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 #endif
  15
  16 static signed char hexindex[] = {
  17     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  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      0, 1, 2, 3, 4, 5, 6, 7, 8, 9,-1,-1,-1,-1,-1,-1,
  21     -1,10,11,12,13,14,15,-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     -1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  24     -1,-1,-1,-1,-1,-1,-1,-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,-1,-1,-1,-1,-1,-1,-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     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
  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,-1,-1,-1,-1,-1,
  32     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
  33 };
  34
  35 static signed char index_64[128] = {
  36     -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
  37     -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
  38     -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,62, -1,-1,-1,63,
  39     52,53,54,55, 56,57,58,59, 60,61,-1,-1, -1,-1,-1,-1,
  40     -1, 0, 1, 2,  3, 4, 5, 6,  7, 8, 9,10, 11,12,13,14,
  41     15,16,17,18, 19,20,21,22, 23,24,25,-1, -1,-1,-1,-1,
  42     -1,26,27,28, 29,30,31,32, 33,34,35,36, 37,38,39,40,
  43     41,42,43,44, 45,46,47,48, 49,50,51,-1, -1,-1,-1,-1
  44 };
  45
  46 #define char64(c) (((unsigned char) (c) > 127) ? -1 : index_64[(unsigned char) (c)])
  47
  48 /*
  49  * Decode two quoted-pair characters
  50  */
  51
  52 int
  53 decode_qp (unsigned char byte1, unsigned char byte2)
  54 {
  55     if (hexindex[byte1] == -1 || hexindex[byte2] == -1)
  56         return -1;
  57     return (hexindex[byte1] << 4 | hexindex[byte2]);
  58 }
  59
  60 /* Check if character is linear whitespace */
  61 #define is_lws(c)  ((c) == ' ' || (c) == '\t' || (c) == '\n')
  62
  63
  64 /*
  65  * Decode the string as a RFC-2047 header field
  66  */
  67
  68 /* Add character to the destination buffer, and bomb out if it fills up */
  69 #define ADDCHR(C) do { *q++ = (C); dstlen--; if (!dstlen) goto buffull; } while (0)
  70
  71 int
  72 decode_rfc2047 (char *str, char *dst, size_t dstlen)
  73 {
  74     char *p, *q, *pp;
  75     char *startofmime, *endofmime, *endofcharset;
  76     int c, quoted_printable;
  77     int encoding_found = 0;     /* did we decode anything?                */
  78     int between_encodings = 0;  /* are we between two encodings?          */
  79     int equals_pending = 0;     /* is there a '=' pending?                */
  80     int whitespace = 0;         /* how much whitespace between encodings? */
  81 #ifdef HAVE_ICONV
  82     int use_iconv = 0;          /* are we converting encoding with iconv? */
  83     iconv_t cd = NULL;
  84     int fromutf8 = 0;
  85     char *saveq, *convbuf = NULL;
  86     size_t savedstlen;
  87 #endif
  88
  89     if (!str)
  90         return 0;
  91
  92     /*
  93      * Do a quick and dirty check for the '=' character.
  94      * This should quickly eliminate many cases.
  95      */
  96     if (!strchr (str, '='))
  97         return 0;
  98
  99     for (p = str, q = dst; *p; p++) {
 100
 101         /* reset iconv */
 102 #ifdef HAVE_ICONV
 103         if (use_iconv) {
 104             iconv_close(cd);
 105             use_iconv = 0;
 106         }
 107 #endif
 108         /*
 109          * If we had an '=' character pending from
 110          * last iteration, then add it first.
 111          */
 112         if (equals_pending) {
 113             ADDCHR('=');
 114             equals_pending = 0;
 115             between_encodings = 0;      /* we have added non-whitespace text */
 116         }
 117
 118         if (*p != '=') {
 119             /* count linear whitespace while between encodings */
 120             if (between_encodings && is_lws(*p))
 121                 whitespace++;
 122             else
 123                 between_encodings = 0;  /* we have added non-whitespace text */
 124             ADDCHR(*p);
 125             continue;
 126         }
 127
 128         equals_pending = 1;     /* we have a '=' pending */
 129
 130         /* Check for initial =? */
 131         if (*p == '=' && p[1] && p[1] == '?' && p[2]) {
 132             startofmime = p + 2;
 133
 134             /* Scan ahead for the next '?' character */
 135             for (pp = startofmime; *pp && *pp != '?'; pp++)
 136                 ;
 137
 138             if (!*pp)
 139                 continue;
 140
 141             /*
 142              * RFC 2231 specifies that language information can appear
 143              * in a charset specification like so:
 144              *
 145              * =?us-ascii*en?Q?Foo?=
 146              *
 147              * Right now we don't use language information, so ignore it.
 148              */
 149
 150             for (endofcharset = startofmime;
 151                         *endofcharset != '*' && endofcharset < pp;
 152                                                         endofcharset++)
 153                 ;
 154
 155             /* Check if character set can be handled natively */
 156             if (!check_charset(startofmime, endofcharset - startofmime)) {
 157 #ifdef HAVE_ICONV
 158                 /* .. it can't. We'll use iconv then. */
 159                 *endofcharset = '\0';
 160                 cd = iconv_open(get_charset(), startofmime);
 161                 fromutf8 = !strcasecmp(startofmime, "UTF-8");
 162                 *pp = '?';
 163                 if (cd == (iconv_t)-1) continue;
 164                 use_iconv = 1;
 165 #else
 166                 continue;
 167 #endif
 168             }
 169
 170             startofmime = pp + 1;
 171
 172             /* Check for valid encoding type */
 173             if (*startofmime != 'B' && *startofmime != 'b' &&
 174                 *startofmime != 'Q' && *startofmime != 'q')
 175                 continue;
 176
 177             /* Is encoding quoted printable or base64? */
 178             quoted_printable = (*startofmime == 'Q' || *startofmime == 'q');
 179             startofmime++;
 180
 181             /* Check for next '?' character */
 182             if (*startofmime != '?')
 183                 continue;
 184             startofmime++;
 185
 186             /*
 187              * Scan ahead for the ending ?=
 188              *
 189              * While doing this, we will also check if encoded
 190              * word has any embedded linear whitespace.
 191              */
 192             endofmime = NULL;
 193             for (pp = startofmime; *pp && *(pp+1); pp++) {
 194                 if (is_lws(*pp)) {
 195                     break;
 196                 } else if (*pp == '?' && pp[1] == '=') {
 197                     endofmime = pp;
 198                     break;
 199                 }
 200             }
 201             if (is_lws(*pp) || endofmime == NULL)
 202                 continue;
 203
 204             /*
 205              * We've found an encoded word, so we can drop
 206              * the '=' that was pending
 207              */
 208             equals_pending = 0;
 209
 210             /*
 211              * If we are between two encoded words separated only by
 212              * linear whitespace, then we ignore the whitespace.
 213              * We will roll back the buffer the number of whitespace
 214              * characters we've seen since last encoded word.
 215              */
 216             if (between_encodings) {
 217                 q -= whitespace;
 218                 dstlen += whitespace;
 219             }
 220
 221 #ifdef HAVE_ICONV
 222             /*
 223              * empty encoded text. This ensures that we don't
 224              * malloc 0 bytes but skip on to the end
 225              */
 226             if (endofmime == startofmime && use_iconv) {
 227                 use_iconv = 0;
 228                 iconv_close(cd);
 229             }
 230
 231             if (use_iconv) {
 232                 saveq = q;
 233                 savedstlen = dstlen;
 234                 q = convbuf = (char *) mh_xmalloc(endofmime - startofmime);
 235             }
 236 /* ADDCHR2 is for adding characters when q is or might be convbuf:
 237  * in this case on buffer-full we want to run iconv before returning.
 238  * I apologise for the dreadful name.
 239  */
 240 #define ADDCHR2(C) do { *q++ = (C); dstlen--; if (!dstlen) goto iconvbuffull; } while (0)
 241 #else
 242 #define ADDCHR2(C) ADDCHR(C)
 243 #endif
 244
 245             /* Now decode the text */
 246             if (quoted_printable) {
 247                 for (pp = startofmime; pp < endofmime; pp++) {
 248                     if (*pp == '=') {
 249                         c = decode_qp (pp[1], pp[2]);
 250                         if (c == -1)
 251                             continue;
 252                         if (c != 0)
 253                             *q++ = c;
 254                         pp += 2;
 255                     } else if (*pp == '_') {
 256                         ADDCHR2(' ');
 257                     } else {
 258                         ADDCHR2(*pp);
 259                     }
 260                 }
 261             } else {
 262                 /* base64 */
 263                 int c1, c2, c3, c4;
 264                 c1 = c2 = c3 = c4 = -1;
 265
 266                 pp = startofmime;
 267                 while (pp < endofmime) {
 268                     /* 6 + 2 bits */
 269                     while ((pp < endofmime) &&
 270                            ((c1 = char64(*pp)) == -1)) {
 271                         pp++;
 272                     }
 273                     if (pp < endofmime) {
 274                         pp++;
 275                     }
 276                     while ((pp < endofmime) &&
 277                            ((c2 = char64(*pp)) == -1)) {
 278                         pp++;
 279                     }
 280                     if (pp < endofmime && c1 != -1 && c2 != -1) {
 281                         ADDCHR2((c1 << 2) | (c2 >> 4));
 282                         pp++;
 283                     }
 284                     /* 4 + 4 bits */
 285                     while ((pp < endofmime) &&
 286                            ((c3 = char64(*pp)) == -1)) {
 287                         pp++;
 288                     }
 289                     if (pp < endofmime && c2 != -1 && c3 != -1) {
 290                         ADDCHR2(((c2 & 0xF) << 4) | (c3 >> 2));
 291                         pp++;
 292                     }
 293                     /* 2 + 6 bits */
 294                     while ((pp < endofmime) &&
 295                            ((c4 = char64(*pp)) == -1)) {
 296                         pp++;
 297                     }
 298                     if (pp < endofmime && c3 != -1 && c4 != -1) {
 299                         ADDCHR2(((c3 & 0x3) << 6) | (c4));
 300                         pp++;
 301                     }
 302                 }
 303             }
 304
 305 #ifdef HAVE_ICONV
 306         iconvbuffull:
 307             /* NB that the string at convbuf is not necessarily NUL terminated here:
 308              * q points to the first byte after the valid part.
 309              */
 310             /* Convert to native character set */
 311             if (use_iconv) {
 312                 size_t inbytes = q - convbuf;
 313                 ICONV_CONST char *start = convbuf;
 314
 315                 while (inbytes) {
 316                     if (iconv(cd, &start, &inbytes, &saveq, &savedstlen) ==
 317                             (size_t)-1) {
 318                         if (errno != EILSEQ) break;
 319                         /* character couldn't be converted. we output a `?'
 320                          * and try to carry on which won't work if
 321                          * either encoding was stateful */
 322                         iconv (cd, 0, 0, &saveq, &savedstlen);
 323                         if (!savedstlen)
 324                             break;
 325                         *saveq++ = '?';
 326                         savedstlen--;
 327                         if (!savedstlen)
 328                             break;
 329                         /* skip to next input character */
 330                         if (fromutf8) {
 331                             for (++start, --inbytes;
 332                                  start < q  &&  (*start & 192) == 128;
 333                                  ++start, --inbytes)
 334                                 continue;
 335                         } else
 336                             start++, inbytes--;
 337                         if (start >= q)
 338                             break;
 339                     }
 340                 }
 341                 q = saveq;
 342                 /* Stop now if (1) we hit the end of the buffer trying to do
 343                  * MIME decoding and have just iconv-converted a partial string
 344                  * or (2) our iconv-conversion hit the end of the buffer.
 345                  */
 346                 if (!dstlen || !savedstlen)
 347                     goto buffull;
 348                 dstlen = savedstlen;
 349                 free(convbuf);
 350             }
 351 #endif
 352
 353             /*
 354              * Now that we are done decoding this particular
 355              * encoded word, advance string to trailing '='.
 356              */
 357             p = endofmime + 1;
 358
 359             encoding_found = 1;         /* we found (at least 1) encoded word */
 360             between_encodings = 1;      /* we have just decoded something     */
 361             whitespace = 0;             /* re-initialize amount of whitespace */
 362         }
 363     }
 364 #ifdef HAVE_ICONV
 365     if (use_iconv) iconv_close(cd);
 366 #endif
 367
 368     /* If an equals was pending at end of string, add it now. */
 369     if (equals_pending)
 370         ADDCHR('=');
 371     *q = '\0';
 372
 373     return encoding_found;
 374
 375   buffull:
 376     /* q is currently just off the end of the buffer, so rewind to NUL terminate */
 377     q--;
 378     *q = '\0';
 379     return encoding_found;
 380 }