *pp && *(pp+1); pp++) {
 169                 if (is_lws(*pp)) {
 170                     break;
 171                 } else if (*pp == '?' && pp[1] == '=') {
 172                     endofmime = pp;
 173                     break;
 174                 }
 175             }
 176             if (is_lws(*pp) || endofmime == NULL)
 177                 continue;
 178
 179             /*
 180              * We've found an encoded word, so we can drop
 181              * the '=' that was pending
 182              */
 183             equals_pending = 0;
 184
 185             /*
 186              * If we are between two encoded words separated only by
 187              * linear whitespace, then we ignore the whitespace.
 188              * We will roll back the buffer the number of whitespace
 189              * characters we've seen since last encoded word.
 190              */
 191             if (between_encodings) {
 192                 q -= whitespace;
 193                 dstlen += whitespace;
 194             }
 195
 196 #ifdef HAVE_ICONV
 197             /*
 198              * empty encoded text. This ensures that we don't
 199              * malloc 0 bytes but skip on to the end
 200              */
 201             if (endofmime == startofmime && use_iconv) {
 202                 use_iconv = 0;
 203                 iconv_close(cd);
 204             }
 205
 206             if (use_iconv) {
 207                 saveq = q;
 208                 savedstlen = dstlen;
 209                 q = convbuf = (char *) mh_xmalloc(endofmime - startofmime);
 210             }
 211 /* ADDCHR2 is for adding characters when q is or might be convbuf:
 212  * in this case on buffer-full we want to run iconv before returning.
 213  * I apologise for the dreadful name.
 214  */
 215 #define ADDCHR2(C) do { *q++ = (C); dstlen--; if (!dstlen) goto iconvbuffull; } while (0)
 216 #else
 217 #define ADDCHR2(C) ADDCHR(C)
 218 #endif
 219
 220             /* Now decode the text */
 221             if (quoted_printable) {
 222                 for (pp = startofmime; pp < endofmime; pp++) {
 223                     if (*pp == '=') {
 224                         c = unqp (pp[1], pp[2]);
 225                         if (c == -1)
 226                             continue;
 227                         if (c != 0)
 228                             *q++ = c;
 229                         pp += 2;
 230                     } else if (*pp == '_') {
 231                         ADDCHR2(' ');
 232                     } else {
 233                         ADDCHR2(*pp);
 234                     }
 235                 }
 236             } else {
 237                 /* base64 */
 238                 int c1, c2, c3, c4;
 239                 c1 = c2 = c3 = c4 = -1;
 240
 241                 pp = startofmime;
 242                 while (pp < endofmime) {
 243                     /* 6 + 2 bits */
 244                     while ((pp < endofmime) &&
 245                            ((c1 = char64(*pp)) == -1)) {
 246                         pp++;
 247                     }
 248                     if (pp < endofmime) {
 249                         pp++;
 250                     }
 251                     while ((pp < endofmime) &&
 252                            ((c2 = char64(*pp)) == -1)) {
 253                         pp++;
 254                     }
 255                     if (pp < endofmime && c1 != -1 && c2 != -1) {
 256                         ADDCHR2((c1 << 2) | (c2 >> 4));
 257                         pp++;
 258                     }
 259                     /* 4 + 4 bits */
 260                     while ((pp < endofmime) &&
 261                            ((c3 = char64(*pp)) == -1)) {
 262                         pp++;
 263                     }
 264                     if (pp < endofmime && c2 != -1 && c3 != -1) {
 265                         ADDCHR2(((c2 & 0xF) << 4) | (c3 >> 2));
 266                         pp++;
 267                     }
 268                     /* 2 + 6 bits */
 269                     while ((pp < endofmime) &&
 270                            ((c4 = char64(*pp)) == -1)) {
 271                         pp++;
 272                     }
 273                     if (pp < endofmime && c3 != -1 && c4 != -1) {
 274                         ADDCHR2(((c3 & 0x3) << 6) | (c4));
 275                         pp++;
 276                     }
 277                 }
 278             }
 279
 280 #ifdef HAVE_ICONV
 281         iconvbuffull:
 282             /* NB that the string at convbuf is not necessarily NUL terminated here:
 283              * q points to the first byte after the valid part.
 284              */
 285             /* Convert to native character set */
 286             if (use_iconv) {
 287                 size_t inbytes = q - convbuf;
 288                 ICONV_CONST char *start = convbuf;
 289
 290                 while (inbytes) {
 291                     if (iconv(cd, &start, &inbytes, &saveq, &savedstlen) ==
 292                             (size_t)-1) {
 293                         if (errno != EILSEQ) break;
 294                         /* character couldn't be converted. we output a `?'
 295                          * and try to carry on which won't work if
 296                          * either encoding was stateful */
 297                         iconv (cd, 0, 0, &saveq, &savedstlen);
 298                         if (!savedstlen)
 299                             break;
 300                         *saveq++ = '?';
 301                         savedstlen--;
 302                         if (!savedstlen)
 303                             break;
 304                         /* skip to next input character */
 305                         if (fromutf8) {
 306                             for (start++;(start < q) && ((*start & 192) == 128);start++)
 307                                 inbytes--;
 308                         } else
 309                             start++, inbytes--;
 310                         if (start >= q)
 311                             break;
 312                     }
 313                 }
 314                 q = saveq;
 315                 /* Stop now if (1) we hit the end of the buffer trying to do
 316                  * MIME decoding and have just iconv-converted a partial string
 317                  * or (2) our iconv-conversion hit the end of the buffer.
 318                  */
 319                 if (!dstlen || !savedstlen)
 320                     goto buffull;
 321                 dstlen = savedstlen;
 322                 free(convbuf);
 323             }
 324 #endif
 325
 326             /*
 327              * Now that we are done decoding this particular
 328              * encoded word, advance string to trailing '='.
 329              */
 330             p = endofmime + 1;
 331
 332             encoding_found = 1;         /* we found (at least 1) encoded word */
 333             between_encodings = 1;      /* we have just decoded something     */
 334             whitespace = 0;             /* re-initialize amount of whitespace */
 335         }
 336     }
 337 #ifdef HAVE_ICONV
 338     if (use_iconv) iconv_close(cd);
 339 #endif
 340
 341     /* If an equals was pending at end of string, add it now. */
 342     if (equals_pending)
 343         ADDCHR('=');
 344     *q = '\0';
 345
 346     return encoding_found;
 347
 348   buffull:
 349     /* q is currently just off the end of the buffer, so rewind to NUL terminate */
 350     q--;
 351     *q = '\0';
 352     return encoding_found;
 353 }