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, *endofcharset;
  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             /*
 131              * RFC 2231 specifies that language information can appear
 132              * in a charset specification like so:
 133              *
 134              * =?us-ascii*en?Q?Foo?=
 135              *
 136              * Right now we don't use language information, so ignore it.
 137              */
 138
 139             for (endofcharset = startofmime;
 140                         *endofcharset != '*' && endofcharset < pp;
 141                                                         endofcharset++)
 142                 ;
 143
 144             /* Check if character set can be handled natively */
 145             if (!check_charset(startofmime, endofcharset - startofmime)) {
 146 #ifdef HAVE_ICONV
 147                 /* .. it can't. We'll use iconv then. */
 148                 *endofcharset = '\0';
 149                 cd = iconv_open(get_charset(), startofmime);
 150                 fromutf8 = !strcasecmp(startofmime, "UTF-8");
 151                 *pp = '?';
 152                 if (cd == (iconv_t)-1) continue;
 153                 use_iconv = 1;
 154 #else
 155                 continue;
 156 #endif
 157             }
 158
 159             startofmime = pp + 1;
 160
 161             /* Check for valid encoding type */
 162             if (*startofmime != 'B' && *startofmime != 'b' &&
 163                 *startofmime != 'Q' && *startofmime != 'q')
 164                 continue;
 165
 166             /* Is encoding quoted printable or base64? */
 167             quoted_printable = (*startofmime == 'Q' || *startofmime == 'q');
 168             startofmime++;
 169
 170             /* Check for next '?' character */
 171             if (*startofmime != '?')
 172                 continue;
 173             startofmime++;
 174
 175             /*
 176              * Scan ahead for the ending ?=
 177              *
 178              * While doing this, we will also check if encoded
 179              * word has any embedded linear whitespace.
 180              */
 181             endofmime = NULL;
 182             for (pp = startofmime; *pp && *(pp+1); pp++) {
 183                 if (is_lws(*pp)) {
 184                     break;
 185                 } else if (*pp == '?' && pp[1] == '=') {
 186                     endofmime = pp;
 187                     break;
 188                 }
 189             }
 190             if (is_lws(*pp) || endofmime == NULL)
 191                 continue;
 192
 193             /*
 194              * We've found an encoded word, so we can drop
 195              * the '=' that was pending
 196              */
 197             equals_pending = 0;
 198
 199             /*
 200              * If we are between two encoded words separated only by
 201              * linear whitespace, then we ignore the whitespace.
 202              * We will roll back the buffer the number of whitespace
 203              * characters we've seen since last encoded word.
 204              */
 205             if (between_encodings) {
 206                 q -= whitespace;
 207                 dstlen += whitespace;
 208             }
 209
 210 #ifdef HAVE_ICONV
 211             /*
 212              * empty encoded text. This ensures that we don't
 213              * malloc 0 bytes but skip on to the end
 214              */
 215             if (endofmime == startofmime && use_iconv) {
 216                 use_iconv = 0;
 217                 iconv_close(cd);
 218             }
 219
 220             if (use_iconv) {
 221                 saveq = q;
 222                 savedstlen = dstlen;
 223                 q = convbuf = (char *) mh_xmalloc(endofmime - startofmime);
 224             }
 225 /* ADDCHR2 is for adding characters when q is or might be convbuf:
 226  * in this case on buffer-full we want to run iconv before returning.
 227  * I apologise for the dreadful name.
 228  */
 229 #define ADDCHR2(C) do { *q++ = (C); dstlen--; if (!dstlen) goto iconvbuffull; } while (0)
 230 #else
 231 #define ADDCHR2(C) ADDCHR(C)
 232 #endif
 233
 234             /* Now decode the text */
 235             if (quoted_printable) {
 236                 for (pp = startofmime; pp < endofmime; pp++) {
 237                     if (*pp == '=') {
 238                         c = unqp (pp[1], pp[2]);
 239                         if (c == -1)
 240                             continue;
 241                         if (c != 0)
 242                             *q++ = c;
 243                         pp += 2;
 244                     } else if (*pp == '_') {
 245                         ADDCHR2(' ');
 246                     } else {
 247                         ADDCHR2(*pp);
 248                     }
 249                 }
 250             } else {
 251                 /* base64 */
 252                 int c1, c2, c3, c4;
 253                 c1 = c2 = c3 = c4 = -1;
 254
 255                 pp = startofmime;
 256                 while (pp < endofmime) {
 257                     /* 6 + 2 bits */
 258                     while ((pp < endofmime) &&
 259                            ((c1 = char64(*pp)) == -1)) {
 260                         pp++;
 261                     }
 262                     if (pp < endofmime) {
 263                         pp++;
 264                     }
 265                     while ((pp < endofmime) &&
 266                            ((c2 = char64(*pp)) == -1)) {
 267                         pp++;
 268                     }
 269                     if (pp < endofmime && c1 != -1 && c2 != -1) {
 270                         ADDCHR2((c1 << 2) | (c2 >> 4));
 271                         pp++;
 272                     }
 273                     /* 4 + 4 bits */
 274                     while ((pp < endofmime) &&
 275                            ((c3 = char64(*pp)) == -1)) {
 276                         pp++;
 277                     }
 278                     if (pp < endofmime && c2 != -1 && c3 != -1) {
 279                         ADDCHR2(((c2 & 0xF) << 4) | (c3 >> 2));
 280                         pp++;
 281                     }
 282                     /* 2 + 6 bits */
 283                     while ((pp < endofmime) &&
 284                            ((c4 = char64(*pp)) == -1)) {
 285                         pp++;
 286                     }
 287                     if (pp < endofmime && c3 != -1 && c4 != -1) {
 288                         ADDCHR2(((c3 & 0x3) << 6) | (c4));
 289                         pp++;
 290                     }
 291                 }
 292             }
 293
 294 #ifdef HAVE_ICONV
 295         iconvbuffull:
 296             /* NB that the string at convbuf is not necessarily NUL terminated here:
 297              * q points to the first byte after the valid part.
 298              */
 299             /* Convert to native character set */
 300             if (use_iconv) {
 301                 size_t inbytes = q - convbuf;
 302                 ICONV_CONST char *start = convbuf;
 303
 304                 while (inbytes) {
 305                     if (iconv(cd, &start, &inbytes, &saveq, &savedstlen) ==
 306                             (size_t)-1) {
 307                         if (errno != EILSEQ) break;
 308                         /* character couldn't be converted. we output a `?'
 309                          * and try to carry on which won't work if
 310                          * either encoding was stateful */
 311                         iconv (cd, 0, 0, &saveq, &savedstlen);
 312                         if (!savedstlen)
 313                             break;
 314                         *saveq++ = '?';
 315                         savedstlen--;
 316                         if (!savedstlen)
 317                             break;
 318                         /* skip to next input character */
 319                         if (fromutf8) {
 320                             for (++start, --inbytes;
 321                                  start < q  &&  (*start & 192) == 128;
 322                                  ++start, --inbytes)
 323                                 continue;
 324                         } else
 325                             start++, inbytes--;
 326                         if (start >= q)
 327                             break;
 328                     }
 329                 }
 330                 q = saveq;
 331                 /* Stop now if (1) we hit the end of the buffer trying to do
 332                  * MIME decoding and have just iconv-converted a partial string
 333                  * or (2) our iconv-conversion hit the end of the buffer.
 334                  */
 335                 if (!dstlen || !savedstlen)
 336                     goto buffull;
 337                 dstlen = savedstlen;
 338                 free(convbuf);
 339             }
 340 #endif
 341
 342             /*
 343              * Now that we are done decoding this particular
 344              * encoded word, advance string to trailing '='.
 345              */
 346             p = endofmime + 1;
 347
 348             encoding_found = 1;         /* we found (at least 1) encoded word */
 349             between_encodings = 1;      /* we have just decoded something     */
 350             whitespace = 0;             /* re-initialize amount of whitespace */
 351         }
 352     }
 353 #ifdef HAVE_ICONV
 354     if (use_iconv) iconv_close(cd);
 355 #endif
 356
 357     /* If an equals was pending at end of string, add it now. */
 358     if (equals_pending)
 359         ADDCHR('=');
 360     *q = '\0';
 361
 362     return encoding_found;
 363
 364   buffull:
 365     /* q is currently just off the end of the buffer, so rewind to NUL terminate */
 366     q--;
 367     *q = '\0';
 368     return encoding_found;
 369 }