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

   1
   2 /*
   3  * fmt_rfc2047.c -- decode RFC-2047 header format
   4  *
   5  * $Id$
   6  *
   7  * This code is Copyright (c) 2002, by the authors of nmh.  See the
   8  * COPYRIGHT file in the root directory of the nmh distribution for
   9  * complete copyright information.
  10  */
  11
  12 #include <h/mh.h>
  13 #include <h/utils.h>
  14 #ifdef HAVE_ICONV
  15 #  include <iconv.h>
  16 #  include <errno.h>
  17 #endif
  18
  19 static signed char hexindex[] = {
  20     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  21     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  22     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  23      0, 1, 2, 3, 4, 5, 6, 7, 8, 9,-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     -1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  27     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
  28 };
  29
  30 static signed char index_64[128] = {
  31     -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
  32     -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
  33     -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,62, -1,-1,-1,63,
  34     52,53,54,55, 56,57,58,59, 60,61,-1,-1, -1,-1,-1,-1,
  35     -1, 0, 1, 2,  3, 4, 5, 6,  7, 8, 9,10, 11,12,13,14,
  36     15,16,17,18, 19,20,21,22, 23,24,25,-1, -1,-1,-1,-1,
  37     -1,26,27,28, 29,30,31,32, 33,34,35,36, 37,38,39,40,
  38     41,42,43,44, 45,46,47,48, 49,50,51,-1, -1,-1,-1,-1
  39 };
  40
  41 #define char64(c) (((unsigned char) (c) > 127) ? -1 : index_64[(unsigned char) (c)])
  42
  43 static int
  44 unqp (unsigned char byte1, unsigned char byte2)
  45 {
  46     if (hexindex[byte1] == -1 || hexindex[byte2] == -1)
  47         return -1;
  48     return (hexindex[byte1] << 4 | hexindex[byte2]);
  49 }
  50
  51 /* Check if character is linear whitespace */
  52 #define is_lws(c)  ((c) == ' ' || (c) == '\t' || (c) == '\n')
  53
  54
  55 /*
  56  * Decode the string as a RFC-2047 header field
  57  */
  58
  59 /* Add character to the destination buffer, and bomb out if it fills up */
  60 #define ADDCHR(C) do { *q++ = (C); dstlen--; if (!dstlen) goto buffull; } while (0)
  61
  62 int
  63 decode_rfc2047 (char *str, char *dst, size_t dstlen)
  64 {
  65     char *p, *q, *pp;
  66     char *startofmime, *endofmime;
  67     int c, quoted_printable;
  68     int encoding_found = 0;     /* did we decode anything?                */
  69     int between_encodings = 0;  /* are we between two encodings?          */
  70     int equals_pending = 0;     /* is there a '=' pending?                */
  71     int whitespace = 0;         /* how much whitespace between encodings? */
  72 #ifdef HAVE_ICONV
  73     int use_iconv = 0;          /* are we converting encoding with iconv? */
  74     iconv_t cd = NULL;
  75     int fromutf8 = 0;
  76     char *saveq, *convbuf = NULL;
  77     size_t savedstlen;
  78 #endif
  79
  80     if (!str)
  81         return 0;
  82
  83     /*
  84      * Do a quick and dirty check for the '=' character.
  85      * This should quickly eliminate many cases.
  86      */
  87     if (!strchr (str, '='))
  88         return 0;
  89
  90     for (p = str, q = dst; *p; p++) {
  91
  92         /* reset iconv */
  93 #ifdef HAVE_ICONV
  94         if (use_iconv) {
  95             iconv_close(cd);
  96             use_iconv = 0;
  97         }
  98 #endif
  99         /*
 100          * If we had an '=' character pending from
 101          * last iteration, then add it first.
 102          */
 103         if (equals_pending) {
 104             ADDCHR('=');
 105             equals_pending = 0;
 106             between_encodings = 0;      /* we have added non-whitespace text */
 107         }
 108
 109         if (*p != '=') {
 110             /* count linear whitespace while between encodings */
 111             if (between_encodings && is_lws(*p))
 112                 whitespace++;
 113             else
 114                 between_encodings = 0;  /* we have added non-whitespace text */
 115             ADDCHR(*p);
 116             continue;
 117         }
 118
 119         equals_pending = 1;     /* we have a '=' pending */
 120
 121         /* Check for initial =? */
 122         if (*p == '=' && p[1] && p[1] == '?' && p[2]) {
 123             startofmime = p + 2;
 124
 125             /* Scan ahead for the next '?' character */
 126             for (pp = startofmime; *pp && *pp != '?'; pp++)
 127                 ;
 128
 129             if (!*pp)
 130                 continue;
 131
 132             /* Check if character set can be handled natively */
 133             if (!check_charset(startofmime, pp - startofmime)) {
 134 #ifdef HAVE_ICONV
 135                 /* .. it can't. We'll use iconv then. */
 136                 *pp = '\0';
 137                 cd = iconv_open(get_charset(), startofmime);
 138                 fromutf8 = !mh_strcasecmp(startofmime, "UTF-8");
 139                 *pp = '?';
 140                 if (cd == (iconv_t)-1) continue;
 141                 use_iconv = 1;
 142 #else
 143                 continue;
 144 #endif
 145             }
 146
 147             startofmime = pp + 1;
 148
 149             /* Check for valid encoding type */
 150             if (*startofmime != 'B' && *startofmime != 'b' &&
 151                 *startofmime != 'Q' && *startofmime != 'q')
 152                 continue;
 153
 154             /* Is encoding quoted printable or base64? */
 155             quoted_printable = (*startofmime == 'Q' || *startofmime == 'q');
 156             startofmime++;
 157
 158             /* Check for next '?' character */
 159             if (*startofmime != '?')
 160                 continue;
 161             startofmime++;
 162
 163             /*
 164              * Scan ahead for the ending ?=
 165              *
 166              * While doing this, we will also check if encoded
 167              * word has any embedded linear whitespace.
 168              */
 169             endofmime = NULL;
 170             for (pp = startofmime; *pp && *(pp+1); pp++) {
 171                 if (is_lws(*pp)) {
 172                     break;
 173                 } else if (*pp == '?' && pp[1] == '=') {
 174                     endofmime = pp;
 175                     break;
 176                 }
 177             }
 178             if (is_lws(*pp) || endofmime == NULL)
 179                 continue;
 180
 181             /*
 182              * We've found an encoded word, so we can drop
 183              * the '=' that was pending
 184              */
 185             equals_pending = 0;
 186
 187             /*
 188              * If we are between two encoded words separated only by
 189              * linear whitespace, then we ignore the whitespace.
 190              * We will roll back the buffer the number of whitespace
 191              * characters we've seen since last encoded word.
 192              */
 193             if (between_encodings) {
 194                 q -= whitespace;
 195                 dstlen += whitespace;
 196             }
 197
 198 #ifdef HAVE_ICONV
 199             /*
 200              * empty encoded text. This ensures that we don't
 201              * malloc 0 bytes but skip on to the end
 202              */
 203             if (endofmime == startofmime && use_iconv) {
 204                 use_iconv = 0;
 205                 iconv_close(cd);
 206             }
 207
 208             if (use_iconv) {
 209                 saveq = q;
 210                 savedstlen = dstlen;
 211                 q = convbuf = (char *) mh_xmalloc(endofmime - startofmime);
 212             }
 213 /* ADDCHR2 is for adding characters when q is or might be convbuf:
 214  * in this case on buffer-full we want to run iconv before returning.
 215  * I apologise for the dreadful name.
 216  */
 217 #define ADDCHR2(C) do { *q++ = (C); dstlen--; if (!dstlen) goto iconvbuffull; } while (0)
 218 #else
 219 #define ADDCHR2(C) ADDCHR(C)
 220 #endif
 221
 222             /* Now decode the text */
 223             if (quoted_printable) {
 224                 for (pp = startofmime; pp < endofmime; pp++) {
 225                     if (*pp == '=') {
 226                         c = unqp (pp[1], pp[2]);
 227                         if (c == -1)
 228                             continue;
 229                         if (c != 0)
 230                             *q++ = c;
 231                         pp += 2;
 232                     } else if (*pp == '_') {
 233                         ADDCHR2(' ');
 234                     } else {
 235                         ADDCHR2(*pp);
 236                     }
 237                 }
 238             } else {
 239                 /* base64 */
 240                 int c1, c2, c3, c4;
 241
 242                 pp = startofmime;
 243                 while (pp < endofmime) {
 244                     /* 6 + 2 bits */
 245                     while ((pp < endofmime) &&
 246                            ((c1 = char64(*pp)) == -1)) {
 247                         pp++;
 248                     }
 249                     if (pp < endofmime) {
 250                         pp++;
 251                     }
 252                     while ((pp < endofmime) &&
 253                            ((c2 = char64(*pp)) == -1)) {
 254                         pp++;
 255                     }
 256                     if (pp < endofmime && c1 != -1 && c2 != -1) {
 257                         ADDCHR2((c1 << 2) | (c2 >> 4));
 258                         pp++;
 259                     }
 260                     /* 4 + 4 bits */
 261                     while ((pp < endofmime) &&
 262                            ((c3 = char64(*pp)) == -1)) {
 263                         pp++;
 264                     }
 265                     if (pp < endofmime && c2 != -1 && c3 != -1) {
 266                         ADDCHR2(((c2 & 0xF) << 4) | (c3 >> 2));
 267                         pp++;
 268                     }
 269                     /* 2 + 6 bits */
 270                     while ((pp < endofmime) &&
 271                            ((c4 = char64(*pp)) == -1)) {
 272                         pp++;
 273                     }
 274                     if (pp < endofmime && c3 != -1 && c4 != -1) {
 275                         ADDCHR2(((c3 & 0x3) << 6) | (c4));
 276                         pp++;
 277                     }
 278                 }
 279             }
 280
 281 #ifdef HAVE_ICONV
 282         iconvbuffull:
 283             /* NB that the string at convbuf is not necessarily NUL terminated here:
 284              * q points to the first byte after the valid part.
 285              */
 286             /* Convert to native character set */
 287             if (use_iconv) {
 288                 size_t inbytes = q - convbuf;
 289                 ICONV_CONST char *start = convbuf;
 290
 291                 while (inbytes) {
 292                     if (iconv(cd, &start, &inbytes, &saveq, &savedstlen) ==
 293                             (size_t)-1) {
 294                         if (errno != EILSEQ) break;
 295                         /* character couldn't be converted. we output a `?'
 296                          * and try to carry on which won't work if
 297                          * either encoding was stateful */
 298                         iconv (cd, 0, 0, &saveq, &savedstlen);
 299                         if (!savedstlen)
 300                             break;
 301                         *saveq++ = '?';
 302                         savedstlen--;
 303                         if (!savedstlen)
 304                             break;
 305                         /* skip to next input character */
 306                         if (fromutf8) {
 307                             for (start++;(start < q) && ((*start & 192) == 128);start++)
 308                                 inbytes--;
 309                         } else
 310                             start++, inbytes--;
 311                         if (start >= q)
 312                             break;
 313                     }
 314                 }
 315                 q = saveq;
 316                 /* Stop now if (1) we hit the end of the buffer trying to do
 317                  * MIME decoding and have just iconv-converted a partial string
 318                  * or (2) our iconv-conversion hit the end of the buffer.
 319                  */
 320                 if (!dstlen || !savedstlen)
 321                     goto buffull;
 322                 dstlen = savedstlen;
 323                 free(convbuf);
 324             }
 325 #endif
 326
 327             /*
 328              * Now that we are done decoding this particular
 329              * encoded word, advance string to trailing '='.
 330              */
 331             p = endofmime + 1;
 332
 333             encoding_found = 1;         /* we found (at least 1) encoded word */
 334             between_encodings = 1;      /* we have just decoded something     */
 335             whitespace = 0;             /* re-initialize amount of whitespace */
 336         }
 337     }
 338 #ifdef HAVE_ICONV
 339     if (use_iconv) iconv_close(cd);
 340 #endif
 341
 342     /* If an equals was pending at end of string, add it now. */
 343     if (equals_pending)
 344         ADDCHR('=');
 345     *q = '\0';
 346
 347     return encoding_found;
 348
 349   buffull:
 350     /* q is currently just off the end of the buffer, so rewind to NUL terminate */
 351     q--;
 352     *q = '\0';
 353     return encoding_found;
 354 }