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

   1 /* fmt_rfc2047.c -- decode RFC-2047 header format
   2  *
   3  * This code is Copyright (c) 2002, by the authors of nmh.  See the
   4  * COPYRIGHT file in the root directory of the nmh distribution for
   5  * complete copyright information.
   6  */
   7
   8 #include "h/mh.h"
   9 #include "fmt_rfc2047.h"
  10 #include "check_charset.h"
  11 #include "h/utils.h"
  12 #ifdef HAVE_ICONV
  13 #  include <iconv.h>
  14 #endif
  15
  16 static const 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 const 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 whitespace = 0;         /* how much whitespace between encodings? */
  79 #ifdef HAVE_ICONV
  80     iconv_t cd = NULL;
  81     int fromutf8 = 0;
  82     char *saveq, *convbuf = NULL;
  83     size_t savedstlen;
  84 #endif
  85
  86     if (!str)
  87         return 0;
  88
  89     /*
  90      * Do a quick and dirty check for the '=' character.
  91      * This should quickly eliminate many cases.
  92      */
  93     if (!strchr (str, '='))
  94         return 0;
  95
  96 #ifdef HAVE_ICONV
  97     bool use_iconv = false; /* are we converting encoding with iconv? */
  98 #endif
  99     bool between_encodings = false;
 100     bool equals_pending = false;
 101     for (p = str, q = dst; *p; p++) {
 102
 103         /* reset iconv */
 104 #ifdef HAVE_ICONV
 105         if (use_iconv) {
 106             iconv_close(cd);
 107             use_iconv = false;
 108         }
 109 #endif
 110         /*
 111          * If we had an '=' character pending from
 112          * last iteration, then add it first.
 113          */
 114         if (equals_pending) {
 115             ADDCHR('=');
 116             equals_pending = false;
 117             between_encodings = false;  /* we have added non-whitespace text */
 118         }
 119
 120         if (*p != '=') {
 121             /* count linear whitespace while between encodings */
 122             if (between_encodings && is_lws(*p))
 123                 whitespace++;
 124             else
 125                 between_encodings = false;      /* we have added non-whitespace text */
 126             ADDCHR(*p);
 127             continue;
 128         }
 129
 130         equals_pending = true;
 131
 132         /* Check for initial =? */
 133         if (*p == '=' && p[1] == '?' && p[2]) {
 134             startofmime = p + 2;
 135
 136             /* Scan ahead for the next '?' character */
 137             for (pp = startofmime; *pp && *pp != '?'; pp++)
 138                 ;
 139
 140             if (!*pp)
 141                 continue;
 142
 143             /*
 144              * RFC 2231 specifies that language information can appear
 145              * in a charset specification like so:
 146              *
 147              * =?us-ascii*en?Q?Foo?=
 148              *
 149              * Right now we don't use language information, so ignore it.
 150              */
 151
 152             for (endofcharset = startofmime;
 153                         *endofcharset != '*' && endofcharset < pp;
 154                                                         endofcharset++)
 155                 ;
 156
 157             /* Check if character set can be handled natively */
 158             if (!check_charset(startofmime, endofcharset - startofmime)) {
 159 #ifdef HAVE_ICONV
 160                 /* .. it can't. We'll use iconv then. */
 161                 *endofcharset = '\0';
 162                 cd = iconv_open(get_charset(), startofmime);
 163                 fromutf8 = !strcasecmp(startofmime, "UTF-8");
 164                 *pp = '?';
 165                 if (cd == (iconv_t)-1) continue;
 166                 use_iconv = true;
 167 #else
 168                 continue;
 169 #endif
 170             }
 171
 172             startofmime = pp + 1;
 173
 174             /* Check for valid encoding type */
 175             if (*startofmime != 'B' && *startofmime != 'b' &&
 176                 *startofmime != 'Q' && *startofmime != 'q')
 177                 continue;
 178
 179             /* Is encoding quoted printable or base64? */
 180             quoted_printable = (*startofmime == 'Q' || *startofmime == 'q');
 181             startofmime++;
 182
 183             /* Check for next '?' character */
 184             if (*startofmime != '?')
 185                 continue;
 186             startofmime++;
 187
 188             /*
 189              * Scan ahead for the ending ?=
 190              *
 191              * While doing this, we will also check if encoded
 192              * word has any embedded linear whitespace.
 193              */
 194             endofmime = NULL;
 195             for (pp = startofmime; *pp && *(pp+1); pp++) {
 196                 if (is_lws(*pp))
 197                     break;
 198                 if (*pp == '?' && pp[1] == '=') {
 199                     endofmime = pp;
 200                     break;
 201                 }
 202             }
 203             if (is_lws(*pp) || endofmime == NULL)
 204                 continue;
 205
 206             /*
 207              * We've found an encoded word, so we can drop
 208              * the '=' that was pending
 209              */
 210             equals_pending = false;
 211
 212             /*
 213              * If we are between two encoded words separated only by
 214              * linear whitespace, then we ignore the whitespace.
 215              * We will roll back the buffer the number of whitespace
 216              * characters we've seen since last encoded word.
 217              */
 218             if (between_encodings) {
 219                 q -= whitespace;
 220                 dstlen += whitespace;
 221             }
 222
 223 #ifdef HAVE_ICONV
 224             /*
 225              * empty encoded text. This ensures that we don't
 226              * malloc 0 bytes but skip on to the end
 227              */
 228             if (endofmime == startofmime && use_iconv) {
 229                 use_iconv = false;
 230                 iconv_close(cd);
 231             }
 232
 233             if (use_iconv) {
 234                 saveq = q;
 235                 savedstlen = dstlen;
 236                 q = convbuf = mh_xmalloc(endofmime - startofmime);
 237             }
 238 /* ADDCHR2 is for adding characters when q is or might be convbuf:
 239  * in this case on buffer-full we want to run iconv before returning.
 240  * I apologise for the dreadful name.
 241  */
 242 #define ADDCHR2(C) do { *q++ = (C); dstlen--; if (!dstlen) goto iconvbuffull; } while (0)
 243 #else
 244 #define ADDCHR2(C) ADDCHR(C)
 245 #endif
 246
 247             /* Now decode the text */
 248             if (quoted_printable) {
 249                 for (pp = startofmime; pp < endofmime; pp++) {
 250                     if (*pp == '=') {
 251                         c = decode_qp (pp[1], pp[2]);
 252                         if (c == -1)
 253                             continue;
 254                         if (c != 0)
 255                             *q++ = c;
 256                         pp += 2;
 257                     } else if (*pp == '_') {
 258                         ADDCHR2(' ');
 259                     } else {
 260                         ADDCHR2(*pp);
 261                     }
 262                 }
 263             } else {
 264                 /* base64 */
 265                 int c1, c2, c3, c4;
 266                 c1 = c2 = c3 = c4 = -1;
 267
 268                 pp = startofmime;
 269                 while (pp < endofmime) {
 270                     /* 6 + 2 bits */
 271                     while ((pp < endofmime) &&
 272                            ((c1 = char64(*pp)) == -1)) {
 273                         pp++;
 274                     }
 275                     if (pp < endofmime) {
 276                         pp++;
 277                     }
 278                     while ((pp < endofmime) &&
 279                            ((c2 = char64(*pp)) == -1)) {
 280                         pp++;
 281                     }
 282                     if (pp < endofmime && c1 != -1 && c2 != -1) {
 283                         ADDCHR2((c1 << 2) | (c2 >> 4));
 284                         pp++;
 285                     }
 286                     /* 4 + 4 bits */
 287                     while ((pp < endofmime) &&
 288                            ((c3 = char64(*pp)) == -1)) {
 289                         pp++;
 290                     }
 291                     if (pp < endofmime && c2 != -1 && c3 != -1) {
 292                         ADDCHR2(((c2 & 0xF) << 4) | (c3 >> 2));
 293                         pp++;
 294                     }
 295                     /* 2 + 6 bits */
 296                     while ((pp < endofmime) &&
 297                            ((c4 = char64(*pp)) == -1)) {
 298                         pp++;
 299                     }
 300                     if (pp < endofmime && c3 != -1 && c4 != -1) {
 301                         ADDCHR2(((c3 & 0x3) << 6) | (c4));
 302                         pp++;
 303                     }
 304                 }
 305             }
 306
 307 #ifdef HAVE_ICONV
 308         iconvbuffull:
 309             /* NB that the string at convbuf is not necessarily NUL terminated here:
 310              * q points to the first byte after the valid part.
 311              */
 312             /* Convert to native character set */
 313             if (use_iconv) {
 314                 size_t inbytes = q - convbuf;
 315                 ICONV_CONST char *start = convbuf;
 316
 317                 while (inbytes) {
 318                     if (iconv(cd, &start, &inbytes, &saveq, &savedstlen) ==
 319                             (size_t)-1) {
 320                         if (errno != EILSEQ) break;
 321                         /* character couldn't be converted. we output a `?'
 322                          * and try to carry on which won't work if
 323                          * either encoding was stateful */
 324                         iconv (cd, 0, 0, &saveq, &savedstlen);
 325                         if (!savedstlen)
 326                             break;
 327                         *saveq++ = '?';
 328                         savedstlen--;
 329                         if (!savedstlen)
 330                             break;
 331                         /* skip to next input character */
 332                         if (fromutf8) {
 333                             for (++start, --inbytes;
 334                                  start < q  &&  (*start & 192) == 128;
 335                                  ++start, --inbytes)
 336                                 continue;
 337                         } else
 338                             start++, inbytes--;
 339                         if (start >= q)
 340                             break;
 341                     }
 342                 }
 343                 q = saveq;
 344                 /* Stop now if (1) we hit the end of the buffer trying to do
 345                  * MIME decoding and have just iconv-converted a partial string
 346                  * or (2) our iconv-conversion hit the end of the buffer.
 347                  */
 348                 if (!dstlen || !savedstlen)
 349                     goto buffull;
 350                 dstlen = savedstlen;
 351                 free(convbuf);
 352             }
 353 #endif
 354
 355             /*
 356              * Now that we are done decoding this particular
 357              * encoded word, advance string to trailing '='.
 358              */
 359             p = endofmime + 1;
 360
 361             encoding_found = 1;         /* we found (at least 1) encoded word */
 362             between_encodings = true;   /* we have just decoded something     */
 363             whitespace = 0;             /* re-initialize amount of whitespace */
 364         }
 365     }
 366 #ifdef HAVE_ICONV
 367     if (use_iconv) iconv_close(cd);
 368 #endif
 369
 370     /* If an equals was pending at end of string, add it now. */
 371     if (equals_pending)
 372         ADDCHR('=');
 373     *q = '\0';
 374
 375     return encoding_found;
 376
 377   buffull:
 378     /* q is currently just off the end of the buffer, so rewind to NUL terminate */
 379     q--;
 380     *q = '\0';
 381     return encoding_found;
 382 }